25

The third-century ‘crisis’ (c. ad 250–70)

The destruction of the port

One of the more perplexing discoveries made during the study of London’s Roman port is that its waterfront quays were deliberately dismantled. There are several reasons why London’s waterside might have been neglected, but it is distinctly curious to find that its timber revetments were systematically removed. The river embankment was a vital part of the urban topography that did far more than serve a working harbour: it defined the city boundary, articulated public buildings along the waterfront, and housed warehouses and workshops that operated independently of river traffic. Why destroy an integral, useful and expensively assembled part of the urban landscape?

The evidence itself was excellently summarized by Trevor Brigham more than 30 years ago. ¹ The upper timbers of revetments along the north bank of the Thames were removed almost everywhere they were accessible, in a destruction that extended from the westernmost quays next to Queenhithe, past the site of London Bridge, all the way downstream to the Old Custom House and Sugar Quay. At most locations the retaining tiebacks were crudely axed through to release the massive horizontal timber baulks which were removed down to contemporary Mean High Water level. This left the lowermost timbers exposed to the river, which chased a tidal channel before depositing a thick band of river-lain silts and gravels. There were minor variations to this picture: in some places the baulks were removed in their entirety and in others the tier immediately above the water level also survived. But most accessible baulks were removed for a distance of 1 kilometre along the waterfront. Only the box quay at the Old Custom House, at the eastern end of the port, may have remained substantially intact. ² The common pattern of disuse, cutting-back, erosion, and deposition suggests a single episode of destructive re-engineering.

We have a rough idea of when this happened. Several strands of evidence show that London’s harbour was still busy into the 240s. The waterfront quays continued to handle imported Samian assemblages until this date (p. 314), with continued investment in London’s transportation infrastructure until at least c. ad 241–3 (pp. 320–1). As we shall shortly see, various classes of archaeological find show that patterns of supply into London changed c. ad 255, perhaps following the redundancy of London’s harbour. We do not know how swiftly the quays were dismantled, but finds associated with subsequent river erosion date to the middle of the third century. The revetments were certainly cut back before c. ad 270, since erosion preceded the construction of temporary quays that were themselves destroyed by the riverside wall c. ad 275 (p. 339).

It is harder to work out why it happened. It is unlikely that the old waterlogged timbers were recovered for reuse. The baulks included a significant proportion of perishable sapwood unfit for structural use, whilst grit and age would have made the remaining heartwood difficult to work. ³ In any case we know of no contemporary projects that demanded such a massive and unusual supply. Some timbers may have been used as fuel, but it is impossible to believe that fuel shortages were sufficient cause for large-scale destructive salvage. Damian Goodburn suggests that the main object of the exercise was to convert the quays into a defensive bank. This is consistent with the way in which the tie-backs behind the facing baulks were chopped back to form a slope, with the foreshore channel a defensive ditch. It seems likely that London’s waterfront quays were converted into a barrier after the port’s redundancy.

The atrophy of long-distance supply

This presents us with two different but related phenomena to account for: the failure of the port, and its replacement by crudely fashioned defences. There are several reasons why traffic through London might have declined in the third century. The port’s role in supplying the army was eclipsed by alternative sites, such as York and South Shields, and coastal shipping meant that fewer goods were carried by road. But these changes took effect earlier in the century, at a time when London’s harbour remained busy. The port’s decline may instead have been provoked by tidal changes that hindered larger ships from reaching London’s quays. Sediments at Queenhithe show that the Thames remained weakly saline upriver of London Bridge until the end of the second century but an accumulation of foreshore deposits at Baynard’s Castle indicates that the river had ceased to be scoured by tides in the mid-third century. ⁴ It has been argued that the falling water levels directly caused the dismantling of the quayside between ad 250 and 270, although this seems likely to be an over simplification of a more complex situation. ⁵ The problems of bringing cargoes ashore without the aid of a rising tide were solved at many inland river-ports throughout the Roman world. Why not at London? It has additionally been suggested that this downward drift of the tidal head encouraged the relocation of the port facilities down-river to Shadwell where a harbour may have been established in a back channel of the Thames behind the island of Wapping. ⁶ Shadwell could not, however, have functioned on anything like the same scale as London. It lacked the large urban workforce needed to man the docks, and the range of storage and transportation facilities that London enjoyed. If Shadwell’s docks were important, then road access to London would have been critical. But the Shadwell road was probably blocked from direct access to the city when the town wall was built. ⁷ Vehicles may have deviated to enter town through Aldgate, but it is difficult to believe that serious volumes of goods were made to traffic this longer route. There is nothing to suggest that London’s earlier handling capacity was replaced at Shadwell, or indeed elsewhere along the Thames before the seventh century. ⁸

London’s port is more likely to have failed because of a decline in long-distance trade. Until the middle of the third century, Britain received significant quantities of olive oil and wine shipped in large amphorae. This supply was directed towards areas of military demand, reaching the British Isles by transportation routes along the Rhone and Rhine. Some goods were destined for consumers in London and more passed through town en route to northern garrisons. Most oil was transported in large globular Dressel 20 amphorae from Baetica in southern Spain, but their supply fell away sharply c. ad 255/260. ⁹ London’s later olive oil imports were more likely to arrive in North African amphorae but at massively reduced levels. ¹⁰ It also appears that fewer wooden wine barrels were imported, since these ceased to be reused to line London’s wells. ¹¹ Many types of pottery, including Samian and mortaria, similarly failed to reach London in earlier quantities. If these containers and tablewares provide a measure of the scale of long-distance supply then this had declined massively in the mid-third century.

The reasons for this decline are much debated. Studies of Dressel 20 amphorae have concentrated on three possible factors: disruptions in long-distance supply provoked by the establishment of the breakaway Gallic Empire; changed dietary preferences that reduced demand for Mediterranean imports; and a reduction in troop numbers that drove a collapse in military demand. ¹² Contributory factors include the reorganization of estate production in Spain and the disappearance of a military redistributive system that subsidized transportation costs. Where large quantities of supplies were moved to military consumers, as previously, then other goods followed in their path. ¹³ Spare capacity came into existence in the months between campaigns and public building programmes, in the holds of vessels returning from commissioned deliveries, and in marginal spaces within larger shipments. This capacity enabled other forms of transaction, and engendered an institutional architecture that facilitated trade, carrying luxury imports to a wide range of British consumers.

We have already explored how London’s fortunes may have been shaped by city’s role as an entrepôt for annonary supply (p. 195). The third-century institutionalization of the annona militaris extended the in-kind allowances of food, fodder, clothing and equipment offered to soldiers and civil servants. ¹⁴ This accompanied a massive reduction in Rome’s willingness or ability to move such provisions over long distances. As Simon Esmonde Cleary has observed, the state no longer found reason to provide its soldiers with a Mediterranean diet reliant on long-distance supply. ¹⁵ Garrisons were instead expected to support themselves through local taxation and cash purchases. The effects can be seen in a pattern of import-substitution as the army turned to locally available resources. Imported wine was replaced by beer, oil by lard and tallow, and Mediterranean fish sauce by North Sea substitutes (above p. 216). Botanical evidence shows that fewer exotic fruits were imported to London and these were replaced by locally grown produce. ¹⁶ British kilns produced fine wares to substitute for missing imports, and Roman London came to rely on pottery carried down-river from manufactories in Oxfordshire and Surrey. ¹⁷ Jerry Evans has shown how regional patterns of pottery supply were profoundly impacted by such changes. Disruptions to the bulk supply of goods within the administered annona and frumentationes had a knock-on effect on other traffic. The contraction of long-distance supply had a massive impact on London’s economy and is likely to have been the main reason for the redundancy of its port.

A slump in regional productivity?

In related changes, several places involved in the large-scale extraction of raw materials around London also ceased operation in the late third century. Closures have been identified amongst ironworking sites of the Weald, including those at Bardown and Beauport Park. ¹⁸ The bathhouse at Beauport Park saw continuous use until the middle of the century and the coin assemblage included numerous issues down to Decius (ad 249–51), but an absence of later coinage suggests that the building was abandoned in the 250s. ¹⁹ It was then stripped of reusable materials and left to ruin. Although these sites in the High Weald are unlikely to have supplied London, serving the south-coast instead, a pattern of declining productivity can also be traced in areas that sent iron by road to London. ²⁰ Places associated with the transportation infrastructure developed around the ironworking industry were closed or neglected in the late third century. ²¹ Malcolm Lyne has suggested that changed patterns of exploitation may have encouraged potters from iron producing areas to relocate to more accessible areas on the margins of the Weald in Hampshire and Kent. ²² Although modest levels of production continued on smaller satellite sites, the Weald witnessed a marked settlement decline. Some iron working sites around Berkhamsted in the Chilterns north-west of London were also abandoned during the third century, suggesting that similar forces were at work in other landscapes. ²³

The contraction of the iron-working industry may have followed earlier over-exploitation of fuel and ore, exacerbated by the costs of maintaining the transport infrastructure. But the timing of this decline fits within the wider patterns of failure. It may also have coincided with the ending of the involvement of the fleet in the procurement chain. ²⁴ The Classis Britannica disappears from the historical and epigraphic record towards the middle of the century. The last known reference, dated ad 244–9, is found in a curriculum vitae inscribed on a tombstone from Arles. ²⁵ The Classis Germanica is also unattested after this date, suggesting the reorganization of naval commands on both sides of the Channel. This may have involved closing the naval base at Dover, which was probably evacuated and its defences slighted in the mid- to late third century. ²⁶ Was London’s port, perhaps controlled by the fleet after its Severan restoration, closed in a similar exercise? ²⁷ It must be acknowledged, however, that we have no direct evidence for the fleet’s disbandment. Its disappearance is reconstructed from absences in the historical and archaeological record, at a time when such records were increasingly sparse. It is also possible that London’s harbour remained a procuratorial responsibility, and we might instead look to a restructuring of this office to account for revised approaches to its management. ²⁸

Pottery export from the Thames estuary also witnesses a moderation in the scale of surplus extraction. Large storage vessels in North Kentish shelly ware, poppyhead beakers in Upchurch Ware, and black burnished ware (BB2) became scarce from the middle of the century, and were no longer exported in quantities by c. ad 275. ²⁹ There was a sharp decline in the quantities of BB2 present in London assemblages dated ad 230–50/60, as amongst material recovered from Leadenhall Court when these products also stopped reaching Hadrian’s Wall in significant quantities. ³⁰ This may have followed a decline in the numbers of salterns operating in Kent and Essex, and likely that the estuary was no longer producing elevated volumes of salt and salted goods for distribution along east-coast supply routes. ³¹

The stone quarries of north Kent were also less busy. London’s needs were greatest during the building of the early second-century town wall, but Kentish ragstone continued to serve the needs of construction projects such as the Shadwell bathhouse and the Walbrook Mithraeum into the 240s. Many later building programmes recycled materials quarried from London’s disused public buildings in a significant shift in approach to the procurement of building supplies. These changes may have been the consequence of reduced demand combined with a readier availability of materials from redundant buildings, but meant that river-barges no longer made frequent trips to bring ragstone to London from the Medway.

London’s major construction projects had long drawn on timber procured from managed woodlands, but there also appears to have been a hiatus in such supply in the middle of the third century. A list of dendrochronologically dated timbers shows that London regularly imported freshly felled timbers until ad 252/253, which is the felling date of timbers used to repair roadside drains at One Poultry. ³² These were the last timbers we know to have been imported before the construction of the riverside wall some 20 years later. This interruption appears to find confirmation in the study of timbers felled for the construction of a large public building ad 293/294 (below p. 350). These timbers all started growing at dates between ad 191 and ad 254 (particularly ad 217 to ad 234) in managed woodlands where regular pollarding was practised. ³³ It is striking that none started growth after ad 254. The implication is that the trees from which these later timbers were sourced had stopped being coppiced, suggesting that these woods were not used to source building timber from c. ad 254.

In sum, there is a consistent pattern of declining productivity within the London region, discernible within its managed woodlands, iron-working sites, stone quarries and salterns. ³⁴ These extraction industries relied on the transport infrastructure established around London and its port. It is hard to decide on the extent to which the loss of transport capacity was a consequence or cause of declining regional productivity. What is clear is that the mid-third century marked a radical change, in which capacity was lost rather than enhanced.

Signs of urban stress

Changing arrangements for local supply may also have contributed to the increased frequency of coin ‘forging’ in third-century London. Debris from the local manufacture of copies of third-century coins has been recovered from at least three locations. We have already referred to the moulds used to forge coins from as early as c. ad 225 found in the wall turret at Warwick Square (p. 309). ³⁵ Others were found in a ditch in Bermondsey Square in 1998. ³⁶ The largest and best studied assemblage, consisting of fragments of more than 800 carefully made moulds, was recovered from the city ditch at 85 London Wall. ³⁷ These moulds produced copies of silver denarii and copper-alloy dupondii and asses dated from Trajan to Trebonianus Gallus (ad 251–3), working from at least fifty-nine silver denarii and fourteen copper alloy coins, roughly one tenth of annual pay at the time. These coiners were working into the mid-250s if not beyond. Most of this production took place beside the city wall and was probably under military control. ³⁸

These copper-alloy coins were comparatively worthless, and were probably made to compensate for shortages in small change rather than for profit. ³⁹ Low value coin facilitated small denomination payments for produce and labour, as when hauliers required cash payment on delivery (p. 199). ⁴⁰ London’s garrison would have included soldiers housed in the towers and gatehouses along the walled circuit who were responsible for directing traffic through city gates where shipments could be received and portoria taxes raised. Coin manufactured on the wall met a demand for small-change to lubricate such transactions at a time when imported supplies were inadequate.

There are signs of under-investment in some elements of London’s urban infrastructure. The latest burials at Eldon Street contained grave goods of c. ad 250, but the road and cemetery were not much used afterwards. ⁴¹ It has also been suggested that roadside properties beside the main Walbrook crossing at One Poultry saw reduced use, evidenced by an unusual scarcity of coin issues of the 260s. ⁴² A hiatus in building activity was even more clearly evident on the extensive Lloyd’s Register site, where at least five late second or early third century buildings were demolished in the mid-third century and the site left vacant for a couple of decades before redevelopment on a completely new layout c. ad 270. ⁴³ Some city buildings were fire damaged in the mid- or late third century. Most devastatingly the forum basilica caught fire and its roof collapsed. ⁴⁴ Pottery spot-dated after c. ad 250 was found beneath this collapse. The fire probably spread into adjacent areas: the road along the north side of the basilica was covered by destruction debris, charred timbers associated with late third-century material collapsed into wells opposite the entrance to the forum, and buildings to its east were reportedly damaged in a mid-third-century fire. ⁴⁵ We cannot be certain that these burnt deposits traced a single event, or how widely the forum fire had spread. It is worth noting, however, that two buildings at One Poultry were also burnt. These were covered by destruction debris containing Alice Holt and Farnham pottery, which suggests that site clearance took place after c. ad 270. ⁴⁶ Further-flung sites showing signs of fire destruction include a site on the banks of the Fleet, at Old Bailey, where the building provisionally identified as an octagonal temple (above p. 266) was destroyed around 270 before being replaced by a large multi-roomed building with ragstone walls and opus signinum floors. ⁴⁷ South of the river, the roadside complex containing a shop, bakery and granary alongside the main road to London Bridge was also fire-damaged around the middle of the third century. ⁴⁸ It is likely that these different observations represent separate outbreaks of accidental fire, but we cannot wholly dismiss incendiary unrest as a cause (see p. 352).

Reassessing London at a time of crisis

This chapter has summarized evidence for a series of changes that contributed to the redundancy of London’s port. These took place during the period of wider disruptions known as the third-century crisis: a time of political instability, plague, barbarian invasion, and economic turbulence. The utility of the term ‘crisis’ has recently been questioned, given the variations of circumstance, but in many spheres the problems were acute. ⁴⁹ The historical narrative tends to start with the assassination of the emperor Severus Alexander in ad 235, after which there was a rapid succession of brief reigns and damaging frontier wars. In the north-west provinces matters reached a breaking point around ad 258, when the emperor Gallienus weakened control of the Rhine frontier by transferring troops to Pannonia. ⁵⁰ Franks and Alemanni from beyond the Rhine, who had been a threat throughout the 250s, used this opportunity to raid deep into Gaul. Order was restored by troops commanded by Postumus, who then rebelled from Gallienus’ rule to establish a Gallic Empire (Imperium Galliarum) that divorced Gaul, Britain and Spain from the rest of the empire for some 15 years. ⁵¹

As we have already noted, the establishment of the Gallic Empire may have interrupted flows of long-distance supply. There were several other contributory factors to this decline. Studies of the third-century crisis are particularly exercised by the evidence of a currency debasement, contributing to the collapse of the Augustan monetary system in the 260s. ⁵² This was also a turning point in architectural history: few monumental buildings were erected and the practice of setting-up of commemorative inscriptions largely ceased, witnessing a retreat from the earlier values of civic euergetism. ⁵³ Research by Kyle Harper has drawn attention to the way in which Rome’s military and economic setbacks may have been provoked by another pandemic that swept the empire in the 250s. ⁵⁴ This plague, suggested by some to have been influenza or viral fever, cruelly echoed the disruptions of the Antonine plague a century earlier. It is known to historians as the plague of Cyprian, after the bishop of Carthage who described its impact in letters written c. ad 252, and can tentatively be traced back to an outbreak in Alexandria in Egypt c. ad 249 from where it reached Rome early in ad 251. The historical accounts imply recurring pestilential outbreaks, but we cannot be certain that they shared a common cause. The evidence is uncertain and contested, but Harper builds a credible case for considering this pandemic to have been critical to the disruptions of the period, adding exogenous shock to a system that was already under considerable stress.

This begs the question of how much London might have been affected by these problems. ⁵⁵ There is no reason to believe that Britain was other than a peaceful part of Postumus’ Gallic Empire but the evidence presented here suggests that London was more profoundly changed than is sometimes assumed. ⁵⁶ The closure of the port can in part be seen as a consequence of the unsustainable nature of its earlier growth. Massive flows of annona goods had followed an infrastructure of military supply that was built around campaigns of advance and supported the exceptional concentrations of Roman forces found in the frontier regions. These met the strategic goals of imperial and provincial government, giving London its political and economic purpose. The disruptions of the mid-third century brought these arrangements to a rapid and decided close.

The dating of these changes is key to understanding them. The record of woodland management and timber supply suggests that London continued to expand its urban infrastructure until c. ad 254. A continuing concern with securing local supplies by payment in cash may have prompted officers responsible for manning the city walls to coin small-change to make good on shortages in official supply until at least ad 253. In the mid-250s, however, work ceased at several larger-scale production sites within the region, including ironworking sites in the Weald and salterns in the Thames estuary. This may reflect a diminution in the level of imperial engagement in the productive economy. This was also when the annona was reformed in ways that reduced the traffic of Mediterranean goods, and in the late 250s London’s port ceased to handle the large-scale import of amphorae and fine-wares. We can assume that shipments of perishable goods, such as grain and leather, were similarly reduced.

In the late 250s, fewer ships travelled the seas and fewer cargoes took to roads and rivers. This may have been the combined consequence of government policy, military insecurity and economic recession, but it certainly involved a major loss of capacity. This may have been caused or compounded by labour shortages, as wars and plagues reduced manpower. Some troops were withdrawn from Britain to support Gallienus’ frontier wars. In ad 255, men of Legio XX, formerly of the British garrison, set up a dedication at Mainz. ⁵⁷ British legionaries were also part of Gallienus’ army at Sirmium in Pannonia in the 260s, and soldiers from Britain went east with Valerian’s field army in ad 258. ⁵⁸ We have no measure of the scale of these troop movements, but the soldiers withdrawn from Britain didn’t return and their loss may have contributed to the abandonment of vici outside forts on the northern frontier. ⁵⁹ It is distinctly possible that demands placed on the forces stationed in Britain at this time resulted in the redeployment of ships from the fleet, and of official personnel managing estates within London’s hinterland. Death and desertion may have compounded losses. ⁶⁰ Parallels can be drawn with a depopulation of towns in Illyricum c. ad 260, which the historian Zosimus blamed on the combined weight of barbarian invasion and plague. ⁶¹ As we have already observed in describing London’s late second-century contraction, Labour shortages are readily provoked by pandemics which not only supress military recruitment but reduce harvests and depress taxation in kind. Food insecurity is a likely outcome.

Several strands of evidence suggest that these failures of supply took place soon after ad 254, making them contemporary with the historical record of the westward spread of the plague of Cyprian. We have no evidence to show that the plague reached London, but it is difficult to believe that this cosmopolitan city would have totally escaped its effects. Whatever the immediate cause, the failures of production and supply represented a major rupture in London’s history. It was no longer a vital hub in military and administered supply and its port ceased to serve government needs.

This helps us to understand the redundancy of the port, but it doesn’t tell us why it became necessary to replace London’s carefully engineered waterfront quays with a crude defensive bank. A speculative explanation builds from the proposition that there was a massive loss of naval capacity following the depletions of war and plague c. ad 254. This would not only have reduced the administration’s ability to maintain annona supplies, but risked surrendering the waters around Britain to piracy. ⁶² Frankish raiders were supposedly a problem to Rome throughout the 250s, and Postumus’ issue of coins carrying the legend Neptvni Redvci in ad 262/263 is thought to have marked a naval victory won against such forces. ⁶³ London’s clumsily executed waterfront defences could have been an emergency response to naval threat in the late 250s or early 260s. Whatever the cause, London’s defended waterfront proclaimed the continued importance of the city as a seat of power and a place worth defending.