23

Severan revival (c. ad 180–225)

London in the late second century

We know little about London in the 170s and 180s. The diminished city built less and consumed less, and a consequent dearth of evidence leaves us hanging on a patchy historical record. ¹ Dio described a military setback of the early 180s when Britain was invaded from the north, and subsequent operations may have inspired an issue of victory coins in ad 184 and justified Commodus’ adoption of the title Britannicus Maximus. In following years the British legions were twice charged with backing usurpers, implying episodic unrest. Some of these actions may have echoed on London’s streets, although not in ways we should expect to trace from archaeological evidence. There is passing intrigue, however, in the fact that some houses were destroyed by fire in the closing decades of the second century. ² But we have no certain way of distinguishing between accident and arson, and these are more likely to have been local house-fires than anything of wider consequence.

One of these fires, identified in excavations at 7–11 Bishopsgate, destroyed a kitchen that contained a Samian table service assembled c. ad 175. This shows that London continued to receive fresh cargoes of Samian in the years after the Antonine contraction, as more amply illustrated by broken warehouse assemblages heaped on the Thames foreshore at Three Quays House and New Fresh Wharf. ³ At least one of these had been put together c. ad 170. This Samian trade filled ships crossing from Gaul, and hundreds of Samian tableware sets of the 180s have been dredged from the sea at Pudding Pan Rock, off Whitstable in Kent, probably from cargoes carried by vessels lost en route to London. ⁴ London evidently remained a staging point on an east-coast supply-route, as traced by distributions of north Gaulish fine wares extending from the continent to York. The prayers of shippers involved in such commerce are commemorated on altars dedicated at sanctuaries in the Scheldt estuary (above p. 268). ⁵ These record the concerns of a merchant community similar to that present in London earlier in the century, although there is little to suggest that the city remained a particularly important centre for their activities. London’s quays were no longer dispatching bulk supplies onto the roads radiating from the city, and the town had no need of a workforce equal to that of its busier infancy. As a consequence there was no significant rebuilding on plots vacated earlier in the century.

Despite its diminished state, London’s role in coastal trade gave it access to a variety of luxury imports. It remained Britain’s principal ruling city and an important market where wealth and power concentrated. Martin Henig has suggested that the nearby villa at Lullingstone in Kent was the rural retreat of the provincial governor Publius Helvius Pertinax. ⁶ Pertinax governed Britain in ad 185–6, before rising briefly to the imperial throne following Commodus’ assassination ad 193. Henig’s argument draws on the evidence of two marble busts found in the villa’s cellar thought to show Pertinax and his father. The villa’s aristocratic credentials are further demonstrated by an exceptional cornelian intaglio of victory, once set in a gold signet ring held by a man of equestrian or higher rank. The bust identified as Pertinax had been decapitated, probably with an axe. This could have taken place during disturbances in Britain whilst he was governor, or in an act of damnatio following his deposition and murder in Rome. The exaggerated overthrow of the images of deposed emperors was a useful act of political theatre when regime-change precipitated a public realignment of loyalties.

The chaos that followed the murder of Pertinax carried another governor of Britain, Clodius Albinus, towards the imperial throne. ⁷ Albinus had probably taken office in Britain by ad 191, where he was one of three provincial governors’ to be acclaimed emperor in ad 193. Albinus’ claim to power derived from the legions under his command—in Caerleon, Chester, and York—but he is likely to have ruled from London. After Severus secured Rome, Albinus tactfully accepted the more junior position of Caesar before later reasserting his rival claim. In ad 196, he crossed to Gaul with his armies, only to be defeated by Severan forces near Lyon in February ad 197. According to the authors of the Historia Augusta, Albinus’ head was cut off and paraded on a pike before being sent to Rome, whilst his decomposing body was mangled by dogs before being thrown into running water. The victorious Severus is reported to have acted to make sure that affairs in Britain were settled, before turning his attentions east. ⁸ A decade later, however, Severus came to Britain to complete its conquest, arriving here with new forces and accompanied by his sons Caracalla and Geta. ⁹ This ambitious military venture was cut short by his death in York in February 211, leaving Caracalla to agree a British peace and return to Rome with the imperial party.

These various events were of enormous consequence for London. For the first time since Hadrian’s visit the city had hosted emperors and their retinues. The campaigns launched by Albinus and Severus, although headed in opposite directions, involved a massive traffic between Britain and the continent. These, it can be suggested, provided the likely spur for a new era of waterfront building along the Thames.

The late second-century waterfront

Excavations at various riverside sites along the north-bank of the river have exposed a complex series of late second- and early third-century waterfront revetments. Their presence shows that after many decades of neglect London’s port was revived and enlarged, implying a significant increase in traffic (Fig. 23.1). Several discrete phases of improvement have been identified from this evidence, some of which are better understood than others. In the first of these, the substantial Hadrianic quays between the mouth of the Walbrook and London Bridge were replaced by slightly built revetments of superimposed horizontal boards retained by timber uprights. These advanced the embanked area some 4–6 metres onto the foreshore. ¹⁰ Frustratingly, none of the timbers preserved sapwood from which an exact felling date can be calculated, and some are likely to have been reused from nearby Antonine constructions. One of the timbers used at New Fresh Wharf was dated ad 188–223, and had probably been felled by ad 204, whilst late second-century pottery was associated with the revetments at Swan Lane. ¹¹ Since these works are likely to have preceded the more heavily engineered programme of c. ad 197, discussed below, we can suggest that the slightly built waterfront was probably assembled after ad 188 but before ad 197. Other parts of London’s infrastructure were repaired in these years, as indicated by a felling date of ad 191 on timbers used in a Walbrook channel at Angel Court. ¹² These works signalled the end of a long interruption to investment in London’s public infrastructure.

Fig. 23.1 A proposed reconstruction of London c. ad 245. The late Severan quays replaced more temporary constructions likely to date to the 190s. Drawn by Justin Russell.

Why then, did the Hadrianic facilities need replacing at this particular point in time? And why were the new quays built with inferior engineering? It is entirely credible that the earlier port was no longer fit for purpose following its neglect. Tidal regression may also have handicapped ships coming upriver towards London, making it harder to moor alongside the quays and encouraging shippers to build further into the river (below p. 327). Despite this the late second-century Samian trade was accommodated in unimproved parts of the waterfront during the 170s and 180s, when boats could still sail upriver and beach on the foreshore. The decision to build new facilities is likely to have been provoked by a more pressing need than this routine trade. It is tempting, therefore, to suggest that it stemmed from Albinus’ plans for a naval expedition to Gaul. This campaign almost certainly involved the largest movement of people and goods between Britain and the continent since the Claudian invasion. The need to assemble and dispatch campaign supplies at short notice would account for a hurried programme of engineering along this prime stretch of London’s riverfront. ¹³

Other phases of waterfront improvement were soon to follow. The north bank was eventually transformed by a series of more substantial quays, stretching from Queenhithe in the west to Tower Hill in the east. ¹⁴ An early component was built at Queenhithe, upstream from the mouth of the Walbrook. Large squared oak timbers were laid in horizontal courses to form an embankment at the westernmost end of the port, using timbers felled in ad 197. ¹⁵ This is a date of enormous resonance, being the year of Albinus’ defeat in Gaul. These quays must have been built following the arrival of Severus’ newly appointed provincial governor, Virius Lupus, in the aftermath of the battle of Lyon. ¹⁶ Albinus had withdrawn all campaign-ready troops from Britain, leaving Lupus under considerable pressure to establish Severan authority in this formerly restless province. The location of the Queenhithe quays further upriver than earlier facilities, suggests a renewed interest in using London as a springboard for moving goods inland along Watling Street. Incoming traffic is likely to have included troops and materials aimed at reinforcing the forts at Chester and Caerleon, perhaps directed through London since overland routes were safer outside the normal sailing season. Traffic through the Queenhithe quays may have hastened a revival of the surrounding district. Timber and clay-walled workshops built over the disused bathhouse at Huggin Hill in the third century housed glass and iron working. ¹⁷ The site of the baths was now given over to craft production that may have helped equip troops in transit.

Carefully constructed quays of a different type were built around the mouth of the Walbrook and Lorteburn near the Tower of London. At both locations, new waterfronts were built of large timber baulks laid in tiers to form two parallels rows of boxes. This form of caisson construction may have been preferred where larger areas of low-lying ground needed reclamation. It is not possible to date these quays closely, although it seems likely that they belonged to a slightly later phase of improvement than the works at Queenhithe. The double-box quay along the east side of the Walbrook was investigated at Cannon Street Station, where it reclaimed some 25 metres of the former river channel at the foot of the ‘governor’s palace’ complex. It incorporated reused timbers felled ad 169–89 and another felled no earlier than ad 185. ¹⁸ Further box constructions at Dowgate Hill House may have lain between two separate channels forming the mouth of the Walbrook. ¹⁹ The quays built near the mouth of the Lorteburn, investigated at Custom House and Three Quays House, were formed from 1.5 metres square openwork boxes that must have supported plank decking. ²⁰ The quay at Custom House was pottery dated to the end of the second century, predating the more substantial extension of c. ad 209–24 described below. The latest structure at Three Quays house included timbers felled ad 197–224 and incorporated early third-century East Gaulish Samian. These late second- and early third-century building works combined to considerably enlarge the working port. New quays were inserted in previously under-developed areas at the port’s western and eastern ends, and in deep-water areas at the confluence of the Walbrook and Lorteburn, significantly increasing London’s ability to handle large volumes of shipping.

The later Severan waterfront

Initially the central section, either side of London Bridge, was left in a more primitive state. Here the late second-century post-and-plank revetments, suggested to have been installed during Albinus’ government of the city, were retained for several years. They were eventually replaced by the construction of a more uniform and durable embankment in the early third century. This consisted of a new terrace of massive oak beams trimmed to standard sizes, set in five superimposed tiers, and held in place by a framework of tie-back braces. This was set 4.4–5.2 metres in front of the earlier post-and-plank revetments either side of London Bridge. ²¹ Underpinning piles were driven into the foreshore and mortised to the underside of the base timber in an innovation that characterized this particular construction programme. Clay and gravel dumps infilled areas behind the riverfront, although some may have been left as openwork structures. Similar construction techniques were used to infill a gap in the waterfront between the Queenhithe quays built c. ad 197 and the mouth of the Walbrook, and to extend the waterfront eastwards towards the Tower. ²² These quays were constructed in the first quarter of the third century, with timbers from New Fresh Wharf and Billingsgate Lorry Park suggesting a date within the period ad 209–24. ²³ Although the engineers made more economical use of timber than in first and early second-century building, the extended length of the waterfront meant that unprecedented quantities of timber were required, necessarily resulting in extensive woodland clearances around London. Damian Goodburn has noted that the bulk of trees felled for a massive construction project in ad 293/294 (below p. 350) had started growth between ad 217 and ad 234, concluding that the parent trees had grown in woodlands cleared by heavy felling for a large London project at this time. ²⁴ Although this is a somewhat indirect source of evidence, it seems credible that the busiest period of port rebuilding took place in the later Severan period starting c. ad 215/220.

The waterfront expansion created new land in front of the old quays, allowing the authorities to regain control of access to the river and over goods passing through the city. Early third-century timber-framed buildings behind the quays at Seal House and Swan Lane added warehousing capacity. An important group of lead seals was found on the Thames foreshore in front of the quays at Billingsgate. ²⁵ These included imperial seals of Septimius Severus with Caracalla and Geta, a provincial seal of Britannia Superior, one of the Arles customs station and a wine importer’s private seal. These seals had marked cargoes exempt from port duties and their distribution follows official supply. London accounts for 14 of the 32 imperial seals found in Britain, and Billingsgate was evidently an important port of entry for official cargoes. The facilities here were also equipped to receive unusually heavy goods, using lifting equipment set over wooden crane bases. ²⁶ These would have helped in the landing of massive wooden barrels used to import wine from the Rhineland, which were generally too heavy for normal handling. Parts of these barrels were reused to line wells, and one found at Drapers’ Gardens had batch and volume marks which show that it had contained 935.45 litres of wine (105 modii and 11.5 sextarri) from a place called Capriacum. ²⁷

A dozen barrels this size would have filled the hold of a river barge such as that found in the mouth of the Fleet during building works near Blackfriars in 1962. ²⁸ Peter Marsden’s studies show this to have been a flat-bottomed vessel designed to sit on the river bed at low water. It was about 18.5 metres long, with a beam of about 6.12 metres and a gunwale height amidships of about 2.86 metres. It had been made from sawn oak planks nailed to a frame, caulked with hazel shavings and pine resin, following Romano-Gallic ship-building traditions and probably using trees grown in south-east England. The deck presented a large hatchway for goods to be lowered into the hold, with the mast step set forward in the hull and rudders and living quarters at the stern. A worn coin of Domitian showing the goddess Fortuna holding a ships rudder was set into a recess at the base of the mast-step socket for luck. The ship’s timbers were felled ad 130–75, probably before ad 156, whilst the absence of repair suggests that the boat was unlikely to have been much over 20 years old when it sank. It was well suited for river work, but its woodwork was riddled by the saltwater teredo worm showing that it travelled the estuary to the coast. Since the wreck was found upstream of London Bridge this suggests the presence of a drawbridge to let it pass. The barge contained 26 tonnes of Kentish ragstone within its half-filled hold, and had perhaps sunk in a collision whilst supplying stone for an Antonine or Severan building project.

Parts of another boat were found at Guy’s Hospital in 1958 and re-examined in 2010. ²⁹ This broad barge-like vessel was at least 16 metres long, clinker-built with oak planks nailed to a skeleton formed of keel and frame with the joints also caulked with hazel shavings and pine resin. It had been designed to rest on the foreshore at low tide, with a low freeboard that would have left it totally unsafe for the sea. The hulk had been abandoned in a channel of the Thames where it was preserved by an accumulation of river mud containing pottery dated c. ad 190–225, suggesting that it also served London’s port in the late second or early third century.

Riverside buildings

Several buildings were built or restored behind the working waterfront, combining to present an impressive panorama to those approaching by river. It is not clear whether this was intentionally achieved, or the incidental product of different projects attracted to the waterside. It is notable, however, that the new buildings hugged the river. The westernmost constructions were perhaps the most important. The riverfront terrace here may have housed a temple precinct here from c. ad 165 (above p. 265), which was extensively rebuilt in the early third century. ³⁰ Alterations included the addition of new colonnades enclosing one or more courtyards, and a river-facing apsidal exedra giving architectural emphasis to the river vista, using timber piles felled ad 205–32 (Fig. 23.2).

Fig. 23.2 The ‘Period I’ complex at the Salvation Army International Headquarters, a possible sacred precinct of the early third century (QUV01: after Bradley and Butler 2008). Drawn by Justin Russell.

Sculptured blocks reused nearby in the later riverside wall at Baynard’s Castle may have derived from monuments within this precinct. ³¹ Some limestone carvings had formed a small free-standing arch about 7.5 metres wide, 1.15 metres deep and little over 8 metres high (Fig. 23.3). This was too narrow to have stood over a road and probably marked a ceremonial gateway. The sides of the arch were flanked by standing divinities: parts of Minerva and Hercules are recognizable. A frieze above the arch included busts of gods, perhaps including Saturn, Sol, Luna, Mars, Mercury, Jupiter, and Venus, representing the days of the week. The spandrels contained sea monsters and busts of the seasons. These classical features find parallel in the design of the temple façade at Bath, with stylistic links to works in Gaul and the Rhineland. Tom Blagg’s careful study concluded that the ornament was no earlier than late Antonine and was probably Severan. This is consistent with dating evidence obtained from the nearby foundations. Other carved stones derived from a decorative screen, 6.2 metres long, 1.4 metres high and c. 0.55 metres wide, decorated back and front with three pairs of niches containing standing figures of gods. Vulcan and Minerva, Diana and possibly Mercury, and Mars can be identified, with features that suggest the presence of Hercules, Summer or Autumn, and Luna. It is not clear if this screen, dated to the second or third centuries, stood in the open or within a building. The figure of Minerva on both arch and screen had her spear turned down which is best read as symbolizing a province at peace. This would be consistent with the situation that applied in Britain after Caracalla’s peace of ad 211, but anomalous at dates earlier in the century. ³²

Fig. 23.3 A ceremonial arch reconstructed from sculptures found reused in the riverside wall at Baynard’s Castle and probably deriving from the religious precinct in the south-west quarter (BC74: drawn by Sheila Gibson and reproduced kind permission of Museum of London Archaeology). The figure shown to the left of the arch is Minerva, shown with her spear turned down.

Two inscribed altars of Lincolnshire limestone were also recovered from the riverside wall (Fig. 23.4). ³³ According to Mark Hassall’s translation one read ‘Aquilinus the emperor’s freedman and Mercator and Audax and Graecus restored this temple which had fallen down through old age for (or to) Jupiter best and greatest’, although the dedication to Jupiter is not certain. The other announced that ‘In honour of the divine house, Marcus Martiannius Pulcher, deputy imperial propraetorian legate of two emperors ordered the temple of Isis…which had fallen down through old age, to be restored’. These inscriptions commemorated the rebuilding of temples likely to have been located within the religious precinct. The governor Pulcher is otherwise unknown, but the reference to two emperors suggests that he held office at some point between ad 221 and ad 259, since the style of lettering seems late for the joint-rule of Caracalla and Geta. Hassall suggests that these altars were likely to have been erected under Trebonianus Gallus and Valerian in ad 251–3 or Valerian and Gallienus in 253–9. An earlier date is, however, possible and Anthony Birley considers the joint reign of Elagabalus and Severus Alexander in ad 221–2 a possibility along with other episodes of joint-rule. The dedications show that the temple restorations relied on the patronage of government officials. Another sculpture reused within the later riverside wall showed four seated mother goddesses (matronae), of a type who featured prominently in the cults of Cisalpine Gaul and the Rhineland. Their worship in Britain reflects on the imported origins of both dedicators and deities. ³⁴

Fig. 23.4 Altar to the restoration of a temple of Isis that had fallen down made by the governor Marcus Pulcher (MMARTIANVSPULCHER) (BC74: drawn by Sue Hurman and reproduced by kind permission of Museum of London Archaeology).

A different type of building, also dated to the third century from associated pottery, was set on a terrace some 40 metres to the west of the Walbrook’s confluence with the Thames. ³⁵ A large terrace wall, was fronted by brick piers which responded with a second line of piers to form an aisled building. The reuse of tiles stamped by the procuratorial office hints at a public involvement in its construction. The presence of tesserae, painted wall-plaster and hypocaust box-flue tile amongst demolition debris implies the presence of high-status reception facilities within the building. One possibility is that this was a portico or guild building associated with the nearby quays.

On the other side of the Walbrook the monumental building complex known as the ‘Governor’s palace’ was extended and rebuilt in the late second or early third centuries, perhaps following the construction of the new quays alongside the Walbrook. It was perhaps at this time that the pool was filled-in, and the large ‘hall’ on its north side replaced by smaller rooms with hypocaust floors. ³⁶ This phase of rebuilding might provide a context for many of the hypocausts and mosaics found nearby. If this had been a bathhouse, as suggested above p. 138, it may have been largely rebuilt during London’s Severan renewal.

Another monumental building was set over a terrace further downstream of London Bridge. Excavations in Botolph Lane found a splendidly preserved stretch of a roofed masonry culvert, large enough for a person to crawl along, which drained towards the Thames. ³⁷ A silt trap beneath a manhole shaft was made from timbers felled ad 176–221. The culvert was an integral part of a monumental structure overlooking the river. Its masonry construction testifies to the importance of running water to the architecture, and associated pottery includes types of vessels commonly associated with ritual use. The new structure may have incorporated running water, perhaps forming a nymphaeum: a monument type favoured in Severan patronage. ³⁸

A short distance further downstream Peter Marsden recorded a small east-facing rectangular masonry structure behind the waterfront at St Dunstan’s Hill. ³⁹ This was possibly a small shrine. We don’t know when it was built, but it might belong to early third-century restoration. An inscribed slate panel found in Walbrook Street in 1954 came from another monument of this period, although the incomplete text leaves its exact date and significance uncertain. It commemorated a victory, and may have been set up by the Province of Britain. It has sometimes been reconstructed as a reference to Trajan’s Dacian victory, although the style of the lettering makes a later date more probable. Anthony Birley suggests a reconstruction that would allow it to refer to Caracalla’s German victory of ad 213. ⁴⁰

Two marble statue heads found in London appear to portray members of the Severan ruling family, and perhaps graced the public architecture of the revived city. ⁴¹ The head of a young boy carved in Italian marble, found at an unknown location in London, has provisionally been identified as a portrayal of the young Caracalla. Another, found in the Thames, probably showed his younger brother and co-emperor Geta after ad 205 when he was first elected consul. This image had been mutilated in antiquity, involving the ‘amputation’ of the shoulders and sides of the chest with a chisel and a horizontal decapitation gash at the back of the neck. This is the sort of thing that might have happened following his damnatio memoriae after murder by Caracalla in December ad 211.

The town wall

London’s town wall was the grandest and most enduring of its Roman monuments, following an ambitious urban circuit that may have been first defined in the early Antonine period. It stretched over 3 kilometres from Blackfriars in the west to the Tower in the east, enclosing an area of 133 hectares, but leaving the riverfront open. ⁴² This made for a walled city larger than any other in Roman Britain, exceeded by only four in Gaul. The wall stood over trench-built foundations of puddled clay and flint capped by a layer of ragstone rubble (Fig. 23.5). The ground level on the outer face was marked by a plinth of large chamfered red sandstone blocks from Kent, with a triple facing-course of tiles at the corresponding level inside. Above this, squared blocks of ragstone were laid in regular courses to form the sides of a concrete wall some 2.0–2.7 metres wide. A double or triple course of bonding tiles was laid through the wall at intervals of five or six courses, at which points there was usually an offset on the inner face to reduce the thickness of the wall. The greatest height of surviving wall is 4.4 metres but it is estimated to have originally stood 6.4 metres high. The tile course at this point is likely to have supported a parapet protected by a crenelated breastwork formed from round-topped coping stones. A V-shaped ditch, 3–5 metres wide and up to 2 metres deep, was dug 2.7–5.2 metres outside the wall. The earth from this was piled in a bank against the inner face of the wall to make a 2-metre-high rampart. A gravel road surface found beneath this rampart at both America Square and Cooper’s Row preserved cart ruts from the construction traffic that had carried building materials from ships docked along the waterfront. ⁴³

Fig. 23.5 Profile of the masonry town wall (adapted from Maloney 1983, reproduced by kind permission of Museum of London Archaeology).

Culverts were placed where the wall crossed streams and a channel of the Walbrook was carried through a 1-metre-wide arched opening at Blomfield Street protected by a rail of five iron bars. ⁴⁴ The north-west angle of the town wall reused the corner of the Cripplegate fort, whose walls were thickened to bring them into line with those of the city. Rectangular chambers, no more than 1.83 metres wide and 3.35 metres long, built onto the inside of the wall probably housed timber stairs giving access to the parapet walk. ⁴⁵ There were four principal gates into town probably with double portals: the Colchester road passed through Aldgate; Ermine Street through Bishopsgate; Watling Street through Newgate; and the road to Westminster through Ludgate. ⁴⁶ Some may have been built as free-standing structures before the construction of the wall (above p. 273). Additionally, the northern and western gateways of the Cripplegate fort remained in use, and posterns are conjectured to have existed at Aldersgate, Aldermanbury, Moorgate, and Tower Hill, combining to provide access at intervals of 250–350 metres. ⁴⁷ A tile-coursed ragstone structure north of Ludgate Circus formed a contemporary masonry bridge across the Fleet. ⁴⁸

Material associated with the construction of the town wall is consistently dated ad 190–230. ⁴⁹ Substantial pottery assemblages of this date came from the ramparts and construction road at America Square. ⁵⁰ A worn coin of ad 183–4 found beneath the thickening of the external wall around the Cripplegate fort at Falcon Square could not have been lost much before c. ad 190, and provides a terminus post quem for the wall’s construction. ⁵¹ The wall may, however, have been built some decades later. This can be inferred from what we know of the chronology of waterfront development. Tens of hundreds of boatloads of building material needed landing on the riverbank for loading onto the carts that trundled their way along haulage roads to supply the masons. It is likely that the Roman engineers followed the sensible expedient of improving the riverfront handling facilities before building the wall. The quays at the easternmost end of London’s port were best placed to serve the haulage road found at America Square, and formed a later addition to the early Severan waterfront. Elsewhere these later Severan waterfront developments are dated ad 209–24 (above p. 300). The evidence is indirect, but weighs in favour of a construction date in the first quarter of the third century.

Debris from the reproduction of Roman coins found in a wall-turret at Warwick Square confirms that the wall was in place by c. ad 225/230 (p. 331). ⁵² This manufacture involved taking impressions from official issues onto clay moulds that were broken up when the replicas were extracted. The coins that were copied had been minted ad 201–15 and the absence of later types suggests that the ‘forger’ was at work in the following decade. An early third-century date would also account for finds recovered from the associated ditch at London Wall. ⁵³ These included several unbroken funerary pots made of black burnished ware fabrics dated ad 160–230 which held offerings of chickens and coins. The archaeological evidence combines to indicate that London’s town wall was built after ad 190 but was standing by ad 225/230. It is also more likely to have been built after the waterfront improvements of ad 209 onwards than before.

It is estimated that somewhere between 82,000 and 86,000 tonnes of material were used to build the wall, chiefly ragstone quarried from near the River Medway in Kent. ⁵⁴ If the Blackfriars barge was representative of the river vessels of the time, then the wall required 1720 barge-loads of 50 tonnes apiece. Each delivery required a week-long round-trip, involving two days for each leg of the river journey and one day each loading and loading. A fleet of a dozen barges would have taken the best part of 3 years to deliver the stone, although a more compressed timetable involving a larger fleet is equally likely. Most bricks used within the wall were made in local fabrics, probably from kilns set up for the purpose. The hemispherical coping stones used in the wall crenellations were made from limestone that matches sources near Boulogne, perhaps supplied by the Classis Britannica from its base there. An exercise of this scale required imperial sanction, and might have been managed by the procurator drawing on vessels within the command of the Classis. Even if the work was undertaken by civic authorities employing private shippers and local resources, it is difficult to imagine that a commitment of this scale would have been permitted whilst Severus was campaigning in Caledonia ad 208–11. Severus’ need to retain capacity for the supply and transport of his forces must surely have taken priority. ⁵⁵

We can therefore conclude that London’s wall was probably built after Caracalla’s peace of ad 211 had ended forward operations. Surplus capacity could then have been redeployed into peacetime engineering projects. Severus and Caracalla were committed to a high level of public expenditure to secure dynastic legitimacy and new masonry walls were built around the main legionary fortresses and forts and supply-bases on the northern frontier at around this date. ⁵⁶ The design of London’s defensive circuit drew inspiration from this military architecture, notably in the use of the distinctive chamfered plinth at the base of the wall. ⁵⁷

London renewed

London’s town wall wasn’t built in hurried response to political crisis, but confidently enclosed a larger area than necessary and left the riverfront undefended. It was a carefully planned monument that secured the city’s revival, making an assertive statement that anticipated the late antique conceptualization of the urbs as a place encircled by walls. In a sense, the town wall took on the argument of the Hadrianic fort built a century previously, but amplified and extended to embrace the city entire. From this point forward there was a military aspect to the way in which London presented itself to the world.

For all of its ideological force, the wall wasn’t only an argument of stone but a device that framed patterns of life. It channelled travellers to supervised entry-places, policing those drawn to the city. Rome’s Aurelianic wall, built later in the century, were allegedly used to exclude people wearing barbarian dress and long hair, and we can similarly imagine London’s privileged community seeking to separate themselves from outside folk. ⁵⁸ Rome’s walls also facilitated controls over goods in transit, enabling tax collection at city gates. ⁵⁹ These were an important source of urban income, and the lead seals found at Billingsgate testify to their importance. These tolls and taxes would have offset the costs of policing the walls and gates.

Town walls marked privileged status, and their building required the emperor’s consent. ⁶⁰ Whilst there is no certain relationship between political status and the right to build a city wall, walled towns were more likely to have been coloniae. It is possible that London now gained this honorific status as part of a wider package of reforms made when Britain was divided into two separate administrations (Fig. 23.6). According to Herodian, the province was split into Britannia Inferior and Britannia Superior by Severus c. ad 197/198. ⁶¹ Since inscriptions show that Britain was undivided in ad 212, this reform was probably not implemented until after Severus’ death, probably under Caracalla c. ad 213/214. ⁶² Britannia Inferior came to be ruled from York, which was described as a municipium in events of ad 211 but as a colonia in an inscription of ad 237. ⁶³ The implication is that York was promoted at around the time of the reform of the provincial administration. A similar arrangement may have applied to London on the creation of Britannia Superior. The presence of the tombstone to the speculator Celsus (p. 220), dated to this period because his affiliation to Legio II Augusta included the title Antoniana which was not bestowed before ad 213, has been seen as proof that London was recognized as a capital city of equal prominence. London’s promotion would have been most easily affected after ad 212, when Caracalla extended Roman citizenship to free inhabitants of the empire in his Constitutio Antoniniana. ⁶⁴ This served to blur distinctions between resident incolae, who were citizens of other cities, and those whose status relied on their ownership of estates in and around Londinium.

Fig. 23.6 Britain in the early third century showing the likely extent of the separate administrations formed for Britannia Superior and Britannia Inferior (governed from London and York respectively), along with the emerging network of defended coastal sites some of which later formed the command of the Saxon shore and some other late Roman walled towns. Drawn by Fiona Griffin.

The building of the city walls and adornment of waterfront temples and public buildings revitalized London at around the time of these reforms. We do not know how much of this architecture was the work of a newly empowered civic administration, borrowing ideas from military engineers, and how much the product of a direct imperial patronage, but the scale of operations hints at the involvement of military engineers and transports. Severus’ confiscation of estates from Albinus and his supporters may have won funds that could be lent to such building projects, and may have secured access to grain, oil and wine that could then form part of the annona. ⁶⁵ Severus was supposedly the first emperor to establish regular distributions of oil at Rome, increasing state control over processes that previously relied on market arrangements. Similarly conceived policies may have inspired London’s restoration, in particular at the port. This investment in improving the flow of goods, whether through imperial estates or other private domains, may have prompted changes within London’s hinterland, accounting for sporadic evidence of early third-century building works within the estuary and Weald. ⁶⁶