28
Reinforcing the urban defences
For the first two centuries of London’s existence, Rome’s interests were written on the waterfront, where intermittent harbour improvements served the machinery of power. London’s last great phase of quayside expansion ended as the city walls were built, and the history of later public architecture was largely a history of urban defences. London retained a pivotal role in Rome’s network of control, but was now presented as a strong-point confronting threat and insecurity. Three main phases in the evolution of the masonry wall can be described. The massive landward circuit was built in the early third century, perhaps c. ad 215/225, followed by the construction of the riverside wall in the late 270s (Chapters 23 and 26). Both works can be viewed as acts of rehabilitation following periods of urban dilapidation, undertaken when high politics gave Rome reason to re-engage with affairs in southern Britain.
In this chapter we turn our attention to the third phase of defensive architecture, when new towers were added to the walled circuit (Figs 28.1 and 28.2). The period preceding the construction of these towers may also have been one of relative neglect: most fourth-century building activity is provisionally dated before c. ad 325 or after c. ad 350 and the second quarter of the century appears distinctly quieter. This hiatus in activity would be consistent with the reduced importance accorded the north-west provinces after Constantine’s attention turned eastwards in the 320s (above p. 356). Soon after the middle of the century a series of D-shaped solid projecting towers, known as bastions, was built along the eastern third of the town wall. 1 Traces of fourteen have been recorded between the Walbrook and the Tower, and the locations of another seven can be conjectured from their regular 53 metre spacing. Their foundations were 5.8–7.9 metres wide, projecting forward from the wall for 4.4–5.6 metres and encroaching onto the line of the earlier V-shaped town ditch. A broad U-shaped town ditch was dug in replacement, adding defence in depth. Evidence of this ditch is scarce, since largely lost to medieval recutting, but a section north of the gateway at Ludgate was 3 metres deep and 10 metres wide. 2 The towers would have stood taller than the existing wall parapet, providing elevated platforms from which archers or small bolt-firing ballistae could direct enfilading fire along the wall. 3
Fig. 28.1 A proposed reconstruction of London c. ad 370, although not all of the illustrated streets and buildings retained from earlier phases are likely to have remained in use by this date. Drawn by Justin Russell.
Fig. 28.2 The foundations of bastion 10 on London Wall at Camomile Street, incorporating spolia from earlier Roman tombs (from a drawing by Henry Hodge on behalf of the London and Middlesex Archaeological Society and first published in Price 1880).
It is not certain that all solid towers were contemporary, but their regular spacing shows them to have been the product of a single plan. This was probably decided on after c. ad 350. At Duke’s Place Peter Marsden recorded the 1-metre-high foundations of a bastion dug into deposits containing fourth-century finds, where material deposited against the structure included fourth-century pottery and coins dated ad 364–75. 4 Subsequent excavations by John Maloney recovered a coin of Constans (ad 341–4) from the V-shaped town ditch. Since the ditch was backfilled in preparation for the construction of the tower the coin provides an indirect terminus post-quem for its construction.
The eastern towers drew on a mix of materials, principally Kentish ragstone, set over chalk and flint foundations, and incorporating material from redundant funerary monuments. Some effigies appear to have been ritually laid to rest and Richard Hingley suggests that they had been selected for their religious potency. 5 This use of redundant tombs was also a pragmatic one. London had no local building stone, and these monuments were ready quarries. 6 The quarrying of stone from the amphitheatre also started soon after ad 355, with coins of ad 367–74 present in robber trenches and demolition debris. 7 Like other Romano-British amphitheatres the arena had ceased to host games and displays during the fourth century. 8 The supposed mansio in Southwark was also dismantled and stone extracted from its foundations, with mid-fourth-century pottery and a coin of ad 347–8 present in a robber trench. 9 The materials recovered from these redundant buildings could have supplied various construction projects, including the building of towers and gates along the landward wall. 10
The series of solid towers did not continue west of the Walbrook. Although bastions were eventually built here, these were hollow structures of likely medieval date. A pair investigated at King Edward Street appear, however, to have been late Roman: one with a solid base built when the V-shaped ditch was still partly open. 11 Improvements to the town gates may also have been made, as at Aldersgate where a double gateway was built with boldly projecting D-shaped flanking towers that may have supported raised artillery platforms. 12 The decision not to build towers at most locations west of the Walbrook may have been due to the presence of an extensive marsh, Moorfields, which had formed after the second-century town wall was built. 13 Towers were also absent around the Cripplegate fort, where the closely spaced gates and corner towers were perhaps deemed sufficient.
The works of this period might have entailed closing off the c. 75 metres long gap at the western end of the riverside wall. Excavations at Baynard’s Castle showed that this stretch lacked the piled foundations characteristic of late third-century work, and the ragstone blocks were set directly into the clay subsoil. There were no tile courses within the body of the wall, and the fact that it eventually collapsed northwards indicates that there was no earth bank behind. 14 Southwark remained largely undefended, although the protection afforded by river channels might have been enhanced by temporary earthworks around the bridgehead. A ditch measuring over 1.5 metres deep and 5 metres wide found aligned east-west at Tooley Street contained fills dated to the second half of the fourth century. 15
London’s new defences followed contemporary fashion in military architecture, designed to give defenders a tactical advantage in the event of hostile assault. 16 As in earlier times, the enhancement of the walled circuit conveyed a message of authority and stability and the serried towers signalled London’s urban status to ships approaching upriver. 17 James Gerrard suggests that this symbolic statement was the essential purpose of the architecture, since gaps in coverage betray a neglect of defensive considerations. The distribution of towers may also, however, have been tactically inspired. They were placed along parts of the wall more easily approached over open ground, but missing along the stretches of wall facing Moorgate, the Fleet and Thames, where river and marsh insulated the city from attack. 18 The decision to manage without towers at these places reduced construction costs, and may have acknowledged the limited availability of soldiery to man the defences. This was, therefore, a lower-budget effort than the earlier town wall. London possessed an unusually extensive walled circuit compared to the reduced urban enceintes of late Roman Gaul, which it had inherited from an earlier and more expansive time. Its length made the circuit difficult to defend, but the decision not to retreat to a smaller area, achievable by retreating to an urban core east of the Walbrook, may have been dictated by construction costs. The building of London’s towers is likely to have involved a series of pragmatic choices balancing defensive considerations against what could be achieved. 19 They offered both appearance and substance: a credible and sustainable promise of security, representing the abiding reality of revived Roman power.
The modernization of London’s defences was a decided strategy implemented soon after the middle of the fourth century. Whilst it remains unfashionable to connect campaigns of wall construction with known historical events, we cannot escape the fact that programmes of investment on this scale were politically inspired. 20 It has long been thought likely that London’s bastions were built in a programme of refortification that Ammianus Marcellinus attributed to Count Theodosius in ad 368. 21 This remains an attractive hypothesis, although the archaeological dating is insufficiently precise for certainty and John Casey alternatively suggested that bastions were added to Romano-British urban defences in building campaigns under Magnentius c. ad 350. 22 It seems unlikely, however, that Magnentius’ rule generated resources for the restoration of Romano-British towns. 23 His bid for power is more likely to have entailed removing troops from the province, reducing rather than reinforcing the available manpower.
London was directly mentioned in accounts of campaigns undertaken in the 360s. In ad 360, the commander (magister militum) Flavius Lupicinus made directly for London to take council when he was sent to Britain to respond to attacks from Picts and Scotti in the north, showing that the city was the appropriate administrative centre for such operations. Similar considerations applied during the disorders of ad 367 that Ammianus described as a barbarian conspiracy. 24 In this account the emperor Valentinian turned to Count Theodosius to restore control. Theodosius crossed to Richborough where he waited on troop reinforcements before marching on London, intercepting bands of plunderers before relieving the imperilled town. London is where he established his operational base, from which he was credited with restoring Britain’s damaged ‘cities and strongholds which had been founded to secure a long period of peace, but had suffered repeated misfortunes’.
This presents a plausible context for upgrading London’s defences, and Theodosius’ forces would have brought manpower and resources to the exercise. A significant proportion of the Valentinian coins found around London were copies, likely to have been made in order to supplement official supplies, lubricating essential transactions and empowering officials returning to the state payroll. 25 Ammianus’ account also provides the first reference to the fact that London was renamed Augusta, a title that can only have been awarded as an imperial mark of honour. 26 Coins minted in London down to ad 325 carry mint-marks that imply, through omission, that it had not been renamed by this date. The grant of the title was perhaps decided on by Valentinian at the time of Theodosius’ restoration. 27
London and the Roman administration
Regardless of the exact circumstances that resulted in the building of London’s bastions, they declared the continued importance of the late fourth-century city. Simon Esmonde Cleary has shown that the late antique walled cities of northern Gaul housed key institutions of the Roman administration. 28 There is a strong correlation between the possession of walls and the presence of military and state personnel, but a comparative indifference to the needs of civilian urban communities. London’s defences are likely to have been similarly inspired. Some officers and departments operating out of fourth-century London are described in the lists of civil and military officials presented in the Notitia Dignitatum, which although amended in the fifth century reproduced elements of earlier arrangements. 29 Following third-century reform Britain was no longer ruled by a governor and procurator but divided into a series of smaller provinces, each with its own government (p. 311). Collectively these formed the Diocese of Britain that was overseen by a vicarius based in London who answered to the Praetorian Prefect of the Gauls. The Notitia lists nine department heads on the staff of the vicarius—managing taxes, records, correspondence and intelligence—along with assistants, secretaries and notaries. It also described London as the seat of the treasury and head of diocesan tax administration. 30 Each of these officials would have been attended by a large private household, combining to draw hundreds of notables to the fourth-century city.
Although Britain’s principal military commands followed the northern frontier London may also have been the seat of operations for a mobile field army (comitatenses) commanded by the Comes Britanniarum. Although we lack any direct archaeological evidence for the presence of the Comes in London, and the history of the command is uncertain, this would not have been an unusual arrangement. This post wasn’t allocated a fixed base in the Notitia, and Mark Hassall suggests that it campaigned wherever required and was billeted accordingly. 31 Roman troops were routinely stationed in strategic cities where civilian populations could be pressed into housing and feeding them. 32 London was a sensible home for the Comes when not on campaign and his forces would have eased the problems of finding soldiery to place on the city walls.
London might also have been the dominant site in the command infrastructure developed under the leadership of the Comes litoris Saxonici per Britanniam. 33 As has been noted (above p. 313) aspects of the design and supply of the suburban establishment at Shadwell showed close links with sites along the Saxon Shore, and some operations would have been best managed from the diocesan capital at London. The development of these commands, and of the coastal frontier itself, gave London much of its raison d’être.
A grave in the eastern cemetery at Mansell Street may have been the resting place of a senior figure from this community of Roman officials (Fig. 28.3). The body of a man more than 45 years old was buried in a wooden coffin with items of dress normally restricted to the upper echelons of late Roman imperial society. 34 The surviving tokens of his rank included an unworn belt set (cingulum) consisting of a large chip-carved buckle, strap-end, and three end-plates; and a rare crossbow brooch of the type shown on the Monza ivory diptych thought to depict the general Stilicho. 35 The belt was designed to go around a tunic decorated with panels of silk worn over embroidered trousers, and the brooch served to fasten a heavy cloak (chlamys) at the shoulder. By the fifth century this had become the attire of Rome’s senior military and state officials, following barbarian-influenced fashion. 36 The crossbow brooch is a type normally thought to have been introduced c. ad 390, but an unpublished radiocarbon date from the burial itself points to an earlier date with a high degree of probability (95.4 per cent) that the individual died within the period 237–365 cal ad. 37 Given the presence of the brooch it is likely that the burial took place in the 360s rather than earlier in this range. At such an early date these items would have been at the very cutting edge of imperial fashion, and we can assume that deceased had only recently been issued the symbols of his rank when he died. It has been suggested that this style of dress may have evoked the credentials of late Roman authority in Britain without necessarily identifying office holders, but the items buried at Mansell Street are perhaps more likely to have accompanied someone who had genuinely held high office. 38 If so, we might be looking at someone who in the mid-360s held one of the posts listed in the Notitia. Intriguingly the strontium and lead isotopes within his dental enamel indicate that this high-ranking individual had grown up in the London area, showing that some government officials were locally recruited. This wasn’t the only person to have been buried in London with elements of the ‘Germanic’ style of official dress. Another relief-cast buckle in the ‘chip carved’ style was found in the cemetery at West Smithfield, and similar items have been found at Fulham Palace and Enfield where officials may have been placed to command satellite sites. 39
Fig. 28.3 The burial of a man dressed as a high-ranking late Roman official, with a crossbow brooch and belt set (cingulum), from the cemetery at Mansell Street (adapted from Barber and Bowsher 2000. Reproduced by kind permission of Museum of London Archaeology).
Wealth within and beyond the walls
Although the city wasn’t densely populated, with housing spread thinly across an overlarge urban footprint inherited from second-century expansion, several properties were refurbished in the late fourth century. The early Roman warehouses at Pudding Lane were still standing, and industrial slag and ash suggests that some operated as workshops. Six successive floor repairs were laid in one room, the earliest of which contained a coin of Constantius II (ad 348–51). A small bathhouse found behind this structure was refurbished with new heated rooms and a small bath late in the fourth century, with a coin of Valens (ad 367–75) found in construction levels. 40 New floors were also laid within the tower of the residential complex at Plantation Place, where the latest surfaces date after c. ad 370. 41 For the most part these works involved the redecoration of existing structures, but they illustrate an urban vitality sustained over several decades. These works ensured that London retained a stock of high-status housing and productive workshops, meeting the needs of the government officials stationed here.
It is possible that London had fewer skilled carpenters, following changes to the mechanisms of timber procurement earlier in the century. A well at Drapers’ Gardens, coin-dated after c. ad 335 and associated with pottery as late as c. ad 350, was built using the labour-intensive ‘native’ styles of woodworking rather than more sophisticated Roman carpentry. 42
A few unusually important buildings were built within the city. One of these was investigated at One Poultry where a large property with tile-faced concrete walls was built at a prime location on the south side of the main road and flanked by a large brick culvert. Coins of ad 340–6 were found in material dumped before the building’s construction, and the presence of Portchester D pottery confirms that it was occupied in the second half of the century. 43 We don’t know what purpose it served, but its scale hints at a public use. It stood in an area that had previously housed granaries and bakeries and it is tempting to suggest a continued association with administered food supply.
Another important property was marked by a vast mosaic floor, measuring 7.4 by 7.25 metres, found at Old Broad Street in 1854. This pavement was later displayed in the Crystal Palace at Sydenham where it was lost in the fire of 1936. 44 It incorporated red, white, yellow, and grey geometric designs, with a central panel showing Bacchus accompanied by his leopard and canthari filling the corners. The pavement belongs to a stylistic group described as the ‘Corinian Saltire’ school by David Smith. 45 The artisans producing these designs normally worked for West Country patrons and the London floor was almost certainly made by the same craftsmen who laid a mosaic in a villa at Halstock in Dorset coin dated after c. ad 350. 46 The scale and quality of the London pavement indicates that it adorned a grand reception room, probably a dining room or audience hall similar to those of the richest luxury villas. The West Country connection is curious. Was this the residence of an emissary or official representative of the government of Britannia Prima attending the London court of the vicarius? Its design shows the continued currency of Bacchic iconography. By this date Orpheus was recognized as a saviour figure within his mysteries, and several of Britain’s later Roman mosaic pavements contained allegorical references to this Orphic interpretation of Bacchic cult. These ideas informed the design of the Old Broad Street pavement, without necessarily meaning that the room was a place of cult worship. This late antique Orphism, drawing also on Neoplatonist philosophy, exercised an important influence in the elaboration and evolution of Hellenized Christianity, as a consequence of which the worship of Bacchus and Christ were sometimes entwined in the fluid soteriological theologies that preceded Christian orthodoxy. 47
Changes were also made to the temple precinct on London’s southern approaches at Tabard Square. 48 In the fourth century the temenos adjacent to the north temple was redesigned, and an unusual building set along its east side. A couple of potsherds dated after c. ad 330 were recovered from the level at which its construction trenches were dug. This rectangular structure was decorated, but contained no heated rooms, and presented a façade formed by two projecting side towers or wings connected by a large corridor or portico. This reproduced the winged-corridor layout favoured in Romano-British villas. In this particular context the projecting rooms might have been suitable places to hosts cult feasts, and an abundance of pig bones, a sign of high-status dinning, were recovered from the temenos ditch. 49 Surfaces within the walled area were renewed into the middle of the fourth century, although the statues or columns set over the plinths had been removed. 50 These changes witness the continuing evolution of liturgical practice within London’s pagan spaces until the middle of the fourth century. It is possible that an earlier emphasis on the architectural furniture of open areas as a stage for cult sacrifice had given way to a religious practice concerned with select gatherings around shared meals and mysteries. The increasingly private nature of cult celebration may have turned attention away from other public temples in later Roman London.
London’s first Christian communities
This evidence for a diversity of pagan belief in fourth-century London contrasts with a relative scarcity of evidence for Christian worship. The cult would have found likely adherents within London’s cosmopolitan communities, and some people in Britain had probably identified as Christian from as early as the second century (above p. 324). These faithful are difficult to identify. Prior to the fourth century, most Christian worship took place in undistinguished domestic spaces. The eucharistic and baptismal practices that gave structure to the faith were easily accommodated within the dining rooms and baths of late antique houses. Similarly there was no particularly Christian approach to burial, beyond a concern to preserve the corpse for bodily resurrection that was shared with other faiths. 51 A fluid and inventive syncretism of ideas and symbols, where much rested on deliberately ambiguous allegorical references, makes it difficult for the uninitiated to distinguish between the art and iconography of different communities of believers.
The most exclusively Christian symbol was the Chi-Rho monogram, combining the first two Greek letters for the name of Christ, which gained currency after its early fourth century adoption by Constantine. A Chi-Rho was engraved onto the base of a pewter bowl found at Copthall Court, perhaps marking its use in eucharistic practice, and another was scratched onto a potsherd found in Brentford. A partial Chi-Rho graffito was also scratched onto the rim of a fourth-century mortarium found at Brandon House in Southwark. 52 These fourth-century finds are the earliest archaeological evidence of Christian affiliation from London, although we cannot know how knowingly these symbols were deployed once they became associated with Constantinian authority.
Twelve oval lead alloy ingots, weighing between 0.8 and 8.5 kilos each, found in river dredging and along the foreshore near Battersea Bridge also carried Chi-Rho stamps alongside other identifiably Christian components. 53 They were executed in a style that is unlikely to be later than the fourth century and stamped as the product of Syagrius. This may have been a member of the prominent Gallic family of that name that held high political office in the late fourth century. The stamps probably marked the ingots as official consignments under Syagrius’ control, dating after Theodosius made Christianity the religion of the state. Whilst they may have been lost in transit along the Thames, James Gerrard convincingly suggests that they arrived here in dumps used to reclaim marshland during the mid-nineteenth-century construction of Battersea Park, using material removed during the construction of the Shadwell basin. It is consequently possible that these ingots derived from a hoard buried near to the Shadwell estate speculatively identified as the residence of a high-ranking official (p. 313).
The written sources offer further detail on London’s first Christian communities, but come with their own difficulties. A late sixth- or early seventh-century Martyrologium identified a martyr-bishop named Augulus or Agulus from a city in Britain identified as civitate Augusta. 54 The implication is that Agulus was a bishop of London killed in a persecution before Christianity was recognized as licit in the Edict of Milan of ad 313. A list of British clergymen recorded as attending a church council at Arles in ad 314 was headed by a bishop from York, followed by bishop Restitutus from London (ciuitate Londenensi), and bishop Adelfius of the colony of London (ciuitate Colonia Londenensium). 55 A literal reading might suggest that there had been two communities of London: perhaps a colonial foundation within the walled city and a separately constituted community on the south bank. Since the church hierarchy treated each community as the preserve of a single bishop, in a monepiscopate that guarded against dissenting voices, this would require us to believe that London and Southwark were now considered completely separate cities. We have no other evidence for such an unlikely arrangement and Adelfius’ affiliation is therefore widely assumed to be a mistake for Lincoln, since Lindinis is only one vowel-shift away from Londinis.
British bishops participated at other fourth-century church councils, indicating a developed ecclesiastical leadership. The institutions of organized Christianity gave rise to new forms of public architecture in many late antique cities. Episcopal complexes defined the urban seats of powerful fourth-century bishops, and excavations in cities such as Geneva, Milan, and Aquileia show that these were built on a grand scale from the middle of the century. 56 These typically included a basilical church for congregational worship, a baptistery and a palace. Some communities were slower than others to follow this trend, and we cannot be certain that London housed an episcopal complex. Britain failed to match neighbouring provinces in its display of the monuments and symbols of what probably remained a minority religion. London was, however, an important administrative city that was otherwise quick to adopt architectures of Roman imperial authority. During the course of the fourth-century Christianity became the religion of the imperial household and eventually a religion of state. The patronage of emperors contributed to the monumentalization of Christian worship, and most senior appointments to London’s late fourth century administration are likely to have been Christian. 57 It is therefore likely that London’s bishops were sufficiently powerful and politically well-supported to have secured the resources needed to build a basilical church before the end of the fourth century.
Excavations at Colchester House in Pepys Street in 1992 found the heavily robbed foundations of a fourth-century building that might just possibly have been this church (Fig. 28.4). 58 Only the north-eastern corner of the building was found, but it was clearly a substantial public building. Its massive external walls were set over 2-metre-thick foundations dug into the natural gravels and supported by timber piles. Internally a series of square pier-bases were set over timber piles capped with flint and chalk and topped with concrete. These pier-bases were arranged in two rows to form a double aisle, and a short wall marked off the eastern end of the building. The arrangement finds close parallel in the basilica of Santa Tecla in Milan. London’s basilica was provided with a stone and tile floor and a well, whilst fragments of marble and window glass suggest a well-appointed interior. This debris also included limestone slates from the building’s roof: these slates came into occasional use in London after c. ad 350 when the city ceased to import roof tile from Harrold (above p. 344). 59 A construction date after c. ad 350 is also suggested by the presence of significant quantities of Portchester D pottery, which marks later Roman occupation. The structure was set across the projected line of a Roman road on a prominent site on Tower Hill in the south-eastern quarter of town: a formerly suburban area that may have been absorbed within the city when it was extended to the line of the town wall. This was an appropriate location for an episcopal church, expeditiously using a vacant but prominent site on the borders of the settlement. 60
Fig. 28.4 The late antique basilica considered the possible site of London’s episcopal church, as identified in excavations at Colchester House (PEP89: after Sankey 1998). Drawn by Justin Russell.
It has alternatively been suggested that the basilica was a large granary (horrea) designed to hold surplus grain commanded by the annona. Large aisled granaries featured in late antique walled cities and frontier forts. This is not, however, a good fit for the evidence, since granaries were not normally decorated with marble interiors and Tower Hill was not ideally placed for this purpose. Other aspects of the finds assemblage also suggest that this part of Tower Hill was the focus of religious observances in the fourth century. James Gerrard has shown that the coin loss profile from Colchester House finds closest parallel with late Romano-British religious sites, whilst the pottery from beneath the basilica was marked by a concentration of an unusual type of bell-shaped bowl commonly found in votive deposits. 61 Whilst we have no proof that the building was a place of Christian worship, this is a credible proposition and is more consistent with the evidence than the alternatives.
The restoration of London in context
Little of London’s earlier public landscape survived into the second half of the fourth century. Major changes to the urban topography are illustrated by the redundancy of buildings around which civic life had previously revolved: the port, forum, temples, and amphitheatre. There was a marked shift away from symbolic monumentality, and public space was more likely to have been used for industrial production than extravagant civic ceremony. These transformations followed an earlier relocation of power into the private domain, consolidated around a small number of wealthy establishments. The rites of public sacrifice that made London’s temples an important part of the environment were in decline throughout the Roman world, and many expressions of belief had moved into the private sphere. Outside of the domestic arena, urban ideologies were chiefly represented by the city walls, although new places of worship may have created other focal points. Processions no longer attended the games at the amphitheatre, but ceremonial life may have been reconfigured around Christian and other cult practices. 62 It is distinctly possible that the feasts of martyrs inspired new processions and stational liturgies, uniting an episcopal basilica with other places where the remains of martyr-saints might be celebrated. As Simon Loseby has explained, we don’t know how far down this route the towns of Roman Britain had progressed before the collapse of the Roman administration, but the early stages in the evolution of such topographies need not have left significant trace. 63 Sadly we have no direct evidence for any such invention of new ideas of city living, but it is certain that the architecture of earlier forms of public life were largely redundant.
This chapter has surveyed evidence for London’s mid-fourth-century revival, best explained as a centrally directed initiative to refortify the seat of diocesan institutions. This is consistent with what we know of Valentinian strategy, and of the actions taken by Count Theodosius following the barbarian conspiracy. These remain suggestions only, since we lack precise dating evidence. 64 Our purpose, however, is not to find archaeological illustration of the events of ad 367–8 but to seek explanation for one of Roman London’s last major programmes of architectural renewal. Even without Ammianus’ commentary it is reasonable to conclude that these works, like earlier phases of restoration, would have followed investment in the command infrastructure drawing on resources marshalled against military campaigns. History suggests that one of the most important of these campaigns was underway ad 367–8 when Theodosius over-wintered in London at the head of a field army of around 2,000 men. 65
The new works did not embrace London’s port. The earlier phase of tidal retreat appears to have reversed by the fourth century, and returning tides would have eased shipping up-river to the city. 66 It would have been possible to engineer new waterfront quays around gates through the riverside wall, but this did not happen. London was no longer needed as a supply-base. The volumes of goods being moved were massively reduced after the annona reforms of the previous century whilst other sites, including those of the Saxon Shore, had risen at London’s expense (above p. 345).
London continued to import goods for local consumption, drawing on a range of regional sources for its pottery. From c. ad 350 this included Oxfordshire wares, Calcite-gritted wares likely to have been made at kilns near Harrold, and Portchester D wares from Hampshire. 67 Some reached London down the Thames, in a traffic that was also directed towards military sites. 68 Contemporary changes involved a decline in the presence of south coast Black-Burnished wares, perhaps starting from a little earlier in the century, and in grey wares from the Alice Holt kilns. 69 Malcolm Lyne observes that Alice Holt provided more than 60 per cent of the coarse pottery found in early and mid-fourth-century assemblages in London, but a much lower proportions in late fourth-century contexts. He suggests that a negotiator responsible for supplying London with these wares ceased operations after c. ad 370. Since the town was no longer importing much building timber, this may have reduced traffic from Alice Holt adding to the marginal costs of moving pottery vessels and their contents. 70 This period may also have seen a significant decline in volumes of goods travelling the east-coast routes. 71 London and coastal sites in the Thames estuary continued to receive small volumes of continental imports including pottery from the Argonne region of northern Gaul and Mayen ware (also known as Eifelkeramik) produced near Trier in the Rhineland. 72 Much of the pottery reaching town may have followed in the wake of the bulkier supplies of grain, beer or wine consumed by the urban population. Some aspects of London’s supply may also have been driven by official requirements, sustaining links with the Rhineland and fortified coastal sites associated with the Saxon Shore.