Compiled by Robert Philpott


The last decade has seen considerable advances in our understanding of the Roman period in the north west, through new discoveries made since 2006 as well as older projects reported on or published since then.

A number of important excavation reports published since 2006 have enhanced our understanding of key regional sites, in some cases clarifying their significance and status. Most are in the southern part of the region. Excavations at Nantwich (Ch) resulted in the  discovery of a major salt-producing settlement in 2002, providing a coherent context for scattered earlier finds (Arrowsmith and Power 2012) while the military origin of the long-known Roman industrial site at Wigan has been confirmed (GM) (Miller and Aldridge 2011). Other significant excavation reports have been published on Wilderspool (Ch) (Rogers and Garner 2007); Manchester (Gregory 2007; with a summary in Redhead 2011), Middlewich (Ch) (Garner and Reid 2013; Williams and Reid 2008; Zant 2016), Chester (Ward et al. 2012), and Chester amphitheatre (Wilmott and Garner 2017). In the north of the region,   major publications have appeared on Ravenglass (C) (Hunter-Mann 2015) and Carlisle (McCarthy 2017). Important works of synthesis include Carrington’s detailed analysis of the economic basis of the Roman north-west, taxation, production and the interplay between the rural population and the military administration (Carrington 2012). McCarthy’s overview of Carlisle synthesises a mass of published and unpublished data to determine the sequence of occupation and function of various ‘zones’, military, industrial, official in the extra-mural settlement (2017). The publication of papers from a conference organised by CBA North West in 2008 at the Grosvenor Museum, Chester (Saunders 2011) explicitly to explore themes identified in the research agenda for the Roman period, in an acknowledgement of the need to prevent a fossilisation of the research frameworks (Carrington 2011, 7).  

The application of geophysical techniques in extensive programmes of survey at Cumbrian military sites has shed new light on the detail of fort interiors as well as the extent and nature of their adjoining extra-mural settlements, especially in revealing the later phases of construction, as at Ravenglass, Maryport and Low Borrowbridge (all C) (e.g. Biggins 2011; Hunter-Mann 2015; Zant and Rowland 2015; Wegiel and Zant 2016).

While the number of known and investigated rural sites continues to increase, too often they remain visible nodes within wider landscapes of which little or nothing else is known. However, the extensive investigation of some sites is revealing evidence of organised landscapes with features such as ditches, enclosures and trackways. They often do not show up in aerial photography due to extensive pasture and glacial till subsoil, and when found in isolation, these landscape features are difficult to date. However, at Chester Business Park (Ch), and Middlewich (Ch) they form extensive Romano-British field systems of a kind which are not widely known from the southern part of the region and await further assessment to determine how typical they are of the area and period.

The Portable Antiquities Scheme (PAS) has continued to provide an important source of new information for the period (, particularly for rural areas. Since 2006, each county area in North West England (apart from Greater Manchester) has more than doubled the number of recorded Roman finds, with a new total of 3880 for the region by April 2017, although this is skewed somewhat by the records of individual pottery sherds from known Roman sites such as Maryport and Kirkham. The quality of the records has improved considerably over recent years, with a greater emphasis on precise findspots, detailed identifications and high-quality illustrations, the latter crucial for future re-assessment of finds as research continues.

The PAS has also provided important regional data for national artefact surveys, notably coins and brooches, the two most frequent Roman metal finds from the North West. A study of brooches by Cool and Baxter (2016) examines the use of one of the most common type of artefact from Roman Britain using the growing PAS database, assisted by Mackreth’s publication (2011) which records over 15,000 examples. Walton has largely drawn from the PAS dataset of nearly 58,000 Roman coin finds to March 2008 (against a new total of 232,397 in April 2017) to establish changes through time in the distribution of Roman coinage across England (Walton 2012). Inevitably broad-brush national surveys tend to gloss over the complexities of local conditions or circumstances, and case studies rarely focus on the North West, owing to the low level of durable material culture in the region during the Roman period. However, such work is of great value in providing a wider backdrop against which to identify regional patterns and emphasises the broad regional differences in distributions across Roman Britain. Comparison with data from surveys of the Iron Age and early medieval period (such as the VASLE project for metalwork AD 700-1100; is important in identifying long-term regional patterns of economic or social behaviour, such as a dearth of metalwork and coinage in the North West in sharp contrast to the North East. Pervasive patterns of artefact distribution can be related in part to population densities but also to long-term patterns of social interaction and display, and economic development.

Two national projects have attempted to identify and characterise such regional variation in landscape and settlement across Roman Britain. The ‘Fields of Britannia’ project investigated landscape evolution during the late Roman and earliest medieval period (Rippon et al. 2015). However, the palaeoenvironmental sequences and excavated Romano-British field systems whose relationship to the historic landscape can be established, are distributed very unevenly in Britain with a much stronger emphasis on the south and east where they are revealed by commercial developments. More successful for the North West has been the New Visions of the Countryside of Roman Britain project (Smith et al. 2016). This has played an important role in identifying and quantifying regional differences in variation in house forms, settlement types, and the composition of artefact assemblages.

The reassessments of surviving antiquarian collections and records have also produced new insights, notably at Meols (M) (Griffiths et al. 2007). Here, the identification of significant groups of pre-Flavian metalwork, coins and pottery suggests military use of the port in the period before the formal Roman military occupation of the area. Afterwards, it appears to have functioned as a safe haven and trans-shipment port for the legionary fortress at Chester but a distinctive civilian element is indicated by finds such as ear-rings. Detailed study of surviving museum collections shows continued use from the Iron Age to later medieval period on a site now largely lost to coastal erosion. The application of new techniques and mining of datasets has proved an important research tool for the period. Study of LiDAR data has led to a number of new discoveries of Roman road alignments and temporary camps. Stable isotope analysis of human bone is of great value in illuminating questions of diet and population movement, particularly appropriate for Roman Britain (e.g. Eckardt et al. 2009), although its potential is limited by poor survival of human bone in much of the region. The application of rehydroxylation analysis to date ceramics offers the potential to date accurately building materials, of particular value in forts and major urbanised settlements, and to refine dating of pottery production independently of typology or archaeological context, with potentially far-reaching impact on site chronologies.

Key Sites

Roman sites mentioned in text (created by Rachael Reader)

Roman Environment

Little fresh palaeoenvironmental evidence has been recovered from Roman sites over the past ten years, and there is still a dearth of evidence for agricultural practices across the north west. Two major regional English Heritage surveys, of invertebrates in archaeology (Kenward 2009), and of charcoal and wood (Huntley 2010), provide valuable syntheses of data and sites, and summarise the findings to publication date, but both have identified almost no new Roman sites since 2006 for the region.

Similarly, all the palaeoenvironmental work for the Fields of Britannia Project was undertaken prior to 2006 in Merseyside and Lancashire Plain. Rippon et al. (2015) suggest that there is variation within the land-use across the Western Lowlands in the Roman period with predominantly open landscapes to the south, and greater areas of woodland to the north. A total of 49 sites within this region have yielded late Roman (or broadly dated ‘Roman’) ditches whose relationship to the historic landscape can be established, mostly in the south of the region. Unlike the regions in southern and eastern Britain, however, the dearth of post-Roman material culture means that there are no instances where early medieval pottery has been identified within the upper fills of Romano-British ditches (thereby confirming that the ditch remained open). Overall, 59 per cent of the Romano-British sites have boundaries that share alignments or orientation with historic landscapes characterized by medieval Closes or former Open Field (Rippon et al. 2015, 258).

A few sites have produced evidence of crop regimes and the management and selective use of timber for fuel. The Ravenglass (C) vicus excavations produced evidence of iron smithing, through hammerscale, metallic slag and magnetic material. The charcoal from the site indicates selection of hardwood types which sustained high temperatures over long periods when burnt, notably oak and alder with some ash (Miller 2015, 102). Hazel, cherry type, alder and poplar/willow charcoal was also identified and reflects possible utilisation of structural wattle panelling or collection of local woodland resources for kindling for domestic fuel or industrial fires. Only limited quantities of cereals were recovered and included spelt and emmer wheat, possible emmer and six-row barley (probably hulled). This suggests local small-scale cereal processing (Miller 2015, 109). It is likely that the grains are residual from accidental loss during the parching stage of cereal processing, a process which by the Roman period was often carried out in purpose-built cereal drying kilns. Wheat generally requires better soils for cultivation in comparison to other cereals and may signify the presence of such here. However, it is also possible the grain for the garrison and vicus to have been transported from other locations rather than grown in the immediate area. The assemblage also indicates consumption of bread, and perhaps beer and potage, although it is too small to determine dietary preferences. However, the first major deposits of carbonised bread type wheat are found in the Roman period, for example at South Shields (Carruthers 1993).

At the Chester Business Park site, although poorly preserved, the presence of a ditch system suggested control of livestock movement between enclosures. The site was notable for good preservation of environmental remains (mainly macrobotanical and insect remains, indicating presence of animal dung) from three waterholes or wells, dating from the Iron Age through to the Roman period. The ditched enclosure and trackways are a possible indication of a pastoral economy or livestock management on the outskirts of the fortress, of a type paralleled at Metchley in Birmingham (Jones 2011; 2012) and in the upper Thames valley (Booth 2016).

At Jersey Way, Middlewich, the land was initially used for both arable and pastoral farming but a reduction in grazing indicators and increase in crop growth in the pollen evidence suggests subsequent less intensive emphasis on pastoral farming. Charcoal analysis indicated a preference for oak and ash as fuel in the salt-manufacturing process (Zant 2016).

Settlement and land-use

Defence, Warfare and Military activity: Military sites

In Cheshire a number of reports have been published since 2006 on the legionary fortress and associated settlement at Chester (Ch). They include the excavations at 25 Bridge Street in 2001 (Garner et al 2008), within the western and southern extramural settlements undertaken between 1964 and 1989 (Ward et al 2012), and the report on at Gorsestacks immediately to the north of the Roman fortress between 2005 and 2008 (Cuttler et al 2012). Publications are also being prepared for Roman activity found at the former Odeon site on Hunter Street and the eastern part of Gorsestacks. The Hunter Street excavations revealed robbed out Roman walls close to the centre of the Roman fortress. Several phases of road metalling were found with a pair of drains and there were signs of a side road or alley leading north off the main road. An earlier timber phase of construction on the north side of the road was identified, and later sandstone rubble wall had cut through it and sill-beam features. Pottery dated from the late 1st to 4th centuries (Wilson 2015, 309). The first of the reports on the excavations carried out on the site of Chester Amphitheatre between 2003 and 2005, dealing with the evidence for the Roman period, has now been published (Wilmott and Garner 2017).

A number of sites have been investigated within the Roman fortress itself although are only available as grey literature reports, including the former George Street Centre (Dodd 2016) which revealed evidence of Roman clay extraction and burial, Bollands Court (Dodds 2015) where a complex sequence of Roman deposits were recorded, the Weaver Street car park (Garner 2015) where the remains of Roman granaries were preserved in situ, and the west end of Hunter Street (Cresswell and Daffern 2015) where the remains of the western fortress defences and intramural structures were identified and preserved in situ. Notable extramural sites were investigated also to the north at Tower Wharf (Towle 2013) and nos 51-56 Upper Northgate Street (Poole 2013) where Roman ditches, a burial, and wells were recognised.

excavations at the Joint Service Cente, The Wiend, Wigan
Joint Service Cente, The Wiend, Wigan, Greater Manchester (courtesy of GMAAS)

A new insight into the character and status of the long-known site at Wigan (GM) has come from a major programme of archaeological excavation for the Grand Arcade shopping development off Millgate, Wigan (2004-5) (Zant 2008; Miller and Aldridge 2011). The Ship Yard site produced the first identified remains of a Roman fort at Wigan. Waterlogged deposits in one of two V-shaped ditch included numerous wooden pegs of oak, typical of those used by Roman army for tents, and probably dated to the late 1st century (Miller and Aldridge 2011, 28-30). The second and larger excavation at Stairgate yielded a substantial stone-walled Roman building containing three hypocaust chambers), a cold bath, dressing areas and colonnade (with stone column fragment). Externally there were rubbish pits and remains of an aqueduct. The excavators concluded this was a bath house for a cavalry fort and dated late 1st century AD to around AD 160. A little 3rd-century pottery was recovered from this site, suggesting the site was not completely abandoned.

Further evidence for the fort was found at the Joint Service Centre on the west side of Millgate. The previous interpretation of timber workshops was re-assessed and it was concluded that the structural evidence related to a late 1st century AD Roman barrack blocks, thus lying within the Roman fort. After demolition in the early 2nd century, a series of iron-working hearths was constructed, used for iron smithing. There is no sign of 3rd or 4th century Roman activity on this site (Zant 2011; Miller and Aldridge 2011). The site has been partially published, with a monograph planned.

At Castleshaw Roman Fort, Oldham (GM), the site was re-excavated and showed that there were no defensive ditches on the east side of the fort, even though a rampart was present. A possible road and evidence for structures outside the rampart suggest the presence of a previously unrecognised annexe. Within the fort, the archaeological remains were multi-phased with evidence for timber buildings, roads and drains, hearths, workshop floors and industrial waste pits. It would appear that, following abandonment, parts of the northern half of the former fort were re-used in the later fortlet phase. The east gate was exposed, and a section of road was recorded where it left the gateway. There were at least two phases of Roman road, the first belonging to the timber and turf fort dated AD 79. The gate structure was supported by large oak posts set in substantial post-pits (Wilson 2015, 306-8). A Conservation Management Plan was produced for the scheduled monument (see; Nash et al 2014; Wilson 2015).

An aerial photograph of the remains of Castleshaw Roman Fort being excavated.
Castleshaw Roman Fort 2014 excavations, Oldham, Greater Manchester (courtesy of Friends of Castleshaw Roman Forts and SUAVE Aerial Photography)

At Vicarage Fields and Quay Meadows, Lancaster (L), geophysical survey and community excavations have been directed by the ‘Beyond the Castle’ project. The Quay Meadows work revealed what may be Roman riverside activity and a well within the NE wall of the fort (link to project).

In Cumbria a large body of new work over the last decade has consisted of extensive geophysical surveys at military sites, shedding new light on the extent and layout of the extra-mural settlements and some fort interiors. At the coastal fort of Ravenglass (C) a community archaeology project was complemented by a magnetometry and resistivity survey which indicated that more of interior of fort survives than previously supposed. Four barrack blocks (or four pairs of smaller blocks) which occupied the east end of the fort were identified, complementing the six already known from previous excavations on the west side. A possible principia was also identified (Adcock 2013). This arrangement may be explained by division of fort into two, with four cavalry turmae housed in the east section and six infantry cohorts housed in the west. This is typical for an equitata cohort (standard infantry with additional cavalry) and is supported by the Ravenglass diploma of AD 158 which records the equitata cohort here. The barracks alignment suggests that the front of the fort faced west towards the sea, which supports the 1970s conclusions regarding the north gate and via principalis but it is contradicted by earthwork evidence which shows that north and south gates lay towards the east end of the fort (Hunter-Mann 2015).

Geophysical survey at Ambleside fort (C) identified a number of features including internal structural remains, and external features, such as former defences. The outline of the fort is visible, along with some internal structural elements, including the via principalis and two of the corner turrets. Some defences and ramparts are also visible. Externally, the outer defences are particularly clear in the resistivity data, especially at the north-east corner. Settlement evidence may be present, particularly in the field to the north of the fort beyond the outcropping rock, but the nature of the responses makes interpretation difficult and conjectural. A feature of particular interest is a linear response with wide halo just to the north of the fort platform, suggested to be a buried structure. There is evidence for the line of the road out of the fort through Borrans Park. There may also be evidence for settlement in this area, although the data is complex and there are no obvious patterns (Taylor 2013).

At the auxiliary fort of Maryport (C)with its extensive adjacent civil settlementnorth and north-east of the fort,several excavations have taken place following geophysical survey. One was a community archaeology excavation aimed at answering a series of academic research questions (Zant and Rowland 2015)Geophysical survey between 2000-2004, and again in 2010, on the fort and extra-mural settlement revealed a highly detailed plan of the settlement and showed a series of regular building plots extending for several hundred metres on either side of the road leading from the north-east fort gate (Biggins and Taylor 2004; Biggins et al 2011, Fig 2).

Excavations to the immediate north of the Roman fort at Stanwix (C) recorded a cobbled area which might be a parade area or possibly a market (Martin 2010). A military-type ditch, running north-south, was superseded by an extensive cobbled surface, laid in the 2nd century or later, which produced a total of seven Roman coins dated from AD 119-337, with a well nearby. A rectilinear timber building appears to have post-dated the ditch, as did the extensive cobbled surface, which was interpreted as a possible parade ground or market area, although the suggestion of an area ‘for quartering animals before procurement to various parts of the Roman military’ is attractive, given its location close to the putative line of road through the fort at Stanwix and the Wall, and access point to the major urban centre of Carlisle. The cobbled area to have remained in use from the 2nd to at least the mid-4th century.

At Low Borrowbridge (C) geophysical survey by OAN in 2015 revealed in detail internal features of the visible fort including the principia (headquarters building), praetorium (commanding officer’s house) and a possible granary (Wegiel and Zant 2016). Resistivity and magnetometer surveys to the south and south-east of the fort have identified ditched enclosures.

Despite the substantial nature and distinctive signature of many Roman military sites, new sites have continued to come to light since 2006, notably the relatively slight remains of short-lived temporary camps. In the north of the region a newly discovered temporary Roman camp at Castle Lane, Castlerigg (C) wasidentified by geophysical survey and trial excavation through one of its ditches, and was thought to be a broadly 1st-century marching camp (Graham 2010).

Geophysical survey at Caermote, Bewaldeth (Cu) suggests an initial larger fort layout with four gates and four roads. The survey has also revealed a road with apparent built structures on either side running north-south in the smaller fort and a possible corner tower in the south west (Graham 2008). Examination of LiDAR data has revealed several new sites, including a possible temporary camp at Bewcastle (C) (Wilson 2016, 303, fig. 11), and a possible Roman fortlet and road NE of Farndon (Ch) (Hardwick 2017, 36). The existence of a possible fort at Wayoh Bridge, near Bolton (GM), was also confirmed by LiDAR (Wilson 2016, 317).

Rural settlement and Landuse

The New Visions of the Romano-British Countryside project aims to ‘allow us to assess the integration of settlements in different parts of Britain with the Roman provincial economy and provide a new characterisation of the Romano-British countryside’. The emphasis is placed on the contribution of commercial archaeological work in advance of development, in particular that only reported in grey literature, to the understanding of the Roman British countryside ( The creation of a comprehensive national database of over 2500 excavated rural settlements embracing England and Wales enables systematic and quantifiable comparability both within and between regions, although partly due to a paucity of excavated rural sites in northern England, the survey classed extra-mural settlements outside forts as ‘rural’. Publication of this major national survey (Allen et al. 2015), remains in progress at the time of writing and will result in an important series of volumes (Smith et al. 2016; Fulford and Holbrook in prep) and journal articles.

Within the project, the North West falls uncomfortably into two zones, the southern part within the ‘Central West’, which embraces not only Merseyside, Greater Manchester and Cheshire but extends overall from West Yorkshire and Derbyshire to Herefordshire. Further north, Lancashire and Cumbria together form part of a trans-Pennine ‘North’. In a key finding for our region, the survey demonstrates the uneven geographical coverage, with much of the ‘North’ and ‘Central West’ markedly under-represented for investigated Romano-British rural settlements.

In our part of the ‘Central West’ region the overwhelmingly predominant rural settlement form is the discrete enclosure, with few settlements of more complex form. Villas are very uncommon and modest in size, and many rural settlements retain the traditional Iron Age roundhouse form, with a movement towards rectilinear structures from the 2nd century AD. Nucleated settlements, set along arterial roads, contain a much higher proportion of rectangular buildings (e.g. Wilderspool (Ch) with no fewer than 36 examples; Smith et al. 2016, 294).

In the ‘North’, Lancashire and Cumbria rural settlements are often well preserved as earthworks especially in upland areas, but suffer from poor chronological control, with few datable artefacts (Smith et al. 2016, 308-322), placing an emphasis on extensive and rigorous programmes of radiocarbon dating to establish occupation sequences. Relatively low development pressures from building or agriculture have led to few opportunities for excavation. Previous assumptions about settlement morphology and dating, in particular that rectilinear settlements are Romano-British in date, have proved unreliable (e.g. Knight et al. 2012, 29). Villas and civilian towns and villages typical of southern Britain are all absent, while nucleated settlements are associated with almost all the permanent forts of the north.

For the North West, against the backdrop of these generalised patterns, excavation and fieldwork are beginning to reveal a more nuanced view of complexity within rural settlement. The consistent pattern of an absence of recognisable building remains from some rural sites, often with apparently blank enclosure interiors, may indicate that some enclosures were intended not for habitation but some function such as livestock enclosures. However, there may have been forms of building either with very shallow foundations or timbers set on the former ground surface of which traces will survive only in the most favourable circumstances. Great care will be needed to recognise these ephemeral features. There are different settlement types in the countryside, with growing evidence of a mixed population of diverse origins. For example, Court Farm, Halewood (GM) (excavated in 1996 but not yet published) is a rare example in the region of a small unenclosed nucleated settlement, with oval buildings of a type which has a restricted distribution along the Mersey estuary. Intra-regional patterns of artefact assemblages reveal differing levels of engagement with the Roman economy, with the hinterland of the fortress at Chester apparently more integrated than the civitas capital at Wroxeter to the south (Smith et al. 2016, 300-306). Amongst the main characteristics of rural settlements north of Mersey are the continued preference for circular buildings and very low levels of artefact assemblages. Coins are absent from the sites, and other finds are very sparse. Rural pottery assemblages in the NW were dominated by jars, and some mortaria, with very few jugs or cups. The use of locally produced wares as well as a frequent preference for BB1 ware – durable cooking vessels, thick hand-made vessels –reinforces the impression that we are seeing here the descendants of the native Iron Age population, who have only slight interest in adopting Roman practices and manners as displayed through material culture.

The principal rural sites or landscapes investigated in the last decade are considered next, moving geographically from south to north. Investigation of the remains of Roman rural settlement close to Chester has been rare until recently. Now, significant work in the hinterland of Chester has shown the extent of a centrally planned and organised landscape up to 3km from the fortress. Two main sites are the Roman settlement and associated field system at Saighton Camp, Huntingdon (Ch) and an area at Chester Business Park (Ch) south 3km of the fortress (Wood 2016; Holgate and Lightfoot in prep.).

At Saighton Camp, archaeological investigation has revealed part of an extensive and complex Roman period settlement, divided into a number of enclosures. Within the enclosures, traces of several structures were identified, although only one building could be defined with any degree of certainty. Located at the northern edge of the settlement, it consisted of a rectangular post-built structure, 14m by 6.5m in size and made up of ten roughly-paired posts, with a possibly internal partition and further smaller posts possibly associated with it beyond its northern side. It straddled a ditch which appeared to mark the northern edge of the recorded sub-rectangular enclosures and so post-dated them. Although its function was unknown, pottery from the building gives a late 4th-century date for construction, making it the latest identified feature (Wood 2016).

The site displays extensive evidence for use of Romanised ways of food preparation and eating, and for buildings using a range of stone and ceramic building materials. In addition, a number of items (including two stone altars, together with ceramics) suggest religious practices and ceremonies were carried out here. Assessment of the recovered artefacts suggests that the site was occupied during the Roman period from the late 1st century until the late 4th century AD, and possibly beyond.

The types of overtly ‘Roman’ materials used or consumed on the site, and the possible religious activity mark out the settlement as unusual and significant. The excavations at Saighton were therefore an important opportunity to add to existing knowledge of the range and diversity of settlement close to the important Roman site at Chester (Deva) (Wood 2016). Since the publication, the discovery through LiDaR of a Roman fort at Saighton provides a context for the unusual finds assemblage.

Chester Business Park, excavated by Network Archaeology in 2003, revealed an area of landscape 3km south of the legionary fortress of Chester. Occupation began in the Iron Age, which was subsequently cut by a co-axial ditched field system, which appears to have been laid out in the early Roman period. A metalled Roman road ran east-west along the southern edge of the site. Finds of Romano-British date were sparse including five lead weights, a jet bead, two melon beads, a hone, a few nails, a fragment of quern and part of an inscription, but no brooches or coins (Holgate and Lightfoot in prep). Pottery dated from the late 1st to late 3rd or early 4th century, with the emphasis on the 2nd century.

These sites indicate the nature and extent of land division in the vicinity of the fortress; probably laid out in the late 1st century and displaying a pattern of formal co-axial land boundaries, served by a wide trackway which may imply provision of grazing along the verges (S Stallibrass pers. comm.).

Elsewhere in the southern part of the region, discrete enclosed rural settlements remain the predominant settlement form and have been examined at a number of locations. Aerial reconnaissance has been moderately successful in Cheshire, identifying some sites through cropmarks. A broad distinction can be made between rural sites with Iron Age antecedents and those which appear to be established de novo. Of the sites which can be seen to occupy, or re-occupy, Iron Age settlement locations, the character of the archaeological deposits and finds assemblages means that continuity of occupation is often hard to distinguish from episodic use. Some settlement locations, such as Irby, appear to have been used repeatedly, if not necessarily continuously from later prehistory into the medieval period. Further examples of discrete enclosure sites extend the geographical extent of excavated examples into Lancashire, which previously had few examples of Romano-British rural sites.

At Poulton (Ch), close to the river Dee, an Iron Age site which has produced radiocarbon dates from the 4th-1st centuries BC also saw activity in the Roman period. There is evidence of field boundaries, enclosure ditches, and small-scale industrial activity, along with a 1st-4th century AD ceramic assemblage which includes local and imported wares. Building material and small amounts of glass have also been recovered but as yet no intact structural remains (K. Cootes pers. comm.). Although the main settlement focus in the Roman period has not yet been identified at Poulton, the character of the assemblage and building materials suggests a possible veteran settlement (J. Axworthy pers. comm.).

A rural site at Irby, Wirral (M) excavated 1987-1996, was finally published in 2010 (Philpott and Adams 2010). An Iron Age phase produced a substantial assemblage of Cheshire VCP, as well as features with mid Iron Age radiocarbon dates, but contemporary structures were difficult to identify. Romano-British occupation followed from the late 1st or early 2nd century through to the late 4th century and probably beyond. A palisaded enclosure was succeeded by a ditched enclosure, to which a second ditched enclosure was then appended. The earliest structures were roundhouses, but in the later Roman period rectilinear and subrectangular buildings were constructed within the interior. The intensity of occupation, which resulted in a dense mass of post-holes, rendered the identification and dating of individual structures problematic.

A rural site at Town Farm Quarry, Norley which was excavated in 2003 (Cooper and Speed 2009) consisted of two areas of Roman features. The first to the north-east consisted of an enclosure defined by slots, ditches and lines of pits, with a concentration of pits within the enclosure, and clusters of post-holes which probably belonged to one small rectangular building, and a line of post-holes probably from one side of a large structure, the rest of which had been removed by quarrying. Associated Roman pottery was confined to the mid-late 2nd century. To the south-west of this concentration was another set of ditches on the same alignment, forming an enclosure with entrance, and part of another, so despite the lack of finds, a broadly contemporary date was postulated. Small amounts of hammerscale indicate smithing in the vicinity, but no other metal finds were recovered.

A prehistoric and Roman settlement at Southworth Quarry, Winwick wasevaluated in 1993, and excavated in 2003 and 2013 (Philpott et al. 1993; Cowell 2010; Moore et al 2014). The finds suggested that the enclosure was occupied in the mid-2nd century, but internal features were few and heavily plough-damaged. The enclosure ditch had been recut in some sections. Finds in both the original and recut ditch dated to the mid-2nd century indicating a relatively short occupation. Pottery from the ditch included samian ware, BB1 from Dorset, Severn Valley ware and local oxidised wares; a fragment of South Spanish amphora was found. Parallel with the enclosure ditch were a number of stakeholes and possible postholes. Several internal features were noted, irregular pits, post-holes, a pit with probable wood or wicker lining, and linear features; the latter probably formed a truncated structure, while a section of curvilinear gully is possibly a heavily truncated ring ditch (Moore et al 2014, 23). Overall, the finds assemblage was very small, with a few metal finds and some possible iron slag. Plant remains were heavily dominated by oats (grains, awns and floret bases) with only two barley grains and one wheat (hexaploid bread wheat type). Wild plants were present in small quantities, notably possible wild turnip. Southworth, as Norley, had suffered heavily from plough damage, which had removed all occupation deposits apart from the remains of deeper cut features. Both sites, as with many others in the region on clay or sands and gravels, also had poor survival of bone.

Previously unknown Romano-British remains were uncovered at Congleton Road in Sandbach (Ch). They consisted of enclosure ditches, and limited evidence of a structure. There were the possible remains of a sheep race and environmental evidence retrieved from the ditches suggested a local economy which included animal husbandry. The relatively small quantity of Romano‐British artefacts from the site, would suggest that the main focus of settlement was elsewhere (Wardell Armstrong 2015).

excavation of a sub-circular feature at Port Salford, Greater Manchester
Sub-circular feature at Port Salford, Greater Manchester (courtesy of Salford Archaeology)

A Romano-British farmstead on a sandy promontory above the river Irwell at Port Salford (GM) was examined in 2008 and again in 2012‐13. A network of gullies and ditches formed a series of small rectangular enclosures, each roughly 20m by 45m, with a trackway at the western end of the site and possible remains of roundhouses. Artefacts were scarce from this settlement, suggesting that the main focus of the farmstead may have been to the north. Finds included a glass bead, a fragment of a rotary quern and a small number of Roman pottery sherds, including a Black Burnished Ware jar, a Grey Ware jar manufactured in Cheshire and fragments of a mortarium, from kilns in Cheshire (perhaps Warrington). These items dated to the 2nd to 4th century AD and may have been acquired at the civilian settlement outside the Roman fort at Castlefield in Manchester.

On the upland margin, the Mellor Heritage Project (GM) produced evidence for the movement of goods east to west across and outside the region, with traded pottery from Derbyshire and the Cheshire Plains. The finds suggest occupation from the 1st through to the 4th century – the only settlement in Greater Manchester to be occupied right through the Roman period. Within the Iron Age enclosed settlement, with its substantial enclosure ditch and series of intercutting Iron Age roundhouses, Romano-British period structures are absent, although the quantity and date of the artefacts indicates prolonged occupation, with Roman coins from 1st to mid 3rd century, several brooches, and pottery including Derbyshire Ware, Cheshire Plains oxidised ware, BB1 and samian. The location of the site on the upland margin enabled occupants to acquire ceramics from east and west of the Pennines (Leary 2005). Small quantities of Derbyshire ware have been recorded at Middlewich and Warrington but the core distribution area is east of the Pennines (Connelly 2005, 102; Hearle 2011).

Mellor Iron Age enclosure ditch with Roman pottery in upper fill, Stockport, Greater Manchester (courtesy of GMAAS)

Small-scale trial excavation in 2010 at Burton, Wirral (Ch) produced evidence of Roman activity within a circular enclosure, with Black-Burnished 1, late 3rd-mid 4th century Mancetter–Hartshill mortarium, and an undated sherd of prehistoric pottery in a primary fill (Philpott pers. comm.). Less than 200m to the north a subrectangular enclosure was excavated producing a clay oven and at least three structures. The pottery assemblage suggested two occupation phases in the late 1st-2nd century and mid-3rd to 4th century. Other finds were sparse, with very little metalwork, two quern fragments and two glass beads. The two enclosures are unusually close; the curvilinear form of the southern enclosure, together with fragment of prehistoric pottery in the ditch, argue for a pre-Roman Iron Age date. A Roman date for subrectangular forms of enclosure has been suggested in the lowland NW of England and may in this case relate to intensification of agriculture in the zone around the fortress at Chester. The hints of prehistoric activity on the northern enclosure site, such as the presence of VCP, may predate the construction of the enclosure itself.

Metal detecting over many years at a site in South Wirral (location confidential) has produced a large assemblage of metal finds, including at least 20 brooches, 10 coins, a stylus, 32 lead spindle-whorls; and other ‘Roman’ metalwork, with a hint of late prehistoric origin from a single cup-headed pin of Iron Age date. Such an assemblage represents a rural settlement well integrated into the local Roman economy, perhaps supplying the legionary fortress with textiles. Such a site may be a candidate for occupation by veteran or entrepreneurial Romano-British inhabitants perhaps within the legionary prata (cf. Mason 1988).

Few securely dated Romano-British rural settlements have been discovered or investigated in Lancashire. One exception is Barker House Farm, Lancaster on the SW Campus of Lancaster University, excavated by OAN in 2003 (OAN 2004) (and due for publication, Zant and Bagwell in prep) on a low promontory overlooking the river Wyre (348297 456500). A circular ring gully 12m in diameter containing a 9m diameter ring of ten post-holes had an entrance to the east; to the south of this was a circular ‘ditched enclosure’ consisting of two segments of ditch with opposed 5m wide entrances. Three phases of gully were identified, while a pit which cut the gully near the terminal contained a smithing hearth bottom. There were also linear arrangements of postholes interpreted as fence lines. The farmstead was bounded by a ditch 38m to the west beyond which was a large water hole. Some further Romano-British activity occurred outside the main excavation area. Finds were very scarce. The first building had a single Romano-British sherd from a posthole (206), and a fragment of beehive quern was recovered from a pit within the building. The circular enclosure produced another beehive quern fragment with a central large posthole and is interpreted as a possible stock enclosure. The boundary ditch has C14 dates from the 1st-3rd century AD, with only a tiny assemblage of five sherds of Roman pottery (one BB1, three grey ware, one amphora) despite its proximity to the Roman fort at Lancaster. The site once again was plough-damaged and all occupation surfaces were lost (OAN 2004; Zant and Bagwell in prep).

On the Wyre Estuary Pipeline/Garstang Road East, Poulton-le-Fylde, a settlement site with several ring-gullies dated to the Romano-British period was excavated in 2008-9 (Smith et al. 2016, 317), and further Romano-British remains were discovered in the area in the adjacent field in 2014 (WA 2014), the latter consisting of ditches (possibly the continuation of an enclosure ditch found earlier), and ring-gullies (including one probable roundhouse) associated with the Romano-British settlement found previously. A further possible ring-ditch on the hill in the NE corner of the site may belong to a separate Romano-British settlement. No finds associated with the Romano-British features were recovered in limited excavations in the 2014 site.

Dutton’s Farm, Lathom, near Ormskirk continued to be investigated after an interim publication in 2005 (Cowell 2005). At least four roundhouses were found, one dated to the 1st century BC and another smaller roundhouse had Romano-British pottery in the gully representing a shift in the settlement focus. About 100m to the west there are several inter-cutting broad trackways for local estate or agricultural use rather than public roads, one of which had a small early 2nd century coin hoard (closing with Hadrian) with tile and Romano-British pottery.It is hard to be sure of the total sequence of intercutting tracks, though the others were not as substantial as the one with Roman material in it. The presumed 2nd-3rd century farm indicated by pottery and other finds was not located. There are thin scatters of Romano-British pottery elsewhere in the enclosure that may pinpoint it. The small Roman ceramic assemblage of only 98 sherds by 2005, consisted of BB1, Oxfordshire ware, local sandy wares, probably from Wilderspool, and ceramic tile (Cowell 2005, 70).

Romano-British rural settlements, and their Iron Age predecessors, in some upland areas have proved elusive, but extensive surveys suggest these locations may genuinely have only sparse settlement. Thus, in the Western Lake District, the National Mapping Programme has identified few monuments that can be attributed to the Iron Age or Roman period within the aerial photography and LiDAR project, particularly in comparison with the earlier periods, and surprising given the proximity of Glannoventa, the Roman fort at Ravenglass. This dearth of Romano-British sites does, however, reflect the existing archaeological record (Deegan 2016, 11).

Targeted research projects aimed at understanding the chronology of poorly dated types of site have been undertaken in upland regions. The Matterdale Archaeological Project, a programme of site reconnaissance and excavation taking place in the area around Glencoyne Park and Matterdale, near Ullswater, in the Lake District National Park, focussed on improving the archaeological record by identifying Bronze and Iron Age monuments, as well as investigating Iron Age – Roman interaction. In the north-eastern Lake District a report prepared in 2013 on archaeological work in 1997 and 1998 at Baldhowend, Matterdale (C) aimed at understanding an unenclosed settlement, a type of site little researched or understood in Cumbria (Hoaen and Loney 2013). A topographic and geophysical survey at Baldhowend Farm (Hoaen and Loney 1999a, b; Loney and Hoaen 2000) found remains which were interpreted as a small, possibly unenclosed, settlement, with an adjacent field system and hollow-way. At the northern edge of the site this field system appears to overlie an earlier phase of activity, perhaps consisting of two hut circles. Selected structures of the ancient settlement were investigated by excavation, including a rectilinear building with evidence of Romano-British metalworking, confirmed by radiocarbon dating, while also recovering environmental data relating to the land use and economy of the inhabitants, to provide a dated sequence of landscape and settlement change for the late prehistoric period to immediate post-Roman period. The investigations at Baldhowend formed part of a wider programme of work in Matterdale (Loney and Hoaen 2000; 2007; Brennand et al. 2006; Quartermaine and Leech 2012, 20).

Investigation of an enclosed settlement at Glencoyne Park (C) showed through radiocarbon dating that occupation began at the start of the first millennium BC. The enclosure wall was constructed in at least two phases and with two different building techniques with a phase of rebuilding at the start of the Roman period in Cumbria. The first phase is dated to some time after the start of the first millennium BC. The second phase is not yet dated but appears to have fallen into disuse before the Roman period. Geophysical evidence revealed external buildings and courts which indicates that the enclosure wall went out of use. Internally some of the range of cut features probably belong to the latest phase before the Roman period. The Roman period sees extensive construction and reconstruction, including a gravel platform for the construction of a stone footed roundhouse, a cobble courtyard and a flagged area. This phase appears to have lasted for some considerable time as the house walls appear to be remodelled on several occasions and the entrance orientation altered (Loney and Hoaen 2010).

The Matterdale project revealed substantial earthworks forming parts of field systems, some at least associated with a known late prehistoric/Romano-British settlement, but others less easy to date. The remains identified represent extensive and significant evidence for activity in the valley from at least the late prehistoric period onwards, possibly continuous. It demonstrates the extent to which these remote valleys were occupied, and it fits well with evidence for large-scale settlement in Glencoyne and elsewhere (Greenlane Archaeology 2013).

The settlement at Gossipgate near Alston (C) is the best-preserved of about 20 ‘native’ settlements in the vicinity of Whitley Castle Roman fort (3.5km to the north west), identified by English Heritage’s ‘Miner-Farmer landscapes of the North Pennines AONB’ project (Knight et al 2012; Wilson 2011, 344). In 2010, a detailed field survey was undertaken which revealed an agglomerated settlement of about 18 irregular compounds, each containing evidence of several circular buildings, most facing into a central yard. The settlement is surrounded by a field system of former arable plots, including cultivation terraces so narrow they can only have been worked by hand, while the site is overlooked by enclosed pasture. Dating evidence is sparse, but a beehive quern built into a field wall could indicate settlement before 50 BC, while two Roman pottery sherds from molehills within house sites indicate occupation lasted into the 2nd or 3rd century (Wilson 2011, 344). The complex and accreted nature of the settlement indicates development over an extended period.

At Maryport (C) a square enclosure, 60 by 60 m, and probably a farmstead serving the Roman fort of Alauna 250m to the north was revealed (Wilson 2011, 345) in association with Roman pottery of 2nd and 3rd-century date. No associated field boundaries were identified.

Other major settlements

Cheshire’s major Roman settlements have also seen significant fieldwork and publication of the results since 2006. Publication and synthesis of excavations carried out prior to 2006 has also been achieved. In Nantwich (Kingsley Fields), the excavation of the extensive industrial settlement by the University of Manchester Archaeological Unit in 2002 has now been published (Arrowsmith and Power 2012). Excavations revealed a previously unknown Roman road at Kingsley Fields, linking the settlement to the main Roman road network. Along the road, evidence for the collection and storage of brine and production of salt, with a series of enclosures, buildings, a well and a cluster of cremation burials; waterlogged conditions provided high quality preservation of organic materials including timber tanks. Development along the road began the Hadrianic period and intensified in the later 2nd century, ending in the AD 180s, before a second major phase began in the early 3rd century and ending largely by the mid-3rd with only small-scale brine collection continuing in the late 3rd and 4th centuries.

The results of the archaeological work in Middlewich (Ch) from the 1960s onwards have recently appeared in a synthetic article (Garner and Reid 2012), and reports on excavations at King Street in 2001 and Jersey Way in 2013 have now been published (Williams and Reid 2008; Zant 2016). The auxiliary fort formed the core of settlement from AD 70-130, probably for a cohors quingenaria peditata of 480 men, though the unusually high proportion of silver coins led Shotter to postulate a legionary detachment. The extra-mural settlement extended for up to 20 ha and was laid out in well-defined ditched plots, which extend into agricultural landscape. The buildings are strip in nature, many post-in-trench, with some post-in-the-ground. Occupation in the extra-mural settlement diminished after 200 but occupation continues until at least the 360s. Salt extraction was a key industry, and tanks and heaths and ovens for drying the brine have been recovered, but other industries include iron smithing.

Middlewich timber brine tank with well-preserved artefacts, Cheshire (courtesy of Oxford Archaeology Ltd)

At Stockton Heath (Ch) excavations in 2007 on the site of the Roman road and associated ribbon development to the south of the industrial settlement at Wilderspool have been published (Dodd et al 2010; Rogers and Garner 2007). Excavations since the 1930s have enabled the progressive understanding of a section of the Roman industrial settlement. At Loushers Lane, Wilderspool a previous excavation had identified a possible suburban villa. However later excavations, including in 2012, suggested the presence of several Roman stone buildings. In 1976, excavations to the east of the site revealed an east-west lane serving a series of plots defined by ditches, which was interpreted as an ‘apparently unplanned’ ribbon development of the 2nd and 3rd centuries (Wilson 2015, 311). The adjacent area to the north was investigated in 2014, revealing an east-west ditch, probably a boundary parallel to the lane found in 1976. Pottery from the ditch fills included late Roman shell-tempered ware, Huntcliff-type ware and Crambeck ware, suggesting infilling in the late 4th century. The presence of quantities of building materials appeared to derive from the large Roman stone building, suggesting it was ruinous, falling into disrepair in the 4th century. A small clay oven and sandstone wall foundations were found on the eastern edge of the site. The 4th-century occupation was interpreted as agricultural/domestic occupation, including cooking jars, bowls and mortaria. A stone spindle whorl suggested textile processing and hammerscale indicated blacksmithing. A low yield of charred cereal grains from the ditch fill suggested crop-processing (Wilson 2015, 311).

The work provides confirmation that occupation of Wilderspool continued into the late 4th century at least on this part of the settlement, although the usual caution should be applied to the end-date of occupation since there is no certainty that occupation ended when the latest pottery was deposited.

Civilian settlements attached to military sites

A national research synthesis of extra-mural sites has recently been completed which brings together a wide range of different types of evidence (from epigraphic to geophysical) to consider activities and functions of the extra-mural spaces and the people living in them (Murphy-Smith 2016).

Excavations on five areas within the Roman vicus at Manchester (Mamucium) during 2001-2005 have been published (Gregory 2007). The excavations revealed a ditch, possibly of military origin and defining an irregular military annexe, with an early vicus protected by a palisade and ditch further to the north, flanking either side of the Ribchester road. The ditches of both the annexe and vicus were infilled in the early 2nd century and the vicus expanded over the area. The road was then lined with a series of timber strip buildings. There was a building which may have served as a mansio, and a hybrid classical/Romano-Celtic temple. Industrial activities included smithing and tanning. There were also a series of field and allotment boundaries, with possible evidence of vegetable cultivation (Carrington 2011, 107). The vicus appears to have been abandoned in the mid-3rd century, although the fort may have remained in occupation later.

Chester Road, Manchester, was excavated in 2008, on the opposite bank of the river Medlock to the Roman fort and vicus, revealing evidence of a ditched enclosure, pits and agricultural ditches (PCA 2009), and adding considerably to our knowledge of the extent and character of the Roman settlement. The earliest Roman activity of early to mid-2nd century AD saw a series of boundary ditches delimiting small, regular plots of land set out to the south of the Roman road. The plots contained probable quarry pits, filled with refuse, although one yielded a well-preserved sandstone altar, dedicated by Aelius Victor to the mother goddesses of a German tribe known to have provided auxiliary units for the Roman army. The altar was probably set up as a roadside shrine and was presumably disposed of in the pit when obsolete. Antiquarian discoveries in this area have long indicated that the southwestern approach to the Roman fort and associated settlement had particular religious significance. The extramural settlement attached to the Roman fort extended beyond the River Medlock by this time, indicated by the presence of domestic refuse.

Chester Road excavations, Manchester (courtesy of GMAAS)

In the late 2nd century a change in layout saw a more substantial boundary system with larger ditches, possibly delimiting the south western extent of the extramural settlement beyond the Medlock at this time. In the early-mid 3rd century smaller boundaries were re-introduced marking plots with traces of possible structural features, such as beam slots and postholes, while refuse pits yielded domestic refuse that presumably originated from nearby habitation. A substantial ditch bounding the south-western side of these plots may have delimited the extent of the extramural settlement at this time.

The excavation yielded a modest sized assemblage of Roman pottery, comprising local wares, Romano-British traded wares and imported material such as samian ware and Spanish amphora. Several ‘small’ finds were also recovered, the majority being of domestic or structural function. Faunal remains and palaeoenvironmental evidence were scarce due to the acidic nature of the subsoil, which is not conducive to the survival of such material (PCA 2009).

As mentioned above, geophysical surveys have revealed, particularly in Cumbria, the layout and extent of the extra-mural settlements, but the limited excavations subsequently carried out show that, inevitably, the surveys focus on the latest phases, they are better at identifying masonry buildings and ditches than relatively ephemeral timber structures, and do not reveal the full complexity of the occupation phases. They have greatly enhanced our understanding of the extents and diversity of extra-mural settlements although several do not reach the furthest extents of the areas of activity and almost none of them put the roads and extra mural settlements and cemeteries into their landscape settings.

At Roman Papcastle (C), a large-scale geophysics and evaluation programme undertaken by Grampus Heritage with the assistance of Wardell Armstrong Archaeology, has uncovered amongst other things a substantial extra-mural settlement, a bathhouse, bridge, and mill (

Roman mill at Papcastle, Cumbria (courtesy of Cumbria County Council)

At Ambleside (C) vicus, geophysical survey identified potential boundary ditches delineating a feature following a course from the lake inland, parts of the fort’s ramparts and ditches, extra-mural settlement and settlement inside the fort. Substantial areas north of the fort appeared blank on the survey, although some features further north may indicate the location of furnaces. There is no conclusive evidence of earlier phases of the fort other than that identified by Collingwood to the north. Responses in the south of the survey area do not exactly match the ramparts reported by Collingwood but they could represent the earlier phase. Other possible features cross the northern ramparts at an oblique angle and these may represent earlier archaeological features. This explanation is not strongly supported by the geophysical evidence, however (Taylor 2013).

Excavations at Ravenglass (C) revealed evidence for an extensive vicus to the east of the fort, mostly in the form of large timber strip buildings alongside roads and streets. These were accompanied by evidence for domestic and industrial activity, including iron working and possibly also lead. One particular building on the slopes east of the vicus may have been associated with metalworking. It was difficult to identify phases of activity within the vicus, although one possible enclosure may pre-date the vicus. Occupation within the vicus appears to have commenced around AD 150, and most pottery was dated to 2nd/3rd century AD. Although there is some late 4th century the usual range of 4th century pottery recovered from this fort is absent from the vicus, suggesting that occupation of this part of the extra-mural settlement had ended by c. 300 AD. The small assemblage suggests that the military vicus conforms to the usual pattern of late 3rd century abandonment, with the fort occupied until the end of the Roman period (Hunter-Mann 2015).

At Brougham (C), pipeline development allowed investigation of the little-known extra-mural to the south-east of the fort, revealing three phases of occupation, beginning in the late 2nd/early 3rd century and continuing into the second half of the 4th century, or possibly recommencing in the late 4th after a period of reduced activity (Zant and Clapperton 2010). An area of ditched field boundaries, associated with trackways, pits and wells or water holes was identified. This was adjacent to a cemetery area where possible cremations and pyre debris were found. Brougham appears to have had an ‘extended vicus’, where settlement was concentrated at several foci with less intensive occupation between, and dispersed over a relatively large area around the fort.

Ritual, Religion and Ceremony

Temples, Altars and other inscriptions

With the exception of burial which is considered separately, new evidence of religious or ritual activity is limited and largely from military contexts. At Chester Road, Manchester a well-preserved gritstone altar found in a rubbish pit also contained a fine Samian bowl depicting a hunting scene dated c. AD 180. The altar was dedicated by Aelius Victor to the continental mother goddesses of the Cananeftis (north Rhineland) (PCA 2009, 68-72). It is thought that the altar once stood next to the main road to Chester but was later moved. The strip of land beside Chester Road, where shrines, temples and mausolea might be expected, had been badly disturbed by 20th century development.

Late 2nd century altar from excavations at Chester Road, Manchester (courtesy of GMAAS)

At Maryport (C) in 2013-2014 two stone temples, one circular, the other square and of classical style, were investigated and found to be contemporary with several altars originally found in 1870. The temples stood within a cobbled enclosure 50m by over 95m, of which three sides were confirmed. Re-examination in 2011-12 of the findspot of 17 2nd-century altars showed that rather than ritual burial, as first interpreted, the altars were re-used as packing in a series of large post-pits. The pits supported a substantial late Roman (or early post-Roman) timber building. It was suggested the altars were originally free-standing within the cobbled enclosure along with a large structure of which substantial foundations were identified. Under the cobbled surface were Roman ditches, interpreted as part of a temporary camp (Wilson 2016, 304).

At Ravenglass fort (C), fragments of one certain and one possible face-jar with general parallels at York and in north-east England and fragments of a pipeclay figurine may suggest the presence of a shrine in the vicinity (Hunter-Mann 2015).

Burial practices

A number of excavations on cemeteries in the last decade have contributed to the poorly understood subject of burial practices in the region. Recovery of data for burial practices is constrained by acidic soils in the North West which are inimical to bone survival. The chief exceptions are in the limestone regions and certain urban or waterlogged contexts where particular soil conditions enable bone to be preserved.

With a few significant exceptions, notably Brougham (C) (Cool 2004) and Low Borrowbridge (C) (Hair and Howard-Davis 1996), where relatively isolated military communities maintained distinctive burial traditions over many decades, the Romano-British cemeteries of North West England have not been extensively investigated or published. The extra-mural cemeteries at Carlisle and Chester have been re-evaluated in the light of new discoveries. A significant addition is the synthesis of information on Lancaster’s cemeteries (Iles and Shotter 2009). Here the publication of research into both historical and more recent discoveries of Roman (and Bronze Age) cremations, has attempted to disentangle the incomplete 19th-century reports of cremations, maps the findspots, assesses the surviving evidence, including antiquarian accounts and unpublished notes, , and reports on recent finds.

A number of features of Lancaster’s cemeteries are found elsewhere in northern England. The use of single urns, often in a black-burnished ware jar, for cremation burials, with any additional vessels apparently deposited on the pyre, appears to be a feature characteristic of burials in the Roman north. Unurned cremations often have only token deposits of cremated human bone rather than the whole collection. Occasional hobnails are present but other furniture is rare.

At the Arla Foods Depot, Lancaster, a rectangular enclosure had been laid out in late 1st or early 2nd century along the Roman road south of Lancaster. Although used for cremation burial, rectilinear enclosures along roads out of forts are known from Carlisle and elsewhere. The partition of the landscape near the fort may have been intended originally for agricultural or pastoral use, but their changing function may have been response to varying pressures on land in the vicinity (UMAU 2007).

Cremation cemeteries have been investigated at several Cumbrian sites. That at Beckfoot (C), c. 350m to the south-west of the auxiliary fort and in the proximity of Milefortlet 15, was subject to trial excavation in 2006 by Oxford Archaeology North in advance of increasing coastal erosion. Three phases of burials were found, from the late 1st/early 2nd to 4th century (Howard-Davis et al. forthcoming 2017). Beckfoot had previously produced unusual evidence of busta, pyre sites, in one case with military equipment and a funeral bed (Caruana 2004). A previously unrecorded cremation cemetery was identified on a flat-topped hill NE of the line of a possible Roman road at Maryport (C); the pottery associated with the cremations dated to the 3rd century AD (Wilson 2011, 345).

At Birdoswald (C) a cremation cemetery was investigated in 2009 following a major landslip. A series of Hadrianic to late 2nd century cremations includes two busta, box burials and bone veneers. Two probable 5th century inhumation graves were found, which blocked the earlier entrance to the burial enclosure and may be contemporary with the post-granary timber hall inside the fort; however, the bone did not survive (Wilmott 2010).

Part of the Roman cemetery on Botchergate, Carlisle was excavated in advance of development. The early Roman features relate to the military expansion of the area, as they appear to predate the formalised route of Botchergate. After the establishment of the Roman road on Botchergate and the Roman civilian settlement, a formal cemetery was created which was in use in the late 1st-2nd century. A total of 19 cremation pits was recovered with the remains of at least 46 individuals, located within organised burial plots along the street frontage. Of particular significance was the discovery of several cremation urns and a large number of accessory vessels (Jackson 2015).

The cemetery appears to have been relatively short‐lived and the function of the site had completely changed by the early‐mid 2nd century AD. It is likely that this change in emphasis at the site related to the continued expansion of the civilian settlement, and probably represents the establishment of Botchergate as a suburb of the Roman town. This phase of activity was unusual in that it was largely represented by two circular buildings, rather than the usual rectangular strip buildings typical within Roman civilian contexts and noted elsewhere along Botchergate. One of these circular buildings also provided evidence of possible small‐scale industrial activity. Whilst the environmental evidence has ruled out metal working, it does suggest that processes such as dyeing or glass working may have been undertaken within the structure. The final phase of significant Roman activity appears to have occurred during the mid‐late 2nd century and possibly into the 3rd century AD. The transition between this and the preceding phase appears to represent a planned change rather than abandonment and subsequent reoccupation of the area, and probably relates to the continued development of the Botchergate suburb. It is probable that the remains dating to this phase represented strip buildings along the street frontage with back plots and a series of enclosed areas located to the rear. The site appears to have been largely abandoned during, or shortly after the early 3rd century AD, which was marked by the presence of accumulated soils blanketing most of the site, suggesting the area reverted to ‘open fields’ on the periphery of the settlement at this time (Jackson 2015).

In the southern part of the region, little new burial evidence has come to light. However, in Cheshire, three cremation groups excavated beside the road at Stockton Heath, Warrington belonged to a roadside settlement that was also associated with pottery kilns (Dodd 2006). At Kingsley Fields, Nantwich a similar roadside location had produced a small group of three cremation burials (Arrowsmith and Power 2012). A re-consideration of the burial evidence in extra-mural areas of the legionary fortress at Chester has led to re-dating of burials at Infirmary Field (Heke 2012).

A preliminary attempt at synthesis onRoman cremation practices for northern England has been published, based largely on sites at Beckfoot and Herd Hill (C), and two outside the NW region at Lincoln and Malton (Thompson et al. 2016). Focusing on the technology of cremation, the study examined the cremation efficiency and pyre technology of the process through Fourier Transform Infrared-Attenuated Total Reflectance (FTIR-ATR) analysis of the cremated bones. The project concluded that the condition of the cremated bone reflected a consistent Roman military practice in northern England in terms of the pyre technology and in the highly selective retention of the human remains.

Unburnt bone survives poorly in many of the region’s soils so inhumations are under-represented in the Roman cemeteries of the north-west. One exception is the remains of at least 28 individuals from the cave at Doghole Cave, Haverbrack (C), probably Cumbria’s largest assemblage of human bone dating from the Romano-British period. Excavations in the 1950s (Benson and Bland 1963) were followed up by Hannah O’Regan (University of Nottingham) as a result of damage to deposits by caving (Wilkinson et al. 2011). Here, although the human bone survived in the limestone environment, post-deposition disturbance had severely disrupted the burials and none remained in situ. Careful retrieval of the remains through sieving of deposits produced a suite of Roman artefacts, including hobnails, and personal ornaments including glass and gold-in-glass beads, copper-alloy and silver bracelets, typical of later Romano-British inhumation burials elsewhere in Britain. An unusual suite of faunal remains is currently undergoing analysis. The identification of such burial rites in a rural location distant from the nearest forts may point to the settlement of former military personnel and their families in the countryside. Planned isotope analysis may contribute further to a study of origins of the individuals (H. O’Regan pers. comm.).

The potential for the application of stable isotope analysis of human bone to the North West is discussed by Stallibrass (2011, 114-116) but such studies are hampered by the poor preservation of bone in the relatively acidic soils common in the region (Smith et al. 2016, 390-1, fig 12.5). Nevertheless, the technique has been applied to a small number of inhumations, with more planned for the future. A sample from a male inhumation recovered from a bathhouse hypocaust at Papcastle in 2012 demonstrated strontium and oxygen isotope compositions that are compatible with a local Lake District origin with a diet typical for Roman Britain (Evans et al 2014). Stable carbon isotope analysis of an inhumation from Leasowe near Meols (M) found that the individual had a diet in which marine protein played only a very low part (Griffiths et al. 2007, 350-351). Isotope analysis is also planned for a sample of the Dog Hole Cave (C) inhumations (H. O’Regan pers. comm.).

Trade, Exchange and Interaction : Industry

The 2006 assessment noted the extensive evidence for industrial production across the region, including the working of iron, lead and copper-alloy, pottery and glass manufacture, and production of leather and salt. The function of the Lancashire and Cheshire nucleated settlements, including Walton-le-Dale, Nantwich, Middlewich and Wilderspool, in supplying the military in the frontier region has a distinct profile of activity, in which markers such as coarse pottery demonstrate a major decline in the early 3rd century and only low level activity into the 4th.

Since 2006, the great majority of evidence for industry during this period derives from metalworking, particularly iron, which was practised across a wide range of Roman settlement sites. Iron working is the most readily identifiable industrial production process apart from pottery manufacture due to its copious waste products. Minor evidence for iron smithing, possibly of Roman date was recovered from Stormy Point, Alderley Edge (Ch). Traces of evidence for possible Romano-British or medieval iron smithing were recovered along with evidence for medieval activity close by, probably at Saddlebole (Mottershead and Wright 2008).

At Ravenglass (C) iron-working waste, including hammerscale and hearth bottoms, was noted in levelling deposits associated with the east-west road, though no in situ smithing activity was identified. High temperature processes are thought to be industrial in origin, including iron working and possibly also lead, but the source is not clear. One building on the slopes east of the extra-mural settlement may have been associated with metalworking (Cubitt 2015).

Following synthesis and publication of major excavations in Cheshire, production of salt has been identified as an important industrial activity, especially in the ‘wich’ towns of Cheshire, e.g. Nantwich and Middlewich (Arrowsmith and Power 2012; Williams and Reid 2008). Limited evidence for other industries, such as glass, tile and copper alloy manufacturing, has also been recovered.

Further evidence for Roman tile manufacture was recovered in 2005-7 from Ochre Brook, Tarbock (M), in the field to the north of the farmstead enclosure which was excavated in 1993 (Cowell and Philpott 2000). A palaeochannel of the Ochre Brook contained fragments of Roman roof tile and metalworking debris showing the channel began to silt up in the Roman period. A further example of the only dated tile stamp from Roman Britain was recovered (Cowell 2012). Adjacent to the channel was a sandy terrace which had extensive evidence of metal working waste, alongside small abraded tile fragments, interpreted as a smith’s disposal area next to the channel, in hollows and a series of intercutting pits, which also included hammerscale and iron smithing waste. Pottery was scarce (13 sherds) and no structures were found. Nearby two Iron Age pits were found 500 m to the SW of the Romano-British farmstead, providing evidence of thinly dispersed activity across the landscape, indicated the earlier use of the landscape.

At Ribchester (L) excavations in 2016 identified a clay floored building with evidence of use as a workshop in the form of a hearth, kiln fragments, slag, and waste from glass working (S. Stallibrass pers. comm.). The date for this activity is currently unknown.

Rural copper-alloy metalworking waste from a rural site in south Wirral was examined by McIntosh and Ponting (2011).

At Whitley Castle, just over the Cumbria border in Tyneside, a detailed earthwork and geophysical survey of the Roman fort, its extra-mural settlement and field systems has concluded that the fort was constructed to oversee the production of lead and silver (Ainsworth and Went 2009).

Trade, Exchange and Interaction:Metal artefacts – coins, brooches and other objects

The two most common types of metal artefact from the North West have been studied in national surveys. The Portable Antiquities Scheme (PAS) has been useful for North-West focused studies, and a number have been conducted over the past ten years which use the data recorded by PAS. Many of these studies have not only noted the intra-regional differences, but also compared the material with patterns observed across other areas of England and Wales.

Walton has examined the distribution of Roman coinage across England with an analysis of changes through time, largely based on the dataset of nearly 58,000 Roman coin finds to March 2008 recorded by the Portable Antiquities Scheme (Walton 2012). (In April 2017 the total stands at 232,397). However, methodological issues emerge from such studies. Walton’s took a base line of 20 coins per parish (Walton 2012, 19-20), but the PAS records have no parishes which qualify in the Cumbria, Lancashire, Greater Manchester or Merseyside and only a handful in Cheshire (Fig. 4). Amalgamating with other data sets of coin finds produces only a small number of parishes above the base-line, a methodology which significantly under-represents the distribution if not the density of coin finds in the North and West of England. Conclusions include the fact that the military north was supplied predominantly with denarii to facilitate payment of the army, with some bronze coinage to enable smaller transactions within the military community (Walton 2012, 55). A cluster of Cheshire Republican and early Imperial coins has been identified around Middlewich and a smaller one around the legionary fortress at Chester (Walton 2012, 65-66, fig. 35). However, the finds distribution in the Roman north which is predominantly rural may indicate rather a conscious preference on the part of the native population for silver coinage, perhaps as one of many prestige goods, a pattern suggested for Scotland (Walton 2012, 55).

Against this national picture, detailed studies flesh out the complexities of local intra-regional patterns to provide a nuanced regional picture of Roman coin use. In the North West, David Shotter’s ongoing research into the Roman coin finds has been published in a series of four volumes, the latest supplement in 2011 (Shotter 2011). Shotter uses historical finds alongside new material recorded through the Portable Antiquities Scheme to focus in detail on the implications of chronological and spatial patterning of coins finds. The analysis of individual site profiles helps determine the origins of settlements and fluctuations in levels of activity, while the pattern of chance or individual finds indicates the varying levels of integration into the market system across the region.

As one index of the degree of integration into the Roman economy (Shotter 2011, 86), the large areas of the rural north and west of England which were either devoid of coinage or show very limited coin loss throughout the Roman period were not strongly monetised in the sense of habitual coin use (Walton 2012, 168). However, the persistence of long-lived patterns which determine native coin use is revealing. There is a striking similarity in the overall distribution of Iron Age coin finds and the dense coin distributions of Roman Britain, a pattern which persists into the early medieval period with sceattas (Walton 2012, 168-9; Fig 120). There are thus strong and persistent patterns within coinage, and wider material culture, from the Iron Age through to the early medieval period, which appear to denote deep-rooted attitudes to wealth and value, and which indicate different economic value systems operating in the western and northern Britain as opposed to eastern and southern regions.

Roman finds recorded on the PAS database by county (total 3880 finds by April 2017)

A distinctive pattern of artefact finds in PAS with a high ratio of brooches to coins, by comparison with other areas of the country, has been highlighted for Cheshire and other West Midland counties (Staffordshire, Shropshire, Herefordshire) (Pearce and Worrell 2016, 363). These were all areas which did not have a tradition of coin manufacture and use in the late Iron Age and were low-level users in the Roman period. These distributions are allied with differences in settlement pattern – an absence of formal temples apart from the forts and nucleated settlements, and an absence of villa estates. Cool and Baxter have pointed out the need to examine the social distribution of finds across different site types (2016, 73), and the biases inherent in the data need to be made explicit. They observe that the rural brooch use in the North is relatively low compared with other regions, and a chronological pattern differs from the rest of Britain in starting later, in the mid –late 1st century AD and peaking later, with sharp decline in brooch use in the early 3rd century and a further decline in mid century (Cool and Baxter 2016, 85-86).

The analyses show that the Cheshire and north Wales assemblages bear a closer similarity to the North region than to the West Midlands (Cool and Baxter 2016, 82, fig. 3). McIntosh and Ponting’s (2014) study of Wirral brooches has identified interesting patterns. First, the type has a core area strongly associated with Wirral and Cheshire extending into NE Wales, where the brooches are found almost exclusively on rural sites; but two other concentrations in the military zone of northern frontier, and native Scottish and northern England sites. The distribution suggests different mechanisms of trade and exchange which resulted in dissemination of the type beyond its assumed core area, through both military networks, and perhaps through contacts between native groups. They have also identified subtle indicators including preferences for certain colours which may convey messages of kinship or group identity which are now difficult to interpret.

For the most part, the PAS records objects from what are currently rural areas and recent research has identified many of the biases of survival, retrieval and reporting inherent in the data (e.g. Walton 2012, 27-30; Robbins 2013; 2014). These studies provide a valuable corrective to earlier archaeological studies which emphasised investigation of known ‘sites’ at the expense of the countryside. The PAS is beginning to yield significant results in terms of rural settlement which are low in material culture. Supplemented by the dispersed but largely rural material from the PAS, the artefact use and a range of material culture assemblages can be assessed over the region. However, heavy truncation by the plough may have had an impact on the survival of artefacts.

The scheme is an actively used resource which helps to shape our understanding of the complex regional questions within the North West and the archaeological potential is beginning to be exploited through research into specific artefact types, such as regional types of Roman brooch (McIntosh and Ponting 2014), while national surveys show that the North West lacks certain types of artefact common elsewhere (Worrell 2008). Thus, cosmetic mortars show small concentration in Cheshire, while button-and-loop fasteners occur in Cheshire and around northern Lancashire/south Cumbria, with major differences from the trans-Pennine and Midland patterns.

Positive working relationships with metal-detectorists have stimulated reporting of hoards enabling some to be investigated under archaeological conditions (Oakden 2015). The Knutsford hoard had been dispersed into ploughsoil but excavation by Cheshire West and Chester and NML archaeologists recovered the pattern of dispersal from the original findspot but also fragments of the pottery vessel in which the hoard was deposited.

The Knutsford hoard and related finds provide new insights into identification of elites and their cultural preferences. The Knutsford hoard contains ostentatious, highly decorated petalled trumpet brooches, in silver gilt, in the second half of the 2nd century which indicate the presence of individuals in the countryside who had the ability to amass significant wealth. In addition, the brooch has a strongly western and northern distribution (Mackreth 2011, 120). The employment of archaic decorative patterns may represent a native elite consciously marking and reinforcing their ancestral identity through the use of La Tène III forms of decoration. Other markers of adherence to traditional practices may include the presence of fire-cracked stones in the ditch of the short-lived rural site at Southworth Quarry appear to provide an indication that the use of organic containers for heating water pot-boilers continued into the Roman period in Britain. The persistence of circular houses on rural sites, and their occasional appearance within the nucleated settlements, as an example with opposed entrances at Walton-le Dale in the mid 2nd century which appears to have been a blacksmith’s workshop or forge (Iles 2011, 42-43, fig. 6). A roundhouse in the earliest phase of settlement at Wilderspool may have been an enclosed farmstead overtaken by the expansion of the nucleated settlement.

Two of the three La Tène brooches from the Knutsford hoard, deposited c AD190s (courtesy of Robert Philpott)

Trade, Exchange and Interaction : Pottery and other artefacts. Production and trade

Pottery, as the most prolific type of artefact found on Roman sites, has traditionally been important for establishing chronologies and identifying patterns of production and trade. In addition, distinctive consumption patterns are evident, with ‘military type’ supply for industrial and military settlements and rural assemblages having different compositions. The last decade has seen the publication of finds assemblages from a range of site types as well as comparative inter-site studies and more detailed research into regional manufacture and supply.

Military sites and urban centres provide the largest and most wide-ranging assemblages in the region. The potential of finds assemblages from forts to inform on trade, interaction, acculturation and chronology is exemplified by that from Ravenglass fort (C). Despite its small size, the pottery assemblage from Ravenglass is important as fully-quantified assemblages from this period are rare in the region. The nearby Muncaster kilns provided a substantial proportion of the cooking wares at Ravenglass into the second half of the 2nd century AD, but seem to have been supplanted by major industries from further afield by the 3rd century as such industries and trade developed, perhaps reflecting the rise of civilian enterprises as military production diminished (Bidwell and Croom 2015). There is one example of a fine-oxidised ware dish of probable Hadrianic date in an Upper German tradition. Black-burnished ware from south-east Dorset was common across the site and indicates that this was an important supply source throughout the life of the vicus. Pottery from the Severn Valley, the West Midlands and the Lower Nene Valley was also well-represented in the assemblage, with a dominance of Mancetter-Hartshill mortaria in the 3rd century being a common feature in assemblages from north-west England (Hartley 1991, 173). Some of the thin-walled oxidised wares may be later 2nd- to 3rd-century local imitations of Severn Valley ware, similar to vessels found in a kiln dump at Carlisle (Swan et al. 2009).  The quantity of amphorae at Ravenglass is typical of coastal sites, indicating the importance of sea trade and connections with the wider Roman trading network, and the ubiquity of samian across the site also testifies to this. The date range of the pottery indicates that occupation of the vicus continued until at least c. AD 270, with an upper date limit suggested by the absence of East Yorkshire wares, which became common in forts along Hadrian’s Wall from the late 3rd century AD. Two fragments of a face jar with applied bosses and incised lines offer a parallel with a hitherto unique face jar from Old Penrith (Braithwaite 2007), with the applied bosses distinguishing them from other face jars where the bosses are pushed out from the body of the vessel. (S5449) The presence of imported pottery from the continent, and from potteries in the province on the coast (or with access to the sea) strongly suggests coastal trade and presumably the importance of Ravenglass as a port. Other imported materials, such as jet and Skiddaw Slate probably also reached Ravenglass by river and sea rather than by road. (S5449)

Other major assemblages included Wigan (GM) which consisted of over 2000 sherds of pottery, including samian ware, South Spanish olive oil amphorae, black-burnished vessels from Dorset and locally-produced grey and oxidised wares. Three glass vessels, a glass bangle and window glass fragments were also present. Pottery kilns have also been identified at 7a Fisher Street, Carlisle (C), producing wasters and potters’ stamps as well as kiln structures (Johnson et al. 2012).

Comparative work on ceramic assemblages from rural sites, such as Leary’s studies of Mellor and Norley (e.g. Leary 2009), have benefited from an increase in the sample size, and recurring patterns are beginning to emerge. Thus, rural sites in Cheshire display in general higher levels of pottery use than those north of the Mersey, or indeed those further south in the Cornovian heartland around Wroxeter (Smith et al. 2016, 300-306), suggesting closer integration into the market for mass-produced traded goods in the area around the legionary fortress and canabae (Carrington 2012, 399-400).

A consistent suite of pottery ware types is present amongst the military, civil and rural settlements (cf. Webster 2011). Assemblages are dominated after AD 120 by Black-burnished ware (BB1), and oxidised wares from the Cheshire-Lancashire Plain; with low levels of samian (examined in detail by Ward 2011). The power of the marketing networks of the major pottery production centres can be seen in the 2nd century with the rise of Mancetter-Hartshill mortaria and Nene Valley products. Dominant wares of the later 4th century largely consist of Huntcliff and calcite-gritted wares, while shell-tempered wares from the South Midlands from the later 3rd century. The influence of coastal trade is evident in the strong showing of BB1, in both rural and military sites, and South Spanish amphorae. Within the overall patterns, detailed variation due to local geographical patterns is exemplified by the rural site at Mellor on the margin of the Pennine upland area which has Derbyshire ware showing interaction with the production and marketing areas to the east of the Pennines.

Peter Webster (2011) has recently discussed the mechanism for regional manufacture and supply alongside trade and importation of pottery, and sale by local merchants or itinerant traders at regional centres such as Chester, Lancaster and Walton-le-Dale from where it was distributed to the rural hinterland. The network of local pottery manufacturing in the immediate vicinity of military and associated civilian population centres was either established by the army or provided sufficiently large markets to stimulate potters from outside to move in to meet the demand (Webster 2011, 60-61). He suggests some Flavian pottery production was private enterprise on the back of tile manufacture. Webster addresses the question of the identity of the potters who supplied military centres from Holt or elsewhere. He notes that pottery production was not exclusively for military use and argues that a low-level military bureaucracy was engaged in purchasing from private enterprise and a multiplicity of suppliers, as indicated in the Vindolanda tablets. By the early 2nd century the importation of BB1 in large quantities seems to have spelt the demise of the Flavian local manufacturers and during the 2nd century larger producers began to capture large swathes of the market (Nene Valley; Mancetter-Hartshill mortaria).

Webster (2011, 68) argues that the concept that there were separate military and civilian markets is probably false and it is likely that merchants will have regarded any settlement of any size as a potential marketing opportunity. Thus, rural settlements will have acquired their pottery from merchants based in military or civil markets. The overall trade was dominated by a small number of large merchants who controlled extensive distribution networks, at first including the continent, but after demise of much imported wares, throughout Britain.

A national survey by Willis (2011) attempts to move samian studies beyond use for dating to compare samian use at different types of site, military sites, extramural settlements outside military installations. The character of samian assemblages was seen as a sensitive indicator of wealth and status of sites. At two military sites, Birdoswald and Castle Green Carlisle, a remarkably high proportion of samian persists through the early Roman period, suggesting through the role of Carlisle as a transport node the ware was shipped in in quantity for redistribution to Hadrian’s Wall and its corridor (Willis 2011, 181-182). The high levels of samian use in extramural settlements suggest they were ‘closely articulated with the exchange systems of the Empire via the road system and association with a fortress/fort, but additionally, a large proportion of their occupants are likely to have been accustomed to Roman practice and material culture’ (Willis 2011, 182).

Margaret Ward has examined the samian supply to the North West (2011), using a statistical profile for each site to investigate questions of site chronologies, status and lifestyles and trade patterns (2011, 99). She draws close parallels between the methodology of numismatists and samian specialists, both artefact types are closely datable in manufacture though production, circulation and use-life patterns may diverge and need to be judged for coins at least on their own terms. The small size of samples means they are open to distortion through individualistic events, importation of vessels as personal possessions or small groups of vessels as special orders.

Trade, Exchange and Interaction : Roman roads

The main outline of the Roman road network in north-western England has been subject to research for over 150 years (cf. Watkin 1883), although there remain significant gaps in detailed alignments. However, the recent move to make LiDAR open access has contributed significantly to new archaeological research across the NW. The technique captures the slight remains of surviving earthworks of road alignments which may not appear as cropmarks or shadow marks on aerial photographs. Examination of datasets is confirming the presence of previously uncertain alignments or in some cases, revealing them in unexpected locations. Significant new information on a number of routes has been recovered by David Ratledge. They include the long-sought route of the Roman road from Ribchester to Lancaster along a previously unsuspected alignment. The road ran from Ribchester to Catterall near Garstang, where it was seen to join the northern route towards Lancaster (Wilson 2016, 311-315).

Other routes which have seen minor adjustment or confirmation of accepted alignments or infilling of gaps include Ribchester to Kirkham (Margary 703), Ribchester to Elslack (Margary 72), Ribchester to Burrow-with-Burrow (Margary 7c), and Lancaster to Burrow-with-Burrow (Margary 705), the latter providing a confirmed context for the milestone found in 1803 at Artle Beck, Caton (; Wilson 2015, 303-307). Revised or new alignments provide a context for previous discoveries such as the probable funerary statues from Burrow Heights near Lancaster. In Cumbria, Toller has used LiDAR data to confirm the course of a road from the Roman fort at Low Borrowbridge, near Penrith, to Kirkby Thore. This is a missing part of a well known road called the Maiden Way that continues towards Whitley Castle and Carvoran Roman Fort, Northumbria, (near Hadrian’s Wall). LIDAR also provides evidence for a new section of this road past Kirkland (

Several sections have been excavated across Roman road alignments, with the largest site being the Wigan to Manchester Roman road at Wentworth School, Salford (GM) (Murphy 2013) where about 25m length of a gravel metalling was exposed, sealing a layer of burnt heather. Investigations of the Roman road between Wigan and Walton-le-Dale north of Wigan have been undertaken. The route from Wigan to Manchester has also been examined at Ellesmere Park in 2005 and Amberswood in 2003 (Miller and Aldridge 2011, 20-21).

Wentworth High School Roman road excavation, Salford, Greater Manchester (courtesy of Salford Archaeology)

Minor roads have been discovered in large area excavations, as at Saighton Camp near Chester where the crossroads of two trackways were revealed (Wilson 2015, 309-310), while a minor trackway was identified at Dutton’s Farm, Lathom (L) (Cowell 2005).

Another minor road was partially excavated in Kentmere (C), revealing different construction techniques, employed due to the nature of the terrain. The excavation confirmed the presence of a metalled surface, which corresponds to surviving earthworks in the immediate area. However other features were also revealed, associated with the prevention of flooding or possibly to stop the encroachment of peat onto the road. Although the ditches identified are common in the construction of Roman roads, the banks may be unique to this site, possibly necessitated by the difficult terrain (Greenlane Archaeology 2006).

Late Roman-early post Roman transition

The breakdown of economic systems after the end of formal Roman administration in AD 410 makes it difficult to recognise very late Roman or early post-Roman deposits archaeologically, although new studies of the northern frontier region have identified certain artefact types, including military equipment, brooches and pottery, which have been recognised as continuing into the 5th century (see the papers in Collins and Allason-Jones 2010a). The scarcity of all types of these metal artefacts in the North West by contrast with east of the Pennines is attributed in part to much lower level of metal-detecting and excavation in the former region, but also biases in modern settlement (Collins and Allason-Jones 2010b, 134).

New evidence for occupation spanning the very late Roman to post-Roman period has come from some north-west military sites. The classic site is Birdoswald (Wilmott 1997) but in default of diagnostic artefacts careful stratigraphical analysis has shown that activity in this period persists in Lancashire forts, although only restricted areas of those sites have so far been examined. Some military sites certainly appear to have been occupied until the end of the Roman period (e.g. Ravenglass). Elsewhere, activity appears to continue into the post-Roman period at Ribchester (L) fort in excavations by UCLAN beginning in 2015, while a newly identified fort in West Lancashire (location confidential) has produced late deposits including an oven or hearth and stone platforms, the latter apparently for structural foundations, which overlie deposits with Crambeck/calcite-gritted wares dated AD 360-400+. Some structural platforms overlie internal fort roads.  It remains uncertain how these places functioned, whether as lineal descendants of the military garrisons, with regional commanders protecting the local agricultural population in return for a customary levy, or as part of a wider administrative network, (Wilmott 2000).

The re-interpretation by Roger White of stratigraphic sequences in Chester at Abbey Green, Hunter Street and Hunter’s Walk highlights the problem of assigning accurate dates to activity which is broadly book-ended by the end of the Roman period and the medieval period. He argues that the presence of late Saxon pottery has been used to date features or phases often based on only slender associations  and some activity may be 5th- or 6th-century in date, especially at Abbey Green which has East Mediterranean amphorae, North African pottery and horse-head military metalwork which would fit potentially into that period (2007, 187-189).

Part of the central area of the fort including the principia at Carlisle appears to have continued in use into the very late Roman period, and well into the 5th century, perhaps by non-military personnel, while adjacent streets were used as a market (McCarthy 2017, 10, and Table 2).

A late 4th or early 5th century military belt plate from Meols, a site with a coin list ending with Magnus Maximus (383-388), hints at late Roman military activity on the Irish Sea coast (Griffiths et al. 2007). This is one of the few sites in the North West where artefacts suggest continued activity through the early medieval period, supported by the hinterland distribution of 6th century Byzantine copper coins in north Wirral (Philpott in prep.). Other evidence is less coherent. An isolated pit, or possible kiln, at Nether Wasdale (C) provided a very late Roman or immediate post-Roman period date. This is exceptional in the context of the site and the region, and in association with the early medieval settlement containing evidence for metalworking the site is potentially of national significance (Oxford Archaeology North 2016). A late hoard of silver siliquae from Dutton (Ch), with coins minted in the mid 370s onwards, is the only siliqua hoard from the NW with a secure provenance (Oakden 2015, 52-53).

Many rural sites in the North West, where datable, have a peak in the 2nd century with a decline in the 3rd but very few can be proved to continue to the end of the Roman period (Nevell 2011, 48-50, fig. 3). Rural sites with late Roman or potential early post-Roman occupation include Mellor and Irby, Wirral; at the latter stratigraphic analysis, C14 dating and finds, including late 4th pottery, suggest occupation continued beyond the end of formal Roman occupation. Within the Saighton enclosures, the most clearly defined building (discussed above) consisted of a rectangular post-built structure, 14m by 6.5m in size and made up of ten roughly-paired posts, with a possibly internal partition and further external smaller posts. It appears to post-date the series of recorded sub-rectangular enclosures, and construction was dated by pottery to the late 4th century date, making it the latest identified feature (Wood 2016). On the face of it the decline in presumed agricultural units, and production, may be a consequence of lower demand due to a reduction in the size of the late Roman military garrison in the region, but a suggestion that this may be in part a reduction in visibility through a decline in engagement with the Roman money economy is also possible. Only tight sequences of radiocarbon dates will resolve the questions over the duration of site occupation during materially poor periods. 


The period since 2006 has been productive in increasing our understanding of the Roman North West. We are, however, still in the basic data-collection phase for many of themes.

Some areas of Roman archaeology in the region have seen major advances over the last decade. Extensive geophysical surveys illustrate the scale of some of the extra-mural settlements of forts but so far they have seen very little detailed investigation, essential before we begin to look at questions such as their chronology, economy, and material culture– and to investigate wider theoretical issues of identity, population origins and movement, economic integration or not with the rural hinterland and the military establishment.

As for rural settlement, a number of new sites have been investigated in detail, and although some broad patterns emerge, notably the presence of roundhouses, perhaps unsurprisingly there remains much variety in detail in the occupation histories. Continuity from the Iron Age to Roman period is hard to demonstrate at most sites, although it is uncertain whether the apparent pattern of intermittent occupation in the Iron Age is a result of the difficulty of recovering detailed occupation sequences or reflects the genuine situation.  Site investigation shows that many rural settlement in the lowland parts of the region are heavily ploughed with the loss of much internal detail within enclosures, enhancing the value of those few sites such as Irby, Wirral which retained complex vertical stratigraphy.

Distinctive patterns within the rural distribution of material culture across the region are emerging both through the PAS data, largely composed of metalwork, and through studies of pottery from site assemblages. There is wide variety in the adoption of the most common artefact, pottery, which may touch on marketing and distribution networks. In addition, different parts of the region demonstrate different access to, or attitudes to, metalwork such as brooches, other personal items and coinage. The latter suggests factors such as social hierarchies and rural economies based on pastoral rather than arable farming restricted opportunities for social display of material culture and resulted in an economy in many areas based on now-intangible commodities rather than coinage.

An important development is the growing evidence for occupation at the end of the Roman period or beyond, with C14 dating allied to careful stratigraphical analysis producing a growing list of sites where early post-Roman occupation is suspected or proven. Together with foci of settlement such as suspected at Burscough and Ribchester, the Chester evidence may indicate continued occupation of Roman forts, while rural sites, and former Iron Age hillforts such as Helsby, may represent an alternative locus of elite power.

Publication of major sites has advanced considerably since 2006 but the tendency to consign important site reports to grey literature is an inadequate response to the duty to make accessible the results of archaeological work.

Of multi-period significance

There are some constraints particular to certain sites in the region. A core principle of development–led work in the historic core of the city of Chester (Primary Zones as defined in the Chester Archaeological Plan) is the preservation in situ of significant archaeological deposits through careful foundation design, which reduces the impact of new works on archaeological strata to the absolute minimum. This has limited the amount of new archaeological information from the city with regard to the Roman and all subsequent periods. However, the CUAD, has seen vast amounts of disparate archaeological data from Chester integrated into the Historic Environment Record and used to define archaeological Character Zones. The data also form the basis of the Chester Archaeological Plan which is part of the Evidence Base for the Cheshire West and Chester Local Plan. The results of this project have proved invaluable in the continued provision of robust development-management advice.

Some major gaps remain –animal bone survives poorly at many sites, except certain types of military/urban locations, denying us the opportunity to investigate such questions as dietary preferences, livestock management and butchery practices. Inimical soil conditions mean that unburnt human burials are notable by their absence, so data on inhumation burial practices are scarce and the opportunities for the developing areas of isotope and genetic analyses which allow us to investigate population movement and composition are accordingly lacking at most places, although isotope analysis is planned for burials from Dog Hole Cave, Haverbrack (C) (H. O’Regan pers. comm.).

Extensive landscape surveys are often multi-period in approach and intend to characterise the archaeological resource irrespective of date. One difficulty is adequate characterisation of the multi-period landscapes which often have undated features or landscape elements. Extensive investigation of landscapes in the northern upland areas encounters the problems of establishing the origins and duration of use.

Leave a Reply