The Saxon Shore Forts

Many years ago on a school trip to Portsmouth harbour we were told by our history teacher that Portchester Castle, Portus Adurni, was evidence of the Anglo-Saxon invasion. Yes, built in the late 3rd century these Roman forts along the south-east corner of Britain were established to deal with Saxons that were settling on the shoreline; that’s why they are called the Saxon Shore forts he stressed. Later, in the early 5th century when the Romans left Britain, he added, the Saxons were left unchecked and run amok across the country, as recorded by the 6th century historian Gildas, to create the Anglo-Saxon kingdoms, as recorded by Bede.

I have always found the concept of a mass Anglo-Saxon invasion a rather extreme theory; yes there were Germanic settlers in Britain but I’ve always considered it more of a cultural change rather than large scale population displacement. However, I believe it is still taught as such in school today. You could interpret my old history teacher as being correct really in meaning that when the Saxons first arrived in Britain it was the beginning of the event. As far he was concerned their arrival was associated with the Saxon Shore, hence the name.

The Roman Shore Forts of Roman Britain

There are eleven sites around the south-eastern coast of Britain, from Portchester in Portsmouth Harbour to Brancaster on the Wash, that are designated as the Roman Saxon Shore forts by modern historians. Nowhere else on the British coast do we find such a cluster of Roman forts, constructed on harbours as at Portchester or on river estuaries as if to guard the watercourses into the Roman province.

The term ‘Saxon Shore’ first appears in the Notitia Dignitatum, a Latin document dated AD 395, revised around AD 420, and lists just nine forts on the British coast under the command of the Count of The Saxon Shore. The majority of the forts were constructed in the late 3rd century, some earlier in first half of that century. Several of the fortifications on the east coast have suffered badly from coastal erosion and stone robbing leaving little of the original stonework remaining; in the case of Brancaster, in Norfolk, the presence of the fort can now only be detected by landscape features where the walls once stood and at Walton Castle, Suffolk, the site is now completely submerged under the North Sea.                                                                                                                                                                                                                                                                                                                             

The Saxon Shore forts today:

Brancaster (Branodunum)
This Roman fort was constructed to guard the southern point of The Wash. Today the site is just a field, the east ditch can be seen as a marked depression, the fort is often revealed by parched summer grass using aerial photograph over its buried ramparts, revealing the positions of gates and traces of internal buildings. The fort was rectangular with rounded corners, the classic ‘playing card’ shape typical of early Roman forts in Britain, and had walls 2.7 metres thick. Robbed facing stones from the Brancaster Shore Fort were re-used (together with other later material) in the south chancel of the church of St Mary the Virgin at Brancaster. It is likely there was another fortification on the northern side of The Wash near Skegness.

Burgh Castle (Gariannonum?)
Situated on a low hill above the river Waveney, the remains of Burgh Castle are now some distance inland. In Roman times the fort would have been on the shoreline of a large sheltered inlet guarding the “Great Estuary”, the site of Great Yarmouth was then sea.

The walls, 3.5 metres thick and in places still 4.5 metres high, enclosed a 2.4 hectare site. The walls with six projecting towers or bastions, possibly added later to the original design, remain on three sides. A fallen bastion can be seen on the south wall, and on the east wall a vertical break between the bastion and the adjacent section of perimeter wall shows they were built as separate components, perhaps a later addition, the bastion is only bonded to the wall at the very uppermost courses.

Burgh Castle

Much of the stonework from Burgh castle has been re-used in later local buildings. The round bell tower of the church of St Peter and St Paul, Burgh Castle, displays Roman masonry such as flint, brick and tile that has been robbed from the nearby Shore fort.

At nearby Reedham the church of St John the Baptist displays large quantities of Roman building material, principally brick and tile. This, and other finds in the locality, provide evidence that a substantial Roman site once stood here. Situated on the shores of the Great Estuary it was probably a lighthouse or watchtower serving the local Shore forts.

Bradwell (Othona?)
Situated on the Dengie Marshes, Othona was located to control the estuaries of the Blackwater and Colne which provided access to the Roman city of Camulodonum (Colchester). Little has survived of the Roman fort into modern times: sections of three walls, one with a bastion, survived until the 18th century when they were demolished. The eastern wall has never been located. 

The 7th-century chapel of St Peter-on-the-Wall at Bradwell-on-Sea stands on the site of the Roman fort’s west gate. Evidently much of the stonework of the fort was robbed in the Saxon period; the north wall of the chapel shows evidence of re-used Roman masonry. Limestone blocks of 'Clipsham Pink Stone' used in the Shore fort and recycled at St Peter's chapel were imported over 300km (190 miles) from quarries to the north-west of the Roman town of Durobrivae (Water Newton, Cambridgeshire).

Reculver (Regulbium)
Guarding the northern end of the now lost Wantsum Channel, a seaway that separated the Isle of Thanet from mainland Kent, stands the Roman Shore fort of Reculver. About half of the site has been lost to the sea yet substantial parts of the walls and the impressive south gate still survive. The fort had an almost square plan with rounded corners.

The Twin Towers of Reculver

The east wall of the fort at Reculver displays clear evidence of later stone robbing, the entire facing stones have been removed leaving just the core rubble constructed from beach flints exposed. At the site of the south gate are three large blocks which once formed the base of a monumental arch. The medieval church of St Mary at Reculver, a landmark to mariners known as the “Twin Towers Reculver” originally sited at the centre of the Shore fort, now stands within a few metres of the coast owing to coastal erosion. The Shore fort was originally 1km from the coast.

Richborough (Rutupiae)
One of the most impressive remains of a Shore fort can be seen at Richborough where the wall circuit survives on three sides and massive ditches crossed by the line of Watling Street provides a glimpse into the days of Roman Britain. The west wall remains are 7m high, but its opposite, the east wall, has fallen into the river Stour.

Richborough

The fort at Richborough was the second of a pair guarding the Wantsum Channel as Reculver stood at the northern end Richborough at the southern end. Richborough has been identified by historians as the place where the Roman invasion force landed in AD 43. The fort developed from an early military camp into a port before later fortifications were added during the 3rd century. 

The north wall was built in two parts, possibly by two different units, as the break in the sections of construction can clearly be seen. Much re-used stone from earlier Roman buildings was employed in this portion of the defences when the site was refortified in the late 3rd century. The monumental blocks of the foundations of the west gate were recycled from an earlier structure on the site, probably the first century monumental arch. The so-called ‘Great Monument’, the only Roman building in Britain to have been covered externally in marble, was replaced by a watchtower, surrounded by triple ditches during this later rebuild.

Dover (Dubris)
The Romans constructed a fort at the mouth of the river Dour, then a wider estuary, for the classis Britannica, a fleet based in the English Channel, its inception probably for the invasion of Britain in AD 43. Archaeological evidence for the presence of the fleet relies mainly on tiles stamped ‘CLBR’ found on the east Kent coast and London. The fleet is not heard of after the mid-3rd century, probably disbanded and located at smaller coastal forts on either side of the channel. The classis Britannica fort at Dover was abandoned around AD 225 and a new fort was built in the late 3rd century.

The Shore fort at Dover was built over the north-east corner of the foundations of the classis Britannica port, which in turn had been built over an earlier unfinished fort. Exposed sections of the Shore fort include part of the south wall with tower. The lighthouse, known as the ‘pharos’, now stands within the medieval castle to the west of the church of St Mary in Castro. The first three quarters of its height is Roman masonry, the top quarter being a later repair, now 19m high but thought to have originally been 24m tall. The pharos is one of a pair, the other stood on the western side of the Dour estuary, known as the Western Heights, now all but lost among the Napoleonic fortification.

Lympne (Lemanis)
The Shore fort was positioned on the slopes of an ancient degraded sea cliff, originally overlooking a wide natural harbour, now Romney Marsh in Kent. The fort may have occupied the site of an earlier fort of the classis Britannica, Portus Lemanis, as mentioned in the Antonine Itinerary

Today the tumbled ruins lie scattered on the hillside owing to slumping of the clay slope on which is was built. It appears the fort was built in the shape of an irregular pentagon but landslips have broken and moved huge chunks of the walls down the slope so that its original form is no longer obvious. Several sections of wall survive, up to 6 metres tall, but mostly leaning and displaced. Eight semicircular bastions have been detected, probably of a total fourteen, the remains of one on the north wall provides a rare instance where the facing stones and brick still remains in situ. The site appears to have been abandoned around AD 350.

Pevensey (Anderida)
Pevensey is the largest and one of the most impressive of the Shore forts. The walls form an irregular oval shape, enclosing 4 hectares with a later Norman castle added to the south-east corner. The fort is now 1km from the sea but when it was built it sat at the end of raised tongue of land surrounded by coastal marsh.

Pevensey Castle

Its projecting D-shaped towers were a feature of this style of Roman military architecture that appeared in the late 3rd century. The 3.6m thick walls were used later to define the outer bailey of the Norman castle. The lower portion of the bastion on the north wall at Pevensey is clearly of Roman construction, however, the part of this tower that projects above the perimeter wall is almost certainly a post-Roman addition.

Portchester (Portus Adurni)
Perhaps the finest example of all the Shore forts is at Portchester, at the head of Portsmouth Harbour, Hampshire, with, it is said, the best-preserved Roman fort walls north of the Alps. Unlike most of the Saxon Shore forts in Britain the coastal setting at Portchester is little changed from the time of its construction in the late 3rd century with the sea licking the fort on the east side as it did in Roman times.

Portchester

The 3.1m thick and 6m circuit walls enclosing a regular square plan is complete except where they were altered in the 12th century by the addition of the Norman keep and medieval gatehouse in the north-western corner. A church now sits in the south-east sector. Fourteen of the original twenty projecting D-shaped bastions of the Roman fort survive. The bastions are hollow unlike those at Lympne and Pevensey. The walls were laid on a timber and flint raft, the timbers cross-braced as at Pevensey. 

Occupation continued beyond the end of the Roman period, with three main phases during the late 3rd and early 4th centuries, with noticeable breaks between these stages. Whether a civilian population took over the fort or a militia formed a continuing military presence is uncertain.

In addition to the nine forts listed in the Notitia Dignitatum (above), there are a further two Roman forts on the east coast of Britain which must have been part of these coastal defences at some time.

Caister-on-Sea
A second fort guarded the Great Estuary, along side Burgh Castle but omitted from the Notitia Dignitatum, although its defences are of similar construction to other Saxon Shore forts. This raises the question of whether Burgh Castle or Caister was the ‘Gariannonum’ listed in the Notitia Dignitatum? The estuary has greatly silted up since Roman times and much of the fort is covered by modern housing, but a short section of the south wall and south gate can still be seen. 

The forts at Burgh Castle and Caister may have formed a pair, or Burgh may have superseded Caister. Individually or as a pair, they certainly guarded a large estuary of several rivers, to prevent undesirables from penetrating inland to the heart of Roman Britain. 

Walton Castle
The Roman fort at Walton Castle, near Felixstowe, Suffolk, is now totally lost to the sea through coastal erosion in the 18th century, but we know of its existence from antiquarian accounts. The wall of the fort at Walton Castle shown on an 18th-century copy of a drawing originally made in 1623 displays a clear similarity to Burgh Castle with circular bastions and rounded corners which suggests a construction date similar to Burgh Castle and Bradwell. It fills a strategic gap in the coastal defences on the east coast, in Roman times commanding the entrance to the estuary of the Stour and Orwell. The fort at Walton Castle has sometimes been identified with Portus Adurni (Portchester) or Othona (Bradwell).

These eleven forts on the east and south-east coasts of Britain have much in common and certainly formed a defensive chain in it’s entirety. They were all located on low-lying ground, close to sheltered anchorages at major estuaries and inlets. Some seem to have functioned as pairs, such as Reculver and Richborough at opposite ends of an important waterway, and Caister and Burgh Castle on opposite sides of a large estuary. 

But guarding the coast from what? Debate remains over the correct interpretation of the term ‘Saxon Shore’ – did it describe the shore settled by Saxons, or the shore attacked by Saxons?

References:
Leonard Cottrell, The Roman Forts of the Saxon Shore, HMSO, 1964.
Nic Fields, Rome’s Saxon Shore, Osprey, 2006.
Valerie Maxfield (ed), The Saxon Shore: A Handbook, University of Exeter, 1989.
Andrew Pearson, The Roman Shore Forts, The History Press, 2010.
Stephen Johnson, The Roman Forts of the Saxon Shore, BCA, 1979.
D E Johnston (ed), The Saxon Shore, CBA Research Report No: 18, 1977.
Donald White, Litus Saxonicum, University of Wisconsin, 1961.

* * *

Petuaria Part III: A Lost Saxon Shore Fort?

The Saxon Shore
The first excavators of Brough on Humber, Philip Corder with Rev. Thomas Romans, who explored the site in the 1930s saw the site as a military or naval fortified base controlling the River Humber and access to York. Corder speculated that it may have been part of the chain of coastal defences known as the Saxon Shore forts. John Wacher carried out excavations at Brough between 1958-61 and found no evidence for a significant civilian settlement and agreed with Corder that it was a military site, which he also considered included a naval base which may well have been part of the Saxon Shore defences.

Wacher’s chronology for the settlement at Brough on Humber commences in the 1st century as an auxiliary fort that was abandoned around AD 80. However, he considered that during this period it was likely that a military supply depot was maintained there. In the later 3rd century the fort was reoccupied perhaps as an outlier of the Saxon Shore defensive chain, and then finally abandoned in the last quarter of the 4th century; the oldest coinage dating to the reign of the usurper Magnus Maximus (383-388).

Situated on a natural inlet of the river Humber, the military base at Brough would have fulfilled two functions: it controlled the approaches to the estuary linking to several major rivers that provided access to the heartland of Britannia; the same function was performed by the Saxon Shore forts at Brancaster on the Wash and Reculver on the Thames estuary.

Secondly, it functioned as a port serving the Roman forts at York and Lincoln and surrounding areas such as the fort at Malton. Brough also served as the northern terminal of the Humber crossing linking York and Lincoln by the Roman Road known as Ermine Street. However, it must be conceded that, to date, neither side of the crossing has produced evidence of a harbour. Wacher attributes this to the Late Roman marine transgression which would have submerged the existing structures. 

During the 4th-5th centuries AD this significant geological phenomenon resulted in several Romano-British sites on the Lincolnshire marshes being completely covered over by several metres of marine alluvium. For example, The Romano-British sites at Scupholme and Ingoldmells were found beneath more than three metres of alluvium.

The Roman Shore Forts
The so-called Saxon Shore is a series of Roman defences along the south and east coasts of Britain. It is mentioned in just one official document, the Notitia Dignitatum (Register of Dignitaries), thus the sole source for its existence.

The (unknown) compiler of the Notitia Dignitatum listed the official titles of all civil and military posts within the Roman Imperial administration around AD 395. The document consists of two sections, the Eastern Empire, and the Western Empire, revised and updated c.420. The western section lists the office of Comes litoris Saxonici per Britannias, the Count of the Saxon Shore for Britain, and is the only historical source for this otherwise unknown military command.

The Notitia records nine forts under the command of the Comes litoris Saxonici per Britannias forming a chain of coastal defences stretching from the Wash at the northern end and along the east and south coasts to the Solent; from Branodunum (Brancaster) in Norfolk to Portus Adurni (Portchester) in Hampshire and includes Othona (Bradwell), Dubris (Dover), Lemannis (Lympne), Gariannum (Burgh Castle), Regulbium (Reculver), Rutupiae (Richborough), and Anderida (Pevensey). To this list the Roman forts at Caister-on-sea (Norfolk) and Walton Castle (Suffolk) have been added by modern historians to make a total of eleven Saxon Shore forts. Could Petuaria (Brough on Humber) have been the twelfth?  

The forts are typically located on inlets or estuaries, all points of potential incursion into the Roman province. Large parts of the forts at Richborough, Reculver, Burgh Castle, and Bradwell-on-Sea have been lost to coastal erosion with Walton Castle now entirely submerged.

Walton Castle, near Felixstowe, is a unique case, now totally submerged under the sea owing to coastal erosion, however, old descriptions by antiquarians seem to confirm it was part of the chain of defences. The location and features make it almost certain that Walton Castle was a Roman Shore fort, being quadrangular in plan, with flint and mortar walls bonded with brick with round bastions. Strategically positioned to control access to the estuary of the Stour and Orwell. It is claimed that there are traces of a signal station system between Walton and Burgh Castles.

However, owing to a lack of positive evidence Walton Castle has not been universally accepted as a Shore Fort. The counter argument is that Walton was Portus Adurni rather than Portchester. Yet it is perfectly possible that by the time the Notitia Dignitatum was compiled the fort at Walton had been abandoned under reorganisations under Theodosius. 

Sentinels of the Waterways
Caister and Burgh Castle stood either side of the ‘Great Estuary’ controlling access in to the Broadlands rivers, (Yare, Bure and Waveney) behind Yarmouth, which was then a sandbank. Richborough and Reculver stood sentinel over the two ends of the Wantsum channel, a strait separating the Isle of Thanet from the north-eastern Kent, a major shipping route in Roman times. 

If, as seems likely, some of the Roman Shore forts were sited in pairs at main points of possible incursion to rivers or estuaries, as above, then we could expect to find another Roman port on the northern side of the Wash (Metaris Aestuarium), opposite Brancaster which guarded the southern side, perhaps near Skegness as the Roman road from Lincoln (Lindum) runs there. The 16th Century antiquarian John Leland makes reference to a ‘walled town’ and a ‘castle’ off the coast of Skegness. But any structures there were claimed by the sea long ago. 

‘Old Skegness’, as the local people refer to the Old Roman site, now lies about half a-mile off the coast after being swallowed by the sea in the 1500s following storms and floods. In addition to the clues mentioned by Leland of a lost Roman town or fort, sailors were still reporting encountering an old church steeple out there in the 17th century.

We could also expect to find a Shore fort guarding the Humber, such at Brough on Humber, and access into to York.

Dating the Forts
The Roman shore forts can be divided into two groups by age and design. The first group shows similarities in size and internal layout with most of the Roman forts in Britain.

The early group of forts comprises the two forts on the Norfolk coast, Brancaster and Caister, and Reculver in Kent. These forts are considered to have been constructed during the early 3rd century AD owing to their conventional architecture, the classic ‘playing card shape’, square sides with rounded corners, common in early Roman military architecture. Evidence from excavations supports the early date. In the days of early Roman Britain the forts were rarely attacked with the legions preferring to meet the enemy in the open field where their superior battle tactics and discipline proved dominant.

The second group, Burgh Castle, Bradwell, Reculver, Dover, Lympne, Pevensey, Portchester, were constructed during the late 3rd century. The principal differences between these later forts and the early group is the incorporation of new aspects of Roman fort architecture that originated in the east of the Empire to resist siege warfare against the Sassanids in the early 3rd century.

The addition of external bastions (turrets) provided the fort (or town wall) with additional defensive capability against an enemy directly attacking the face of the fort walls with battering rams or siege engines. With bastions external to the fort walls the attackers could now be subjected to artillery crossfire along the long axis of the wall.

Burgh Castle (Gariannonum) east wall bastion
(Creative Commons)

These novel features included the construction of thick walls (up to 3.5 m at Pevensey), bonded with brick courses, over 7m tall, whereas the earlier forts had thinner walls supported by earthen inner ramparts. The later forts all displayed variability of plan, from square as at Portchester, oval at Pevensey, to pentagonal at Lympne. Yet the signature features of these later forts is the external bastions on the fort walls. 

The majority of the external bastions incorporated into the Saxon Shore forts were circular, or semi-circular, solid masonry ‘drums’ as at Burgh Castle, Bradwell, Pevensey and Lympne. However, Portchester was constructed with hollow bastions which may indicate a shortage of building materials, it was not finished, or, being at the western end of the chain of Shore forts, was only considered likely to be suffer minor attacks.

The river wall and bastions completed the defensive circuit of Roman London, as with the later Saxon Shore forts and other fortifications in the north and west constructed at this time, all appear to have been built in a hurry.

Wacher considered that there was a close comparison between Brough on Humber (Petuaria?) and the Saxon Shore fort at Burgh Castle (Gariannum) in Norfolk. Wacher suggests that although the original plan at Burgh may have been pre-Carausian, that is before AD 286, the additional constructions including thicker walls, external bastions and towered gateways, were in place prior to the Theodosian restructuring following the Barbarian Conspiracy of AD 367. It is therefore likely that both forts were updated by Carausius, the usurper emperor (r.286-293) in the late 3rd century. He argues that the pattern of development at Brough is typical of the chronology of military fortification rather than civil settlement. 

Carausius
Marcus Aurelius Mausaeus Carausius, a Menapian of humble birth, was an officer in the Roman military service, a helmsman by profession. He had received recognition for fighting against the Bagaudae, groups of rebel peasants in the western parts of the later Roman Empire, who arose during the Crisis of the 3rd Century and persisted until the very end of the Western Empire.

Around 285-6 Carausius was appointed by Maximian, co-ruler with the emperor Diocletian, to deal with Franks and Saxon pirates who had ‘infested’ the Channel and were plundering the coast of northern Gaul. Carausius was given command of a fleet, or given authority to assemble his own, based at Gesoriacum or Bononia (Boulogne-sur-Mer) to counter the threat.

Carausius was remarkably successful but was accused of letting the pirates complete their raids, then attacking them to take their spoils for himself. Infuriated by this news Maximian ordered Carausius’s death. In 286, in response Carausius declared himself Emperor with the title of Augustus, over Britain and northern Gaul, issuing coins showing himself as co-emperor alongside Maximian and Diocletian.

Coin portraying Carausius alongside Diocletian and Maximian as joint emperors
‘CARAUSIUS ET FRATRES SUI’  (Carausius and his brothers!)

Several attempts by Maximian to remove Carausius failed, and in 290 he was finally acknowledged as ruler of Britain. On 1st March 293, Constantius was made caesar by Maximian. The two men, together with Diocletian and his caesar, Galerius, formed the 'tetrarchy'. Constantius was assigned to rule Gaul with his first instruction to remove the usurper Carausius. Shortly later he captured Carausius’s mainland base at Gesoriacum. Carausius pulled back to Britain but later that year was assassinated by his finance minister, Allectus. Three years later in 296 Allectus was killed in a battle with Asclepiodotus, the praetorian prefect, who landed on the south coast to recover Britain.

Brough on Humber: A Saxon Shore fort?
Although Wacher saw Brough on Humber as a naval base there has been no archaeological link established with the classis Britannica, the Roman fleet, probably constituted for the Claudian invasion in AD 43, but no longer documented after the mid-3rd century. Tiles marked ‘CLBR’ associated with the Roman fleet have never been found north of London.

However, if as seems likely that Carausius constructed the later Shore forts, and enhanced the defences of the earlier forts, he must have had an operational fleet. It is probably more correct to say that Carausius planned the Shore forts and they were completed by his successor Allectus. 

Other bases and Roman London were re-fortified at this time while new defensive shore forts were being established at Porchester while others at Brancaster and Lympne were being restored or repaired. At nearby Malton (Derventio), 23 coins of the Carausian period were found at the North-East gate of the fort during the 1927-30 excavations indicating it was probably rebuilt by Carausius. There was also significant reconstructions at Brough on Humber at the same time. 

Relatively recent archaeological excavations by the Petuaria Revisted Project unearthed evidence of another bastion in 2023, and coins from the Carausian period, which confirms that the fortifications were enhanced in the late-3rd century. But was Brough on Humber a Saxon Shore fort? To answer that question it is necessary to establish what exactly was meant by the term 'Saxon Shore'.

There is little agreement with regard the meaning of the term ‘Saxon Shore’ and the function of the forts: were they constructed at strategic points to guard against attack from Saxon pirates; or was the Saxon Shore so-named because it was settled by Saxons or the forts manned by Germanic mercenaries. Carausius is known to have used Germanic troops to operate his fleet, perhaps he settled some in his chain of Shore forts and his successor Allectus had an army of Franks defending London when Constantius attacked in 296.

However, there is little evidence of coastal raiding by Germanic pirates on the south-east coast or Picts down the east coast before the 4th century as evidenced by the five fortlets (signal stations) constructed on the North Yorkshire coast at Huntcliff, Goldsborough, Ravenscar, Scarborough and Filey during the Theodosian restructuring. Furthermore, Germanic settlers were not seen on the Humber until the early 5th century.

A more recent argument suggests these forts may have served as military ports rather then as a defensive system. It’s location on the bank of a major river estuary makes it likely Brough on Humber (Petuaria) was a naval port. But there is little evidence for anything beyond a fortified supply depot importing and exporting products to and from Britannia Inferior (Northern England and the Midlands).

The location of Brough on Humber on the East Yorkshire coast sets the Roman base too far north to be considered a true ‘Saxon Shore’ fort which were established to defend the south-eastern corner of Britain, the Wash to the Solent.  Its absence from the Notitia Dignitatum does not support the case for Brough either, yet as we have seen above, only nine forts are listed in that document when eleven sites have been identified.  

Yet we know Brough on Humber was refortified during the Carausian revolt of the Late 3rd century.  Therefore it was likely included in Carauius’s coastal defence system. Carausius would not have used the term ‘Saxon Shore’ as this term did not appear until the compilation of the Notitia Dignitatum in the Late 4th century. By this time, some of the forts along the shore may have been abandoned during the defensive reorganisations under Theodosius following the Barbarian Conspiracy of AD 367, and subsequently not included in the document. According to Wacher’s phasing the military base at Brough on Humber was finally abandoned in the last quarter of the 4th century.

* * *

Petuaria Part II: Usurpers and Red Herrings

 One of the few Roman urban sites known in East Yorkshire, Brough-on-Humber is situated close to the north bank of the Humber estuary, 11 miles from Kingston-upon-Hull and less than 30 miles from York. The Humber Estuary is the largest coastal plain estuary on the east coast of Britain, providing access to several major rivers including the Don, Aire, Ouse, Trent, Hull and Ancholme.

During the Roman era a fortfied site at Brough occupied this strategic position, situated on the shore of the Humber estuary the site was perfectly situated to guard access by river, which was then navigable, into the heartland of Britannia. Northwards the land route from Brough went to the legionary fortress at York (Eboracum) and southwards across the ferry route to Lincoln (Lindum). Undoubtedly at some stage it would have functioned as a supply depot for imports for the military bases of the north. Exports would also have gone through Brough to the wider Roman world, for example lead ingots mined from Derbyshire have been found at the site.

Brough lies on the natural approach line from Lincoln from where Petillius Cerialis launched his campaign against the Brigantes in AD 71-72. Cerialis was an experienced general who had campaigned against the Britons ten years earlier as commander of Legio IX Hispania when he was engaged in the Boudican revolt. The Roman historian Tacitus records that the entire infantry force was massacred during that event, while Cerialis escaped with his cavalry. Ten years on, Cerialis was back in Britain with the newly raised Legio II Adiutrix and headed north to Lincoln to join Legio IX Hispania. II Adiutrix remained at Lincoln, while Cerialis advanced with IX Hispania to York, crossing the Humber at Brough.

The lack of known Roman forts in east Yorkshire suggests there was no great hostility from the Parisi toward the advancing Romans. The Roman fort, usually identified as Petruaria, appears to have been the first significant construction at Brough-on-Humber around AD 70, although slight traces of previous Late iron Age occupation have been detected.

Archaeologist John Wacher asserts that the early fort at Brough was evacuated in the late 70’s, possibly during Agricola’s reorganisations. After being abandoned the first Roman fort lay vacant for the next 40 years or so but the supply depot appears to have been retained. Early in the Hadrianic period, around AD 125, the fort was re-occupied for a brief period, possibly in anticipation of the next phase of construction when a new extended defensive circuit erected.1

Wacher sees these new fortifications as providing a base for the Classis Britannica (the Roman British fleet). He concedes that it is possible that at this time the civitas was reconstituted under civilian government not far from this base; it being extremely unlikely that both would have operated out of the same site. During the later 3rd and early 4th century the site was refortified with the rampart being replaced by a stone wall with external bastions.2

It is often assumed that the post-fort occupation on the site was the civitas capital of the Parisi, but there is far from general agreement on this which is based on interpretation of an inscription (RIB 707) found in 1937 and the ancient geographer Ptolemy who refers to Petuaria as the ‘polis’ of the Parisi.  [See; Petuaria: Civitas or Vicus?] 

Wacher found no evidence for a significant civilian settlement during excavations at Brough between 1958-61 noting many features that were not typical of a civitas capital, such as a lack of urban sophistication,  e.g. absence of public buildings or a planned street system, and argues that it is unusual for a town to be enclosed so early, and its turf and timber ramparts have more in common with contemporary military works rather than later town defences. Wacher sees the extensive Iron Age settlement at North Ferribly extending westwards as far as Brough with no other major settlement east, west or north of the military base, as the probable site of the vicus.3

RIB 707

Wacher argues that excavations have shown an almost exclusively military chronology for the sequence of defensive circuits and that the inscription (RIB 707) relating to the status of a British capital is completely unrelated to any structural remains.4

The inscription indicates that by AD 144 the vicus Petruariensis had a junior magistrate of a class only usually found in towns ranked as civitas capitals, and implies the existence of a theatre which would have been provided only after the provision of more functional public buildings such as a forum, basilica and a bath-house;5 despite promising ground scans the remains of these key buildings have, to date, not been found.

The identification of Petuaria with Brough-on Humber is almost entirely dependent on a Roman inscription found by the first excavators of the site Philip Corder with Rev. Thomas Romans during the 1930s. In 1937 their foreman Bertie Gott uncovered the stone inscription under the Burrs Playing Field.6

2018 GPR survey of Burrs Playing Field, Brough (Malton Museum)

In 2018 a ground-penetrating radar (GPR) survey was carried out at the Burrs Playing Field, the same field that had produced the inscription, and showed a D-shaped feature (marked 'G' above). 

Historic England granted consent to open a trench over the D-shaped feature in 2020, but the excavators found only a hearth that contained a burnt coin dating to c.AD 330 and rubble, evidence that any structures on this feature had been systematically demolished and its building materials used for masonry elsewhere.

However, this layer had later-3rd- and 4th-century coins and pottery associated with it, making it unlikely that it relates to a mid-2nd-century theatre. As they went below this layer they found further stony layers interpreted as successive surfaces of the courtyard. Yet it must be conceded that following two seasons of excavations over the D-shaped feature, the theatre remains elusive.7

Spolia
The ancient and widespread practice of ‘Spoliation’ (from the Latin for 'spoils') was common in late antiquity (250 - 750 AD) when entire redundant structures were known to have been comprehensively demolished and the materials used in the construction of new buildings. 

The practice survived into the Late Middle Ages, evidence can be seen at Lanercost Priory where Roman masonry from Hadrian’s Wall was robbed and used in the nearby construction of the abbey in the 12th century. One such stone built into the abbey wall displays a Roman inscription, recording the presence of Legio VI Victrix on the Wall.

An example of this practice from the Roman period in Britain can be seen in two sandstones that were uncovered at Bainbridge in Wensleydale in 1960 which record the rebuilding of a fort on Brough Hill in the 3rd Century. One of the stones bears an inscription which is one of the longest and largest ever to be found in ancient Britain, designated RIB 3215.

The inscription stones were initially built into the wall of the barracks adjacent to the east gate of the fort at Brough-by-Bainbridge, North Yorkshire, (probably the Roman fort of Virosidum?) and record work carried out by Cohors VI Nerviorum. The Nervian cohort completed extensive construction works at Brough-by-Bainbridge c. AD 205–7 including the building of four barrack blocks, the relocation of the east gate and its enclosure described in an inscription as the bracchium caementicium.

RIB 3215

Clearly the inscription RIB 3215 had become redundant when the principia (headquarters) was partly demolished to provide space for a timber building probably to accommodate the commanding officer.8

The inscription stones, RIB 3215 and RIB 3216, were later reused, upside down, in the 4th century as stone packing in the foundations for a Roman road and covered over for the next 2,000 years. The excavator Brian Hartley states that the larger stone had been re-used at least once before, since the inscribed face was partly covered by mortar. 

The history of the fort at Brough-by-Bainbridge seems somewhat similar to Brough-on-Humber. As a place name “Brough” is thought to have originated from the Old English word “burh” meaning "fortress". Occupation at Brough-by-Bainbridge began in the governorship of Agricola, it may have had an earlier phase, but the visible remains date to AD 90–105. Firstly, an earth and timber rampart was constructed which around AD 190 was rebuilt in stone. Sometime between AD 120 and AD 160 the site had been abandoned. Then we come to the time of inscription (RIB 3215), the early 3rd century, when parts of the fort were rebuilt in by the Cohors VI Nerviorum. Reconstruction took place across the whole site in the late 4th century, then pottery evidence indicates a late abandonment.

A Naval Base?
Wacher suggests the inscription at Brough-on-Humber is possibly a red herring that has misled archaeologists for decades. Notably, it was found incorporated into the masonry of an early-4th century military building, its original location unknown. The robbing of the inscription stone suggests the theatre, if it ever existed, and other associated buildings such as the forum and basilica, were redundant by the late 3rd century, purposefully demolished and the masonry used in re-fortifications of Wacher’s naval base. 

Wacher may well be correct in arguing that Brough was a failed civitas. He concludes that the town never seems to have been fortified at a time when urban defences were generally being constructed as the site no longer warranted it, and views Brough as another failed town similar to Chelmsford. He adds that when this happened disused buildings were typically robbed of their masonry to use in construction of the new defences.

Wacher considers the naval base at Brough was abandoned around AD 360 when some inhabitants of the town moved inside the fortifications and occupied the south-west corner until at least the end of the 4th century.9

The Petuaria Revisited Project team uncovered evidence of continued Roman activity in Brough on Humber into the late 4th and probably 5th centuries AD, beyond that previously thought, and seemingly at variance with earlier determinations about the end of occupation at Brough as given by Wacher, who suggested that it had become redundant by the 4th century. Yet, it must be noted that the bulk of the coins found at Brough date to the mid-4th century, in line with Wacher’s suggestion.

However, there was a significant cluster of coins from the time of Carausius and Allectus, the usurper emperors who reigned sequentially for the period AD 286-296. Carausius had declared himself as emperor in AD 286 and set Britain as his base. It seems likely that he would take advantage of British naval bases following his previous command of the Roman fleet in the English Channel. Wacher agreed with the earlier excavators Philip Corder with Rev. Thomas Romans who saw the site as a military or naval fortified site possibly re-constructed under Carausius. However coins do not provide evidence of his presence at Brough or that it was indeed a naval base yet it may hint at a connection.

An external tower, or bastion, found at Brough in 2023, resembles similar examples in Roman London, as well as those of the Saxon Shore forts, which were apparently built to protect the south and east coasts of Roman Britain from seaborne raiders.

New fortifications of this nature were typical of the 3rd century Roman Britain and tends to support the theory that Brough was a form of naval base on the shore of the River Humber, controlling access to York, and possibly part of the chain of coastal defences known as the Saxon Shore forts.10

 
Notes & References
1. J.S. Wacher, Excavations at Brough-on-Humber 1958- 1961, Society of Antiquaries of London No. XXV (1969).
2. Ibid.
3. Ibid.
4. Ibid.
5. Adam Rogers, Roman Towns in Late Roman Britain, Cambridge University Press, 2011.
6. Wacher, op.cit.
7. The Past: Petuaria revisited: searching for Brough-on-Humber’s lost Roman theatre
8. Roman Inscriptions of Britain: RIB 3215. Imperial dedication to Septimius Severus, Caracalla, and Geta. 
9. John Wacher, The Towns of Roman Britain, BCA edition, 1976, p.393-97.
10. The Past: Petuaria revisited: searching for Brough-on-Humber’s lost Roman theatre

 * * *

Have they Found Merlin’s Grave?

MERLIN came late to the Arthurian story introduced by Geoffrey of Monmouth in the 12th century yet there are traces of a prophet named Mryrddin from at least the 10th century in Wales and then John of Cornwall, a contemporary of Geoffrey, produced his own Prophetia Merlini (Prophecy of Merlin).

Geoffrey had written his own Prophetiæ Merlini which is believed to have been circulated as a stand alone work, Libellus Merlini, prior to incorporating the prophecies into his later work the Historia regum Britanniae, originally titled De gestis Britonum, c.1138. Debate continues as to whether John of Cornwall copied Geoffrey’s work or vice versa, yet there are subtle difference between the works which hints at their independent knowledge of a common legend.

Geoffrey returned to Merlin around 1150 with his Vita Merlini, or 'The Life of Merlin', in which he transformed the boy prophet and magician Merlin Ambrosius, who brought Stonehenge to England in the Historia, as a wildman of the woods who had gone mad after a battle in the north. It is generally agreed that the Vita Merlini was influenced by the Welsh Myrddin poems which pre-date Geofffrey, so we can be fairly certain he did not create the Welsh prophet.

Two Merlins
Geoffrey’s later Merlin of the Vita seems to be a consolidated character. We can guess that at some point after writing the Prophetiæ Merlini and incorporating it into the Historia, Geoffrey became aware of a northern wildman tradition from which he mixed components with the Welsh Myrddin tradition.

It would appear that the idea of two Merlins was well known in the 12th century. In The Itinerary of Archbishop Baldwin through Wales, Gerald of Wales (1146-1223), writes:

"There were two Merlins; the one called Ambrosius who prophesied in the time of king Vortigern, was begotten by a demon incubus, and found at Caermardin, from which circumstance that city derived its name of Caermardin, or the city of Merlin; the other Merlin, born in Scotland, was named Celidonius, from the Celidonian wood in which he prophesied; and Sylvester, because when engaged in martial conflict, he discovered in the air a terrible monster, and from that time grew mad, and taking shelter in a wood, passed the remainder of his days in a savage state. This Merlin lived in the time of king Arthur, and is said to have prophesied more fully and explicitly than the other." [Book II, Ch.8]

This later Merlin, also known as Myrddin Wylt or Myrddin Sylvestris, shows some striking parallels with the tales of a northern wildman said to have lived in the Caledonian forest during the 6th century who is said to have gone mad after witnessing a horrific battle, associated with the Battle of Arfderydd, and fled to the forest where he was cursed with the gift of prophecy. This theme is also apparent in the Welsh Myrddin poems.

The Northern Wildman
The Merlin legend was a late addition to Scottish literature but it is claimed to be based on the northern wildman tradition of Lailoken that developed earlier in Scotland associated with St Kentigern (Mungo) from 6th century Strathclyde and mentioned by Jocelyn of Furness in his 12th century Life of St Kentigern.

The story of Lailoken appears in two versions written in Latin that are found in the 15th century MS Cotton Titus A xix. The manuscript includes the only surviving fragment of the Herbertian Life of Kentigern, an abridged version of Geoffrey of Monmouth’s Vita Merlini, and the Vita Merlini Sylvestris (the Life of Merlin of the Forest).1 The two tales of Lailoken are contained in the Life of Merlin of the Forest and generally referred to as "Lailoken and Kentigern," (Lailoken A Text), and "Lailoken  and Meldred," (Lailoken B Text). The two texts are neatly summarised by Basil Clarke2:

A Text
“Kentigern met a wild man in a desert and prayed for him at Kentigern's church by Glasgow. The Wildman had been driven mad by an accusatory vision in the sky during a battle. Later he used to appear (apparently unrecognised) at Kentigern's church by Glasgow, where he was a nuisance with his wild prophecies. At length he demanded the last sacrament from Kentigern, who tested his sanity by asking three times about his death and received conflicting answers (that he would be stoned and clubbed, that he would be pierced by a stake, and that he would be drowned). 

Under pressure from his clerics, Kentigern gave the sacrament. Lailoken then predicted the death of a king, a bishop and a lord within the year. The same day Lailoken was killed in the triple way he had predicted, being set upon by shepherds near Dunmeller. The clerics grasped the point and the story ends with their distress.”3

Lailoken's death is described in the A Text, [as soon as he had received the sacrament he] “rushed away like a wild goat breaking out of the hunter's noose and happily seeking the undergrowth of the wilderness………… it came to pass that on the same day he was stoned and beaten to death by some shepherds of King Meldred, and in the moment of death had a fall, over a steep bank of the Tweed near the fort of Dunmeller, on to a very sharp stake which was stuck in a fish pool. He was pierced through the middle of his body with his head bent over into the shallows, and so yielded his spirit to the Lord as he had prophesied.”

B Text
“Petty-king Meldred of Dunmeller captured Lailoken to hear him prophecy: Lailoken stayed mute for three days. When the queen entered court, he laughed. The disclosure of the queen's adultery, through the leaf in her hair, followed. But before disclosing this Lailoken predicted his own death in a few days, obtained a promise that his body should be buried where Pausail Burn meets Tweed and prophesied about the re-integration of the British nation.

The queen failed to discredit Lailoken, and plotted his death. A few years later he was set upon by shepherds at the instigation of the queen while he was passing Dunmeller at sunset on the same day that he had received the last sacrament, was killed as predicted, and was buried by the king, as he had been promised. (The time discrepancy - a few days, a few years - is not cleared up).”4 

Lailoken's death in the B Text: "As he had predicted and as it is recorded above, so we have heard was his end accomplished. It is said that the king handed over his lifeless corpse for burial in just that place which he had chosen while he lived. Now that fort is some thirty miles from the city of Glasgow. In its plain Lailoken lies buried.

Pierced by a stake, suffering by a stone and by water,
Merlin is said to have met a triple death.”

The author(s) of the two tales were seemingly keen on linking the northern wildman with the prophet of Welsh poetry, Geoffrey's Merlin. The first tale, Lailoken and Kentigern, asserts the association with Merlin, “He was known as Lailoken, and some say he was Merlin” and the tale of Meldred and Lailoken is alternatively referred to as a 'Scottish Tale of Merlin'.5

The tradition of Merlin’s grave near Dunmeller appears to have been widely known in the 13th century as Thomas the Rhymer (Sir Thomas de Ercildoun) prophesied:

“When Tweed and Pausayl meet at Merlin's grave,
Scotland and England shall one monarch have” 

Based on this northern tradition, Lailoken was identified with Merlin, and buried near the village of Dunmeller (modern Drumelzier), near the point where the Powsail Burn joins the River Tweed. Across the Tweed from Drumelzier is a spot called Merlindale. Indeed, 'Merlin’s Grave' has been marked on maps since the 18th century, although no archaeological remains have ever been found at the site.

The Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS) record says:

"Merlin's Grave" (Site): According to legend which is at least as old as the 15th century, the wizard Merlin was buried 200 yds NNW of Drumelzier Church, on the level haugh close to the right bank of the River Tweed. No structural remains are now to be seen, or have ever been recorded, at the place in question, but it is possible that the tradition may have been originated from the discovery of a Bronze Age cist.”6

The Search for Merlin’s Grave
The legendary figure of Merlin has been associated with Drumelzier in the Tweed Valley in the Scottish borders for many years. It would appear then that the location of Merlin’s Grave is no secret either and has been known since at least its appearance in 12th century literature, with oral accounts probably circulating well before that. 

Today the ruins of the medieval Tinnis Castle stand on a prominent rocky knoll overlooking the Upper Tweed valley, less than half a mile north east of Drumelzier. Underneath the ruined castle walls lies the remains of a hillfort whose ramparts enclosed the summit, this was a ‘nucleated fort’ type of hillfort unique to Scotland between c. 600-1000 AD. 

The Scottish tale of Merlin is centred in and around Dunmeller, from which the place name Drumelzier is said to derive, where he was imprisoned by a 6th century King, suffered a triple death and was buried by the banks of the Tweed. The village of Drumelzier is today still overlooked by the ruins of Tinnis Castle, said to be the location where Lailoken was held captive by Meldred (B Text). Is this tradition based on fact or a legend that migrated to the area? 

As we have seen above, the RCAHMS record states that no structural remains have ever been known at the site of Merlin’s Grave but the tradition may have originated from a Bronze Age burial cist, a stone-lined grave covered over with another slab, in the Tweed Valley.

In 2022 the Drumelizer’s Hidden Heritage Project team led by Ronan Toolis with volunteers pulled from across Scotland set out to investigate the archaeological roots of the Merlin legend at Drumelzier. In November that year a geophysical survey by GUARD Archaeology was carried out in a field to the north of Drumelzier with the objective to see what lay beneath the surface on the spot marked as Merlin’s Grave on maps. Nothing was detected at this spot but only a short distance away to the south-east in the same field an anomaly was found where there appears to be some form of archaeological remains, a possible grave, under the surface. Only excavation will reveal the true nature of the anomaly but permission has not been given to dig at the site. Perhaps this is the Bronze Age burial cist?

Across the Tweed excavations at Thirlestane Barrows, discovered a square barrow, dated to the late 3rd and late 6th centuries AD, constructed over the graves of two individuals of exceptional elite status. It is not known if this burial is related to the anomaly on the other side of the river.

At Tinnis Fort, which overlooks Merlin’s Grave, excavations revealed that this prominent hillfort was occupied around the late 6th to early 7th centuries AD, exactly the time when the Lailoken and Meldred story is set.

Tinnis was reoccupied on the site of a Late Iron Age hillfort which had been destroyed by an immense fire that reached such extreme temperatures that the ramparts were vitrified, a process in which temperatures reach a height that is capable of fusing stones together. At least 60 other vitrified sites are known in Scotland from around this period.

The extreme temperatures required to achieve this has led to a multitude of theories such as some ancient super-weapon such as a laser, or the ramparts were purposefully set on fire to strengthen them.

Scottish hillforts of this time had roofed structures many metres high, constructed on the rampart walls. Archaeologists have proposed that the extreme heat may have been the result of the burning of this timber superstructure with the fire raining down on the stones and heating them up like an oven. At Dun Deardail in Glen Nevis tests had shown blocks of molten stone were formed in anaerobic conditions without oxygen and likely caused by a “tremendous heat from above”.7

However, the question remains as to why these structures were set alight; it has been suggested it was a deliberate act of destruction at the end of the active life of the fort, or perhaps a long forgotten attack.

Following the destruction of the Iron Age hillfort at Tinnis by fire, the hilltop was reoccupied around 200 years later when the early medieval fort was built during a period when this part of the Scottish Borders was under the rule of the kingdom of Strathclyde. It may just be an odd coincidence but the one hillfort associated with the local Lailoken/Merlin legend dates to exactly the same period as the story is set, especially when hillforts of the post-Roman period in this area are relatively rare. Perhaps there is some truth in the tale of Merlin's death in Scotland? But who was Merlin: was he a historical character or just a figure of literature?

The legend of Merlin that has come down to us is clearly a mixture of history, literature and folklore. It is suggested that the Merlin-archetype may well have been based on a real person (Lailoken?) known to St Kentigern, who was driven mad by a vision during a battle and remembered in the literature of the Strathclyde Britons. This northern lore, including tales of the battle of Arfderydd and the wildman of the Caledonian forest, migrated to the Britons of Wales and inspired the Welsh Myrddin poems and later Geoffrey of Monmouth. Thus, Lailoken is proposed as the root of the Merlin legend.8

However, the archaeological evidence unearthed in the Tweed Valley does not prove that the local story was true but does raise the possibility that the legend may have originated in Drumelzier itself.

Notes & References
1. Tim Clarkson, Scotland’s Merlin: A Medieval Legend and its Dark Age Origins, Birlinn, 2016 pp.40-45.
2. Basil Clarke, The Life of Merlin, Geoffrey of Monmouth: Vita Merlini UWP, 1973, pp.24-25.
3. Clarke, p.24.
4. Clarke, p.25
5. Extracts of the relevant texts can be found in the appendices to Tim Clarkson, Scotland’s Merlin, Birlinn, 2016:
Lailoken (from Jocelin of Furness Life of St Kentigern), p.137.
Lailoken and Kentigern (from Vita Merlini Sylvestris), pp.138-142.
Lailoken ans Meldred (from Vita Merlini Sylvestris), pp.142-144.
Myrddin Wyllt, extracts from poems attributed to Myrddin, pp.148-154.
6. The Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS): "Merlin's Grave"
7. Archaeologists solve ancient mystery of '˜melted' Iron Age fort – The Scotsman
8. Clarkson, pp.131-136.

Further Reading

• The results of the two year study were published last week in Archeology Reports Online. 
• Archaeologists uncover new evidence of Merlin legend in Scotland  - The National, 3 September 2024
• Unearthing ancient Tweeddale: ‘Merlin’s Grave’ and other lost stories embedded in the landscape - The Past
• Drumelzier’s Hidden Heritage project
• GUARD Archaeology: Unearthing Ancient Tweeddale

 

* * * 

Petuaria: Civitas or Vicus?

The Petuaria ReVisited Project, currently investigating the Roman site at Brough on Humber in the East Riding of Yorkshire, has been making the news recently: ‘Dig sheds light on Roman history of town’, (BBC News, 3 August 2024)

The news article displays a rather splendid artist's impression by graphic designer Mark Hoyle showing how the fortifications of Roman Petuaria at Brough on Humber may have looked, demonstrating the imagined scale of the fort, immediately reminiscent of the Roman coastal fort at Portchester. 

This Roman site, located about 12 miles west of Hull on the north bank of the Humber, has a rather unclear history with its full size and status yet to be determined. 

The site is mentioned by the 2nd century astronomer, mathematician and geographer Ptolemy from Alexandria, as the tribal centre (polis) of the Parisi, the only ‘town’ he attributes to their territory.  Located east of the Brigantes, the Parisi were an Iron Age people of the Arras Culture, named from the early 19th century discovery of chariot burials at Arras near Market Weighton, East Yorkshire. Since those early excavations over twenty chariot burials have been discovered in the region.

It is generally agreed that the settlement at Brough on Humber started life as a Roman fort, Petuaria, around 70 AD, then abandoned by 125 AD.  The historian and archaeologist Sheppard Frere claims earthwork defences were added during the Hadrianic period (117 AD to 138) for a short re-occupation which he suggests could be connected with the time when Hadrian became emperor as his biographer writes "the Britons could no longer be held under Roman control" which Frere interprets as indicating there was war in Britain.1

Frere suggests the early fort may have been a 30-acre site for a legionary vexillation or group of auxiliary regiments organised as a battle group, as at Derventio, modern Malton.2

By the middle of the 2nd century military occupation of the site had ceased and a civilian town was arising over the site of the Roman fort. Frere suggests that the Roman site at Brough, strategically placed on the north bank of the Humber estuary, may have remained in use as an army supply depot.3

The civil settlement expanded across the site of the fort replacing it with a walled settlement surrounded by a turf and timber rampart. The site was refortified during the later 3rd and early 4th century with a stone wall. Archaeologists believed that many of the original Roman buildings had been destroyed as the town developed over the years. The site certainly has a complex history which has led to confusing interpretations. 

The Roman walls at Brough have been known about since the 1930s, but recent archaeological excavations by The Petuaria ReVisited Project have led to new discoveries with ground penetrating radar (GPR) revealing the remains of ditches, roads and walls and located what it is believed to be a Roman road in a nearby garden. 

Petuaria marked the southern end of the Roman road known now as Cade's Road, which ran roughly northwards for a hundred miles to Pons Aelius (modern Newcastle upon Tyne). The section from Petuaria to Eboracum (York) was also the final section of Ermine Street. 

The Roman Roads of Britain (Roman Roads Research Association) 

In 1936 a stone-lined burial was found on the western side of the Roman road to York, as with the usual Roman custom interred outside the town walls. The grave contained an iron-bound wooden bucket and two sceptres, which has led to identification of the inhumation as a priest. The two sceptres had been intentionally bent and broken in preparation for their owner’s new life in the Otherworld. The intentional destructive treatment of grave goods is a native burial rite, found at many other sites across Britain, much has been found in the river Thames for example. Frere adds that nothing could illustrate better the dual character of the Romano-British civilisation; outwardly Roman, yet inwardly it remained Celtic.4

The Petuaria ReVisited Project team, including more than a hundred volunteers, are also re-visiting the Roman defences at Burrs playing field in the centre of the modern town. Recent geophysical surveys and excavations have shown many structures lie under the field. 

During the 1930s archaeological excavations were carried out at Burrs Playing Field (formerly known as Bozzes Field), which revealed traces of a stone wall, buildings and a sequence of fortified structures. However, the highlight of these excavations was the discovery of a stone inscription found lying on its edge 30m south-west of the east gate of the Roman town in 1937. Catalogued as RIB 707, the inscription refers to the dedication of a theatre and has been interpreted as referring to the capital, or the civitas, of the Parisi. The identification of Petuaria with Brough-on-Humber has depended almost entirely on this inscription.

Ptolemy, who probably never visited Britain, became confused with the location of his so-called polis of the Parisi. Rivet and Smith note that Ptolemy correctly records a distance from London to Petuaria of 170 miles, but the ancient geographer logs the distance from Eboracum (York) to Petuaria as 45 miles, when the true distance is 30 miles.5 Ptolemy’s distance would put his polis somewhere on the east coast of Yorkshire.

One would expect the location of Petuaria to be clear in the Roman Road network, as the roads go there, yet with regard to The Antonine Itinerary: ITER I, Rivet and Smith write that after York serious problems arise.

ITER I runs from the Roman fort at Bremenium (High Rochester), north of Hadrian’s Wall, to Praetorio, somewhere in east Yorkshire. Following Dere Street the route runs along the eastern side of the Pennines to York. As Rivet & Smith note, the route south of York is not clear and has been the subject of much debate over the years. After York (Eboracum) the route is recorded as travelling to three forts Derventione, Delgovicia then Praetorio. For many years Derventio has been identified as the fort at Malton, on the river Derwent, which the Notitia Dignitatum records as garrisoned by the Numerus Supervenientum Petueriensium.6

The British archaeologist and specialist in the translation of Latin text and epigraphy Roger Tomlin proposes that Numerus Supervenientum Petueriensium was a military unit, where Petueriensium would appear to refer to the fort of Petuaria, whose soldiers could have been reinforcing the fort at Malton.7 It follows that owing to the proximity of Petuaria to Malton the assumption has manifested that Derventio must be Malton. 

However, the counter-argument is that Derventio is not Malton, but the Roman site south west of Stamford Bridge, also on the Derwent. It follows that if Derventio is Stamford Bridge, then Malton must be Delgovicia.8

The lies confusion with Praetorio that has been identified as Petuaria (Brough on Humber) because of the similarity of the names. Rivet & Smith argue that Praetorio is not actually a proper name, rather it is a descriptive term meaning an official residence. They also suggest Delgovicia could be a site near Wetwang and conclude that Praetorio should be accepted as a corruption of the name Petuaria since the roads lead there.9 

However the distances don’t work; the Itinery records 26 miles from Praetorio to Delgovicia (Malton), then a further 13 miles from Delgovicia to Derventio (Stamford Bridge), a total of 39 Roman miles. Whereas Petuaria is 26 Roman miles from Stamford Bridge, and 33 from Malton. As a solution Praetorio has been identified as Bridlington on the east Yorkshire coast.10 

And just to confuse matters further, Petuaria appears in the Ravenna Cosmography as Decuaria which according to Rivet and Smith is best explained as a simple miscopying.11

The Inscription 
The inscription has been translated as:

RIB 707: "For the honour of the divine house of the Emperor Caesar Titus Aelius Hadrianus Antoninus Augustus Pius, father of his country, consul for the second time, and to the Divinities of the Emperors, Marcus Ulpius Januarius, aedile of the vici of Petuaria, presented this [new] stage at his own expense." 12  

RIB 707 (Roman Inscriptions of Britain)

From this we can see the town possessed a theatre whose new stage was presented by a Roman named M Ulpius Ianuarius, aedile of the vicus of Petuaria, who set up the tablet in honour of Antoninus Pius (emperor from AD 138 to 161).13

The Theatre
The theatre is important because it links the site to the inscription and would confirm its place in history. Yet, as much as archaeological investigations have searched it has remained elusive for many years. Frere states that this theatre is known only from the inscription (RIB 707) and the physical remains are yet to be discovered.14 Needless to say, its discovery would be a major find for The Petuaria ReVisited Project.

In 2018 specialist ground scanning equipment revealed the remains of buildings with rooms arranged around a courtyard, interpreted as a forum and a substantial D-shaped feature under the playing field; was this the long lost theatre referred to in the inscription?

In 2020 a trench was cut over the D-shaped feature. At the northern end of the trench, the team uncovered a hearth that contained a burnt coin dating to around AD 330. Following two seasons of excavations over the D-shaped feature the team has failed to find the theatre. With the absence of the theatre one must question if the inscription originated at this site or was brought in from another location?

However, they have found evidence of continued Roman activity in Brough into the late 4th and probably 5th centuries AD, longer than previously thought, contra Wacher (see below), who suggested that it had faded out of use by the early 4th century.

Civitas?
The reign of Antoninus Pius saw great advances in the towns of Britain; at Brough a small town was developing on the site of the recently evacuated Roman fort at Petuaria, and a theatre was built during his reign. 

Theatres were relatively rare in Roman Britain, only three others are definitely known, from the major settlements at Verulamium (St Albans), civitas capital for the Catuvellauni, Durovernum, (Canterbury), civitas capital of the Cantiaci and Camulodunum (Colchester), the first Roman capital of the province. A ceramic theatrical mask found at Cataractonium (Catterick) does not necessarily confirm there was a theatre at the civil settlement, it may simply be from a theatre-goers personal collection. 

Accordingly, based on the inscription, the theatre at Brough has been interpreted as indicating it was also a major settlement, possibly the civitas capital of the local tribe, the Parisi.

In Roman Britain civitas capitals were the administration centres of local level government, consisting of local people installed by the Romans to control their own tribal areas. Many of these may have originated as pre-existing Iron Age settlements; it is estimated that there were over 20 Iron Age tribal territories, of which between 11 and 16 had defined civitas capitals and awarded a tribal suffix such as Calleva [Atrebatum] (Silchester) and Venta [Belgarum] (Winchester). 

However, line 8 of the inscription RIB 707 refers to Petuaria as a vicus (village) not a civitas, yet the dedication was ordered by an aedile, who maintained public buildings and was responsible for entertainment. This was a significant position, with a higher role than seems necessary for a mere `village`, hence the argument that Petuaria was the civitas capital of the Parisi. And as we have seen above the ancient geographer Ptolemy refers to Petuaria as the ‘polis’ of the Parisi.15 

Petuaria Roman fort, after Halkon 2013 p.132 & Ottaway 2013 p.174
(probable line of RR2e based upon cropmarks) (Roman Roads Reseach Association)

The limestone slab is broken along the bottom edge and on the right-hand side therefore lacking the right pelta decoration and a few letters in each line. The argument in support of civitas status follows that a letter ‘C’ apparent on the left side-panel and a letter ‘P’ on the missing right-side of the stone, has been interpreted as c(ivitas) P(arisorvm); the civitas of the Parisorum. However this interpretation has been rejected and the letter ‘C’ on the left-hand pelta said to be purely decorative.16

A vicus was a civil settlement that became established around a Roman military site, predominantly forts, as with Petuaria. There is a clear reference to ‘vici’ on the inscription and there is no evidence that Petuaria was ever granted a tribal suffix as found at other civitas capitals as noted above.

The name ‘Petuaria’ is thought to be derived from a root meaning ‘four',17 and may mean that it was the vicus of the fourth pagus, an administrative term designating a rural subdivision of a tribal territory. If this is correct it is possible that Marcus Ulpius Januarius held the position of aedile over all four subdivisions of the territory of the Parisi.

John Wacher, who excavated the site between 1958-1961, has argued that a series of characteristics contrast Brough from other civitas capitals where he sees a more military character for the site. These include the military nature of its defences and a lack of organised street system and early urban development.18

Wacher argued that that none of the excavated structures within the fortified area had parallels with domestic buildings normally found in Roman towns. He saw the site as military with an extensive naval base overlying the earlier fort. Wacher concludes that the occupation of Petuaria seems always to have been military and naval rather than civil and suggests the vicus Petuariensis may have been 3 miles downstream at North Ferribly.19

Wacher's Phasing for Brough-on-Humber20

This naval base, he suggested, was probably connected with the classis Britannica, the Roman Fleet operating in the English Channel. This interpretation has not been universally accepted as there is some doubt if ships of the classis Britannica were ever based as far north as the Humber as the fleet is mainly associated with the south east of Britain as attested by the profusion of stamped tiles “CL BR”21 and, to date, no tile-stamps of the classis Britannica have been found north of London.22

Although Wacher leaned towards a purely military site with a naval port,23 the current status of the site is unresolved; the origins of Petuaria appear to begin with the Roman fort but the development of the associated vicus is unclear. It is not even certain that Brough was ever known as Peturaria Parisorum or the civitas of the Parisi. And we cannot confidently identify it from The Roman Road Itinerary Iter I. 

It is significant that the stone bearing the inscription (RIB 707) was not found in its original setting where it had been erected about 140 AD but had been re-used in a 4th century building in the naval base.24 And the location of the theatre, if it was indeed ever at Brough, currently remains elusive.

The artefact assemblage is far from conclusive and the available evidence allows the site to be interpreted as a fort, a naval port and civilian settlement over sequential periods of occupation, but not necessarily in that order. It is anticipated that ongoing excavations by the Petuaria ReVisited Project will clarify the matter.

Notes & References:
1. Sheppard Frere, Britannia: A History of Roman Britain, 1967, Routledge & Kegan Paul (BCA edition, 1973), p.126.
2. Frere, p.98
3. Frere, pp.249-251
4. Frere, p.303
5. A.L.F. Rivet & Colin Smith, The Place-Names of Roman Britain, Batsford, 1979 (reprint ed. 1981), p.119.
6. Rivet & Smith, p.155-56.
7. RSO Tomlin, Numerus Supervenientium Petueriensium, pp.74-75, in JS Wacher, Excavations at Brough-on-Humber 1958- 1961, Reports of the Research Committee of the Society of Antiquaries of London No. XXV (1969).
8. Creighton 1988, pp. 401-2 – quoted by Roman Roads Research Association (RRRA), The Roads of Roman Britain, The Antonine Itinerary – Iter I
9. Rivet & Smith, 1979, p.155.
10. The Roads of Roman Britain, RRRA, The Antonine Itinerary – Iter I 
11. Rivet & Smith, p.437.
12. RIB 707. Dedication to the Divine House of Antoninus Pius, Roman Inscriptions of Britain,~Roman Inscriptions of Britain 
13. Frere, p.303
14. Frere, p.244
15. Rivet & Smith, p.142: Ptolemy writes"Camunlodunum [Colchester], next to whom, beside the gulf suitable for a harbour, are the Parisis and the city of Petuaria".
16. RIB 707. Dedication to the Divine House of Antoninus Pius, Roman Inscriptions of Britain,~Roman Inscriptions of Britain
17. Rivet and Smith, p.437-38: The name is a femimine of British *petuario = fourth (Welsh pedwerid, pedwyr-yd). 
Compare with the Middle Welsh poem Preiddeu Annwn from the Book of Taliesin in which Arthur  leads an expedition to the Welsh Otherworld to take a cauldron from “the Four-Peaked Fortress, four its revolutions” (ygkaer pedryuan, pedyr ychwelyt)
18. J.S. Wacher, Excavations at Brough-on-Humber 1958- 1961, Society of Antiquaries of London No. XXV (1969).
19. Wacher, The Towns of Roman Britain, Batsford, 1978, pp.394-97.
20. The Status of the Roman Settlement, 7.0 Discussion by K. Hunter-Mann, with D. Petts, The Roman settlement at Brough-on-Humber, Internet Archaeology.
21. Historic England Research Records: Petuaria Roman Town
22. 2481. Tile-stamps of the Auxilia: Classis Britannica, The Roman Inscriptions of Britain 
23. Wacher, The Towns of Roman Britain, Batsford, 1978, p.395
24. Rivet & Smith, p.438.

* * *