Bamburgh Castle: St Oswald’s Gate

Our current area of excavation is St Oswald’s Gate and the outworks that lie beyond it. The gate and outworks connect Bamburgh the fortress to Bamburgh Village, and to the wider landscape and seascape beyond. The low lying natural cleft in the bedrock was the entrance to the site from prehistory and has some of the least altered medieval structures available to be investigated today. The archaeological work below has been undertaken intermittently over a number of years and includes the fortress defences, the routes that would have led to the village and farmland beyond and even includes a now lost port! Our current excavation is exploring a medieval tower with a well within its basement room. Last season we also identified a second postern gate (that led outto the village) and a stone base or foundation that might even be early medieval in date. We will continue investigating all of these this year.

The route of least resistance: the origins of the earliest known entrance into Bamburgh Castle

*St Oswald’s Gate looking south-east from one of the mounds that could be part of a siege castle*

Bamburgh owes much of its long historical relevance to its location and geological form. A natural rock fortress with flat surfaces on the summit where buildings could be constructed, but also of a size that allows for it to be defensible. We now know it was occupied as a fortress from prehistoric times, as our earliest radiocarbon date suggests construction activity and occupation from the late Bronze Age (10^th century BC).

The rock that forms the fortress extends some 320m west-south-west to east-north-east and 78m south-south-west to north-north-east. The highest point is in what is now the Inner Ward and is just over 45m above sea level. The bedrock undulates downwards towards the north and reaches its lowest point, 20m above sea level, at what is the base of the cleft of St Oswald’s Gate. This cleft is a natural fissure, 3.4m wide at the top that narrows down to 1.6m wide at the base, though it seems likely that the cleft has to some extent been widened in the past as only one side is formed of naturally solidified lava and the other is rough and angular and appears to have been cut away to some extent. The ground surface outside the entrance is 12m above sea level showing that the height difference between the surfaces within the fortress and the external ground is at is lowest here. The cleft is the earliest recorded entrance to the fortress, mentioned in writing in the later 8^th century AD. Given that this route would have offered the path of least resistance, and easiest way up onto the castle rock, it seems highly likely that this was always the way in to the fortress from prehistory up to the 12^th century.

*Layout of the castle with trenches noted*

One of the features that likely made Bamburgh a desirable high status fortress is the ready access to both land and sea. Immediately to the north of the gate there is a low lying area of ground that we know from the earliest map evidence was formerly a tidal inlet, now cut off from the sea by the developing sand dunes. Written accounts indicate that the inlet acted as a tidal beach-harbour for part of the medieval period, up to the middle 13th century. Past and current work to understand this area better and how it changed in importance over time is listed below.

A Defended Entrance

The area of the fortress within the gate has been the focus of investigation by the BRP from almost the beginning of the project, as one of the principal trenches (Trench 1) was sited to investigate inside the entrance and to find traces of early defences.

Research was frustrated at first by the presence of an intrusive and extensive construction trench from an early 20^th century wall that encloses most of the West Ward. This had removed much of the evidence of the earlier defences down to boulder clay and bedrock level. Thankfully, in the area of the gate the medieval wall had survived (albeit it seems refaced in the post medieval period) and here we found evidence for earlier defensive structures preserved immediately inside the wall. This was because the later medieval wall had been constructed to overlap the break of slope at the edge of the plateau. This was to show a more impressive, and taller, wall face out to the world without the need of too much masonry. As a consequence the rear faces of earlier defensive structures, that had been constructed entirely on the top of the rock and close to the edge, survived on the inside of the later wall. Here we discovered traces of early timber defences, that were later replaced in stone, a series of substantial timber, then stone buildings, were also found aligned on the gate cleft and have been interpreted as the lodge of a gate-ward who controlled access to the palace.

*Excavated post-holes shown in the foreground and a small section of the rubble wall foundation still in place at the top of image*

The Small Port

The castle is the subject of many paintings dating back to the 18^th and 19^th centuries and those that depict it viewed from the north show a much more intimate connection with the sea than today. They frequently depict the tide coming much closer to the base of the rock and occasionally showing boats in the inlet near St Oswald’s Gate. This small tidal inlet is also depicted on the early mapping, from the 1^st Edition Ordnance Survey (c. 1870) where it is labelled ‘Postern Nook’. It is easy to imagine the shallow draft ‘longships’ of the early medieval period being drawn up in the this tidal inlet and able to provide access to the sea, which was much more of a highway than a barrier in early periods. This crude and awkward little harbour seems to have been in use up to at least the early 13^th century when we read of instruction sent to the Sheriff of Northumberland from the government to proceed to Bamburgh and arrange the release of ships that had been held there by the community.

*The northern end of the castle at St Oswald’s Gate showing the likely tiday beach area before the modern dunes developed*

Development of the Defences

The main early medieval structures within Trench 1

St Oswald’s Gate was well sited to provide access to both land and sea via the modest harbour in the inlet, as a consequence of the low-lying topography making entrance easier. These factors help to explain why the gate remained the principal entrance for much of the fortress’ life and, as a result, adds to the importance of understanding how it developed, linking the archaeology undertaken at BRP Trench 1 within the entrance to the outworks and access routes that led up to the gate.

Within the gate the surviving evidence reveals well constructed timber defences that begin as deep set post-holes, cut into bedrock, that are later replaced by a rubble foundation in a trench that we believe carried a large baulk of timber into which vertical planks would have been set to form a wall. We do not yet know how early the post-holes were cut, but it is easy to imagine the site as a kind of Iron Age fort like those we see on hilltops throughout the north. We know from Bede’s history that Bamburgh was a well defended timber fortress in the middle 7^th century when Bede describes how Penda King of the Mercians (a large rival kingdom to Northumbria in the English midlands) demolished all the timber buildings in the surrounding landscape to pile a huge bonfire against the rock in an effort to burn his way through the defences.

Inside these defences we found evidence of substantial rectangular halls of early medieval date. The earliest of them was 12m by 7m and built with a post-in-trench construction technique that suggests a construction date around the 7^th century. This was later replaced by a stone structure that may have had a lead roof. We can offer a reconstruction like this as we have the outline of the building foundation (really a robber cut from which the stone had been removed for recycling) and a small number of the dressed stone building blocks still in place. The presence of a large and heavy lead ingot found in the trench is our evidence that a lead roof had been melted down, again for recycling.

The view of the outworks from within the castle. The low lying area is the now silted port and the mounds to the left may include surviving elements of the siege mound of 1095

The stone building was robbed out no later than the 11^th to early 12^th century so could be evidence of stone architecture of a pre-conquest date. Evidence that the defensive wall of the fortress had also been replaced in stone before the conquest was not as conclusive but a mortar impression of a stone block set on the bedrock and part of a wall set back beneath the later and wider 12^th century curtain was seen just above the bedrock. This limited evidence was supported by the presence of a 9^th century grain drying kiln set right inside and partly beneath the 12^th century wall. As it was too close to the edge of the bedrock for a timber wall to have been present (timber structures and fire being a poor mix!) but a thinner stone wall could have been present from this time.

The evidence from these early medieval defences and structures shows an investment in functional and impressive architecture that seems to have been intended to impress or even awe those seeking to enter the royal fortress.

The Siege of 1095

One dramatic event that played out around this entrance occurred in AD 1095 when the then Earl of Northumberland Robert de Mowbray rebelled against King William Rufus, the son and successor of William the Conqueror. When Robert refused to attend at court William led an army up to Northumberland and up to the gates of Bamburgh where he realised that the fortress was too strong to take by force and so ordered a siege castle to be built to trap the defenders within. A chronicle tells us that as the earth and timber siege castle was thrown up, by the royal army and local soldiers called up to serve the king, that the defenders shouted abuse at them from the ramparts. The king called the siege castle – Mal Voisin – meaning evil or bad neighbour in Normal French. Given that the intension of the siege castle was to bottle up Bamburgh’s defenders it must have been close to the gate, which was of course St Oswald’s Gate at that time. Indeed, if the chronicle is correct, it must also have been at abuse shouting range! The mounds just beyond the gate, behind what is now the sports pavilion, have to be a good candidate for being the remnants of this castle – assuming that any trace does survive. One of our future aims is to investigate this and see if we can prove it.

A New Entrance and Access Route Defined

The layout of the fortress appears to have been substantially altered after the Norman Conquest and this involved changing the main entrance and access routes. Sadly we do not have a record that directly describes the building of the new Great Gate, so we do not know just when it was constructed. We do have a record of repairs to the gate and drawbridge in 1221, showing that they were old enough to be in need of repair by then, perhaps suggesting a construction date in the 12^th century. It is tempting to see the gate and drawbridge and new entrance route that wraps around the Inner Ward as part of a new style in castle building, called concentric fortification. This was introduced to Europe after the crusades and changed the way castles were built. This again would suggest a construction date sometime in the 12^th century.

Changes to the main entrance also reflects changes in how the castle was used and the role it played in its landscape. Now connected to the land and with the setting of the castle more about the drama of the site and how it impresses those that approach it. Much less focus was placed on the sea as a means of communication from this time.

Tower of Elmund’s Well

Oswald’s Gate remained a functioning entrance but was now not the main entrance. The fact that it retained importance is clear from the construction and maintenance of outworks and a tower beyond the gate. In the records the tower is called the ‘Tower of Elmund’s Well’, and when we look at these and its form (relatively small and rectangular) we can make a case for the tower being constructed by the 12^th century. It is possible that the tower, in some form, is earlier as the name certainly seems to have pre-conquest origins, but only the current fieldwork has the chance to demonstrate if this is the case.

The reason for the complex outworks and the tower is surely at least at first related to the presence of the small port, as the tower is well sited to dominate it. We do have reason to think that the port, which would always have been small and rather limited in use, was replaced by a more practical deeper water port in Budle Bay during the 13^th century. This may fit in with construction or reconstruction of a further postern gate the leads out towards the village.

Post Medieval Decline

As far as we can tell the postern remained in use throughout the medieval period but seems to have gone out of use as an entrance in the 17^th century as animal carcasses were dumped onto the steps leading down to the gate around this time. This is perhaps not too surprising as after the great siege of 1464 the castle falls increasingly into ruin. Local tradition tells us that a different low lying gap in the bedrock of the West Ward, called the Miller’s Nick, was used as a way to climb up into the castle in its ruined state. Given the name it is perhaps understandable that the tradition was that it was the short cut to the windmill.

The ‘Witches Cottage’

When the castle’s fortunes began to rise again towards the end of the 18^th century, the castle and its increasingly rebuilt structures found new uses. It was at this time that the Tower of Elmund’s Well was also repurposed. Expanded into part of a small cottage that seems to have been the dwelling of an apothecary who aided the medical needs of the poor of the district, providing herbal medicines from plants grown in the grounds. The apothecary would have been part of the infirmary within the castle with its Dr/Surgeon, paid for from the Lord Crewe Trust.

The Archaeological Work

The outworks beyond St Oswald’s Gate lie at ground level below the castle rock and are a good distance from the main buildings of the Inner Ward and as such have not been altered as much. Post-medieval work seems to have been just the reuse of the tower as part of a cottage and some rebuilding of the north and west walls that had long fallen. As a result of the area being so little altered it repays archaeological investigation as its medieval phases are much more apparent. In addition, a good understanding of the build sequence should aid in identifying the purpose of the outworks and how this changed over time as the importance of access to the port area and to the village changed with the fortunes of each.

The aerial photograph shows the outworks immediately outside the entrance to St Oswald’s Gate (left side of photo) two different phases of wall – the thin and earlier wall has an arch leading down towards the later cottage and a thicker wall – diagonally across the photo – is a postern gate leading out towards the village

The first trenches dug on the outworks were as far back as 2002. It did show us that the area was important and needed further work, so its good to have been able to concentrate on this area for the last two seasons.

The initial work had concentrated on the tall standing walls immediately outside the gate but had not got very close to the cottage where remains of the tower should also lie. Our first objective when embarking on the new work in 2020 was to be sure that something remained to be found or it could be a rather short dig. Our first task was to clear the steps down to the area where the cottage was known to have stood. This involved the removal of a lot of soil and ivy but secured our access to the new site and reassuringly confirmed that some structures were still present to be uncovered just where the cottage should be.

During the first season of the new work, in 2021, we were able to reveal the full plan of the missing tower and the elements built onto it it later to form the cottage. The surviving cottage structure was a complex entrance that led down to the basement level of the tower that we believe contained the well.

The cottage elements that survive at this level (ground floor and below) comprise two or three phases of the building (seen as different stonework) that form stairs that turn through right angles as they descend down. In the photo below you can see the entrance as it was uncovered being recorded by photography. Behind the two people in the foreground -still hidden at this time was the straight stair that led down to an archway into the medieval tower that was reused as the core of the later cottage.

*The Tower of Elmund’s Well (foreground) and elements of the post-medieval cottage forming the entrance to the tower on the lest side of the photo*

As we began to uncover and empty the rubble fill of the room that we believed was the tower we found -built into two walls of the structure- were carefully constructed (later blocked) narrow arrow-slit style windows. This makes us fairly confident that we are right that this is the surviving medieval elements of the structure.

*Drone shot of current excavations revealing St Elmund’s Tower*

Finding the Well

The excavation of the Tower of Elmund’s Well has proved to be a more productive, investigation than we imagined when we set out. More structure survived below ground than we we expected and with a very complicated, but fascinating, build history with many phases visible. The work has involved digging out the fill and moving it in buckets up the stairs to the spoil heap, but we were able to slowly reveal the stairs down into the well room at the base of the tower and discover the well itself!

Medieval wells seem to come in both circular and square forms and are often quite substantial affairs. We do know from earlier work that the early medieval well, still to be seen in the keep, was circular and in the order of 2m in diameter. The well in the tower here turns out to be more of a cistern type. It is accessed from a platform that forms the floor within the well room, both from the platform and also via a further short flight of two steps curving down from the far left corner. These stop at the second step at a vertical edge that goes down about a further 0.5m (likely more when we finaly bottom the well).

In the rectangular ‘well’ there is on the south side a very crude section of wall constructed of rough and irregular stone. This seems to be deliberately constructed this way to be porous allowing water to enter and fill the well up to the height of the water table. On the east and north (the side of the platform) sides the masonry is neat and squared and regularly constructed and may even be later in date. The east wall that is part of the extensive bastion that enclosed both the tower and wider area is also quite crude in construction, though not to the extent of the south wall.

The platform through the archway at the base of the steps. The well is the square cistern full of water where the pump is lying. The two steps on the right are likely to be there to make reaching the water easier in a dry season.

The water table that corresponds to the water level of the well in use can be seen in an area of staining on the east and north sides. This ‘ancient’ water level seems to be confirmed by a feature seen in the north wall. Here a small circular hole extends level beneath the platform for over a metre (as far as the steel pence peg we had to investigate it could reach). We think this is a drain that leads outside of the tower on the seaward side. If it is a drain then it would prevent the water overflowing up onto the platform during wet parts of the year and the two steps would have aided in reaching the water in dryer times when the water was lower. If we can safely excavate to this level on the outside of the tower then we may be able to confirm the presence of a drain.

As the fill of the well was further removed other features were also revealed. The most interesting and exciting was the discovery of substantial structural timbers seen on three sides of the well so far. These were large and square sectioned and at least on the east side underlay the stone wall that extended above this. This relationship showed that the initial interpretation that the timbers represented a lining within the well structure was not the simple and obvious interpretation. Could the well have first been built as a timber structure? Does it even predate the construction of the tower? These questions remain for the coming summer season in July 2026 when we also hope to take samples for dendrochronology. As the timbers lie beneath the water table there is a very real possibility that the timbers could be early and preserved in these anaerobic conditions.

*The timbers exposed at the base of the well cistern. The water has been pumped out to reveal them temporarily for examination.*

The timbers are at a different alignment to the later masonry, and in places extend beneath them. It is clearly from a different phase and earlier phase. If we were particularly imaginative we might speculate that they could be as early as the 9^th century which is when the name “Elmund” is most attributed in Old English primary sources.

Although the purpose of the well at ground level to the north-west of the rock on which the main castle stands is never explicitly stated in any annal or writing. We can speculate that it may be here a consequence of the small port that once lay to its immediate north. Replenishing water would have been an important for a functioning harbour. The defensive outworks were likely built to enclose this and control not just the port but the access to water. This seems the most logical explanation for the shape of the outworks and the square bastion that extends out to the edge of the break of slope down to the port.

The defences that face the village and the discovery of a second postern gate

The tower containing the well was likely one of the earliest parts of the outworks beyond St Oswald’s Gate and likely helped shape the space that the external defences included. We can also assume that even from the earliest times access out to where the village is today was important. There was an earlier version of the village present from at least the 7^th century and in even earlier times access to the surrounding farmland would have been important. It seems that the stone outwalks that we see today are mostly 13^th to 15^th century in date with alterations made in the 18^th and 19^th centuries when the cottage was built into the tower ruins. There is reference to a wall at the base of the rock from at least 1138, when we are told it collapsed during a raid by a Scottish army resulting in many casualties. As a result of this we have been trenching where space is accessible on the village side of the castle to see if we can find evidence of earlier phases of defence. Perhaps even very much older timber phases of fortification can be found archaeologically.

The access to the village from St Oswald’s Gate today is through a pastern gate that passes through the remains of a wall of substantial thickness that extends from the rock face for several metres to form part of the square enclosure that contains the tower and well. In places the wall still stands several metres tall but is lower and around 3m tall where the current gate access passes through it. The wall is medieval in origin and some features that survive within the gate also appear to be medieval, such as the partially blocked hole for a large square section timber bolt that would have held a timber door closed. The path that leads through this and some of the door facing stones are certainly late post medieval, as are the stairs on the outside. It is even possible that the wall here may have been lowered deliberately to clear the view to Oswald’s Gate behind it from the village as it is consistent in height and very different from the wall to its north.

Investigation of this wall in the area of the gate has already produced an unexpected find in the form of a second gate that had been blocked up and had lain concealed by foliage. This second postern gate has proven to be well constructed and, it seems, altered at least once during its time of use. As well as the final blocking phase that put this entrance out of use there was at least one reconstruction of the entrance itself. The walls that form its sides of the entrance are neat and well constructed masonry, raised in ordered courses to form an entrance 1.3m wide.

Square post settings (30 cm by 30 cm) with vertical sides were set just inside the walls on both sides. They would have held substantial timber uprights that either held a gate or formed the door jambs against which a gate closed. They do seem to match up with a new phase of masonry set against the existing wall just ahead of the post settings. Taken together they would make a new, somewhat narrower, entrance. Also an inconsistent set of stone slabs formed a new floor surface at a somewhat higher level that we cannot help but associate with this later phase of narrowed gate. That the new stone narrowing is wider on the right side (looking out) would be consistent with the door in the postern opening backwards to that side.

*The outwork defences immediately outside St Oswald’s Gate with the second postern gate being recorded (foreground).*

Work this Season (2026)

During recent seasons of work we have exposed and excavated the cottage and tower and revealed the well in the basement of thew tower. This year we will record this futher and further investigate the timber structure at its base.

There is certainly more to be identified and understood regarding the second postern gate and the phaes of the outworks themselves. Last year we we opened a nw open area in front of the standing defences at the base of the castle rock and found extnesive new archaeology in this area that bodes well for some fascinating new discoveries in July and August.

Publications

You can find out more in-depth information by visiting our blog:

Some useful posts include:

We have also published numerous articles on the excavations. Please visit our Publication page for a full list.