Part 5

Destruction and Recovery

19

The Hadrianic fire (c. ad 125–35)

The burning of London

Soon after Hadrian’s visit to Britain in ad 122, London was overwhelmed by a catastrophic fire. This left a distinctive horizon of fire-reddened daub and destruction debris that can be traced across most of the settlement (Fig. 19.1). ¹ It may have been an entirely accidental event like the Great Fire of 1666, but it is also possible that London had been razed in a largely forgotten British war. Evidence of fire damage is particularly intense around London Bridge, coincidentally close to the Pudding Lane bakery where the Great Fire began. ² To the west the Walbrook had failed to act as a fire-break, and burnt debris can be traced either side of the river’s lower reaches. ³ The horizon is not evident further upriver along the Thames waterfront, perhaps because the area was under-developed after the Huggin Hill baths were closed. ⁴ Otherwise destruction engulfed roadside buildings to the westernmost fringes of town, if not to the banks of the Fleet where the quays and mill at the river mouth escaped damage. ⁵

Fig. 19.1 The likely extent of destruction caused by the Hadrianic fire of c. ad 125/6 based on observations of in-situ fire debris. Drawn by Justin Russell.

On the northern side of town the fire skirted the amphitheatre, but left no trace within the arena itself. ⁶ This may have been because the timber superstructure had already been dismantled awaiting rebuilding. More probably, the Hadrianic rebuilding here was necessitated by partial fire-damage, and much of the superstructure could have been burnt without leaving archaeological trace. In either case, the amphitheatre marked the boundary of the built-up area and the northern limit of fire damage. Further to the east fire debris can be traced as far north as the settlement then extended, including areas of suburban development along Bishopsgate. ⁷ Buildings flanking the forum were also burnt although it is not certain that the basilica was destroyed. ⁸ This was damaged by an early second-century fire, but it has been suggested that this was a later event. ⁹ Eastwards the destruction reached Plantation Place and Mark Lane, but not as far as the Tower where there is no certain evidence of occupation before the late 120s. ¹⁰ It also extended north-east along Fenchurch Street as far as Lloyd’s Register and St Katherine Coleman, possibly to 3–4 Jewry Street. ¹¹ A few buildings on the eastern margins of the settlement, set back from street frontages, show no direct evidence of fire damage. ¹² Fire debris is also generally absent from the upper Walbrook valley and Cripplegate fort, which areas may not have seen significant development until after the fire. ¹³ Only a few buildings within the area of destruction show no clear trace of fire damage, but these are isolated examples where the subsequent rehabilitation of partly ruined buildings may have cleared away destruction debris. ¹⁴ Taken collectively, the evidence indicates that little of London north of the Thames survived, leaving a smouldering wasteland over some 64.5 hectares.

The situation in Southwark was different. There is no evidence of a general conflagration, although a couple of buildings may have seen fire damage at about this time. ¹⁵ The high-status building complex at Winchester Palace was comprehensively rebuilt c. ad 120/130 and the floor of one of the structures replaced at this time, a possible granary, was covered by an unusual charcoal layer. ¹⁶ In the absence of a more substantial destruction horizon, Brian Yule suggested that this was a deposit of soot. On the other hand, the horizon included window glass distorted by fire, and subsequent levelling dumps included burnt material. It is possible that the charcoal was a residue from burnt roof timbers, generating a distinctive destruction horizon in the absence of burnt mudbrick and daub.

London’s Hadrianic fire is conventionally dated c. ad 125 by a massive warehouse assemblage of fire-damaged Samian found at Regis House. ¹⁷ Destruction at One Poultry also included numerous potter’s stamps from Samian made at Les Martres-de-Veyre dated c. ad 105–25, and a Pulborough Samian vessel made from a Lezoux mould dated c. ad 125–50. ¹⁸ Dendrochronological evidence from timber structures damaged by the fire confirms, but doesn’t refine, the dating suggested by the pottery. ¹⁹ Archaeomagnetic samples from Regis House have also provided dates of ad 110–30 and 130–80 for the fire. ²⁰ These sources combine to confirm the dating of the fire to the period ad 125/130. A very slightly earlier date is not wholly impossible, but can be considered highly improbable.

Post-fire reconstruction appears to have been underway by c. ad 128. Dumps associated with the replacement of short-lived Hadrianic quays investigated at Suffolk House included scorched and fire-sooted pottery dated ad 120–60, with in-situ fire debris nearby. ²¹ The waterfront had probably been rebuilt following the fire and timbers used to line a box-drain within the new harbour facilities were felled ad 128. Although we cannot discount the possibility that these timbers were reused, this is improbable. Most of the timber used in rebuilding London’s harbour would have been procured for the exercise, and since felled after the fire should provide a terminus ante quem for the event. A substantial jetty was also built over the Thames foreshore at the Tower of London using timbers felled in the winter of ad 126–7. ²² This was an unusual location for waterfront activity, since the site lay some distance downstream from the settlement, in an area that was only built-up later in the second century. It is tempting to suggest that the jetty was a temporary facility for ships coming up-river whilst the port was in ruin.

In sum, the fire is dated c. ad 125/130, and post-fire rebuilding was almost certainly underway in ad 128 if not by the spring of ad 127. This establishes a strong case for dating the fire to ad 125–6. This is also the earliest date that easily accommodates the evidence of fire damage in the north range of the forum basilica. This structure wasn’t built before the Hadrianic period, and was perhaps completed in time for the imperial visit of ad 122 (above p. 165). ²³ The building was subsequently modified, involving the replacement of the nave wall with brick piers, before it was destroyed in a fire of Hadrianic date. The road built alongside the basilica’s north range had been resurfaced twice before this fire. If the basilica was completed ad 119/122, and a minimum of 2 or 3 years elapsed between each road repair, it is hard to date the fire much earlier than c. ad 125.

This review of dating evidence is complicated by the fact that we cannot be certain that London only suffered one Hadrianic fire. A report on excavations at 10 Gresham Street proposed two successive periods of activity ending in Hadrianic conflagration, both dated c. ad 130 by assemblages with diagnostic forms in Verulamium region white-ware. ²⁴ The presence of successive fire horizons is not, however, securely demonstrated by the published evidence. Clearance operations may have resulted in fire debris being redeposited, converting a single event into several archaeological horizons. Elsewhere in London there is overwhelming evidence for a massively devastating conflagration that was capable of leaping 8-metre-wide fire-breaks presented by the Walbrook and wider streets, and there is good reason to see this as a single event dated c. ad 125/126.

Accident or arson?

Cautious scholarship has generally preferred to treat the Hadrianic fire as a terrible accident. ²⁵ In such a fire, however, one would expect to find buildings upwind of the fire’s starting point that escaped destruction. This doesn’t appear to have been the case, and the comprehensive nature of destruction along areas of ribbon development leading into the city hints at deliberate action. London’s Hadrianic destruction also gave rise to unusual post-fire interventions which mirrored those which took place after the rebellion of ad 60/61. Activities after the Boudican revolt included the construction of a fort, the building of a track across the Walbrook valley to bypass the town centre, a scatter of fragmentary human remains on the city borders suggestive of retributive corpse abuse, and the heavily engineered enhancement of the port infrastructure (pp. 99 and 110). An identical pattern of activity can be described from London’s Hadrianic archaeology. This embraced the building of a fort, the engineering of a road to bypass the town centre, a spike in the volume of human remains scattered on the city borders, and the rebuilding of the port and public buildings. The details of this archaeological ‘signature’ are considered in the following chapters, but these similarities suggest a common context for both Neronian and Hadrianic rebuilding.

This encourages us to explore the possibility that London was destroyed by war c. ad 125/126. The very idea of a war at this time runs counter to the view, drawn from Edward Gibbon’s confident opinion, that ‘the empire flourished in peace and prosperity’ during Hadrian’s reign. ²⁶ History may, however, have been over-kind to Hadrian. Although we have no detailed sources for political developments in Britain during the second century there is the clear testimony of a letter written to the emperor Marcus Aurelius by his tutor Cornelius Fronto in ad 162, which refers to a British slaughter of Roman soldiers during Hadrian’s reign that stood comparison with the disastrous war in Judea. ²⁷ This was a serious matter: according to Cassius Dio, 580,000 men were slain during the Judean revolt. ²⁸ A late fourth-century life of Hadrian also refers to the difficulties he met in keeping Britons under Roman control. ²⁹ These sparse references are sufficient to suggest the historical reality of a British revolt during Hadrian’s reign, but tell us nothing about where, when or why it happened. Until recently most scholars were inclined to see this as an event preceding Hadrian’s visit to the province and confined to the northern frontier. ³⁰ Inscriptions describing the military careers of Maenius Agrippa and Pontius Sabinus encourage a slightly different reading of events. These officers participated in an expedition to Britain likely to have been mounted to quell the revolt. Other details in Sabinus’ career path make it difficult for this expedition to have taken place before ad 124. ³¹ Events along Hadrian’s Wall also suggest some form of crisis c. ad 123/4. A tombstone from Vindolanda commemorated a war casualty of the period, and the building of the wall was interrupted and modified. ³² John Casey has drawn attention to successive coin issues from the mint at Alexandria that appeared to celebrate Roman victories obtained in ad 124/125 and ad 125/126, which he argued were won in the British campaign. ³³ The celebration of victories in successive years suggests a prolonged war, involving a setback or different theatres of action. The second victory, obtained in the months prior to September ad 126, is appropriately dated to have been won in a war that encompassed London’s burning in ad 125 or 126. ³⁴

There is no direct proof that other parts of southern Britain was affected, although Hadrianic fire debris found in Staines and Brentford might chart a route of destruction west of London. ³⁵ It should come as no surprise, however, that London would have been sacked if Rome lost control of the province. The city’s role in raising tax and rent, as the seat of government, and as the wealthiest place in Britain, made it a prime target.

London was in the midst of change at the time of the fire, perhaps following reforms made on Hadrian’s visit. Change was also evident in the surrounding landscape which was extensively farmed into the second century ad. ³⁶ Zoological and palaeo-botanical remains suggest a greater emphasis on cultivating spelt wheat and cattle farming, accompanied by improvements in cattle livestock. ³⁷ Farmers appear to have been maximizing agricultural production, perhaps in response to more elevated demands of tax and rent. Wheel-thrown black-burnished vessels produced on both sides of the Thames estuary started arriving in London early in the Hadrianic period, probably by ad 120, contributing to the archaeological definition of a new period of ceramic supply. ³⁸ These vessels also appear in quantities on Britain’s northern frontier in the following decades, witnessing a marked increase in traffic along the east coast. ³⁹ It has been suggested that some of this was carried by the Classis Britannica, since the fleet was directly involved in building work on the wall and must have been an important presence along the coast. These pots were probably moved such distances because they contained valued commodities, probably including salt and salted produce. The use of black burnished pots to package such goods imitated earlier production in Dorset, which packaged salt or salted goods from an area likely to have been under military control. The east coast trade drew on a significant increase in production within the Thames estuary. New kilns either side of the river, in south-east Essex and northern Kent, ramped up their seasonal production of black-burnished vessels. ⁴⁰ Excavations at Dagenham Farmstead illustrate the introduction of new pottery kilns, involving radical new technologies and production techniques, c. ad 125. ⁴¹ These changes within the Thames estuary are likely to relate to new lines of supply associated with the building of Hadrian’s Wall, the garrisoning of which relied on taxes and rents raised elsewhere. ⁴² Production responded promptly to new military demand in the north, perhaps managed by the procurator’s London office. We can speculate that the increased exactions implied by this evidence might have provoked tax rebellions similar to those of the revolt of ad 60/61. ⁴³ Unrest may also have followed the example of the rebellions in Judea, coming to a head after Hadrian’s tour of the western empire was broken off in ad 123 when an emergency is believed to have summoned him east. ⁴⁴

The Cripplegate fort

If London had been sacked in a British revolt this might explain why a new military garrison was subsequently stationed at the north-west corner of town. Excavations by Peter Grimes after the Second World War traced the plan of an imposing early second-century stone-walled fort (Fig 19.2). ⁴⁵ The defensive perimeter was formed by a high stone wall backed by an earthen rampart, fronted by a V-shaped ditch about 3 metres wide and 1.5 metres deep. A second, outer, ditch may have been dug along its western side. ⁴⁶ Peter Marsden’s excavations next to the Guildhall extension found evidence of two phases of fort construction since the original ditch was comprehensively backfilled soon after it was dug, presumably because the associated earthworks had been slighted. A new ditch of identical profile was then dug along the same line to restore the defensive circuit, before this ditch was in turn silted up c. ad 140–80. ⁴⁷ In sum it appears that plans for the construction of the fort were cancelled before being swiftly revived.

Fig. 19.2 The second-century Cripplegate fort (after Shepherd 2012). Some detail of ill-preserved buildings immediately to the north of those shown is omitted. Drawn by Justin Russell.

The corners of the fort were reinforced by internal towers and smaller rectangular turrets set along the line of the perimeter wall. ⁴⁸ The west gate, now preserved within the London Wall car-park, consisted of a double portal flanked by square towers, with each carriageway about 2.75 metres wide. ⁴⁹ Some of the stones used in its construction were salvaged from an earlier monument, probably a tomb. ⁵⁰ Other gates would have been placed at each of the fort’s cardinal points. The fort measured 220 by 215 metres, covering an area of c. 4.7 hectares, which was less than a quarter of the size of contemporary legionary fortresses. A recent survey of contemporary garrisons suggests a mean density of about 260 men per hectare making the Cripplegate fort the right size for a force of about 1,200 men. ⁵¹ The fort’s interior is incompletely studied, but was laid out around a T-junction, with a principal entrance at the south. A 5-metre-wide intervallum road followed the inside of the fort’s wall and a grid of lesser streets (via vicinaria) separated barrack blocks and other internal buildings. Parts of eight timber-framed buildings set over masonry dwarf-walls have been identified within the southern part of the fort (praetentura). Assuming that the internal layout was essentially symmetrical, this allows us to extrapolate the presence of eighteen or twenty-two buildings in this southern area. ⁵² Most were rectangular, around 52 metres long and up to 10 metres wide, containing rows of rooms fronted by a corridor. Separate mess-groups are assumed to have inhabited each room, combining to provide housing sufficient for an infantry century of about 80 men within each block. Other, perhaps shorter, buildings were aligned east west. The different design suggests a different purpose, and these may have housed smaller cavalry units (turmae). Neither style of building appears to have included officers’ quarters. Objects associated with the buildings include armour fittings from lorica segmentata, of the sort principally used by legionaries. The range of finds also suggests the presence of auxiliary cavalry and the fort is likely to have been garrisoned by a mixed unit. The rear of the fort is largely unexplored but likely to have included further barracks with working and storage facilities.

A Hadrianic date is indicated by Samian from within the fort’s earthen ramparts. Hadrianic fire debris hasn’t been identified inside the fort, although layers of charcoal and burnt daub were recorded beneath its bank. ⁵³ It is therefore likely that the fort was built after the fire, although an earlier date cannot be discounted. The fort was both larger and more substantial than that built after the Boudican revolt, following the fashion for stone building that characterized second-century public architecture. It presented an argument of permanence and control that must have carried particular resonance in the aftermath of the fire. It was also an unusual feature, since urban garrisons were exceptionally rare. A cohort had been stationed at Lyon to protect the mint, and Vespasian based another at Carthage to secure annona supplies. ⁵⁴ These stations protected Rome’s interests at places of exceptional strategic importance. London’s stone-built fort might indicate that it too belonged to this select group of imperial cities. ⁵⁵

It has alternatively been argued that the Cripplegate fort housed soldiers serving on the governor’s staff and bodyguard. ⁵⁶ These functions might have employed some 1,000 men, divided between infantry and cavalry. The fort was slightly over-large for such a force, but the additional space might have accommodated other military personnel serving in London. This requires us to assume, however, that soldiers previously garrisoned elsewhere in London were relocated into the fort when it was built. If this was the case, we ought to find new centurions’ quarters to match the town houses where these officers previously lived (p. 142). There is no such evidence, and senior personnel serving with the Governor are unlikely to have moved into the fort. The soldiers stationed within the fort were regular troops rather than senior officers and quartermasters. It is also reasonably certain that the fort was abandoned at a time when the governor and his entourage continued to be London based, showing that the fort wasn’t necessary to this arrangement. ⁵⁷ There must have been an important strategic or political reason for building the fort at this point in time, and the Neronian fort built after London’s destruction in ad 60/61 offers a clear precedent. The Cripplegate fort could have been built in similar response to political circumstance, involving the settlement of a new garrison after revolt rather than in anomalous administrative display.

The only other fort known to have been built in southern Britain during Hadrian’s reign was attached to the base of the fleet, the Classis Britannica, at Dover. ⁵⁸ There are important similarities between the Cripplegate fort and the Dover fort. The construction of both appears to have been interrupted, perhaps in changing strategies similar to those noted on Hadrian’s Wall. The likelihood that Dover’s fort was built by the fleet deserves particular attention, given the contemporary use in London of rare examples of tiles carrying the CL.BR stamp produced in the fleet’s Sussex tileries. These bases would not only have helped secure and police the supply routes, responding to vulnerabilities exposed in earlier conflict, but brought new manpower and capacity to the operations. Seen in this light the Cripplegate fort should not be seen as simply a defensive feature, but as a hub for marshalling troops and supplies. ⁵⁹

Post-fire rebuilding

London’s public buildings were rebuilt promptly after the Hadrianic fire. This is the likely context for the construction of a splendid new masonry amphitheatre: the largest such building in Britain and able to accommodate over 10,000 spectators (Fig. 19.3). ⁶⁰ A new oval arena was enclosed by a masonry wall surmounted by an iron railing, which retained an earthen bank supporting rising tiers of wooden seats. The eastern entrance was rebuilt as a 7-metre-wide tunnel, flanked by two small chambers with sliding doors to release wild beasts into the arena. ⁶¹ This ambitious rebuild followed the installation of the new garrison in the Cripplegate fort and may have been necessitated by fire damage to its Flavian precursor. Like the fort, it proclaimed a restoration of Roman order in the wake of London’s destruction. One of three small inscribed lead tablets found in later destruction addressed a curse to the hunter goddess Diana, often associated with the Roman amphitheatre in the guise of Nemesis, who is likely to have had a shrine here. ⁶² The cult was linked to both city protection and retributive justice exercised killing of criminals and beasts, celebrating the supremacy of Roman authority. The forum basilica may also have been restored, reusing its pre-fire walls, and a portico paved with a herringbone-patterned tile floor added along the east side of the complex. ⁶³

Fig. 19.3 The Hadrianic masonry amphitheatre (GAG87: after Bateman et al. 2008). Drawn by Justin Russell.

The Thames waterfront was also rehabilitated. Substantial quays, formed by tiers of large squared timber baulks retained by dovetail jointed tiebacks, were built between London Bridge and the Walbrook. As already noted, the earliest may have been under construction from c. ad 128, although pottery dating suggests that others may not have been completed for the best part of a decade. ⁶⁴ As in the Neronian period the more substantial facilities were built upriver of the bridge. Simpler post-and-plank revetments of late Hadrianic or Antonine date occupied other locations along the north bank, reaching as far upstream as the mouth of the Fleet and perhaps as far downstream as the Tower. ⁶⁵ These may have been the product of various local initiatives. At some point in the early second century, new quays subsumed the temporary pier-base associated with the bridge over the Thames, suggesting its replacement. A new bridge may have been built in masonry, accounting for broken Roman tile found when piles for London Bridge were dug in 1832. ⁶⁶ A curse inscribed on lead and addressed to Metunus, a Vulgar Latin spelling of Neptune influenced by its spoken Celtic pronunciation, was found on the adjacent foreshore, giving rise to the suggestion that the bridge housed a shrine to the ocean god. ⁶⁷ Clusters of Roman finds recovered from river dredging might represent votive offerings made to the Thames. ⁶⁸

The streets leading to the bridge were given monumental aspect by masonry porticoes. Arcades flanking the street from London Bridge to the forum were built of walls and pier-bases set over chalk footings cut into Hadrianic fire debris. ⁶⁹ South of the river, along the stretch of Watling Street beneath Borough High Street, pier-bases were built for a similar purpose using bricks in a fabric dated ad 120–60. ⁷⁰ These covered porticoes can be seen as part of an empire-wide fashion of Hellenistic inspiration that took-off in the Hadrianic and Antonine periods. ⁷¹ The street architecture drew activities out of the forum and along London’s axial streets. Rather than the geometric uniformity of the street grid, second-century urban design emphasized the importance of colonnaded streets encouraging movement between public spaces, forming an architectural armature suited for street festivals and processions. ⁷²

Differences in structural detail suggest that London’s porticoes were commissioned piecemeal from different builders, perhaps in private investment. For the most part, however, reconstruction after the Hadrianic fire is likely to have relied on official support and engineering, perhaps benefitting from imperial patronage following the example of Hadrian’s contribution to the restoration of other war-damaged cities. ⁷³ Water supply was presumably restored, and this may have been the date for the insertion of a water supply-pipe along the southern side of the Colchester road at 60–3 Fenchurch Street. ⁷⁴

Most areas of housing were restored after the fire, although some plots remained vacant. ⁷⁵ At Watling Court rainwater puddles formed over the destruction debris, but needn’t have marked the passage of more than one wet winter. ⁷⁶ The slightly built timber houses that subsequently occupied the site failed to reproduce the architectural sophistication of the earlier town houses. Builders reverted to types of construction favoured in the Neronian city, perhaps husbanding resources whilst capacity was stretched. Elsewhere post-fire rebuilding involved some amalgamation of earlier properties, although earlier boundaries were generally maintained. ⁷⁷ The impression is of a city rebuilt in haste, but along former lines. Some areas near the waterfront may have been treated more ambitiously. At Regis House the opportunity was taken to redesign the arrangement of warehouses, and the scale of clearance and rebuilding hints at the involvement of a central authority operating on a grand scale. ⁷⁸

The upper Walbrook ‘vicus’

Although some areas were not as densely populated as previously, the establishment of a new district in the upper Walbrook valley suggests a new influx of settlers (Fig. 19.4). The northern margins of the Flavian city reached to the borderlands of the upper Walbrook valley, a basin of open land dissected by small streams that had attracted sporadic activities from the close of the first century. ⁷⁹ At 10–12 Copthall Avenue a drainage channel feeding into the Walbrook employed a timber felled in ad 107, whilst a timber building at 55–61 Moorgate housed a wooden tank associated with industrial production. ⁸⁰ Nearby were the robbed foundations of a large masonry building that fell into disuse c. ad 120, and from about ad 110/120 pottery kilns at 20–8 Moorgate made vessels of types associated with the Roman military. ⁸¹ But the main phase of expansion into the upper Walbrook valley was Hadrianic. An ambitious engineering programme introduced new streets between the Cripplegate fort and Walbrook, including at least three aligned north-south and perhaps two east-west. Marshy areas were drained and reclaimed before gravelled road-surfaces were laid over timber and turf causeways flanked by timber drains. ⁸² Pottery from these construction levels establishes their Hadrianic date and a terminus ante quem for the layout of the new district is provided by timbers felled ad 129 used in roadside buildings at Drapers’ Gardens. ⁸³

Fig. 19.4 A proposed reconstruction of London c. ad 130, also showing the distribution of ‘Walbrook skulls’ and other detached crania found in Hadrianic contexts. Drawn by Justin Russell.

The new streets, between 5 and 8 metres wide, were set on a slightly different alignment to the earlier street grid, perhaps orientated on the Cripplegate fort instead. The scale of the engineering suggests military involvement. ⁸⁴ The district presented characteristics similar to those of a fort vicus, housing industrial production likely to have been stimulated by the demands of the fort garrison. ⁸⁵ These included facilities for preparing leather goods and glassworks: which industries were relocated into this part of town and production was increased. The district also contained its own taverns, indicated by pottery assemblages dominated by flagons at 12–18 Moorgate. ⁸⁶ It was probably developed in tandem with the creation of the fort early in post-fire recovery (Table 19.1). Parallels can be drawn with the situation after the Great Fire of 1666, when London’s displaced population was resettled in camps north of the ruined city, although the Hadrianic expansion had more to do with servicing the military-run administration than looking after homeless Londoners. ⁸⁷

Table 19.1 A suggested timeline of events affecting London in the period ad 125–65