6

A supply-base (c. ad 43–52)

After the conquest

The discovery of a large defended enclosure of likely Claudian date beneath the Roman city has revived the proposition that London was built in the summer of ad 43 to house the invasion forces. This hypothesis accommodates new-won evidence and is consistent with both historical sources and topographic context. The Roman army set up a camp somewhere on the banks of the Thames and nowhere was better suited for their purpose then the Cornhill site.

Whilst it is of more than passing interest to fix the point and purpose of London’s first creation, the argument in favour of a military origin doesn’t require us to abandon the idea that other processes helped make London a town. Far from it. The evidence of London’s earliest Roman stratigraphy shows this to have been a complicated process: involving at least two phases of major investment that gave London its essential shape. There were significant differences in the scale of ambition of these two exercises, but both involved planned components suggestive of central direction. The argument developed here, presented as a model for future testing, is that an initial foundation of c. ad 48 made London an important supply-base before it was remodelled into a fully fledged town c. ad 52.

The disuse of the defensive enclosure on Cornhill is our starting place. If we are correct in identifying this as a military site, it makes sense to see the slighting of London’s defences as an act of decommissioning when the army departed (p. 53). It is also likely that any pontoon bridge over the Thames would have been dismantled, allowing the redeployment of the boats involved in forward campaigns. It is hard to believe, however, that London simply ceased to exist. The strategic value of the Thames crossing was increased, not diminished, by Rome’s advance, and the bridge could have been replaced by a ferry pending more permanent arrangements.

We have no certain archaeological evidence of London in the 5 years after the conquest, but continuity of occupation is likely at late Iron Age sites in Southwark and Bermondsey. There is a suggestive concentration of early Roman military finds at 15–23 Southwark Street, and London could easily have been controlled from the south-bank in these years. ¹ This would account for a higher representation of Claudian coinage in Southwark’s assemblages, as compared to sites north of the river, finding parallel in conquest-period supply ports at Fishbourne, Fingringhoe, Sea Mills, and Richborough. ² The creeks around Southwark remained a sensible destination for shipping coming up the Thames, at a secure point for organizing the inland movement of goods and supplies. There may, however, have been a hiatus in the occupation of the Cornhill site on the north side of the river. Some temporary structures might belong to this period, such as a timber building found at the base of the archaeological sequence at 5–12 Fenchurch Street and ditches pre-dating the earliest streets at Plantation Place, but the evidence is inconclusive. ³

The streets of London

In contrast we have a more substantial body of evidence for what came next. Lacey Wallace has made a close study of the evidence for the earliest permanent settlement on the north bank. Using conservative criteria, she identified nearly 3,000 archaeological assemblages pre-dating the Boudican fire from fifty-six different sites. ⁴ The earliest activities were associated with clearing and draining the land, and quarrying gravel for building the first streets. These quarries were dug without regard to the needs of later users, creating a pockmarked landscape that needed landscaping and consolidation before house-building could proceed. This was a lazy choice, since it would not have been difficult to relocate quarries into areas destined for back-plots. We can conclude that the road engineers were largely indifferent to the needs of the eventual owners of properties established along these streets, suggesting that engineering took place before property interests were fixed or by powers remote from the concerns of those who settled here.

The best dating evidence for these first streets comes not from the core of the settlement, but from one of the main approach roads. Excavations at One Poultry, 50 metres west of a Walbrook crossing, found metalled surfaces of the main west road. ⁵ The earliest surface was laid over timber culverts that diverted water from springs rising in the Gresham Street area. These culverts were part of the primary road engineering and incorporated at least one timber felled in the winter of ad 47/48. Others felled the following spring or summer were reused in later repairs. These timbers were procured for road-building in the ad 48 construction season, presenting a credible date for the creation of London’s first permanent street system.

According to Wallace the culvert cut into a deposit that contained products of the first Romano-British pottery kilns, including locally manufactured Verulamium region wares. ⁶ This might imply that these kilns started production immediately after the conquest. But the excavation report is less definitive, describing sherds of a Verulamium whiteware mortarium from adjacent landscaping. This landscaping might have followed the insertion of the culvert, allowing pottery production to have commenced c. ad 48 in time for a new-made mortarium to be introduced in a foundation ritual. Rubbish beneath this first road included leather panels likely to have come from tents, perhaps those of the army of conquest. ⁷ The likeliest source of this material is the defended site east of the Walbrook, reaching Poultry by means of a crossing established before the metalled road surfaces were laid.

At this date most roads were army-built, but since the focus of military attention lay elsewhere, engineering capacity would have been stretched. By c. ad 48, however, Rome had invested in the construction of a road system that consolidated London’s position at the heart of the transportation network. Streets converging on the Thames linked ports in the south-east with areas of military deployment to north and west. The road from Richborough to the Thames formed the trunk of this communication system which branched once the river was crossed, heading to pre-Roman centres at Colchester and Verulamium (St. Albans), and areas of military operations in the Midlands and Wales. The road to London Bridge has been investigated in Southwark, where it was laid over log causeways across lower areas and buildings erected alongside are dated c. ad 50. ⁸ Pottery assemblages from roadside sites at Brentford and Old Ford show similarities with those associated with the initial Roman presence in London, including grog-tempered fabrics following Iron Age traditions. ⁹ These finds may shadow conquest phase activities, as Rome secured important river crossings along routes where metalled roads were later built.

The need to cross the Thames and bring supplies inland gave London abiding strategic importance. The point where the roads came together to cross the river was the hub. ¹⁰ The main streets within the settlement formed a central T-junction, taking the road that climbed Cornhill from the site of London Bridge as its trunk (Fig. 6.1). This was London’s main north-south axis throughout its Roman history: identifiable as a cardo maximus if we follow modern misuse of the term imported from Roman surveying. ¹¹ It rose to meet an equally important east-west axial route, similarly described as the decumanus maximus. When first laid out this road was no more than 4 metres wide which was barely sufficient for two-way traffic. ¹² The roads were primarily designed for waggon travel, facilitating the overland movement of goods. River transport along the Thames would have been equally important, allowing ships to reach deep into southern Britain. There was, however, no contemporary investment in waterfront facilities: Roman harbours could be simple affairs, where larger ships anchored mid-stream off-loading their cargoes into smaller vessels that beached on the foreshore. ¹³

Fig. 6.1 A proposed reconstruction of London c. ad 48 (based in part on Wallace 2016). Drawn by Justin Russell.

The vast quantities of grain transported by the army required marshalling yards and storage-facilities where hauliers could congregate. In the absence of waterfront facilities, this may have taken place more centrally. An area about half the size of a football pitch (at least 25 metres across and 37 metres deep) north of the central T-junction, had been de-turfed and levelled in preparation for a hard-wearing gravel surface. ¹⁴ Lacey Wallace has estimated that its construction involved between 180 and 380 person days’ work. ¹⁵ This was the heart of London, where a monumental forum was later built. The central hardscape was integral to London’s original design, establishing a place where people, goods, livestock and vehicles could assemble. ¹⁶ Early London had no evident need of the grand public buildings usually associated with city centres: the open space mattered more than the buildings around it. Excavations at 168–70 Fenchurch Street, east of the metalled area, showed that the first buildings were modest timber-framed and earth-floored structures, separated by narrow north-south alleys barely wide enough for a single cart to pass. ¹⁷ These were probably built before c. ad 52. An early U-shaped ditch, up to 1.8 metres deep, found west of the gravelled area at Lombard Street might have formed a sacred boundary (temenos) to a ritual or administrative enclosure west of the forum hardstanding. ¹⁸ A parallel can tentatively be drawn with the larger late Iron Age or early Roman enclosure beneath the forum at Verulamium, the purpose of which is also unknown. ¹⁹

In addition to the narrow backstreet found east of the central hardstanding, a 3- to 4-metre-wide north-south road dating to the first phase of settlement was found at Plantation Place. ²⁰ This lay close to the eastern boundary of the settlement and was perhaps a perimeter road similar to the intervallum road of a Roman fort. A row of narrow timber buildings, each 25 metres long and 3.5 metres wide, was erected alongside. Almost all of the early buildings and property boundaries within this Cornhill settlement shared the strictly orthogonal orientation established by the principal streets. London was laid out on a regular grid from the outset.

At the gates

Shallow ditches dug along the line of the slighted defences at The Walbrook suggest that the settlement boundary was reasserted before ad 60, although no longer as a defensive feature. ²¹ This perimeter, perhaps defining something akin to a city pomerium (the space beyond the walls), would have held functional and symbolic significance: the ritual marking of boundaries gave cities a sacred identity from which threatening impurities could be expelled. ²² The establishment of a formal boundary might account for the way in which roads into London narrowed on entering the orthogonal core. Whilst the roads approaching London were 6–14 metres wide, those inside the initial settlement were generally only 4 metres wide. ²³ A bottle-neck occurred where traffic congregated at entrances, suggesting the presence of gates. If so, London’s main west-gate was near the site of St Stephen Walbrook and its east-gate beneath Minster Court. These locations, equidistant from the central T-junction, marked the limits of the orthogonal settlement. ²⁴ Traffic was notoriously difficult to manage in Roman towns, and gates created checkpoints to restrict access. ²⁵ The wider approaches allowed waggons to halt beside the road during tax inspections or to transfer cargoes onto pack animals better suited to narrow streets. ²⁶

The broadways into London were flanked by large V-shaped ditches with a characteristic square ‘ankle-breaker’, or cleaning-trench, at the base. ²⁷ The ditch on the north side of the Colchester road was set nearly 20 metres back from the metalled street surface. It had a brief life and was backfilled before being covered by buildings destroyed in ad 60. Finds from its fills are consistent with a Claudian date and include the bone handle-grip of a legionary sword. Claudian ditches also chased the road west from London over Ludgate Hill. These too were set back from the metalled surfaces, leaving a gap of at least 6 metres. At one point the ditch was preceded by a fence or hedge that similarly separated the street from its surrounds. ²⁸ The ditches improved local drainage and provided quarried materials for the raised streets, but were over-engineered for this alone. They also protected traffic within a wider corridor that may have included livestock droveways flanking the metalled highway. Other ditches may have defined temporary annexes on the borders of the settlement, similar to those outside the early fortress at Exeter. ²⁹ The areas outside the orthogonally planned core were sensible locations for compounds where livestock could be corralled. This was perhaps the motive for the excavation of V-shaped ditch set perpendicular to the Colchester road, some 60 metres to its south, at Northumberland Alley. ³⁰ A row of pits along the east side of this ditch might have supported a palisade for added security.

Two unusually early burials on the borders are sometimes treated as evidence of late Iron Age occupation. If so, they might have been associated with the pre-Roman farmsteads described in the Chapter 4. It is more likely, however, that they belonged to the first phase of Roman settlement. One occupied a shallow grave on the southern island next to Borough High Street. ³¹ The body of a man, perhaps between 26 and 35 years old based on the evidence of tooth wear, had been placed in a flexed position with his head at the north end of the grave raised to face south. The grave contained no funerary goods, but was sealed by a layer containing early Roman pottery. According to interim reports this was cut by three parallel rows of post-pits then buried during the construction of the main road to London Bridge no later than c. ad 55. Finds from the site included a bone wrist-guard of late Iron Age type. ³² The burial was placed beneath the main road into London at a point where it deviated as it passed a river inlet and took a new heading towards London Bridge. The post-pits may therefore define an important structure, perhaps a gate. It is consequently possible that the burial was an early Roman foundation deposit facing those approaching London from Kent.

The other early burial was found at the Tower of London. The body of a youth, perhaps aged 13–16, was placed with his legs flexed in a shallow grave cut into foreshore deposits containing abraded Iron Age pottery. ³³ A Carbon-14 date of ad 70+/-70 leaves it uncertain if this was a late Iron Age or early Roman burial but the presence of a copper alloy finger-ring of Roman type hints at a post-conquest date. It is notable that both of these unusually early inhumations were buried in a flexed position. It is also worth drawing attention to the importance of the Thames as the main route into London. Might the foreshore burial have also been a foundation deposit placed to face those arriving into London?

Who built London and why?

The decision to build a hardstanding at the site of the forum and to set out carefully protected broadways and compounds beyond the gates was integral to the design of the post-conquest settlement. This emphasis on traffic management and security betrays something of the thinking behind London’s establishment. Preparations for road-building appear to have been put in hand shortly after the arrival of a new provincial governor, Ostorius Scapula. Scapula is thought to have reached southern Britain late in ad 47 and is credited with shaping the new territories into a Roman province. ³⁴ His objectives were tied-up with plans for further military action in the Midlands and Wales. Scapula’s frontier wars relied on supply routes along Watling Street, and this gave London considerable strategic importance. This was the place where coastal traffic intersected with roads that headed inland. The decision to invest in London’s facilities, perhaps made in the winter of ad 47–8, anticipated the needs of following campaigns. It may also have been part of a wider programme of investment in the development of the infrastructure of the pacified parts of the province following the suppression of an Icenian revolt described as having occurred in ad 47. According to Tacitus, Scapula also established a veteran colony to replace the fortress at Colchester, following the redeployment of LegioXX in ad 49. This is also a likely date for the foundation of a chartered town at Verulamium and the establishment of London may have been part of a concerted programme of urban foundation at this time. ³⁵

It is open to question, however, whether the London of c. ad 48 was intended as a town. It lacked the civic and elite architecture that normally formed part of the urban package, although this may have been intended to follow. The early site had more in common with the settlement at Richborough, which is usually identified as a supply-base. This too was a place where a Claudian defensive enclosure was slighted, probably in the early 50s, before a regular grid of metalled roads was laid out. Timber warehouses and granaries that flanked these streets were used for the storage and distribution of supplies needed by the Roman army. ³⁶ Ptolemy was later to describe Richborough and London in identical terms and there is no evidence to suggest that the two sites were differently conceived or held different status at this date. ³⁷ It is only because of their divergent later histories that they are understood so differently. In the early 50s, both places primarily served the needs of Roman military supply.

Roman military campaigns commonly involved establishing operational bases behind the frontier, where equipment and provisions could be stockpiled for onward movement. River ports were inevitably favoured, and the practice of converting abandoned fortifications for this use was well established. ³⁸ As Jonathon Roth has shown, these bases served a variety of ancillary functions. They contributed to the manufacture and repair of armour, were places where cavalry remounts were stabled, housed the army’s money and documents with the personal baggage of commanders and officials, and were holding sites for hostages and captives. This made them into secure locations where forces could be gathered before campaigns were launched. These functions required an investment in the settlement infrastructure and gave such bases a distinct urban character, regardless of their formal standing.

It is likely that London’s main roads were built by the army. Absent any pre-existing self-governing community, it is difficult to recognize any alternative authority with the need or means to take oversight over the exercise, and the engineering skills employed were those learnt in military service. Objects found beneath the metalled surfaces of the first roads into London may have held symbolic and votive significance, perhaps placed as foundation deposits rather than casually lost or discarded. ³⁹ These included a fist-and-phallus amulet, a symbol of masculine virility used to ward off the evil eye and popular with Roman soldiers, found beneath the surface of the main road through Southwark. A young sheep or goat buried under the earliest buildings at Fenchurch Street, east of the central hardstanding, may also have been a sacrificial offering of this period.

The decision to build London required the support of the provincial government, perhaps facilitated by the fact any redundant military site would have been owned by the emperor. The initial programme betrays signs of haste, both in the distribution of gravel quarries and in the way that the town centre needed extensive re-planning soon after. This is consistent with the unsettled circumstances that faced Scapula on his arrival. The likely involvement of army engineers in setting out London’s first streets, and the use of the site as an operational base, does not make London a military site and the distinction between military and civilian is not straightforward. Private contractors (publicani and negotiatores) were closely involved in the supply of the army, particularly in providing sea transport. Roman commanders benefitted from the support of private parties, including those attracted to the lucrative opportunities of public contracts, whose involvement reduced demands on a short-handed colonial administration. ⁴⁰ For example, Caesar placed a private citizen, C. Fufius Cita, whom he described as a businessman (negotiator) in charge of logistics (res frumentariiae) during his Gallic campaigns. ⁴¹ Many recent studies of Roman London have emphasized the importance of merchants and civilian traders to its early history. ⁴² This argument has been championed by Martin Millett, in his reluctance to ascribe agency to the military in the Romanization of Britain (p. 29). In Millett’s view the principal dynamic for the growth of London came from a community of opportunistic Gallic traders, although he recognizes that London must have housed many soldiers. ⁴³ These issues will be explored further below, but an important distinction needs to be drawn between understanding how London came into being and how it was populated. As Mike Fulford has observed, whilst the majority of London’s inhabitants might have been civilians their reason for being there was dictated by imperial policy. ⁴⁴ The argument developed here is that London was designed to meet the needs of army supply rather than for the convenience of traders, although these interests were so closely aligned that the distinction is largely academic.

London in the geography of Roman Britain

The Roman conquest changed Britain’s political geography to London’s advantage. London was placed on borderlands between pre-Roman polities, where the absence of a pre-Roman seat of power gave the new colonial authorities considerable operational freedom. The infrastructure of the new province converged on the optimal site for the economic and political domination of southern Britain. There were few powerful landowners and no population centres whose interests needed accommodating. Officials and merchants could operate with greater latitude than elsewhere. Land could be annexed and impositions made without disrupting the existing social order. These freedoms reduced the risk of conflict, and increased the rewards available to new settlers.

Rome had long recognized the value of borderland riverside sites. This finds ancestry in the origin stories of Rome itself, founded on the Tiber frontier between Etruscan and Latin peoples. ⁴⁵ The choice of the site of London finds closer parallel in the colonial foundation at Lyon (Lugdunum). Here an earlier Gallic settlement at the confluence of the Rhône and Saône gained political centrality through the conquest of Gaul, resulting in the foundation of a Roman colony by the governor Plancus to house a community of citizens expelled from Vienne. ⁴⁶ The city was subsequently developed into the main base from which the Roman forces on the Rhine were supplied, and transformed into an emporium by Agrippa and Augustus in preparation for campaigns in Germany. ⁴⁷ This involved the construction of new facilities on the navigable rivers and a network of roads, largely for military purposes. The main difference is that at Lyon these functions accrued to a formally constituted city, designed to accommodate resettled citizens. There was no need to give London an equivalent political identity. Other Roman riverside cities at politically liminal locations included Cologne (Colonia Claudia Ara Agrippinensium) and Mainz (Moguntiacum) on the Rhine, and York (Eboracum) on the boundary of the Brigantes and Parisi at the confluence of the rivers Ouse and Foss. ⁴⁸ Nicholas Purcell has described how these rivers were neutral spaces between debatable terrains, but also arteries in the inter-regional negotiation of trade, supply and government. ⁴⁹ In all cases the cities were used as instruments of imperial domination, but drew authority and legitimacy from the potency of their river locations.

London was in many regards an unusual case. Most Roman cities drew either on the urbanizing ambitions of pre-existing communities, or came into being as organized settlements of colonists supported by land grants within a surrounding territorium. Some sites, Lyon and Colchester amongst them, could be both: where a new colony was settled onto an earlier centre. London appears to have been neither. As a consequence, we can imagine a more tenuous link between land-ownership and social power than was normal. If London came into being without the involvement of a resident, land-owning elite society this helps to explain the later vulnerability of the site to cyclical growth and contraction. Government offices and business interests enjoyed fewer ties to the land, and were naturally transient.