27

City of emperors (c. ad 285–350)

The ‘British empire’

London’s revival as a seat of Roman government conferred power on those who commanded here, eventually bringing it the pretensions of an imperial capital under the usurper Carausius and his successor Allectus. London had already been the likely setting of an attempted rebellion by a provincial governor in ad 280 or 281. Anthony Birley, extrapolating from the writings of the Byzantine historian Zonaras, suggests that this ambitious governor was a Moor adding to the body of evidence for North African influence in later Roman London. ¹ In ad 286, the Menapian Marcus Carausius, who had been appointed to marshal Rome’s northern fleet against Frankish and Saxon piracy, seized authority over Britain and parts of Gaul. ² The terms of his appointment imply that the fleet had needed rebuilding, and his subsequent propaganda suggested an abiding concern with coastal piracy (p. 334). His administration established new mints, including one at London, to produce coin to pay the army. The images on these coins celebrated his fleet, naval prowess, and the divine support of Neptune and Oceanus. They also emphasized his role as ‘Restorer of the Romans’ and argued his legitimacy as a partner in rule with the joint-emperors Diocletian and Maximian. ³ In March 293, however, these more widely recognized emperors appointed two junior colleagues to form the shared rule of the Tetrarchy. Constantius Chlorus, one of the junior emperors within this arrangement, swiftly restored Rome’s authority over Gaul. Carausius was then deposed and replaced by Allectus as ruler in Britain.

It was during Allectus’ reign that work started on one of later Roman London’s more ambitious building projects. The foundations of a monumental architectural complex have been revealed in several excavations near the approaches to the Millennium Bridge (Fig. 27.1). ⁴ New terraces behind the riverside wall stretched 150 metres along the waterfront. The riverside terrace established a platform for the construction of two large buildings, probably intended as temples, over a mass of timber piles driven into the London clay that had been felled from the winter of ad 293/294 into the early summer of ad 294. ⁵ A chalk raft with a timber lattice set over these piles supported massive foundations of reused ashlar blocks capped by petit appareil masonry walls. These techniques had been deployed in Gaulish walled circuits in the 260s and 270s before use at Portchester and Pevensey. These similarities suggest that the craftsmen assigned to the London project were redeployed here after completing improvements to fortifications along the Saxon Shore, indicated by the use of piles felled ad 293 in the defences at Pevensey. ⁶

Fig. 27.1 Temples and public buildings in the south-west quarter of London after the building programmes of Allectus (PET81, QUV01: after Williams 1993 and Bradley and Butler 2008). Drawn by Justin Russell.

The two buildings erected along the Thames waterfront had exceptionally broad foundations, 8.5 metres wide, on their southern sides. The side walls were also very substantial, measuring about 3.75 metres across, with the rear formed by a narrow wall terraced into the hillside to the north. No superstructure survived over these foundations and it is fairly certain that the buildings were never finished. Tim Bradley and Jon Butler suggest that the broad southern foundation supported a rectangular podium measuring c. 20.5 by 8 metres, with the other walls forming the sides of an enclosed courtyard to the north. This seems improbable. The enclosed areas are more likely to have lain below the temple cella where deep foundations were not required, with the broader southern foundation designed to support the columns and walls of the portico and pronaos of the temple facade. If so, these were classical prostyle temples, slightly over 20 metres wide and 25 metres deep, facing south to the river. The riverside wall might have been partially dismantled to give access to the temple steps, or the structures towered above this wall on a raised terrace. In either case the new buildings were designed to be seen from the river, presenting commanding views over a waterway previously cut adrift from the city. A broad flight of steps at the east end of the terrace established a grand entrance to the precinct where a small bathhouse may have been located.

The construction of the riverside terrace involved the demolition of an earlier precinct that had probably housed a temple to Jupiter (above p. 303). It is possible that one of the new temples rehoused Jupiter since a close association between Jupiter Capitoline and Roman imperial authority made this a popular dedication under the tetrarchy. ⁷ It is tempting to identify the other building as a temple to Sol Invictus. A temple to the unconquered sun was built in Rome in ad 274 after Aurelius’ promotion of this cult to official status. Several emperors identified closely with Sol Invictus and he featured on issues of Allectus’ London coinage. Since he was the only state deity to receive such acknowledgement he is perhaps the likeliest to have been honoured with a new temple. ⁸ Few temples were built anywhere in the late third century, but they featured in the architecture of imperial palaces in the tetrarchic capitals. ⁹ The best known is Diocletian’s palace at Split, which incorporated a temple to Jupiter. Allectus is particularly likely to have drawn inspiration from the palace at Trier, and the architecture of these imperial palaces suggests that he intended to arrange temples, treasuries, armouries and mints around a vast palace complex in the south-west quarter of London. ¹⁰ The northern boundary of this area was formed by the terrace walls along Knightrider Street that might have been the site of a circus. ¹¹ Many late Roman imperial ceremonies were orchestrated in the circus, which featured prominently within the tetrarchic palaces at Milan and Thessaloniki. Allectus may have decreed himself a riverside palace to form the ceremonial centre for his imperial rule, flanked by a circus to the north and temples to the south. Whatever the master-plan, this building programme rivalled the contemporary tetrarchic capitals. It drew on engineering capacity developed around the preceding programme of urban and coastal fortification, and shows that Allectus intended to rule permanently from London.

The temples were set to dominate the Thames, proclaiming that London no longer needed cautious defending but advertising an imperial authority rooted in Roman history and sanctioned by its gods. This ambition was wholly hubristic, since in ad 296 Constantius reconquered Britain. Historical sources describe an invasion fleet that sailed in two divisions: one from the Seine under Asclepiodotus the other from Boulogne under Constantius. Asclepiodotus’ ships landed in the Solent, where Allectus was engaged and killed. Constantius’ crossing was delayed and he reached London just in time to halt the sack of the city by Allectus’ Frankish mercenaries. ¹² The event left no certain mark on the city, although some of the fire damaged buildings described above (p. 332) might possibly have been looted at this time. Constantius’ rescue of London was acclaimed by panegyrists and celebrated on a commemorative gold medallion found in the suburbs of Arras in 1922. ¹³ This was part of a hoard buried in a vessel that contained at least twenty-five medallions, eight probably struck to commemorate the recovery of Britain, assembled by an official rewarded for his services by Constantius and his son Constantine. The most famous of these medallions, weighing 53 grams, illustrated Constantius’ arrival in London in 296, proclaiming him as restorer of the eternal light (redditor lucis aeternae). It showed the emperor on horseback approaching the gate of a city identified as London (LON), whose female personification knelt with arms outstretched to receive him. A manned war-galley occupies the foreground.

The County Hall boat, whose discovery in 1910 caused such a stir (above p. 15), might have been involved in this campaign. ¹⁴ The hulk abandoned in a river backwater was more than 3 metres wide, with a rounded hull and a draught of around 1.8 metres. Neither end survived, but the boat is likely to have been between 19 and 26 metres long, with a capacity of between 50 and 100 tonnes. It was carvel-built, following Mediterranean rather than local shipbuilding traditions, with the oak planks fastened edge to edge by mortice-and-tenon joints. Dendrochronological analysis establishes that these timbers came from trees grown in south-east England, felled after ad 287. Four coins were found within the hulk, the latest of Allectus minted between ad 293 and 296. There was no trace of the mast-step, and the surviving evidence doesn’t let us know what type of vessel it was. Peter Marsden finds evidence of a decked hull which would rule out the possibility that it was an oared ship, like the one shown on the Arras medallion, for want of headroom below deck. The absence of footrests and oarports also implies that it wasn’t a warship. ¹⁵ It remains the case, however, that boats of this type of construction were rare and it is unlikely to have been a merchant vessel. It may, therefore, have been one of the ships built as part of Carausius’ restoration of the fleet. According to the panegyrist, Carausius had a great many ships built in the Roman fashion (in nostrum modum), which fits both the style and date of the County Hall discovery. ¹⁶ This ship might alternatively have been added to the fleet after Britain was restored to the Roman Empire in ad 296, as Marsden thinks likely, although shipbuilding would have been a less pressing priority at this time.

Within the city

London changed little in the early fourth century. A panegyrist described Constantius’ transfer of skilled workers (artifices), with whom the provinces of Britain were overflowing, to help in construction at Autun. ¹⁷ London needed fewer construction workers now that Allectus’ unfinished palace was surplus to requirements, but remained sufficiently important to retain the mint established under Carausius. This continued to produce coins in silvered-bronze bearing mint marks LON or LN down to ad 325. ¹⁸ Constantius also returned to Britain in ad 305 to campaign beyond the Antonine Wall, before dying in York in ad 306 which is where his son, Constantine, was acclaimed emperor. ¹⁹ London briefly remained an important command centre during Constantine’s ascent to power, marked by coins issued to celebrate his visits in ad 312 and 314. ²⁰

Most of London’s late third-century town houses survived into the fourth century, and a few were renovated. ²¹ Rear-extensions were added to roadside buildings at One Poultry, one of which contained a hypocaust, plunge bath and mosaic pavement. ²² Piles used in this construction were felled ad 302–34, making it the latest tree-ring dated Roman building known from London. A feature of the architecture of this period may have been the addition of strongly built towers within important domestic compounds, some of which were now walled (Fig. 27.2). A large early fourth-century tower was built at the end of the western wing of the town house at Plantation Place. ²³ This was a massive construction, measuring some 11.5 metres by 9.25 metres, set over chalk foundations 1.7 metres wide and 3.2 metres deep. It was semi-basemented with mortar floors that included coins dated 307–10 and 330–5. Plaster from decorated niches and openings included fragments of yellow and green vegetation. This was a lordly addition, signalling a new approach to the demonstration of social status following a trend towards the creation of palatial private residences that characterized late Roman towns. ²⁴ Another semi-basemented tower may have been attached to a property in Southwark at 4–26 St Thomas Street, where stone walls set over piles enclosed a cellar some 6.7 metres long by 4.4 metres across. ²⁵ A curious foundation, more than 4 metres wide, built across the line of a disused road south-east of the amphitheatre temple precinct around the end of the third century may also have been associated with a large private property encroaching onto public space. ²⁶

Fig. 27.2 The late Roman town house at Plantation Place, adapted from the third-century town house shown as Fig. 21.2 with the addition of heated baths and a substantial tower (FER97: after Dunwoodie et al. 2015). Drawn by Justin Russell.

The Walbrook temple remained busy: the entrance steps were worn from use and the nave floor needed relaying nine times. For 50 years it housed the mysteries of the Persian god for whom it was built. A marble slab found on a late floor surface carried a dedication that translates as ‘for the welfare of our August Emperors and most noble Caesar, to the god Mithras and the Invincible Sun from the east to the west’. The details indicate that this was inscribed between May ad 307 and May ad 308. ²⁷ Soon after, the building needed extensive repair. The columns that separated the nave from the aisles were removed, the nave floor raised and the west end rearranged. Prior to this rebuilding, dated c. ad 310–20 from associated finds, important cult icons were collected together and buried beneath the nave floor. ²⁸ Others were dispatched into the waters of the Walbrook. One of the buried items was the marble head of Mithras whose discovery caused such excitement in 1954 (above p. 16). It had been broken asunder before burial by a blow to the left side of the neck, perhaps in a deliberate decapitation to exorcise the simulacrum’s power before it was laid to rest. Cult images were mutilated in the destruction of other Mithraea in the Roman world, and it has been argued that militant Christians were responsible for iconoclastic attacks. ²⁹ We have absolutely no evidence to suggest that Christians were responsible for any destructions of pagan sites in London. In some regards the dismemberment recalls the Bacchic myths of sparagmos and earlier rituals of corpse abuse (above p. 254), and the head of the god may have followed a similar path of dispatch to other worlds. Mithras and his companions may have been ritually sacrificed as the building was prepared for a new type of worship. New altars replaced the earlier cult image, and the open layout accommodated different liturgical and ritual activities. Martin Henig has suggested that the front of the apse was converted to form a baldacchino covering a new shrine. ³⁰ New floors were laid and a small votive offering containing coins to ad 313–18 placed beneath timbers associated with the construction of a stone altar. Finds from in front of the apse included coins to ad 341–6 which provide a terminus post quem for the latest use of the temple. Objects associated with these floors suggest that the temple had been redesigned around Bacchic worship, which at this date may also have been Orphic. A small marble statuette of Bacchus and his companions was inscribed ‘HOMINIBUSBAGISBITAM’: you give life to wandering men. A circular silver casket and a silver bowl cut up into pieces were probably been hidden in a space in the north wall of the temple. Jocelyn Toynbee recognized that the casket was probably of late third- or fourth-century Mediterranean manufacture, decorated with hunting scenes in exotic landscapes of Bacchic inspiration. This was perhaps a cista mystica, to infuse and drug wine. ³¹ Other statues were found in the debris to the sides of the building; two marble figures found here had been hacked down to the torso in further ritual dismemberment, with their limbs perhaps removed for use as votive offerings. ³²

Redundancies

The pace of construction slowed in the early fourth century, leaving London little need to import building timber. Mosaic floors were sufficiently scarce to suggest that London is unlikely to have permanently housed skilled mosaicists. ³³ London became less important to Constantine when he moved his headquarters from Trier to Serdica in ad 316/317, and as a likely consequence London’s mint ceased production in ad 325. ³⁴ No further imperial visits are known to have taken place, apart from a brief and unexplained visit by Constans in ad 343. ³⁵ The lead seal of Constans found on the Thames foreshore near Billingsgate perhaps arrived with goods moved to London on this occasion, almost certainly the last time that a ruling emperor visited the city. ³⁶

London’s massive forum basilica was torn down in the late third or early fourth century. Already, from the mid-third century, the portico along its eastern side had been replaced by ephemeral structures used by metalworkers. ³⁷ The nave floor was later sealed by thick mud and roof collapse, and most of the basilica was demolished to its foundations as its building materials were quarried for use elsewhere. The cleared site was then levelled, to be covered with dark grey silts in the early fourth century. Similar deposits extended over the road to the north of the basilica showing that this too was closed for use. Several of London’s other public buildings may have been made redundant. By the end of the third century parts of the building in the area of Cannon Street Station, the so-called ‘governor’s palace’, were dilapidated. ³⁸ In Southwark parts of the supposed mansio were demolished by the early fourth century, whilst the baths attached to the residential complex at Winchester Palace were demolished at the end of the third century. ³⁹ Other parts of these buildings remained in use, with coin evidence suggesting that some occupation continued into the late fourth century.

London no longer needed all of its administrative buildings. Tetrarchic reforms may have encouraged the relocation of bureaucratic departments into the palaces of new rulers, leaving public buildings under-used. The forum was no longer a clearing point for shipments that passed through London. The shops and stores that surrounded its courtyard saw diminished use, finding fewer tenants once the flow of annona supplies dried and local provisioning replaced bulk importation. Similar changes occurred in other Romano-British towns. The main hall of the basilica at Silchester was given over to metalworkers from the late third century, and Wroxeter’s forum was largely abandoned after a late third-century fire. ⁴⁰ Mike Fulford suggests that some public buildings were casualties of regime change, following the demotion of monuments associated with the failed Imperium Galliarum. ⁴¹ The changes were, however, part of a more widespread shift in the use of urban space. An important study of rubbish in Rome finds that there was a distinct change in distribution mechanisms used to support the annona and frumentationes after the mid-third century. ⁴² Up until this date the amphora that carried these supplies were discarded at centralized public locations, especially Rome’s river port. Subsequently, however, there was a more widespread pattern of distribution that bypassed the earlier public locations. Something similar can be read from distributions of amphorae in Roman London. In earlier phases these were disproportionately concentrated around the port and forum, followed by a more dispersed pattern in later antiquity. ⁴³ London’s forum and port may have played an important role in the redistribution of annona and frumentationes until the middle of the third century, but ceased to do so after the reorganization of supply in the latter half of the century.

After building materials had been reclaimed and recycled, the sites of public buildings were converted to other uses, often housing metal workshops. Late Roman timber-framed building with hearths used for small-scale industrial production were built over the abandoned site of Allectus’ temples. ⁴⁴ These activities, dated c. ad 340, present such a contrast with earlier civic architecture that they are sometimes dismissed as a form of squatter occupation. But the settlement of metalworkers suggests urban regeneration. Their workshops generated income through rents and taxes, and secured vital industrial production. Much of the late antique economy was subordinate to military requirements which relied on the manufacture and repair of weapons and armour in both state-run and private workshops (fabricae). ⁴⁵ Whatever the impulse, it resulted in the transformation of civic space, as marginal activities were drawn into central locations.

The buildings made redundant in the late third and early fourth centuries were those most closely associated with the earlier public administration, perhaps particularly those managed through the office of the procurator. Some places used as settings for civic ceremony may have shown greater resilience. This might account for the retention of the apse at the east end of the forum basilica after the rest of the building was demolished, perhaps saved from destruction as a setting for imperial ritual and civic memory. ⁴⁶ As elsewhere, we can assume that ceremonies designed around the public demonstration of loyalty to the emperor and Rome continued. ⁴⁷ London would have witnessed processions and banquets, celebrating the important feast-days that reinforced the relationship between urban communities and political authority. Coins from the latest arena surfaces date around ad 340, suggesting their continued use and repair. It is unlikely, however, that the amphitheatre hosted many gladiatorial combats at this time. From the second century these buildings were chiefly used for animal shows, particularly bear-baiting. The wider popularity of this cruel sport probably accounts for the recovery of the remains of brown bears from some sites in later Roman London. ⁴⁸

Southwark

In contrast with the city, Southwark remained unwalled. It is not certain that London Bridge survived and it has been suggested from the dates of coins found in the Thames that the crossing closed c. ad 330. ⁴⁹ The evidence is, however, inconclusive and might simply witness the disuse of a bridgehead shrine from which coins were dedicated to the river. River traffic must still have been landed cargoes along Southwark’s ageing waterfront, supplying the industrial district on the north island with fuel and raw materials from Hampshire and Surrey. This area continued to house metalworking and smithing, although it is not clear if these activities relied on imported or recycled iron. After a busy end to the third century, an early fourth-century contraction is implied by some sequences. This was followed by an upsurge in activity around the middle of the fourth century with metalworking and smithing attested by hearths, copper-alloy and iron objects, and slag. ⁵⁰ A timber gate built across the newly resurfaced road may have controlled access to the industrial quarter, where production may have responded to official commissions involving shipbuilding or naval repairs.

Several wells of this date have been identified and Richard Hingley has drawn attention to the distinctive nature of the assemblages associated with their disuse. ⁵¹ These were probably the product of ceremonies celebrated at the moment of their closure (below p. 385). An important collection of cult items was recovered from the fills of a well beneath Southwark Cathedral, presumably derived from buildings standing here in the late third and early fourth century. ⁵² Most of the items had been made in the second or third century, and included a funerary chest portraying a woman reclining on a couch holding a bunch of grapes. Some might have come from a mausoleum, but others are more likely to have been found in a shrine. The reclining figure had been decapitated and burnt before being deposited in the well, offering a parallel to the cult items buried in the Mithraeum. Similar processes may have been at work, in which changing belief systems resulted in the fourth-century dispatch of items that had served outmoded third-century worship.

The temple precinct on the southern border, at Tabard Square, was divided into two separate enclosures, accommodating the different liturgies of the two temples. ⁵³ Three plinths along the western side of the precinct supported statues, altars or columns, whose presence is indicated by fragments of stone sculptures and the left foot of a larger-than-life bronze statue. Another foundation opposite the north temple may have supported an altar or column.

The countryside

The fourth century was the great era of villa building in Roman Britain. The mosaics in these country houses show a sophisticated and cultured understanding of the philosophical arguments of the time, drawing on the literary allusions and iconography used to express paideia. ⁵⁴ This was particularly the case in south-west Britain, where an influx of ideas, money and perhaps people, took inspiration from the imperial capital at Trier. This area is thought to have been detached from the command of London to become the separately governed province of Britannia Prima, perhaps in the political settlement made by the incoming regime of Constantius Chlorus. ⁵⁵ Villas and their mosaics illustrate the presence of a wealthy landed gentry exercising power from country estates, in a settlement pattern inherited from the later Iron Age. ⁵⁶

Despite its political importance London failed to attract an equivalent share of spectacular grand country houses. Although several villas are known from the region, these were modestly appointed. Some may have been abandoned around the middle of the third century, probably including Ashtead in Surrey and perhaps, temporarily, the villa at Lullingstone in Kent. ⁵⁷ Overall, however, villa densities in north Kent and on the Sussex coast peaked in the later third century, and declined thereafter. ⁵⁸ The principal villas within London’s wider orbit remained those of the Darent valley in north Kent (above p. 178), several of which prospered. Some incorporated structures likely to have been used as granaries, with unusually ample facilities at Lullingstone, Darenth, and Farningham. ⁵⁹ These estates may have come to play a more important role in the collection and storage of tax in kind. Villas in Essex were also built or enlarged, notably the large courtyard villa at Chignall St James, near Chelmsford, where new enclosures may have supported ranching and stock-breeding. ⁶⁰ Although these farms are likely to have fed London, their owners may have remained socially and politically attached to neighbouring cities such as Colchester and Canterbury.

A smaller group of suburban villas is more likely to have housed individuals involved in London’s political life. Shadwell remained one of the most important of these. ⁶¹ Programmes of renovation to the bathhouse started c. ad 275, resulting in five different phases of remodelling over a 50-year period. Coin-finds point to peaks of activity through the 280s and 290s as the baths were enlarged and improved. Aspects of the coin supply find parallel in the Saxon Shore fort at Reculver, reinforcing the impression that Shadwell was the seat of an official mediating between the London-based government and the spreading estuary and coastal regions beyond. ⁶²

We can chart the distribution of the rich and powerful through the stone sarcophagi in which some were buried. These stone tombs were imported to the region at considerable expense in the late third and fourth centuries, and are usually found close to the likely locations of suburban villas. Examples include the tomb of Valerius Amandinus, designed for display within a mausoleum, from Thorney Island in Westminster. ⁶³ Others have been found at St Martin-in-the-Fields, Lower Clapton and Ratcliffe. The early seventeenth-century discoveries at Ratcliffe included a stone sarcophagus and lead cist decorated with scallop-shells, with impressive grave-goods. ⁶⁴

Similar sarcophagi are occasional finds in London’s extramural cemeteries. ⁶⁵ One found at Haydon Square in 1853 was sculpted with the bust of a young man, designed to be displayed within the burial chamber of a mausoleum. A more recent example from Harper Road, on the borders of Southwark, consisted of a sarcophagus housed within a rectangular mausoleum defined by chalk foundations which contained a burial carbon-dated around ad 328. A widely publicized find was made at Spitalfields Market in March 1999, where a limestone sarcophagus, one of two from the site, contained a lead coffin that had been decorated with scallop shells symbolizing the journey to the underworld originally connected with the god Bacchus. ⁶⁶ Inside the coffin was the body of a woman in her late teens or early twenties who had been buried in the mid- to late fourth century. The study of her aDNA and lead isotopes in her teeth suggests that she spent her childhood in southern Europe, possibly Rome, before coming to London. She was laid to rest on a blanket with her head on a pillow of bay leaves, and clothed in a silk damask tunic decorated with gold thread that may have incorporated purple wool. Residues show that the body had been prepared for burial with imported pine and mastic resins, and she was accompanied by goods that included a jet box and glass phial that hint at a belief system rooted in Bacchic approaches to salvation.

The settlement at Old Ford 4 kilometres north-east of London, where the road to Colchester approached the crossing of the river Lea, may have expanded at this date. Water-meadows alongside the Lea offered grazing for cattle and horse, and the roadside settlement may have managed these valuable territories and been a holding station for livestock destined for town. ⁶⁷ The growth of the site, and its extensive burial grounds, dates largely after ad 270 when there was a shift towards higher cattle frequencies, meeting London’s renewed demand for beef, horn and leather. Cattle was driven into town for slaughter: where butchery shows an increase in the distribution of meat taken off the bone. ⁶⁸ Rebuilding took place at other roadside sites, including Enfield, Brentford and Staines, often on plots left open since the end of the second century, and involved the introduction of buildings with masonry walls and tile roofs. ⁶⁹ It has been suggested that the revived fortunes of small towns and roadside sites was a consequence of their role as collection points for tax and rent drawn on for annona supply. ⁷⁰ This might account for the development of a late third-century settlement on the south bank of the Thames at Thamesmead, where a masonry building is suggested by building debris. ⁷¹