At the end of our 2021 dig season, we outlined the background and focus for the dig season here. In this blog post we look at the results of the dig.
A number of the older surviving plans of Bamburgh Castle depict the Tower of Elmund’s Well amongst the outworks beyond St Oswald’s Gate, so its location has long been known. In addition some some older aerial photographs (probably from around the 1950s to 1960s) seem to show that the cottage that was built into its ruins stood well above ground, and was even partly roofed at that time. There was some nagging doubt just what might survive today, as the area was covered with ivy and was not very accessible. So it was a matter of some relief that after even a few hours of ivy removal it became clear that stone structures did survive beneath the ivy above ground level. It would have been a much less interesting excavation season had we been trying to find a robber trench from which the stone structure had been entirely removed.
The outworks and the tower foundation lie at a much lower level than the main castle, you descend some 10m through St Oswald’s Gate and down to the areas where the tower stood. The route today is via a series of steps of varying date from quite modern to worn steps that may well be of early post-medieval date. There are two route ways (they split outside St Oswald’s Gate) one towards the village green and the other towards the tower and the port beyond to the west. How old these routes were was one of the questions we have posed as part of our investigations. It seemed likely that they both dated back to at least the later medieval period, even if the steps themselves were more modern, but some form of these routes are likely to date back even earlier as we know St Oswald’s Gate was in use from at least the early medieval period and perhaps the late Bronze Age.
One of our first tasks once we were on site was to clear the steps down to the area we were to work in. The lower part of which were covered in soil and ivy and required quite a bit of work. This would ensure we had a reasonably good access way to the site. I am sure the climb each tea-break was good for us, even if it did not feel that way.
It was natural to start clearing and investigating from the base of the steps northwards into the area where the cottage and tower stood. This means the first discovery was the end of a wall, that appears to extend back to the slope of the bedrock, and is likely to represent part of the wall that closed off the seaward side of the outworks on the north side. This stub wall had facing stones on the outside and some rubble and core work behind it but the other facing stones, that would have been expected on the opposite side, were missing. It would have been quite a wide section of wall had the other side been present and just possibly may have been a remnant of medieval date, given its form.
The fact that the wall we had just uncovered ended in a deliberately constructed face, on the west side, strongly suggested that we had a small gate present between the wall and the cottage. In fact the plan of the ancient parts of Bamburgh Castle compiled by the Antiquarian Cadwallader Bates for the 1st Lord Armstrong in 1895 shows a path in just this area, passing by the cottage east wall and then along the north wall veering off at its end towards the beach. There is no depiction of the wall end that we had found but in all other aspects it seems to confirm the presence of the route-way, and by inference the gate. The plan of 1803 showed the wall from the tower back to the bedrock as complete without a gate, but then this plan also shows the steps and path in a different area and neither map seems to be definitive, though may reflect changes in access arrangements between their compilation.
As we had a good idea of where the cottage and earlier tower lay, from the older plans of the site, we were able to start to reveal the top of the wall lines fairly quickly. Starting from the area of the gate, through to the beach, we were able to trace the top of the wall, westwards to the corner where it returned to the south. Tracing the wall top in the other direction (southwards) we discovered an area where the wall appeared to become more like rubble than an in place structure. This under excavation turned out to be an entrance, unsurprisingly right in front of the current steps down from St Oswald’s Gate.
To the immediate south of the entrance we also identified a wall along the south side that we at first considered might mark the southern wall of the tower. This proved not to be the case when we realised that this wall had one face forming the side of the entrance but that there was no outside face just core-work and sand. The wall had been built up against the sand subsoil (or what at the moment we think is subsoil) as what we call a revetment. It was not all that substantial and did not extend very far to the west making us see it as a late addition to the structures and only part of the cottage. Further investigation within the entrance, removing rubble and soil fill, revealed a set of steps down into the cottage, which we now realised survived more substantially below ground than we expected. More of this below.
Interlude – the enigma of Area A
Whilst the investigation of the cottage/tower area was our main focus for the season we also had questions concerning a short length of stone wall that lay to the south of the, still standing, main closing wall of the outworks. Whilst only a few courses high it survived over some 7m in length and was broadly parallel to the southern wall of the outworks. As it was relatively narrow it would be easy to dismiss it as of late post-medieval in date. We excavated a trench at its base back in 2002, which revealed three or more courses of very substantial stone foundations below ground level. This put the idea in our heads that it just might be earlier and of medieval date. Helping with this interpretation we have the earlier phase of the medieval outworks (the multiphase wall with the arched entrance through it) that still stands to a good height that is also relatively narrow in width. It remained possible that this short length of wall could be associated with this multiphase wall. If it represented an earlier version of the Postern wall it would likely extend across in front of it and all the way to the bedrock. We sited a trench to see if its line continued there below ground. Whilst this trench did produce some medieval pottery it has failed, so far, to reveal a wall or the trench from which a wall had been robbed, despite the trench being substantially extended. It is fair to say that this wall remains enigmatic and we will have to try harder next year to find some answers.
And back to cottages and towers
Whilst the wall investigation beyond the outworks was only adding to our confusion the investigations at the cottage / tower had identified four stone steps that led down into the cottage through an entrance from the base of the stairs that lead back to St Oswald’s Gate. A landing at the base of these stairs turned you round ninety degrees to the door of the cottage. Traces of the door survived as a stone door jamb on the east side, with some rather rotted timber that had formed the door frame, and a threshold stone.
Inside the threshold three further stone steps led down deeper into the cottage, that we were now realising survived to quite a depth below the current ground level. Here a further ninety degree turn pointed you towards the interior of the cottage proper, where evidence of a further door was seen in form of much more rotted timber and rusted iron nails, that marked a second door-frame.
The presence of two doorways so close together was unexpected and may be explainable if we imagine one being in use later than the other. The two plans that we have that depict the structure in some detail may show that this is the case. The later plan, Bates’ plan of 1895, appears to depict the outer of the two entrances in use and also shows the short flight of steps into the structure from the base of the steps down from St Oswald’s Gate. If we want to push the interpretation of the plan as far as we dare it also suggests that the wall at the south side of this entrance that has only the one face was just there to revet (and hold back) the mound of sand that the structure was dug into.
The earlier plan, from 1803 shows the entrance as rather different. The revetting wall was absent and a set of steps entered the structure from the south. The east wall of the entrance is hardly depicted and we might infer that the second entrance was in use then, even though it is not clearly depicted. There is no gate out to the beach on the north side and the steps and path down from St Oswald’s are shown in a different area heading down in a straight line towards the closing wall on the west side of the outwork. It is possible that this means the plan has been simplified but other evidence may support it as accurate and suggest that the route was realigned in the later 19th century (see below).
The 1803 plan also shows a set of additional steps down from within the second, and probably earlier door, leading down to the well-room. It seems safe to call it that as this plan also depicts the well-head itself within the room. This does suggest that the room here is perhaps at a basement level and excavation of the rubble fill already shows it to be more than 1m below ground level. If we are to see any part of the cottage structure as the most likely candidate to be old and part of the Tower of Elmund’s Well, this surely is it.
The final area of investigation was along the closing wall of the outworks, between the tall standing south wall and the cottage / tower. It is shown as a solid structure on the plan of 1803 and not depicted on the Bates plan of 1895, though it is on the 1st Edition Ordnance Survey of c.1860 so may have been collapsed and partly covered by the end of the 19th century. We have revealed that this wall survives in this area, close to and below ground level. Within the trenches excavated it certainly shows different phases of stone structure and in some places rubble and mortar foundations, stepping downwards to the north, towards the beach. As depicted on the earliest plans this wall does not seem to have been anything like as wide at its base as the southern closing wall, and as both were built on sand this may explain why only the southern wall remains standing to substantial height to this day. There appears to be a possible blocked opening through this wall a little to the south of the cottage / tower and this is interesting given the different line of the steps down the slope from St Oswald’s Gate shown on the plan of 1803. Perhaps this will prove to be a gate out towards the port area when we get to investigate it more next season.
Looking forward to next year the most exciting discovery that remains is to get to the floor level of the ‘well-room’ and find what remains of the well itself. How was it built and how was it lined? It must surely have been lined to have stood open for any time as it clearly was in places excavated through sand and down into the boulder clay. By uncovering much more of the surviving masonry, and some investigation of the foundations, we will hope to identify more evidence of the different phases and hopefully gain an insight into the date of some of these components. Perhaps we will even solve the puzzle of the enigmatic wall in Area A!