The Military Town and Bathhouse of Birdoswald Fort on Hadrian’s Wall

Introduction

Birdoswald is one of the most thoroughly investigated forts on Hadrian’s Wall, with 21% of the interior of the fort excavated and published since 1987, and a long history of research before that date. The fort cemetery was also the subject of a major excavation in 2009. This means that there is a clear understanding of the various phases of occupation and their dating from the foundation of the fort in the Hadrianic period (around 120- 138AD) to the 5th century AD, and probably beyond.

The parts of the complex which remained virtually unexplored were the extramural areas. Geophysical survey undertaken by the late Alan Biggins and David Taylor in the late 1990s revealed just how extensive and intensive settlement in these areas was. This survey, followed by their similar work at Maryport, revolutionised our view of the size and complexity of such settlements throughout the Roman frontier zone. What was missing was detail, which could only be recovered through excavation.

The work described here was a joint project between Newcastle University and Historic England, designed to excavate areas to the east, west and north of the fort to establish the relationship between phasing and activity in these areas and the fort interior, and to examine the types of activity that took place. The project, which ran seasonally from 2021 to 2024 (with a double season in 2021 because of the cancellation of 2020 due to COVID) was also a training excavation for Newcastle undergraduates, some 250 young people gaining excavation experience.

The north

Two areas were excavated to the north of the fort, and therefore to the north of Hadrian’s Wall. One (C) , on the berm of Hadrian’s stone wall, proved that this sector did not feature the obstacle pits found on the berm in the eastern 12 miles or so of the Wall.

The second (D), led by the geophysical results, showed occupation, featuring clay sills for timber buildings. A large material culture assemblage indicated that domestic and industrial activity took place here, and one of the buildings was identified as a smithy.

This should not cause surprise; the Roman frontier was a zone, not a line , and a road, the Maiden Way, ran from the north gate of Birdoswald to the outpost fort of Bewcastle, seven miles to the north, and these structures were probably associated with this road.

The west

To the west, a trench (E) measuring 60 x 15 metres was excavated 130 metres from the fort, where geophysical survey suggested that the road from the west gate expanded into an oval open area resembling a medieval marketplace with structures on each side . Excavation confirmed this basic layout, but occupation north and south of the space differed. To the north were long, narrow stone-founded strip buildings which probably originally had timber superstructures. Their narrow ends faced onto the street. To the south a large timber building was succeeded by a complex of industrial kilns. The date of this change of use has not yet been established.

The east and the bathhouse

Two areas were excavated to the east. One (B) was sited to enclose a single strip-house property, identified from the geophysical survey and facing onto the road emerging from the east gate. This gate had been excavated in 1855, establishing that the northern portal had been blocked in antiquity. The excavated building had two phases; the first respected the primary wide road, but the second fronted onto the narrowed road associated with the blocking of the portal, which was probably contemporary with the blocking of a portal of the west gate during the first third of the third century. The building was likely to be commercial in function, with a prime position on the street and near the fort.

The final area (A) was the most complex and productive of new information. Like all other areas it was sited with relation to the geophysical survey but was also deliberately sited to examine a substantial building first discovered in 1932 by Sir Ian Richmond. He had described a stone building 20 feet square, standing 13 courses in surviving height. He was unable to continue beyond this observation before being inundated by an inrush of ground water. Subsequently this structure was very tentatively interpreted as a signal tower.

The building proved to be part of a large building complex with several phases and structures, and its depth and stratigraphic complexity meant that this part of the excavation was returned to and continued during all four excavation seasons.

It was not until the third season of excavation that it was possible to demonstrate conclusively that Richmond’s building was in fact part of the fort bathhouse. Importantly the excavated area exposed the stoking chamber, the furnace, and part of the hypocaust of the hot bath (caldarium).

The construction of this facility was carefully planned. First a deep trench was cut in which a substantial stone lined drain was laid in order to take away ground water. This drain ran under the floor of the stoking chamber, which was itself semi- subterranean, with just over a metre of the walls beneath the contemporary early Roman ground surface. This allowed space for the stoking of the furnace providing sub-floor heat to the hot room of the baths and allowed the heated floor to be level with the exterior gravel surfacing.

Uniquely in the Roman Empire, the iron fire beams provided over the flue survived intact. These were over a metre long and 150 millimetres square in section and were immensely heavy. In use they would have supported a semi-cylindrical bronze or lead boiler called a testudo (not to be confused with the Roman shield formation), which would have been closed at the fire end, but open to the hot bath (alveus). Water within the bath and testudo would have circulated over the fire, reheating the bath water. The bath itself was supported on tile stacks (pilae), and the walls were heated by way of ceramic flue tiles (tubuli) set in the walls.

Water supply

To the north of the bathhouse was a row of buildings of unknown function, which seem to have been frequently remodelled. Along the western edge of these buildings ran a narrow trench in which were found the iron rings which would have connected lengths of timber water pipe. This split into two branches north of the bathhouse, one of which ran off to the east. The main line of the pipe, however ran directly to a small stone-lined shaft against the wall of the bathhouse. This shaft led through the wall and was clearly the means by which water entered the building. Piped water would accelerate as it dropped into the shaft and would probably have been siphoned to the level required by the boiler though a system of now-vanished metal pipes. Full analysis of the engineering possibilities remains to be considered.

For piped water to be supplied, a consistent source would be required most probably led to the site by an aqueduct. There is suggestive evidence, in the shape of a primary culvert in Hadrian’s Wall that such a facility was provided from the beginning of the fort’s history, though this suggestion needs to be further pursued.

The end?

Limited excavations in the extramural areas of other forts have generated the idea that these settlements were abandoned in the later 3rd century. One aim of the present excavation was to establish whether this applied to Birdoswald as well. Despite the extent of the excavation, nothing later than the later 3rd century AD was found. This indicates that, despite the size of the military town, it was abandoned, leaving the fort to remain in occupation for at least a century and a half. The bathhouse itself was demolished to a consistent level in antiquity, and the flue and stoking chamber deliberately filled with demolition debris. Analysis of these intriguing aspects have only just begun.

The impact of the research

Further information

Wilmott, T 1997 'Birdoswald: Excavations on a Roman Fort on Hadrian's Wall and its successor settlements, 1987-92', English Heritage Archaeological Report, 14 (Swindon)

Wilmott, T 2001 Birdoswald Roman Fort; 1800 years on Hadrian’s Wall, (Tempus, London)

Wilmott, T, Cool, H and Evans, J 2009 ‘Excavations at the Hadrian’s Wall fort of Birdoswald (Banna), Cumbria: 1996-2000’, in Wilmott, T (ed) Hadrian’s Wall: Archaeological Research by English Heritage, 1976 – 2000 (English Heritage: Swindon), 203- 395

Wilmott, T forthcoming: The cemetery of the Hadrian’s Wall fort at Birdoswald, (CWAAS Research Series)

Download Issue 27 of Historic England Research as a PDF magazine

You can download this article and the other articles in this issue as a PDF format magazine.