BRITAIN

MICHAEL FULFORD

Almost the entire period between the accessions of Vespasian and Septimius Severus was dominated by military affairs in Britain. Even during those relatively short periods when little or no campaigning is recorded as having taken place it can be argued that developments elsewhere in the civil zone of the province were conditioned by events on the frontier. Recently, greater emphasis has been placed on the achievements of the native aristocracy where the development and pace of romanization is seen to reflect more the latter’s attitudes and aspirations towards Rome rather than the results of policies imposed by the invader. However, it is difficult to isolate civil developments from a framework imposed by the progress of conquest and its associated administrative structure.

Our written sources are dominated by Tacitus and, in particular, by the biography of his father-in-law, the Agricola, which also includes a commentary on political developments in Britain before the latter became governor. The Histories also offer some useful insights into the Civil War of 68—9 and relations with the Brigantian client kingdom during that period and at the start of the 70s. Second-century historical sources are much more limited, but nevertheless, and in conjunction with a rich epigraphic record, offer a framework for the principal developments on the northern frontier: the construction of Hadrian’s Wall and the Antonine Wall. Although the recently discovered series of documents from Vindolanda offer vivid insights into military organization and military life in the north at the very beginning of the second century, they do not add materially to the historical narrative. The latter has had graphic illumination from the archaeological record in the form of the physical remains of the frontier systems, forts and the legionary fortresses, but only the association with closely datable inscriptions, or strong coin evidence, allows good correlation with the historical narrative. While there has been a long tradition of trying to associate the military history of Roman Britain as derived from written sources with the archaeological record of forts, fortresses, and the like, much of the basis for this is inherently speculative, depending as it does on a series of tenuous links, more often than not linked to ceramic evidence, which, in its turn, is dependent on correlations with otherwise independently dated deposits. Whereas a coin will almost invariably provide a terminus post quem.

559

The pottery that is dated by association is usually broken and distant by an unknown period from the date of its manufacture. Thus the date of a deposit of, for example, new sigillata, such as the crate containing South Gaulish sigillata buried by the destruction of Pompeii in a. d. 79, will probably be much closer to that of its manufacture than a deposit of broken sherds of identical material from a rubbish pit associated with coins which reveal varying degrees of wear. Thus, although there is a strong expectation of developing chronologies of material culture that will allow a close reading and interrelationship with the historical narrative, it is an unrealistic one. Except where particular circumstances prevail, such as with dendrochronologically dated sequences, it is usually difficult to work in archaeological periods of less than about twenty-five years. The Flavian governorships of Cerialis, Frontinus and Agricola provide good examples of the difficulties of correlating the archaeological and written records in that there can be no easy resolution of precisely who was responsible for conquering which territories or for the construction of which forts. Even in the case of Agricola, where we have more information from Tacitus about the extent of his conquest in Scotland, there is considerable uncertainty over what should be attributed to him, rather than his predecessors and successors.

In any case, while the written sources, including the epigraphic material, largely focus on frontier affairs, there is almost no evidence to relate to the civil province and the development of administrative structures, cities and the romanization of the island. Here we are almost entirely dependent on archaeological evidence. In this connection, and within the period with which we are concerned, much has been made of one brief reference to the civil policies of one governor, Agricola, with whom much has been linked in the archaeological record.1553 In effect such correlations cannot be sustained by the evidence. Indeed, the one inscription of a. d. 79 from a civil, public building which mentions Agricola (the forum-basilica at Verulamium) records the completion of a structure which can only have been started some years before his appointment; the initiative was not his.2 Thus, although the non-militarized part of the province probably consumed as much, if not more, energy and resources as the frontier zone, we are dependent on a slow acquisition of material evidence to characterize the full extent of the achievement in the romanization of the province during the period between 69 and 193.

I. the frontier

At the outset of our period only three legions remained in Britain, legio XIV having been withdrawn in 66 or 67 from its base at Wroxeter for service in the East. Changing legionary dispositions in Britain point to substantial retrenchment after the suppression of the Boudican revolt: the fortress at Usk, probably occupied by legio XX, had also been abandoned sometime after 65. Thus, revised arrangements for the supervision of south Wales were necessary and evidence from Gloucester points to the foundation of a new legionary fortress there from the late 60s. This is likely to have been occupied by legio II Augusta, although evidence from its base at Exeter suggests continuation there of military occupation into the 70s. Legio XX, another possible candidate for Gloucester, is likely to have occupied Wroxeter in place of legio XIV. Legio IX remained at Lincoln.3 This disposition is likely to have been in place during the governorship of Vettius Bolanus (69—71) whose military resources did not apparently allow him the means to intervene in the client kingdom of Brigantia where Cartimandua had been usurped by Venutius.4

This task was left to Petillius Cerialis, with whose appointment a fourth legion (II Adiutrix) was restored to Britain.5 His brief from Vespasian was clearly to tackle the problem of Brigantia and to advance the conquest which had effectively been suspended since the Boudican rebellion. That policy was followed by the two succeeding governors, Iulius Frontinus (74—8) and Gnaeus Iulius Agricola (78—84). Initially, under Cerialis, the thrust was to the north, into the territory of the Brigantes, and certainly as far north as the Solway—Tyne line.6 The legionary fortress at York (for legio IX) appears to have been founded in the early 70s, leaving legio II Adiutrix at Lincoln at the outset. Frontinus, on the other hand, resumed the conquest of Wales, concentrating first on the southern tribe of the Silures and establishing a new base for legio II Augusta from c. a. d. 75 at Caerleon, while Agricola completed the conquest of the tribes of north Wales in the first year of his appointment. This was accompanied by the provision of a new fortress for legion II Adiutrix at Chester which was under construction in 79.7 Agricola’s tenure of office was longer than usual and extended through the Principate of Titus into that of Domitian, and it saw the conquest of northern Britain almost completed. The foundation in 84 of the legionary fortress at Inchtuthil on the River Tay for legio XX symbolized the military achievement of that governorship.8 This success was short-lived; even during his term of office some troops had had to be withdrawn to serve under Domitian against the Chatti. Following the major invasion of Roman territory by the Dacians in 85, further reinforcements from Britain were required to serve on the Danube. This necessitated the withdrawal of legio II Adiutrix by about 86 or 87, and certainly by 92, along with an uncertain

3 Webster (1988)passim; Manning (1981). 5 Attested at Lincoln, RIB i 253, 258.

4

Tac. Hist. 111.45.

6 Dendrochronological dating gives certainty of a Roman fort at Carlisle from a. d. 72/3: Caruana (1992); McCarthy (1991). 7 RIB ii 3, 2434. i—3. 8 Hanson (1987); Pitts and St Joseph (1985).

Number of auxiliaries. This loss of troops called for the precipitate abandonment of the unfinished fortress at Inchtuthil and, over the next fifteen to twenty years, also of the network of auxiliary forts stretching back through Scotland down to the line of Stanegate (Carlisle to Corbridge) and the Rivers Tyne and Solway. Legio XX was redeployed to Chester in place of legio II Adiutrix. Whether it is appropriate to describe the line of forts which had stabilized along Stanegate and down the Cumbrian coast by about 105 as a frontier is unclear. Either a lack of resources — the commitment to Dacia — or a deliberate disregard for Britain determined a policy of consolidation on the part of Trajan. This is exemplified by the programme of rebuilding in stone of the existing legionary fortresses at Caerleon, Chester and York as well as many of the associated auxiliary forts.1554

It is usually accepted that the campaigns of the Flavian governors were prompted by military necessity and there is, not surprisingly, a certain amount of evidence for this in Tacitus’ accounts concerning relations with Brigantia. The divisions within the client kingdom of Cartimandua, who lost control to her husband Venutius, may in particular have invited military intervention, but was total conquest of Britain the appropriate solution? The decision by Trajan to abandon so much of what Agricola had won rather implies that concerns about renewed warfare on the part of the natives were not paramount; the loss of northern Britain, whether temporary or not, was not a major strategic issue. This also seems to be borne out by the evidence for the posting of legio IX Hispana outside Britain sometime between 108 and the early 120s.1555 Until the arrival of legio VI, probably in 122, the island may have been left with only two legions for ten or fifteen years. Other possibilities might also be canvassed to explain the apparently enormous scale of resource devoted to Britain under the Flavians. Awareness of the power of the legions on the Rhine in influencing the rule of the empire would have been very fresh in Vespasian’s mind; the diversion of the energies of his best generals in a series of major campaigns safely beyond the continental shores was surely a significant factor in deciding on policy towards Britain. Native considerations were, perhaps, a lesser influence.

The writing tablets found at the auxiliary fort of Vindolanda on Stanegate, which date to the closing years of the first and the opening years of the second century, offer us some important insights into frontier life and the deployment of a garrison. The documents contain relatively few references to fighting or campaigning; in one instance the natives are referred to derogatively as Brittunculi. Other documents refer to the out-stationing of substantial numbers of troops; one detachment at Corbridge, another at the governor’s headquarters in London. This hardly provides us with a picture of a garrison on full alert and at full strength ready to confront a dangerous and numerous enemy. Equally, the references to the long-distance procurement of a wide range of supplies implies that lines of communication through the northern military zone were relatively secure behind the frontier line. More pertinent still, perhaps, is the correspondence of the fort commander’s wife which also gives no hint of inse-cunty."

In this context it is difficult to construe the motives behind Hadrian’s decision to construct the great 130 kilometre linear barrier between Solway and Tyne, with further extension along the Cumbrian coast. However, Hadrian’s biographer refers to difficulties at the beginning of his reign and coins struck in 119 commemorate a victory in Britain.12 These may well have followed the withdrawal of legio IX, thus coinciding with a period when Britain was lightly garrisoned. Nevertheless, Hadrian’s decision also reinforces indirectly that of his predecessor to withdraw from Scotland, in that no strategic sense was seen in prosecuting conquest further north. Given that comparatively little data have so far been assembled on the size of the native populations, and that what exist so far point to relatively low densities, a frontier defence of this magnitude seems out of all proportion to its likely defensive function. Nevertheless, we should recall that this would have been just about the time when a new generation of fighting age would have emerged in the north. In one area, namely manpower, the linear frontier may have been a more economical proposition than the conquest and occupation of Lowland Scotland. In other respects the costs of construction and maintenance of a system with an even distribution of forts, milecastles, turrets and a continuous curtain wall can only have been very substantial, particularly in the developed state of the monument, and may possibly have exceeded the costs of a looser disposition of forts which also embraced the Lowlands of Scotland. Indeed, in terms of possible manpower requirements, it has been estimated that only an additional 4,000 or so troops would have been necessary to hold all the north of Britain, including the Lowlands, and the successor frontier system to Hadrian’s Wall, the Antonine Wall.13 As a further indication of the scale of the undertaking, the building of Hadrian’s Wall involved all three British legions (one of which, legio VI Victrix) had been brought over by Nepos, presumably to replace legio IX Hispana) as well as the Classis Britannica and numerous auxiliary detachments. It was a prestigious undertaking; a strong symbol of the power of Rome to face an enemy of uncertain size and strength.

That Hadrian’s Wall may have had little to do with confronting a particular military threat from the north is perhaps best brought out by the decision of his successor, Antoninus Pius, to build a shorter, linear frontier between the Clyde and Forth in Scotland. Although less than half the length of its predecessor and requiring a garrison about 60 per cent of the estimated establishment of Hadrian’s Wall, additional forces were required to police the Lowlands and to man a series of outpost forts to the north of the new wall. Thus, whatever savings might have been made by the construction of a shorter barrier, these were, as we have seen above, more than offset by the new establishments in the Lowlands and to the north. The new wall was begun almost at the outset of Antoninus’ reign and was probably completed by 143. In many respects, it represents an advance on its predecessor, not least in the economy of its construction where turf was used rather than stone, and in the variable size and closer spacing of its accompanying forts. The reasons for the move northward and the abandonment of as massive an undertaking as Hadrian’s Wall remain obscure. There is evidence neither of unrest at the end of Hadrian’s reign nor of inadequacy in the strength of military dispositions; nor were any economies made in setting up the new frontier and its associated forts to north and south. The closeness of the policy change with the accession of the new emperor prompts the conclusion that the decision was determined by the political situation in Rome and the need for Antoninus to earn prestige by military success rapidly consolidated by a distinctively new frontier arrangement.

The subsequent history of the Antonine Wall is somewhat problematic. There seemed extensive support for a short break in occupation in the mid-150s with Hadrian’s Wall being temporarily put back into commission. However, close examination of the evidence from the Antonine Wall does not bear out this hypothesis and the case for its continued occupation seems stronger.14 Nevertheless, an inscription from the Tyne recording the arrival of reinforcements from the German armies during the governorship of lulius Verus, and thus assignable to the mid-i 50s, combined with coin issues of 154—5 showing Britannia subdued, suggest that there were problems in the north at this time.15 That these were overcome is indicated by the evidence for the continued retention and occupation of the Antonine Wall into the early or mid-i6os. The date of the final abandonment is unclear for, while the evidence of the imported sigillata from Gaul, the most abundant source of dating evidence, makes it clear that the two walls were never held together for any length of time, it does not provide close dating for the abandonment of the Antonine Wall.16 Taken with the coin evidence, it seems clear that the turf wall was certainly held up to the beginning, or earliest years, of Marcus Aurelius’ Principate, but some

Garrisons may have been kept in place until the 170s. As with the abandonment of Hadrian’s Wall, it is not clear what role — if any — native opposition played in the decision and, once again, it is tempting to associate such a major change with the new policies of Marcus. The problems presented by the Marcomanni and Sarmatians on the Danube frontier posed a serious challenge and, as with the Dacians before, a certain strategic redeployment of resources away from Britain may have been necessary.

Nevertheless, the troubles of the 150s perhaps provide evidence for the beginning of a shift in the balance of power towards the native societies of the north. This is further strengthened by the events early in Commodus’ reign when Dio reports that the greatest war of his principate was fought in Britain.17 Attributed to 181 with coins celebrating victory in 184—5 ,itis possible that these troubles contributed to the final abandonment of the Antonine Wall. Archaeological corroboration may be found in the burnt and destruction deposits from forts of Hadrian’s Wall and its hinterland. However, given the problems of dating outlined above, while it is possible that some or all of this evidence may be attributed to the Commodan war, equally it may date to the period of the usurpation of the governor Clodius Albinus in 196—7, or to another otherwise unrecorded event. Although there is no independent evidence for trouble after the removal of troops from the island to fight Severus for the empire, the latter would have provided the natives with an opportunity. This undoubtedly provided part of the context for Septimius Severus’ decision to regain the initiative in the north in 208, some fifteen years after coming to power. By then the northern tribes had grouped themselves into new confederacies, the Maeatae living close to the Wall, and the Caledonii beyond, but it is likely that the origin of these power groups dates back to the time of Commodus or Aurelius. A more united opposition in the north required a more forceful response.

II. URBAN DEVELOPMENT

At the beginning of the period of our survey the cities of Britain were recovering from the aftermath of the Boudican rebellion. The three certainly known to have been destroyed, Colchester, London and Verulamium, have each produced evidence of slow recovery with delays in their redevelopment of a decade or more.18 Equally, the extent of city life at the beginning of the Flavian period was limited to the south-east, the area which, before the Boudican rebellion, had effectively become demilitarized. This included the kingdom of Cogidubnus, whose boundaries are uncertain. It appears to have been centred at Chichester and probably extended north and west to embrace the territories which, on the king’s

Death in the late 70s or 80s, became the civitates of the Atrebates and the Belgae. The civitas of the Cantiaci may also have formed part of his domain.1556

The decision to press the conquest forward from 71 inevitably led to the redisposition of forces and the cessation of military control over extensive tracts of the island. The movement of the legions is perhaps the key to understanding the potential for reorganization in this period. Legio II Augusta had certainly moved to Caerleon by about 75, thus releasing the south-western peninsula from military control and allowing for the foundation of Isca as civitas capital of the Dumnonii at Exeter on the site of the abandoned fortress. Whether the foundation of Durnovaria (Dorchester) belongs to this period, or a little earlier, is not clear; recent excavations suggest the latter.1557

Civil government of the civitas Dobunnorum was based on Cirencester (Corinium) whose street grid and major public buildings appear to be of Flavian origin. A similar date can be assigned to the establishment of the street grid at Leicester (Ratae), even if the provision of public buildings such as the forum-basilica and public baths belonged to the second century. A similar sequence of slow development has been observed at Venta Icenorum (Caistor-by-Norwich). This gradual process of city development is not confined to those which were either physically destroyed or politically implicated in the Boudican rebellion; Durovernum (Canterbury) has revealed little evidence of major public building projects before the closing years of the first century. This conservatism is also reflected across the Thames in the territory of the Trinovantes where, if the original intention had been to make Caesaromagus its civitas capital, that role effectively seems to have been ceded to the colonia at Colchester (Camulodunum). None of the buildings that might be expected of a tribal capital have so far come to light at Chelmsford.1558

The physical evidence from the south as a whole does not wholly conform with the model for the development of the cities of Roman Britain in the Flavian period which takes as its lead the well-known Tacitean passage where Agricola is said to have encouraged the provincials to build ‘templa, fora, domos’.1559 Presumably, as confidence gradually attached to the success of the conquest of Wales and the north up until the military retrenchment after the mid-8os, a corresponding enthusiasm grew for investment in the civilian infrastructure. The withdrawal of legio II Adiutrix and the gradual pullback from Scotland may have shaken confidence. Thus, although there are some strong cases for the associations of buildings with the Flavian dynasty (as at Cirencester, Exeter and Verulamium), we should be cautious about the overall speed of urbanization in this period. Even where the case for Flavian development has been strong, as at Calleva (Silchester), further investigation shows a more complicated pattern. While the core of the street grid, the amphitheatre and the forum-basilica are certainly of Flavian date, the latter was a timber construction, only replaced in masonry in the early to mid-second century.23

Changes of policy towards the northern frontier may explain the decision to found two new coloniae at Gloucester (Glevum) and Lincoln (Lindum) following evacuation by their respective garrisons. Providing the prospect of additional support for the civil zone, these were in place by the reign of Nerva. Just as earlier with Colchester, there was extensive utilization of the existing fabric, such as theprincipia and barrack blocks, of the legionary fortress, indicating economies of expenditure without parallel in the foundation of continental coloniae?4

The extension of urbanization continued into the second century with the establishment of new civitates closer to the military zone. Although the military situation allowed for the foundation of Caerwent (Venta Silurum), Carmarthen (Moridunum), Wroxeter (Viroconium), Brough (Petuaria) and Aldborough (Isurium) by the end of the first century, there is little evidence for this happening so early.25 Although, as previously with Agricola, there is a temptation to associate these initiatives with Hadrian’s visit to Britain, largely on the basis of the forum inscription at Wroxeter which dates to his principate, the archaeological evidence does not normally allow us to make precise associations.26 However, recent work at Caerwent supports a date in the second quarter of the second century for the construction of the forum basilica there.27 The setting up of new civitates was patchy; it is by no means clear, for example, why the Demetae in the west of Wales were favoured in addition to the Silures and not, for example, the extreme southwest beyond Exeter which was neglected. The Brigantian territory is believed to have extended over much of the north, but was served at the outset by only one centre at Aldborough, situated like Caerwent close to a legionary base, in this case York. The north-west, whether Brigantian or not, was not the subject of separate provision unless or until Carlisle became the centre for the Carvetii somewhat later.28 What determined the policy behind the establishment of the civitates is unclear; how much was attributable to local rather than central initiatives is not certain. In any case final decisions rested with the imperial authorities. The epigraphic evidence as a whole is not particularly helpful; there is little of it, and, as with the case of Wroxeter, there is general anonymity on the question of responsibility

For the building programme(s). This suggests that individual munificence was rare, but whether costs could have been covered by the resources of the civitas without remission of taxation has not been established. If the subsequent development of villas gives us some index of regional prosperity, it would seem that the peripheral civitates would have had much less available surplus than the south and east; here, at least, governmental assistance is indicated.

While much has been discovered about the overall history of the cities of the province and their public buildings, very little is known about the urban population and the governing elite. Although the more complete plans of, say, Verulamium, Silchester and Caerwent show the widespread existence of town houses, it is now clear that many of these structures only emerged from the late second or third century onwards. Only in the colon-iae and at Verulamium do we have much evidence for the origins of this development. In the former, at Colchester up to Boudica, and Gloucester, we find the continued use of barrack accommodation adapted for civilian use, until by about the early to mid-second century evidence of the first purpose-built shops-cum-workshops and town houses emerges. At Verulamium small, rectangular, timber-framed town houses, occasionally appointed with mosaics and decorated wall plaster, emerged in the course of the first half of the second century. A disastrous fire in about 15 5 seems to have precipitated a gradual programme of rebuilding in masonry thereafter.1560 The first-century predecessors of the town house at Verulamium and elsewhere are ill-understood, thus reinforcing both our lack of knowledge of the origins of the urban aristocracy, and the impression that investment in the towns was directed first towards the civic fabric, rather than urban housing. In London, where a different pattern of development might in any case have been anticipated on the basis of the different role it played in the province, the earliest housing seems to have consisted of series and sequences of simple, rectangular, narrow-fronted, timber-framed buildings, some of which were decorated with mosaics and wall plaster. Evidence for Mediterranean-style atrium-type housing of the kind that eventually emerged in Gloucester is lacking in first - and second-century London.1561

There are few indications of what activities took place in towns, but the excavations at Verulamium in Insula XIV alongside Watling Street show a thriving commercial quarter where shops-cum-workshops fronting the street were redeveloped over successive generations until the disastrous mid-second-century fire.31 Other towns such as Silchester show comparable patterns of development influenced by the main through-roads/ streets of the town. London, in particular, with the recent publication of excavations in the west of the city, exemplifies this pattern. Off the principal thoroughfares the evidence is patchy, but again London and Verulamium point to the existence of artisanal activity — properties with hearths and ovens — in these quarters, prior to the construction of the first town houses. The picture is one of some commercial activity and vigour gradually giving way from the mid - to late second century onwards in the central blocks of insulae of towns like Verulamium and Colchester to a pattern of more residential occupation with commercial activities focused on the main thoroughfares and around the centre of the town.32

In London the evidence for the decline of the commercial/artisanal aspects of the city is more marked, and the transition to the residential pattern is less clear. Indeed it has been suggested that the second half of the second century represents the beginning of a marked decline in the fortunes of the city.33 To some extent there may be a point of comparison with neighbouring cities like Verulamium, inviting the question as to how far the economic life of the cities was locally based, and how far it rested on the wider provincial or imperial economy. This theme will be considered further below.

So far little has been said of an urban dynamic; a state of affairs which correlates closely with the lack of detailed urban investigation. We have seen that, even with the provision of public buildings, progress was slow and by no means uniform. Yet there is one area where there is considerable uniformity, and that is defence. Only the coloniae, reusing or refurbishing their existing fortress defences in the case of Gloucester and Lincoln, or with a new, purpose-built wall at Colchester, were defended with stone walls before the second century. However, towards the end of the second century almost all other towns of civitas rank and above were protected with earthen ramparts and ditches. Although the archaeological evidence cannot allow precision as to exactly when these defences were added — and, given that the dating rests on a terminuspostquem everywhere, it is possible that there could be variation of between two or three decades — the consistency in the manner of their construction is striking. Furthermore, the phenomenon is not just confined to the major towns; most of the lesser towns on the principal road networks were similarly provided. Altogether it looks as if this programme was the result of a common policy, possibly engendered by apprehensions over the northern wars of Commodus’reign, or by the departure of Clodius Albinus and the army in 196. Hurried though the defences appear to be, there does not seem to have been a need to enclose a reduced area; as far as was practical entire street grids were embraced and only ribbon developments along the principal roads in and out of the towns were excluded.34

32 Frere (1972) and (1983); Crummy (1984), (1992).  33 Perring (1991) 76—89; Milne (1995) 71—5

34 Burnham and Wacher (1990); Maloney and Hobley (1983); Frere (1984); Fulford and Startin (1984); Fulford (1984a).

The urban analysis has so far been almost exclusively confined to the chartered towns and civitas capitals, but other centres also emerged from the late Iron Age onwards. These are known collectively as the ‘small towns', even if very little is known about their urban character. The more important of these were those which developed on the road network, which, with its focus on London, predominantly post-dates c. a. d. 55. Such towns, as we have seen, were defended at the end of the second century. A major consideration in the development of these sites will have been the provision and maintenance of mansiones and other establishments necessary for the cursuspublicus. Excavation at Chelmsford and elsewhere suggests that this provision belongs to the later first century onwards. Some sites on the principal road network, such as Bath (Aquae Sulis), developed for other reasons as well (for the development of the spa, see below). Off the road network, development depended on special factors, such as we find at Charterhouse-on-Mendip (lead-mining centre). A characteristic of all these towns is the lack of town planning and public buildings. Development was focused on the principal thoroughfare and tended to be linear, but side lanes emerged towards the centre of the larger examples such as Water Newton (Durobrivae), Cambridgeshire.1562

III. RURAL DEVELOPMENT

That lack of clarity about native elites is further borne out by the countryside where the evidence for romanizing in the form of villa developments is unevenly distributed spatially and chronologically. Just as with the town houses, the emphasis in our understanding lies in the third and fourth centuries and a record largely derived from early investigations. Comparatively few extensive excavations have taken place where attention has been paid both to the potential of the stratigraphy and the possibility of discovering timber-framed buildings or other activities beneath masonry successors.

The patchy nature of the evidence for the first and second centuries is partly a reflection of where fieldwork has been carried out. One area where there has been a considerable investment of effort in recent decades is the countryside around Verulamium, where we can begin to identify a pattern of villa development from the late first or early second century, sometimes with evidence of late Iron Age predecessors. The latter is the case with Gorhambury on the outskirts of the city, whereas at Boxmoor and Gadebridge Park to the north the evidence shows the gradual development of a villa from the late first century onwards. In the case of the latter, the provision of a bath-house, probably accompanying a timber-framed house, was followed by the setting up of a masonry villa in the second century, subsequently aggrandized in the Antonine period with projecting wings. What evolved here is comparable to the sequence at, and the scale of, Gorhambury.1563

More elaborate villas are to be found elsewhere in the south-east, but they are exceptional. Among these may be included the ‘palace’ at Fishbourne, less than 2 kilometres west of the civitas capital of Chichester. About A. D. 75 a very large courtyard building closely related in style and grandeur to counterparts in Italy was built. It replaced a Neronian building, possibly a baths building, which was also elaborate. The architecture was sophisticated and the building was richly decorated with mosaics and painted wall plaster and the central courtyard was graced with an ornamental garden. Although this residence compares well with the larger country houses at the centre of the empire, a number of other, outstanding villas of similar date and also substantially ‘built as one’, have been identified at locations such as Rivenhall in Essex, Eccles in Kent and Angmering in Sussex.1564 While these can be seen as the product of major capital investment from the outset, the reverse is true of the Verulamium group and others like it which gradually developed over a period of time.

With one or two exceptions the area where we can identify early villa development is in the south-east. Elsewhere the aristocracy, whether native or immigrant and presumed to be associated with the urban developments of the Flavian period and the second century, is much less visible architecturally in the countryside. The native farmstead at Whitton, South Glamorgan, and probably situated in the territory of the Silures, had taken on a more romanized appearance by the late second century with rectangular buildings replacing round houses, but there was no sophisticated architecture or internal decoration. A similar situation prevails among the northern civitates where, at sites like Dalton Parlours or Rudston, there is little evidence of romanization before the third century. Although it is difficult to draw the boundaries between these peripheral civitates and the military, and thereby non-urbanized, zone, there is certainly no clear distinction in terms of rural romanization to be made between them. In these areas the priority was investment in the establishment of the urban (albeit modest) centres.1565

In parallel with the gradual emergence of villas in the countryside is a corresponding development of rural shrines. The characteristic ground-plan of these buildings comprises a double square: a central cella with enclosing ambulatory. Although there is some evidence for parallels for rectangular structures as cult buildings in the mid - to late Iron Age, as at

Heathrow and the hillfort at Danebury, their plans do not coincide so closely with those of the Roman temples which emerge from the later first century onwards; indeed, there is comparatively little evidence of them before the late first century. As with the villas, the earliest examples of temples are recorded from the south-eastern counties. In two cases, at Weycock Hill, Berkshire, and Wanborough, Surrey, the origin of temple buildings seems to post-date the deposition of ‘temple’ treasures, which consist of hoards of, for the most part, Iron Age coins. The association of some Roman coins suggests that these collections were not deposited until at least the mid-first century a. d. Where there is evidence of masonry succeeding timber structures, as at Harlow, Essex and Hayling Island, Hampshire, the latter appear to be of latest Iron Age date. Beyond the south-eastern core the evidence for a late Iron Age origin is not so clear or consistent, and, as at Uley, Gloucestershire, or Lowbury Hill, Oxfordshire, the date of temple buildings cannot be placed any earlier than the latest first century or first half of the second century a. d. With the lack of firm evidence for a history of development from native forms, we have to consider whether these rural shrines represent a strategy of legitimation of ownership on the part of new landowners amassing country estates in the securer parts of Britain.1566

This difficulty in identifying a romanizing native elite over large areas of the province before the late second or third century must raise doubts about the scope of its involvement, whether political or financial, in the early romanization of the province outside the south-east. Such opacity is nowhere more evident than at Bath (Aquae Sulis) where the Neronian— early Flavian period saw the construction of a fine classical-style temple adjacent to a large bathing complex built to exploit the hot mineral springs. The identity of all the early visitors to the spa is hard to seek, but dedications on altars and tombstones from the city provide a valuable source. Predominantly, these record soldiers and those associated with the cult; the wider family is scarcely in evidence.1567 In the scale of its investment the spa complex at Bath rivals that at Fishbourne, or, to a lesser extent, the other de novo foundations in the countryside, while in the towns, the comparison is with the provision of public buildings such as thermae and fora basilicae.

At the level of the individual the contrast between the evident affluence in material terms between those living in southern towns, such as Colchester, London and Verulamium, and those in the countryside is striking. Even in the south-east, villa development is relatively rare before the later second century, and there is a large measure of continuity in settlement patterns and structures, as well as material culture, from the late preRoman Iron Age. A recent survey of the results of rescue excavation on the river gravels of southern England, and particulary on the Upper Thames, revealed settlement continuity from the middle Iron Age up to the late first/early second century a. d. The tendency thereafter was for only those settlements founded at the very end of the Iron Age to continue on and develop into small romanized farms or villas in the third and fourth centuries, as at Barton Court in Oxfordshire.1568 It is not easy to explain settlement abandonment unless it relates to a reordering of the countryside as villa estates began to grow, or to a drift of settlement into the towns. The former may account for the development of rural nucleated settlements such as Chisenbury Warren in Wiltshire which developed rapidly from some point between the mid-first and mid - to late second century a. d., perhaps as a planned estate village, and in a landscape where otherwise there seems to have been a major break between those settlements which were occupied from the early to the mid - to late Iron Age and those which developed in the first or second century a. d. Catsgore in Somerset, where there is evidence of a romanized, nucleated settlement from the early second century a. d., offers another possible example of an ‘estate’ village. Stanwick on the River Nene in Northamptonshire may be similar, but here there is also evidence of a later Iron Age predecessor.1569

As can be seen from the foregoing account, much of our understanding of the countryside is derived from single-site excavations, often on villas before the advent of stratigraphic and area excavation; there has been very little effort to understand the landscape as a whole and the full range of the settlement hierarchy. This is an urgent necessity if we are to understand better the relationship between the emergence of the romanized landscape and the patterns of desertion, adaptation and reorientation of the non-romanized countryside.

Iv. economy

Although mineral exploitation played an important role in the economic life of the province, with major development of the iron (the Weald and Forest of Dean) and lead (Mendips, north Wales, Pennines) extraction and processing industries, with a lesser emphasis on gold (south-west Wales), tin (Cornwall) and copper (Anglesey), agriculture remained the pivot of all economic activity.43 The fecundity of the island may well have been an influence in the original decision to invade, but it is not clear how far production of a surplus could keep pace with the demands placed upon it by the army and the developing civilian infrastructure of the province.

Theoretically the potential was there, but there is no evidence that society was capable of delivering it from the start, or that the infrastructure was in place to manage distribution. On the contrary, there is evidence for the importation of basic requirements such as grain in the first and second centuries.44

There can be little doubt that the principal areas of consumption of resources in Britain in the first and second centuries were the military with its fortresses, forts and linear frontier schemes, the programme of town-and road-building, and the support of the communities engaged in metal extraction and production. Whereas the campaigning of the Flavian period, with all its subsequent provision of fortresses and forts, was undoubtedly a costly affair, substantial investment in the frontier systems continued through the second century. While much of the work, as well as the procurement of certain raw materials such as lead, or iron from the Weald, were undertaken at the outset by the army and therefore represented no extra labour cost, the supply of other materials and the feeding of the troops involved expenditure outside the closed military circle. Although many of the basic requirements could probably have been met from within Britain, it is clear from the evidence of amphorae and other imported pottery, that a wide range of commodities flowed into the island from the Mediterranean, Gaul and Spain in the first and second centuries. The economic irrationality of these trade or supply systems, such as the numerical superiority in Britain of the more remotely produced central Gaulish sigillata as opposed to the eastern Gaulish types, or of Baetican as opposed to Gaulish olive oil, has been observed before, and similar patterns can also be found with goods of British origin within Britain itself.45

This is most striking in the second century when the northern frontier can be shown to be at the end of supply routes, identifiable by ceramic tracers, whose origins lie in the south of England as well as the midlands (e. g. Dorset black-burnished category i, Thames Estuary black-burnished category 2, Colchester, Mancetter and the Nene Valley). More local sources do not appear to have been much utilized. Although the distributions of the various types of pottery have this northwards distortion, civilian markets were served as well; but it is clear that the military was their most powerful magnet and was largely responsible for the coastwards location of the major industries. This seems to imply that in the balance between the development of the civilian infrastructure and the maintenance of the garrison, the emphasis was in favour of the military. Only in the second half of the second century do specialist industries develop, such as the Oxfordshire or Alice Holt pottery industries, whose market seems to be entirely confined to the civilian sphere; but in this case the distributions are

Quite narrowly circumscribed. One implication which can be drawn from the developments of the second century when, for ceramic supplies, the northern frontier appears to draw on sources from as far away as the south of the island, is that this represents a situation where the whole of the urbanized area of the island is exploited for the support of the army and frontiers, and that this represents a real change from previous arrangements.46

In the first century we have greater reliance on imported goods — graphs for samian supply from London and elsewhere show the peaking of south Gaulish sigillata in the latter part of the century. The internal evidence shows the importance of London as the principal node in a centralized supply organization. Brockley Hill pottery (between London and Verulam-ium) can be shown to be widely distributed to the northern frontier from the Flavian periods onwards; it points to a road-based supply from London, where goods were collected for transit to the north and north-west.47 In the second century, however, marine-based supply systems utilizing east - and west-coast routes, became more important. Given its involvement in the construction of Hadrian’s Wall and the exploitation of Wealden iron, it is likely that the Classis Britannica played an important role in the supply of the northern frontier.48

By the end of the second century the military and civilian structures of Roman Britain were firmly established. After the decision to abandon the Antonine Wall, the frontier arrangements in the north remained essentially unchanged. Equally, although there may have been minor adjustments to the organization of the civitates, there was no significant expansion to their number, and, correspondingly, no new towns were founded after the first half of the second century. Emphasis on the development of the public aspects of towns and on infrastructure such as the road network in the second half of the first and early second century was followed by evidence for expansion of conspicuous consumption at the level of the individual household. The town house and the countryside are the principal areas to show substantial development later, but even here the framework for that expansion was securely in place by the elevation of Septimius Severus.