Dramatic changes at a grand Roman bathhouse

Why was the monumental Roman bathhouse at Silchester demolished in the 1st century AD, only to be rebuilt on an even grander scale? Mike Fulford guides us through the latest excavations at the Roman town.

Image overlooking the excavation of the baths at Silchester
Overlooking this summer’s excavations at the Roman town of Calleva Atrebatum (Silchester, Hampshire), which revealed new clues to how the settlement’s grand public baths evolved.

Just over a century ago, Edwardian antiquarians excavated the remains of the monumental Roman bathhouse at Calleva Atrebatum (today, Silchester in Hampshire). In 1905, they published a detailed plan of the baths’ layout – but they were rather less meticulous in recording their interpretations of the finds, or any evidence for their date. For the last two years, Reading University’s Archaeological Field School has been working to fill this gap, re-exposing carefully targeted areas of the structures to shed new light on their purpose and their phasing. We described the first part of this work in CA 343, and returned to the site in June and July of this year.

A plan of the site showing the 2018 and 2019 trenches overlying the Edwardian trenches
The site was last excavated by Edwardian investigators in 1904-1905. Here, the present project’s 2018 and 2019 trenches are overlaid on the antiquarian plan.

As in 2018, our chief aims were to investigate the buildings’ remains in order to gain a better understanding of their development over time, and how the materials used in their construction had changed; and to explore the deposits that had accumulated beside the structures to find out more about how they were used (and by whom!) over the baths’ estimated 400-year lifespan. To this end, over the summer we investigated two new areas of the bathing complex, as well as a section of the site’s Iron Age defences (see CA 356 for more on these earlier features).

One of our target areas lay in the south-east corner of the palaestra, an open-air exercise yard surrounded by a roofed colonnade called a peristyle. To the west, our other focus was a pair of linked rooms with hypocausts (underfloor stacks of mortared bricks, called pilae, that allowed hot air to circulate and heat the space above), which our antiquarian predecessors had described as a tepidarium or warm room. By extending this trench to the east, we gained a complete east– west transect across the middle of the building, and the results have been illuminating. Not only do we now have a much-clearer idea of how the baths evolved c.AD 45-85 (we can now propose three major phases of building, followed by some major and many minor adjustments to the last phase), we have also uncovered the remains of what may be the earliest masonry building known from the Roman town.

The remains of the earliest masonry building found at Silchester
These traces are thought to be the earliest masonry building yet found at Silchester: the sturdy brick-built foundations of a structure possibly dating to the reign of Claudius (aD 41-54). A well-preserved wood-lined drain was found against these remains.

The first hints of this early structure emerged in 2018, when we uncovered a section of sturdy brick-built foundations, more than 1m wide, beneath the robbed-out north wall of the Late Roman tepidarium. When these were re-exposed in 2019, we could see that they also continued round to the east, and that they were fronted by a well-preserved plank-lined water channel. Assuming that the overall plan of this structure was echoed by the footprint of the (much later) tepidarium, we were looking at a building measuring some 9m by 6.4m – but what was it for? For now, its function remains a mystery, but one possible clue comes from the fact that it lies right next to the substantial Iron Age defensive ditch known as the ‘Inner Earthwork’. Could this structure have been used for storing water lifted from the ditch to supply an adjacent bath building?

Intriguingly, the bricks used in its construction are of dimensions and fabric not previously recognised at Silchester – something that strongly implies that it was standing before, and respected by, the Neronian (AD 54-68) and later phases of the baths. Given that we still know so little about it, it is probably premature to think of this as part of a very early civic, as opposed to private, bath building, but its place in the construction sequence points to it being of an early, probably Claudian (AD 41-54), date. Strictly speaking, we have no independent dating evidence for it other than that its construction would have required the removal of part of the Iron Age defensive rampart – but whatever purpose and whatever community within the town this structure served, it survived as an integral part of the later civic baths complex until the 4th century, when it was demolished down to its foundations and replaced by a tepidarium.

The outline of the Claudian building with the later hypocaust that was built over it
The outline of the ‘Claudian’ building – excavated remains are shown in yellow, and projected walls in red. The later hypocaust that was built over it and was excavated in 2018 is shown superimposed in green; assuming that the tepidarium followed the footprint of its predecessor, this suggests that the early structure measured 9m by 6.4m.

When, then, was the main bathing complex built? While we were re-excavating the earliest phase of the baths’ frontage in 2018, it struck us that the remains were made entirely from materials from one source: the Neronian brick and tile works at Little London, Pamber, which lies just to the south of Silchester. Given that the building also lay at an eccentric angle, aligned not to the later street grid, but to the Iron Age ditch, this provides a powerful clue that what we perceive as the town’s first public baths date to the reign of Nero.


This picture became clearer during our latest season of excavations, when we revealed elements of the foundations of an early phase of bath building, which could be linked to the original frontage. This structure was monumental in scale, but had been almost completely demolished and replaced by a new building with an even larger footprint. With the demolition of metre-wide walls to create the new build, the scale of change was very substantial, raising the question of whether it was ever completed. At present we cannot be certain of this, as we have only explored a relatively small proportion of the baths complex (for example, we do not yet know how far south this early phase extended), but it does appear that the structure was at least nearing completion.

Photograph of the site with the Neronian bath complex outlined in red and the 'Claudian' building in yellow
The town’s first public baths are thought to date to the reign of Nero (AD 54-68). The footprint of the monumental complex is superimposed over this photo of the site in red; the ‘Claudian’ building that was found nearby is shown in yellow.
The floor of a room with bricks laid in a herringbone pattern
The Neronian baths seem to have been at least nearing completion at the time of their demolition: one room was floored with opus spicatum, bricks laid herringbone fashion.

A key clue to this was the traces of fine wall plaster that remained on the inside face of the robbed-out walls of a room on the east side of the structure (this room was extended and rebuilt with a hypocaust in the complex’s second phase), while its western neighbour had been floored with opus spicatum (bricks laid herringbone fashion). These observations suggest that at least part of the building had been floored, roofed, and finished before it was demolished. Some elements – such as the front elevation of the first phase, which was simply extended to the east and west, and part of the south wall of the palaestra – were retained and incorporated into the later rebuilding, but overall the destruction was so radical that it is not possible to identify the individual elements (for example, the frigidariumor cold room, tepidarium, and caldarium or hot room) of the baths’ first phase, other than the frontage with its fine portico and adjacent latrine block.

What prompted this very significant change of heart? The construction (or part-construction) of a civic baths building would have been a very expensive exercise, particularly given that in 2018 we saw how materials like Greensand were brought in large quantities from relatively distant sources, probably the western Weald, in preference to the much more locally available flint (which was also used in the first phase, notably in the latrine block). To then demolish whatever had been completed and build again on a larger footprint required more than just the same level of investment as before.

We can only speculate as to the reasons. For example, if the project was commenced in the early part of Nero’s reign, was it interrupted by the Boudican rebellion of AD 60/61 and then suspended in its aftermath as other more pressing needs were addressed? Or, if its construction had been intended to promote renewed confidence in the province after the rebellion, was work curtailed by Nero’s death and the confusion of the civil war of AD 68-69? How then would this building, we assume unfinished, weather several years of neglect before work could be restarted? Perhaps its condition required a new start? Given its low-lying position within the town, it may have suffered damage from flooding – at the very least, the new build took the opportunity to raise the complex’s floor levels by about 0.5m.

Trench 4 hypocausts
An overview of the Trench 4 hypocausts which vividly illustrates the complete transformation that the site underwent.

Calleva lay within a client territory ruled (with Roman consent) by Cogidubnus. On the far side of his kingdom, just outside what is now Chichester in West Sussex, we find a parallel with the ‘proto-palace’ at Fishbourne, which was started in Nero’s reign but then abandoned (perhaps also unfinished), demolished, and replaced by a much larger and more impressive building, the Flavian ‘palace’. A further connection between the baths at Silchester and the ‘protopalace’ is their shared use of the same building material, Greensand, in neatly shaped blocks. We are no clearer as to the reasons for the radical changes put in place at Fishbourne, but might the shared history of the two projects suggest a common explanation?

This is an extract of an article that appeared in CA 358. Read on in the magazine. Click here to subscribe

Leave a Reply