4

Before London

Landscapes of origin

There is a persistent literary fancy that would make London a more ancient and venerable place than the evidence allows. ¹ Such wishful thinking, however poetic, has no basis in fact. There is no archaeological trace of Brutus and his fabled fleet, no Celtic fortress to explore on the banks of the Thames. Our problem, as John Kent so perceptively observed, is that ‘we are so used to thinking of the site of London as destined by nature to be the focal point of England’s political and economic entity that it requires a considerable effort to envisage those times when it was otherwise’. ² We need to remember that the gently flooding Thames was a formidable frontier. In their description of the medieval river, John Baker and Stuart Brookes show how it both defined and divided territories, carving a boundary that forced overland routes towards rare crossing points. ³ This medieval landscape owed a considerable inheritance to the pre-Roman geography. London only became a commanding site when the competing kingdoms of southern Britain were brought under unified government by the force of Roman conquest. Because of this the story of London starts, inevitably, with Rome.

The fact that there was no pre-Roman city does not, however, leave us contemplating a prehistoric wilderness. The political revolutions of the late Iron Age reconfigured a landscape that was already ancient. ⁴ Stone tools show that people arrived in Britain during the early Stone Age, or Lower Palaeolithic, with anatomically modern humans entering the scene some 38,000 years ago. Some of the first people to glimpse the Thames were transient hunter-gathers whose Mesolithic sites have been identified close to London in Bermondsey. Farming communities were slow to establish themselves, but there is patchy evidence for their presence from c. 4500 bc. These pioneers left trace in north Southwark in the late Neolithic and early Bronze Age, where buried land surfaces preserve marks left by the wooden ards they used to till the land. A network of wooden trackways eventually crossed the estuary wetlands, and the Thames became a focus of ritual activity. ⁵ Bronze Age swords and weapons recovered from its waters are the likely legacy of votive ceremony, as may also be the case for the prehistoric human crania occasionally found on the river bed.

Pollen indicates that fields were cleared for small-scale arable farming in a lightly wooded pastoral landscape around the turn of the first millennium bc, a full 1,000 years before the Roman invasion of Britain. ⁶ Late Bronze Age ring-forts at Carshalton and near Heathrow became regional centres of power. ⁷ Some pottery of this period has been found on sites within the City where it is associated with pits and occupation surfaces. These discoveries were mapped by Nick Holder and David Jamieson, who identified three clusters that might mark the sites of farmsteads buried beneath the Roman town. ⁸ The fact that sufficient has been recovered to identify these Bronze Age sites, but nothing so tangible to attribute to the following centuries, shows that settlement densities reduced after a Late Bronze Age peak. There was no continuity of occupation, and the City of London and its environs were sparsely populated throughout the Iron Age.

Metropolitan London was born of the Thames: raised into being by a topography that made this the lowest point where a fixed river crossing was readily engineered. The dark river, the Roman Tamesis, rises west in the Cotswolds whence it flows east to the sea. London itself sits some 80 kilometres (50 miles) inland, close to the river’s tidal head. The Thames, navigable deep inland, afforded ready access to England’s coast, facing the European Continent where the busy Rhine emptied into the North Sea. Its estuarine reaches benefit from twice-daily tides that draw rivercraft inland and, by turn, flush them seawards on the ebb. The earliest Roman foreshore at London Bridge is rich in species of diatom that only thrive in brackish water, indicating that the river was tidal to this point. ⁹ Similar evidence shows that it had also been tidal a little further upstream at Westminster in the Bronze Age, although this may no longer have been the case by the early Roman period since water snails on the Roman foreshore belong to freshwater species. Low tide at London Bridge must have been below OD in the first century ad, since quarry pits were dug into the foreshore at this level, whilst waterfront structures suggest that land above 1 metre OD was expected to escape flooding. ¹⁰ The tidal range was therefore between about 1 metre below OD and 0.5 metres above OD, although spring tides would have extended this by an extra 0.5 metres in both directions.

The meandering river settled to its present position well before the Roman conquest, but occupying a much broader channel than it does today. At low tide the Roman Thames was 275 metres wide at its narrowest, compared to the modern embanked river which is only 200 metres across. At flood tide it spread a kilometre wide, drowning mudflats on the south bank. Critically a few sandy gravel eyots stood proud of the tide, separated by braided river channels that carved inlets and creeks. Two areas of firmer ground, now buried deep beneath Borough High Street in Southwark, formed a pinch-point in the river channel, above and below which the river widened considerably (Fig. 4.1). These precious islands became stepping stones to London Bridge: the lowest fixed crossing of the Thames in Roman and Medieval times.

Fig. 4.1 The rivers and topography of central London before the city was first established (based in part on work by David Bentley, amended to take account of research by Myers 2016 and Ridgeway et al. 2019). Drawn by Justin Russell.

Opposite, the Thames cut directly against two brickearth capped gravel hills. Geologically these are known as part of the Taplow terrace, with the brickearth belonging to the Langley Silt complex. David Bentley’s interpretation of an 1841 survey of surface relief, supplemented by modelling from archaeological and engineering investigations, allows a detailed reconstruction of the pre-settlement topography. First published as the Ordnance Survey map of Roman London, this is a vital starting point in understanding the Roman settlement. ¹¹ The City was contoured by three watercourses which dissected the drift geology to form two steeply sloped hills overlooking the river. The western hill, Ludgate Hill, now crested by St Paul’s Cathedral, had a summit at c. 13 metres OD. This was flanked by the River Fleet to its west and the Walbrook stream to its east. The eastern hill, rising to about 12 metres OD, lay between the Walbrook and a smaller stream known as the Lorteburn. ¹² This area, aptly known as Cornhill, is now home to Leadenhall market.

The Fleet, with its headwaters on Hampstead Heath, was the largest of these rivers flowing into the Thames, occupying a substantial valley that is still easy to follow on the ground. The lost Walbrook, which bisected the Roman city, has been the subject of several recent studies. ¹³ Stephen Myers describes two principal branches of the river that came together north-east of Finsbury Circus. A substantial western stream, 3.4 kilometres long, originated in springs at Barnsbury and Highbury Fields on the Islington ridge. A slower-flowing eastern tributary was fed by springs at Hoxton and Holywell Priory in Shoreditch. The combined river then flowed through the city for a distance of approximately 0.9 kilometres, receiving additional water from springs rising near the Barbican and around Gresham Street, before entering the tidal Thames at Dowgate, just west of Cannon Street Station. This lower stream was originally up to 7 metres wide, although later reclamations pressed it into a 3-metre-wide channel. Insects and plant materials show that the upper Walbrook was a slow moving watercourse before London was built. Areas of rough pasture with scattered trees flanked the river, where the presence of dung beetles suggests that herbivores grazed. Pollen shows that alder grew along the valley floor, where the marshy banks included bracken, rush and sedges. The wider landscape included woods of oak with some hazel on the higher and drier slopes, but was dominated by long grass pasture. Some local cereal cultivation was also in evidence. South of the river, in Southwark, was a low lying marshy fen-type environment with little woodland cover, but areas of oak dominated woodlands, with some alder carr, lay further to the south. ¹⁴ This was the landscape that awaited Rome, at a time when the climate was probably a little warmer and drier than it is today. ¹⁵

Southern Britain before Rome

Although London wasn’t established until after the Roman conquest, southern Britain’s pre-conquest political geography is a proper starting point in seeking to understand how and why it came into being. The late second and early first centuries bc witnessed busy contact between south-eastern Britain and north-eastern Gaul, at least at the level of aristocratic society. ¹⁶ New funerary practices involving urned cremations were adopted in some regions, alongside the use of wheel thrown and grog-tempered pottery that imitated Gallo-Belgic types and may have been introduced by immigrant potters. ¹⁷ Italian wine and Italic-type bronzes were imported, and new fashions of dress from the near continent involved a distinctive use of brooches (fibulae) as dress-fasteners. Collectively these features are known as the Aylesford–Swarling tradition, itself part of the wider late La Tène cultural package, which came to characterize elite sites in Kent, Essex, and Hertfordshire: territories later identified as Trinovantian and Catuvellaunian. Goods connected with feasting were given emphasis, highlighting the importance of hospitality in building networks of clientship. Whilst we cannot rely on Roman sources to have understood or accurately described the societies that they encountered, Caesar’s description of patron–client relationships in Gaul seems credible. ¹⁸ Politics involved factional rivalries and alliances within and between ruling dynasts supported by armed retinues. Payments of gold and precious metal, in the form of high-value coinage, was used to reward loyalty or rendered as tribute. These transactions account for the presence of continental coins in Britain from c. 200 bc, and for the British copying of Gaulish prototypes from the middle of the second century bc onwards. ¹⁹

Julius Caesar’s description of the settlement of Belgae in coastal Britain before the Gallic wars has been seen as a possible stimulus for the introduction of aspects of Gallic social practice represented by the Aylesford–Swarling package. ²⁰ The concept of a single, large-scale, Belgic invasion has, however, long fallen from favour. The changes in southern Britain took place over a longer time-frame, involving an extended history of cross-Channel interaction. ²¹ The fact that we have no convincing evidence for pre-Caesarian mass migration should not, however, blind us to the fact that there is abundant evidence of people moving between Britain and Gaul before, during and after the Roman invasions. Alliances and affiliations, as well as rivalries and hostilities, extended across the English Channel with ease. These connections involved immigration from Gaul to Britain c. 25–15 bc, and a network of clientage and kinship relations between Britain and Gallia Belgica. This made the aristocrats of northern Gaul important intermediaries in the spread of Mediterranean influences into Britain, both before and after the Roman conquest. The subsequent reception of Roman social practices in southern Britain, and adaptation to Roman political rule, were influenced by these established links with Gallo-Belgic society.

Caesar’s campaigns in Gaul brought Roman armies to the English Channel, resulting in his invasions of 55 bc and 54 bc. ²² Until recently no certain trace of these expeditions had been identified, but excavations by the University of Leicester at Ebbsfleet in Kent between 2015 and 2017 found a large defended enclosure similar to Caesarian fortifications in Gaul and associated with military equipment of the right date. ²³ Caesar’s expeditions enhanced his reputation in Rome, as was their purpose, and created a model for the subsequent use of British conquest to burnish the credentials of imperial rule. They also established the principle of Roman authority in Britain, giving local rulers reason to value Roman protection. ²⁴ Caesar described the political groups he encountered in southern England using the term civitates. Tacitus’ subsequent descriptions of Germanicus’ campaigns in Germany mentions Roman contact with the regules (little kings) of Southern Britain who ruled over gentes (peoples) such as the Iceni or the Silures. ²⁵ These pre-Roman polities are sometimes referred to as ‘tribes’ in modern accounts, although the appropriateness of this term is questionable. ²⁶ The sources suggest that these peoples were sometimes ruled by kings whose power could be inherited within a lineage, and who were able to enter into alliances, raise armies and conduct wars.

Caesarian arrangements laid the foundations for the emergence of two major powers in southern Britain, sometimes termed the Eastern and Southern kingdoms, which came to prominence in the Augustan period. ²⁷ Distributions of coins issued by the pre-Roman rulers of southern Britain have been used to reconstruct changing areas of territorial influence and suggest lines of dynastic succession, although might be better viewed as illustrating fragmented sets of social networks rather than territorially coherent entities. Their design showed a close awareness of Roman political and religious symbolism, and John Creighton has suggested that British kings were recognized as clients of Rome where members of the ruling dynasties had spent time as guests or hostages. ²⁸ Some rulers adopted the Latin title rex, and used Latin to claim descent from earlier rulers in apparent confirmation of political arrangements described by Roman sources. Coins were minted at sites of political authority that the Romans knew as oppida. These were usually multi-focal sites where banks, or dykes, enclosed relatively large open areas and including sites of elite burial, ritual activity and craft production.

Coin distributions and Roman texts shows that the ‘Eastern kingdom’ was associated with the House of Cunobelin, with important oppida at St Albans (known to the Romans as Verulamium of the Trinovantes), and Colchester (Camulodunum of the Catuvellauni). Patterns of elite consumption and burial within this region were remarkably similar to those of Reims and Trier in northern Gaul. ²⁹ Suetonius described Cunobelin as king of the Britons (Rex Britannorum), implying that he ruled various peoples. ³⁰ Colchester may have been the principal centre of his royal power, but important ritual functions also took place at high-status secondary centres, such as Heybridge near Maldon in Essex. ³¹ These were perhaps the seats of aristocrats and lesser kings on whom Cunobelin depended. The Southern kingdom, with its main seat of power at Chichester is generally associated with the Atrebates, ruled by the house of Verica before the Claudian conquest. A city was established at Silchester (Roman Calleva) in the last quarter of the first century bc. ³² The material and biological evidence suggests that this was a planned and colonial settlement, involving immigrants from the northern parts of Gallia Belgica, probably Amiens or its environs.

Late Iron Age London

Coins naming Tincomarus and Verica are distributed across a coherent territory south of the Thames, representing the Southern (Atrebatic) Kingdom, whilst the gold coinage of Cunobelin occupies an area largely north of the Thames but which included northern Kent. ³³ There was little overlap between these distributions, and London was peripheral to both (Fig. 4.2). There is a genuine scarcity of coins of the post Caesarian Iron Age in the London region, where the absence of the coinage of Cunobelin is notable. This is not simply the consequence of low levels of metal-detectorist finds in a built-up area, since other metalwork doesn’t show the same pattern. ³⁴

Fig. 4.2 The political geography around the place where London was founded, as suggested by the evidence of material culture. The distribution of the coins of Tincomarus and Verica marks the extent of influence of the ‘Southern Kingdom’ (Atrebates), whilst the presence of the gold coins of Cunobelin reflects a dependency on the ‘Eastern Kingdom’ (Catuvellauni/Trinovantes) ruled from Colchester (information drawn from Fulford and Timby 2000 Fig. 238). The late Iron Age ceramic zones, shaded, are those identified by Thompson (2015) and suggest the extent of discrete territories within Cunobelin’s ‘Eastern Kingdom’. Drawn by Fiona Griffin.

The Iron Age coins found nearer to London are more usually the cast ‘tin’ potins that pre-date the middle of the first century bc. Their distribution has been used to suggest that a political centre was established near the Thames west of London, perhaps in the vicinity of Brentford or Kew. Occupation in Putney close to the confluence of the rivers Beverley and Thames, represented by Iron Age pits and ditches, is the most recently identified candidate for this hypothetical site. ³⁵ There is no evidence, however, that any such place retained importance beyond the c. 60/50 bc date suggested by the coinage. Some hoards of potin coins found west of London, including one from St James’ Park, are also thought to date from the approximate period of the Caesarian invasions of 55/54 bc. ³⁶ Similarly, whilst there are other defended Iron Age sites along the Thames, including the Middle Iron Age fort at Uphall Camp beside the Roding and a late Iron Age site at Woolwich Power station enclosed by massive V-shaped ditches, these appear to have lost importance by the time of Caesar’s expedition. ³⁷

Despite the difficulties of seeing the Iron Age landscape through the confusion of later metropolitan clutter, we have sufficient evidence to be confident that the dearth of evidence in Greater London reflects on a real scarcity and is not the product of later disturbance or a reduced intensity of research. Grog-tempered pottery, showing Gallo-Belgic influence, has a similar pattern of distribution to Trinovantian/Catuvellaunian coinage, with London on the fringe of its area of use. ³⁸ Aylesford-Swarling style wheel-thrown pottery, of the type abundant in Kent, Essex and Hertfordshire, is largely absent in Greater London, which area has few cremation burials and none with rich grave goods. ³⁹ We can conclude that the London basin was peripheral to the ‘kingdoms’ associated with the Houses of Cunobelin and Verica. ⁴⁰ This in turn suggests that the Thames may have been a contested frontier zone between the emerging polities. Substantial earthworks known as Grim’s Dyke that lie close to Brockley Hill might have originated as a late Iron Age territorial boundary, possibly the southern frontier of polity based on the pre-Roman oppida at St Albans (Verulamion), but the dating evidence is inconsistent and an early medieval date is perhaps more likely (below p. 396). ⁴¹

London may have been a backwater before the Claudian invasion but it was no wilderness. Palaeoenvironmental evidence, especially from Southwark, describes a pre-Roman agricultural environment involving some cereal cultivation. There is a growing body of evidence for the rural settlements from which the area was farmed, some of which were occupied at the time of the conquest. Our reading of this landscape is complicated by the fact that pottery assemblages following Iron Age traditions are found in contexts associated with the first Roman presence in London (pp. 41–2 and 53). These include rare examples of grog-tempered jars and other forms normally associated with the late pre-Roman Iron Age. ⁴² Precise dating is difficult because these kinds of pottery continued in use until the Flavian period, and in many cases a post-conquest origin can be surmised from the contexts in which they were found. It is consequently difficult to use the distribution of these finds to map the distribution of pre-Roman settlement.

There are sites on the south bank of the river, however, where people were living around the time of the conquest (Fig. 4.3). Excavations near Bermondsey Abbey have located the likely site of a small Iron Age farmstead represented by pits containing Iron Age pottery and loom-weights. ⁴³ Bermondsey eyot, perhaps an island at high tide, lay about 1 kilometre south-east of the later site of London Bridge and appears to have been occupied at most periods from the Middle Iron Age onwards. Louise Rayner’s study of the pottery identified several vessels of Gallo-Belgic type, along with wheel-made vessels copied from imported Gaulish beakers. ⁴⁴ Sherds in greensand-tempered fabrics are thought to be products of the Medway valley of a type that had gone out of use by the conquest. The range of types, and the absence of the Romanized products normally found in London assemblages after c. ad 50, suggests that these features date within the first half of the first century ad. Evidence of contemporary occupation has been found on the neighbouring Horsleydown eyot. ⁴⁵ Two parallels ditches that defined a trackway leading towards the river contained late Iron Age pottery likely to date to the early first century along with cereal remains suggestive of local arable farming. The presence of a couple of pits, one of which contained a large storage jar in late Iron Age or early Roman North Kent shell-tempered ware, and an oven are indicative of a small settlement occupied around the time of the conquest.

Fig. 4.3 Late Iron Age settlements around the site where the Roman city was founded. It is not certain that all were occupied at the time of the Roman conquest. The shaded areas flanking the principal rivers include both an inter-tidal zone and water-meadows subject to rarer flooding. Drawn by Justin Russell.

Finds of this approximate date have also been found in Southwark, chiefly on its northern island. ⁴⁶ Most come from Roman deposits and may have been introduced after the conquest, as is the case for the seven Iron Age coins found here. One of the more recent discoveries is a potin coin of the Cantii, likely to have been issued in the first century bc. ⁴⁷ More certain evidence of pre-Roman occupation was identified at 15–23 Southwark Street. ⁴⁸ This site lay on the southern side of the north island beside a small tidal creek that later housed a Roman wharf. It attracted settlement as early as the Bronze Age, generating a rare and important group of Beaker period pottery. Later pre-Roman activity was represented by a ditch, gullies, and timber constructions that included a circular structure. The gullies contained late Iron Age grog- and sand-tempered pottery, whilst pottery reworked in Roman deposits included coarse fabrics likely to date to the Middle Iron Age. The stratigraphic complexity of the earliest features allows the identification of several different phases of pre-Roman settlement and a bronze coin of Greek Massalia came from one of these features. ⁴⁹ The bronze fittings from the leather sheath for a Roman dolabra, or military pickaxe, were found in the upper fills of one of the Iron Age gullies. ⁵⁰ John Creighton has drawn our attention to the possibility that it reached Southwark before the conquest after the adoption of Roman-style military dress by the upper echelons of pre-conquest society. ⁵¹ Such items could have accompanied British or Gallic veterans returning from participation in the Gallic wars as mercenaries and kinsmen. The site at 15–23 Southwark Street is about 1.5 kilometres distant from Bermondsey Abbey so these two Iron Age sites were separate but neighbouring settlements. Southwark was more evidently linked to waterfront activity but cannot be mistaken for a significant port. We simply do not have the range or volume of imported goods, from either the site or its surrounds, to stand comparison with the known trading sites of the period.

Richard Hingley suggests that activities in Iron Age Southwark included the exposure of corpses to the water, in rituals that drew on a history of depositing metalwork and human crania in and around the Thames, which practices could have survived to influence choices made in the creation of the city. ⁵² The sacred properties of watery places, at the interface between material and spiritual worlds, were widely appreciated (below pp. 110 and 254). The river’s tidal head, where fresh and salt water co-mingled under the pull of the moon, was undoubtedly a singular and suggestive place. It is harder, however, to demonstrate any continuity of practice. Most evidence for the manipulation of disarticulated human remains took place on the margins of the Roman city after its establishment, and there is no evidence for late Iron Age corpse exposure in Southwark. The prehistoric votive deposits of metalwork were considerably earlier and concentrate further upstream. Different stretches of the river attracted votive practices before and after the conquest, but there is no evidence to suggest a continuing tradition where London was built.

Some pottery from the north bank might also derive from late Iron Age occupation, but it is difficult to distinguish between material that pre-dates and post-dates the Roman conquest. A pit containing late Iron Age flint-tempered pottery was found at 14–18 Gresham Street, and another, containing a single sherd of possibly Iron Age pottery, was found at Plantation Place. ⁵³ In both cases a cautious reading would see these as early Roman features backfilled before post-conquest production and importation introduced significant quantities of more closely dateable material. This is the likely explanation for the Iron Age character of pottery found in Roman-style ditches at Walbrook House discussed in the next chapter. ⁵⁴ Redeposited and residual finds of similar material have been noted in stream channels in the upper Walbrook valley. ⁵⁵ These suggest early occupation nearby, but not necessarily of pre-conquest date. Abraded sherds of flint-tempered pottery, probably of Iron Age date, were also recovered from river-silts on the foreshore at the Tower of London, with similar material in a large pit dug into the foreshore. ⁵⁶ These finds might illustrate the nearby presence of a pre-Roman settlement, perhaps a farmstead on the south-facing slopes of Tower Hill, but we cannot rule out the possibility that the material arrived here after the conquest. A burial cut into these deposits has also been argued to be a late Iron Age find, although a post conquest date may be more likely (below p. 68). In all cases the quantities involved are miniscule and the dating imprecise, leaving no certain evidence of Iron Age building activity.

An important group of unusually early first-century pottery was found at Pinners’ Hall, at the corner of Great Winchester Street and Old Broad Street. ⁵⁷ Here a shallow ditch, perhaps a field or enclosure boundary, was dug across a plateau of higher ground east of the Walbrook and north of the first Roman settlement. These finds remain unpublished, but according to notes by Barbara Davies the ditch contained a simple shell-tempered bead-rim jar, sherds of friable briquetage and an early product of the Highgate Wood kilns. The presence of hand-made forms, and the use of grog, shell, and flint to temper the clay, are characteristic of late Iron Age assemblages, whilst the Romanized products ubiquitous in London by c. ad 50 were absent. A feature thought to be a contemporary well was found south of the ditch. This might mark the site of pre-Roman settlement, but could equally date shortly after the Roman conquest. This latter option seems marginally more likely.

There is more certain evidence of late Iron Age settlement in further-flung areas. One was found where the church of St James Clerkenwell stands on a hill above the east bank of the river Fleet and a little over 1 kilometre north-west of the urban site. ⁵⁸ Pits and ditches with Iron Age pottery contained fragments of daub from timber buildings. The absence of Early Iron Age forms indicate that the site wasn’t occupied prior to c. 200 bc, and the presence of glauconite-rich or greensand-tempered sherds shows similarities to the Bermondsey eyot material, suggesting that these sites were in contemporary occupation prior to the conquest. ⁵⁹ Building material from later features indicates that occupation continued into the Roman period. This was probably another small Iron Age farmstead that survived the invasion, continuing to exploit the advantages of a well-watered site on a south-facing slope.

Another place with Iron Age origins was probably located some 2 kilometres west of the Roman city at St Martin-in-the-Fields, next to Trafalgar Square, on the slopes of a low hill beside a stream and overlooking a bend in the Thames. ⁶⁰ This site is better known for its late Roman and early medieval finds, but a small pottery assemblage suggests that it may have been settled prior to the conquest. This included a large shell-tempered storage jar and a bead-rimmed chaff-tempered jar, similar to the Southwark and Clerkenwell finds. A timber beam-slot may have been part of a pre-conquest building, although an early Roman date cannot be excluded. These were probably the traces of a pre-Roman farmstead that survived through the Roman period, growing in importance in late antiquity. No other late Iron Age sites are known from London’s immediate environs, with the next nearest some 10 kilometres away in the Lea Valley.

Other farms probably await discovery, but the intensity of fieldwork in Greater London makes it likely that we have a reasonably complete picture of the settlement landscape encountered by Rome. This shows that a few high-status rural sites were established on the south bank of the Thames. These places in Southwark and Bermondsey shared in the material culture of other north Kentish sites fringing Cunobelin’s Eastern kingdom. Both may have been occupied at the time of the Roman conquest and remained important thereafter. Finds from 15–23 Southwark Street might witness the involvement of a local leader in military campaigns in Gaul, where service in the auxilia would have resulted in the adoption of aspects of Roman military dress although these could have reached Southwark in other ways. The evidence from the City, north of the river, is less conclusive. Although a few pottery assemblages with late Iron Age characteristics have been identified, these could derive from satellite activities around the initial Roman presence (Fig. 5.2). A couple of small rural settlements were found a little further away, at Clerkenwell and St Martin-in-the-Fields. These were naturally advantaged sites, on south-facing slopes with sources of fresh water, which accounts for continuities of occupation into the Roman period and beyond (below p. 393). These late Iron Age sites give no hint of the imminent foundation of a great city. They were not places of wealth and power with far-reaching trading connections. In his review of the pottery used around the time of the conquest, Paul Tyers concludes that ‘we have assemblages that echo more than one of the surrounding areas, but without any one of them predominating. Pre-Roman London’s status as an area without a strong indigenous material culture seems to be confirmed’. ⁶¹ The Thames remained a borderland until the arrival of Rome.

There is one last piece of evidence to draw on in our consideration of pre-Roman London: its name. Rome first knew the place as Londinium. The earliest historical citation is found in Tacitus’ description of the revolt of ad 60/61, and a writing tablet addressed to Londinium internally dated to ad 62 was found at the Bloomberg headquarters. ⁶² The etymology of the name leaves little doubt that it was a Latinization of Londonion, which belonged to the British Celtic language that the Romans would have heard at the time of the conquest. ⁶³ The fact that London has a Celtic place name doesn’t make it a Celtic settlement or require the name to have been used before the conquest. According to Theodora Bynon it incorporated a collective-forming suffix and a lexical base which names an object or feature of the local environment. ⁶⁴ Its exact meaning is unknown, and suggestions that it might derive from a British compound of Llyn-don ‘the fort by the lake’ or from a Celtic personal name have not withstood scrutiny. Bynon cautiously proposes a derivation from the Celtic landa, or ‘low-lying open land’. This would, of course, be a name better suited to the mud-flats south of the river than the hills to its north.

Ptolemy, a Greek geographer writing in the second century ad but drawing on sources likely to date to the first century, mentions Londınion as a poleis of the Kantioi, along with Richborough (Rutupiae) and Canterbury (Durouernum). ⁶⁵ Until recently most scholars were reluctant to consider London a Kentish town, since it was built north of the river. ⁶⁶ Martin Millett has, however, been more sympathetic to the idea and opinion has turned in favour of accepting Ptolemy’s identification. ⁶⁷ If communities on the south-bank were friendly to Rome this offers a route for converting a British place name for Southwark into the name of a Roman town on the other side of the river. An honorary arch erected in ad 54 recorded the surrender to Claudius of eleven kings of barbarian gentes, presumably local leaders hastening to preserve their estates from the confiscations imposed on those defeated in war. ⁶⁸ We can imagine that small kingdoms in Kent were fairly easily detached from their loyalties to Cunobelin’s heirs and quick to side with Rome.

Why London was Roman

The reason that London is seen as a Roman foundation is because it lacks the archaeological footprint of other pre-conquest sites. There was no oppidum here to compare with the late Iron Age sites at Colchester, St Albans, Chichester, or Silchester, all of which became important Roman towns. We have no comparable high-status burials or elite-consumption assemblages, and no surrounding distribution of high-value coins describing a sphere of political influence exercised from the site. Reports on pre-conquest imports of arretine ware to London have been shown baseless, the product of spurious attributions made by dealers of antiquities. ⁶⁹ Not only does London compare poorly with the territorial oppida, but it lacks the artefactual signatures and settlement densities of secondary sites within the region such as Heybridge in Essex. In any case, there was no need for a political centre at this location. Prior to the Roman conquest the lowlands of south-east Britain were politically fragmented and the Thames estuary a backwater.

The train of events that lead to the Claudian annexation of Britain had origins in the political settlements that followed Caesar’s earlier invasion, and Rome’s recognition of friendly kings in Britain. ⁷⁰ The ostensible reason for invasion, according to Cassius Dio, was the flight across the Channel of Berikos, usually identified as Verica the king of the Atrebates. This was a pretext for military intervention aimed at restoring Verica to his throne and forcing the submission of the Catuvellauni whose political capital was at Colchester. Claudius’ decision to prosecute this war was also influenced by the fact that military success was the best way for an emperor to establish his credentials as commander in chief and head of state. Rome’s militarized systems, and the need for new emperors to win personal and dynastic legitimacy, made it expedient to follow in Caesar’s illustrious footsteps. Such were the circumstances that brought Roman troops back to Britain, and established the conditions for the foundation of London.