Part 8
30
The desertion of the city
We know that some of London’s houses were occupied until the very end of the fourth century, but there is no certain evidence that any were actively maintained in the early fifth century. A hoard of copper-alloy coins scattered over the latest floors of the furnace room and corridor of the domestic baths at Billingsgate was assembled after ad 395 since it included issues of Arcadius and Honorius.1 It had probably fallen from a place of concealment in the east wing of the building, implying that the house was still standing around ad 400. This late phase of occupation was also reflected in the presence of part of a Palestinian amphorae. A few other sherds of these exotic amphorae have been recovered from sites east of the Walbrook, as at New Fresh Wharf.2 They were rare imports, perhaps carrying wine from the Holy Land for religious liturgy. Floors associated with late Roman pottery have also been recorded at Pudding Lane and in the former Mithraeum, hinting at occupation into the early fifth century but not proving it.3
We cannot put a timeline to the continued use of the buildings standing at the end of the fourth century, and some slight timber constructions of this period may have escaped attention. In the previous chapter, we reviewed the difficulty faced in finding secure dating evidence after the close of the fourth century, but the little we have is consistent with the argument that most of the city was soon derelict following the trends of closure and abandonment evident in the last quarter of the fourth century. This remains the likeliest explanation for our failure to identify later phases of repair and reuse within the houses that were still occupied at the end of the fourth century. A dearth of early Saxon finds from within the walled city leaves little doubt that it was largely abandoned by ad 450. This cannot be a failure of recognition, since early Saxon material has been recovered from London’s less thoroughly explored hinterland (below p. 396).4
London’s abandonment followed the failure of the provincial government.5 By the end of the fourth century Trier was no longer a capital city and the seat of the Praetorian Prefect transferred to Arles around ad 395.6 This placed distance between Britain and sources of authority, leaving it increasingly isolated. Local policing by settled groups had largely replaced mobile standing armies, eliminating the need to secure long-distance supply. The garrison in Britain was no longer supplied with bronze coinage and the flow of new coin reduced to a trickle after mints were closed at Arles, Lyons, and Trier c. ad 395.7 By the Theodosian period (ad 388–402) coin-use was largely restricted to internal transactions within cities and forts. It is important to note that earlier failures in supply had provoked local episodes of copying, meeting a demand for low denomination coinage for essential transactions (above p. 331). Significantly, this did not happen after ad 395. London may have continued to use the small change already in circulation, but the failure to compensate for the decline in official supply shows that cash-based exchange mattered less.8
So why was there no longer an equivalent need of coin for local purchase to pay hauliers, supplement supplies, and facilitate market activity? We must conclude that London housed a different and smaller community with little need for such transactions. The people still living here were perhaps more easily fed from the immediate hinterland, where a greater degree of local self-sufficiency reduced the need to buy goods in the marketplace or pay for services. By the end of the century, London had a negligible role to play in raising taxes and rents to support the Roman administration. Few grain waggons trundled the damaged streets. It is doubtful that taxes were routinely raised at the city gates, or port duties charged on cargoes beached on the foreshore. It is unlikely that the walls were regularly manned, since the feeding of the garrison would have brought carts back onto the roads and put small change into soldiers’ purses. Supplies of luxury goods diminished, and traffic shrank to the point where bridges and roads were no longer maintained.
The immediate train of events leading up to Britain’s political rupture with Rome started ad 405/406 when massed barbarian forces crossed the frozen Rhine.9 In the ensuing chaos, mutinies in Britain raised a succession of usurpers culminating in the proclamation of Constantine III in ad 407 who, following earlier example, assembled a fleet and army to cross to Gaul.10 Constantine’s doomed campaign diminished whatever was left of the professional military in south-east Britain, perhaps including the command of the Comes Britanniarum. According to Zosimus the British subsequently expelled the remaining Roman officials, taking up arms on their own account to free cities from barbarian threat in ad 409.11 A complete vacuum in the Roman command structure might be implied by a letter written by Honorius, the emperor of the west, advising the cities (poleis) of Britain to look to their own defences, although there are some doubts as to whether this communication was not directed to Bruttium in southern Italy instead.12 The instruction is not inconsistent, however, with the situation in Britain and might indicate that there were no Roman officials here who could be charged with this duty. Compounding the earlier redundancy of civic institutions with the abdication of Roman military authority, and for the want of any abiding social or economic function, London had no purpose. Like other Romano-British cities, it lost its role in the extraction of resources from dependent territories and as a centre of power. Desertion accompanied the withdrawal of the political and military presence.13
Many histories of the end of Roman Britain stress a political rupture c. ad 410, but it can be argued that the Roman diocesan government had struggled to recover from the break in authority under Magnus Maximus in ad 383. After that date, Britain was only intermittently governed through the Gallic prefecture, totalling only 8 years in all. London’s faltering political role during this quarter-century of weakened central authority may have contributed to the loss of urban infrastructure. Urban lifestyles evaporated with the urban institutions that engendered them.
Rural continuities
Although the walled city was largely deserted in the early fifth century, the surrounding countryside fared differently (Fig. 30.1). A few late antique sites, especially in Southwark and Westminster, present compelling evidence for continuity into the fifth century, and formed focal points within the post-Roman landscape.
Fig. 30.1 London and its surrounds in the early fifth century (showing also principal features of the later Saxon settlement). The settlement pattern shows a striking degree of continuity with that of the late pre-Roman Iron Age, as shown on Fig. 4.3 (p. 40). Locations of pottery and finds with early fifth-century characteristics are marked + and identified by site-codes. Drawn by Fiona Griffin.
It is likely, however, that Shadwell shared in London’s fate.14 It is not possible to fix the precise date that the baths here were last fired up, but the destruction of associated structures seems to date from c. ad 375, and the nearby tower or mausoleum was covered by demolition debris containing finds dated after ad 365. A crude well was dug to supply fresh water in the later stages of the settlement’s life, when the piped water had presumably failed. This contained a ‘closure deposit’ consisting of complete and broken ceramic vessels, with a high proportion of red-slipped tablewares from around Oxfordshire, and a copper-alloy bowl. Pottery from the site included fabrics dated after ad 370 including North African and Palestinian amphorae.15 It seems likely that the patterns of high-status social behaviour, signalled by the continued operation of the bathhouse, were not sustained until the end of the century, by which time the site was largely deserted. The fortunes of Shadwell, like those of London itself, had perhaps relied on the continued presence of the military administration. Some roadside sites around London may have followed a similar trajectory of change, although evidence is sparse. Coins found on the clay floor of a cellar at Old Ford witness its continued use beyond c. ad 383, with others of Honorius (ad 395–402) within its backfill. This late activity might follow from the continued exploitation of the Lea valley for grazing, and chart official traffic along the Colchester road, but needn’t have continued long into the fifth century.16
Our best evidence for fifth-century continuity comes from St Martin-in-the-Fields, some 2 kilometres west of London. The presence of an important late-antique site here was suggested by the discovery of Roman brickwork and two stone sarcophagi during church building in ad 1722–5.17 More recent discoveries included a limestone sarcophagus containing human remains carbon-dated ad 390–520 and a tile kiln.18 The kiln had a double flue with stacked floor tiles and an outer wall of chalk blocks, and archaeomagnetic dating has placed its last firing within the period ad 400–450. Wasters derived from tile production included roofing tile and combed box-flue supplying specialist materials for building heated rooms and tiled roofs at the dawn of the fifth century.19 These products are likely to have been used to enhance Roman-style villa architecture at a time when the necessary building materials could no longer be obtained from supplies routed through London.
The sarcophagi are likely to have been set within a mausoleum of the sort associated with other late antique suburban villas around London, in order to commemorate members of the aristocratic family seated here. This high-status Roman site became the focal point for a small early medieval cemetery, represented by a group of at least ten inhumations on the north side of the church. One of these was accompanied by a Saxon jar of a style likely to date from c. ad 430–70. We cannot know if this evidence represents a direct continuity of use of the site as a place of burial for an established community, or the reuse of a site held to be significant by later communities because of its Roman remains. Michael Garcia’s PhD thesis describes how the early medieval reuse of Roman mausolea created an artificial continuity with the Roman past. Saxon churches built following St Augustine’s foundation of an English church at the end of the sixth century sometimes favoured sites of Roman tombs because these were places where the graves of martyrs were thought to be found, in an appropriation of late antique monuments that gave new Christian sites legitimacy and the illusion of continuity.20 This may explain why several of London’s medieval churches stand over important late antique buildings. Important examples include St Bride’s, St Andrew Holborn, All Hallows Barking, Bermondsey Abbey, St Michael, and St Peter Cornhill, although the reuse of these sites may also have been influenced by the availability of building materials that could be quarried from Roman ruins. Most of these places probably became Christian through later re-invention, but we cannot dismiss the possibility that a few were places where saints had been venerated at the beginning of the fifth century. The graves of martyrs became places of worship on the borders of cities in Gaul and Italy in the late fourth and early fifth centuries.21 St Germanus’ visit to the tomb of St Alban suggests that the cult of saints was introduced to Britain by the end of the fourth century, and had probably developed along the same trajectory as on the continent.22 It is therefore credible that places were found for the veneration of London’s martyr saints, but since we have no archaeological evidence for their existence there is little point in speculating further.
What the evidence recovered from Al Telfer’s investigations at St Martin-in-the-Fields tells us with greater certainty is that some farms and villas near London outlasted the city. Small quantities of early Saxon pottery, dated to the late fifth century, have been found on several such sites and although this is not certain evidence of continuity of settlement, it seems likely that some farmlands around London continued in cultivation through much of the fifth century. Just as the Roman conquest had a limited effect on patterns of settlement within London’s more rural hinterland, so too did the failure of its Roman administration.
Two recent discoveries from just outside the walled area probably belong to this period. A coin-hoard of nineteen gold solidi and 114 silver silique, closing with issues of the house of Theodosius c. ad 402, was buried in the fills of a boundary ditch associated with the cemetery on the west side of Ermine Street about 800 metres north of the town wall.23 This is a fifth-century treasure, and Peter Guest suggests that hoards of this type were assembled in the post-Roman period.24 A single silver coin found within dark-earth in excavations on the southern borders of Southwark may also date to the early fifth century. This coin of Valentinian I (ad 364–7) had been clipped to remove silver from its edge so as to leave the imperial portrait intact.25 It is thought that the fashion for clipping coins in this way began in earnest after the usurpation of Constantine III in ad 407, lasting for a few decades. It is significant that nothing similar has been found within the walled town, despite the greater intensity of archaeological research within this area. The former city seems to have been positively avoided, perhaps vacated because of its ideological redundancy in a fragmented post-Roman landscape. Roman cities had no place where Roman rule was rejected.
Early Saxon settlement
A similar pattern can be traced from distributions of late fifth-century Saxon pottery. The chaff-tempered fabrics of this period are difficult to date, but occur in small numbers immediately outside the Roman town whilst largely absent from within the walled circuit. Recently recovered material shows two principal areas of activity: one along the Fleet valley and the other on the southern shores of Borough Channel. A Saxon settlement at Clerkenwell, beyond the Roman cemetery in Smithfield and 400 metres north of the Roman town wall, is implied by a cluster of eight pits containing Germanic-type pottery dated c. ad 450–550 and a Saxon burial.26 The people here may have been the source of four sherds of fifth to sixth-century pottery found in the ditch outside the city wall at Aldersgate.27 A few sherds of similarly dated material have been found on the west side of the Fleet, at St Bride’s and possibly at the Inner Temple.28 Small-scale early Saxon activity on the other side of the Thames, next to the Borough Channel, is indicated by small assemblages of early Saxon pottery from a ditch and pit at Lant Street, which included a sherd of igneous tempered pottery with a characteristic kind of slip (schlickung) dated ad 400–600, and by material from Trinity Street.29 Somewhat larger quantities of early Saxon pottery have been recovered from Bermondsey Abbey, where elevated numbers of later fourth-century coins have also been recovered.30
Most of these late fifth- or early sixth-century sites had been settled in the Roman period, and those at St Martin-in-the-Fields, Clerkenwell, and Bermondsey may have originated as late Iron Age farmsteads. These were naturally advantaged sites, occupying well-watered locations of strategic aspect near to navigable waterways. There may have been continuity of settlement in or around these places. There was no particular incentive for local farmers to abandon their fields simply because the Roman government had evacuated their London offices. The use of Saxon styles of pottery is not necessarily a marker of ethnic identity and may have been a cultural and economic choice, whilst the survival of place names with Celtic and Latin elements around London might indicate a continued British presence.31 In sum, there is no evidence of a breakdown of order in the countryside, and those who farmed the land may have continued to do so.
Whilst evidence for a fifth-century presence beyond the city walls is thin, it exists. This contrasts with an almost complete absence of comparable evidence from within the walled area. The only stratified post-Roman artefact from the city confidently dated to the fifth century is a Germanic-style saucer brooch found among fallen roof tiles in the frigidarium of the ruined Roman bathhouse at Billingsgate.32 We do not know how the brooch reached this spot, but Billingsgate is likely to have remained a suitable point of entry into London. Finds on the foreshore suggest that official cargoes were unloaded here until the middle of the fourth century, implying the presence of a water-gate at a beaching spot for rivercraft coming upriver to London (above p. 356). The brooch is identical to ones from graves in early Anglo-Saxon cemeteries established following the movement of people up the Thames and its tributaries from c. ad 450 onwards. The pattern of early Anglo-Saxon settlement is best explained as the result of immigration into the area after the failure of the Roman city, along with some adoption of the dominant immigrant culture by surviving rural communities. The main fifth- to sixth-century occupation sites were new settlements: villages in the river valleys housing farming communities of a few dozen people.33 They had little or nothing to do with Roman London. This is also likely to be the case for the ring of early Anglo-Saxon settlements and cemeteries south and east of London that John Morris suggested had been placed to defend the approaches to sub-Roman London.34 These date no earlier than the mid-fifth century, by which time the city was uninhabited. The pattern of settlement is likely to have been determined by the arrival of immigrants seeking suitable sites for settlement, where the Adventus Saxonum followed after the Roman government had already failed.35
Earthworks on the borders of the London region might mark territorial boundaries that emerged from the post-Roman fragmentation of political authority, but are too poorly dated for this to be established. Grim’s dyke stretches west from Brockley hill to the river Pinn, and perhaps west to Manor Farm, Ruislip.36 It survives up to 6.9 metres wide and 1.8 metres deep, alongside an earth bank and forming a substantial barrier. Excavations in Pear Wood, near Brockley Hill, found fourth-century material beneath the bank, although the south-western part of the earthwork, near Pinner, might have been thrown up in the pre-Roman Iron Age. It followed a line that might plausibly have been the boundary of a British Kingdom in the Chilterns. A bank and ditch on the east side of the Cray Valley may have similarly marked the northern boundary of a Kentish polity.
London after Londinium
London itself remains largely absent from the historical sources of this period. There are no reliable references between its late antique listing in the Notitia Dignitatum and ad 601 when Pope Gregory wrote to Augustine proposing London as the primary See.37 The Anglo-Saxon Chronicle presents a single fifth-century entry referring to London, as a place to which Britons fled after defeat at the hands of the Jutish war-leader Hengest at Crecganford (identified as Crayford) in ad 457.38 The entry is suspect, but even if it drew on an historical event, it tells us little of use.
A few other early medieval finds from within the walled area suggest that the city attracted sporadic interest through the sixth century. Two Frankish ceramic vessels were found at Gresham Street and Christ’s Hospital and a Merovingian-style buckle loop was found in a twelfth-century context at the Guildhall.39 London may have been uninhabited for over a century by the time these finds reached here, although it seems likely that land within the walls continued to be cultivated. Much may have been grazed. Liberated from the need to pay Rome’s taxes and rents arable production may have declined, shifting emphasis from producing grain to raising livestock.40
The political impetus for the re-occupation of the city was provided by the foundation of a monastic community at St Paul’s in ad 601 by the missionaries sent by Pope Gregory. Gregory proposed the establishment of two ecclesiastical provinces in Britain, at London and York, inspired by the former importance of these imperial cities. These expectations may have been the catalyst for the re-emergence of the walled city as a place of consequence, generating a scatter of middle Anglo-Saxon finds.41 This was not, however, an urban site. London’s seventh-century urban revival focused instead on the area of the Strand, east of the river Fleet, where new roads and houses were developed next to the site at St Martin-in-the-Fields.42 This was Lundenwic, the busy emporium described by Bede which grew to cover an area of over 60 hectares, almost half the size of the former walled town. The reasons for the growth of this entrepôt are similar to those that encouraged the Romans to build London: it lay on the orbit of several neighbouring polities, in a geographically opportune spot with access to the sea, river, and overland travel.
The restoration of the old walled city didn’t take place before the ninth century. This is generally attributed to the policies of King Alfred the Great although recent research suggests that widespread revival didn’t take place until the late tenth or early eleventh century when London Bridge was probably rebuilt.43 The surviving city walls framed and shaped the medieval city, which thrived on a fortunate location where the river was readily bridged and its tides drew the world to England. Whichever the case, this is the start of a different urban history. The medieval city was a new foundation that inherited little from the Roman city beyond the idea, still very much with us to this day, that this was a place from which Britain could be governed.