The Romano-British period

An Assessment

Simon Esmonde Cleary

Introduction

Like the Roman god Janus, the West Midlands region faces both ways in Roman Britain. It straddles the two archaeological provinces, the south and east with its archaeology of towns, villas, temples, burials, multitudinous artefacts, and the north and west with its forts, vici and ‘native’ settlements, and a low level of artefacts (Fig 4.1). As such, it is in many ways a microcosm of the province of Britannia as a whole. Therefore, how we interpret the Roman West Midlands will need to draw on a wide range of approaches to the archaeology of the period, but if some sort of synthesis is achieved it may have wider relevance than just this region.

A ‘traditional’ resource assessment would in many ways read like a scaled-up excavation report, with sections on settlements and structures, various types of artefacts – coins and pottery first for their dating value, then metal objects ordered by value and aesthetic qualities, then stonework and finally the environmental evidence – reflecting the history of development of the specialisms and all too often their implicit ranking. There would also be discursive sections on relevant topics such as military aspects, towns and urbanism, villas and rural settlement, ritual and religion. Here, rather different approaches will be proposed and a different structure developed in order to try to put in place an assessment that both develops out of the regional database and also responds to current approaches to the archaeology of the Roman period. The need for this became clear from the seminars held as part of the process for producing the frameworks, which showed that the manifestations of the Roman period in the region were much more varied and complex than the ‘old’ categorisations allow for.

The first important approach is to realise that it is crucial to respond to the nature of the region and its database rather than seek to force it into schemes devised in other areas. The region does have one feature which largely unifies it and sets it apart from many other areas: it is resource-rich, not just in agricultural terms but also in minerals; the implications of this need to be thought through as we try to understand its Roman-period manifestations. Secondly, we are now in a position to understand that the discrepancies in the visibility of the Roman period across the landscape of the West Midlands are the results of conscious and unconscious choices made in the past more than of the vagaries of modern recovery. Thus, any new attempt at a resource assessment must embody this variability at its core. Thirdly, the database is increasing not only in quantity but in range, for instance in the results of the Portable Antiquities Scheme for our understanding of artefact distribution across the region (cf Worrell 2005), or in the growing importance of the data from environmental and other specialisms.

Map showing distribution of Roman sites in the area
Fig 4.1 Site distribution map of the West Midlands in the Roman period

Accordingly, it is proposed here that there are four major features of the Roman-period archaeology of the West Midlands which can provide the framework for a different sort of resource assessment, an assessment which will not only seek to elucidate our existing knowledge of the archaeology of the region, but will also help set agendas and provide frameworks for the programming, execution and study of future work on the archaeology of this period. The four over-arching themes of this resource assessment are:

  • Resource Mobilisation.
  • Assessing the Evidence.
  • Assessing the Gaps.
  • Tradition and Innovation.

It is to be hoped that the relevance of each of these is clear: the first responds to a defining feature of the region’s Roman-period archaeology; the second to the existing and possible future strengths in the database; the third attacks a major problem both in the archaeology of the region for this period and in how we assess the evidence we do have; the fourth seeks to address these major differences while at the same time having more regard to chronological structure. One feature of all four is that they are designed to be essentially integrative, pulling together techniques and analyses from a variety of different materials and approaches not only to benefit from their insights but also to try to counteract any tendency to compartmentalisation. Thus within each of the three major themes are the various minor themes to articulate them. It is hoped that these themes will be of use to curators, contractors and consultants in the field, in order to justify, structure and deliver individual projects, so the themes do need to be ones which they find credible and practicable. This is the more so as most of them are not Romanists by training (still less inclination).

Resource mobilisation

One of the defining features of the archaeology of Roman Britain is the extent to which natural and human resources were exploited, especially compared with the preceding and succeeding periods. A result of this is the vast amount of physical evidence for the Roman period: settlements, buildings, artefacts. But over and above this it can be argued that the West Midlands and surrounding areas were a zone where the extraction, processing and movement of natural resources, both mineral and agrarian, were of particular importance.

Minerals

The West Midlands contain a variety of mineral resources. The most famous and unusual are the brine springs at Droitwich (cf Woodiwiss 1992). These were already a well- established resource by later prehistory, but excavation has shown a very considerable investment in extraction during the Roman period. Some sense of the volume and extent of the distribution of this salt in the earlier Roman period is given by the briquetage, though this still needs to be fully synthesised and published to give an impression of the range and scale of contact. Use of briquetage seems to be essentially a phenomenon of the first half of the Roman period, wooden containers perhaps taking over thereafter. A range of metals were systematically exploited. Argentiferous lead was won around Shelve in western Shropshire (cf Jones and Mattingly 1990, 186-8), though judging by practice elsewhere it was the silver content that yielded most value with the lead as a useful by-product. Iron was clearly a major resource. In the south-west of the region lies Ariconium/Weston-under-Penyard, apparently one of the principal centres of the huge iron-mining and -smelting area which extended south into the Forest of Dean. We understand painfully little about the mechanics, organisation and chronology of this industry: in itself it could form an important research objective. Iron-smelting was also a major concern in Worcester where there is evidence for activity on a large scale (Jackson 2004, 100-05). However, there is also the evidence for smaller-scale bloomeries at rural sites in the area; how are the two related? The Worcester iron centre also provokes questions about the source of the ore (whether the iron-fields in areas such as Staffordshire were exploited in the Roman period remains an open question) and about the destination of the product. The siting of Worcester suggests the use of the Severn, so perhaps the end-users of the iron were outside the region.

Mineral resources are an area where the modern administrative boundaries serve us ill. It has already been pointed out that Ariconium is only part of a much larger zone extending south out of the region. Just over the northern part of the boundary with Wales lies Llanymynech, probably an important source of copper. North of the region is the area around the legionary fortress at Chester, including sites such as the lead- workings at Pentre Farm, Halkyn and the salt-workings at Middlewich, Nantwich and Northwich. To what extent does the northern part of the West Midlands region lie in the resource-procurement zone for Chester? North-east of this region is of course the major lead-mining area of the Peak District. Adding these sites to the West Midlands sensu stricto gives one of, if not the, largest regions of mineral exploitation in Roman Britain. It is also worth noting the amount of stone quarrying which would have gone on in the region to provide for all the stone-built structures of the period. Much of this would have been small scale and transitory, called into being by the needs of individual projects and then extinguished. However, the requirements of major centres, Wroxeter above all, probably engendered longer-term complexes to exploit the various stones used in the public and private buildings. The south-eastern part of the region, lying near the good freestone of the Cotswolds, would also have been an area of quarrying.

Our evidence for the manufacture of objects in the region is relatively limited. The overwhelming bulk of the evidence is for pottery manufacture (cf Swan 1984), though even for this the evidence is not that great when compared with other regions. The best-investigated industry lies just inside the eastern border of the region at Mancetter- Hartshill, an industry particularly well-known for its widely-distributed mortaria and parchment wares, and operating from the turn of the 1st and 2nd centuries through to the end of the Roman period (cf Tyers 1996, 123-25). The other significant production was of the wares grouped under the appellation ‘Severn Valley’, originating south of our region and characterised principally by oxidised storage and drinking vessels (Timby 1990 for origins; cf Tyers 1996, 197-99). Unlike the Mancetter-Hartshill industry, this was not a tightly-grouped production centre but rather a series of smaller and larger- scale centres scattered across the region and associated with other productions such as the Malvernian kilns. A few other centres are known, such as the 2nd-century mortaria made at Wroxeter, whose distribution is markedly military and almost entirely outside our region. Otherwise, penetration of the region by wares from other production centres is low, apart from the south-eastern zone in the Cotswold/Avon area. Reasons for the low take-up of pottery production and products over most of the region will be discussed below.

Pottery apart, there are isolated instances of manufacture such as brooches and glass at Wroxeter (Houghton and Ellis 2006) or the two-piece moulds for casting bronze fittings from Rocester (Ferris in Esmonde Cleary and Ferris 1996). Of course, wood-working and textiles would have far outstripped any other manufacture in the region both in terms of labour and of output; unfortunately almost none of the products survive, so we have no means of gauging their overall scale nor their impact on other areas of the economy such as agriculture (woodland management, sheep-rearing, dyestuffs).

The organisation of mining, quarrying and manufacture reflects types and levels of social organisation, so seeking to examine the organisation of these industries in the region is not just of interest in understanding them, it is also of wider interest in understanding the nature of the region’s society. Traditionally, the winning of metals, particularly argentiferous lead, in Roman Britain has been seen as one of two systems. One was top-down, essentially created and staffed by the army in direct exploitation. The other sees smaller-scale operations by civilian lessees (franchise-holders) working the ores for the silver and lead on the state’s behalf and making their percentage on the deal. A related debate has been in progress over the structure of pottery production, especially the large ‘industries’. Were these organised top-down by entrepreneurs who articulated a number of specialists into a form of production line, a model drawn from the evidence for something of this sort in the Gaulish samian production centres? Or were they structured essentially by production site, and the ‘industry’ is the net aggregate of these individual enterprises?

In turn this relates to the often imprecise terms ‘craft’ and ‘industry’. ‘Craft’ is seen as something essentially small scale (eg household) and low-tech, whereas ‘industry’ is centrally organised and profit-driven. However, is there such a neat division in pre- industrial societies? Is the activity long-term and year-round or, as has again been suggested for pottery production, sporadic and/or seasonal? Environmental evidence could be decisive in answering such a question. So it may be that attempting to unravel the structures, systems and cycles of the exploitation of mineral resources in the region could lead onto important insights into the structures of the society which undertook this work.

Another feature of these industries seems to be that across the Roman period there is little in the way of technological or other innovation (unlike, of course, later periods), reflecting one of the major perceived characteristics and weaknesses of the ancient world (cf Greene 1994). Innovation, or the lack of it, is a reflection of complex causalities. However, one regularly identified is the social matrix; are societies ‘conservative’ or ‘traditionalist’ and thus inimical to change, or do they allow for the dissent and experimentation that promote change? Can we approach closely enough the social structure and social ideology of the Romano-British West Midlands to be able to tackle these big questions?

Agriculture

It is not possible to tackle here the whole field of arable and pastoral agriculture in the Roman West Midlands. Instead, one can raise two questions which may be worthy of further research from a variety of angles. The first is a resource question. One of the most fundamental resources in the period was timber/wood: for construction, transport, furniture, weapons, tools, fuel; its availability was essential for all other activities. Additionally of course, it is not just a question of any old wood. Different purposes required different types of wood of different scantlings and ages, and these did not just become available without thought. Woodland and wood were a resource that had to be managed and planned for in the long term.

Our general picture of the southern half of Britain in the Roman period is of a land already largely deforested, indeed the north was also undergoing continuing deforestation. The English Heritage Environmental review notes that the south and east in particular may have carried even less woodland than they did in modern times. On the other hand, it also notes that one of the few coleoptera spectrums to indicate the presence of woodland is that from Stourport, and also that some of the wood from late Roman and early medieval water-logged contexts in the region suggest growth starting within the Roman period. This is in no way a basis for firm conclusions, but it does at least suggest an avenue for future research. Was the West Midlands an area where timber was a major and managed resource, some of which was distributed outside the region to timber-poor areas? Clearly, a range of environmental evidence and techniques could address this problem as opportunity presented itself. Such a picture would of course accord well with the thesis developed above for the region as a net exporter of other raw materials. Woodland, of course, might help explain away some of the gaps in the settlement distribution maps: there really was no one there. The second feature of a comparison of the West Midlands with areas further to the east and south is the apparent low take-up of introduced plant and tree species in this region. The picture overall seems to be one of a mixed agrarian regime, concentrating on the staple crop of spelt wheat and the standard livestock of cattle, sheep and pigs.

One way of reading this might again be in the light of the region as a resource-procurement zone. These are precisely the resources needed by the Roman army, certainly in the first two centuries or so of the province, maybe also later, to feed, clothe and equip its troops. With the forts of Wales to the west and Chester and the garrisons of the north-west, were the requirements of military supply constraining the agriculture of the region? They are the recipients of much of the output of the Wroxeter mortaria production (see above) and, as so often, pottery may be standing as a proxy for the transport of bulk supplies such as grain. Indeed, much of the West Midlands might be seen as the southern end of a military-dominated super-region encompassing the north-west all the way up to Hadrian’s Wall. The long-distance distribution of Severn valley ware as far as Hadrian’s Wall could be, as with other industries such as that at Colchester (Peacock 1982, 102), an archaeologically-visible proxy for long-distance movement of bulk supplies. If substantial value was being removed from the region, this may be part of the explanation as to why the more costly forms of investment in Roman-style buildings and objects was limited. Or, to reprise another theme, are we seeing the effects of a ‘conservative’ or ‘traditionalist’ social organisation and ideology, which had little or no place for widespread integration of these novelties into the existing regime?

Military involvement

Military archaeology was a major focus of investigation in the region in the 1960s and 1970s under the leadership of Graham Webster. The central concern then was the identification of military sites of the conquest and consolidation (usually through aerial photography cf Frere and St Joseph 1983, Pt.2); these were then trenched to recover dating material in order to assimilate the site to a narrative of the conquest structured by individual governorships (cf Webster 1981). As a result, the region is now a locus classicus for the study of the temporary fortification or ‘marching camp’ (Welfare and Swan 1995). However, the wider enterprise fell into disfavour because of increasing reservations over the idea of using the patchy historical sources to constrain the archaeology, and because of increasing doubts over the possibility of coins and samian being able to yield the sufficiently precise dates required.

Mention of the possibility that the resource mobilisation of and from the region is influenced by military dictates raises the question of whether there is evidence for this. There is a small but significant group of sites and material which may provide a partial answer. The Bays Meadow villa at Droitwich is unique in Britain for its defences. This has long been hypothesised as some sort of official establishment for overseeing the saltworks, in which case these would be some form of imperial monopoly or resource. It has more recently been proposed that the establishment was a mansio of the cursus publicus (Hurst 2006, 239-41). In a way, this does not affect the argument, since evidence from the wider empire suggests that mansiones also acted as sites for resource extraction. So what of other mansiones in the region such as Tripontium and Wall? Is it significant that the latter were also the sites in the late empire of two of the burgi along Watling Street, which have also been proposed as sites for the extraction and/or protection of matériel in transit (cf Gould 1999)? For the same time-frame, there is the large granary or store-building at Coulter’s Garage/Gateway Supermarket, Alcester also suggested to be in some way official (Cracknell 1996, 4-41). This was demolished to make way for the town defences; such ‘small’ town defences may also have been to do with safeguarding installations, personnel or matériel of interest to the late Roman state.

Many towns and other sites in the region have yielded military equipment and fittings of the early or late empire. Traditionally the early ones have been explained away as residual from underlying conquest-period forts. More recently, both early and late material has been accounted for by soldiers in transit. But the surviving documentation from the Roman army (including the Vindolanda tablets) shows clearly that troops were regularly posted to ‘civilian’ areas in order to secure supplies and commissariat, both under the early empire and with the late empire’s increasingly oppressive methods of securing its own existence. In general, a case can be made for the mobilisation of resources being a, perhaps the, defining feature of the archaeology of the Roman period in the West Midlands. It may also be that many of the resources were then exported from the region, in particular to supply and service the long-term military garrisons to the west and north of the region.

Assessing the evidence

It was noted above that the traditional site categories (fort, town, villa…) have served us well enough as a means of ordering emerging data; to continue to depend on them risks perpetuating their epistemological weakness, which is that they impose a divisive rather than an integrative approach, more concerned with cataloguing than with articulating the classes of information. Our understanding of the region should benefit from structuring the evidence according to new themes, ones which necessarily involve combining evidence from more than one of the traditional categories. This would also have the advantage that such approaches can be applied to the existing database as well as contributing to the analysis of future work both on-site and off-site. Here it is proposed that the major themes should be Identity and Community and Ritual and Religion, and in each case the relevance of these will be illustrated through reference to examples from the Roman-period archaeology of the West Midlands.

Identity and community

Identity has become a central concern of archaeologists of the Roman period, subsuming and building on previous concerns with ‘Romanisation’ (cf Millett 1990). This sought to explain the visible changes in the archaeological record consequent upon assimilation into the Roman empire in terms of the degree of adaptation to Roman ways by indigenous peoples, and was long the implicit agenda in the West Midlands. It has been criticised as a linear and teleological view of the past, and it is now preferred to regard ‘Roman- ness’ as one of a suite of possible ways in which people constructed and expressed their identities, alongside other factors such as age, gender, status, and membership of communities. Identity of course works at various levels: the individual; the family; the kin or ‘tribe’; the functional community; the ‘people’, though all these, of course, overlap and an individual can have multiple identities depending on context – ‘situational identity’. All can be expressed through a variety of different types of material evidence. This therefore gives us the opportunity to approach the ‘visible’ archaeology of the south and east of the region through a new series of themes, each of which allows different types of evidence to be exploited. Here a series of ‘levels’ will be outlined, ranging from the general of ethnic and cultural identity down to the particular of the construction of the individual’s identity.

Ethnicity and cultural identity

The identification of ‘incomer’ and ‘indigenous’ has been one of the central questions of the study of Roman Britain, particularly at its beginning and its end. Likewise, the cultural changes wrought by these incomers and their cultural impact are central to the study of the period. For the earlier part of the Roman period, it is cultural rather than ethnic identity which is the central concern; ethnic identity is usually seen as bound up with the army and its largely Mediterranean-derived legionaries and provincial auxiliaries. Though they would have had a considerable short-term impact on the region, subsequent changes in the archaeological record are investigated through changes in cultural frameworks rather than ethnic identities. At the other end of the Roman period, of course, the emphasis is reversed; the question of ethnicity is central, with identification of ‘Germanic’ or ‘British’ ethnicity long seen as the key, with cultural markers an epiphenomenon of those ethnicities.

In the West Midlands it is the south-eastern zone which exhibits most clearly the impact of Roman-style cultural markers, through relatively well-known ‘small towns’ such as Alcester (cf Mahany 1994; Cracknell and Mahany 1994; Cracknell 1996; Booth and Evans 2002), Droitwich (Woodiwiss 1992; Hurst 2006), Kenchester (Wilmott 1980), Rocester (Esmonde Cleary and Ferris 1996), Wall (Gould 1964; Round 1992) or Whitchurch (Jones and Webster 1969). Other such sites are less well understood, such as Blackwardine (Brown 1990) or Chesterton-on-Fosse (cf Booth 1996, 37). Problems of definition between ‘urban’ and ‘rural’ are exemplified by sites such as Tiddington (Palmer 1982; Biddulph 2005). It is the same south-eastern zone that has produced many of the villas from the region and it is the zone where coins, pottery and other artefacts are found in relative abundance. In sum, this Cotswold/Avon zone is properly part of the wider culture-province of the south and east of Roman Britain and betokens the uptake by some of the population of the economic and cultural innovations available consequent upon the incorporation of the island into the Roman empire. It is not, though, seen as betokening any major change in ethnic composition of the population. The chronology of these changes is as yet poorly understood, but as elsewhere the turn of the 1st and 2nd centuries seems to mark a major threshold of innovation, developing through the 2nd and into the early 3rd centuries, with the later 3rd and 4th centuries marking something of a change in urban, villa, religious and funerary patterns.

Interestingly, it is the same zone which at and after the collapse of Roman rule in the early 5th century shows further change, in particular with the disuse and disappearance of Roman-style settlements, structure, artefacts and technologies and with the appearance of ‘Germanic’ material, above all the artefacts and pottery present in certain burial sites (Ford 1996; 2003). Here the dominant interpretation has been in terms of changing ethnicity, the arrival of Anglo-Saxon incomers, rather than of cultural adaptation, though this is beginning to change.

Burial and identity

When assessing the composition of the archaeology of the region in general and the south-eastern zone in particular, it is worth pointing out that there is one major site-type about which we know little, to our loss: burials and cemeteries. These certainly existed at urban sites such as Alcester (also Wroxeter in the north-western zone) and the rural site at Wasperton, though their presence elsewhere is more open to question, perhaps explaining their poor showing in the literature. The osteological and artefactual analysis of burials for information on age, gender, status, way of life and cause of death is well-established and could make major contributions to studies of individual and group identity. But even more promising are rapidly-developing techniques of mtDNA and stable-isotope analysis, which could have a huge amount to say about the areas of origin of individuals and thus their ‘ethnic’ identity. This would be particularly important for the identification of incomers in the early Roman period, but also at the end of that period where ethnic identity was crucial both at the time and to modern study of the period.

The potential to distinguish between indigenous and incomer from the 4th century onwards would be critical in assessing the transition from the Roman West Midlands to the Anglo-Saxon kingdom of Mercia. To what extent was this a matter of the imposition of an alien culture or of the assimilation of that culture by the indigenous population (as had happened throughout the Roman period)? The cemetery of Wasperton (Crawford 1982; 1983) has already been proposed as a type-site for the indigenous assimilation model on the basis of its artefacts and burial rites, and the cemetery at Stretton-on-Fosse (Warws) (Ford 2003) may be another case in point. Stable-isotope analysis from such sites should be far more definitive.

The above can only be a brief ‘taster’ of how realigning our categories of interpretation can refocus our perceptions of many aspects of the more visible archaeology of the Romano-British period in the region, but it should be sufficient to show the value of such an exercise both for interrogating the existing database and for devising future fieldwork and research.

Community

Membership of functional communities can also be visible in the archaeological record. A very obvious one for Roman Britain is the army. Though the direct military occupation of this region was relatively brief, the West Midlands were, as has been seen, in all probability affected by the longer-term garrisons in Wales and from Chester northwards. The discussion of the Roman army as a ‘community’ has recently moved forwards rapidly (cf James 1999; 2001); this was not just the community of the serving soldiers themselves but also all their servants and slaves, their wives and children and the traders and others who earned their living from supplying the army. Such approaches allow a more balanced appreciation of the impact of the early military occupation in general and of particular important sites such as the fortress at Wroxeter, for they do not confine themselves within the defences. The military are, of course, one of the most archaeologically visible and distinctive groups in the landscape of Roman Britain as regards settlement-, building- and artefact-types, thus giving rise to the traditions discussed above of writing a form of military history. However, questions over the relationship of military and civilian and the assumption of an essentially male military world are increasingly being raised, so evidence for the presence of women, children, grooms and slaves inside the forts needs to be assessed. What of the presence of ‘feminine’ items of apparel or of child-sized items such as shoes within a fortress such as Wroxeter? How does this change our reading of such sites?

The possible longer-term military impact on the region through resource procurement was outlined above. However, as well as grain and quadrupeds such as sheep, pig, goat and horses, what of bipeds? How did the Roman army’s need for manpower impact on the region; did men and women leave the region? There is the evidence for a cohort of Cornovii at Newcastle-upon-Tyne in the later Roman period (Not. Dig. Occ. XL), or there is the tombstone of a Cornovian woman married to a soldier serving at Ilkley (RIB 639). In return, the men in garrison at Chester or in the Welsh forts may have had a considerable input into the demography of parts of the region. Also, it was argued above that the presence of items of military equipment at towns and other sites in the region might relate to the presence of military personnel in these ‘civil’ areas in the longer term, perhaps to do with resource procurement, perhaps representing veterans?

Buildings and the household

At the level of the family, there is another type of evidence which is starting to produce useful discussion: house-plans. Considerable work on the houses of Pompeii has opened up a number of models for discussion which might usefully be applied to the house-plans of Roman Britain (cf Wallace-Hadrill 1994). The Pompeian evidence suggests that the major criterion conditioning the patterning of these houses was access along a public to private axis; how far one was allowed to penetrate was a function of one’s status vis- à-vis the head of the household? Curiously, gender and age were much less visible in these houses, with no identifiable women’s quarters or nursery areas. Slave and service areas were, though, visible, reinforcing the importance of status as a determinant of layout. Such analyses of the surviving building plans from urban sites such as Alcester, Kenchester and Wroxeter and the villa sites of the region, could be most revealing. On the other hand, Smith (1978; 1997) has proposed an alternative scheme whereby building layout is patterned by kin relationships, his multiple occupancy model. Romano-British houses do not have to conform absolutely to a set pattern, indeed variation may be very revealing of peculiarly British concerns; for instance, it may be that criteria such as age and gender are more visible in Britain than in Italy. These sorts of analysis may of course be extended to the study of settlement layouts, both at the smaller scale such as ‘farmsteads’, but even up to the ‘landscapes’ of towns, at all of which functional and other zonings may be discernible.

Artefacts and the individual

For the individual, the most archaeologically visible statement of identity is dress, which betokens to the viewer age, gender, status, and membership of wider communities (eg the military or a particular ethnic or social grouping). The Roman period is of course rich in dress accoutrements, particularly brooches, belt-fittings and items such as pins, widely distributed on occupation sites in the south-eastern part of the region. Recently, the study of such material has been advancing considerably through theoretically informed questions being put to the rich database (eg Crummy and Eckardt 2003). As yet, this region has not benefited much from such approaches, though the south-eastern zone appears on the edge of some recent distribution maps, showing the potential of this type of study for the area. The evidence for brooch manufacture from Wroxeter does give particular types where the distribution of the finished product could yield important information on the extent of contacts from the town and allow questions to be put over the mechanisms of distribution; trade? group identity?

Diet and definition

One of the ways in which the Roman military community was distinctive in the earlier Roman period was in its diet (cf Grant 2004). This is of course only part of a wider topic of the ways in which the types, preparation and consumption of food is used to define people’s self-definition as part of groups or peoples (‘foodways’). There has been some work on the ‘Romanisation’ of diet through the introduction of new crops, herbs, beverages, butchery practices, and in the preparation and consumption of food and drink, including changes in the related pottery and other vessels. The English Heritage environmental reviews for the region suggest that in fact it was one in which there were few introductions of new foodstuffs, but on the other hand the evidence of pottery and metalware suggests that in the south-eastern zone at least the ways in which food was consumed changed markedly, with the Roman period characterised by vessels for the individual preparation and consumption of food and drink. The osteological evidence for changing preferences in the Roman period away from sheep towards cattle is visible also in the West Midlands, though the butchery evidence is so far scarce and remains to be synthesised in detail.

Ritual and religion

The West Midlands is a region poorly endowed with the sorts of evidence traditionally used to structure consideration of religion in Roman Britain, principally temples (to locate where worship took place) and inscriptions (to show what sorts of deities were being worshipped). A roughly classical temple is known from Wroxeter (Bushe-Fox 1914) and a very probable ‘Romano-Celtic’ one from aerial survey at Kenchester (Wilmott 1980), and others will have existed at ‘small towns’ such as Alcester and Droitwich. In the rural areas there is the temple complex at Coleshill (Magilton 2006), a probable temple known from aerial photography near Craven Arms, the very different evidence from Croft Ambrey (Stanford 1974, 131-55) and the carvings from Wall (Henig 2004, 51-3). Apart from the small group of inscriptions from Wroxeter (mainly military tombstones) the region is essentially anepigraphic. However, the recent work at Orton’s Pasture, Rocester, has identified what is most probably a ritual or religious site (Ferris et al 2000), albeit one lacking the traditional appurtenances of temple structure or inscription. This identification was made possible by the recent development of alternative ways of identifying ritual or religious behaviour and the continuing development of new ways of thinking about religion.

This has again been much influenced by developments in prehistory, particularly the realisation that much deposit formation and associated presence of artefacts at Iron Age sites was not solely functional or random rubbish, but the product of processes of ‘structured deposition’. Particular types of ‘special’ deposits have long been recognised in Romano-British archaeology, most obviously burial, but also deposition of objects at temples or in temple precincts, and also the burial of hoards of coin and plate (though these have tended to be analysed more for their contents than their contexts). However, what is becoming clear is that there is a variety of Roman-period contexts where the deposition of material is clearly not random or rubbish. Contexts such as pits and wells, where the backfills have clearly taken place over time, and with the inclusion of objects and of animal and human remains in significant associations, are a good example of this. For instance, the presence of altars and other material in the backfill of wells in the extra-mural areas of Alcester strongly suggests this sort of practice (Mahany 1994, 113, 144). Roman period pits at Wasperton contained unusual deposits, including in one instance a stone inscribed FELICITER associated with two sets of antlers (Crawford 1981, 124). Ditches also, particularly boundary ditches, are now recognised as places for unusual deposits such as human burials, for instance at Tiddington (Palmer 1982) or the household deity from Berwick Alkmund Park (White and Gaffney 2003). The list could be multiplied, but what is important is that such locations are now recognised. It is also becoming important to be more attentive to the artefactual and other material from such locations, for instance finds of human bone, traditionally either ignored or assumed to be from disturbed formal burials, but now increasingly realised to be purposive (cf Esmonde Cleary 2000).

If the phenomenon of the structured deposition of selected artefactual and other material is now more widely recognised, how is it to be interpreted? The approach to these deposits has often been spatial, looking at them in relationship both to the features in which they are contained and these features in relationship to the site as a whole or in part. Spatial analysis tends to lead to cosmological interpretations, trying to use these features as a guide to how their creators divided up and assembled the world around them. It can also reveal structures of binary opposition by place (inside/outside), but also by criteria such as gender and possibly age, suggesting that it may be possible to approach other religious themes such as rites of passage. Furthermore, the artefacts and other material represented in these deposits are often also present in formal human burials and cemeteries, arguing that we are looking at a suite of ritual activities across a wide range of religious contexts.

It should be clear that all major types of evidence – settlement, structural, ritual, funerary, artefactual and environmental – have their contribution to make to these over-arching themes (as well as to others such as Resource Mobilisation and Tradition and Innovation). Furthermore, the themes entail the use of suites of evidence from across the major types in new ways and combinations, thereby increasing the range and sophistication of the analyses and militating against continuing to run in the same well-established ruts.

Assessing the gaps

Our knowledge of the archaeology of much of the West Midlands region remains patchy and, consequently, our understanding even more so. The characteristic definition of this zone is in negative terms; it lacks the sorts of archaeological evidence which characterises the civil archaeology of the south and east of Britain. Wroxeter is the only major town, and has of course been the focus of large-scale excavations over the last century (cf White and Barker 1998, ch 1), with more recent geophysical surveys filling out remarkably our picture of activity within the defences (Gaffney and Gaffney 2000). Kenchester is the only ‘small town’ which conforms with those seen further south and east, indeed it is a classic example of the type (Wilmott 1980). Otherwise, putative ‘small towns’ such as Blackwardine and Leintwardine remain poorly characterised. Wroxeter (in particular) and Kenchester each have a group of satellite villas, with Acton Scott, Whitley Grange and Yarchester among the better-known near Wroxeter and Magna Farm and New Weir near Kenchester, but these apart, this type of rural settlement is generally absent. So also is any settlement that looks like the ‘villages’ to be found in parts of the south and east (cf Hingley 1989, ch 6). Largely absent too are temples and shrines of ‘Romano-Celtic’ type (apart from a probable ‘Romano-Celtic’ temple near Craven Arms), as too, by and large, are burials (cf Philpott 1991, passim). Artefacts too are thin on the ground, be it from settlement sites or more generally in the landscape with, for instance, very few coin hoards from the area. The Portable Antiquities Scheme is beginning to add some dots to the distribution maps for counties such as Herefordshire and Shropshire (eg Worrell 2005), but these remain few relative to other areas of Roman Britain. It is important not to over-schematise the differences; there is a penetration of Roman-style markers into this zone, but only patchily as regards both the type of evidence and its geographical distribution, so the type and distribution of this material need further definition and explanation.

Photograph taken for an aircraft of a field with square marks on the ground showing settlement.
Fig 4.2 Statements through the extent and complexity of settlement earthworks: Chirbury (Salop)

Interpretatively, such a region was all too often either largely ignored in favour of the more archaeologically visible regions of the south and east and the military north; or else it was written off as a ‘failure’ because it did not evince the assimilation to Roman-style practice implicitly assumed to be the goal of any self-respecting provincial (the élite, at least), remaining at a ‘peasant’ or ‘prehistoric’ cultural level. Another approach, which may still have some validity, was to interpret the difference in terms of what little is known of the political geography of Roman Britain, with the south-eastern zone belonging to the civitas of the Dobunni and the north-western to that of the Cornovii. The ‘fit’ is, of course, not exact, with Kenchester belonging to the Dobunni despite the low level of Roman-style archaeology in this area.

However, as will be seen below in the Tradition and Innovation discussion, there is much more to it than that. This essentially negative characterisation restricts our understanding, so the zone needs to be characterised in more positive terms which reflect what is there, rather than what is not.

What we are left with is essentially a series of landscape areas archaeologically characterised either by enclosed (Fig 4.2) or unenclosed settlements and by field- systems, or by a lack of anthropogenic features. As a result of survey projects such as that by the RCHM(E) (Whimster 1989), the Wroxeter Hinterland survey (Gaffney and White 2007) or the Marches Upland Survey all depending heavily on aerial photographic data, there is now a considerable body of systematised data to be drawn on. These can be supplemented by the county period assessments drawn up as part of the present Research Framework exercise (White in prep). One avenue of approach might be to define a number of areas within the broader zone on the basis of physiographical characteristics – elevation, drainage, solid geology, drift geology – creating essentially a number of areas of differing agricultural potential (arable, pastoral, forestry). This would define relatively small-scale landscape blocs (what the French might call pays) susceptible to individual analysis in terms of their natural resources, and onto this the map of human activity could be overlaid. Some of these landscape blocs would also be susceptible to current techniques of landscape analysis through GIS, as has been tested in the Wroxeter Hinterlands Project, or to phenomenological approaches for highly distinctive blocs such as the Stiperstones.

Essentially, discussions have had to be based on morphological characteristics of the settlements located by aerial photography (size, shape). These have shown that there is a small number of general correlations that can be made: rectilinear enclosures are more common in the eastern part of the zone than the western; larger and more elaborate enclosures (eg multivallate) tend to be more common in the river valleys than on the surrounding uplands; so also do unenclosed settlements, rare outside the valleys. There would therefore seem to be a broad distinction between the major river valleys such as the Severn and the Lugg and their adjacent uplands. It must be emphasised that these are trends, not hard-and-fast divisions. Unfortunately, due to the paucity of excavation at these sites, it is at present not possible to understand the variations and explain the reasons for them, or instance whether they are related to factors such as date, status, agricultural regime or social structure (cf Wickham 2005, for an illuminating discussion of a predominantly pastoral society in early medieval Ireland, its social formations and its archaeological manifestations). Nevertheless, it is clear that the broad zone of the north and west of the region is susceptible to being broken down into these smaller areas or blocs.

In broad terms also, it would seem to be the case that the relatively thin and poor soils of much of the upland meant that these areas were much more suited to pastoralism than to arable agriculture. The latter, of course, could be carried out profitably on the richer soils of the river basins, perhaps accounting in part for the observable differences in settlement morphology and for the presence in these areas of evidence for field-systems and other patterns of enclosure. Indeed the two might have been intimately linked in patterns of seasonal short-range transhumance, so even parts of the lowland landscape might have pastoral elements alongside arable. It has been argued for instance (White and Gaffney 2003, 223) that the pattern of enclosures visible in the immediate vicinity of Wroxeter is better linked with cattle rearing than with crop cultivation. The landscape of the north and west of the region may therefore have been one dominated by pastoralism. Given the demands of the long-term garrisons of the Roman army just outside the region for leather, wool, meat and bone, pastoralism might well have been the ‘rational’ response to the landscape. It may also be that some of the areas currently devoid of settlement or field-systems really are ‘blanks’, having been areas of managed woodland, responding to the military and other demands discussed above.

The north and west of the region is a zone likely to remain difficult to understand until such time as a great deal more fieldwork, particularly extensive excavation with particular emphasis on the recovery of environmental data, has been undertaken. This is particularly evident in the difficulty of tracking change through time at the level above the individual site; this may relate to the suggestion made below of a traditionalist society where change happened over the longer rather than the shorter term. Nevertheless, even at present it is possible to see the overall outlines of a system by which this much more problematic database can be interrogated. It is also worth remembering that even in the south and east of the region there are extensive areas where the archaeology is of ‘farmsteads’, field- and enclosure-systems and apparent ‘blanks’, reminiscent of the north and west, so even there we must not allow our vision of the archaeology to be dominated solely by the highly visible towns and villas.

Tradition and innovation

The division of the archaeology of the West Midlands into the two broad zones of the south and the east and the north and the west is one which of course is an expression of deeper social and ideological structures. In order to approach these and thus to try to explain, at least in part, the divergences in the archaeology, this section of the discussion will address explicitly a theme already touched upon: tradition and innovation. In broad terms it could be said the south and the east exhibit more innovation in their archaeology because of the adoption of a range of settlement and artefact types derived from provincial Roman practice, whereas the north and the west are more traditional in eschewing such developments. However, tradition and innovation should not be taken as proxies for inertia and for progress, this would be too simple. Each is the result of a series of positive choices rooted in the social and ideological structures and it must be our task here to try to unravel and understand these structures. Therefore, this section will also be the most explicitly chronological. The discussions above have tended to be a-historical in the sense of presenting the evidence undifferentiated by date; here the time-line will be much more important. This also means that it is this section which will allow us more easily to address the questions of how the Romano-British archaeology relates to what immediately preceded and succeeded it – late prehistory and the early Middle Ages. It is proposed here to start by considering the ‘traditional’ areas of the north and the west of the region. In part this is to get away from the assumption that it is change (‘Romanisation’) that is the more interesting; in part it is because it might well be easier to describe and understand innovation if we have an understanding of the base-situation out of which it develops and with which it is to be contrasted.

The extreme shortage of ‘Roman’ settlement-, structure- and artefact-types, as well as of evidence for religious practice and burial in this zone is striking. Comparison with the medieval period reveals the potential wealth of the zone, so the explanation is not determined by environmental or other natural factors; it must be anthropogenic, the results of explicit or implicit human choices. The reasons for the lack of uptake of Roman- style material culture have been the subject of considerable debate in the last decade or so (eg Metzler et al 1995; Webster and Cooper 1996; Mattingly 1997; Mattingly 2004), provocatively summarised by Whittaker (1997, 145) as ‘If received enthusiastically, the process is called ‘Romanization’; if less so, ‘resistance’’. ‘Resistance’ has enjoyed a certain vogue as a supposed explanation, but it is, of course, a term currently implying a strong political agenda; it is also problematic in that it is essentially negative in its construction of motivation. ‘Resistance’ is in fact not only a reaction against something, more importantly perhaps it is a statement in favour of something else, most often the status quo ante. If we can characterise the grounds for the people of much of the Romano-British West Midlands to prefer to retain the established order of things, then not only may this explain (away) ‘resistance’, more helpfully it may aid in establishing further ways of understanding and interrogating the existing archaeological record and thus of formulating future fieldwork and research.

A useful comparison and point of entry for a different understanding of what was (not) taking place in the region in this period is afforded by arguments arising out of the situation in the Netherlands, encompassing essentially the Roman province of Germania Inferior. There in the Roman period there was a similar situation to that found in Roman Britain – a heavily militarised frontier zone full of military installations and personnel co-existing with an immediate hinterland with few visible signs of ‘Romanisation’, then further to the south in the province of Gallia Belgica a ‘Romanised’ landscape of towns, villas, temples etc. Rather than polar oppositions of ‘Romanisation’ and ‘resistance’, Dutch workers have proposed an explanation more grounded in the nature of the archaeology of the pre-Roman and Roman periods. It is argued (for a clear and convenient summary see Roymans 1996) that in the Late Iron Age economic and social worth (particularly for adult males) was grounded in cattle and warfare, because of the particular environmental conditions of the lower Rhine and the delta. This created a very particular ideology of society and of male values. Religious and ritual practice were also heavily conditioned by these factors, further reinforcing the ideological structures of the society. These were all reflected in the archaeology of settlements, building and deposition.

After the Roman conquest, the martial ideals could be perpetuated through service in the Roman army, and indeed the province yielded an exceptional number of auxiliary units for such a small area, particularly the multiple cohorts of Batavians. But on the other hand, the economic, social and ideological/religious formations promoted by the centrality of cattle and pastoralism meshed poorly with ‘Roman’ social and ideological/religious practice, not necessarily through antipathy or resistance, but simply because these made little or no sense in the existing structures. Thus, the hinterland of the Germania Inferior remained notably un-Romanised. Interestingly, even all the men who had served in the Roman army and had adopted Roman military identities become largely invisible when they retire and return home, resuming their ‘indigenous’ identities. It is only further south, in the grainlands of Gallia Belgica, that Roman-style practices took root.

The moral of this for the West Midlands is clear and instructive. If the north and west of the region (the lands of the Cornovii?) is one where pastoralism was an important (the dominant?) agrarian mode, then we may be looking at a similar situation. The late prehistoric archaeology of the zone is even more fugitive than the Roman-period, but it may be that the importance of pastoralism was pronounced, and the presence of all those hillforts suggests that warfare (however ritualised) may have been a significant activity. If so, then we may hypothesise along the lines proposed for Germania Inferior. The theatre of warfare may have been transferred to the Roman army, not so much in the short term of the conquest, but in the longer term as the units in garrison in Britain increasingly turned to local recruitment. Thus, we may have an aristocracy for the north and west of our region (peoples such as the Cornovii), it is just that we have been looking for it in the wrong place. Instead of looking for it only within the region at sites such as Wroxeter, we should perhaps be looking to Chester, to Wales and to the military north and for its expression in the form of the Roman army. If these men retired back to their homeland, then it may be that like the Batavians in Germania Inferior, they resumed their indigenous identity rather than perpetuating their Roman military identity; the presence of a military discharge certificate (‘diploma’) at Wroxeter is interesting in this regard. So here we may have another aspect of the West Midlands as a resource-procurement zone, as a recruiting-ground for the Roman army. However, this may have had further demographic and social impacts on the region by removing from it many of its most innovative young men, thus helping perpetuate the status quo. It may also explain why at least one Cornovian woman, commemorated at Ilkley, left her homeland for a husband.

But for those who stayed at home, the Dutch analysis may also offer a model to help explain the archaeological evidence of this north-western zone. The population of the zone remained attached not only to the largely pastoralist economic basis of their life as developed in later prehistory, but also to the social and ideological structures founded upon that basis. So, when they came in contact with Roman concepts and material, these were essentially alien to them and had little meaning or value within their frame of reference; as a result they by-and-large did not adopt them, or adopted only specific elements which they found useful/meaningful. A specific instance of this may be the site at Whitley Grange where the architecture and values of the ‘villa’ and of Roman-style bathing and dining were adopted (all markers of aristocratic status in the wider Roman world), but pottery by contrast was not significant (only 325 sherds). Such a model allows us to forge a possible understanding of the rationales which lay behind the near-total absence of Roman-style material culture over much of the region. Moreover, it allows us to do this through an appreciation of the particular nature and consequences of the zone’s archaeology rather than through the imposition onto it of teleological categories such as ‘Romanisation’ or ‘resistance’. What is more, ‘tradition’ is not a default position arrived at through inertia, it is rather a positive choice and one which, moreover, has continuously to be redefined and restated. This may mean that there is observable change over time, not stagnation, as successive generations redefine ‘tradition’.

Photograph taken from an aircraft of a landscape with multiple fields and the line of a Roman road, hills in the background.
Fig 4.3 The impact of Rome: the Roman road from Wroxeter to Leintwardine framed in upland and lowland landscapes.

If the argument proposed above has merit for the zone of the region which we have characterised as ‘traditional’, what contribution can this make to an understanding of the motives for and expressions of the ‘innovative’? Again, the nature of the Late Iron Age evidence of the south and east of the region will be crucial for an understanding of what follows. Are there observable differences in the environmental evidence for economic basis (e.g. arable rather than pastoral) or for the social structures of the zone and consequently in the patterns of settlement and of artefact use and deposition? If so, then this may have been a zone more open to the adoption of innovative practice consequent upon the arrival of Roman-style practices. Clearly, in gross terms this was a zone which in the Roman period did refashion its ideologies, allowing features such as towns, villas and objects to become widely diffused as meaningful signifiers of identity. There is one medium which may allow us to detect and calibrate such changes (or lack of them) not only in the south-eastern zone, but across the whole region more widely. This is the most prolific of all Romano-British goods and one which has so far been (curiously) absent from this discussion: pottery.

Pottery and cultural change

In the West Midlands, as elsewhere in Roman Britain, the principal use of pottery has for a long time been for dating, and this use will undoubtedly and rightly continue.

Nonetheless, it is two other features of pottery that must concern us here: overall distribution and trade. The huge variability over the region in the incidence of pottery in general and of fabrics and of functional types in particular, is of course one of the most noticeable and characteristic features of Roman-period archaeology. For example, two comparable projects, the Arrow valley pipeline in Warwickshire (Palmer 2000) and the M6 Toll motorway in Staffordshire, both produced linear transects across the landscape and a range of settlements and other indications of human activity. However, the disparity in the quantity and range of pottery between the two projects was remarkable, and when the latter is published it should be possible to make precise comparisons. Additionally, clearly it should be possible on present data to produce distributions which can be manipulated (eg by trend surface analysis) to show the extent of the south-eastern zone with strong presences of ceramics and equally importantly the fall-off line between that zone and the north-western one. Beyond this there would of course be islands of pottery use such as early military bases at Mancetter and Wroxeter or later towns at Kenchester and Wroxeter, but this would again emphasise the exceptional nature of such ceramic islands in a sea of general indifference.

As well as gross presence/absence patterns, the distributions of wares from particular production centres should also be revealing. It was noted for Rocester, for instance (Esmonde Cleary and Ferris 1996), that its ceramic relations lay to the east and south (Derbyshire, Mancetter products) rather than the west (Severn Valley wares), suggesting different zones of economic and/or social interaction characterised by the variations in the ceramic assemblage. However, pottery as a commodity transported over longer or shorter distances can stand as a proxy for the distribution of other things. The presence of Severn Valley wares on Hadrian’s Wall is probably a proxy for the long distance transport of grain or other matériel from the West Midlands to the military north, and we have also noted the distribution of 2nd-century Wroxeter mortaria largely to military sites. It could also be argued that the movement of pottery stands as a proxy for the movement of ideas and information. The very long distance movement of samian, the regional distributions of major fine wares, and the more local distributions of coarse wares could indicate the penetration not only of goods but also of ideas, and their distribution within the West Midlands shows the areas more open to such interaction and those to which it was not a matter of importance (cf. Peacock 1982). The tradition of study of Romano- British pottery for chronology would mean that variations in these patterns would be at once apparent. Another approach is through the spatial distribution of functional types. In particular, which vessel types do manage to penetrate into the north and west of the region and when? Do these types tell us about aspects of Roman-style practices making some headway in the region, in domains such as the storage, preparation, and consumption of foodstuffs and beverages?

The Roman period and the longer term

To conclude this essay the relatively short Roman period will be set in the longer durée of its adjacent periods to see to what extent the Roman interlude conformed to longer term rhythms of the region’s archaeology or stood outside them (Fig 4.3).

At some levels, the archaeology of the Romano-British period does show distinctive features. Perhaps most important is the great upsurge in resource exploitation and transformation which characterises the period. The archaeology of the later prehistoric and early medieval periods show nothing like this; individual resources, notably the Droitwich brine-springs, were exploited, but there is no evidence for the sort of systematic and intensive exploitation of a whole range of mineral and agrarian resources that it has been argued here is a determining feature of the Roman period. If one is to look for parallels to such behaviour, then it comes a millennium and more later with the later medieval period, leading of course to the early modern and Industrial Revolution periods when the winning and transformation of the natural resources of the West Midlands again come to mark and define the archaeology of the region.

Equally, other features of the Roman period mark a break with what came before and after; this is particularly true of elements such as the creation of an urban hierarchy, masonry construction, and the relatively widespread use of artefacts in the ‘Roman’ style. This was very much the case for the south-eastern zone of the region and of course marks its assimilation into the wider context of the south and east of Roman Britain. These are features which recur in the archaeology of the region, starting in the 7th century AD but not developing fully until the end of the first millennium AD or after. Overall, the Roman period sees the West Midlands far more integrated into the wider economic and social order of the province of Britannia and beyond it the wider imperial military, economic and cultural framework. This is, of course, the archaeology which suffers a catastrophic collapse in the first half of the 5th century, the archaeology linked to the Roman economic and cultural system. It is now widely accepted that the reasons for this collapse lie ultimately in the wider provincial and imperial frameworks, that the parts of society which maintained the towns, villas, coins, pottery, etc., were precisely the parts which suffered the deepest, irreversible crisis with the cessation of Roman rule in Britain (Esmonde Cleary 1989; cf Wickham 2005, 306-10). Interestingly, it is the south-eastern zone of the region that also seems most receptive earliest on to ‘Germanic’ material culture, particularly evinced by grave-goods in what is now southern Warwickshire and shading into eastern Worcestershire: are we again seeing an expression of the social and ideological divide in the modern region? What of the differences between the archaeology of the Arosaetna as opposed to the Magonsaetan or the Wreoconsaetan?

On the other hand, the archaeology of much of the West Midlands through the Roman period seems to fit better with that of much of the first millennium BC and the later first millennium AD. This is the poorly visible archaeology of a rural landscape operating apparently at local levels with little evidence for economic or social complexity or innovation. The archaeology of the north and west of the region in the Roman period looks little different from that for centuries before and after; the longer-term formations of the agrarian economy and society were what characterised so much of the region for over a millennium, if not deeper into prehistory. In this case, then, the contrasting face of the archaeology of the region may help us to another view of the ending of Roman Britain and of the relationship of the Roman to the early medieval. It can be argued that over much of the region and for most of the population, the terminal crisis of the wider Roman system in Britain at the beginning of the 5th century was not central to their lives. The longer-term rhythms of the agrarian landscape continued as probably did the social and ideological frameworks based on them, even if economic formations such as taxation and other resource extraction ceased. Paradoxically, therefore, the late survival of sites such as Wroxeter (Barker et al. 1997) and Whitley Grange may be an expression of the persistence of the ‘un-Romanised’ society in which they were embedded rather than of the survival of ‘Roman’ practices and values. In this case, it is not surprising that the archaeology of the early medieval period in the West Midlands is so fugitive; it is continuing a long-established tradition.

It is not until the 7th century that this part of the region sees the appearance of ‘Germanic’ material culture. Historians (cf Gelling 1992; Bassett 2000) are now pretty much agreed that the Anglo-Saxons arrive into a landscape peopled by Britons, who are almost as fugitive in the surviving documentary sources as they are in the archaeology. The heartlands of Mercia noticeably lay in the eastern part of the region, the western part remains much more shadowy to us. Again, it is not until late into the first millennium or later that many of the characteristic features of medieval archaeology become established in the region (for instance, widespread pottery use is a late development, cf Vince 1988), quite probably under the influence of the southern kingdom of Wessex, with the last elements to be put in place perhaps being associated with another set of intruders from overseas, the Normans.