The 2017 Roman Specialist Group consisted of Nick Hodgson (Tyne & Wear Archives & Museums)(Convenor), Lindsay Allason-Jones (independent), Rob Collins (Newcastle University), Jim Crow (University of Edinburgh), David Heslop (independent), Frances McIntosh (English Heritage), Jennifer Proctor (Pre-Construct Archaeology). The following specialists or members of the previous specialist group were consulted: Mike Collins (English Heritage), Paul Frodsham (independent), Richard Hingley (Durham University), Don O’Meara (Historic England), Steve Willis (University of Kent).
The advent of Roman rule in the North-East of England had a profound qualitative and quantitative impact on the archaeological record. A suite of new site types appeared, particularly those related to the Roman military infrastructure, and in many areas there was a significant change in the availability of material culture. For the first time, written sources, both literary and epigraphic, become available so that from the 1st century AD there is evidence for the names of individuals, places and political and ethnic groups. This combination of written evidence and a significant body of highly diagnostic material culture (particularly ceramics and coins) allows the archaeology to be explored at a chronological resolution not practical for earlier periods.
Hadrian’s Wall is the iconic Roman monument in the North-East, its national and international importance being reflected in its World Heritage Site status. Due to the Wall’s outstanding importance, the large-scale heritage management issues it raises, and the sheer quantity of material relating to it, as well as the fact that it crosses two regions (North-East and North-West), the North-East Regional Research Framework does not tackle it comprehensively. The Wall is instead subject of its own separate research framework (Symonds and Mason 2009), which is being revised in parallel to the current revision of the North-east framework. However, the impact of the Wall on the region both in Roman times and in modern archaeological terms means that it cannot be ignored, and frequent reference will be made to it in what follows.
The study of the remains of the Roman period has a long tradition in the region. The first important account was by William Camden in his Britannia, published in Latin in 1586 (first English language edition: Camden 1610). Periodically revised by successive editors with contributions from various local agents, this remained the main source for the Roman antiquities of the region until the early 18th century, when John Horsley’s monumental Britannia Romana (1732) was published. Epigraphy and observations of visible remains form the backbone of research in this period, which frequently describes details that have since disappeared.
The first significant archaeological excavations on the Wall were carried out by John Hodgson, who started work at Housesteads in 1822. Hodgson was also the first to date the Wall to Hadrian; previously the Vallum was believed to be the earliest defence, supplemented by a wall built during the reign of Septimius Severus. Anthony Hedley and John Clayton carried out other excavations at the same time. Hedley worked at Vindolanda in the early 1830s but died in 1835, before a report on his work could be submitted to the Society of Antiquaries of Newcastle. Clayton bought up much land along the Wall and dug many sites including three milecastles, several of the forts, Coventina’s Well, and carried out a decades long programme of excavation at the fort at Chesters, in the grounds of his own country seat (McIntosh 2014).
The Wall by no means monopolised excavation projects in the nineteenth century. An important local patron was the Duke of Northumberland who, in 1852, funded excavation at High Rochester in Redesdale. As well as commissioning the first accurate survey of the Wall, carried out by Henry MacLauchlan between 1852 and 1854, Northumberland also commissioned a survey of Dere Street (then known as Watling Street) by MacLauchlan in 1850-51. This included plans of several forts north and south of the Wall, such as Lanchester, together with records of finds and associated earthworks. Other very important nineteenth century surveys and excavations of Roman forts included Risingham (1843-49), South Shields (1875-77), and Binchester (1878-80), the latter two published to high standard for the time.
As well as excavation and survey there was also an increased focus on epigraphy. The most notable early scholar in this field was John Collingwood Bruce, who produced the Lapidarium Septentrionale (1875), an overview of all the inscriptions and sculptures known at the time (Collingwood Bruce 1875; Breeze 2003). He also wrote The Roman Wall (1851), the third edition (1867) of which was the first widespread popularisation of Hodgson’s theory of a Hadrianic date for the Wall. In 1863 he published his Handbook of the Roman Wall, which continues to be updated (14th edition edited by David Breeze, 2006).
The major period of broadly scientific excavation began in the 1890s, including Robert Carr Bosanquet’s excavations at Housesteads (Bosanquet 1904). The ambitious excavation of the Roman town at Corbridge (1906-1914), published in considerable detail, stands on the cusp of the nineteenth-century era of pre-stratigraphic excavation and the emerging scientific approach. Work by John Pattinson Gibson on Turret 44b led to a series of collaborations between the major figures of Roman archaeology in the region, including F. Gerald Simpson, Robin Collingwood and Ian Richmond. They would go on in the 1930s to work out the basic structural sequence of Hadrian’s Wall that still holds good today. In the late 1920s and early 1930s a burst of activity saw the foundation of the North of England Excavation Committee, founded by the Society of Antiquaries of Newcastle in 1924 to encourage ‘under proper supervision the excavation of sites in the North’. Due to its Newcastle base it tended to focus on sites on the eastern half of the Wall. Also in the east, the improvement of the Military Road (B6318) resulted in a series of rescue excavations.
The importance of this work was recognised by Durham University who, in 1924, appointed F. G. Simpson as its director of excavation, a post he relinquished in 1931 to allow Eric Birley to be appointed as a lecturer. Birley was to stay at Durham for 40 years, directing his students to research on a wide range of aspects of the Wall and the Roman army (for an overview of his life see Dobson 1998). Important milestones in the study of the areas north and south of the Wall in this period were, for Durham, a synthetic study by Petch (1925) and Kenneth Steer’s PhD thesis The Archaeology of Roman Durham (1938) (which remained standard until Brian Dobson’s re-consideration of 1970); for the north, Richmond’s The Romans in Redesdale (1940).
As this brief account demonstrates, the early phases of investigation into the Roman archaeology of the North-east were dominated by individuals, though in the later 19th and early 20th century the role of societies, particularly the Society of Antiquaries of Newcastle, became increasingly important. State intervention only began in the 1920s when parts of the Wall were taken into Guardianship. The work of the Ministry of Works, Department of Environment and latterly English Heritage and Historic England, in the recording and conservation of the archaeological resource has carried state participation in the study of Roman archaeology in the area through to the present day. The presence of two departments of archaeology in the region, at Durham and Newcastle, has also influenced scholarship, and scholars such as Eric Birley and Ian Richmond shaped their departments as centres for the study of this period. More recently, major work has also been carried out by local authority museums, notably Tyne & Wear Archives & Museums, with recent excavation and curation of South Shields and Wallsend, as well as extensive exploration of the urban sections of the Wall. Work by independent scholars has also been significant; the long-term campaign of excavation by the Vindolanda Trust has been on a very large scale and has produced a number of internationally significant finds, such as the Vindolanda tablets.
From the 1960s there were a number of overviews of the Roman archaeology of the region. The first was Peter Salway’s The Frontier People of Roman Britain (Salway 1965), one of the first major attempts to consider the evidence for the civilian as well as military settlement, but now almost entirely overtaken by new knowledge (much of it gained in a series of geophysical surveys carried out by the late Alan Biggins, David Taylor and their colleagues) of the form, extent and date of occupation of the civilian settlements attached to Roman military sites. The long history of research on the Wall has been summarised in several publications, most recently Breeze 2014.
For an overview of the earlier work on the Wall, Birley’s Research on Hadrian’s Wall (1961) remains a standard text, but can now be supplemented by further historiographical study of Roman archaeological research in the region, in which the study of the subject has become a subject of study in its own right. Recent papers which explore the history of research on Hadrian’s Wall and the North-East include work by David Breeze (2003) on the role of John Collingwood Bruce and on the development of thought about Hadrian’s Wall (2014) and, reviewing the way the significance of the Wall has changed for people over time, Richard Hingley’s Hadrian’s Wall – a Life (2012).
Since the 1980s major excavations, carried out for a variety of managerial, tourist-promotional and research reasons have continued to take place at some major fort sites (South Shields, Wallsend, Vindolanda, Birdoswald, Binchester). However, a conservation philosophy with a presumption against research excavation for its own sake has meant that since 1990 curatorial bodies have not encouraged or promoted much in the way of intrusive investigation into the history of the Wall itself and its minor installations, which consequently remain relatively poorly documented and understood, especially where the sites are unthreatened. It is also notable that the major excavation work at fort sites from the mid-1980s has been undertaken by bodies (e.g. the National Trust, English Heritage, the Vindolanda Trust, Tyne & Wear Archives & Museums) other than the two Universities with archaeology departments in the region. However, since the introduction of PPG16 and associated planning guidance (currently the National Planning Policy Framework) in 1990 there has been an increase in interventions into the Roman-period archaeology of the region. These have helped expand our knowledge of Hadrian’s Wall, particularly in urban areas where previous knowledge even of the actual course of the Wall was patchy and where the pace of development has been greatest. Most dramatic, however, has been the way that developer-funded archaeology has cast light on rural settlement away from the Wall, bringing to light previously unknown or archaeologically invisible site-types and funding landmark publications as well as excavations of threatened sites. When the first version of the present framework was published it could still be considered that the only really substantial work on civilian rural settlement was that carried out by George Jobey, whose pioneering and almost single-handed exploration had dominated the study of Iron Age and Roman-period native settlement, predominantly north of the Wall. Upland survey work, such as on the Otterburn training area and by Denis Coggins in the North Pennines, had also revealed many settlements, though without excavation it had proved difficult to date them closely. The evidence base for sites of this type has been hugely expanded by developer-funded archaeology and the conclusions reached by George Jobey can be regarded in a different light. Indeed, a new dimension has been added to our understanding of the density of settlement in lowland areas and of interaction between the Roman military and the rural population, and old categories of study are breaking down: where once ‘Iron Age’ and ‘Roman’ archaeologists appeared to operate in parallel but self-contained worlds (‘Roman and native’ in the parlance of the 1980s), the need is now felt to understand social development in the Roman north-east holistically, and for the region to be regarded as a dynamic frontier-zone, and not merely a remote area where the presence of the military hindered ‘normal’ Roman provincial development.
Within the North East, these developments have led to much recent discussion of the Iron Age-Roman transition, which can be most clearly observed through rural settlement, where the primary settlement form of the eastern lowlands of the region did not radically change with the arrival of the Romans in the later-1st century, but appears not to have been maintained after the consolidation of the frontier on the Tyne-Solway Isthmus (Hodgson et al. 2012; Proctor 2009). It is important to note here major recent developments just outside the region that will eventually have far-reaching implications for the understanding of the Iron Age-Roman period transition in the north-east: these are the final publication of the Stanwick research project (Haselgrove 2016) and the major excavation, still in progress at the time of writing, that has preceded the A1 widening at Catterick/Scotch Corner.
At the opposite end of the Roman period, there has been a considerable preoccupation with the increasing evidence from Hadrian’s Wall and other Roman forts, towns, and villas across northern England attesting to various activities in the fifth century and beyond (Collins 2012, 2017).
Since the first edition of this Research Framework some research tools have become available which will stand as significant landmarks in the study of the Roman archaeology of the region. The Hadrian’s Wall National Mapping Programme (NMP) has mapped the whole of the Hadrian’s Wall World Heritage Site (WHS) and its landscape setting, synthesising information available from aerial photographs and other sources. The results have not been formally published, but have been incorporated into the new 1:25,000 Archaeological map of Hadrian’s Wall, published by English Heritage in 2010, the first authoritative published map of the Wall since the last (1972) edition of the old Ordnance Survey archaeological map, which the new one has taken as its model.
Of works of reference relating to inscriptions and sculpture it is important to note that Roman Inscriptions of Britain has now been extended to embrace discoveries up to 2006 (Tomlin, Wright and Hassall 2009). Work is currently in progress on the final volume of Corpus Signorum Imperii Romani, covering the hinterland of Hadrian’s Wall (Northumberland south of the Stanegate, Tyne and Wear, County Durham, Cumbria, Lancashire and Derbyshire). Sculpture from Corbridge and the Wall east of the North Tyne up to 1974, and from the Wall further west up to 1986, are covered in CSIR volume I, fascicules I and 6 respectively. In the past there have been few general works that address the rural population of the region (but see Hingley 2004 and Hodgson 2017, the latter a monograph on Hadrian’s Wall but attempting to place the Wall in a regional context and to assess its effects on the rural population). Now, the region has been covered by the monumental Roman Rural Settlement project, a synthesis of the excavated evidence for rural settlement in Roman England and Wales, including the flood of new information since the introduction of developer funding in 1990 (Smith et al. 2016; online resource at http://archaeologydataservice.ac.uk/archives/view/romangl/), although this remains relatively limited for the north-east compared to southern Britain.
A series of initiatives at both regional and national level provide research agendas from which to work. At a local level the Research Framework for Hadrian’s Wall (Symonds and Mason 2009), which provides a research structure for the study of the Wall in both the North-West and the North-East, is now undergoing its own process of revision.
Still of value at a regional level are the papers covering the Roman period in the Past, Present and Future volume which arose out of a conference held in Durham in 1996, with useful stocktakings on military sites (Crow 2002), non-military sites (McCarthy 2002, identifying the need for macro-level study of the wider Roman-period landscape and environment combined with micro- level or site-specific analysis of data, including site formation processes and the study of cultural assemblages), and of finds (Allason-Jones 2002, calling for further work on military equipment from the North-East, placing it in its national and international context, and more work on the investigation of the poorly researched topics of cemeteries and vici).
At the national level, the most influential contribution has been Britons and Romans: advancing an archaeological agenda (James and Millett 2001), a volume of collected papers which arose out of a session at the Roman Archaeology Conference sponsored by English Heritage (e.g. J. D. Hill 2001; James 2001; Allason-Jones 2001; Evans 2001). As well as highlighting particular research topics, there are suggestions here for structural initiatives, such as the training of finds specialists and improved publication of ‘grey literature’, issues that are also echoed in this document. Similar suggestions were made in Town and country in England: frameworks for archaeological research (Perring et al. 2002) which laid out an agenda for work on urban archaeology. Although its Roman case study focused on Essex and Colchester, that volume presented a series of methodological recommendations that have relevance for the North-East.
Recommendations have been issued by the Study Group for Roman Pottery (SGRP), including a national research framework (Willis 2004) and a regional overview for the north of England (Evans and Willis 1997). The former highlighted a series of research themes that are still germane, including trade, supply and distribution, chronology, continued work on samian pottery, Roman/native interaction, pottery and the organisation of the Roman army, functional trends, site status, spatial patterning, social/cultural identity, ritual sites, Roman pottery production and the end of the Romano-British economy. The regional review highlighted similar themes and also identified a number of backlog sites as priorities for publication. Of the latter Piercebridge was seen as particularly important and was published in 2008 (Cool and Mason 2008). A series of publications of Roman military and of civilian/indigenous sites in the last twenty years has improved the pottery information available but from a low starting point where very little was available reported to modern standards, hence priorities endure. There remain, for example, few samian assemblages from the northern frontier published to current standards. Since 2012 the Scottish Archaeological Research Framework has provided a living online resource for Scottish archaeology, including an overview and set of research questions for the Roman period: these have much overlap with and relevance for the North-east England region.
The degree to which the Roman occupation contributed to the patterns of clearance indicated by pollen evidence has remained controversial, as the imprecision inherent in radiocarbon dating makes it difficult to distinguish between pre- and post-conquest clearances at a high degree of resolution. Dumayne (1992; 1993a; 1993b; 1994) has argued for vegetational change as a result of the presence of the Roman army, but the majority view is that the increase in agriculture in the region dated to the Late Iron Age and was not decisively related to the impact of Roman settlement (Tipping 1997; McCarthy 1995; 1997). Chris Fenton-Thomas (1992) concluded from the pollen evidence from 13 pollen core sites between the Tyne and the Tees that forest clearance pre-dated the Roman period and moved out from a south-east core-zone throughout the period, only reaching a peak at some locations in the 5th century AD. The pre-Roman clearance of vegetation was also suggested via pollen analysis from ditch fills at Vindolanda (Manning et al. 1997). Cores from the areas closest to the forts of north-west Durham also show clearance beginning in the late pre-Roman Iron Age, indicating that it was not related to the presence of the Roman military. The same controversy occurs with woodland regeneration after the end of the Roman period: some commentators have claimed this as a phenomenon in the region specific to the early-fifth century, but for others it is impossible to be so precise and the most we can say is that woodland regenerated, with much local variation, between AD 400 and 800 (Dark and Dark 1996; Collins 2012, 134-7; Huntley et al. 2009).
Significant plant macro-fossil evidence is now published from non-military sites at Faverdale (O’Brien 2012) and the Roman villa site at Ingleby Barwick on the Tees (Huntley and O’Brien 2013), where the assemblages were dominated by spelt wheat and hulled barley, which recent excavations north of the Tyne have shown to be typical of pre-Roman Iron Age crop husbandry regimes throughout most of our region (e.g. Hodgson et al. 2012, 203), contra Van der Veen (1992), who had argued that emmer wheat predominated north of the Tyne. Earlier assemblages from non-military sites include Dubby Sike, Upper Teesdale (Co. Durham), which produced no evidence for cereals, suggesting possible seasonal occupation (van der Veen 1986), Thornbrough Scar (Northumberland), which produced evidence for possible rye cultivation, and Catcote (Teesside), which produced barley and wheat (van der Veen 1983).
Despite the number of Roman military sites, they yield surprisingly few significant assemblages, of which the burnt deposits from a granary at South Shields are one example (van der Veen 1988b). The assemblage from the Roman fort at Newcastle produced wheat, heather, barley, oats and arable weed seeds, as well as more exotic material such as coriander and fig seeds (Huntley and Daniell 2002). Some waterlogged remains are known from Peel Gap, which implied the presence of fen meadows, grassland and heathland (Huntley 1989a). In the well-preserved pre-Hadrianic layers at Vindolanda identification of sedges enables some distinction between different areas used for grazing or hay production (Huntley 2002). From the same site the examination of cereal bran has allowed inferences to be made into the context of cereal consumption based on waterlogged bran remains (Britton and Huntley 2011). These assemblages are primarily from forts associated directly with Hadrian’s Wall, and there are to date no comparable assemblages from military sites north or south of the Wall.
Turning to faunal remains, a recent development is the availability for the first time of significant assemblages from non-military sites, notably Faverdale, where the animal bone was taken to suggest that the site was involved in the production of stock, including horses (Rielly 2102). Further assemblages are published from the villa at Holme House, Piercebridge (Co. Durham) on the very southern border of the region (Gidney 2008), and Quarry Farm, Ingleby Barwick (Gidney 2013). For the military sites, significant amounts of bone have been recovered from South Shields (Younger 1994; Stokes 1996; 2000), where a large assemblage can be associated with dated contexts, and also from Wallsend (Gidney 2016), Vindolanda and Corbridge (Hodgson 1967; 1968; 1977). Away from Hadrian’s Wall, bone assemblages are published from Binchester (Cussans and Bond 2010) and Piercebridge (Rackham 2008). A further important source of faunal evidence will be forthcoming from the more recent excavations at Binchester.
The faunal evidence for the region has also been the subject of some important discussion, notably the suggestion (Stallibrass 2009) that cattle may have been driven over long distances to supply Hadrian’s Wall and other military markets, which has received remarkable corroboration from isotope analysis of cattle bones from 3rd century contexts at South Shields which suggests that the animals had originated in Cumbria or even south-west Scotland (Waterworth 2014). This has profound implications for the relationships the Roman state might have had with central authorities in areas north of the Wall who had control over the production, collection and distribution of cattle (cf. now Mercer 2018, 204-16). There is limited survival of insect remains, probably due to the lack of deep, well-preserved deposits, such as those in York and Carlisle. A single insect assemblage from a pit at the Roman fort at South Shields was reported by Osborne (1994), and there is a report from Faverdale (Davis 2012). An unpublished assemblage of insect remains form Housesteads is currently stored at Corbridge Roman site. There is also an excellently preserved assemblage from a fill of a late-Iron Age or early Roman ditch terminal at the Flodden Hill rectilinear enclosure (Kenward 2001) in the northern extremity of the region. The remains from this site were used as part of a wider national study which demonstrated a milder climate than the present day, the remains from Flodden indicating temperatures similar to those of modern Kent (Kenward 2013).
Although Roman archaeology in the North-East has traditionally been site centred, there is increasing evidence for the wider Roman-period landscape, although much of the surviving networks of fields, cord rigg and settlements revealed by aerial photography (Frodsham 2004, 57-59; Gates 1997; 1999; 2000) still cannot be closely dated. However, at individual site level, the move towards larger, open-area excavations has helped to place settlements in their wider landscape. Excavations at Pegswood (Northumberland) and Newcastle Great Park (Tyne and Wear) show how settlements sit in networks of fields and paddocks; a similar pattern has also been recognised further south at the villa of Ingleby Barwick. There are also traces of native field systems beneath the fort at Wallsend (Hodgson 2003), in active use when the fort site was selected in the 120s. Unpublished field systems recorded at Newton Bewley may be indicative of the later-Roman landscape in the developed province south of the Wall. Field systems and complexes of enclosures have now been recorded in detail on the outer fringes of several military vici, by means of both geophysical survey and excavation (e.g. Biggins and Taylor 2004; Proctor 2011).
The basic layout of the Roman road network in the North-East remains that laid out in Margary’s Roman Roads in Britain (first published 1955; revised ed. 1973). Dere Street runs the entire length of the region from the Tees to the Scottish border. The first major research on Dere Street was undertaken by Henry MacLauchlan (calling it Watling Street) who carried out a survey of the road in 1850 and 1851 (MacLauchlan 1852). The road ran north from York through Aldbrough (8a-b; all road numbers are those provided by Margary 1973) before entering County Durham at Piercebridge. It then ran north to Binchester (8c), Lanchester, and Ebchester (8d) until it reached the Tyne at Corbridge. It then headed roughly north-west through Redesdale passing by Risingham, High Rochester and up into the Cheviots. At Chew Green, at the head of Coquetdale, it drops down into Scotland heading towards Newstead and ultimately Lothian.
One of the main west-east routes over the Pennines came off Dere Street at Scotch Corner and ran west through Greta Bridge, Bowes and over Stainmore through Rey Cross before reaching Brough-under-Stainmore and the Eden Valley (82). There was also a link road running from Bowes through Barnard Castle joining up with Dere Street around Bishop Auckland (820).
Just north of Binchester at Willington a road (83) branched north-eastwards towards Durham; a 150m length of this has recently been excavated and published (Scott 2015). The road continued north from Durham to Newcastle via Chester-le-Street (80b). Another road, often known as Cade’s Road (80a), after the Durham antiquary who first suggested its course in the late 18th century, is supposed to have crossed the Tees around Middleton St George and then headed north to Durham to join 80b see Dobson 1970). Cade’s road has not been proved by excavation, although the newly discovered and excavated settlement at East Park, Sedgefield is on its traditionally accepted line. A spur road, known as the Wrekendyke, ran in a north-eastern direction from 80b at Wrekenton to South Shields (809). A short stretch of Roman road is also known in the North Pennines running from Stanhope to Egglestone (821).
The Tyne-Solway gap was an obvious west-east communication route, and two Roman roads ran across it. The earliest was the Stanegate (85a), running west from Corbridge. No continuation east of Corbridge has ever been traced. A second east-west route, the Military Way, was eventually built as a communications route for Hadrian’s Wall, closely accompanying the Wall from Wallsend westwards (86a, 86b).
North of the Wall the Devil’s Causeway (87) forks out north-eastwards from Dere Street at Bewclay to run via Longframlington to the Tweedmouth area, where a Flavian or Trajanic fort is to be expected. Only one of the several forts that must have existed along this route, all abandoned by the time Hadrian’s Wall was built, is known for certain, at Low Learchild at the crossing of the Aln. This route was surveyed by Henry MacLauchlan in the late 1850s (MacLauchlan 1864). Dere Street and the Devil’s Causeway are linked further north by a west-east route running from High Rochester to Whittingham (88). A current research project is revealing Roman artefacts in association with the Devil’s Causeway, particularly as it crosses rivers (Allason-Jones 2016). This picture of the basic road network described here cannot claim to be comprehensive or final. There are many other observations and records of supposed Roman roads, particularly in the County Durham area, ranging from antiquarian snippets to lengthy reports produced in recent years by amateur groups and archaeological societies, which can only be evaluated on the basis of complicated research and further fieldwork.
The sites of a number of Roman bridges are known in the region, some well studied (Dymond 1961; Bidwell and Holbrook 1989; Holbrook 1996; Fitzpatrick and Scott 1999; see also Moorwood and Hodgson 1992). In particular the region is rich in evidence for architecturally ambitious bridges with stone arched superstructures, a rarity in the south and midlands (where timber construction was the norm), to be explained by investment in major building schemes on a military frontier and its associated highways that were conceived of in monumental terms. Major stone bridges are recorded in varying amounts of detail at Chesters, Corbridge, and Risingham. Piercebridge has a complicated sequence of several bridges. In addition to bridges it is likely that rivers were also often forded, for example, the Tees is fordable at Neasham and Barnard Castle.
Maritime transport was undoubtedly an important form of communication in the Roman period, and it is likely that most bulk cargo was carried by ship. In spite of this, no port or harbour facilities have been firmly located, though on circumstantial grounds the existence of port facilities at the coastal supply-base of South Shields can be regarded at certain. Here Roman finds which wash up periodically on the Herd Sand beach may emanate from the wreck of a Roman ship that came to grief entering the Tyne in the later 2nd century AD (Bidwell 2001). Another possible wreck is known from Hartlepool Bay (Swain 1986). A collection of objects from the beach at Carr House Sands, Seaton Carew, is likely to be a domestic midden, although it could also be all that remains of a foundered Roman cargo ship (Swain 1986). It should also be noted that the recent discovery of villa sites distributed along the Tees Valley suggests that the river was navigable and may have been used for the transport of agricultural products.
The north-east region stands alongside parts of Scotland and Wales in possessing one of the richest resources relating to temporary Roman military works to be found anywhere in the empire (Welfare and Swan 1995; Richmond 1940). Although some are known from County Durham, most notably Rey Cross on the Stainmore pass (Flavian in date), they predominate in Northumberland and along the Hadrian’s Wall corridor. Some of the many camps known along the line of Dere Street relate to the Flavian conquest of the north, including Bellshiel, Birdhope, Chew Green, Dargues and possibly Silloans (Welfare and Swan 1995). Others, like the remarkable complex at Chew Green, high on the Cheviots, where the earthworks of a series of related camps and forts still survive in excellent condition (Keeney and Richmond 1937) relate to the long-term use of this military highway and supply-line. Aerial reconnaissance has recently added a new dimension to our knowledge, with the recognition of lines of march represented by camps along the Devil’s Causeway, and along the Tweed Valley (Gates and Hewitt 2007), the latter route being particularly interesting as a campaign route where no constructed Roman road was ever built. Where modern excavation on camps has taken place it has tended to be in a rescue rather than a research context, as at Rey Cross (Co. Durham) (Vyner 2001). Dating is often uncertain, and the potential for camps to contain datable structural features, now demonstrated in Scotland, is unexploited in our region.
The origin and nature of the Stanegate road and its associated structures have been much debated. The traditional interpretation was that this was a defensive frontier created following the withdrawal by the Roman army to the Tyne-Solway isthmus in AD 105. In the 1970s and 1980s, however, this interpretation was questioned (e.g. Daniels 1970; Dobson 1986) and the arrangement seen as a secured communication route rather than a fortified limes. More recently, the argument for the role of the Stanegate system as a preclusive chain of fortifications has been restated (Hodgson 2000). Poulter (1998) has pointed out that the road itself appears to be one of the latest elements of the system, making it unlikely that the primary purpose of the forts was to defend the road. The forts at Corbridge, Vindolanda and Carlisle were rebuilt simultaneously around 105, and this may mark their refurbishment as part of the constitution of the new line following the abandonment of southern Scotland, which took place at that time. For the fullest recent discussion of the Stanegate as a frontier ‘system’, see Symonds and Mason 2009, 10-33.
The road itself ran from Corbridge to Carlisle (and possibly west to Kirkbride), and a series of forts stand along its course, including, in the North-east region, Corbridge, Newbrough, Vindolanda, and the fortlet at Haltwhistle Burn, carefully excavated in 1908 (Stobbs 1997). It is possible that the small fort at Washingwells was also related to the Stanegate, though no road has ever been found linking it to Corbridge and it has no firm dating evidence (Holbrook and Speak 1994; Casey and Howard 2010). It has been tentatively suggested that the pre-Hadrianic frontier may not have followed the Tyne east of Corbridge, rather being represented by the Devil’s Causeway, with the Northumberland coastal plain only being abandoned when Hadrian’s Wall was built (Hodgson et al. 2012, 212-13).
Following the building of Hadrian’s Wall the Stanegate continued in use as the most direct arterial route across the Isthmus. The major fort sites at Corbridge and Vindolanda continued to be occupied, the former becoming the major town of the eastern frontier zone, the latter remaining in military occupation.
The earliest evidence for occupation at Corbridge is the Flavian campaign base fort at Red House (Hanson et al. 1979), probably abandoned around AD 87/88 when the new fort was constructed to the east, where evidence for at least four superimposed forts has been found (Dore and Bishop 1988), the latest evacuated around 160. The site then developed into a legionary supply-base and civil town, occupied until the end of the Roman period.
Vindolanda has been the site of a number of antiquarian investigations, and, in the 20th century, work by Eric Birley (1930-36), Robin Birley (1949- 69) and major campaigns principally by the Vindolanda Trust from 1970 (Bidwell 1985; Birley 1977; 1995; 2000; 2002). Numerous phases of fort construction have been recognised varying in size from around 1.4ha in Period I (c. AD 85-90) to greater than 2.8ha during Period IV (AD 104/105-c. 120), with the first stone fort in the sequence dating from the Hadrianic or mid-Antonine period. From the largely pre-Hadrianic timber fort sequence have come the internationally important well-preserved wooden writing tablets (Bowman and Thomas 1983; 1994; 1996), mostly published and the largest collection of new Latin documents from anywhere in the western empire.
In the early-3rd century the fort went through an unusual phase when, following demolition, a series of at least 300 small stone roundhouses were built across the site. They were replaced after a short time by another fort, and subsequent phases of building and demolition followed. At a very late date an apsidal building was built over the eastern side of the commandant’s house; this may be a Christian church. There is some indication that the east wall was repaired in the early 5th century through the construction of an earth bank. The status of the late-Roman fort found beneath the church at Newbrough (Wright 1958) remains uncertain, and no trace of the pre-Hadrianic auxiliary fort that might be expected here as ever been seen.
As a World Heritage Site and the premier Roman archaeological monument in the region, Hadrian’s Wall has been intensively researched and the body of archaeological literature is enormous and ever growing. Even a bald summary of recent developments would take up a disproportionate amount of space in a multi-period regional research framework, and here it is intended to highlight the key elements of the resource and signpost the most significant attempts to synthesise information. Here the importance of the new Archaeological Map of the Wall (2010) should be highlighted, and special attention drawn to Breeze 2006, the most authoritative and up-to-date description of all of the visible and buried remains, with the only site-by-site bibliography available for the Wall. The Hadrian’s Wall Research Framework (Symonds and Mason 2009), currently undergoing revision, naturally synthesises much past and current research on the Wall. Recent research is summarised in the handbooks for the decennial Hadrian’s Wall Pilgrimage (Bidwell 1999; Hodgson 2009 – another will appear in 2019). The standard history of the Wall using the archaeological evidence is Breeze and Dobson’s Hadrian’s Wall (first published 1976; 4th edition, 2000); see also Hodgson 2017 for a recent historical account and general study.
The exact course of the Wall has always been unclear in some of the more built-up areas around its eastern end. Recent excavations, however, have transformed our knowledge about Hadrian’s Wall in urban Tyneside, including for several areas, such as Shields Road, Byker, a true understanding of the course of the Wall for the first time (e.g. McKelvey 2003). For Newcastle upon Tyne itself, where much of the exact location of the Wall remains uncertain, a new compilation of knowledge (Heslop 2013) now superseded previous accounts. It remains true that with the exception of the Westgate Road milecastle discovered in 1985, no minor structure (milecastle or turret) can be located with confidence between Wallsend and Turret 7B at West Denton, a situation now more realistically recognised on the new ‘Archaeological Map’. This does, however, include significant recent discoveries such as the defensive obstacles on the berm, found since 2000 at several places in urban Tyneside and possibly an addition to the general anatomy of the Wall (Bidwell 2005). The excavation of a newly revealed 50m length of the Narrow gauge Wall just west of Wallsend fort, completed in 2016, probably represents the most comprehensive examination of the internal structure and construction methods of the Wall carried out anywhere (publication in prep.). The planning and construction order of the Wall in this area remains controversial (Hill 2001; Bidwell 2003).
Further west, where for over 25 miles the Wall curtain lies beneath the modern ‘Military Road’(B6318), there has been little initiative or opportunity since the 1980s to investigate the curtain and minor structures (but note Wilmott 2009). Nevertheless here and in the central sector, the widely accepted construction order and timetable of Wall-building has undergone much reconsideration on the basis of the re-examination of older records: Symonds 2005 and Grafstaal 2012 are examples of new approaches based on meticulous use of the existing information resource. There is indeed an extensive record and earlier discussion of work on the minor installations of the Wall which still repays study (e.g. Bellhouse 1969; Brewis 1932; Charlesworth 1977; Crow 1991b; Dobson 1986; Hill and Dobson 1992; Hill 1997; 2001; Welfare 2000; Wilmott 1999). Including South Shields but excluding the Stanegate sites, 12 of the forts along the Wall lie within the North-east region, with highly variable states of knowledge and preservation. They might be grouped as follows:
Of these South Shields is of outstanding interest because of its conversion into a supply-base in the early-3rd century. Newcastle is a late-2nd or early-3rd century addition to the series of Wall-forts, with internal arrangements that are therefore non-standard and not fully understood. Wallsend and Benwell were ‘normal’ Wall-forts, founded under Hadrian for auxiliary units. Newcastle lies under a Norman castle (and before that an Anglo-Saxon cemetery and presumed monastery). The others were levelled and severely damaged during the agricultural revolution and subsequently subsumed under modern building or industrial developments. At South Shields and Wallsend long-running excavations since the 1970s following the demolition of overlying housing mean that there is paradoxically a greater knowledge of the structural sequence and overall phase plans of these sites than at most better preserved Wall-forts in rural areas, and they have produced large assemblages of finds which come from recorded and often closely dated archaeological contexts. Work at Wallsend is almost fully published (Hodgson 2003; Rushworth and Croom 2016), that at South Shields partially so (Allason-Jones and Miket 1984; Bidwell and Hodgson 1999; Bidwell and Speak 1994; Gillam and Dore 1979; Miket 1983; Hodgson and Bidwell 2004; Hodgson 2009, 61-72, with further refs.). Even at Benwell, excavated in the 1930s, there is considerable knowledge of the overall fort plan due to the inspired rescue work of Simpson and Richmond, while Newcastle, much more fragmentarily known, has considerable published information (Snape and Bidwell 2002; Heslop 2013). All of the Tyneside forts have also yielded information about the extent and nature of their extra-mural settlements, in all cases evidently preserved beneath the modern buildings that still surround the fort sites, although the urban research environment makes it impossible to recover overall plans. Published work at South Shields includes a rare cemetery excavation (Snape 1994) and one of the few published modern excavations in a vicus (Snape et al. 2010).
These have all suffered to a greater or lesser extent from stone-robbing and agricultural erosion. Even with the geophysical plans of Halton Chesters (Berry and Taylor 1997; Taylor et al. 2000) and Carvoran (Biggins and Taylor 2016), which presumably show the barrack accommodation in its 3rd or 4th century state, overall knowledge of the plans of these sites in different phases of their history is very poor. Influential and probably erroneous historical conclusions were based in the past on very limited interventions at Rudchester (Newman et al. 1973) and Halton Chesters (Simpson and Richmond 1935; Jarrett 1959). Work at Halton Chesters by John Gillam of 1960-1 is now fully published, with revised conclusions (Dore 2009). Recent field survey has taken place at Rudchester (Bowden and Blood 1991), while the geophysical work at Halton has illustrated the extent of its vicus.
The fort at Carrawburgh has seen investigation by Eric Birley in the 1930s and by David Breeze and Dorothy Charlesworth in the 1960s, focusing mainly on the defences and the headquarters building (Birley 1935; 1961, 175-178; Breeze 1972; Charlesworth 1967). Here despite Coventina’s Well, and the Mithraeum, discovered in 1949, knowledge of the extra-mural settlement is poor.
Knowledge of Great Chesters (Aesica) relies almost entirely on excavation in the 1890s (Gibson 1903). The defences were re-examined in the 1920s (Hull 1927), with further work in the 1950s (Birley 1961, 191). Many of the finds have been published, including the important Aesica hoard (Allason-Jones 1996; Charlesworth 1973). Carvoran (Magnis) is the most westerly fort in the region. It stands at the point where the Stanegate and the Maiden Way join. A site here presumably predated the Wall and the visible fort, built in 136-7, is separated from the Wall-system by the Vallum. Geophysical survey has failed to reveal a proposed larger pre-Hadrianic fort but has revealed an extensive civil settlement clustering around the fort and lining the Stanegate (Biggins and Taylor 2016). Evaluation trenching by the Vindolanda Trust has shown the fabric of the fort defences to be poorly preserved, but the geophysical survey suggests good internal preservation. As with most of the forts in this section, our knowledge of its history remains based largely on antiquarian observations and the epigraphic record.
These two forts owe the visibility of their extensive stone remains to the
19th century excavation campaigns of John Clayton. There has been only the most limited twentieth century excavation work at Chesters, usually related to the consolidation of the exposed structures (though the fort interior and limited areas of the vicus have been geophysically surveyed), so our knowledge of this site depends almost entirely on the antiquarian and epigraphic record, and on the remains that can be studied on the ground. The remains are of exceptional importance architecturally, including some of the best preserved remains of fort gates, barracks, baths and one of the finest examples of a closely dated, Hadrianic principia to be seen anywhere in the Roman empire. An extensive vicus is known principally from air photography.
Housesteads (Vercovicium) has been the site of extensive but by no means total excavation and is relatively well understood, with both excavation work and geophysical survey extending into the vicus and fort environs. Following early excavation by Hodgson, and Clayton’s clearances, a systematic attempt to recover the fort plan using trial-trenching was carried out by Robert Carr Bosanquet (Bosanquet 1899; 1904), the resulting plan standing, for much of the 20th century, as the type-site for an auxiliary fort. Further work took place at the site in 1930s (Birley 1937; Birley and Keeney 1935; Charlton and Birley 1934; Charlton et al. 1933; Hedley et al. 1933) and then long-term work developing into ambitious area-excavation took place from the 1960s (Charlesworth 1975; 1976; Crow 1989; 1988; Tait 1963; Wilkes 1960; Wilkes and Leach 1962), the most recent work now definitively reported and all knowledge synthesised in the monumental report by Rushworth (2009). See also Allason-Jones 2013, where it is suggested that civilian occupation of the vicus might have continued through the Antonine Wall period. Both Chesters and Housesteads are important not only for their histories as Wall-forts, but for their relationship to the Wall as first planned, that is, without forts. At Chesters, Turret 27a and a length of Broad Wall had to be removed to make way for the fort, which straddles the Wall. Housesteads also overlay a turret (36b) and although the fort was not designed to project north of the wall, as at Chesters, it overlay the turret and Broad Wall foundation, and when the Narrow Wall was eventually completed in this sector it was brought up to abut the northern corners of the fort on a completely different, more northerly course than that of the original Broad Wall. A considerable interval in that area between the abandonment of Broad Wall and completion of Narrow Wall (during which the fort was constructed) is implied.
The general distribution of forts in County Durham is, not surprisingly, closely linked to the road network. Although the last few years has seen the final publication of reports on landmark excavations of the 1970s and 80s at Binchester and Piercebridge, and some significant fieldwork and survey at Binchester and Lanchester, in general the internal anatomy and history of the ‘hinterland forts’ is more obscure than on the Wall. The remote fort of Whitley Castle, in the south-east corner of Northumberland, which supervised lead-mining in the Roman period, has been subject to a major analytical survey and geophysics has outlined its interior plan (Went and Ainsworth 2013). Bidwell and Hodgson 2009 is a recent gazetteer and discussion of all military sites southof the Wall. The County Durham forts fall into three principal groups related to routes:
Heading south from Corbridge, a series of forts is situated along Dere Street. The fort at Ebchester (Vindomora) has mainly been built on, though some areas of the ramparts are visible. Small scale excavation has indicated several phases, four in timber and three in stone, extending from the Flavian period to the later-4th century (Harper et al. 1964; Reed and Maxfield 1975).
The fort at Lanchester (Longovicium) is upstanding but has seen almost no scientific excavation. Small-scale work was carried out here by Kenneth Steer (Steer 1938; 1939; Swinbank 1953) and more recently there has been geophysical survey and excavation of a related cemetery (Casey et al. 1993; Turner 1990). The geophysical survey has spectacularly revealed the fort plan, and combined with the epigraphic evidence (Collingwood and Wright 1965, nos 1083, 1093, 1091-92), suggests a mid-Antonine date for the first phase of the fort, which was occupied until the end of the Roman period. This work is now complemented by geophysical survey that has revealed an extensive vicus lining Dere Street east of the fort (Hale et al. 2010).
Binchester (Vinovium), standing on the Wear, was of Flavian origin. The first significant excavations here took place in the 19th century (Hoopell 1879), then between the wars, and later between the 1950s and 1980s (Dobson and Jarrett 1958; Jarrett 1960; Ferris 2010). As well as excavation, there have been several recent geophysical surveys carried out at the site, the latest (in 2004, 2007 and 2009) revealing an even larger predecessor of the known 3.9ha fort and an exceptionally clear plan of the extensive vicus (Geoquest 2004). Between 2009 and 2017 the most ambitious campaign of open area excavation ever to be mounted in one of the ‘hinterland’ forts has explored barracks, roads and defences within the eastern quadrant of the fort, part of the vicus and extra-mural baths (extraordinarily well-preserved) and part of a later-Roman cemetery. When synthesised, the results will add considerably to our understanding of the Roman north in general. The final fort, Piercebridge, commanded the crossing of the Tees. The very large known fort is of 3rd-century date, though earlier military activity extending back to the Flavian period is to be expected. A civilian settlement lay along the main road to the east of the fort, extending along both sides of the Tees. Small-scale excavations took place on the fort and civilian settlement between 1934 and 1964 (Richardson 1934-36; Harper 1965- 68) and more extensive excavation was carried out from the early 1970s onwards by the late Peter Scott, now definitively published (Cool and Mason 2008). Substantial quantities of Roman finds, mainly coinage and metalwork, have been recovered from the river at Piercebridge, forming ritually deposited assemblage of international importance (Walton 2008).
The only fort on this road known to date is Chester-le-Street (Concangis). Extensively though fragmentarily investigated where opportunities have arisen in the overlying modern town, it is only in recent decades that the exact location and layout of the fort walls have been appreciated (Gillam and Tait 1968; Evans et al. 1991; Bishop 1993; Turnbull 1994, 2; 2003, 2). The foundation date of the site is uncertain but generally thought to follow the re-establishment of Hadrian’s Wall in the 160s AD. The stages by which the known stone fort developed are equally unclear. A turf and clay rampart has been found preceding the west side of the known fort and, apparently in use until the later-3rd century (Bishop 1993), this possibly represents the west defences of the earlier fort, or an annexe belonging to an earlier and smaller stone fort. As at Binchester the fort housed a cavalry ala in the 3rd century. The stone fort in its final large form dates to the late-3rd century. The final stone fort is now known, thanks to work by a local amateur group, to extend further south than previously thought, giving the fort a rectangular rather than square shape and increasing its area to over 3ha (Mason 2010). The same group has rediscovered a building with hypocaust reported in the mid-19th century (Featherstonehaugh 1855), almost certainly the extra-mural baths, 80m south of the south-east angle. Development control work has located an outlying settlement extending over 1km along the road running north from the fort (Platell 2014).
Apart from the two main north-south roads, the other major axis of communication is west across the Pennines into Westmorland. The route used by the Romans ran over the Stainmore Pass, and roughly followed the route of the modern A66. As well as the military activity on Stainmore itself, there were two forts on this route at Greta Bridge and Bowes.
The fort at Greta Bridge, as the name suggests, was situated at a crossing point on the River Greta. The fort survives as an impressive earthwork but apart from a small, recent evaluation excavation in the northwest corner (NAA 1996), there has been little work on the fort itself, and its history remains obscure. An important excavation in its related vicus is however fully published (Casey and Hoffman 1998). The 1.7ha fort at Bowes (Lavatrae) was founded during the Flavian conquest of c. 71, and rebuilt in stone in 130-33. A very extensive civilian settlement extended along the road bypassing the fort. Bowes underwent major refurbishment with its stone walls and principia rebuilt in the late-3rd or early-4th. Earlier research on Bowes (Wooler 1913) is now complemented by the extended reconsideration and final publication of excavations by Brian Hartley and Sheppard Frere in the 1960s (Frere and Fitts 2009).
As with the forts to the south of the Wall, the northern outpost forts are also closely related to the road system. The foundation for the study of these northern forts remains Ian Richmond’s article ‘The Romans in Redesdale’ (1940). Most of the major forts are located along Dere Street, the main route taken by Roman forces into Scotland.
Heading north from Corbridge, the first fort after crossing the Wall was Risingham (Habitancum), scene of ambitious excavations in 1840-49, of which there is some record and which yielded a rich haul of inscriptions and architectural information. Since then there has been only small-scale excavation in the 1930s (Richmond 1936), but the fort interior has now been graphically revealed by geophysical survey (Biggins et al. 2014). Geophysical survey has been carried out across the floodplain in an attempt to locate the course of Dere Street, producing inconclusive results (Anderson et al. 1992).
The pre-Hadrianic Dere Street fort of Blakehope (Birley 1961, 240-2) has seen only small-scale investigation.
Further north, at High Rochester (Bremenium), the earliest excavation, ordered by the Duke of Northumberland, took place in the mid-19th century, providing one of the earliest recorded plans of a fort interior to be recovered anywhere in the Roman empire. Small-scale excavation was carried out by Richmond in 1935 (Richmond 1936; 1940). More recent survey geophysical has revealed details of the western annexe (Hancke et al. 2004), but survey and small-scale excavation has revealed the notable absence of any significant civilian settlement related to the fort (Crow 2004).
Risingham and High Rochester represent a remarkable resource, being amongst the best preserved and least disturbed fort sites in the empire, rich in epigraphic material, and they are revealing of conditions north of the Wall: the forts are small compared to those south of the barrier, and a tight-fit for the size of their epigraphically attested units (nominally 1000 strong , plus irregulars); they lack evidence for external civilian settlements; and the internal structures are unusually closely arranged, the whole presumably a response to the isolated and vulnerable location of these forts.
Fewer forts are known related to the Devil’s Causeway, perhaps simply a matter of archaeological visibility, as their lifespans were too early and short for them to have seen stone construction: further sites along this road may remain to be discovered. The only certain fort is Low Learchild, which may have a Flavian origin and was perhaps occupied until the Hadrianic period. Small-scale excavation in the 1940s and 50s has been supplemented by more recent geophysical survey (Crawford and Richmond 1949; Birley 1961, 244-5; Anderson et al. 1992) which suggests that the fort is significantly larger than Richmond originally estimated. Another possible military site on this road has recently been located at Wooperton Quarry, where a series of pits containing a large quantity of early Roman pottery have been discovered, along with some possible structural remains (Carter 1998; 1999; Gates and Hewitt 2007). An early fort (and perhaps a port) in the Tweedmouth area probably awaits discovery. An analytical survey of the fortlet at Longshaws, near the Devil’s Causeway crossing of the Font, is now published (Welfare 2011). Ongoing analysis of historic finds and Roman artefacts found through metal-detecting in north Northumberland has revealed appreciable amounts of material along the coast just south of Berwick-upon-Tweed and in the Norham area.
There is no firm evidence from the north-east coast to date of any extension of the well-known chain of fort-tower combinations provided along the North Yorkshire coast in the late-4th century – the so called ‘Yorkshire coast signal stations’. The possibility that this frontline coast between South Shields and the Yorkshire installations was not watched and guarded by the late Roman era (if not before) seems implausible given other protective measures. This apparent absence may be real, but is equally likely to be due to loss or obscuring as a consequence of coastal erosion and urban/industrial development, together with a lack of concerted searching along the Durham coastal margin. Dobson had speculated on the existence of a system similar to that of the Cumbrian coast, pointing to finds from Monkwearmouth and Seaham (Dobson 1970, 35). A late Roman reformation of such defences may be implied by piecemeal finds that cluster at several ‘high point’ locations near the coast: late Roman pottery from the north side of Pincushion Dene, south of Ryhope, from the area of the former Vane Tempest coal mine to the north of Seaham, from the area of the modern cemetery on the southwest side of Seaham, at Beacon Hill north of Easington Colliery (at 85m OD), at Yoden, Peterlee (100m OD) and from Backhills Gill, Horden (Willis 1998). Roman objects found at Whitburn in the late 1970s may indicate a site under the gunnery range (Allason-Jones, pers. comm.). While some of these collections could relate to settlements their topographic settings would be consistent with ‘Signal Stations’. A naval unit placed by the Notitia Dignitatum at Arbeia – probably South Shields – attests to concern with coastal patrolling and defence. Systems of inland signal stations have also been suggested (Richmond 1951). A possible signal tower on the Stainmore Pass has been excavated (Annis 2001), though the extent to which any meaningful signals could have been passed is debateable, and the excavator prefers to call them simply ‘towers’. Like the East Yorkshire examples they are late-4th century and seemingly only had a short period of use.
Although the extra-mural civilian settlements of traders, soldiers’ dependants and veterans – vici – attached to forts have been noted above where their remains are a conspicuous feature of an individual site, some general remarks are appropriate. The North-east region has one of the richest resources available in the empire for the study of such communities. While there are no published large-scale excavations of vici, knowledge of the extent and complexity of these settlements has been greatly advanced in recent years by a series of geophysical surveys, noted above for individual sites. These surveys have allowed a sophisticated comparative study and typological classification of north-British vici which places them in an empire-wide context (Sommer 2006). The presence of vici can be assumed at all fort sites in our region, even where physical traces have not been seen, except in the case of the outpost forts of Risingham and High Rochester: at the latter site a search for the vicus concluded that it had never existed.
It is now widely accepted that most vici had drastically contracted or had been almost wholly abandoned by the end of the 3rd century. The coin-list from the only really extensively excavated example, at Vindolanda, suggests that it was abandoned by about 270; excavated samples of other military vici on the Wall (South Shields; Wallsend; Halton Chesters; Housesteads; Burgh-by-Sands; Maryport) have produced only stray 4th-century finds. The story is much the same at the few excavated vici of the forts south of the Wall, but there are some exceptions (see under Towns). In the later-Roman period some trading and marketing activities continued to take place at forts, but apparently inside the walls. The substantial number of coins found in the fort at Newcastle (dating from 270s to 364- 375), for example, suggests that this may have been a market site (Brickstock 2002). Similar phenomena have been recorded at Wallsend and Vindolanda.
Traditionally, archaeological research across much of the North-East has been dominated by the highly visible Roman military sites, and in the civilian sphere by visible remains in remote upland areas. Technical advances, and especially the growth of developer-funded archaeology in 21st century, are rapidly changing this situation. Since the first version of this resource assessment there has been an enhanced awareness of formerly invisible rural settlements in lowland areas, and crucially, the ability to date them using scientific techniques, so that it is possible to characterise the rural landscapes of the Roman period in a way that was not possible before. It is important to stress, however, that the preliminary conclusions drawn in what follows are based on what remains a very small sample of excavated and scientifically dated sites. Despite the fact that modern survey and evaluation techniques regularly encounter previously unknown examples of a now familiar form of late-Iron Age/early-Roman period enclosed settlement, the number of sites whose sequence is well-dated by multiple radiocarbon measurements remains very small, so for all the familiarity of the site type it is essential that as many as possible are excavated in the future – and an appropriate number of radiocarbon dates obtained – to allow a fuller scientifically dated picture of the rural settlement history of the region to be assembled. Of the excavated sites, only a tiny fraction of those with extensively excavated enclosed or unenclosed nuclei have seen any serious attention paid to their environs: subsidiary enclosures, trackways, and outlying compounds and field systems. This is an area where knowledge is very poor. Only more extensive work on a seemingly familiar site types will still yield significant information about the outlying areas of what are now understood to have been extremely complex settlements with multiple peripheral functions and activity areas.
The Tees has traditionally been seen as the northern boundary of Roman villa landscapes, but there is increasing evidence for a distribution of villas in the Tees lowlands, with one partly excavated example at Ingleby Barwick, Stockton now published in detail (Willis and Carne 2013). Here the development of the villa began in the mid-2nd century, and the complex was in use throughout the Roman period. Other villa sites are known at Holme House near Piercebridge (Cool and Mason 2008, 126-57), and Dalton-on-Tees (just outside the region in North Yorkshire), while these discoveries tend to rehabilitate the remains long known at Old Durham, only 22km south of Hadrian’s Wall, as being those of a villa (Romans et al. 1944).
An important and pioneering excavation of an indigenous rural site in the County Durham was that at Thorpe Thewles (Heslop 1987). This site showed continuity from the Iron Age into the 1st century AD, and saw an initial enclosure around a central house being filled in as the settlement expanded. However, by the mid-2nd century the site was abandoned and succeeded by field systems, suggesting settlement shift and landscape reorganisation at this time. Further examples are now available of rural sites developing along new lines in the earlier-Roman period. A site at Faverdale, near Darlington saw the rapid development in the Hadrianic period (around 120-140) of a roundhouse settlement into an enclosure complex making conspicuous use of Roman pottery, metalwork and building technologies (Proctor 2012a). There may possibly be an undiscovered villa building at this production site, which was probably supplying the military market.
It is therefore possible to discern a horizon of change in the immediate hinterland of Hadrian’s Wall which could speculatively be connected to the permanent presence of the frontier and the need to supply the military communities of the region. The Ingleby Barwick villa and Faverdale point in this direction, as does a previously unsuspected well-preserved linear settlement revealed by survey and excavation at East Park, Sedgefield. This site too has been shown to have developed in the second century, not long after the Wall was built. Extensive geophysical survey has revealed much detail of the morphology of this site, resembling what in southern Roman Britain would be termed a roadside settlement (Hale 2010). This type of site has no Iron Age antecedents in the region.
Nonetheless it is still difficult to characterise rural settlement of the later-Roman period (3rd and 4th centuries) in County Durham. Catcote (Teesside) is an example of an Iron Age site that continues in use through the Roman period. Excavations on the site revealed a rectangular stone building, possibly a grain store. Other sites with possible later-Roman developments include Dixon’s Bank and Bonny Grove Farm, Middlesbrough (Cleveland) (Annis 1996), and Newton Bewley (Teesside) (Platell 1999), notable for its field system evidence. Much of this evidence awaits publication in detail, but available knowledge (relating to the Roman period despite the title of his book) has been usefully synthesised by Sherlock (2012, esp. 113-136 and Appendices 1, 2 and 4). Later Roman occupation is also evidenced at Apperley Dene on Dere Street, originally interpreted as a Roman fortlet, before re-excavation showed it to be a re-used Iron Age enclosure (Greene 1978).
It is equally difficult to know the fate in the later-Roman period of settlements in the upland areas of the North Pennines such as Forcegarth Pastures and Dubby Sike, high in Teesdale (Fairless and Coggins 1980; 1986; Gidney and Coggins 1988), or an iron-smithing site with 2nd century AD material recently excavated on Bollihope Common, in Weardale (pers. comm. Rob Young). The lack of either later-Roman material or the emergence of a different form of settlement might reflect either abandonment or simply the isolation of a remote area. The only contemporary settlement to be excavated in the Wall-Vallum corridor is still Milking Gap (Kilbride-Jones 1938), which contained late-1st to 2nd century Roman pottery. This site was probably abandoned when the Wall was built. Other similar sites in the area, such as Fold Hill, near Sewingshields, and Green Brae, Crindledykes (both discovered during the aerial photographic survey of the Hadrian’s Wall corridor: Gates 2004, 238-239) may well also have been in occupation when the construction of the Wall was ordered. Thornborough Scar, a roundhouse settlement near Corbridge excavated over 30 years ago, was in occupation until the second century but remains unpublished.
The predominant rural settlement type north of the Wall has been strongly characterised in the past as the ditched and banked rectilinear enclosure (sides typically 40-50m long) containing a few roundhouses, sometimes a large central house. The buildings were of timber on the coastal plain, sometimes of stone in the uplands. Knowledge of these sites relied heavily on the particular research interests and fieldwork of one scholar, George Jobey, who was responsible for identifying and investigating many of them (e.g. Jobey 1960; 1978). In more recent times this work has been supplemented by aerial photography (Gates 1997; 1999; 2000; 2004), which has revealed extensive landscapes, but much of the dating of the newly discovered sites has been based on assumptions about settlement morphology and has lacked confirmation by excavation. There are also morphological variations within the region. Settlements close to Hadrian’s Wall, particularly in North Tynedale and Redesdale, and on the Northumberland coastal plain, tend to be rectilinear in shape, in contrast to pockets of curvilinear settlements which occur in places west of the Devil’s Causeway and in parts of the Cheviot uplands.
Even where excavated, until recently the settlement sites of the pre-Roman Iron Age in the region were impossible to date closely because they produce very few datable artefacts: the scatter of Roman objects from these sites have tended to be the main basis for dating, leading to a general belief that roundhouse settlements continued in occupation to either side of the Wall, indeed achieving their most developed form in the Roman period. The last 10 years has at last yielded new information about native settlement north of the Wall, thanks largely to large-scale developer-funded archaeology in lowland areas (Proctor 2009; Hodgson et al. 2012; numerous other sites currently being investigated or prepared for publication). This work has shifted the balance away from the uplands and documented dense late Iron Age rural settlement in the coastal plain areas. Crucially, the financial resources made available through developer-funding mean that it is now possible to date sites which are poor in artefacts, by means of extensive programmes of radiocarbon dating examined within frameworks of Bayesian statistical analysis (Hamilton 2010; Hamilton et al. 2015) These newly available techniques offer the prospect of revisiting regional collections from the later 20th century to establish the potential for a retrospective carbon dating programme based on samples of bone, macrofossils or carbonized residue on pottery.
All earthwork-enclosures which have been radiocarbon dated in this way appear to have been formed around 200 BC as the latest phases on palisaded or unenclosed open roundhouse settlements continuously occupied since the late Bronze Age. Considerable social complexity is implied by the range of settlement plans now available: the late-Iron Age settlements often occurred in pairs or clusters, and there is increasing evidence for outworks, subsidiary enclosures and field systems. A social hierarchy is implied by the discovery of occasional contemporary small unenclosed roundhouse settlements (not previously known in the region) and agglomerations of small-ditched enclosures that may have been dependent on the heavily enclosed sites that up to now have monopolised the archaeological record.
Recently discovered and excavated sites such as West Brunton and Pegswood continued in use beyond the Roman conquest of the 1st century, and this is the case at other recently excavated enclosures north of the Wall, such as Station Road, Wallsend, which have produced 1st and 2nd century Roman pottery. At Hetha Burn, in the remote uplands, a settlement with two round houses expanded into a village with around ten houses (Burgess 1984), a process continuing into the early-Roman period.
Nevertheless, the radiocarbon evidence and structural evidence combined suggests that at most lowland pre-Roman Iron Age settlements north of the Wall, either occupation ceased by some date in the 2nd century AD, or the sites were completely reorganised along new lines. The imposition of Hadrian’s Wall from the 120s therefore seems to bring about dislocation and social change in the area to its north. The two settlements excavated at Newcastle Great Park (East and West Brunton) were in use up to the early 2nd century on radiocarbon evidence, but were abandoned before more than a trace of Roman ceramic material could reach them (Hodgson et al. 2012). At Pegswood (Northumberland), there was continuity from the Iron Age into the Roman period, but by the 2nd century AD there was a significant phase of replanning (Proctor 2009) with the former settlement of roundhouses and small-ditched enclosures swept away and replaced by a possible stock enclosure. An Iron Age salt production and trading centre at Needles Eye, Berwick (Proctor 2012b) ceased to function before or around the time of the Roman conquest, but some kind of occupation on the site continued up till the 2nd century AD.
A significant gap remains in the present archaeological resource in that it is still difficult to recognise sites north of the Wall with evidence for occupation the 3rd or 4th centuries. One exception is Huckhoe (Jobey 1959), which had rectangular buildings and a pottery assemblage that appears to continue into at least the 3rd century AD. The re-ordering of traditional society which had evidently been a consequence of the permanent imposition of Hadrian’s Wall on the region may have led to the emergence of new centres of authority, a shift in the economic basis of society (from agriculture to cattle-rearing or hose-raising perhaps), and the emergence of less substantial and therefore less archaeologically recognisable forms of settlement. One recent discovery which may indicate the character of later-Roman site types is at St George’s Hospital, Morpeth, where a slightly ditched enclosure of irregular shape, apparently used for stock collection, has produced radiocarbon dates spanning the Roman period. It has some resemblance to the possible stock enclosure which replaced the traditional Iron Age site at Pegswood. The rearing and collection of livestock raises the possibility that the leaders of these communities north of the Wall were supplying the Roman military, through a tax or trade mechanism (Mercer 2018). Possibly relevant in this connection, approximately 1 mile northeast of the Wall fort of Halton Chesters and immediately to the south of the Devil’s Causeway junction with Dere Street, the fields of Great Whittington have produced an abnormal amount of Roman artefacts (Collins and Biggins 2013). This includes single finds of coins and a small coin group dating to the early-mid 5th century (Collins 2008), metal vessels, militaria, dress accessories, and religious objects. Limited geophysical survey did not reveal traces of any structural remains, but the spread of artefacts is considerably wider than was feasible to survey. While the exact nature of the activity around Great Whittington cannot be confirmed, it can be hypothesised that the area served as a potential settlement or market area immediately north of the Wall and proximal to two important roads.
Roman period occupation is thought to have occurred at some long-abandoned hillfort sites, including West Hill (Frodsham 1999; Oswald 2008, 27-34), where a D-shaped enclosure containing roundhouses – sometimes described as a ‘village’ – superseded the hillfort defences. Unfortunately, excavated dating evidence from such sites is scarce and there is ambiguity about whether such developments commenced in the late-Iron Age or the Roman period. There are a few cases where hillforts appear to have been slighted or abandoned in the early Roman period, such as Dod Law (Smith 1988-89), and a recent study has discussed the possible refortification of Yeavering Bell (Hope-Taylor 1977, 267) by a Brittonic community in the sub-Roman period (Miket 2013). Finally, it should be noted that rural settlements, whether of Iron Age or Roman date, are no longer confidently ascribed to recorded ‘tribes’ and there is no archaeological support for the idea of the Tyne as dividing line between the ‘Brigantes’ and the ‘Votadini’ (Allason-Jones 2009; Hodgson 2012, 211; Haselgrove 2016).
Until recently it was assumed that there was no early development of towns in the north-east, and during the 2nd century the forts and their vici substituted for the kind of urban infrastructure of small towns and roadside settlements that developed in the south of the province. However, a hitherto unknown site was revealed by chance in 2001-3 following metal detecting at East Park, Sedgefield. Survey and selective excavation revealed a linear settlement extending for over 500m along ‘Cade’s Road’. The site layout comprises ditched compounds containing rectilinear timber structures, presumably property plots and activity areas arranged along the road, plus side tracks and ancillary enclosures, in a ladder-settlement form (Hale 2010). Limited excavation to date has revealed evidence for industrial production, including pottery manufacture and metalworking. Although yet to be fully analysed, the ceramic assemblage indicates activity beginning in the second century and continuing for much of the Roman period. There is no evidence for a military presence but this substantial site seems likely to have been involved in military supply. The discovery of the Sedgefield site makes it highly likely that others may exist in the County Durham/Teesside area.
Corbridge was probably constituted as a civil town by the early-3rd century (as its equivalent in the west, Carlisle, certainly was) and may well have been an administrative centre for much of the region. Unfortunately the civil town (as opposed to the military base from which it developed, and the military enclave which continued to exist at its core) is only known from antiquarian excavation and air photography, although in 2017 it was geophysically surveyed for the first time. Civilian settlements attached to military bases just south of the region at Catterick and Malton have continued activity in the 4th century, long after most military vici had disappeared, suggesting that they survived and prospered as regional market centres in an area where there were few civilian towns. The same might be true of Binchester in County Durham. These quasi-urban centres, along with established towns, like Corbridge, probably provide the answer to the question of where civilian traders and soldiers’ dependants resided for much of the 4th century.
Because of the lack of a pre-Roman Iron Age pottery tradition in the region that would suit the army’s requirements, pottery production was carried out locally by the army in the early (especially pre-Hadrianic) period before external sources of supply for non-fine wares became established. There has been considerable advance in recent years in recognising and categorising the coarse grey, fawn or brownish-yellow coloured wares that the army at first produced in its own kilns. From the beginning the army seems to have been accompanied by communities of civilian traders and manufacturers, who also carried out these activities on its behalf (camp-followers rather than indigenous inhabitants of the region), whose productive activities are to be sought in the vicinity of military bases such as Corbridge and Binchester. However, there is now increasing evidence that production for the military market rapidly became widely dispersed in rural society, at least south of Hadrian’s Wall, for example the discovery that products of the prolific 2nd century mortarium maker ANAUS were made at Faverdale (Proctor 2012a), and that pottery, presumably for the military market, was manufactured at Sedgefield, where some of the few actual kilns known in the north-east have been identified (Gallagher 2004); details await publication. Early kilns have also been recorded at Piercebridge (Swan 1984, 87). Roman building materials such as roof tiles and hypocaust box tiles were also acquired by or made in these rural communities, as in 2nd century Faverdale. Tiles for the military were always made locally (kilns were also probably located at South Shields and Binchester (Gillam and Dore 1979, 29-32, figs 9-10)), but from the Hadrianic period onwards the army was increasingly supplied with coarse pottery and finewares from industries outside the region.
Hides and other animal products were also potentially supplied from distant rural communities rather than the immediate locality. The environmental conditions from Vindolanda have led to excellent preservation of leather, giving a clear indication of the material requirements and the sheer scale of leather working required by a military installation, including both shoes, horsegear and fittings (van Driel-Murray 1989; 2001), and at Vindolanda the writing tablets explicitly attest animal and agricultural products being transported over a considerable distance (for example, from Catterick) to Vindolanda. Across the region as a whole, leather working is not as clearly attested in later-Roman times because later drier levels are not conducive to preservation.
Like leather-working, ferrous and non-ferrous metal-working was carried out at a wide range of military and civilian sites, widely attested by furnaces, hearths and residues at all periods. It is ubiquitous at fort and vicus sites and occurs at milecastles, such as Sewingshields, and even turrets (26a, 18B) (Allason-Jones and Dungworth 1997) and native sites, such as Huckhoe (Jobey 1959). A gold and silver-smiths’ workshop has been identified in a 3rd century extra-mural context at South Shields and there is evidence of gold-working at Halton Chesters. A late-2nd century hoard of Roman coins north of the Wall at Longhorsley shows evidence of melting down for re-use (Abdy 2003), possibly for recyling into objects such as brooches.
Actual metal extraction sites have not been identified, though it had been suggested by Woller (1924) that Roman evidence should be present in the region, while Coggins (1986) suggested a pre-Roman date for lead extraction in the Pennines. The main function of the fort at Whitley Castle in the north Pennines was almost certainly to supervise lead mining and possibly silver extraction, well-attested in this area in medieval times. Five lead sealings (found at Brough-under-Stainmore) of the unit stationed at Whitley Castle are inscribed ‘metal(la)’ – ‘the mines’. Environmental studies in Weardale have not revealed evidence for Roman lead working, suggesting instead the beginnings of metal extraction in the medieval period (Mighall et al. 2013). This study has not, however, been followed by further research of this nature in the Alston area where Whitley Castle is located. There may have been some ironstone mining at Skelly Braes, Birtley (Richmond 1955, 159), and iron smelting may have taken place at the native settlement at Tower Knowe (Jobey 1973b), though there are difficulties in distinguishing between smithing and smelting slags (Hedley 2004, 310).
There is also significant evidence for quarrying. Most obvious are the Roman quarries associated with the construction of Hadrian’s Wall. The discovery of a Hadrianic coin hoard at Thorngrafton, near Barcombe Down quarry near Vindolanda, indicates the working of quarries (Birley 1963), as does the discovery of a Roman inscription on a quarry face there (Wilson 2003). Epigraphic evidence from Haltwhistle Burn and Fallowfield Fell is also indicative of quarrying during the Roman period (Collingwood and Wright 1965, nos 1442, 1680). Without additional indicators of date like these it is difficult to evaluate the chronology of simple quarries, however, and much has probably been destroyed by later workings. Some limestone was clearly burnt for use in mortar, and lime-mixing pits have been recorded in detail at South Shields. A Roman limekiln was also found at the Knag Burn, near Housesteads (Simpson 1976, 152-156). Recent analysis of mortar samples from Wallsend suggests that the lime may have been brought all the way from the central sector of the Wall.
Stone was not only quarried for constructional purposes; its other major use was the manufacture of querns. The number of published querns available for study in the region has almost doubled in the last decade. Major survey work by David Heslop (Heslop 2008) on the distribution of querns in the south of the region (the former county of Cleveland and the Durham districts of Darlington, Sedgefield and Teesdale), and ongoing work by Heslop and John Cruise extending to the rest of the region, has enhanced our understanding of querns since the last resource assessment. Among beehive querns and disc hand querns local sub-types have been recognised. The ability to distinguish mechanically powered millstones from hand querns has developed, with the former seen as largely a later-Roman phenomenon. Data for beehive querns shows an interesting fall-off from south to north, with the north-east having a modest density (150) compared to the Yorkshire heartland (1,130), but even lower numbers in the borders and Scotland.
The sites of two watermills close to Hadrian’s Wall were identified by F .G. Simpson at Haltwhistle Burn and possibly Willowford (Simpson 1976, 32-43, 49-50); a channel which is almost certainly a mill-race was inserted into the bridge abutment at Chesters.
Research by Willis has collated the evidence for the production and distribution of salt in the Iron Age of the region as evidenced by transport briquetage (Willis 2016). This study showed an abrupt end to the trade at around the end of the 1st century AD. An important report on the Needles Eye enclosure site at Berwick (Proctor 2012b) which added to our knowledge of pre-Roman salt production and exchange in the region, showed that the site was out of use by the beginning of the Roman period. Paradoxically our knowledge of where and how salt was later procured in the Roman era remains very poor. It would seem that salt was distributed by different means and containers from the start of the Roman period in the region and may have been supplied in barrels, coarse ware jars, or in other organic containers from outside the region.
There is a little evidence for Roman glass production in the North-East. At the beginning of the period bangles were being produced from re-cycled Roman glass in the region or in adjacent areas to north and south. They are an especial feature of the period around the conquest and the half-century or so afterwards, and are commonly found on both military and indigenous sites. In later times evidence for production is limited to re-cycling of glass cullet (Price 2002), while evidence for secondary glass production (i.e. working pre-made glass into objects) in the North-East comes only from Corbridge and from Binchester, where twelve fragments of glassworking waste were recovered, one with indications that it had been blown using an iron blowpipe. Although found in late and post-Roman contexts it is more likely that these fragments belong to the late-3rd or 4th century (Price 2002, 90). Presumably glass for building projects (mainly window glass) was produced locally along with other building materials.
Coal mining, either using bell pits or open-cast methods, must have taken place as coal is known from a number of Roman sites, such as Housesteads, where it was used in iron smithing and working copper alloy (Starley 1996). Coal was widely used as fuel at South Shields where, however, it could have been gathered from the beaches.
Recent analytical work at Newcastle University has shown that there was industrial activity at South Shields working black rocks, including jet imported from Yorkshire, shales fromlothian and Derbyshire, and cannel coals from the Northumberland Coal Measures north of Hadrian’s Wall (Allason-Jones and Jones 1994; Allason-Jones 2003; 2005). Jet was also worked at Binchester. There is also evidence for armlet manufacture at Halton Chesters exploiting the Coal Measures available there (Allason-Jones 2002b, 116).
Owing to the fact that in Roman Britain ‘classical’ forms of religious expression were predominantly employed in urban and, above all, military contexts, the region possesses an internationally important corpus of religious inscriptions and sculpture, which has been studied synthetically at various times (e.g. Breeze and Zoll 1995; Irby-Massie 1999; Breeze and Dobson 2000, Appendix 3). A range of gods is recorded on these altars and other sculptural fragments, including the imperial cult, traditional Roman gods (e.g. Jupiter, Mercury), eastern gods (e.g. sculpture relating to Jupiter Dolichenus from Corbridge, and altars to the same god at Chesters), ‘Celtic’ gods (such as Cocidius and Antenociticus), and gods probably of Germanic origin, such as Veteris.
Architecturally, a number of temples are recorded from the line of the Wall itself. These include temples of classical type, attested by architectural fragments at Corbridge (Hodgson 2010: the actual buildings have probably not been found, despite Richmond’s belief that certain rectangular buildings in the central area were temple podia). Small, non-classical temples are also known, for example that dedicated to Antenociticus at Benwell (Tyne and Wear). At Vindolanda the remains of a pre-Hadrianic temple have been excavated; this is the only example of a ‘Romano-Celtic’ temple (concentric cella and ambulatory) from the North-East. It was out of use by the mid-2nd century, and after demolition the site became a focus for burial instead.
At Carrawburgh a shrine to the water goddess Coventina included a possible stone structure or precinct and a masonry-lined well. This was the focus for extensive ritual deposition from the mid-2nd to the late-4th centuries. The votive deposits included over 13,000 coins, bronzes, stone altars, pottery, glass, leatherwork, jet and shale (Allason-Jones and Mackay 1985). Nearby, there was a shrine to the Nymphs and Genius Loci (Smith 1962).
Mystery cults are represented on the Wall, both epigraphically and architecturally. There is relatively extensive evidence for Mithraism. Mithraea are known from Housesteads (Daniels 1962), Rudchester and Carrawburgh (Northumberland) (Gillam and MacIvor 1954; Richmond et al. 1951). All three sites have also produced related epigraphic and sculptural evidence, as has Lanchester. Finds of charred pine cones from Carrawburgh are significant as rare in situ evidence for the use of pine in religious ceremonies.
Although Christianity is traditionally not believed to have been strong in the army, the pagan practice of erecting inscribed altars (and presumably maintaining the temples that went with them) had ceased by the end of the 3rd century, although the pagan deposition at Coventina’s Wall continued through the 4th century. The practice of Christianity in military circles is attested, probably as early as the 3rd century, in the form of finger rings with Christian symbols, as at Binchester. Possible churches have been recognised at Vindolanda, Housesteads, and South Shields. Their exact date is not clear however, and they could potentially be post-Roman (Bidwell and Speak 1994, 44-46, 103-104; Crow 1989; Birley et al. 1999, 22). Chi-rho symbols are also present on silver vessels recovered from the Tyne at Corbridge (Petts 2003, 122).
Away from Roman forts a number of other smaller religious sites are known, though probably still military in nature. Two altars to Vinotonus have been recovered from a site on Scargill Moor (Co. Durham), together with the remains of two structures, presumably simple temples (Wright and Richmond 1948). North of the Wall a small carving of a naked horned god, probably Cocidius, has been found carved onto living rock at Yardhope, with some evidence for a simple associated structure (Charlton and Mitcheson 1983). A similar carving has also recently been discovered near Chesters fort.
The religious life of the indigenous population, and of the later Roman rural population, whether indigenous or of immigrant descent, remains stubbornly difficult to see in the archaeological record, although there are examples of ritual or votive deposition reflecting wider Roman and indeed Iron Age practice found in Britain and northern Europe, often associated with watery contexts. Coventina’s Well provides a clear example of this. A remarkable assemblage of metal and other objects recovered by divers from the Tees at Piercebridge indicates a site of intense and sustained ritual deposition (Casey 1989; Walton 2008), at its peak in the 3rd century. The discovery of a series of Roman silver plate objects from the Tyne around Corbridge and Bywell, including the famous Corbridge lanx and several other silver cups and vessels, suggests that at least one silver plate hoard may have been placed in the river (Nicholson 1995; Petts 2003). Fraser Hunter has drawn attention to wider north-eastern context of the practice of ritual hoarding, including the deposition of patera (e.g. Capheaton) (Hunter 1997; cf. Allason-Jones 2010). Assessment of historic finds now lost or in regional collections and material currently being found by metal-detectorists has shown significant assemblages north of the Wall at Great Whittington (Collins and Biggins2013), Chesters, Adderstone, Norham, and Newham Bog, although it is unlikely that all of these objects and hoards were deposited for ritual purposes. In addition there are various recorded examples at rural sites of animal burials (e.g. at Ingleby Barwick) indicating ritual practice. When the pre-Roman Iron Age site at Pegswood was closed down in the 2nd century, the event was marked by a conspicuous ritual placement of quernstones (Proctor 2009; see also Willis 1999 for the case of Burradon); there might now be enough data to begin to assess how long such obviously Iron Age practices continued across the region in the Roman period.
As with religious epigraphy and sculpture, the region is rich in the inscribed tombstones that are so relatively rare throughout most of non-military Roman Britain, recording the burial of both soldiers and civilians at the fort sites and the town of Corbridge. These were first extensively treated as evidence for a civilian as well as a military population in the study by Salway (1965).
Cemeteries existed at every fort/vicus and urban site, but these are notoriously under-researched, few examples having been accurately located and even fewer extensively excavated. North of the Wall a group of burials, mostly cremations, in a barrow cemetery outside the fort has been explored at Petty Knowes, High Rochester (Mitcheson and Charlton 1984). Predominantly 3rd century in date, the burials were poor in finds. A mixture of inhumations and cremations has been found at Lanchester, dating from the mid-2nd to the late-3rd century. The excavators interpreted these as civilian rather than military burials (Turner 1990). A similar mix of cremations and inhumations has also been excavated at South Shields (Snape 1994), where grave goods indicated one female grave. An early cremation cemetery north of Corbridge, yielding fine grave goods, was excavated in 1974 (Casey and Hoffmann 1995).
If extensive late-Roman inhumation cemeteries with human remains and grave goods to shed light on the origins, health and status of the population existed (as one might expect), or survive, at sites like Corbridge and other important late-Roman centres, these have yet to be discovered. The lack of Roman human bone assemblages is perhaps surprising considering the excellent cemeteries excavated elsewhere in the province, such as York. The only human remains from the cemeteries at South Shields, Petty Knowes, and Lanchester consist of cremated bone, despite the presence of inhumation burials, presumably due to acid subsoils.
One additional site from a military context that deserves to be mentioned is the massive mausoleum or tower tomb at Shorden Brae, Corbridge (Gillam and Daniels 1961). This monument was built at some point in the second quarter of the 2nd century AD. It clearly belonged to an unknown high ranking officer or official. The monument was still standing in the 7th century when dismantled to provide building materials for Wilfrid’s church and crypt at Hexham (Bidwell 2010).
The region has important examples of late-Roman or 5th-century burial. These include the discovery at South Shields of a number of 5th-century burials from a small cemetery outside the south-west gate of the fort, and from inside, two individuals who had been executed and subsequently buried in the ruins of the late-Roman commanding officer’s house (Bidwell and Speak 1994, 45-46, 143-144; Hodgson 1999, 82). Other cases of Roman burials re-using structures are known. Three skeletons were found placed in the apse of the Temple of Antenociticus when it was excavated in 1862 (Simpson and Richmond 1941, 38), and 33 human skeletons were found at the bathhouse at Chesters (Macdonald 1931).
New light on Roman cist burial has also been shed by a recent re-evaluation of a stone-lined grave from Sewingshields (Crow and Jackson 1997) and a number of other cist burials is known from the Wall, all probably of Roman, rather than early medieval date (e.g. Turret 39a at Peel Crag, and Milecastle 9 at Chapel House; Simpson 1976, 100, 102-103, fig 22; Birley 1930, 154, Pl 48). There is far less evidence for burial in rural contexts. Tait and Jobey also list nearly 20 cist burials from southern Scotland and Northern England of Iron Age or Roman date; while not enough examples have been found to indicate that this was ever a majority rite, it certainly indicates that this was part of the wider regional native burial tradition (Tait and Jobey 1971, 61, 66-69). In the Roman period, unlike the southern areas of England, there appears to have been no widespread practice of the deposition of unaccompanied or accompanied cremations except at military sites (Philpott 1991, 221). It is possible, however, that inhumations interspersed with settlement structures (as opposed to being segregated in a formal cemetery) will come to characterise the late-Roman rural settlements in the area south of the Wall. A pair of burials is known from Hartlepool (Daniels et al. 1987) and more graves are known at Newton Bewley (Robin Daniels pers comm), Faverdale (Proctor 2012a) and radiocarbon dated late-Roman examples are now published at Ingleby Barwick (Willis and Carne 2013). In the north of the region, there is an unusual group burial from Beadnell (Tait and Jobey 1971).
Again the Roman military presence in the region has led to the occurrence of a number of extremely rich finds assemblages of this period, but there has also been an increasing concern with categories of material culture produced in the dynamic conditions of interaction in the frontier region. For the later-Roman period the material culture of the frontier zone has been characterised in a series of landmark studies collected in the Finds from the Frontier volume (Collins and Allason-Jones 2010). The Portable Antiquities Scheme (PAS), only operating in the North East for three years from the first edition of the NERRF, has continued to record large quantities of Roman material in the region. Objects of certain and probable Roman production from the North East number 4,625 objects identified in 4,258 records (some records contain multiple objects, such as hoards). Of this total, 2,011 objects are individual coins and 92 hoards (coin and objects), with numerous other types of objects records, such as metal, glass, and ceramic vessels (no=257), brooches (no=225), mounts and studs (no=311), and even militaria such as sword remains and new sculpture (Collins 2013). Full details of these objects can be found on the PAS database, though some analysis has been provided (e.g. Collins 2009; 2014). The main conclusion to date is that Roman material of the 1st and 2nd centuries can be seen in a relatively even distribution across the region (regardless of Hadrian’s Wall), while Roman material of 3rd and 4th century date is far more clustered and found almost exclusively along and south of Hadrian’s Wall. The other conclusion is that material cannot always be attributed to a known or confirmed archaeological site, and that as a region the North East is quantitatively poorer and less diverse in terms of the different types of artefacts recorded than the southern and eastern regions of England. The regional PAS assemblage, however, is relatively richer in material culture associated with the Roman army.
The great number of military sites in the North-East has, unsurprisingly, produced large amounts of Roman pottery. The fundamental importance of the region in Roman pottery studies arises from the fact that samian ware and coarse wares that occur in sealed contexts on military sites have been closely dated according to the accepted chronological sequence of the northern frontier (for example, pottery found in contexts on Hadrian’s Wall which pre-date the move to the Antonine Wall c. 140 must have been current in the period 120-140). The dating of assemblages of pottery by reference to the chronological framework of ‘wall periods’ is reflected in Gillam’s highly influential type series first published in 1957 and still used (with caution) by pottery specialists. These dated groups are fundamental to national and international studies for the chronology of pottery types that they provide, that then assist the dating of other locations and phases across the province/s where they occur. In addition supply to the army from regional and extra-regional sources is a highly informative window upon distribution and trade, and potentially the supply arrangements and their administration. A major study of early military supply in the west of Britain revealed ad hoc arrangements being made ‘fort by fort’ in the first years following conquest and occupation (Darling 1978), similarly seen in early local arrangements in the north (cf. above). Questions of quarter mastering in the north have received only limited attention to date. Evans in a forthcoming paper identifies sub-regional military distributions indicating a large zonal basis to army supply (Evans forthcoming).
The discovery and analysis of several large assemblages of pottery from dated contexts since Gillam’s day means that much of his dating stands in need of revision, and it is now possible to recognise a much greater range of types and to refine their dating. In the north-east military assemblages have remained significant in this respect, particularly from the Tyneside forts, including South Shields, Wallsend and Newcastle (Bidwell and Speak 1994; Snape and Bidwell 2002; Rushworth and Croom 2016), where there has been extensive modern excavation (see above ‘The Roman frontier’). Other important stratified assemblages include those from Vindolanda (Bidwell 1985), Binchester (Ferris 2010) and Housesteads (Rushworth 2009) though there is still scope for further synthetic work on all these major groups of ceramics, with modern study moving beyond mere dating to consider questions of supply, functional analysis, and taphonomy. There are fewer published stratified assemblages from vici: see Casey and Hoffmann 1998 and Snape et al. 2010 for rare examples.
Indigenous sites in the Iron Age tradition north and south of the Wall have generally produced only tiny amounts of Roman pottery, but with the discovery of new Roman period rural site-types south of the Wall has come the publication of some important and well-dated assemblages from rural contexts: Faverdale is particularly important in this respect. Finds from Roman sites, including Faverdale, show that there was some continuity of native pottery production in this region, for example the so-called Local Traditional Ware at Newcastle, which has also been found at Corbridge, South Shields, and Wallsend (Bidwell and Croom 2002, 169-70). One fabric group (LTW Group 1) may have been produced on the Northumbrian coastal plain between the Aln and the Wear; the other (LTW Group 2) probably came from near South Shields or just to the north of the Tyne.
Pottery may also facilitate the recognition of possible external ethnic groups; it has been suggested that ‘Housesteads Ware’ may have been made by Frisian units stationed on the Wall (see now Evans in Wilmott 2009). The significance to be placed on locally made ceramics with North African affinities has been controversial (Swan 1992; 1999; cf. Bidwell and Croom 2016). A statement on the research priorities for Romano-British pottery studies in the north has been prepared by the Study Group for Roman Pottery, who also produced a research framework document for the national study of Roman pottery in Britain (Evans and Willis 1997; Willis 2004).
with Jenny Price
Glass vessels (generally tablewares and containers) and objects (generally bangles, beads and counters or gaming pieces) have come from virtually all Romano-British sites in the region, and very large groups were found in excavations at Binchester, Piercebridge and just over the border at Catterick, all now published. The majority of the vessels were produced in the north-west provinces or in Britain itself, but it is clear that more exotic pieces were also present in the region, such as the mould-blown cup with a Greek inscription from Binchester (RIB II2, no2419.38) and two polychrome mosaic plates from Piercebridge (Price 2002) and Quarry Farm, Ingleby Barwick (Willis and Carne 2013), both of which may have come from Egypt.
Glass bangles (already discussed under Industry) have been found on a wide number of Roman and native sites, and date primarily to the 1st and early 2nd century. They appear to be made from recycled glass. Originally thought to have been Scottish items traded south, work by Jennifer Price on examples found in East Yorkshire shows that their distribution is wider than previously thought, and their place of manufacture is not certain (Price 1988). This class of object is the subject of a major new research project by Tatiana Ivleva.
The Roman forts and vici have produced large quantities of small finds (Allason-Jones and Miket 1984; Allason-Jones 1988; Allason-Jones and Bishop 1988; Allason-Jones and Mackay 1985; Birley et al. 1993; van Driel-Murray et al. 1993; Croom and Snape 1996; Snape 1993; Snape and Bidwell 2002), and since the last resource assessment full publication of the finds from major excavations discussed above at Halton Chesters, Wallsend, Housesteads, Piercebridge and Binchester). The substantial nature of this resource makes it a significant research area, the level of cataloguing allowing more complex, synthetic work to be carried out (Allason-Jones 1995; 2001a; 2001b; 2002a; 2004; 2013; Birley 2002). It is increasingly possible to recognise distinct distributions of artefact types, particularly with respect to buildings within forts. Quantitative and qualitative analysis of some of the more prosaic items is also starting to produce information which is having an impact on finds study throughout the Empire (Allason-Jones 1985; Allason-Jones and Dungworth 1997).
Only small assemblages of objects have been recovered from rural sites, with a clear distinction between sites south of the Wall (such as Thorpe Thewles; Heslop 1987) and north of the Wall (Allason-Jones 1991; cf. Willis 1999). The southern sites seem to contain slightly larger assemblages, while the inhabitants of northern sites appear to have utilised little material culture, although this may be an artefact of more extensive pre-conquest contacts. The new class of rural sites south of the Wall (Faverdale, Ingleby Barwick etc.) have significant published finds reports, but these show a more selective and attenuated uptake of material culture than that of the Roman military and urban sites.
Groups of silver vessels have been found in the region, although not as many as in other regions of Roman Britain. They include a group found in Capheaton in the 18th century, probably dating to the 2nd century AD, the Backworth hoard including a silver patera (found early 19th century), and Corbridge (including the Corbridge lanx) (Craster 1909; Haverfield 1914a). There are a large number of bronze vessel hoards, with examples known from Bishop Middleham, Rookhope (Co. Durham), Ingoe, Whitfield Moor and Prestwick Carr (Northumberland) (Egglestone 1917; Hodgkin 1891; Wright 1969).
Jewellery formed an important element of the Backworth hoard, while the jewellery hoard from Great Chesters included the gilt bronze Aesica brooch considered to be a masterpiece of Celtic design, and several intaglios (Charlesworth 1973). Intaglios found in the region prior to 1978 have all been published by Martin Henig (1978). It is likely that many of these deposits were related to the wider practice of ritual deposition, found in North-East England and elsewhere (see above; Hunter 1997).
A high percentage of the small finds in the region is made from copper alloy, some imported, some manufactured locally. Iron objects are less prevalent, although the catalogue produced by Manning remains an important reference work (Manning 1976). Lead artefacts tend to consist of building and plumbing fragments, although the lead shrine found during excavations at Wallsend is a remarkable survival and shows a more decorative use of lead for religious purposes (Allason-Jones 1984).
Bone and antler artefacts are more commonly found on military sites on the Wall itself rather than on the native sites or the forts north and south of the Wall, although this may be a consequence of the acid soils in the area. Organic material, such as leather, is rare, though particularly fine assemblages are found at Vindolanda (van Driel-Murray et al. 1993).
A detailed resource assessment of coin research relating to Hadrian’s Wall and its environs, in fact applicable to much of the region, can be found in Brickstock 2009. The coin lists from the excavated Roman military sites and the single civil town of the North-east constitute a major numismatic resource – Corbridge for example has an assemblage of over 8,000 coins plus various hoards. These assemblages have benefited from intensive study in recent decades by John Casey and latterly Richard Brickstock, and many are the subject of reports published to modern standards; others remain unpublished or are in preparation. However, there is no modern general overview. The military sites are now supplemented as a source of numismatic data by the increasing numbers of rural sites producing Roman coins, though generally in low numbers (e.g. Faverdale, four coins) and from metal-detector finds, for which the data is readily available through the Portable Antiquities Scheme.
The long history of archaeological endeavour on Roman sites in the region, and the generally high level of material culture from military sites of this period mean that there are substantial museum collections. Significantly, most of the important material has remained in the region.
The two most important collections are those of the Society of Antiquaries of Newcastle upon Tyne, held at the Great North Museum: Hancock in Newcastle, and the English Heritage Hadrian’s Wall Collections. The Great North Museum holds finds from sites in Northumberland and Tyne and Wear (and some material from County Durham), with Roman objects comprising around 40% of the total museum collection. The epigraphic collection includes 200 altars, 65 tombstones, 200 other inscriptions, and 133 other sculptural items. Major items in the museum include the Aesica hoard of jewellery, the bear cameo from South Shields, Mithraic sculpture from Housesteads, Carrawburgh and Rudchester, and material from the Temple of Antenociticus at Benwell. It also holds domestic artefacts and one of the largest collections of Roman jewellery in the country. As well as objects the museum is also home to a major aerial photographic library and the Hadrian’s Wall Photographic Archive, in addition to other extensive archives and the combined Cowen Library of the University and the Library of the Society of Antiquaries of Newcastle upon Tyne.
The English Heritage Hadrian’s Wall Museum’s Collection comprises three separate collections: the Clayton Collection, the Housesteads Collection (managed jointly with the National Trust) and the collection from Corbridge. The Clayton collection at Chesters was formed by John Clayton between around 1840 and 1890 (McIntosh 2014). It incorporates a small amount of material acquired by his family at an earlier date, but also includes material acquired after John Clayton’s death, mainly as a result of F. G. Simpson’s excavations at Haltwhistle Burn and Housesteads (Wallis Budge 1903). This material almost all comes from the central sector of Hadrian’s Wall (Halton Chesters to Carvoran) and was acquired either through Clayton’s excavations, through his ownership of the land, inheritance or deliberate purchase. The greatest number of items is from Chesters fort itself, but the other main sites represented include Nether Denton, Carvoran, Great Chesters, Vindolanda, Housesteads and Carrawbrugh. The collection, entirely catalogued, includes c. 9,000 coins from Coventina’s Well, c. 2,000 coins from the Walbottle Hoard and a further c. 10,000 items, many of them small finds or pottery. There are few coarsewares, but there are larger quantities of samian and mortaria. 372 inscriptions, sculptures and stone objects are displayed alongside a further c. 800 objects. The wide range of material includes the Carvoran modius and the objects from Coventina’s Well, as well as iron tools and weapons from Chesters and fragments of painted wall-plaster from the commanding officer’s house there.
The Housesteads Collection includes objects from the excavations in the 1930s on the vicus, Dorothy Charlesworth’s material from the hospital, and the work by Charles Daniels in the 1970s and 1980s, as well as over 500 fragments of architectural stone.
The Corbridge Roman Site Museum holds material from the 1906 excavations onwards comprising work done 1906-14 and from 1933 to c. 1972 and again in 1980. This all comes from the Corbridge Roman site or its immediate environs, the latter including the Shorden Brae mausoleum, the supply base at Beaufront Red House and the A69 bypass excavations. The material, redisplayed in 2018, includes material from Red House and other important assemblages, such as the Corbridge Hoard. The exhibited material comprises just 5% of the total collections, the rest being stored on site except for the important collection of architectural fragments which is stored at Helmsley, North Yorks. About 50,000 items are catalogued, and cataloguing of the remainder is in progress. Most items or groups of items are inventoried and accessible.
Two major collections relating to North and South Tyneside are held by Tyne and Wear Archives & Museums. For South Tyneside the majority of the collection consists of material and archives from excavations at Arbeia Roman Fort and its surroundings. Much of this comes from Victorian excavations at the site from 1875 onwards, excavations in 1949-53, and from the excavations carried out by Tyne and Wear Museums since 1977. This last constitutes the largest collection of securely stratified material from any site in the Hadrian’s Wall World Heritage Site. The collection also includes the Bruce Library of antiquarian books, archival material and ephemera relating to Hadrian’s Wall, and watercolours by Ronald Embleton. Material from archaeological fieldwork in the district also forms part of the collection.
The bulk of the North Tyneside material is made up of material and archives from the excavations at Segedunum Roman Fort and its surroundings (Wallsend) from 1975-84 and 1988-2016. Material from archaeological fieldwork from the district is also kept at Segedunum.
The Vindolanda Trust owns a substantial and unique collection of site-specific artefacts excavated from Vindolanda over almost half a century. The collection increases annually as a result of the on-going excavation programmes. The following gives an indication of the extent of the Vindolanda Trust’s collection at the present time. It must, however, be noted that individual acquisition numbers, for leather, pottery and bone may refer to composite assemblages rather than a single object. Vindolanda’s collections of textiles, leather, and wooden objects represent the largest single site collection anywhere from the Roman world. Each of these collections contains individual pieces of rare and outstanding quality. As of 2006 there were over 9,800 ‘small finds’, 1,300 coins and 159 stone inscriptions and sculptural fragments, 5,870 leather fragments, 632 textile fragments, 1,580 wooden objects, two tons of bone and six tons of pottery.
Among the smaller museum collections in the region, Alnwick Castle Museum contains mainly objects from the Duke of Northumberland’s lands (Collingwood Bruce 1880). The Roman collection is relatively small, but includes some pottery, a collection of small finds from High Rochester and a number of miscellaneous items, including bronze vessels from Newham Bog and a fragment of a military standard from Halton Chesters.
In County Durham the Bowes Museum holds the finds from a number of excavations, including Binchester, Chester-le-Street sites, Ebchester, Greta Bridge and Piercebridge, as well as the Scargill shrine altars and the paterae from Bishop Middleham, but at the present time its collecting policy is under review and there may not always be practical access to archived material for study. The Durham University Museum of Archaeology, now mostly at the Palace Green Library, holds significant collections of Roman epigraphy, material relating to the work of Eric Birley when he was lecturer of archaeology in Durham, as well as the Oswald-Plique samian collection, which is of international importance.