Part 2

Making London

5

The Roman invasion (c. ad 43)

Debating London’s Roman origins

There may be no more contentious issue to be tackled in this book than the story of London’s origin. The site enters written history in descriptions of the Icenian revolt of ad 60/61 and Tacitus, our main source, goes out of his way to inform us that London wasn’t a Roman colony at this time. ¹ This emphasis on what London was not begs the question of what it was: if it was neither a pre-Roman town nor an early Roman colony, how then did it come into being?

Many studies, especially those written after the Second World War, argued that London started life as a fort built during the Roman invasion. ² A detailed account of this invasion is found in the second-century writings of the historian Cassius Dio, drawn from earlier sources. ³ Forces commanded by Aulus Plautius landed on the south-coast of Britain in the summer of ad 43 in a moment that has generated a considerable literature, lately dominated by a lively dispute over whether the landings took place at Richborough in Kent or on the Solent. ⁴ Wherever the army started its British campaign it soon advanced to the Thames, which is where our interest quickens. After forcing its way across the river, the army halted for a few weeks to await the Emperor Claudius who wished to take personal command of the capture of the enemy capital at Colchester.

This pause in the campaign gave scholars reason to seek evidence for a Claudian fort or marching camp controlling the Thames. The idea of a military forerunner to Roman London also seemed consistent with the example of Colchester, where the first town reused the shell of a disused legionary fortress. ⁵ Lands expropriated by the army would have remained imperial property, facilitating their gift to colonists or civic authorities, leaving a legacy of infrastructure fit for re-purposing. The supposition that London had military origins aligned with contemporary interest in the Roman army as a driving force in the pacification and administration of Britain. It was assumed that civilian communities struggled to find the architects, engineers, and craftsmen needed to build new towns, which deficiencies were made good through the military support and patronage of the provincial government. The impetus for the foundation of towns in Britain was therefore seen as a Roman affair: state-driven, top-down and interventionist. Drawing on such expectations, discoveries at London of V-shaped ditches of a style sometimes associated with Roman fortifications were taken as proof of London’s military origins. ⁶ Building on these assumptions, the idea that London came into existence as the site of a military establishment commanded widespread support into the 1970s. ⁷

At this point, however, a combination of new data and new thinking prompted a radical reassessment. Despite the increased intensity of archaeological investigations following the creation of the DUA in 1973, this work failed to isolate features early enough to belong to a conquest phase establishment. ⁸ London’s earliest coin assemblages were dominated by issues known as Claudian copies. These are more typical of places established towards the end of Claudius’ reign, rather than those of conquest date. ⁹ A relative scarcity of finds of military equipment also weighed against a military origin. The information fitted better with a foundation date c. ad 50 and this made it impossible to argue for direct military involvement, since Roman forces were deployed further inland by this time. Since it was also possible to argue that the original Claudian encampment might be found elsewhere in the Thames basin, perhaps south of the river or in Westminster, opinion shifted towards seeing London as a civilian foundation brought into being after the initial military campaigns.

London was, however, in the wrong place to have been founded at the behest of a local civilian community with urban ambitions. As we have seen, there is no evidence that any such community existed. It consequently seemed likely that credit for London’s creation should fall to the merchants and shippers who followed the advancing forces, and whose needs were served by a strategically located port. This argument was first advanced by Francis Haverfield who concluded ‘that London began, not at the nod of a ruler, but through the shrewdness of merchants’. ¹⁰ Peter Marsden, writing in 1980, returned to this hypothesis, suggesting that London ‘did not have a military beginning as a camp built during the Roman invasion of ad 43, but was founded about ad 50 as a carefully planned civil trading settlement of Roman merchants’. This reinforced the view, gaining ground following Moses Finley’s seminal study of the ancient economy, that earlier studies had over-stated the importance of the Roman state as an actor in urban affairs. ¹¹ Towns were instead seen as largely self-governing entities, directed by local elites who self-identified with Roman power and ideology. This removed the need to ascribe any great importance to the Roman military in the creation of towns. Martin Millett’s work, in particular, emphasized how little evidence there was for a significant official intervention. ¹² The idea that Romano-British cities relied on an inheritance of military engineering became difficult to sustain, and could be dismissed as guilty of ignoring the creative capabilities and political interests of other communities. By 1994, Millett could write of ‘a consensus emerging about the development of the city. Uniquely in Britain, the town seems to have grown up as a planned trading settlement of citizens from other provinces within a decade or so of the invasion’. ¹³ As the busy programme of investigations continued, an absence of evidence seemed evidence of absence, leading to the conclusion that excavation had been ‘sufficiently widespread for us to be sure that had a major fort existed we would know about it’. ¹⁴ Thus, a new generation of scholarship found agreement over the combination of events that brought London into being. The city was seen as the creation of mercantile endeavour, inspired and underwritten by the trading opportunities that attended the making of Britannia. Speculation continued over where the invasion period fort might be found. But since this would have been a temporary establishment, perhaps located to control a pre-existing crossing of the Thames, there was no need to look for it beneath a city built at a post-conquest river crossing. ¹⁵

London’s Claudian defences

The argument seemed settled. But against all expectation it was reopened as a consequence of chance discoveries made in 2007. A team led by Ian Blair, working on a construction site at the junction of Cannon Street and Walbrook known as The Walbrook, found ditches that they identified as having formed the west side of a large fortified enclosure built immediately after the conquest (Fig. 5.1). ¹⁶ This surprising conclusion was not the outcome of a purpose-designed exercise in exploratory research, but the product of the routine investigation of a threatened site. Since the idea that London had originated in a conquest phase fort had long been dismissed the study of its early defences did not figure in any research agenda. This was despite the fact that in 1974, Ralph Merrifield and Harvey Sheldon had drawn on topographic evidence to suggest a likely location for the Claudian encampment, predicting that the only evidence likely to survive would be the buried fort ditch. They concluded that Claudian defences would be found at exactly the spot where Blair and his team were later to make their discovery. ¹⁷ What came as a complete surprise in 2007, had been accurately predicted in 1974. A neglected hypothesis had been put to accidental test and supporting proof obtained.

Fig. 5.1 Details of the Claudian features found at ‘The Walbrook’ (WAO06: after Blair 2010), 7–11 Bishopsgate (ETA89: after Sankey 2002) and Regis House (KWS94: after Brigham and Watson in prep.). Drawn by Justin Russell.

The discoveries at The Walbrook reopen the issue of how and why London was built where it was, reminding us that our models need constant questioning in the face of new data. This return to a long-rejected argument has understandably met with scepticism. Lacey Wallace has strongly re-asserted the reasons why the fort-into-town model is suspect, and questioned the reliability of the evidence used to identify Claudian fortifications. ¹⁸ Her arguments, and a continuing hesitancy to challenge received wisdom on this topic, require us to give close attention to the evidence now before us.

The earliest features at The Walbrook were two parallel V-shaped ditches set on the upper slopes of the east bank of the stream. The ditches were about 4 metres apart: that closest to the river survived for a depth of 0.8 metres and was distinctly smaller than the eastern ditch, which survived to a depth of 1.8 metres. Both contained a square-cut cleaning trench or ‘ankle breaker’ at the base, and were comprehensively backfilled shortly after being dug. The crisply defined and un-weathered profile of the ditches, and the absence of any hill-wash sedimentation, makes it unlikely that they were left open throughout an entire winter. The lower fill of the inner, eastern, ditch included the remains of a storage jar in a late Iron Age ‘Romanizing’ grog-tempered ware: a transitional fabric of unusually early date for a London assemblage. Both ditches also contained late Iron Age flint-tempered pottery. ¹⁹ Since these types of pottery remained in use after the Roman conquest their presence is not proof of an early date. The complete absence of the ‘Romanized’ products ubiquitous in all London assemblages after c. ad 50 provides a more compelling reason to believe that the ditches were early. Only in unusual, and improbable, circumstances would a collection of broken pottery assembled after c. ad 50 remain wholly free of examples of post-conquest imports and local Roman products. An early Claudian date is also implied by the extended sequence of pre-Flavian activities that took place after the ditches were filled. These activities included quarrying on the banks of the Walbrook, followed by the dumping of rubbish from a nearby bone-working industry, and then the digging of a new pair of ditches that reasserted the line of the earlier boundary (below p. 67). After these later ditches had in turn been backfilled a late Neronian or early Flavian road was built over their line (below p. 122). We can be reasonably confident that the earliest ditches were Claudian, and they are more likely to have been backfilled before c. ad 50 than after.

Two similar ditches had previously been recorded at 7–11 Bishopsgate in 1995, and it now became clear that they may have formed the northern boundary of the same enclosure. These large parallel V-shaped ditches, set 2 metres apart, were found at the base of the archaeological sequence and only incompletely investigated. They were aligned east-west, some 150 metres to the north of the site of the early forum. The excavator, Dave Sankey, described both as originally about 2.5 metres wide and 1.4 metres deep, with a square-cut ‘ankle-breaker’ at the base, but accompanying plans show that the ‘inner’ southern ditch was nearly 0.5 metres wider than the northern one. ²⁰ They were also quickly backfilled. The site produced little dating evidence, but the ditches were clearly part of the pre-Flavian topography. Parallels have since been drawn with the fortifications established after the Boudican revolt of ad 60/61, leading to the suggestion that they formed part of a short-lived military post set on Cornhill after the rebellion. ²¹ An earlier date is equally consistent with the evidence. Interestingly, in the light of the parallel it provides with the sequence at The Walbrook, an early Flavian road was subsequently built over the line of the outer ditch. If the ditches at Bishopsgate were contemporary with those found at The Walbrook then these observations combine to identify two sides of a large Claudian enclosure on Cornhill.

Why then should these ditches be identified as defensive? Ditches serve many purposes. Although V-shaped ditches with square ‘ankle-breaker’ cleaning trenches at the base were introduced to many parts of Britain by the Roman army, they were neither exclusively Roman nor exclusively military. The dolabra sheath found in Southwark (p. 41), is a reminder of how the paraphernalia and practices of Roman military engineering might advance independently of the Roman army, and similar V-shaped ditches at Silchester may have been dug prior to the Roman conquest. ²² Lacey Wallace has also drawn our attention to instances where V-shaped ditches were dug alongside later roads into London. ²³ She consequently suggests that rather than being defensive, the ditches might have been roadside features, dug in anticipation of a planned extension to the street grid that was interrupted by the Boudican revolt. This argument doesn’t withstand closer scrutiny. In the first places it assumes that the ditches were designed to flank roads, but they were too close together for this to have been physically possible. Roads important enough to have been marked by major ditches were usually set within an easement 10–15 metres wide. Even the small drains flanking minor roads were usually placed more than 5 metres apart, but the ditches at Bishopsgate were no more than 2 metres apart. Wallace’s suggestion also assumes an interrupted construction exercise, but this does not square with the evidence that the ditches were comprehensively backfilled soon after they were dug. The backfilling was a major exercise, consistent with the deliberate slighting of associated earthworks rather than a suspended building programme. When roads were eventually built at these sites, they were set over the line of the earlier ditches, not between them. No attempt was made in these later phases of roadbuilding to reinstate or replace the Claudian ditches. Not only were the ditches dug without regard to road construction, it is also difficult to identify any need to drain the land in the places they were found. Those investigated by Ian Blair’s team were on well-drained slopes of the river Walbrook, set parallel to the line of the river in such a way that they neither directed water towards or away from it. Most importantly of all, Wallace’s suggestion fails to account for the fact that the paired ditches were of different sizes. This is a characteristic feature of defended enclosures, in which ramparts were protected by a larger inner ditch and a smaller outer ditch. There are no parallels for such a configuration from a roadside context, where it would serve no purpose.

The double-ditch configuration was, however, common in the forts of this period. ²⁴ London’s arrangement was identical to Claudian ditches found at Richborough that some think protected the site of Aulus Plautius’ landing in ad 43. ²⁵ The Richborough defences consisted of an inner ditch 3 metres wide and 1.8 metres deep and an outer one 2.13 metres wide and 1.2 metres deep, with an interval of 1.8 metres between them. This arrangement finds parallel in other temporary Claudian military defences, as at Longthorpe near Peterborough. ²⁶ The ditch systems of these temporary camps, like those of London, were smaller in scale than those of permanent legionary fortresses. ²⁷ No trace of the original ramparts was found at any of these sites, suggesting the application of a consistent policy of slighting the defences of temporary forts on their evacuation.

The two sightings of a double-ditch enclosure at London combine with other topographic evidence to permit the tentative reconstruction of a defensive circuit. The ditches were at the western and northern boundaries of London’s inner core, as marked by the extent of the orthogonal Claudio-Neronian street grid. This grid was symmetrically laid out around the approach road from London Bridge and a central T-shaped junction at the site of the forum. Although this urban topography was the product of later Claudian development, its essentials were dictated by the earlier topography. If so, and assuming the original enclosure to have been both rectangular and symmetrical, we can take a mirror to the evidence to locate the southern and eastern sides of the enclosure.

On this basis the southern boundary of the site, overlooking the Thames, should lie between Cannon Street and Thames Street. A ditch of appropriate form and date has, indeed, been found at exactly this location: at Regis House. ²⁸ This was a substantial east-west aligned feature, 2.8 metres wide and 0.9 metres deep, with an ‘ankle-breaker’ cleaning slot at its base. The east end of the ditch turned slightly south-eastward, where it terminated a little short of the line of the main road to London Bridge. A timber structure, represented by a row of three post-pits set 1.2–1.3 metres apart, stood between the ditch terminal and the road. The posts were on a line parallel to that of the ditch, slightly to its north. They could have supported a timber palisade closing the gap between ditch and road but are more likely to have been the forward side of a flanking tower to a gateway. The configuration is identical to the post-built gateway of the Claudian fort at Wigston Parva in Leicestershire. ²⁹ This would have been the main gate into London, set at the northern end of London Bridge. The natural defences provided by the Thames might account for the presence of only a single ditch along this side of the enclosure. Two timber pales found reused in a nearby waterfront revetment built c. ad 52 may originally have been part of the gateway or a palisade set over the rampart. These pales were made from oak poles, halved by cleaving, and carved at one end to form spear-shaped ‘hastate’ terminals. This was a common military design and the timbers were significantly larger than needed for domestic construction. The sapwood showed no signs of borer damage indicating that they had not been exposed to the elements for long, almost certainly less than 2 years. Whilst we cannot be certain that these southern defences were contemporary with those found to north and west, this is a reasonable supposition given the early date and brief working life of the palisade. These observations and reconstructions identify three sides of a Claudian enclosure, and Merrifield and Sheldon’s original topographic argument suggests the location of the fourth side. This combines to define an area about 610 by 450 metres, with an internal area of nearly 27.5 hectares (Fig. 5.2).

Fig. 5.2 Proposed reconstruction of the Claudian defended enclosure of c. ad 43, illustrating also the locations of discrete finds assemblages with conquest period characteristics (marked + and identified by site-codes). These small groups of pottery came from ditches and pits likely to have been dug early in the Roman occupation, although a pre-Roman origin for some of the material cannot be discounted. Individual finds are not plotted. Drawn by Justin Russell.

The evidence indicates that London was provided with a defensive enclosure of likely Claudian date and military execution. There are three possible reasons for its construction. One is that the army built the defences on behalf of a civilian community, following the example suggested for the Augustan site of Waldgirmes in Germany. ³⁰ If so, the ditches might date from c. ad 48 when London’s street system was engineered (below p. 64). This would suggest that London’s foundation was part of a programme of urbanization promoted by the provincial government in an area where there was no native community to coax into taking this initiative. On balance, however, this seems improbable. A military involvement in building a new town seems unlikely given the other demands on the army at this date (Table 5.1). A town of sufficient importance to warrant urban defences should also have housed civic buildings of the sort not seen in London until the Flavian period. The second possibility is that the ditches enclosed an army-engineered supply-base established soon after the conquest. This too would suggest a date c. ad 48 and might be consistent with other evidence for the establishment of a supply-base at London (Chapter 6 below). Although this remains a distinct possibility, it doesn’t explain why the ditches were backfilled so promptly and failed to include the types of Roman pottery found in assemblages associated with other constructions of this period. The third possibility is that this was the site of a temporary fort. If so, its scale only makes sense within the context of troop deployments at the time of the conquest. We cannot securely date the defences from archaeological finds, but there was no need for a large military post here once the Claudian army had advanced on Colchester. If the defences protected a military garrison, then the history of Claudian military deployment presents no credible alternative to ad 43.

Table 5.1 A suggested timeline of events affecting London in the period ad 43–70

The problem with this hypothesis is that although many sites have been excavated within the fortified area, none has produced evidence securely dated before ad 50. ³¹ The trouble is that a campaigning army leaves little trace and the evidence of proof that we need may be highly elusive. ³² Historical accounts suggest that Aulus Plautius’ army only waited on Claudius’ arrival for a few weeks during the summer of ad 43. Since the site was occupied at a critical point in the Roman campaign, it would have warranted more imposing defences than normal for a marching camp. These building operations would also have kept troops busy during the pause in campaign. But despite the scale of the defences it would have remained the temporary station of an army bivouacked in tents. Not only do tents leave no foundations, but the use of mess kits limited the use of pottery. No buildings are to be expected and the camp would only be traceable by its defences, which is why temporary Roman camps are notoriously difficult to date. The fleeting nature of this occupation would explain why London presents a lower density of military finds than the Claudian fortress at Colchester, and proportionately few of the early south Gaulish Samian pottery types found at permanently occupied conquest sites. ³³ The temporary installations would also have been highly vulnerable to truncation in the subsequent engineering of the Roman town, which in places reduced the pre-urban land surface by more than a metre. ³⁴

London’s Claudian enclosure was about the right size to have housed the bulk of the invading Claudian army. Philip Crummy’s speculative reconstruction of the successor Claudian encampment at Colchester comes up with something of similar size. ³⁵ London’s 27.5 hectare site was certainly larger than normal for a legionary fort. It has been estimated that tents might have made it possible to accommodate some 1,174 men per hectare, in which case the enclosure presented space for over 32,000 men. ³⁶ A lower density of occupation is more likely but it is entirely credible that the enclosure quartered 15,000–20,000 men. This would have been around half of Aulus Plautius’ total force, which is generally reckoned to have included around 40,000 troops, although a lower figure is possible. ³⁷ This seems about right, since a significant part of the army would have been deployed to secure the supply route from the channel forming garrisons at key locations between the south coast landing site and bridgehead of the Thames crossing.

If troops had also been stationed to secure the south bank of the river, this might help explain the discovery of an early Roman V-shaped ditch with an ‘ankle breaker’ at Park Street in Southwark in 1990. ³⁸ The lower fills of this feature included an important assemblage of late Iron Age pottery, also notable for the absence of products of London’s ‘Romanized’ industries, and a single sherd of a fine-ware beaker likely to have been imported from northern Gaul. The final silting of the ditch contained glauconite-rich or greensand-tempered sherds similar to those found in late Iron Age assemblages at Bermondsey and Clerkenwell, as well as two early or middle first-century brooches. The ditch was located some 180 metres to the north of the line of the Roman approaches to London Bridge, close to the western end of the island. The character, date, and context of this feature are consistent with it having been associated with the Claudian settlement of London. A small fortification here would have controlled access to the important pre-conquest and Roman waterfront site at 15–23 Southwark Street (above p. 40). Another early Roman V-shaped ditch was identified at Winchester Palace, parallel to the south-bank of the Thames. This might also have been associated with a Claudian military presence, although it is described as a drainage ditch in the excavation report and its fills included Verulamium region ware that first occurs in archaeological assemblages at the end of the 40s. ³⁹

The Thames crossing

If Aulus Plautius waited on Claudius at London in the summer of ad 43, having taken command of the river crossing, then this might also be the place where his army had first crossed the Thames in pursuit of British forces. Dio described how the Britons fell back on the river at a point near where it enters the sea and at high tide forms a pool. ⁴⁰ He explained that the British crossed easily because they knew where to find firm ground and an easy passage, but the Romans in trying to follow them were not so successful. However, Celtic auxiliaries (Keltoi in the original Greek) swam the river, and other troops crossed by a bridge a little way upstream, after which they attacked the barbarians from several sides and killed many of their number. The Roman army then sustained losses of its own when pursuing the retreating Britons into marshes.

This description fits the site of London, where the islands at Southwark created a pinch-point, above and below which high tides would have pooled before ebbing. There has been speculation that Plautius’ crossing might alternatively have been achieved further downstream. This claim has been most strongly advanced for a crossing from Higham to East Tilbury, where the river might have split into several meandering channels making them individually easier to cross. ⁴¹ There is, however, no credible evidence that the river was ever fordable at this point. The expanse of exposed inter-tidal marshland along the lower Thames has usually been a barrier to all but waterborne traffic along this reach. Early medieval sources and place names indicate that the Thames was then impassable except by ferry all the way upstream to Shepperton, with the lowest all-season ford found further up-river at Wallingford. ⁴² A campaigning army would, of course, have been prepared to risk crossing at places not routinely fordable, and the gentler tidal regime of the early Roman period may have made the river less formidable than it later became. Nonetheless, Dio’s description of troops crossing by a bridge close enough to military action to determine the outcome, along with the need to avoid intractable marshlands on either side of a contested crossing, eliminates most sites in the lower estuary from serious consideration. If his account is to be believed, it is more likely that the Roman forces crossed the river closer to its tidal head, leaving London and Westminster as the preferred candidates.

The idea that the Romans might have crossed the river at Westminster was first advanced to make sense of alignments of Roman roads, specifically elements of Watling Street, which appeared to have been directed towards Westminster rather than the City. ⁴³ It was consequently suggested that Watling Street followed the line of a pre-Roman route that used an Iron Age river-crossing at Westminster. Although this idea was promptly dismissed as improbable by Francis Haverfield it has remained curiously persistent. ⁴⁴ As matters stand, the physical remains of Watling Street south of the Thames have been traced from Deptford to Southwark, but are not found further to the west. The topography shows that the road followed the best line achievable if the purpose was to keep to higher ground (Fig. 14.2). It did not head in a straight line for London Bridge because it skirted tidal inlets and marshes on the south bank, not because it was aiming for Westminster. The line of Watling Street north of the Thames is found beneath Edgware Road, but never further south than Oxford Street where it joined the great west road that exited London through Newgate. The decision to route Watling Street along the line of Edgware Road is likely to have been dictated by a preference to hold to higher ground west of the river Fleet, whose valley made a more direct line to London impractical. The road system does not, therefore, bypass or otherwise ignore the Roman city. Nor did it direct traffic through Westminster.

Whilst there is no need to identify a river crossing at Westminster to make sense of the Roman street pattern, this is not proof that no such crossing existed. There are other grounds, however, for considering Westminster an unlikely place for a ford. There was an important late medieval crossing from Lambeth to Westminster, but this was achieved by means of a horse-ferry. Archaeological investigations on both sides of the river have failed to find any evidence to suggest that the river was regularly crossed here before the ferry was established, and the point where the line of Watling Street is projected to have met the river is too close to its bend to have been suitable for a ford. Excavations on Thorney Island, beneath the Palace of Westminster, indicate the presence of a high-status building and elite burial from the second century, but this is likely to have been associated with a later villa. ⁴⁵ Evidence for an approach road to the supposed crossing has also been sought south of the river, in the gardens of Lambeth Palace. Investigations undertaken for the ‘Time Team’ television series in 1994 drew a blank, and contour mapping showed that most of the land lay below the probable water level in Roman times. Despite the positive gloss provided by a programme editor reluctant to trumpet failure, the archaeologists involved concluded that ‘the case for any river crossing at Westminster appears to be…without any real substance at all’. ⁴⁶ This wasn’t a sensible place to approach the river, and there was no attempt to engineer causeways and roads to bring Roman troops to its banks.

If we return to Dio’s description, this makes it clear that whilst some soldiers swam the river, which is feasible at numerous locations along the river, the main force crossed by bridge. This was the only sensible option for moving an entire army and its supplies over the river. The Greek text described this as a γέφυρας, a term which could embrace a variety of fixed crossings including pontoon bridges formed from ships lashed together. ⁴⁷ This would have suited the needs of the time. In the absence of any archaeological or topographic evidence for a fixed pre-Roman crossing, or evident need for one to have existed prior to the conquest, Dio may have described a bridge built as part of the Roman campaign. It is consequently worth drawing attention to observations at Regis House in 1929–31, where two crossed timber beams were found on the riverbank in the only appropriate location for the landward abutment of a conquest-period pontoon bridge. ⁴⁸ This crossing point exploited gravel islands projecting into the Thames at Southwark to reach a natural gravel promontory beneath modern Fish Street Hill. On topographic grounds, London is the only sensible site for a fixed-river crossing down-river from Staines, where the Roman place name Pontibus shows that a bridge also existed. The configuration of islands in Southwark, facing Cornhill where the river flowed through a narrow channel, uniquely eliminated the need to traverse areas of intertidal marsh. The advantages of the site, and the impracticality of the alternatives, meant that London remained the lowest bridging point of the Thames into modern times.

The need to bridge the Thames is certain to have figured in campaign planning, drawing on intelligence obtained from Caesar’s earlier expedition and Rome’s allies in Britain. Aulus Plautius had a large transport fleet and it would have been logistically and strategically necessary to develop seaborne supply lines along the north coast of Kent and into the Thames estuary. This adds gentle weight to arguments in favour of seeing Richborough as the principal point of entry into Britain. In any case, some ships of the invasion fleet could easily have been lashed together to form a pontoon bridge, the need for which was foreseen.

Once Aulus Plautius had occupied his position on the Thames, he halted his advance and waited on the arrival of the Imperial party. It seems inconceivable that he would have withdrawn south of the river after securing the necessary bridgehead. Dio’s account explains that Claudius joined the troops who were awaiting him near the Thames when he crossed into Britain. ⁴⁹ It also specifies that Claudius was only in Britain for 16 days. This was too short a period to have involved a landing on the south coast: the distances involved were too great to permit the return trip to Colchester within the available time. ⁵⁰ His timetable was, instead, consistent with a sailing that made directly to the welcome of the army waiting on the Thames. The waterfront site at 15–23 Southwark Street, presently a car-park near Borough Market, was an established landing place of exactly the sort where Claudius might have disembarked. A landing in Southwark would have allowed Claudius to lead his forces across the Thames, as described by Dio, to join troops already camped on Cornhill. The timeline is uncertain, but Claudius probably arrived in Britain in late July or early August. ⁵¹ The Claudian party was a large one, including soldiers of the Praetorian Guard, elephants and siege engines, and an extended court of senators and members of the imperial household. It may even have included an additional Roman legion. His triumphal march to Colchester, navigating marshlands flanking the river Lea, would have started immediately. The departing troops would have evacuated their temporary site beside the Thames, leaving it redundant.