6 October 2014
Pompeii of the North:
Excavations at Bloomberg London
and the Return of the Temple of Mithras
Today I am going to talk about archaeological excavations on the Bloomberg London site, formerly known as Bucklersbury House. I was lucky enough to work on the project for several years and am now involved in the analysis and publication. I will give a brief summary of how Bucklersbury House came to be the site of one of the most important discoveries of the twentieth century and then give you some news of our recent work; with the analysis still ongoing there are, of course, still many aspects of the archaeology which require further study but we can already say that it was a record breaking project in many ways and the artefacts and structures recovered will greatly enhance our knowledge of Roman London, and our place within the wider Empire as a whole.
So why was the site dubbed ‘The Pompeii of the North?’ Although this was largely a media invention, there are similarities between the two sites, namely the staggering level of preservation, and the internationally significant amount of Roman material culture revealed. There are also differences of course: the preservation conditions were not the same, as Bloomberg London was a wet site rather than dry. Neither did the Bloomberg site represent one day in time as at Pompeii but instead 250 years of the Roman period. However, although the comparison may have been slightly over-exaggerated, we are nevertheless very proud to have worked on ‘London’s Pompeii!’
Firstly, it is useful to remind ourselves of how MOLA came to excavate the site again recently. Buried archaeology in the UK is protected within the structure for planning legislation, which requires the developer of a site (whether for a road, railway, basement or major City office) to allow sufficient funds and time for the archaeology to be excavated in advance of the development. Usually the only archaeology that is excavated is that which would be directly impacted (or destroyed) by the development, so often there will be a great deal more protected below the new building. This is the case with the Bloomberg site, where we undertook three years of excavations across the site, working closely with the demolition contractors to enable work to progress in many areas at once.
The story of discoveries on the site really starts in earnest after WWII, when the City was destroyed by bombing, leaving large empty bomb sites. The Royal Society of Antiquaries set up the Roman and Medieval London Excavation Committee, led by Professor W F Grimes, to excavate trenches across many of these sites in the years after the War. On Bucklersbury House the RMLEC were intending to study the Walbrook Stream, known to have allowed excellent archaeological preservation. In one of the trenches, located entirely fortuitously as it was the only basement that was not full of bomb rubble, the team found a building with an apsidal end (curved wall). They were given permission to return two years later when Grimes returned to conduct a more extensive excavation, although the archaeologists were only granted a month. Gradually more of the ground plan of this building was uncovered, and on the last day of the excavation the marble head of Mithras was revealed.
A furore ensued, in the media, and amongst the public who visited in their droves. There was also a political furore too with Winston Churchill recalled from his weekend at Chequers to discuss the forthcoming destruction of the Temple.
After much wrangling it was agreed that the Temple fabric would be dismantled and removed from the site to a builders’ yard in New Malden, Surrey. It was to return to the site eventually, as a reconstruction. Notably much of the original stone went missing from the yard, and during the reconstruction plans the archaeologists such as Grimes were excluded from discussions as to how the Temple would be rebuilt. The eventual result was not to Grimes’ liking!
If we now move to the (almost) present day it is thanks to the City of London and English Heritage that we have a condition placed upon the planning permission of the new development –that the redevelopment project should not only include the excavation and publication of buried archaeology but also that it would incorporate the careful dismantling of the 1961 reconstruction and a new reconstruction be built in a purpose-built display space.
Our fieldwork ran over a period of three years, with teams of various sizes, up to a maximum of 55 archaeologists working in the main area at the northeast corner of the site. It was record-breaking for many reasons: the deepest archaeology in London (seven metres of compacted, waterlogged deposits; almost three tons of Roman pottery; over 11,500 registered finds; 3,500 tons of soil removed by hand).
The site straddles the Walbrook stream, an ancient river that was the main topographical feature of the Roman town. It was also crossed by Watling Street, one of the principal Roman roads through South East Britain. The volume and variety of the artefacts recovered from the waterlogged deposits found in the eastern part of the site constitute an assemblage of international significance, covering the period from the founding of Londinium through to the 3rd century AD. The waterlogging of the site is attributable to the Walbrook stream. These conditions also ensured the preservation of timber buildings, fences and floors.
The earliest Roman activity was a series of banks and ditches built within the Walbrook floodplain. The enclosures appear to have been built deliberately as close to the stream as was feasible. The assemblage of artefacts from this phase is significantly rich in military objects providing a link to the Roman army.
Next in the site sequence were significant make-up dumps, often retained within box-revetment structures known as cribs, and timber buildings constructed over these boxes. There was excellent preservation of the structural timbers. Later building phases were of typical waterfront construction, with oak piles below timber-framed buildings with clay walls, some of which had painted wall plaster decoration. A circular building in the middle of these rectangular buildings was a bakery, or large oven. The size of this structure indicates some degree of municipal organisation, as clearly it does not relate to a small residential building but was intended to provide food for a significant proportion of the local population.
Generally the timber buildings were built on a series of low terraces descending towards the Walbrook. On the floor surface within one room of one of these buildings were a number of writing tablets, styli and one seal box; it could be conjectured that this room was an office.
A series of oak piles in trenches at the south of the site were situated along the line of Watling Street: a Roman road known to have crossed the site. A bridge would have been necessary to carry a major road such as this across the Walbrook in this location, and additional piles that may be part of an abutment were excavated from a trench in the far south eastern corner of the site, also along the line of Watling Street. In addition, several small timber structures were found on the site, located along the length of the Walbrook in various trenches, which may have been temporary structures providing access across the stream at various points.
By the early 2nd century AD, the site was occupied by a particularly large timber building which was very likely to be industrial in function and could be a mill but further analysis is required before this interpretation is secure. The building was again constructed on oak piles with clay walls laid onto large oak base plates. A fine timber plank floor was also exposed. There are two very rare survivals of early Roman machinery in oak timber from the landfill associated with this building. One was a probable mill wheel, paddle blade of oak sculpted from a sawn oak plank, unique within the Roman Empire. The other item was part of a lantern gear essential to drive the rotating shaft for the mill stones. This was a disk of radially cleft oakwhich was pierced by six round holes for gear bars and a square axle hole for the drive shaft. A timber-lined drain or culvert ran to the east of this building, possibly providing power to a water wheel, but this remains unproven. Further comparison of this building to known Roman water mills will obviously be crucial to understanding its function, as will a closer look at finds assemblages, environmental information and the varying levels of the natural water table at the end of the 1st century AD, as a mill would undoubtedly require a significant supply of water running at a relatively rapid rate to power its wheels and other mechanisms. If this building does represent the remains of a water mill it would have had two storeys, with the ground floor’s timber planking providing easy removal of water flooding in from the wheel pit to the east. The mill machinery would have extended through the external east wall and joined to the wheel. The mill stones would have been on the first floor, which would have been drier and therefore preferable for the grinding of grain and other foodstuffs.
Timber drains crossed the site throughout the Roman period and these were probably a combination of freshwater supply, industrial run-off and foul water drainage. Environmental analysis of their fills will elucidate this further. The drains included a wide variety of construction techniques, such as lidded timber box drains, drains integrated into building plans, hollowed-out logs, lead linings to timber drains, a possible gutter drain, drains forming right angles and extending around open areas, and drains lined with reused ceramics. Several of these appear to link with each other across wide areas. On initial investigation it appears these drains were associated with individual buildings and their differing construction materials and techniques may be due to the management of drainage having been the responsibility of individual property owners.
This building and several others on the site were destroyed by a fire sometime during the early second century AD, possible the Hadrianic Fire of 120/5 AD, though whether this event was genuinely city-wide is unknown and requires further study. Rebuilding after the fire on the site concentrated upon domestic properties, often with more robust stone foundations, supporting timber walls. Two of these buildings had small ovens to the rear and external tessellated floor surfaces, possibly from porticoes, gardens or similar features.
Later second and third century AD stone buildings were heavily truncated, although some of these buildings were contemporary with the Temple of Mithras, which was constructed around 240 AD. Although our main phase of excavations was not concerned with the Temple, we had uncovered the surviving remains during earlier evaluations. This involved careful re-recording of the sections through the eastern wall/foundation of the Temple drawn by Grimes and Audrey Williams. Sections of the masonry foundations for the north and south sleeper walls and the south external wall of the temple were revealed. The foundations were resurveyed which allowed the location of the temple in relation to modern buildings to be accurately determined. The cuttings made in the 1950s were identified, with the east end of the main 1954 cutting found more or less where predicted. Unexpectedly, an additional North-South trench was also identified, which further study of the archive has indicated was recorded by Grimes during a later visit to the site in 1956.
In addition, further masonry foundations were exposed during our 2011 works. Remnants of five Roman-period walls, which either adjoined the temple or formed part of an extended building complex around it, were revealed. These may relate to the Narthex, the ancillary structure or porch at the West end of the Temple.
Located to the north of the Temple was a timber box-frame well, constructed during, or shortly after, the late third century AD. The fills within the well included organic-rich lower fills. A substantial chalk foundation surrounding the well may have been a later addition related to a structure or building enclosing the well (ie a well house). The well’s final disuse may have occurred in the later fourth century AD, or even in the early fifth century. The basal fills of the well contained several cow skulls, in a recognised Roman ritual of ‘closure’ (thanking the Gods for the water and shutting the well at the end of its use). Further fills contained several other ritually deposited items, including a hoard of fine pewter tableware vessels, a military belt buckle and several large pieces of lead, parts of which may originally have been a water tank.
The survival of organic materials such as leather, textile and wood is particularly noteworthy and important. In Roman Britain, the quantity of well-preserved textiles from Bloomberg London is only exceeded by that from Vindolanda. It far exceeds the number of textile finds from other Roman town in Britain. Consequently, the textiles have an extremely high potential for furthering all aspects of study into Roman textiles in Britain and the Northern Roman Empire.
The 410 fragments of writing tablets constitute a hugely important assemblage, a body of correspondence of international importance with parallels such as the inked tablets from Vindolanda and the wax tablets from Vindonissa (Switzerland). About 90–100 tablets are likely to carry some legible writing and potentially will include some of the earliest documents ever found in Britain. Examination of the tablets is ongoing and though conventional photography – taking multiple images each lit from a different direction – is most likely to prove successful, techniques with ultra-violet light will also be trialed.Palaeographically they have enormous research potential. They constitute a large assemblage, earlier in date than the Vindolanda ink tablets and inscribed in a different alphabet of letter forms, since since this varied according to the medium.
On a general note regarding the artefactual assemblage: the sheer quantity of finds from the Roman period will enable new statistically sound theories to be proposed, in particular regarding the creation or adaptation of chronologies, for specific classes of artefact (brooches for example).
Now we return to the story of the Temple itself, and look at the dismantling of the reconstructed fabric, using diamond-tipped chainsaws, over the cold winter of 2011/2. The threshold stone was the only stone that could be said was definitely in its correct original position and coincidentally was the only stone positively identified amongst the rubble lying in the grass in the New Malden builder’s yard in 1960. This stone was planned by us and lifted very carefully onto pallets to be taken to the new (safer) storage area. Several sections of the concrete infill included 1960s bricks, various hard core and some surprises such as medieval column bases.
The new reconstruction will aim to recreate the first phase of Temple dated to AD240, unlike the original reconstruction which combined several phases. We hope that some of the most important or interesting objects from our recent excavations will be exhibited in the new display space and we are working closely with the designers to provide information on the site within Roman London as well as on the Temple itself.
As well as the artefacts themselves we hope to be able to provide some memories of one of London’s most famous archaeological stories, and are working with Bloomberg on an Oral History project, collecting and celebrating the experiences of people who went to marvel as the Temple was exposed. Anyone who visited the site in 1954 is urged to contact us and share their stories. The contact details are firstname.lastname@example.org; 0207 410 2266.
If you would like to read further information about the archaeological discoveries on the project, meet members of the team involved, or learn about our methodologies then please do have a look at our ongoing blog: The Walbrook Discovery Programme, at
© Sadie Watson, 2014