3. the Romano-British Period Research Agenda (Pdf)

Chapter 3: Romano-British Agenda Chapter 3

The Romano-British Period Research Agenda

by Robert Philpott and Mark Brennand

With contributions by Peter Carrington, David Shotter and Sue Stallibrass

Introduction to be able to find out simply and quickly what work has been undertaken within a particular area, but A review of research into the North West’s archae- there is ultimately no single location where all grey ology in the Romano-British period gives at first an literature, reports and theses are listed, let alone avail- impression of an area relatively well studied. The dis- able for consultation. tribution of known sites in the HERs suggests a rea- sonably widespread and evenly distributed degree of Initiatives knowledge. In reality this distribution is unlikely to be genuinely representative of past settlement and 3.1 There is need for full assessment of the current activity and is biased according to a number of fac- resource beyond the scope of the Research tors such as geology and topography, agricultural Framework process. This work should cover regimes, and the selective nature of archaeological antiquarian work, museum archives, SMRs and research, including both antiquarian work and mod- HERs, and the archives of university depart- ern development-led projects. Despite a considerable ments and archaeological units. number of antiquarian investigations directed towards the archaeology of the period, a significant 3.2 Too much information remains unpublished or proportion of this work simply does not stand up to presented in grey literature only with no usable contemporary scrutiny. The first major gap in our synthesis, and a framework is needed to address current knowledge of this period, therefore, is in un- this regionally or nationally. derstanding the nature of the archaeological dataset; whether the data are genuinely representational, and 3.3 As a minimum, all Romano-British sites exca- where the biases may lie. The research potential of vated should have an entry in Britannia and this the archaeological resource for the period could be requirement should be stated within curators’ greatly increased overall by addressing several key briefs. areas. Unpublished Archives Improving Accessibility of Data A significant body of archaeological data, often exca- It remains difficult to access information about cur- vated under modern conditions, remains beyond the rent research and archives, despite the publication of public domain and is not currently under analysis or summaries of recent work in publications of the So- preparation for publication. This backlog is not the ciety for the Promotion of Roman Studies, the CBA result of any single factor, and the history of some and the Archaeological Investigations Project at post-excavation projects is as long and complicated Bournemouth University. Equally, it is currently dif- as their stratigraphic records. This has limited the ficult to trace all work undertaken in universities, at scope of the Resource Assessment and some of the undergraduate and post-graduate level, or to search problems highlighted below in the Agenda could be for that material by geographical area. There is a need addressed by existing material, if the data were avail-

Archaeology North West Vol 9 (issue 19 for 2007) 55 Research and Archaeology in North West England

Fig 3.1 Conservation of the granaries at Galava Roman Fort, Ambleside, after re-excavation for English Heritage (Jason Wood). able for study. The current backlog of unpublished Chester, Eaton by Tarporley, Manchester, Middle- material covers all types of sites from rural and urban wich, Papcastle, Ribchester, Walton-le-Dale, Wilder- locations, and is evident throughout the region. spool, and numerous smaller interventions. It is ac- Noteworthy examples for this period include ar- knowledged that some of the analyses on this mate- chaeological work on sites at Ambleside, Carlisle, rial are ongoing, and in the long term publications

56 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda will be forthcoming. However, some archives appear genuine reflection of the original distribution. This is to be static, with no apparent movement towards illustrated, for example, by the A5300 road corridor analysis and publication. which ran through an area of Merseyside thought to be devoid of Romano-British settlement. Fieldwork Initiative located three previously unknown sites occupied in the Romano-British period along a 4km stretch of 3.4 An audit and prioritised programme for post road. Repeated aerial reconnaissance in an area per- excavation of key sites is required to deal with a ceived as poor for production of cropmark sites has backlog of unpublished excavations (see above shown the potential to detect the relatively ephem- General Introduction). eral remains of rural or short-lived military sites in the restricted but often unpredictable windows of Chronologies and Dating favourable conditions. While the development control process may enable While pottery dating from the 1st to 3rd centuries is the Research Agenda to be pursued in certain parts fairly widespread in the region, it is not necessarily of the region such as areas of Cheshire or Merseyside found on all sites dating from this period, and reli- where there is intensive development pressure, in ance on datable ceramic material is not suitable or other parts of the region, such as the Cumbrian fells satisfactory to establish chronologies for the majority or upland Lancashire, development is restricted and of rural sites in the North West. Additional sampling takes place at a slower rate so there are fewer oppor- strategies need to be implemented in order to estab- tunities either to investigate known sites or to re- lish chronologically defined stratigraphic sequences. cover information on new sites through developer- There is still uncertainty about the actual foundation funded work. date of many rural sites with what appears to be Romano-British occupation evidence and equally Initiatives many questions about 4th-century or later phasing on both rural and urban sites. There is also considerable 3.6 Proactive programmes of fieldwork and air re- doubt that the end of the supply of durable material connaissance are required if we wish to see sig- culture necessarily occurs at the same time as the end nificant new understanding of rural society and of occupation on rural sites. The view that AD 410 economies, particularly in the uplands, during represents a significant event for the rural population the Roman period. and saw widespread abandonment of the settlements has been rejected and increasingly replaced by a view 3.7 Absence of known distributions should not be that society continued to function at some military regarded as genuine gaps and should be ad- sites as well as in the countryside. The occurrence on dressed positively through site assessments and several rural sites of early medieval occupation recog- evaluations (see above General Introduction). nised only by radiocarbon determinations of features Similarly, genuine absences of relevant material containing residual Romano-British pottery high- in fieldwalking or other interventions should be lights the potential for ‘Roman-period’ sites to con- recorded to contribute towards a fuller picture tribute to the identification of the highly elusive early of overall landuse in the period. medieval settlement. New Research Questions for the Romano- Initiative British Period in the North West

3.5 Many more radiocarbon dates are requiredrom Beyond established models for the chronology of Romano-British sites, for both early and late military sites, towns, industrial settlements and rural phases and most especially for rural sites. Briefs farmsteads, few new questions are being asked of this for development-led projects and project re- significant archaeological resource, or the interaction search designs should require positive discrimi- between the wider population and the military, the nation in favour of programmes of dating, economy or the environment. Aspects of the period, stratigraphic and scientific analysis (see above previously thought to be well understood, such as the General Introduction). military occupation of the region, are in fact imper- fectly understood. In short, there are currently too Identification of new sites many presumptions based on too little material. There is a need to go beyond more traditional ar- There are substantial gaps in the known geographical chaeologies and archaeological methods that are not distribution of rural sites, which are almost certainly a suitable to address the specific problems for this pe- result of the difficulty of locating sites and/or a lack riod in the North West, and to approach the archae- of appropriate archaeological fieldwork, rather than a ology of this period (as with any other) with method-

Archaeology North West Vol 9 (issue 19 for 2007) 57 Research and Archaeology in North West England evidence from this period has been improved signifi- ologies that will elicit significant data. This needs to cantly over recent years, especially in the southern be a top down approach starting with those writing part of the region. It is likely that many more Iron briefs and project designs for work, and implemented Age and Romano-British rural settlements will be and monitored for all aspects of field projects. discovered and excavated during development-led There is an inherent tension in research aspirations investigation, and further analysis will undoubtedly for the region. On the one hand there is a lack of begin to increase the number of sites known to be basic data and knowledge with which to formulate occupied in the Late Iron Age. It is important that new models for research, and on the other hand a these sites are excavated to uniform high standards, need to challenge traditional approaches in order to and that provision is made for a suite of scientific address the major areas of weakness in current ar- and artefactual analyses that are appropriate to the chaeological understanding. The Agenda for the pe- data and address current research questions. The po- riod attempts to achieve a balance between address- tential for accurate dating and phasing of rural sites is ing high level theoretical questions about patterns of available, but this will require targeted programs of social, cultural and economic relationships in the sampling and scientific dating. The potential for arte- North West and identifying research projects with factual dating cannot be relied upon prior to the well-focused aims that are realistically deliverable commencement of site work. given the Region’s distinctive archaeological charac- There has been little integrated research to relate ter and relatively modest infrastructure. the location of Roman military sites to the distribution of the native population. To attempt this for the The Late Pre-Roman Iron Age whole region probably requires far more information about LPRIA settlement in the region than is cur- The key to understanding early Romano-British soci- rently available. Nevertheless, an attempt should be ety is an understanding of Late Pre-Roman Iron Age made, in order to explore methodologies and estab- (LPRIA) society. It has recently been argued that the lish how much relevant data actually exists. Roman army invaded and occupied the North West because the economic and agricultural structure was Initiatives already in place to support a large garrison (Wells 2003, 81) but there is currently little evidence for this 3.8 There is a need for stratified and dated material beyond the environmental record. The sparseness of from all LPRIA sites, including the larger en- material culture on all sites constrains attempts to closed lowland and hilltop sites, the smaller en- reconstruct even general models of the chronology, closed settlements, and seemingly open settle- status and economy of most sites. There is little evi- ments. The assigning of dates to enclosures and dence for social stratification in the LPRIA and in cropmarks on morphological grounds is not many areas equally little sign after the army have secure and rigorous application of sequences of been stationed in the region. The region is ‘poor’ in radiocarbon dates are required to produce se- some respects but this also adds to its regional dis- cure chronologies. tinctiveness (Haselgrove et al 2001, 22-5), and to the challenges faced by archaeologists working here. We 3.9 There is a need to look for any evidence for the have little idea of the LPRIA population density fortification, or re-fortification, of enclosed sites (often presumed to be low), ethnicity and ethnic in the 1st century AD, and for evidence for ei- identity, beyond the use of unhelpful cultural labels ther continuous occupation or abandonment of such as ‘the Brigantes’. sites over the LPRIA to Romano-British transi- There is little archaeological evidence for the physi- tion. cal expression of cultural or tribal identity or social ranking, or understanding of how that might be de- 3.10 The question of identifying regional or tribal tected from the archaeological record. It is not clear identities may be addressed through artefact if natural resources (salt, metal ores, coal, grazing assemblages, building style and other indicators land etc) in the LPRIA were ‘owned’ by individuals to assess the continuation of these aspects or tribal groups, or exploited communally. The through the Roman period. marked difference in the LPRIA material culture across the Pennines, and more subtle differences in 3.11 There is unexplored potential for artefact stud- material culture within the North West region (eg ies to identify distinctive distributions in the either side of the Mersey), hint at cultural groupings type of artefacts in use and the frequency of within northern England which require detailed in- their occurrence. For example glass bracelets or vestigation. certain brooch types which may be indicative of There is a serious lack of information about LPRIA the economic and social networks within which settlement throughout the region. This covers all they were made, distributed and worn. It may classes or types of sites. The apparent invisibility of be possible to define LPRIA ethnic heartlands

58 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda

Fig 3.2 The line of the Roman road known as King Street, Cheshire (Cheshire County Council).

and peripheries in the distribution of Romano- Initiative British sites and artefacts. 3.13 It is perhaps time for critical reassessment of 3.12 A critical assessment of the location of the earli- relevant documentary data for the extent and est military installations and their relationship to organisation of the Brigantes and subsidiary known native settlements or material is re- polities, and to formulate archaeologically quired. A detailed project incorporating the achievable proposals to address questions mapping of known Late Iron Age settlements raised by historical interpretations. and landscapes in relation to known Roman military sites may elucidate aspects of Late Iron Environment Age settlement hierarchy and Roman military strategy. Such work could draw on the current Despite recent overviews for parts of the region (eg SMR/HER and Portable Antiquities Scheme Wells 2003) the existing environmental work for the (PAS) databases in order to interpret economy period is not synthesised or well disseminated. A sig- and political structure from the archaeological nificant proportion, up to 90%, is unpublished or data, and then attempt to reconcile the results within grey literature. There are few rural faunal as- with the documentary sources. semblages, largely due to the poor survival of bone, and it may be productive to look at other sources of Tribal Identity evidence such as lipids on ceramics to determine elements of diet and economy. In the North East, bar- The documentary data attests the ‘tribes’ or civitates of ley seems to be the predominant crop, in contrast the Brigantes, Cornovii and later the Carvetii. In real- with the south, but evidence in the North West is ity we know little about these polities. There has currently sparse. been a major discontinuity between analyses using There is a need to exploit a wider range of environ- historical and archaeological data; each operating to mental analyses. For example, only a very few sites some degree without consideration of the other. The have had analysis undertaken on buried soils (mostly former highlights the ‘Brigantian confederacy’ of along Hadrian’s Wall), although these have generally Cartimandua, straddling the Pennines, with ‘septs’ produced useful results and indicate a complex mo- such as the Setantii. The latter has so far emphasised saic of vegetation and land use. All military installa- the apparent lack of hierarchy visible in settlement tions with ramparts (eg forts, annexes, milecastles, types in the North West. Hadrian’s Wall and the vallum) have the potential to

Archaeology North West Vol 9 (issue 19 for 2007) 59 Research and Archaeology in North West England preserve buried soils. Turves utilised in construction survive on many sites and have been under-studied in terms of recognition, habitat exploitation and uses on site, despite new methodologies being developed to study them. Ditches, often associated with ramparts, may also contain preserved waterlogged remains of plants, sediments and animal remains with similar or complementary evidence. Existing work has demonstrated the rich resource available for geoarchaeological and palaeoecological analysis within lake sediments, peat bogs and coastal Fig 3.3 Denarius of the Emperor Trajan (AD 98– 17) from and estuarine deposits. Natural and anthropogenic Bebington, Cheshire (PAS). changes to inland river valleys, including realignment of river channels in the post-medieval period, have mate favours this type of preservation. Some forts preserved palaeo-channels with high, but unex- near river crossings (as at Ribchester) may have con- ploited, potential for organic survival especially in the siderable waterlogged or other anaerobic deposits. vicinity of settlements. Similarly the peat deposits of Rural sites are less likely to have preserved timbers/ the Romano-British period represent a significant roundwood due to lack of waterlogged material, but untapped resource and only a limited number have some may have occasional wells and pits where or- been investigated for their contribution to under- ganic material may be preserved. Woodland manage- standing changes in climate and vegetation. Lake ment is not well understood, despite the clear need sediments in Cumbria also represent a significant of the military for large quantities of timber and potential resource for investigation of climate change roundwood, especially in the 1st and 2nd centuries. and pollution from industrial activities otherwise un- Studies of roundwood from Carlisle (Huntley 1987; documented. A systematic programme of sampling in 1989) have demonstrated relevance to considerations all these types of wet environment could begin to of land-use/resource acquisition and woodland man- address the question of whether there really was a agement. There are clear indications that the avail- significant increase in land-use intensity and diversity ability of resources changed through time and that in the Romano-British period, and its consequences sometimes acquisition was ad hoc and sometimes for habitats, soils, and patterns of erosion. from managed woodlands. Coastal change during and since the Roman period includes the silting of the Mersey and Dee estuaries, Initiatives creating extensive salt marshes on the Wirral shore and infilling the Pool of Liverpool. The movement 3.14 A programme of analysis should target late and encroachment of sand dunes, which were well Holocene peat and silt deposits, with good dat- established by the Roman period on the soft coast- ing control and using a variety of methods. line of southern part of the region, present largely There is a need to look not only for woodland unexplored potential for the survival of extensive, clearance and the introduction or increase in well preserved buried land surfaces under the dunes, pasture and arable, but also for surface wetness as evidenced at Meols for other periods. Details of indicators relating to climate change and indica- coastal changes are not well understood and the ma- tors of industrial emissions. rine transgression is poorly dated, although this could have had major implications for coastal installations 3.15 Programmes of sampling should target estuaries such as defences and harbours. The extent to which and tidal reaches of major rivers for geoarchae- the coastal resources were utilised in the late Iron ological investigations of river deposits, espe- Age and Romano-British period is uncertain. There is cially the River Dee, with a view to locating the little direct evidence for fishing equipment of the main channel and its depth. Similar studies Roman period and even though molluscs and fish could be undertaken in the Fylde and Solway. have been recognised in the archaeological record, This could also elucidate how far upstream it remains of fish are scarce. may have been navigable and how military sup- There are almost no dendrochronological studies plies including personnel were deployed of the mid to late Roman period but a good dendro- throughout the North West region. chronology master curve has been constructed for 1st and 2nd centuries from Carlisle, one of only two in 3.16 Wherever ramparts or ditches are to be exca- the whole country. Many urban or proto-urban sites vated, a sampling strategy should be considered within the North West contain waterlogged deposits for the recovery and investigation of buried especially wells, deep pits etc that may preserve tim- soils, turves and similar deposits likely to pre- bers or roundwood in situ, and the region’s wet cli- serve pollen, insects and other micro-organisms

60 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda

likely to be indicators of past vegetation, water process of the military advance into the north and quality and landuse. Specialists need to look for the subsequent infra-structure and deployment of the turf material and be aware of what to expect. army is still poorly understood. The discovery of new military sites in recent years 3.17 Roundwood studies could be extended to other suggests we need to be more open minded about the sites to search for consistent patterns or site- topographical layout of the network of occupation. specific ad hoc exploitation of resources. A high Although we may have an outline of military activity, priority should also be given to dendrochro- a great many details still need to be filled in. Many nological studies of timbers, especially those military sites lack detailed chronologies, to elucidate thought to be possibly mid or late Roman in the strategy, chronology and geography of the con- date. There is a need to extend the dendrochro- quest of the region. The nature of any Roman mili- nological master curve to cover the whole of tary activity prior to this is uncertain. There is an ex- the Romano-British period. Studies of timbers, tensive backlog of unpublished material from many dendrochronology and roundwood need to be military sites. Some of this work was undertaken in linked with pollen studies of clearance and de- antiquity, while some was undertaken as rescue work forestation. This could be liked to studies of in the latter half of the 20th century. There is a great other woodland resources and use of land for deal of potential within material already excavated, pasture or for cultivation, with studies of cli- which in some cases has been analysed, but not pub- matic changes and evidence for soil erosion. lished. The recent work on the Brougham cemetery material, excavated in the 1960s, has illustrated the Military Activity high returns than can be achieved with the application of new analytical approaches and interpretation A significant proportion of antiquarian and later re- (Cool 2004). search focussed on the Roman military and the sta- Although many of the known fort sites are sched- tions of the northern frontier. Despite the quantity of uled ancient monuments and under little threat from archaeological research on Roman military sites, the development, they are not inaccessible to investigation. Work is not entirely prohibited, provided re- Fig 3.4 Timber water pipe within the Roman fort at Carlisle, search proposals are properly presented and ade- uncovered during the millennium excavations (OA North). quately funded, as current work at the Chester amphitheatre and Heronbridge demonstrates. The excavations at Birdoswald (Wilmott 1997) amply demonstrate the insights to be gained through large scale modern excavation of a fort site, although the ex- pense of large scale research exercises suggests that large scale work of this kind is unlikely to be undertaken in the future. This does not preclude entirely development-led work on fort sites beneath modern towns, such as Carlisle (and Stanwix), Chester, Lan- caster, Manchester, Middlewich and Ribchester. Many forts may have as yet unknown predecessors beneath known levels or adjacent to known sites. Possibilities include Ravenglass (Potter 1979), Beck- foot and Maryport. There is also the possibility of new fort sites being discovered. Potential areas include Farndon, Nantwich, Wilderspool, Walton-le- Dale, Fleetwood, Wigan, Affetside and the South Cumbrian coast. Non-intrusive work such as geophysics and aerial reconnaissance has enjoyed substantial success within recent years and should be utilised to address such questions. After the initial military advance and consolidation, when supply lines would have presumably been se- cured from the south, the Roman administration would have looked to the north for procuring supplies. The Cheshire and Lancashire industrial settlements were possibly primarily for this purpose. The inter-site relationship between military sites, industrial and manufacturing centres, field systems, rural

Archaeology North West Vol 9 (issue 19 for 2007) 61 Research and Archaeology in North West England sites and transport networks in Lancashire and 3.20 Potential sites of un-located forts have been Cheshire may provide an insight into mechanisms of identified and air reconnaissance and field movement and trade supporting the frontier (see evaluation / survey should be directed to these Settlement, below). Coin distributions are a key indi- (Farndon, Nantwich, Wilderspool, Walton-le- cator of the means of taxation and exchange between Dale, Fleetwood, Wigan, Affetside, Hincaster, the Roman authorities and further analysis of ceramic South Cumbrian coast, Barrow-in–Furness, St assemblages and metalwork may also illuminate the Bridget Beckermet). economics of occupation. Important contributions have been made by Shotter (1993; 1999; 2000), ana- 3.21 The nature of the coastal defences for the west- lysing the distribution of early coins, and Wild (2002), ern seaboard in the 3rd and 4th centuries from analysing that of early Flavian Samian. Rogers’s the Bristol Channel to the Solway Firth and, in (1996) argument of an early date for the mid- regional terms, from the Dee to the Solway, Lancashire north-south road is now out of favour could be investigated in tandem with a pro- and Richardson’s (2004) suggestions of early roads gramme of environmental research aimed at along the Pennine edge south of Manchester have yet improving understanding of coastal and estua- to be critically appraised. rine change. Chester formed a hinge-point in Roman control of western and north-western Britain. The structural 3.22 The ethnicity and country of origin of those history of the fortress (defences, major buildings, serving on the province’s frontier are subjects barracks) reflects the interest of the imperial state in of particular contemporary interest and while the region, whether in (re-) asserting military control bone survival is often poor, every opportunity or making architectural statements of its authority. to analyse surviving human remains should be Over the past 45 years extensive excavations have taken, in conjunction with study of artefacts, been carried out over large areas of the fortress. Nu- personal ornaments, burial practices and epigra- merous preliminary statements of the results are phy, to understand the ethnic origins of units available, but only full publication will allow an un- stationed in the North West. derstanding of the full nature of the evidence and critical appraisal of the current interpretations. Settlement and Landuse The publication scheme drawn up for English Heritage 20 years ago remains valid but resources of The complex inter-relationship between town, fort staff time and funding for external specialists are re- and countryside which developed for the first time in quired to deliver this. There are high-quality strati- the Romano-British period in the North West was graphic and finds archives for many of the excava- the outcome of new political and social relationships, tions, and much analysis has already been done. interactions and systems of communication and ex- However, there is a danger that, as time goes by, that change. Rural communities and their agricultural analysis will become obsolete and need to be re- economies need to be seen in relation to the new done. urban and military centres rather than as separate systems. The vici are in one sense the location of in- Initiatives teraction between the countryside and the town, but there is currently little data on which to base serious 3.18 The publication and dissemination of unpub- examination of the differences between material cul- lished work from military and fort sites should ture and lifestyles in different kinds of community. be treated as a priority. Work on the Chester Recent excavations have helped to improve our post excavation programme requires funding to knowledge of urban settlements but certain types of progress with the analysis that is underway, in site, such as villas, are apparently largely absent order to understand the structural history of the within the region. Manufacturing and industrial cen- Roman fortress at Chester. tres, often sited on navigable rivers, appear to characterise the earlier Roman period in Cheshire and Lan- 3.19 Development-led archaeology on military sites cashire but we still have only a limited understanding in urban contexts provides an important oppor- of their origins, relationship with markets and the tunity to review chronologies and phasing of organisation of production and process flows. The the sequence of fort construction and use. Par- Study Group for Roman Pottery (SGRP) Research ticular attention should be paid to any possible Framework confirms that the bulk of known material traces of early pre-fort phases or later Roman has been recovered from military and urban sites, occupation. This should be complemented by and both rural and vicus sites are poorly represented carefully targeted geophysical survey and re- (Evans & Willis 2002). search excavation on scheduled military sites, as There is little published synthesis of Romano- demonstrated at Birdoswald. British material culture in the region, especially out-

62 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda

Fig 3.5 Well preserved centurion's quarters and barracks of the First Cohort of the Twentieth Legion at Chester, revealed during excavations at Crook Street in 1973/4 (Chester Archaeology). side military sites and small towns. Arguably there is of the uppermost stratigraphic deposits. a crucial distinction between portable artefacts and The Roman settlements at Heronbridge, Middle- building materials and, while both may provide an wich, Nantwich, Northwich, Walton-le-Dale, Wigan index to the influence of Romanisation, the extent to and Wilderspool have all demonstrated excellent which Roman building types and techniques were preservation of archaeological deposits, and most adapted and adopted in different settlement types have produced a wide range of evidence for buildings may be a fruitful area for research. and industrial processes. Development-led excavation is likely to continue on most of these sites and Industrial settlements should address these research questions specifically. The nature of the site at Wigan in particular requires The so-called industrial settlements of Cheshire and clarification. Lancashire appear to have been primarily geared towards processing and producing supplies for the Initiatives army in the north. It has been suggested that the regular layout of both plots of land and buildings 3.23 The publication of recent investigation is re- may be evidence of direct military instigation and quired, and comparison and synthesis of work foundation, and the pottery assemblages have been on industrial centres. likened in character to those from the fort sites themselves (J Evans pers comm). What is lacking is 3.24 Development-led excavation in these centres archaeological evidence for their origin and status, should have clearly focused objectives relevant and of the communities occupying and working in to research questions for this particular class of them. Are these distinct and different from those in settlement. the surrounding countryside? Do their origins clearly post-date the arrival of Roman garrisons in the re- ‘Towns’, Vici and Canabae gion? The full range of productive and industrial processes taking place on these sites and the extent If military influence was the motor of societal change to which they were centres for wider distribution also in north-west Roman Britain, then one would expect remains to be defined. The prevailing view of their that change to be at its greatest in the settlements in later Roman decline should be tested against the re- the neighbourhood of the forts, and most especially sults of recent excavations and by closer examination the legionary fortress. There remain many unan-

Archaeology North West Vol 9 (issue 19 for 2007) 63 Research and Archaeology in North West England swered questions over the decline of vici, as most do not appear to have outlasted the mid 3rd century. There are no full, detailed vicus plans other than those from aerial photography and geophysics, and the organisation, phasing, population and buildings of the vici all require exploration. Again, evidence of the nature of these communities, their ethnicity and interaction with the surrounding rural communities are unexplored. What was the nature of the canabae and other surrounding settlements of the Chester fortress? Closer examination of the upper levels of these sites may assist in explaining what was going on in the later 3rd and 4th centuries and subsequently. The absence of diagnostic artefacts could be evidence of depopulation, but the reasons for this need to be examined. The explanation may be other changing Fig 3.6 Roman military buckle from Tiverton, Cheshire circumstances in the status and material culture of (PAS). these communities. As Dio Cassius commented – did the Romano-Britons become different without Rural Settlement knowing it? The division of settlement types for this Research The known distribution of rural settlement sites, sug- Assessment follows the now-conventional division gesting a sparsely distributed population inhabiting between military vici/canabae and other settlements small, dispersed farmsteads, is acknowledged as in- near military sites, inferring that native life was little complete. ‘Large areas of the North West have not affected by Roman rule. There may in fact be a range been subjected to detailed field-work and have not as of responses and reactions to Roman military occu- yet been covered by aerial reconnaissance’ (Shotter pation, varying over time and geographically across 1997, 104). The existing evidence for rural settlement the diverse region. The Chester environs in particular has been studied by Nevell (1999c; 2004a) and Phil- may present an opportunity to evaluate this presump- pott, but the sample size is small. Major gaps are west tion across a range of site types. A synoptic volume Lancashire, where Dutton’s Farm, Lathom is the on recent excavations and earlier discoveries in the only excavated Romano-British rural site, and South western canabae at Chester is at an advanced stage of Cumbria and the Furness Peninsula, where there is preparation by Chester Archaeology, subject to avail- an extremely low density of visible rural Romano- ability of further resources for completion. There is British settlement. Virtually no rural sites have been material for a similar volume on the eastern canabae, excavated in eastern and southern Cheshire. Targeted where important discoveries of field systems at the programmes of work are required using a battery of urban-rural interface are currently being made. Work techniques to overcome the obstacles of land-use, ongoing at Heronbridge is at the post-excavation geology or nature of settlement which render sites stage and another suburban settlement is known at difficult to recognise. Many of these aspects relate to Saltney. the Iron Age, Romano-British and post-Roman periods and the nature of settlement and economy for a large proportion of the rural populace may have Initiatives changed little between the latter centuries of the 1st millennium BC and the latter centuries of the 1st mil- 3.25 A programme of reassessment of existing ar- lennium AD (above Introduction xx). chives may produce new interpretations of the The lack of modern excavation and scientific dat- later phases on Romano-British towns. Chester ing on upland rural sites means that with few excep- and Carlisle have both provided evidence for 4th tions (eg Hoaen & Loney 2004) little new informa- century activity, although the nature of that ac- tion has been added to that of 1930s surveys to en- tivity is unclear. hance evidence for the chronology of the sites. Un- dated enclosures generally considered to be of late 3.26 Completion of post-excavation work on the prehistoric/Romano-British type are increasingly Chester canabae and re-appraisal of archives of known from the region through aerial reconnais- earlier excavations in other nearby areas could sance, though few have seen sufficient investigation form the basis for a review of the satellite settle- to be able to confirm their date or determine the du- ments around Chester, assessing the interaction ration of occupation. Some have been found to be of aspects of rural and urban communities. Iron Age in origin with either continuous or renewed

64 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda occupation in the Romano-British period. Many low- the region. We still have little understanding of land sites are truncated by the plough and lack the changes in land use and animal husbandry. So little is organic components of the finds and environmental known about the organisation and division of land assemblages. Waterlogged sites therefore have far that questions of estate management and administra- greater potential in terms of the preservation of tion can barely be addressed, and any continuity of structural timber, animal bones evidence, evidence land management from the LPRIA through the pe- for food preparation and consumption, and artefacts riod and into the post-Roman era is almost entirely a in leather, wood and bone. matter for speculation. The subject of the allocation There are still major gaps in our understanding of of land for military veterans and their relationship Romano-British rural settlement, and how the native with the native population also deserves exploration. populations interacted with the Roman authorities. The small number of properly excavated and pub- Current evidence suggests considerable differences lished rural sites in the region means that little is within the region. known of rural building forms, of spatial organisation It is evident that some rural communities in the on rural sites, or of the rural trading economy. As in north of England possessed very few typically Ro- later periods, there has been an over-reliance on ce- man artefacts, and do not appear to be active partici- ramics as a means of dating sites and at times there pants within a ‘Romanised’ society. In contrast settle- has been a failure to recognise the highly residual ments within the Dee/Mersey Basin area appear nature of much pottery. Many rural sites have strong more closely integrated into the Romano-British ceramic profiles in the 2nd to 3rd third centuries AD, market economy than other areas. The culture of the corresponding with a peak in the production and indigenous population needs to be examined in terms consumption of pottery. These tend to have a of the archaeological evidence, rather than in a theo- swamping effect on later deposits. retical model of Romanisation. While one might ex- The transitions at the beginning and end of the pect the demands of the army to be initially reflected Roman period are generally poorly defined by dura- in settlement expansion and increased agricultural ble material culture. At some Romano-British rural activity, followed by recession as that market de- sites, this can be virtually or wholly absent from the clined, it has yet to be shown that this pattern is dis- archaeological record, such as at Barker House Farm, cernible in the character of rural settlement in the near Lancaster University (J Quartermaine pers Romano-British period. comm), whereas some seemingly artefact-sparse sites Rural sites do not exist as isolated settlements in may produce a single finds-rich phase (Hoaen & Lo- the landscape and their wider context is important to ney 2004). Radiocarbon dating programmes are an understanding of the economy and settlement of needed as a priority to examine rural sites.

Fig 3.7 Roman bronze patera from Nantwich (UMAU).

Archaeology North West Vol 9 (issue 19 for 2007) 65 Research and Archaeology in North West England

Initiatives ment of soil and limited opportunities for fieldwalking, metal detecting or the retrieval of chance finds. 3.27 There is an urgent need for work to locate rural Reporting of finds, especially metal objects, to the sites and to investigate potential Iron Age/ Portable Antiquities Scheme (PAS), has been greater Romano-British sites across the whole region, in the southern part of the region. Whilst this may to determine their chronology, economy, char- reflect in part a genuine difference in levels of arte- acter, and to examine the origins of rural settle- fact and coin use across the Mersey it may equally be ment patterns. Whilst the sites of the majority due to the location of the Finds Liaison Officer of Roman forts and towns in the region are (FLO) in the south of the region. The recent ap- probably known, this is far from the case with pointment of a second FLO in the northern part of rural settlements. the region has begun to increase the number of finds, especially from Lancashire, which has often been 3.28 Rural sites need to be considered within their perceived as a black hole in terms of finds reporting. landscape context, by investigation of their field The PAS is currently recording important primary systems and boundaries, the local landuse, to- data, but there is no synthesis at a regional level. It is pography and exploitation of other resources. acknowledged that this is not the role of the PAS In the southern part of the region at least, the officers, but remains an extremely important require- immediate vicinity of sites has often produced ment. The identification and distribution of artefact earlier or later occupation. Thus the core of the types may produce potentially important patterns at a site is only part of the long history of the occu- localised level. There is no systematic publication of pation of a particular location, and a longer the finds (the database is on the internet), though view of the occupation is important to see the articles such as that on brooches from Cheshire by shifting of settlement component of sites. One Nick Herepath (2004) have begun to define the re- approach would be to investigate a sample of gional brooch use. While it is a significant asset to cropmark enclosures, in a variety of topographi- have the information on the PAS database this mate- cal and geographical settings, including an exca- rial requires analysis by finds specialists beyond that vation programme which embraces the area which the current presentation of the raw data alone around the site. can supply. Artefact distributions have considerable potential 3.29 A high priority must be the detailed examina- to distinguish between activity areas, or social use of tion of a sample of well preserved rural sites. It space within individual houses and buildings, within remains crucially important to establish the ori- the environs of buildings, within military stations, gin of sites and the chronology and nature of within the vici, canabae and rural sites. Not only is their occupation and abandonment. In order to there a requirement for much more systematic publi- overcome the problem of recognising occupa- cation of artefacts from the region (eg Buxton & tion phases with no durable material culture this Howard-Davis 2000), but also chance finds. With the work will depend on structured programmes of exception of pottery and coins, artefact studies of radiocarbon sampling throughout the strati- material from the region are sparse. Equally while graphic sequence. The need for such pro- there are many studies of pottery on a site basis, Gil- grammes should be specified by the archaeo- lam’s (1968) extensive overview is now significantly logical curators in order to ensure a consistent out of date. Colin Richardson published many ac- approach. counts of finds reported to Tullie House within the Transactions of the Cumberland and Westmorland Society 3.30 A study of SMR and HER data for finds of di- and Roman brooches have been the subject of sev- agnostic artefacts and hints of associated Ro- eral studies (Snape 1993; Philpott 1999; Herepath man building materials may indicate whether 2004) but otherwise there has been little attempt to the virtual absence of ‘villas’ may be more ap- characterise regional or sub-regional patterns of arte- parent than real and, in more general terms, fact use or manufacture or types (cf Dearne in Der- assist in mapping the scale and effect of Roman byshire). Publication of finds has been undertaken at cultural influence. best on a site by site basis, or all too often not for- mally published at all. This has resulted in a highly Artefacts disparate body of material which is difficult to assem- ble coherently, which in turn makes it difficult to There is generally a low level of material culture out- characterise the artefact use of the region, or to iden- side military and urban centres and its recovery is tify regional patterns of use. subject to a degree of bias. Differential systems of The study of classes of distinctive artefact – land-use, including the presence of extensive tracts of whether metalwork, ceramic or other – has the po- permanent pasture, means little rotation and move- tential to inform questions such as regional or social

66 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda

blage on rural sites – to assess the use of leather, wood, bark, textiles and other organic material. In addition the preservation of structural timbers offers the possibility of establishing with remarkable preci- sion the felling date, and hence a terminus post quem for construction of structures (cf Carlisle: Caruana 1992). The preservation of writing tablets at Carlisle provides information on aspects of Roman life and culture that are of international significance but are invisible to archaeological techniques.

Initiatives

3.31 Systematic publication of excavated assemblages from the region and of chance finds is a priority, in order to characterise contexts, identify regional types and intra- or inter-regional patterns of distribution. The unusually low level of material culture outside the major military/ urban centres in the region means that the publication of individual or small groups of objects is a higher priority here than in artefact rich areas.

3.32 Greater use of new and developing techniques, such as high-quality digital cameras and scan- Fig 3.8 Complete Roman beaker recovered from the base of an ners, should be employed, to enable cost- eroding cliff face at the Beckfoot cemetery site, Cumbria effective and high quality recording of artefacts (PAS). and greater ease of dissemination. identity, economic patterns of manufacture and mar- 3.33 Contexts with high potential for waterlogged keting (see Hartley & Webster 1973, 89-98). The dis- assemblages provide the opportunity for high crepancy between levels of durable material culture in returns in terms of understanding the totality of use in rural sites and nucleated (urban or military) material culture on sites that in this region gen- sites is significant and an assessment of the picture erally produce low levels of more durable mate- would be assisted greatly by corpora of artefacts. In rial. These should be a priority for investigation this respect a re-examination of old museum collec- and proper resourcing of such projects tions has the potential to yield important new in- (including conservation) should not be underes- sights. timated. The Roman material from Meols, found in the 19th century and preserved in collections in five separate Ritual, Religion and Ceremony museums, has brought to light an important group of pre-Flavian finds, prompting a re-examination of the The evidence for LPRIA religion is scant, and re- role of Meols in the early post-invasion period mains so for a large part of the region after the mili- (Philpott et al forthcoming). tary occupation. There are few temples in the region Waterlogged assemblages such as those at Carlisle and none which are not military or urban in charac- (eg Caruana 1992; J Zant pers comm), Nantwich (P ter. The concentration of metalwork (coins) around Connelly pers comm) and Ribchester (Buxton & religious sites does not appear to be a common trait Howard-Davis 2000), and exemplified par excellence within the region and for a large part native religious outside the region at Vindolanda, are of great impor- expression does not appear to have had an archaeo- tance in providing detailed evidence for the organic logically visible form. component of sites which is usually lost on dryland A study of the distribution of evidence of Romano- sites. Celtic deities (interpretatio Romana) and of deities who The general scarcity of durable material culture on were evidently local in origin, and not related to Ro- rural sites means that waterlogged deposits would man equivalents, may contribute to the understand- provide an important corrective to the impression of ing of the impact of Roman culture in different types low levels of material culture of all kinds, and would of communities. Evidence of Christianity is largely begin to enable us to characterise the total site assem- also lacking apart from some portable artefacts and

Archaeology North West Vol 9 (issue 19 for 2007) 67 Research and Archaeology in North West England tombstones. character of burial rites, the variety of burial practice, Initiatives as well as pre-burial and post-burial rites such as fu- neral ceremonies, and post-burial feasting and offer- 3.34 There is scarcely any knowledge of religious ings. Analysis recently published from Brougham sites in this period, apart from that evidenced emphasises the rich potential of statistical analysis of by inference from inscriptions and sculpture/ cemetery assemblages. figurines, and the opportunity to investigate any Burials occurring in cave deposits may be atypical such sites particularly in a rural context should but in some rural areas these represent the only burial be a priority. record from this period since soil conditions often preclude preservation of inhumations (see above In- 3.35 The suggestion of a sub-Roman bishopric based troduction xx). In this context the deposits recovered in Chester and the association of a possible from limestone caves in north Lancashire and south church within the Chester amphitheatre should Cumbria have particular significance, as they rep- be the subject for further research. resent a unique assemblage of human remains and artefacts from the period, although the circumstances Burial of their deposition remain open to interpretation. It is likely that there are also selective depositional prac- Burials can yield valuable information about indi- tices in play. It is by no means certain what viduals and the social groups to which they belong. proportion of the material currently held in the ar- In the North West burial evidence is sparse and very chive actually dates from the Romano-British period, unevenly distributed. Large samples of burials are although the association of artefacts would suggest desirable for the analysis of populations, to identify that numerous cave assemblages have Romano- the development of cemeteries and the rites practised British elements. The current project on material within them over time. However, burial practice may from the Doghole (C) and Dog Holes Cave (L) not involve the use of nucleated cemeteries, but be planned by Liverpool John Moores University will dispersed as single burials or small groups. These are provide an indicator of the potential for further cave nonetheless of value to identify cultural groupings, material. Radiocarbon dating and stable isotope using analysis of burial ritual and new techniques analysis will be essential in this context. Previously such as isotope analysis. Allied with DNA analysis of excavated material from further sites needs to be lo- wider populations in Britain and the continent, analy- cated and assessed. ses of inhumations in towns, military or rural sites may assist in characterising the origin or ethnic affini- Initiatives ties of population groups, to illuminate questions of population movements from elsewhere in Britain or 3.36 Wherever identified, cremated remains should the empire as opposed to those of local origin. receive full and integrated analysis to investigate There are remarkably few examples of funerary the range of practices and their distributions practice from rural areas, beyond the hinterlands of and associations. towns and military installations. Military communities often have unusual or cosmopolitan funerary prac- 3.37 Inhumations are less common survivals and tices, which will not necessarily be shared by other merit intensive study, with macroscopic mor- Romano-Britons, and certain practices continue in phological and metrical analyses. All inhuma- the North West after they have ceased in other areas. tions unaccompanied by dateable artefacts It may be that there are different funerary practices should be subject to radiocarbon dating, as well employed over large parts of the region or even that as stable isotopes and DNA analysis where ap- cremated remains are simply deposited in the ground propriate, in order to place these within a co- and not placed in ceramic urns. herent time-frame and to enable rural burial Overall, we need to be pro-active whenever practices to be characterised in topographical Romano-British funerary material is discovered in a and chronological terms. This may have signifi- rural context, or on the outskirts of the proto-urban cant implications for considerations of how, settlements and vici, as the rural population may not when and where foreign troops integrated with be as homogenous as previously believed. Recent the indigenous populations. studies at Brougham and Carlisle demonstrate a wide range of burial rites and practices, which may reflect 3.38 DNA analysis should be deployed wherever the presence of troops and their entourage from di- feasible, for indications of whether individuals verse geographical and ethnic origins, as indicated by in groups of burials are related and to shed light epigraphy. Cemeteries at military or urban sites have on questions of homogeneity of populations. the potential to inform on cemetery layout and chronological development, the chronology and 3.39 Current work on assemblages from cave

68 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda

Fig 3.9 Excavation of a kiln at Walton le Dale, Lancashire (OA North).

deposits including burials indicates their high of mineral exploitation and patterns of movement of potential and should be pursued further (see the mineral. Despite the region’s role as major coal Introduction pp 25-26). producer in later periods, very few analyses have been undertaken for the North West. Technology and Production Initiative Mineral resources and extraction 3.40 Analysis of the origin of stone for building, fu- The North West is rich in mineral resources such as nerary sculpture and quernstones may help to coal, metal ores, and building stone. The first mine determine patterns of exploitation of resources, shaft of Roman date has been identified at Alderley workshops and schools, and of trade on an in- Edge but the site otherwise lacks evidence of Roman tra- and inter-regional level. activity. There is little other knowledge of the sources of metal ores and other minerals used in the North Industrial processes West in the Roman period or indeed of how extractive and processing industries related and were Industrial production is a feature of many sites in the organised. As with the rural economy, it is unclear North West, notably the proto-urban settlements of whether technology and production were signifi- Cheshire and Lancashire and the Manchester vicus, cantly modified, or simply stimulated, by the de- although evidence is also found on rural sites. While mands of the Roman army and administration. some centres of production are known, details of Examination of river silts in valleys such as the industrial processes and the pattern of movement Ouse in the Vale of York has revealed traces of lead and consumption of both raw materials and the fin- pollution from extraction sites upstream, providing a ished products are still poorly understood. The re- record of landscape change and mineral extraction gion lacks any synthetic study of Romano-British (Hudson-Edwards et al 1999). Extraction sites such industry despite the prominent place this has in the as Alderley Edge and potential sites in Cumbria, as character of many settlement centres. Similarly there well as those outside the region in north-east Wales is a need for the characterisation of the production and Derbyshire, may benefit from application of outputs such as decorative styles of metalwork, their such techniques and, in Cumbria, by the analysis of distribution and where they were made. lakebed silts. Palynological analysis can source coal to Vivien Swan wrote twenty years ago with regard to particular outcrops, allowing us to build up a picture the Wilderspool industry ‘the need now is for excava-

Archaeology North West Vol 9 (issue 19 for 2007) 69 Research and Archaeology in North West England tion of more kilns at Wilderspool and in the north- west Midlands, as well as the discovery and exploration of the potteries near Carlisle or in south-west Scotland which were founded by migrants from Wilderspool around the middle of the 2nd century’ (Swan 1984, 105). Though much has been done to characterise production, typology and chronology of pottery in the region, there almost certainly remain unidentified manufacturing centres in the region. In terms of consumption, especially on rural sites, the recognition of problems of residuality and the limitations of pottery as a purely chronological indicator need to be addressed. Fig 3.10 Roman period wood-lined well at Middlewich, Initiatives Cheshire (Cheshire County Council).

3.41 Romano-British industries, and the communi- of social and economic relations in the North West, ties that were engaged in processing and pro- as outlined above. duction of ceramics, salt and metalwork, have The evidence for Late Iron Age trade is sparse, but been identified throughout the North West and demonstrates some long distance exchange and the are a strong characteristic of the region in this acquisition of high status artefacts. The nature of the period. Research is needed, however, to draw evidence, however, renders the bulk of these transac- together the many and varies sources for this tions invisible. distinctive aspect of the Romano-British period, For the Romano-British period too, evidence for and to formulate an integrated research agenda exchange is largely confined to the proto-urban set- rather than a series of single issue or site-related tlements and larger enclosed rural settlements, pre- research questions. dominantly in the southern part of the region. The presence of the army is thought to have pro- 3.42 Scientific analysis of materials and environ- vided the market for the majority of imported goods, mental sampling should be strategically applied and the monetary currency for trading, and the sug- to support site-specific research into industrial gestion is that a significant proportion of the rural production. population operated outside the monetary system making payments in kind. 3.43 For ceramics there is a need for the application Evidence for exchange is at present mainly re- of more sophisticated analysis than simply as- stricted to pottery, often not present in significant sessing sherd numbers, to look at quantity of quantities outside urban and military contexts, but pottery against volumes of stratified deposits. there is scope for tracking the exchange of other du- Alison Jones of Chester Archaeology produced rable materials such as metalwork. an agenda for Roman pottery studies in the The coin data for the North West has been assem- Dee-Mersey Region in 1994, and the SGRP bled by Shotter. There is an increase in coinage in the Research Framework has also identified a series 3rd century but this does not appear to be the same as of agenda points (Evans & Willis 2002) which elsewhere. There are a significant number of later need to be addressed. coins from Chester, but they are predominantly of low value. The coin data has potential for interpreta- Trade, Exchange and Interaction tion to create new models for economic activity, though models are also needed for the distribution The accepted hypothesis for the north-western prov- mechanisms for other material culture. inces is that the monetisation which followed the The identification of the trading centre at Meols, imposition of Roman rule, and especially the demand evidently a port from the Iron Age onwards, is based for money taxes, stimulated a monetary economy to largely on antiquarian reports and museum collec- some degree, and had a multiplier effect on trade and tions. Erosion has probably destroyed or removed economic activity generally. much of the site there and Roman finds diminished While many questions may be posed theoretically in number during the 19th century. The report on the about the relationship between trade, taxation and finds is currently in preparation. supply to urban and military centres and the impact There is however, considerable potential for associ- of occupation, the limitations of the archaeological ated structures, tracks and field systems to survive in evidence place serious restrictions on any modelling the immediate hinterland behind the modern sea

70 Archaeology North West Vol 9 (issue 19 for 2007) Chapter 3: Romano-British Agenda wall. man period and beyond. The movement of some Ports and Maritime Trade traded goods has been studied along this route, notably pottery including BB1 (Black Burnished Ware 1) Despite having an extensive coastline penetrated by from Dorset, though there has been little study of long estuaries and navigable rivers, disappointingly the traded products, and the direction and destina- little is known of the ports of Roman North West. tion of trade. Inter-regional and international trade Coastal and estuarine sites such as Carlisle, Lancaster, patterns deserve further study. As one of the princi- Chester and Ravenglass are all documented ports, pal coastal regions facing Ireland and the Isle of Man, though Chester is the only site to have produced the trading, diplomatic and military links across the physical evidence for harbour installations. There is Irish Sea deserve exploration (eg Robinson 2000), no direct evidence for the types of vessels employed especially in view of Tacitus’ reference to merchants in maritime traffic and trading in the region so far. visiting Ireland. Nor has the question of riverine traffic received Patterns of overland trade with neighbouring re- much attention although clearly important in the lo- gions to the east across the Pennines, with North cation of forts, settlements such as Heronbridge, or Wales (notably minerals), to the north into Scotland industrial sites such as Walton-le-Dale, Wilderspool and south into the Midlands also deserve considera- or Holt. Chester may have played a major role in the tion, using diagnostic goods such as sourced pottery redistribution of goods to north Wales and the North or other distinctive artefacts. West. No clear research agenda for this theme emerged While Roman roads have received abundant study from the Research Framework process though the as means of transport, the coastal maritime trade has import and export of goods are clearly evidenced been little studied in its own right. The road system from urban and military sites. The reason for this can be verified archaeologically and many aspects of may be because, like industrial production, this as- it have been mapped. In contrast the actual points of pect cannot be addressed regionally or strategically embarkation and disembarkation of people and until relevant material, dispersed in the publications goods from waterborne craft can in many instances and unpublished archives of large numbers of pro- only be postulated. jects, has been identified and brought together. Cur- The west coast trade route which can be glimpsed in rently there is no satisfactory corpus of information the Iron Age continues in operation through the Ro- on which to construct an economic model for trade in North West Roman Britain. Fig 3.11 Potential Roman timber and stone ford in the River Lune, Lancaster (Pete Iles/Lancashire County Council). Initiatives

3.44 Research is needed to draw together the potential evidence for this aspect of the Romano- British period and to formulate an integrated research agenda. Publication of the numerous excavated artefact assemblages recovered since the 1960s must be a regional priority, since they provide some of the raw data for analyses of production, trade and exchange.

3.45 An ongoing distribution plot of coins is required, in relation to date and value, in order to provide a model for trade against which other exchange distributions can be compared.

3.46 Individual, site-based projects which may make a significant contribution in terms of defining trading centres or products should be a priority for research.

3.47 Systematic survey of coastal, estuarine and river environments is required to assess the surviving resource and the potential for Romano-British buried land surfaces and structures. Coastal and river-edge development and engineering projects need to be undertaken with sufficient pro-

Archaeology North West Vol 9 (issue 19 for 2007) 71 Research and Archaeology in North West England

vision for archaeological works or supervision. beyond the traditional end date of Roman military Legacy occupation in Britain. Evidence from within the fort and the civilian settlement at Carlisle would also sug- There is the suggestion that there is a decline in activ- gest occupation into the 5th century. The nature of ity, or material evidence of activity, at a number of the later phases of occupation on military sites and sites in the southern part of the region in the 4th cen- urban sites is critical to this understanding and spe- tury; industrial sites appear to have little pottery and cific provision in project designs could help to ensure few coins for the late period, and little trace of con- it is recognised and investigated appropriately. struction. These are also the deposits most affected by later ploughing. At Chester a ditch filled with Ro- Initiatives man building material produced a single sherd of Saxon pottery indicating the residual nature of the 3.48 The ephemeral nature of the archaeology and Roman material. At Irby, Wirral a single 10th to 12th the low density of artefacts requires greater at- century spike lamp amongst much Roman material in tention to stratigraphic evidence and detailed a building foundation gully was the only datable evi- sampling strategies in the higher levels of sites dence for a whole structural phase. of all types in this period. It remains unclear whether the decline in the volume of 4th-century pottery indicates a reduction in 3.49 New research projects on well-preserved mili- the population and in the scale of activity or a re- tary sites with good indications of later occupa- sponse to a change in availability of pottery in the tion, to complement that at Birdoswald, are late Roman period. Excavation at Birdoswald clearly needed to establish whether this type of 5th- demonstrates continued occupation within the fort century and later activity may be typical or ex- ceptional.

72 Archaeology North West Vol 9 (issue 19 for 2007)