The North-East

CHAPTER 7 THE NORTH-EAST By Martyn Allen The North-East region covers c. 14,282 km², stretching from Tyne and Wear in the north to Lincolnshire and Nottinghamshire in the south. In addition to these counties, the region includes all or parts of Derbyshire, East Riding, South Yorkshire, West Yorkshire, North Yorkshire, and County Durham (FIG. 7.1). THE NATURE OF THE LANDSCAPE The North-East region is made up of fourteen landscape zones based upon Natural England character areas (FIG. 7.2). The geography of the region is heavily influenced by the Humber estuary, which flows into the North Sea. Immediately to the north of the estuary lies Holderness, a coastal marshland that is mirrored FIG. to the south by the Lincolnshire Marshes. The estuary is fed by five major rivers – the Hull, Trent, Don, Aire, and the Ouse – which drain the Vales of York and Mowbray, and the Humberhead Levels. Upland regions include the North York Moors and the chalk hills of the Yorkshire Wolds, which are separated by the Vale of Pickering. The North-East region is divided from the Pennines by the Magnesian Limestone Belt, which is generally covered by fertile and free-draining calcareous soils, while a second limestone plateau is located at the northern point of the region and is separated from the North York Moors by the Tees Lowlands. To the west of the Magnesian Limestone, the Coal Measures of South Yorkshire lie in the Central West project region (see Ch. 8). Although outside the North-East, this landscape zone 7.1. The North-East region in relation to modern county boundaries 242 Romanch7.indd 242 06/09/2016 17:14:10 THE NORTH-EAST FIG. 7.2. Constituent landscape zones of the North-East region contains a distinctive settlement pattern of enclosed farmsteads and field systems which is similar to that found on the limestone, as well as the forts and vicus at Castleford, on the same road as Doncaster and Tadcaster (Bidwell and Hodgson 2009, 133–6). This point serves as a reminder that the boundaries of the project regions should not be seen as clear divisions, but simply as a means for organising the data. THE NORTH-EAST DATASET The dataset for the North-East contains 363 site excavation records, which account for 258 settlements as well as ‘isolated’ field systems, religious sites, burial sites and industrial sites. The site records are distributed unevenly across the region, with the greater concentrations tending to be located in the low-lying vales and on the Magnesian Limestone. The density of site records for each landscape zone is presented in TABLE 7.1, which confirms the high frequency of sites located on the Magnesian Limestone Belt, while other concentrations occur on Holderness, in the Vale of Pickering, and on the North Lincolnshire Coversands and Clays (cf. FIG. 7.3, top). Romanch7.indd 243 243 The high proportion of excavated sites identified on the Magnesian Limestone is mirrored by a high volume of excavations recorded in that area on the National Monuments Records (NMR) database; the effect of developer-funded excavations on the North-East region is readily apparent (FIG. 7.3b). Limestone quarrying on the Magnesian Belt and gravel quarrying on the Humberhead Levels has been responsible for the identification of a number of sites in the east and south of the region. Excavations resulting from pipeline projects have been even more productive, contributing 62 sites (17 per cent) to the regional dataset, many of which are located on the northern side of the Humber Estuary. The vast majority of sites recorded from Holderness, East Riding, for example, were identified during the excavation of the Easington to Ganstead natural gas pipeline (Flintoft and Glover 2009). In addition to excavated evidence, aerial photography and geophysical survey has been fundamental in revealing a range of settlements and field systems, such as the extensive, linear trackway and enclosure system in the Vale of Pickering, where small-scale excavations have demonstrated a late Iron Age/Romano-British 06/09/2016 17:14:18 244 THE RURAL SETTLEMENT OF ROMAN BRITAIN FIG. 7.3. Kernel density of North-East region records (n=363) and all excavation records (1910– 2010) from National Monument Records (NMR) Index (n=3434) (excluding data from York) Romanch7.indd 244 06/09/2016 17:14:31 THE NORTH-EAST TABLE 7.1: NUMBER OF SITES AND DENSITY (PER KM²) BY LANDSCAPE ZONE IN THE NORTH-EAST REGION Landscape zone Area (km²) No. of records Density of records per km² Durham Magnesian Limestone Plateau 447.00 7 Holderness 869.32 32 0.0368 Humber Estuary 285.68 9 0.0315 Humberhead Levels Lincolnshire Coast and Marshes Lincolnshire Wolds 0.0157 1723.53 48 0.0278 859.60 18 0.0209 844.86 17 0.0201 North Lincolnshire Coversands and Clay Vales 1318.61 40 0.0303 North York Moors and Hills 1864.34 18 0.0097 Sherwood 534.57 7 0.0131 Southern Magnesian Limestone 1367.62 59 0.0431 Tees Lowlands 1004.99 24 0.0239 425.50 13 0.0306 Vale of York and Mowbray 1627.16 29 0.0178 Yorkshire Wolds 1108.85 20 0.0180 Vale of Pickering date (Powlesland et al. 2006). A number of studies of the cropmark evidence from the region have also been undertaken, covering the Sandstone of South Yorkshire and North Nottinghamshire (Riley 1980), the Magnesian Limestone Belt (Roberts 2010) and the Yorkshire Wolds (Stoertz 1997). Each of these surveys has confirmed considerable evidence for late Iron Age and Roman-period enclosure, trackways, and field systems, much of which is not represented in the project database, but must be considered alongside it. The distribution of known cropmark sites of possible Roman date has also been mapped by Taylor, who highlighted additional concentrations of enclosed settlement evidence in the Vale of Pickering and on Holderness (Taylor 2007, 24–5, 44). In contrast to areas where evidence for Romano-British rural settlement is more abundant, the North York Moors have received comparatively little attention in terms of survey or excavation. Here, excavated sites are restricted to the periphery of this landscape and only become more numerous again on the Tees Lowland to the north. ROMAN RURAL SETTLEMENT PATTERNS The North-East region contains a number of major urban centres, including the civitas capital of Isurium Brigantum at Aldborough, the possible civitas capital of Petuaria at Brough (cf. Wacher 1969; 1995, 394–8; Hunter-Mann et al. 2000), and the coloniae of Eburacum at York and Lindum at Lincoln. The origins of Roman York and Lincoln lay with the establishment of the legionary Romanch7.indd 245 245 fortresses in the later first century A.D., and a number of other Roman forts and accompanying vici are also known at varying points along the road system, with notable examples located at Malton and Doncaster. These two sites may also be considered alongside the other small towns found in the region, though it is uncertain whether they ever became fully established as walled towns (Buckland and Magilton 1986; Wenham and Heywood 1997). Other walled ‘small towns’ are located at Horncastle, Caistor, Tadcaster (Calcaria), and Catterick (Cataractonium), and it is important to also consider the role of the military in the development of a number of these sites, including Catterick and, potentially, Horncastle (Bidwell and Hodgson 2009, 141–5, 162–4). Of the 258 settlements recorded for the NorthEast region, 213 are farmsteads. However, the percentage of farmsteads compared to other settlement types varies between different landscape zones (TABLE 7.2; FIG. 7.4). For example, in Holderness and on the Lincolnshire Coast and Marshes all the sites recorded are farmsteads, while in the Vale of York and Mowbray and on the Yorkshire Wolds, farmsteads represent only 64 per cent and 47 per cent of sites respectively. Settlements with villa architecture are rare, with only twenty sites recorded for the region. However, many more potential villa sites are recorded on the National Monuments Record (NMR), with a distinct concentration in the Lincolnshire Coversands and Clays between Lincoln and the Humber Estuary (see Villas below). Roadside settlements are represented by thirteen sites and military vici are represented by five sites. The latter only include sites with direct evidence for the 06/09/2016 17:14:40 246 THE RURAL SETTLEMENT OF ROMAN BRITAIN FIG. 7.4. Excavated late Iron Age/Roman rural settlement in the North-East region in relation to Roman roads and urban centres TABLE 7.2: NUMBER Landscape zone Durham Magnesian Limestone Plateau Holderness OF ALL SETTLEMENT TYPES BY LANDSCAPE ZONE IN THE NORTH-EAST REGION Farmstead Farmstead Farmstead Farmstead Villa Roadside Village Vicus Oppidum Total % (all) (complex) (enclosed) (open) settlement 6 30 1 5 2 Humberhead Levels 26 6 Lincolnshire Coast and Marshes 11 1 Lincolnshire Wolds 10 1 Humber Estuary North Lincolnshire 16 Coversands and Clay Vales North York Moors and Hills Sherwood 1 3 3 4 6 3 2 1 1 4 5 17 1 3 4 1 2 3 8 2 1 1 1 Vale of York and Mowbray 16 4 2 8 4 2 213 33 45 1 11 4.3 13 5.0 21 8.1 16 6.2 6 2.3 43 16.7 2 1 23 8.9 1 1 11 4.3 6 1 1 25 9.7 6 1 2 17 6.6 20 14 6 1 10 2.3 30 11.6 1 17 Total 2 3 Tees Lowlands Yorkshire Wolds 6 2 2 3 2.3 30 11.6 1 3 3 6 Magnesian Limestone Belt 39 Vale of Pickering Romanch7.indd 246 15 3 4 1 258 100.0 06/09/2016 17:14:48 THE NORTH-EAST no. of selements in use associated development of a civilian settlement alongside a fort, although this is not always straightforward to determine. For example, the role of the military in the early development of Piercebridge is poorly understood (Cool and Mason 2008). Here, a large, stone-built fort was certainly constructed by the middle of the third century A.D., but appears to have been off-set from the main road and civil settlement. Millett stated that the secondary position of the fort indicated that Piercebridge was a small town rather than a vicus, implying that the settlement developed independently (Millett 1990, 147). However, several authors have since argued, based upon comparisons with other sites in the north, that a first-century A.D. fort must have existed lying somewhere on or close to the River Tees (Bidwell and Hodgson 2009, 147; Ottaway 2013, 106). FIG. REGIONAL CHRONOLOGY Dating evidence was available from 252 sites, mostly through pottery and coinage, and an increasingly important corpus of radiocarbon dates (see Ch. 1). The chronological data show an increase in settlement numbers from the late Iron Age to a peak in the second half of the second century A.D. (FIG. 7.5). Thereafter, a decline can be observed through to the later fourth century A.D., when settlements numbered around onethird of that seen in the late Iron Age. However, there are clear intra-regional differences within this pattern (FIG. 7.6). The settlement chronology in Holderness, for example, shows a distinct decline in settlement numbers from a peak in the late Iron Age. In contrast, settlement numbers peak in the second century A.D. on the Southern Magnesian Limestone and on the North Lincolnshire Coversands and Clays, before 200 180 160 140 120 100 80 60 40 20 0 7.5. Number of settlements in use over time in the North-East region North Lincolnshire Coversands and Clays (n=20) 40 35 30 25 20 15 10 5 0 no. of sites no. of sites 30 25 20 15 10 5 0 no. of sites Holderness (n=30) 16 14 12 10 8 6 4 2 0 Southern Magnesian Limestone (n=42) FIG. Romanch7.indd 247 247 Vale of York and Mowbray (n=27) no. of sites 25 20 15 10 5 0 7.6. Variations in settlement chronology by selected landscape zone within the North-East region 06/09/2016 17:14:51 248 THE RURAL SETTLEMENT OF ROMAN BRITAIN 100 90 80 sites established % of sites 70 60 50 sites abandoned 40 30 20 10 0 Late Iron Age Late 1stC AD (n=142) (n=156) FIG. 2ndC AD (n=187) 3rdC AD (n=167) 4thC AD (n=127) 7.7. ‘New’ and ‘abandoned’ settlements over time in the North-East region b a Phase 4 Phase 3 Roman road N 0 200 m 1:5000 FIG. 7.8. Site plans of Roman Ridge, West Yorkshire (Roberts et al. 2001), showing (a) the late Iron Age/early first century field system, trackway and enclosures, and (b) the line of the Castleford to Tadcaster Roman road declining into the fourth century A.D., while in the Vale of York and Mowbray, settlement numbers also increase into the second century A.D., but show little evidence for decline thereafter, possibly because of the influence of major urban centres at York and Aldborough (see Ch. 12). The intra-regional differences in chronology may have been partly influenced by land in the south coming under Roman control in the later part of the first century A.D. With Brigantia continuing to hold political power, much of the northern part of the North-East can be equated as being culturally ‘Iron Age’ well into this period. Around one-third of the settlements occupied in the late Iron Age continued in use from the middle Iron Age, suggesting that this was a period of both continuity and expansion (FIG. 7.7). In the late first century A.D., only 18 per cent of the 156 sites in use were established ‘new’ at that time, considerably less than the 60% in the East region, for example (see Ch. 6), where there was much greater evidence for settlement expansion in this period. Nevertheless, the construction of new roads and military establishments, particularly along the Romanch7.indd 248 western and eastern fringes of the Humberhead Levels and the Vale of York and Mowbray, appear to have caused some disruption to settlement (Roberts 2010, 71–2). At Roman Ridge, West Yorkshire, in the first century A.D., a field-system with a trackway and two enclosures was transformed by the construction of a Roman road running north from Castleford to Tadcaster (Roberts et al. 2001; FIG. 7.8). The deposition of second to fourth century pottery at the site suggests that continued or renewed domestic activity occurred nearby (ibid.). At 10–14A Hall Gate, Doncaster, evidence for a wattle fence, a gully and a ditch, were later sealed by a road that was constructed around A.D. 70 when the Roman fort of Danum was built (Archaeological Services WYAS 2008). Similar evidence has also been identified at Glen Garth, Hayton, where the construction of a road at the end of the first century or the beginning of the second century led to the reorganisation of a small settlement, which became realigned towards the highway (Halkon et al. 2015). However, while the new road network affected some settlements, cropmark evidence suggests that its impact did 06/09/2016 17:14:53 THE NORTH-EAST 249 N 0 100 m 1:2000 FIG. 7.9. Plan of the twin enclosure complex at Allerton Park Quarry, North Yorkshire (Ross 2009), established in the mid-third century A.D. not result in any substantial reorientation of field systems (Roberts pers. comm.). Some of the changes that took place in the later first century A.D. may have been partly responsible for the increase in newly established settlements in the second century A.D., and it is possible that these changes brought about new opportunities for trade and exchange. At Gibraltar Farm, KingstonUpon-Hull, a settlement established on the northern bank of the Humber in the mid-second century A.D. soon began to receive continental goods, including samian and Moselkeramik, perhaps suggesting that the inhabitants were exploiting river-borne trade links (Tibbles and Steedman 1997; Van de Noort and Ellis 2000). Fewer new settlements originated after the second century A.D., though the establishment of enclosure complexes at Allerton Park Quarry (Ross 2009; FIG. 7.9) and Swaythorpe Farm, Kilham (Mackey 1998), appear to have been exceptions. In the third and fourth centuries an increasing number of settlements were being abandoned, while at Hensall Quarry (Weston 2013) and Crossgates, Seamer (MAP Archaeological Consultants 2001), both the settlements and their associated field systems went out of use, suggesting that wholesale changes in land use were occurring in some areas. It is possible that soil exhaustion was a factor in the decline of some settlements, with Hensall Quarry and Crossgates both being located on sandy, slightly acidic, soils, which may not have been suitable for the cultivation of some crops. In total, 106 sites were occupied in the second half of the fourth century A.D., though little can be said about settlement activity in the early fifth century A.D. Some excavators have suggested that Romanch7.indd 249 the occurrence of very late fourth-century pottery vessels, such as Crambeck wares and Huntclifftype jars, may indicate continuity into the fifth century. However, this is very difficult to prove owing to the kiln sites being abandoned before the end of the fourth century (Corder 1989a; 1989b). Ottaway points out that the Crambeck industry probably ceased once the army was no longer being paid, while the reappearance of regional hand-made wares at this time represents a return to a ceramic style not seen since the Iron Age (Ottaway 2013, 319–20). Unfortunately, these vessel types are difficult to date with any degree of accuracy. Late Roman coinage dating to the end of the fourth century A.D. (Reece period 21: A.D. 388–402) has been found on twelve sites, mostly villas and roadside settlements, with large numbers being recovered from Beadlam villa (Neal 1996), Langton villa (Corder and Kirk 1932), Shiptonthorpe (Millett 2006), and Hayton (Halkon et al. 2015) – the latter two benefitting from extensive metal-detecting surveys. Of course, the recovery of Reece period 21 coinage does not demonstrate evidence for fifth-century activity, and none of these sites otherwise provided evidence that they continued beyond the fourth century. Even at sites where clear evidence for fifthcentury activity has been demonstrated, there is uncertainty whether it represents continuous occupation from the late Roman period. At Malton, the fort was abandoned in the fourth century, though some of the masonry buildings to the south appear to have been re-used in the fifth and sixth centuries A.D. (Wenham and Heywood 1997). At Heslerton, the late Roman ladder settlement appears to have been abandoned prior to the establishment of an Anglian settlement and 06/09/2016 17:14:54 250 THE RURAL SETTLEMENT OF ROMAN BRITAIN cemetery (Powlesland et al. 2006). However, a nearby late Roman shrine continued to be used through to the seventh century, providing a focus for the early development of the Anglian settlement (Powlesland 1998). Some sites were clearly abandoned by the end of the fourth century, but show some evidence that activity recurred in the early medieval period, such as at Wattle Syke (Martin et al. 2013), Sewerby Cottage Farm, Bridlington (Fenton-Thomas 2009), and Melton A63 (Fenton-Thomas 2011). While conclusive evidence for domestic occupation on late Roman sites in the fifth century is generally lacking, lone burials dating to the early medieval period do appear to have been a feature at some sites, such as at Dalton Parlours (cf. I. Roberts 2013, 300). FARMSTEADS: MORPHOLOGY, CHRONOLOGY AND DISTRIBUTION The distribution of all 213 farmsteads in the North-East region is presented in FIG. 7.10, and their combined chronological data are shown in FIG. 7.11. There is very little difference in the total numbers of farmsteads in use between the late Iron Age and the second half of the second FIG. Romanch7.indd 250 century A.D., after which they decline through to the end of the fourth century. In total, 88 farmsteads have been classified in terms of their morphology. Enclosed farmsteads are most common, found relatively frequently on the Magnesian Limestone Belt compared to the low-lying Vale of York and Mowbray and the Humberhead Levels, where only a few excavated examples exist. Naburn in the Vale of York stood as one such rare example (Jones 1988; Roberts pers. comm.), but, more recently, the Asselby to Pannal pipeline has revealed evidence for settlement enclosure and field systems that bear some resemblance to those found on the Magnesian Limestone, the Coal Measures, and the Sherwood Sandstones to the south (Gregory et al. 2013). A number of sites excavated in these landscape zones can be identified as complex farmsteads, such as at Topham Farm, Sykehouse (Roberts 2003). Certainly, complex farmsteads are fairly widespread across the North-East in general, and it appears that the lack of enclosed and complex settlements in the Vale of York and Mowbray may have been due to poor visibility, an issue that is now being partly rectified through the identification 7.10. Distribution of all farmsteads in the North-East 06/09/2016 17:15:02 THE NORTH-EAST of cropmark sites (Ottaway 2013, 62–3). The few open farmsteads in the region are mostly located in the northernmost part, with other sites located on Holderness, in the Vale of Pickering, and on the Magnesian Limestone Belt. As with most other regions, the chronological data from classified farmsteads show an increase in the frequency of complex farmsteads over 251 time, and a reduction in enclosed and open farmsteads (FIG. 7.12). Open farmsteads are predominantly late Iron Age or early Roman in date, and very few are present in the second century A.D. or later. Although some appear to have been abandoned in the early Roman period, such as Thorpe Thewles in the Tees Lowlands (Heslop 1987), others developed into enclosed or no. of farmsteads 160 140 120 100 80 60 40 20 0 FIG. 7.11. Numbers of farmsteads in use in the North-East over time open farmstead (n=16) enclosed farmstead (n=50) complex farmstead (n=33) 80 % of classified farmsteads in use 70 60 50 40 30 20 10 0 FIG. 7.12. Relative frequency of farmstead types over time in the North-East First century A.D. Second century A.D. Third century A.D. (a) Third century A.D. (b) N 0 200 m 1:5000 FIG. Romanch7.indd 251 7.13. Phase plans of High Wold, Bempton Lane, Bridlington (Roberts 2009) 06/09/2016 17:15:04 252 THE RURAL SETTLEMENT OF ROMAN BRITAIN a b N 0 100 m 1:3000 FIG. 7.14. Plans of two types of complex farmstead in the North-East at (a) Faverdale, Darlington (Proctor 2012) and (b) Newbridge Quarry, Pickering (Richardson 2012) complex farmsteads, such as at High Wold, Bridlington (Roberts 2009; FIG. 7.13), and Faverdale, Darlington (Proctor 2012). At Faverdale, an extensive open settlement consisting of several roundhouses with hearths and stock enclosures, developed into a complex farmstead in the second century A.D. (FIG. 7.14a). Defined by a network of conjoined rectilinear enclosures and trackways, with a principal enclosure located on a high spur of land, this type of settlement is a form more typically found in the Central Belt during this period (see Ch. 5). In addition, the construction of a small, two-roomed stone building with a hypocaust and painted-plaster walls, probably a small bathhouse, also suggests some degree of affluence at the site. Romanch7.indd 252 N Trackway Roundhouse 0 100 m 1:3000 FIG. 7.15. Plan of Holmfield Interchange, Site Q, showing late Iron Age/early Roman D-shaped enclosure and the late Roman complex farmstead (Brown et al. 2007) 06/09/2016 17:15:05 THE NORTH-EAST a 253 Enclosed farmstead 0 100 m N 1:2000 b Enclosed farmstead FIG. 7.16. Plans of two late Iron Age/early Roman enclosed farmsteads integrated with a rectilinear field system at Heslington East, (a) Area A1 and (b) Area A2 (Antoni et al. 2009) Romanch7.indd 253 06/09/2016 17:15:06 254 THE RURAL SETTLEMENT OF ROMAN BRITAIN a b Potential earthworks Former ditch line 0 200 m 1:5000 FIG. 7.17. Site plans of Wattle Syke, West Yorkshire, showing (a) the late Iron Age–middle Roman enclosure complex (village), and (b) the late Roman (fourth century A.D.) open settlement (Martin et al. 2013) Second century A.D. Fourth century A.D. N Third century A.D. Fifth century A.D. 0 200 m 1:5000 FIG. 7.18. Phase plans of the settlement at Parlington Hollings, West Yorkshire (Roberts et al. 2001) Alongside those complex farmsteads consisting of a major, sub-divided enclosure, settlements that were characterised by systems of enclosures lining a trackway also became more common in the Roman period. These types of site are commonly referred to as ‘ladder settlements’ owing to their Romanch7.indd 254 linear morphology (e.g. Stoertz 1997). However, they varied considerably in size and in terms of the density of settlement features and material culture they included. The larger examples recorded on the database have been classified as villages, but smaller settlements are better understood as 06/09/2016 17:15:07 THE NORTH-EAST complex farmsteads, as at Newbridge Quarry, Pickering (Richardson 2012; FIG. 7.14b). The 45 enclosed farmsteads in the North-East vary considerably in size and shape, though the majority (56 per cent) are rectilinear in form. They ranged from small, single-ditched enclosures measuring around 0.1 ha, up to the doubleditched, 3.5 ha curvilinear enclosure at Tattershall Thorpe (Seager Smith 1998). While most enclosed farmsteads were fairly simple settlements, some are found to have developed over time. For example, the small D-shaped enclosure at Holmfield Interchange, Site Q, was incorporated into a larger complex farmstead in the late second century A.D. (Brown et al. 2007; FIG. 7.15). Cropmarks showed that the farmstead was associated with other settlements in the area, via an adjoining trackway and fields. The integration of enclosed farmsteads within field systems appears to have been relatively common, as at Heslington East where two small, enclosed farmsteads of late Iron Age/early Roman date were found to utilise field ditches as part of the settlement’s boundaries (Antoni et al. 2009; FIG. 7.16). Further cropmark evidence suggests that some of these systems extended over several FIG. Romanch7.indd 255 255 kilometres, for example at Tickhill and Went Hill, South Yorkshire (Roberts 2010, 32). In the late Roman period, some farmsteads show evidence for substantial reductions in size and complexity. At Wattle Syke, a modestly sized, open farmstead with cellared buildings was established in the late third century A.D. where a much larger, complex farmstead with multiple enclosures had been present in the later first and second centuries, possibly after a short hiatus in activity (Martin et al. 2013; FIG. 7.17). At the nearby site at Parlington Hollings, a complex farmstead and field system was reduced in size in the fourth century, with only a single enclosure and a boundary ditch being maintained (Roberts et al. 2001; FIG. 7.18). While many settlements show evidence for continued activity into the late Roman period, changes, such as those witnessed at Wattle Syke and Parlington Hollings show that this was not uniform. VILLAS IN THE NORTH-EAST Twenty villa sites are recorded in the North-East dataset and these are widely distributed across the region (FIG. 7.19). This pattern is somewhat 7.19. Distribution of excavated villas in the North-East compared to ‘villas’ recorded in the NMR index) 06/09/2016 17:15:15 256 THE RURAL SETTLEMENT OF ROMAN BRITAIN 500 year BC/AD 400 300 200 100 a 0 -100 b ological ed villas in the iest date of nd (B) number me no. of villas in use date of villa construc on 16 14 12 10 8 6 4 2 0 FIG. 7.20. Chronological patterns of excavated villas in the North-East: (a) earliest date of villa construction and (b) number of villas in use over time different from the distribution of villas recorded on the NMR, which shows a dense concentration of sites on the Lincolnshire Coversands and Clays, representing the northernmost tip of the so-called ‘villa zone’ of central England (Mattingly 2006, 480, fig. 17). The disparity between the two datasets highlights the lack of villas that have been comprehensively excavated and published. Of those that are recorded on the database, the quality of information available is highly varied. Some sites can only be regarded as ‘possible villas’, such as Wharram Le Street and Wharram Grange, which were primarily investigated using magnetometry and fieldwalking (Rahtz et al. 1986). Other sites have suffered from a protracted post-excavation process, such as Beadlam villa, where paper archives and material assemblages became widely dispersed, with some since being lost (Neal 1996). However, the situation is beginning to improve with more recent excavations, as with the publication of the villa at Ingleby Barwick (Willis and Carne 2013). Chronological data from villa sites shows that around half were constructed in the second century A.D., with the other half in the third century, and most developed from an earlier farmstead (FIG. 7.20). Although it is not always easy to demonstrate continuity of occupation, seven villas show evidence for pre-villa occupation stretching back to the late Iron Age. This chronological pattern is somewhat different from that of farmsteads, which reduce in number into the fourth century, in part Romanch7.indd 256 reflecting the development of some into villas. However, the small number of excavated villas means that their social and economic importance is difficult to assess. Villas in the North-East appear to have been quite varied in their size and complexity, ranging from the large courtyard building at Scampton, containing at least 40 rooms (Illingworth 1808), to the small, ‘cottage-style’ villa at Rudston (Stead 1980). Unfortunately, there is very limited evidence for the immediate context around villa buildings. An exception can be found at Welton villa, where excavations in advance of gravel quarrying revealed an extensive landscape of enclosures, trackways, aisled barns, corndryers, and a possible shrine/mausoleum (Mackey 1999; FIG. 7.21). Similar evidence for an extensive settlement complex has been revealed around the Roman villa at Ingleby Barwick, where a network of rectilinear enclosures and trackways was laid out on a sand and gravel terrace just south of the River Tees (Willis and Carne 2013). Despite these useful examples, most villas provide very little evidence for how the settlement functioned, which requires greater emphasis on the recovery of finds and environmental assemblages. NUCLEATED SETTLEMENTS Roadside settlements The thirteen roadside settlements recorded in the database are unevenly distributed along the major 06/09/2016 17:15:16 THE NORTH-EAST 257 a Corndryers Site of shrine/mausoleum? Corndryer Timber buildings Corndryer Corndryers Corridor house N Aisled barn Corndryer 0 200 m 1:5000 b N Winged-corridor villa FIG. 7.21. Plans of the villas at (a) Welton Wold (Mackey 1999) and (b) Ingleby Barwick (Willis and Carne 2013) road network, with a particular concentration found between Brough and York, though it is likely that other sites are yet to be identified (FIG. 7.22). Despite their comparative rarity in the NorthEast, roadside settlements probably performed a range of economic roles, though our understanding of their function is very limited. The majority are recognised by contiguous plots of land along the roadside, which are divided in places by trackways running off the main road (Halkon et al. 2015; FIG. 7.23). Most provide evidence for buildings, pits and wells, and some, such as Bainesse, North Yorkshire, have field systems backing onto ‘property plots’ (Wilson 2002). In most cases, roadside settlements in the North-East do not appear to have been enclosed, though it is rare for their full size to be determined, particularly without extensive geophysics or cropmark Romanch7.indd 257 evidence, such as at Shiptonthorpe where the settlement covered around 10 ha (Millett 2006). It is also uncertain whether these settlements had a central focus, which, presumably, would have been important if markets or local administration were a feature. Geophysical survey and small-scale excavation at East Park, Sedgefield, identified a single rectangular timber building sited in an otherwise open space in the centre of the settlement, which was suggested as having a ‘public’ function, perhaps as a shrine or a marketplace (Hale 2010). Chronological data show that some roadside settlements developed from late Iron Age foci (FIG. 7.24), such as the large, double-ditched enclosure identified at Hayton (Halkon et al. 2015), or the ritual complex at Nettleton and Rothwell (Willis 2013). At both these sites, the late Iron Age 06/09/2016 17:15:17 258 FIG. THE RURAL SETTLEMENT OF ROMAN BRITAIN 7.22. Distribution of nucleated settlement and military sites in the North-East N Ro m an ro ad Trackways Possible stone structures 0 200 m 1:5000 FIG. Romanch7.indd 258 7.23. Interpretative plan of the geophysical survey results at Hayton (Halkon et al. 2015) 06/09/2016 17:15:25 THE NORTH-EAST 259 25 no. of sites in use 20 15 10 5 0 FIG. 7.24 Chronological development of nucleated settlement in the North-East elements were incorporated into the Roman roadside settlements, demonstrating their continued significance. A possible late Iron Age shrine could have provided a ritual focus at Rudstone Dale, Newbald, where a timber structure was found to contain ten neonate burials and two animal burials, a calf and a lamb, placed around a central hearth (Wood 2011). Other sites appear to have been founded upon ‘virgin’ ground, as at Winteringham on the south side of the Humber, where early activity is suggested to have been stimulated by the military (Stead 1976). The position of forts nearby at Hayton and Kirmington may also have helped the civilian settlements to flourish in the early Roman period (Halkon et al. 2015; Jones and Whitwell 1991). Of course, sites such as these blur the boundary between roadside settlements and military vici, where civilian settlements are located very close to the fort itself (see below). The decline and abandonment of roadside settlements in the North-East is also poorly understood. Most continued to be occupied into the fourth century A.D., and the lack of fourth-century material at Dringhouses, near York (Ottaway 2011), East Park, Sedgefield (Hale 2010) and Bishop Grosseteste College, Newport (Wragg 1995), is probably due more to the restricted scale of excavations than a genuine lack of activity. At Shiptonthorpe, gravel spreads dating to the late fourth century possibly represent the latest buildings in use at the settlement. Here, Millett suggests that a reordering of space along the roadside points to a reduction in the local population, while the road may have continued to be used into the early Saxon period, influencing the development of the early medieval village at Shiptonthorpe (Millett 2006, 307–8). Vicus settlements As with roadside settlements, excavated evidence from military vici is equally patchy. Only five are included in the North-East dataset, although other known examples such as Castleford lie just Romanch7.indd 259 outside the region; the location of this site on the road between Doncaster and Newton Kyme shows that it would have been an important settlement in the south-west of the region. Excavations at Welbeck Street have revealed part of the vicus at Castleford, including some good examples of early Roman timber strip buildings that front the road leading south from the fort (Abramson et al. 1999). The vicus settlement at Newton Kyme is not included in the database because the very limited excavations carried out at the site have been restricted to the fort, and are unpublished (Bidwell and Hodgson 2009, 136–8). However, aerial photography has shown an extensive roadside development of enclosures, buildings and trackways, running for 600 m south of the fort (Boutwood 1996; FIG. 7.25). Unfortunately, the lack of excavation means that the relationship between the fort and the civil settlement is not well understood. The plan of the settlement at Newton Kyme represents one of the few more extensive overviews of the form and size of a vicus in the North-East. Catterick (Cataractonium) is a comparatively well-understood site owing to extensive excavations over an area where there has been very little modern development (Wilson 2002). Catterick itself is not included in the database because it had clearly developed into a walled town at least by the fourth century A.D., the defences of which were integrated with those of the fort that sat adjacent to it (Bidwell and Hodgson 2009, 144, fig. 59). Two sites that are included in the database as military vici are Doncaster (Danum) and Malton (Derventio). These two settlements are far less well understood compared to Catterick, but each has produced indications that they too developed into ‘military towns’ in the later Roman period. Boundary ditches excavated at Doncaster were suggested to have marked the western extent of the settlement, and it was argued that the recovery of limestone and finely tooled sandstone from the base of two 06/09/2016 17:15:26 260 THE RURAL SETTLEMENT OF ROMAN BRITAIN N Roman fort Possible Neolithic henge monument and/or Roman amphitheatre Vicus settlement Road 0 200 m 1:5000 FIG. 7.25. Plan of cropmarks showing the fort and vicus at Newton Kyme, North Yorkshire (Boutwood 1996) of the ditches could have derived from a ‘defensive’ wall (Buckland and Magilton 1986, 42, 210–13; Bidwell and Hodgson 2009, 131–2, fig. 53). Further, slim, evidence for vicus enclosure has been noted at Malton, between the fort and the River Derwent. Here, several phases of building construction took place during the second and fourth centuries, which appear to coincide with renovations of the fort, signifying substantial investment in the settlement (Mitchelson 1964; Wenham and Heywood 1997). Romanch7.indd 260 In contrast to most military vici in the North region, which were largely abandoned by the end of the third century A.D., particularly along Hadrian’s Wall, the vici in the North-East region mostly continued into the fourth century (Bidwell and Hodgson 2009, 33–4). This seems to reflect the transition of some into regional market centres, which certainly occurred at Catterick and possibly at Doncaster and Malton as well. Such continuity into the fourth century is consistent with the dating of the roadside settlements in this region. 06/09/2016 17:15:27 THE NORTH-EAST The end of the fourth century sees the abandonment of most of the North-East military vici, though there is some evidence for continuity, or re-emergence of activity into the fifth century at Piercebridge (in the far north of the North-East region), where activity contracts to areas within and immediately around the fort (Cool and Mason 2008), and at Malton, where Roman buildings appear to have been re-used in the fifth and sixth centuries (Wenham and Heywood 1997). Other nucleated settlements A number of nucleated settlements that are not associated with the main Roman road system are also present in the North-East. The region contains an important late Iron Age oppidum at Stanwick, North Yorkshire, consisting of an earthwork enclosure covering nearly 350 ha. Dating the site to the middle of the first century A.D., Mortimer 261 Wheeler argued that it was constructed by the Brigantes tribe as a defensive response to the Roman Conquest (Wheeler 1954). However, excavations in the 1980s led to the suggestion that the elaborate entranceways and walled ramparts were more consistent with displays of status and prestige by a pro-Roman client kingdom (Haselgrove et al. 1990). Stanwick was abandoned soon after the conquest, as was the extensive late Iron Age settlement at Dragonby, Lincoln, where the recovery of Roman ballista bolts and spearheads were interpreted as evidence for an attack (May 1996). However, while Stanwick was never reoccupied, activity at Dragonby had been re-established by the end of the first century A.D. Excavations there revealed a flourishing settlement that continued into the fourth century, with evidence for stone-footed, aisled buildings, stockkeeping, craftworking and trade. 45 N 40 Area of excavation 40 45 50 55 40 0 500 m 1:10000 FIG.7.26 Romanch7.indd 261 Plan of cropmarks showing the late Iron Age/Roman ladder complex at Burton Fleming (Tabor 2009) 06/09/2016 17:15:28 262 THE RURAL SETTLEMENT OF ROMAN BRITAIN Ladder settlements have been discussed previously in relation to complex farmsteads, though some expanded to a size where they are better understood as villages. They have been widely recognised through aerial photography of cropmarks in the North-East, particularly on the Yorkshire Wolds (Stoertz 1997). Several examples can be seen to follow the contours of valley slopes, such as at Burton Fleming (Tabor 2009; FIG. 7.26) and Heslerton (Powlesland et al. 2006), both of which originated in the later Iron Age, implying that the local topography was fundamental to their development. Both these sites consist of a large number of contiguous enclosures running along both sides of a trackway, most probably linked to the corralling and long-distance movement of domestic livestock (cf. Giles 2007). It is important to note here that the long trackway identified by Powlesland at Heslerton, which was in use during the late Iron Age, appears to be the same as the hypothesised Roman road (Margary route 816) running between the fort and vicus at Malton/ Norton, along the northern foothills of the Yorkshire Wolds, to the signal station on the coast at Filey. Other than the discovery of small sections FIG. Romanch7.indd 262 7.27 of metalled road surfaces at Malton and Filey, no traces of it have otherwise been discovered in between (Margary 1955, 424–5). It is very likely that the route between the two Roman settlements was the late Iron Age trackway, which continued to be used throughout the Roman period. BUILDINGS The North-East dataset includes records for 644 buildings from 156 settlements, an average of just over four buildings per settlement, close to twice the mean number recorded in the South and East regions. Excavated buildings are widely distributed across the region, being well represented in the Tees Lowlands, but are less concentrated to the south of the Humber and on the Magnesian Limestone (FIG. 7.27). Circular buildings are nearly twice as numerous as rectilinear buildings, though the distribution of the two forms is generally similar (FIG. 7.28). Holderness and the North York Moors lack rectilinear buildings, which may be due to a relative lack of post-conquest settlement in these areas. Sites with circular buildings were more Distribution of all buildings in the North-East 06/09/2016 17:15:36 THE NORTH-EAST 263 FIG. 7.28 Distribution of circular and rectangular buildings in the North-East Romanch7.indd 263 06/09/2016 17:15:51 264 THE RURAL SETTLEMENT OF ROMAN BRITAIN circular rectangular 100 90 80 no. of sites 70 60 50 40 30 20 10 0 Late Iron Age 3rdC AD 4thC AD 7.29. Use of circular and rectangular buildings in the North-East over time common than those with rectangular buildings until the third century A.D., and they continued to be constructed into the fourth century, perhaps reflecting elements of conservatism (FIG. 7.29). At The Bridles, Barnetby le Wold, the construction of a timber roundhouse within an enclosure in the fourth century A.D. appears to have followed a long tradition of circular building use that stretched back into the late Iron Age (Allen and Rylatt 2002). Only two sites with evidence for rectangular buildings date to the late Iron Age, at Faverdale, Darlington (Proctor 2012), and Pig Hill (Northern Archaeological Associates 2004), both in County Durham. The function of these structures is uncertain; the example at Faverdale was located to the south of the main domestic area and it may have been an ancillary building. The construction of rectilinear buildings only becomes more common from the late first century A.D., though the earliest examples that appear to have been used for domestic habitation seem to occur at roadside settlements and military vici. FARMSTEADS Most buildings in the North-East dataset are found at farmsteads, where circular, timber/masswalled buildings were particularly dominant (TABLE 7.3). Although the chronological pattern of building use on farmsteads is similar to that for all settlements, the data show that the uptake of rectangular building was far less pronounced, though circular and rectilinear buildings are recorded in roughly equal numbers of sites in the third and fourth centuries. Rectilinear buildings are better represented on complex farmsteads than enclosed farmsteads, and are almost entirely absent from open farmsteads (FIG. 7.30). This pattern is partly chronological, owing to the increasing number of complex farmsteads present in the middle and late Roman period when rectilinear buildings were also more common. Of the eight complex farmsteads with buildings Romanch7.indd 264 2ndC AD circular 25 rectangular 20 no. of sites FIG. Late 1stC AD 15 10 5 0 complex farmstead enclosed farmstead open farmstead FIG. 7.30. Use of circular and rectilinear buildings on farmsteads of different types dating to the fourth century, none are circular in plan. The roundhouse excavated within the late Roman complex farmstead at Holmfield Interchange, Site Q, is dated by a few sherds of calcite-tempered pottery and it was very uncertain whether this related to the occupation of the house (Brown et al. 2007, 69). The contemporary use of circular and rectilinear buildings on some farmsteads may suggest differences in function and possibly prestige afforded to different areas of the settlement – for example, at Crossgates, Seamer, four late Iron Age, timber roundhouses lay within a sub-divided enclosure and were added to in the first century A.D. by a second enclosure that contained a rectangular, limestone building. Timber buildings vastly outnumber masonry structures on farmsteads, with many circular buildings defined by drainage gullies, and some providing evidence of postholes for internal support. On many sites, these gullies can be seen to cut one another, indicating the gradual replacement of these buildings over time. Postholes that mark out the external walls of roundhouses tend to be short-lived, though a good, late Iron Age example can be seen at Welton Wold (Mackey 1999). The majority of 06/09/2016 17:15:54 THE NORTH-EAST TABLE Site type Circular buildings (no. sites) 7.3: BUILDINGS IN THE NORTH-EAST REGION Rectangular buildings (no. sites) No. circular buildings 105 36 10 10 Villa 1 Roadside settlement 3 Farmstead Farmstead/villa 265 No. rectangular buildings No. timber buildings No. masonry buildings 350 62 384 26 34 54 33 55 8 1 25 0 26 9 10 44 36 17 Vicus 0 3 0 26 4 22 Village 3 1 21 10 23 8 Shrine 0 1 0 2 0 2 Industrial 0 2 0 5 3 2 122 70 416 228 483 158 Total a b N N Wall footings on residual bank 0 20 m 1:500 FIG. 7.31. Plans of drystone masonry structures on the North York Moors at (a) Percy Rigg, Kildale (Close 1972), and (b) Great Ayton Moor (Tinkler and Spratt 1978) rectilinear timber buildings are of posthole construction, though a few have been identified from post-in-ditch foundations, such as those recorded at Thorpe Hall, Eastrington (Wood 2011). Many of these buildings were small and fairly simple structures, and the identification of their remains is likely to be adversely affected by truncation. Stone buildings are very poorly represented on farmsteads, found on only 16 per cent of sites with evidence for structures. In the North York Moors, localised examples of drystone, circular buildings have been identified at Percy Rigg, Kildale (Close 1972), Crag Bank, Kildale (Close et al. 1975), Great Ayton Moor (Tinkler and Spratt 1978), and Roxby site 2 (Inman et al. 1985), though subtle differences in construction technique occur between sites, such as the use of internal paving (FIG. 7.31). These structures all date to the late Iron Age, with those at Crag Bank and Roxby site 2 continuing to be used in the Roman period. Romanch7.indd 265 Most masonry structures on farmsteads are rectilinear in form, and are commonly unmortared, single-room structures, such as those identified in Edlington Wood, South Yorkshire (Corder 1951). Signs of significant investment or elaboration in masonry are very rare. Only two buildings with hypocausts are known from farmsteads in the North-East, one at Heslington East (Roskams and Neal 2012), and another at Faverdale, Darlington (Proctor 2012). The lack of a main villa house at these two sites means that they are classified here as ‘farmsteads’, but they clearly display indications of status and perhaps should not be considered much differently from modest villas elsewhere in social and economic terms. The building at Heslington East was located near a masonry ‘tower’, which was interpreted as a monumentalisation of the western entrance into the settlement. A stone-walled octagonal building at land off Horkstow Road, South Ferriby, may 06/09/2016 17:15:55 266 THE RURAL SETTLEMENT OF ROMAN BRITAIN a b c Possible corndryer N 0 20 m 1:500 FIG. 7.32. Post-built ‘structures at (a) Melton A63 (Fenton-Thomas 2011), (b) Cedar Ridge, Garforth (Owen 1998), and (c) Stile Hill, Colton (Archaeological Services WYAS 1995) Romanch7.indd 266 8 7 6 no. of sites have been a shrine or a memorial (Clay 2006). Such investments in masonry building were exceptional on farmsteads in the North-East. The specific function of most buildings is usually indeterminable without further evidence, though many are likely to have been primarily for domestic use. Possible agricultural buildings have been excavated at Melton A63 (Fenton-Thomas 2011), Cedar Ridge, Garforth (Owen 1998), and Stile Hill, Colton (Archaeological Services WYAS 1995). Dating to the second and third centuries A.D. these buildings were timber, post-built constructions, with the example at Melton A63 formed of closely aligned postholes, indicating that it held a raised floor (FIG. 7.32). All these buildings were located close to corndryers, which perhaps suggests their use as granaries. Other buildings may have been ancillary workshops, such as the late Roman post-built structure at Thorpe Hall (Wood 2011), or the late Iron Age beam-slot structure previously mentioned at Pig Hill (Northern Archaeological Associates 2004), both of which produced considerable evidence for ironworking. Cellared buildings have been identified at six sites, and those at Wattle Syke (Martin et al. 2013) and Dalton Parlours (Wrathmell and Nicholson 1990) have produced evidence that they were used for a range of functions, including crop processing, grain drying, animal skinning, antlerworking, cooking, and smithing. However, at Welton Wold (Mackey 1999) and Sewerby Cottage Farm, Bridlington (Fenton-Thomas 2009), the insertion of corndryers into cellared buildings indicates that some were used more specifically for agricultural purposes. 5 4 3 2 1 0 coage/hall FIG. corridor winged corridor courtyard uncertain 7.33. Villa building forms in the North-East VILLA COMPLEXES The few villas excavated in the North-East contain a total of 81 buildings, the majority of these being ancillary to the main domestic residence. As with other areas of the country, villa houses ranged considerably in size and complexity (FIG. 7.33). Small corridor buildings with minimal evidence for interior refinement have been excavated at Scruff Hall Farm, Drax (Wilson 1965), and Welton Wold (Mackey 1999), while similarly modest striphouses, such as at Rudston (Stead 1980) and Holme House (Harding 2008), incorporated bathhouses and tessellated flooring (FIG. 7.34). The villa at Scampton is a somewhat grander and more complex affair, though, unfortunately, it was largely excavated in the late eighteenth century, and the dating evidence is poor (Illingworth 1808). Aerial photography of the site suggests that the settlement complex comprised an array of 06/09/2016 17:15:57 THE NORTH-EAST a 267 b N c d 0 20 m 1:500 e f 0 100 m 1:2000 FIG. 7.34. Comparative plans of six villa houses: (a) Scruff Hall Farm, Drax (corridor) (Wilson 1965), (b) Welton Wold (corridor) (Mackey 1999), (c) Rudston (strip-house) (Stead 1980), (d) Holme House (strip-house) (Harding 2008), (e) Winterton (courtyard) (Stead 1976), and (f) Beadlam (courtyard) (Neal 1996) Romanch7.indd 267 06/09/2016 17:15:57 268 THE RURAL SETTLEMENT OF ROMAN BRITAIN buildings, including a bathhouse, arranged round two courtyards. The substantial courtyard villa at Winterton, North Lincolnshire, was also subject to antiquarian excavations, though more recent investigations in the 1950s showed that the settlement developed gradually from the late firstcentury A.D. farmstead into a far more sophisticated complex (Stead 1976). By the end of the second century A.D., the settlement boasted several buildings with hypocausts, including a bathhouse, two aisled buildings, a workshop and other structures, together arranged around a central courtyard. A similar complex arrangement of buildings is also found at Beadlam villa, where two winged-corridor houses were set side on, once more with other structures to form a central courtyard (Neal 1996). Aisled buildings have been identified on six villas in the North-East. Most of these appear to have been ancillary to the main villa house, apart from at Chapel House Farm, Dalton-on-Tees, where a winged-corridor villa developed from an earlier aisled structure (Brown 1999). Otherwise, these buildings range from the comparatively large, masonry-footed structures at Dalton Parlours and Ingleby Barwick, to the more modest timber buildings at Welton Wold. Circular or oval masonry buildings have been found at five villas. The example at Langton was tentatively interpreted as a mill, and was associated with the insertion of an apparent threshing-floor nearby (Corder and Kirk 1932). An agricultural or an industrial function may be supported by evidence from Ingleby Barwick, where a large timber, circular structure (c. 8 m diam.), rebuilt in stone during the fourth century A.D., included a compacted earthen floor built over an oven (Willis and Carne 2013). Evidence for iron-smithing in the circular structures at Winterton and Beadlam suggest that some may have been workshops, and were potentially important elements in the economy of these villas (Stead 1976; Neal 1996). NUCLEATED SETTLEMENTS Evidence for buildings at nucleated settlements have been recorded from roadside settlements, military vici and villages, and include a range of different structural forms (TABLE 7.3). Rectangular buildings are far more common than circular buildings at roadside settlements and vici, both in terms of the number of sites and the overall number of buildings, though the use of timber and masonry construction is more equally represented. In contrast, circular buildings are more common at villages, though this only relates to three sites, with Dragonby being the only one with rectilinear structures standing alongside circular ones. Eight of the Dragonby examples were masonry Romanch7.indd 268 constructed, with some showing evidence of central postholes (May 1996). In contrast, late Iron Age and early Roman buildings at Wattle Syke (Martin et al. 2013) and Low Caythorpe (Fraser and George 2013) were exclusively roundhouses. The dominance of rectilinear buildings at roadside settlements signifies the different architectural traditions of these settlements compared to villages. Four of the post-built roundhouses at Rudstone Dale, Newbald, dated to the late Iron Age, prior to the site’s development as a roadside settlement. It is perhaps significant that the only circular structure found during the Roman phase contained the burial of a sheep or goat and a human cremation within its central area, and a neonate within the north-west arc of its putative wall, suggesting that it may have been a shrine (Wood 2011). Architecture at Bainesse (Site 46) consisted of mostly rectilinear, timber buildings, located alongside a smaller number of masonry structures (FIG. 7.35). Their alignment along the road, along with the recovery of ironand boneworking waste, and a wide array of artefacts, suggests that they may have performed a mixture of domestic, commercial and industrial roles (Wilson 2002). The timber buildings found here used a combination of beam-slot and posthole construction, a technique that also appears to have been typical in other roadside settlements, such as at Shiptonthorpe (Millett 2006). Masonry buildings tend to be more common than timber structures at military vici, though this partly reflects an apparent desire for stone architecture in the immediate vicinity of the forts. Outside the third-century A.D. fort at Piercebridge, a number of large masonry buildings were constructed, including a temple dedicated to Jupiter Dolichenus (Cool and Mason 2008). At Malton, masonry buildings were built just south of the fort, enclosed in an annexe between the defences and the River Derwent. In the fourth century A.D., a major rebuilding programme in the area saw the addition of eight new structures, including a town house with a fine mosaic floor and painted plaster, a bathhouse, and a kiln building. Finds recovered from the site included a large number of high-status items, such as hair pins, finger rings, bracelets, toilet equipment, and a considerable quantity of vessel glass (Mitchelson 1964; Wenham and Heywood 1997). These ‘annexes’ are largely restricted to the military vici, having more in common with larger towns than rural nucleated settlement. This finding seems to be consistent with Bidwell and Hodgson’s suggestion that some vici developed into regional market centres, perhaps existing as ‘military towns’ in the late third and fourth centuries A.D., such as at Catterick (Bidwell and Hodgson 2009, 33–4). 06/09/2016 17:15:57 THE NORTH-EAST a 269 N Corndryer Dere Street Projected timber buildings Gravelled surfaces 0 20 m Masonry building 1:500 N b Aisled building Pits Enclosure ditch FIG. 7.35. Plans of buildings at roadside settlements: (a) timber and masonry structures at Bainesse (Site 46) (Wilson 2002) and (b) a beam-slot and posthole structure at Shiptonthorpe (Millett 2006) Romanch7.indd 269 06/09/2016 17:15:59 270 THE RURAL SETTLEMENT OF ROMAN BRITAIN enclosed farmsteads height above ordnance datum (m) height above ordnance datum (m) complex farmsteads 100+ 90-99 80-89 70-79 60-69 50-59 40-49 30-39 20-29 10-19 0-9 0 1 2 3 4 5 6 no. of sites 7 100+ 90-99 80-89 70-79 60-69 50-59 40-49 30-39 20-29 10-19 0-9 8 0 0 FIG. 1 2 3 4 5 6 no. of sites 7 8 3 4 5 6 no. of sites 7 8 100+ 90-99 80-89 70-79 60-69 50-59 40-49 30-39 20-29 10-19 0-9 0 1 2 3 4 5 6 no. of sites 7 8 7.36. Spot height analysis for major settlement types in the North-East LANDSCAPE CONTEXT AND INFRASTRUCTURE TOPOGRAPHY As outlined at the beginning of this chapter, the North-East region consists of a mixture of lowlying Vales, the upland landscapes of the North York Moors and the chalk Wolds of Yorkshire and Lincolnshire. Much of the region, other than the upland areas, is covered with alluvial deposits of clay and till. As has been shown previously, rural settlement tends to favour the fertile vales and the Magnesian Limestone. Excavated sites tend to be more visible on the chalk Wolds compared to other upland areas, such as the peatland of the North York Moors, a pattern that is supported by cropmark evidence (e.g. Stoertz 1997). The settlement pattern is, of course, heavily biased by modern factors (see Chs 1 and 12) as well as cultural influences, such as the development of towns, military sites, and the major roads. However, there does appear to have been a strong relationship between rural settlements and the underlying geology of the North-East region. The preference of farmsteads in the region for low-lying ground is demonstrated by spot height data, which shows that 74 per cent were sited below 60 m OD, and 40 per cent located on land below 20 m OD, though variations emerge for particular types of site (FIG. 7.36). Complex farmsteads are frequently encountered on lowland areas, particularly on superficial till and clay geologies. In comparison, enclosed farmsteads Romanch7.indd 270 2 roadside settlements/vici 100+ 90-99 80-89 70-79 60-69 50-59 40-49 30-39 20-29 10-19 0-9 height above ordnance datum (m) height above ordnance datum (m) villas 1 were more commonly located on higher ground, with a number of sites found on land over 100 m OD. The highest are farmsteads on the North York Moors, such as Percy Rigg, Great Ayton Moor, Crag Bank, the Levisham Moor enclosures, and Roxby sites 1 and 2, sited between 190–280 m OD. These sites tend to include evidence for stock pens and were probably engaged in upland pastoral farming. As seen in other regions, roadside settlements and military vici are consistently located at low levels, especially below 40 m OD, and generally close to rivers, where their forts were strategically sited. Villas, on the other hand, were more evenly distributed at different heights. Beadlam villa, for example, was located on slightly raised ground next the River Riccal, which may have been an important influence on the siting of the villa. In contrast, Langton, Wharram Grange and Wharram Street all occupied higher ground at the edge of the Yorkshire Wolds. TRANSPORTATION: RIVERS, ROADS AND TRACKWAYS The Humber estuary formed an important conduit for the movement of goods and people into and out of the North-East, with the town of Petuaria at Brough and the roadside settlement at Winteringham on the opposite bank long thought to have been of naval importance (Wacher 1969; 1995; Stead 1976). Certainly, epigraphic evidence suggests that the Ouse was navigable and that York was a port of trade (Hawkes et al. 1946, 67; RIB 06/09/2016 17:16:00 THE NORTH-EAST 271 a Trackway Enclosed farmstead Trackway Complex farmstead Trackway 0 400m N 1:8000 b Trackway Trackway FIG. 7.37. Plans of long-distance trackways with associated field systems and settlements at (a) Holmfield Interchange (Brown et al. 2007), and (b) Swillington Common, West Yorkshire (Roberts et al. 2001) Romanch7.indd 271 06/09/2016 17:16:01 272 THE RURAL SETTLEMENT OF ROMAN BRITAIN trackway have recently been identified near Bramham (Weston and Roberts 2015) and near Collingham (Gregory et al. 2013, 138–41). FIELD SYSTEMS Aerial photography has produced considerable evidence for late Iron Age and Romano-British field systems in many areas of the North-East region. The visibility of cropmarks is greatest on the Sherwood Sandstones, the Yorkshire Wolds, and the Magnesian Limestone Belt, where previous studies have demonstrated a variety of field forms (Riley 1980; Stoertz 1997; Roberts 2010). The evidence is far sparser in the Vale of York and on the Humberhead Levels, where the alluvium and glacial clay deposits have created a masking effect, reducing visibility (Taylor 2007, 44), as well as a warping effect that alters the appearance of some field systems (Roberts pers. comm.). A total of 120 excavated sites in the North-East include evidence for field systems. These are widely distributed across the region, with a notable concentration occurring on the Magnesian Limestone Belt. Just over half of all these sites (63) are associated with farmsteads, and two with villas. Often field systems have been identified only from fairly insubstantial traces of ditches, though, where cropmark evidence survives, some settlements can be observed as parts of a much wider agricultural landscape. The complex of ‘brickwork’ fields at Dunstan’s Clump on the Sherwood Sandstones, now seen as a classic example of this type of land management, included numerous settlements interspersed among the extensive pattern of parallel strip-fields (FIG. 7.38). N I.653; RIB III.3195), though archaeological evidence for the putative Roman harbour is limited (Ottaway 1993, 69, 85). In general, however, the extent to which rivers were used as transport routes in the North-East is difficult to assess, owing to a lack of evidence for wharfs/ jetties. At Perrins Cottages, Fiskerton, a possible docking area was tentatively suggested from the identification of a limestone, rubble surface, mixed with quantities of late Roman material, which could have functioned as an access point to the River Witham (Palmer-Brown 1994). Beyond the major rivers, it appears that riverine transport was not important, particularly for the rural population, who were perhaps more dependent on the road system and local trackways. A relatively high proportion of villas were located within 2 km of a major road (60 per cent), implying that easy access to these highways was desired. In contrast, only 22 per cent of complex farmsteads were located that close to the road network, which is perhaps surprising given the clear association of this type of settlement with major roads in other regions in the south and east of the country (see Chs 5 and 6). However, trackways were recorded on 75 per cent of complex farmsteads, highlighting the importance of the transport network beyond the major roads in the North-East. In some areas, cropmark evidence shows trackways running over considerable distances, surrounded by field systems and linking individual settlements (Brown et al. 2007; Roberts et al. 2001; FIG. 7.37). Dating evidence from some settlements suggests that many of these landscapes had continued from the late Iron Age, though two examples of metalled sections of late Roman rural 0 2 km 1:50000 FIG. Romanch7.indd 272 7.38. Plan of Romano-British ‘brickwork’ fields at Dunston’s Clump, Babworth (Garton 1987) 06/09/2016 17:16:02 THE NORTH-EAST Fieldwalking around one of the enclosure settlements produced evidence for manuring in the fields closest to the settlement, indicating their use for arable cultivation (Garton 1987; 2008). Determining whether fields were under pasture or being cultivated is notoriously difficult, particularly since fields may have supported both livestock and crops at different times. Open areas of land have been identified on the Magnesian Limestone Belt by Roberts, who suggested that they may have been for communal use, perhaps for livestock grazing, as at Wattle Skye where a large, open space was encircled by several enclosure complexes (Roberts 2010, 26). The division of open land has also been identified on the Yorkshire Wolds at Cat Babbleton Farm, where Romanperiod pit-alignments have been traced over long distances, perhaps also signifying rights over grazing (Cardwell 1989). Of course, in order to distinguish between pastoral and arable usage of fields, much depends upon the recovery of suitable environmental evidence. On the Humberhead Levels near Doncaster, a rare preserved assemblage of waterlogged wood, pollen, and insect remains, recovered from late Iron Age/early Roman field ditches at Balby Carr, clearly showed that they were used for grazing livestock (L. Jones 2007). Dung beetles were relatively common, while the pollen evidence indicated the presence of grassland with hedges and patches of woodland. Coppiced roundwood, found with tool marks, may also have been used for fencing, perhaps for stock control. The site was waterlogged and appears to have been unsuitable for arable agriculture, which likely occurred on higher, drier ground. Although numerous botanical assemblages from the North-East demonstrate that cereal farming was widespread (see Plant remains below), evidence for the cultivation of specific field systems is more sporadic. At North Thoresby, Lincolnshire, phosphate levels in the fills of a number of regularly spaced ditches were consistent with manuring. The excavators speculated that the site was used for vine cultivation (Webster et al. 1967), though no pollen or macrobotanical evidence was available to support their interpretation, and the layout of the field system may simply have been consistent with drainage. Further possible evidence for horticultural trenches was also identified at Burnby Lane, Hayton, close to the valley floor, though as at North Thoresby, the lack of corroborating environmental evidence means that we do not know how this land was actually used (Halkon et al. 2015). Greater insights into the arable use of field systems could be achieved through the analysis of weed assemblages from farmsteads and field systems, and further research should be directed towards this end. Field systems Romanch7.indd 273 273 and paddocks have also been recorded at roadside settlements, where ditched boundaries extended behind property plots lining the road (e.g. Wilson 2002). The presence of fields at these sites would support Millett’s suggestion that roadside settlements were essentially agricultural and pastoral in character, but also formed central places where produce could be gathered and processed (Millett 2006, 309). The dating of field systems in the North-East, as with other regions, remains problematic, though a considerable number of sites have been shown to originate in the late Iron Age and continue into the Roman period. Field systems were most numerous in the second century A.D., after which they reduced in number into the fourth century. The apparent decline in the number of field systems in use into the late Roman period is a pattern in common with the South and East regions. However, while many Romano-British field systems may have gone out of use, some appear to have influenced the form of the subsequent medieval landscape. Recent analysis of the medieval furlongs at Dunstan’s Clump shows that they were aligned with the primary boundaries of the Roman brickwork fields, which ran parallel to each other in an east–west direction. This suggests that, although the brickwork field system at Dunston’s Clump probably went out of use by the end of the Roman period, the boundaries could have survived as earthworks, providing a template on which the medieval open field system was lain (Rippon et al. 2015, 214). However, while this may have been the case at Dunston’s Clump, the extent to which this was repeated elsewhere is uncertain. SETTLEMENT HIERARCHIES: THE SOCIAL AND ECONOMIC BASIS OF SETTLEMENTS MATERIAL CULTURE It has been stated that trade and exchange in the North-East predominantly occurred on a local basis throughout the Roman period (Ottaway 2013, 146–9, 204–8). Although regional and long distance supply networks were facilitated by the army, there appears to have been minimal economic integration between the population of the countryside and those in urban and military centres. While broad differences in the type and quantity of ‘Roman’ commodities has been pointed out, the growing body of evidence from rural farmsteads is less well understood, and is here outlined further in order to better understand the different forms of settlement in the North-East. 06/09/2016 17:16:02 274 THE RURAL SETTLEMENT OF ROMAN BRITAIN Pottery As in the south and east of England, the most ubiquitous type of artefact recovered from Roman sites in the North-East is pottery. In the early Roman period, hand-made vessels common in the Iron Age continued to be used on rural sites, and there appears to have been little appetite for new ‘Roman’ forms of food preparation or dining customs (Ottaway 2013, 148). Later Roman ceramics, notably the fourth-century wares from Crambeck, are far more widely distributed among rural settlements, though the success of the Crambeck industry is largely attributed to its supply links with the military (Wilson 1989; Tyers 1996, 16, 74; Ottaway 2013, 290–1). The low level of trade in ‘Roman’ commodities between rural farmsteads and the towns and military bases in the North-East can be examined through the recovery of pottery on different types of site. Samian ware has been recovered in significantly higher proportions from roadside settlements and military vici, particularly at the latter where the use and trade of samian can be associated with the military (FIG. 7.39). If anything, the vici data are under-represented since these counts do not include the exceptional ceramic assemblages from Piercebridge and Castleford (the latter is technically outside the region), which skew the data so heavily in favour of vici that the comparatively tiny quantities coming from farmsteads and villas become almost undetectable. However, compared to samian ware, mortarium sherds are recovered in far more equal quantities average no. of sherds per site samian 400 350 300 250 200 150 100 50 0 roadside vicus (n=7) villa (n=8) complex enclosed farm (all) farmstead farmstead (n=90) settlement (n=23) (n=23) (n=13) average no. of sherds per site mortaria 40 35 30 25 20 15 10 5 0 roadside vicus (n=7) villa (n=8) complex enclosed farm (all) farmstead farmstead (n=90) settlement (n=23) (n=23) (n=13) FIG. 7.39. Recovery of samian and mortaria from different types of site in the North-East, based upon average sherd counts per site (not including Piercebridge) Romanch7.indd 274 between different types of site, which might suggest that the adoption and use of mortaria by rural communities was more consistent with those living along the road network. The distinction between the consumption of both samian and mortaria on villas and complex farmsteads, compared to enclosed farmsteads also draws attention. It is perhaps surprising that there is very little difference between the average sherd counts deriving from villas and complex farmsteads, yet both of these settlement types produce much greater quantities than that recovered from enclosed farmsteads. On closer inspection, the data from complex farmsteads is inflated by pottery from three sites: Faverdale, Heslington East, and Burnby Lane, Hayton. These are exceptional sites in terms of their material culture and architecture, and clearly stand out from other farmsteads. These sites remind us that there was probably little difference in terms of social standing and wealth between the higher-status, complex farmsteads and the more modest villas. Although in general a low level of economic integration between settlements on the road network and farmsteads may have been apparent, the data suggest that there existed some variation, with more evidence for trade and exchange being detectable on some farmsteads than others. Small finds The range of objects from the North-East closely mirrors those found in other regions, with quern stones, coins, knives and other tools featuring prominently. At roadside settlements, military vici and villas, most artefact types occur at similar frequencies, and are consistently more common than at farmsteads (FIG. 7.40). Coins, for example, are only recorded at 28 per cent of farmsteads, compared to 87 per cent of roadside settlements and 86 per cent of villas. By far the greatest numbers of coins have come from the roadside settlements at Shiptonthorpe and Hayton (Millett 2006; Halkon et al. 2015). However, these datasets are skewed by large numbers of coins from extensive metal-detecting programmes, alongside coinage recovered from excavations and fieldwalking. The largest brooch assemblage comes from the village at Dragonby, Lincolnshire, which included a considerable number of late Iron Age La Tène types (May 1996). Although coin and brooch assemblages are frequently small or absent from farmsteads, two notable exceptions are Burnby Lane, Hayton (Halkon et al. 2015), and Heslington East (Roskams and Neal 2012). These settlements are, perhaps significantly, located close to major nucleated centres and, as discussed above with regards to pottery, were probably higherstatus farmsteads more akin to villas. 06/09/2016 17:16:03 FIG. % of sites with object present 100 90 80 70 60 50 40 30 20 10 0 % of sites with object present THE NORTH-EAST 100 90 80 70 60 50 40 30 20 10 0 roadside selement/vicus (n=15) villa (n=14) complex farmstead (n=40) farmstead (n=194) enclosed farmstead (n=31) 7.40. Frequency of major artefact categories on different types of rural settlement in the North-East region Items concerned with Roman styles of dress and personal display, such as finger rings, toiletry objects and hair pins are almost completely absent from farmsteads, but are comparatively well represented at roadside settlements and villas. Combs for personal use have been recovered from fifteen sites, almost exclusively nucleated settlements and villas, including wooden examples from Shiptonthorpe (Millett 2006) and Rossington Bridge (Buckland et al. 2001). The only examples of bone combs recovered from farmsteads have occurred at Melton wastewater works, near Brough (Bishop and Westwood 2004), and Chase Farm, North Killingholme (Humberside Archaeology 1991). In contrast to personal-use items, utilitarian objects, such as food-processing items and knives/ tools, are more equally recovered from each of the three main settlement types, indicating the need for these object classes on farmsteads over other forms of material culture. Quernstones are notably more widespread across the Tees Lowlands and the Magnesian Limestone Belt in comparison to coins and brooches. Millstones are also widely distributed across the region, but have been found on only eighteen sites, seven of which are nucleated settlements that may have operated as centres where agricultural produce could be gathered and processed. Other settlements with high numbers of millstones include the village settlement at Wattle Syke on the Magnesian Limestone Belt (seventeen), and the villa at Ingleby Barwick in the Tees Lowlands (six), and these sites may have Romanch7.indd 275 villa (n=14) 275 similarly acted as centres for the centralised processing of grain in these sub-regions. Equipment for textile processing is very well represented at nucleated settlements, including the exceptional quantity of spindlewhorls, loomweights and sewing needles recovered from Dragonby (May 1996), and the two weaving combs and four weaving tablets at Malton (Wenham and Heywood 1997). The concentration of textile-processing artefacts at roadside settlements, military vici and villages may be a reflection of commercial activity, as opposed to household craft, which would involve the centralisation of sheep management, or at least their wool. Other potential indicators of commercial activity are comparatively sparse. Writing equipment has been identified at fourteen sites, mostly villas, though greater numbers of styli and seal boxes have been recovered from nucleated settlements, especially Dragonby (May 1996), Bainesse (Wilson 2002) and Shiptonthorpe, the latter of which contained two silver fir writing tablets (Millett 2006). Weighing objects, such as steelyards, are equally rare, coming from eighteen sites, though farmsteads more commonly produce this type of artefact. Religious objects appear most frequently on nucleated settlements, occurring in their greatest number at Dragonby, which contained two figurines, one of Mars Gradivus and the other of Mars Ultor, as well as ‘votive’ plaques and two Iron Age miniature shields. The votive finds were interpreted as being associated with a Mars cult 06/09/2016 17:16:05 276 THE RURAL SETTLEMENT OF ROMAN BRITAIN centre, or at least were indicative of a shrine (Alcock 1996; May 1996, 603). As well as nucleated settlements, eight farmsteads have also produced religious items, including a sceptre head in the form of Mars from Blackhills Farm and The Hollys, Wickenby (Hall 2008), and a pipe-clay figurine from Millfield Farm,Wheldrake (Robinson 2009), perhaps reflecting religious observance being carried out at individual households. In addition to items of a more ‘classical’ type, a number of sites include apparent religious objects that could have been worn, such as a phallus amulet from Burnby Lane (Halkon et al. 2015), a perforated boar’s tusk from Melton A63 (FentonThomas 2011), and a drilled dog tooth pendant from Newbridge Quarry, Pickering (Richardson 2012). As in other regions, villas clearly stand out in terms of their profile of material culture compared to farmsteads in the North-East. Artefacts relating to security (locks, keys, etc.) and recreation (gaming counters, dice, etc.) are fairly common finds in villa assemblages, but are exceptionally rare on farmsteads, reflecting differences in lifestyle and attitudes towards protecting property. There is little difference between the profiles of material culture of enclosed farmsteads and complex farmsteads, although the range of artefacts from complex farmsteads is slightly greater, with some producing lighting and writing equipment for example. Perhaps more surprising is the poor representation of agricultural tools (ards, spades, and scythes, etc.) and equine/ transport-related equipment (cart-fittings, hipposandals, etc.) at all types of farmsteads, considering that arable farming and livestock husbandry would have been the preoccupation of most of the population. Items associated with ploughing are particularly rare, being restricted to an ard tip recovered from Malmo Road, Hull (Tibbles 1992), and a ploughshare from Gunhills, Armthorpe (Richardson 2008). However, harvesting and horticultural tools, such as scythes, pitchforks, pruning hooks and spade shoes, are slightly more common. ENVIRONMENTAL EVIDENCE The environmental evidence from excavated sites is unevenly distributed across the North-East, with a greater number of animal bone and plant assemblages located in the central and southern parts of the region. These assemblages are predominantly restricted to low-lying land in the river valleys, but also on the Magnesian Limestone and the chalk where the alkalinity of the soils are good for bone preservation, but less so for botanical remains unless they are charred. Romanch7.indd 276 Animal bones A total of 90 well-dated, animal bone assemblages of over 100 identified specimens (NISP) have been recovered from sites located widely across the North-East. The largest assemblages derive from Piercebridge, Dragonby, Winterton villa, Wattle Syke, and Dalton Parlours, though there is a considerable lack of faunal assemblages from the Sherwood Sandstones, the North York Moors, and the Lincolnshire Coast and Marshes, which prevents a thorough examination of intra-regional variability. Despite these restrictions, a broad comparison of animal bone assemblages between different landscape zones shows some distinct variations in the relative frequencies of the cattle, sheep/goat and pig bones (FIG. 7.41). The Vale of York and Mowbray is heavily dominated by cattle bone, being almost twice as frequent as those of sheep/goat, whereas assemblages from Holderness and theYorkshire Wolds produce a higher frequency of sheep/goat remains. This is perhaps a reflection of the greater suitability of the chalk Wolds and the coastal lowlands for sheep-grazing, while settlements in the Vale of York may have been concerned with the supply of cattle to York itself, where their bones have been recovered in considerable quantities (O’Connor 1988, 75–81). An increasing frequency of cattle remains over time also occurred in the North-East, as with other regions in the south and east of England, and is a pattern which has been previously documented from national trends (King 1984; 1999). Sheep/goat remains clearly dominate over cattle in the late Iron Age and in the early Roman period, but are overtaken by cattle in the middle Roman period, and continue to be the most common domesticate into the fourth century A.D. (FIG. 7.41). Within this shift of emphasis, some assemblages continue to exhibit high sheep/goat percentages into the later Roman period, particularly on the Magnesian Limestone at Wattle Syke (Martin et al. 2013) and Parlington Hollings (Roberts et al. 2001), where a preference for mutton appears to have been maintained, perhaps alongside a greater emphasis on wool exploitation. A slight increase in the frequency of pig bones is also evident over time; they tend to be more common on sites in the Vale of York and Mowbray, though exceptionally high proportions of pig bone have been recorded at Cedar Ridge, Garforth (44 per cent) (Owen 1998). Alongside the evidence for intra-regional and chronological variation in the animal bone assemblages, there are also some distinct differences between different site types (FIG. 7.41). While assemblages from villas are disappointingly rare, they do show a clear preference for cattle, with some including a relative abundances of pig 06/09/2016 17:16:05 THE NORTH-EAST cale sheep/goat 277 pig 60 mean %NISP 50 40 30 20 10 0 Vale of York and Mowbray (n=8) Southern Magnesian Limestone (n=15) 60 North Lincolnshire Coversands and Clay Vales (n=10) Holderness (n=8) Yorkshire Wolds (n=8) mean %NISP 50 40 30 20 10 0 LIA (1st C BC-mid 1st C AD) n=7 LIA/ER (1st C BC/AD) n=13 ER (1st-2ndC AD) n=16 MR (2nd-3rdC AD) n=12 LR (3rd-4thC AD) n=19 Mean %NISP 60 50 40 30 20 10 0 villa (n=6) farmstead (complex) (n=18) farmstead (enclosed) (n=10) farmstead (all) (n=60) roadside selement /vicus (n=12) village (n=6) FIG. 7.41. Relative frequency of cattle, sheep/goat and pig: (a) by landscape zone; (b) by period; and, (c) by major settlement type (mean percentages from sites with >100 NISP) bones, most notably at Dalton Parlours (15 per cent) and Holme House (31 per cent in the early Roman phase). The preference for pork on villa sites is also found in other regions and is a welldocumented feature nationally (King 1984; 1991; 1999). Dalton Parlours and Holme House also produced relatively high proportions of sheep/goat bones, which is at odds with assemblages from other villas. It is uncertain whether this reflects a real difference in consumption patterns, or whether the villa assemblages were affected by residuality from earlier phases of occupation. Unlike pottery consumption (see above, p. 274), there is very little difference in the proportion of major domesticates between different types of farmsteads, though there is significant variation between individual assemblages. Faunal assemblages from the villages at Dragonby, Low Caythorpe and Wattle Syke all produced a consistent pattern of high sheep/goat frequencies, though roadside settlements and military vici, while being heavily cattle-dominated overall, also greatly vary between different sites. It may be significant that the highest proportions of cattle bone were identified in assemblages from the vici at Doncaster (Buckland and Magilton 1986; Chadwick and Burgess 2008), Piercebridge (Cool and Mason 2008) and Rossington Bridge (Buckland et al. 2001), which may reflect the importance of cattle to the military (cf. King 1999). In contrast, the faunal material recovered from the roadside settlement at Mount Pleasant Romanch7.indd 277 House, Nettleton and Rothwell, was overwhelmingly dominated by remains of sheep/goat, which was possibly associated with religious activity occurring in the vicinity of the shrine identified at the site (Willis 2013). Few faunal assemblages in the North-East have been recovered from overtly religious contexts, though special mention must be given to the Iron Age chariot burial and nearby shrine at Ferry Fryston (Brown et al. 2007; FIG. 7.42). The chariot burial itself dated to the second century B.C., though radiocarbon dating of cattle remains recovered from the upper fills of the enclosing ditch demonstrated that the monument was a site of ritual significance well into the Roman period, with most of the remains having been deposited in the third and early fourth centuries A.D. (Bates et al. 2007). The bones derived from a minimum of 162 cattle, but were almost exclusively skulls and right forelimbs, articulating from the shoulder to the foot, with little evidence for butchery marks, suggesting careful methods of carcass dismemberment (ibid.). Tooth wear patterns showed that most of the animals were slaughtered at either one-and-a-half or two-and-a-half years old, indicating periods of intensive culling during the late summer/autumn months. It is therefore possible that the slaughter may have been a seasonal event, with deposits accumulating over a period of time. The lack of carnivore gnawing on the bones demonstrated that the body parts were deposited quickly, or at least were protected from 06/09/2016 17:16:08 278 THE RURAL SETTLEMENT OF ROMAN BRITAIN N Early-prehistoric feature Early-prehistoric features Cattle bones placed in ditch Iron Age chariot burial Neolithic henge Possible RomanoBritish shrine l va -ha ha die me tos P 0 50 m 1:1000 FIG. 7.42. Religious monuments at Ferry Fryston showing location of cattle deposits around the Iron Age chariot burial (Brown et al. 2007) scavengers, leading Orton (2006) to suggest that the remains may have been curated for some time prior to being deposited in a single episode. In terms of the sources of origin of the cattle, strontium isotope analysis showed that many were not raised on the Magnesian Limestone, where the site is located, but were imported from further afield (ibid.). The cattle chosen for deposition may have been traded from a variety of sources, or possibly the rituals undertaken at the site were fulfilled by gatherings of people arriving with cattle over long distances (either as live animals for ritual slaughter or as carcass parts). Another striking aspect of the site is the absence of Roman material culture, a feature also raised by Hodgson who suggested ‘that the participants in ceremonies here were detached from Roman provincial culture’ (Hodgson 2012, 52). Although the true nature of the activities being undertaken are uncertain, the Ferry Fryston material is a prime example of the important spiritual role held by livestock throughout the Roman period. Alongside the three main domestic livestock species, a range of other animals make up the remaining fauna in animal bone assemblages from the North-East, the most common of which are equids. Horse bones almost always form less than 20 per cent of faunal assemblages, though this varies from site to site. Of the sites that contain over 10 per cent horse bones, eight are complex farmsteads and only two are enclosed farmsteads. The use of horses for riding perhaps outweighs their use as pack animals, though increasing Romanch7.indd 278 evidence for other equid species is now emerging. Reanalysis of late Iron Age equid specimens from Thorpe Thewles by Johnstone identified evidence for donkey and mule in the assemblage, which she suggested might have been imports from the continent (Johnstone 2004, 246). Other examples of possible mules have been identified at Hayton fort (ibid.) and a late first/second-century A.D. burial in the vicus at Healam Bridge (Ambrey et al. 2011), suggesting that these animals were being utilised by the Roman military. Dog bones are found on 75 per cent of sites, though rarely in any great quantity, unless recovered as associated bone groups, as discovered at Dalton Parlours where two partially articulated dog skeletons were found in a pit deliberately lined with bones from pigs and sheep/goats, possibly forming an elaborate burial rite. Cat bones are far rarer, and have only been identified on 15 per cent of sites. Bone from wild species is equally sparse, though a considerable quantity of red deer bones was recovered from the complex farmstead at Heslington East, suggesting that hunting was undertaken to some degree (Roskams and Neal 2012). Butchery marks found on a red deer scapula at the site indicated that it may have been from a shoulder of venison hung for smoking or salting. It is clear that antler, particularly from red deer, was widely utilised for tool manufacture. In particular, late third–early fifth century A.D. deposits of worked antler at Wattle Syke suggest that this was an important part of the economy of the site during this period (Martin et al. 2013). 06/09/2016 17:16:08 THE NORTH-EAST roadside selement/vicus (n=18) 279 farmstead (n=125) % of assemblages 100 90 80 70 60 50 40 30 20 10 0 FIG. 7.43. Presence of arable crops and other plant taxa from roadside settlements/vici and farmsteads in the North-East Plant remains Of the 153 sites with recorded plant remains most are, unsurprisingly, biased towards the lowland areas of the region, being most common on the Lincolnshire Coversands and Clays (12 per cent), the Southern Magnesian Limestone (12 per cent), the Humberhead Levels (12 per cent), and in the Tees Lowlands (11 per cent). There are a disappointingly small number of samples located on the chalk of Yorkshire and Lincolnshire Wolds, and the alkaline soils here may have hindered preservation. The range and proportion of plant taxa identified from floral assemblages is highly consistent with other regions, with spelt wheat dominating, although barley is the best represented cereal taxa in terms of its overall presence (FIG. 7.43). Barley, however, is rarely found in abundance within samples, exceptions including Site 20–4 (West) on the Asselby to Pannal pipeline, where a late Iron Age pit produced a large quantity of processed barley grain, mixed with a few oats. Only at Dragonby has barley been found in greater quantities than spelt wheat in samples across different phases of the site (May 1996). Emmer wheat is mostly identified on sites south of the Humber, though is also present at a few sites in the Tees Valley. When emmer wheat is identified in relative abundance it is generally from late Iron Age or early Roman samples, such as at Raymoth Lane, Worksop (Palmer-Brown and Munford 2004) and Stenigot Reservoir, Donington-onBain (Lindsey Archaeological Services 1997). In comparison, free-threshing wheat is widely distributed across the North-East, though it consistently features in low proportions. It only appears to have been the dominant cereal taxa in samples recovered from 8–10 High Street, Doncaster, where it was identified alongside rye and oat grains (Chadwick and Burgess 2008). Romanch7.indd 279 As in other regions, oat remains are fairly ubiquitous finds in archaeobotanical samples (identified at 58 per cent of sites), but nearly always represent a very minor proportion of the assemblages. In most reports, the archaeobotanical specialist dismisses the finding of oat as a weed growing alongside other cereals. However, the recent excavation of the complex farmstead at Rossington Colliery, South Yorkshire, found oat remains alongside spelt and barley in abundant quantities, suggesting that it may have been deliberately cultivated and processed in the late Roman period, perhaps as a fodder crop (Roberts and Weston 2016). In contrast, rye is very poorly represented, found on only 8 per cent of sites, and appears to have been a genuinely marginal crop in the North-East, while pulses are clearly more common, present on nearly 20 per cent of sites. Broad beans are the common type of pulse identified, though lentils have been recovered from the Waterdale site at Doncaster (Davies 2013), and cultivated peas were recovered in some quantity from a late Roman ditch at Site 2 on the Asselby to Pannal pipeline (Gregory et al. 2013). Fruit remains have only been identified from 10 per cent of sites, reflecting the fact that they tend to require waterlogged conditions for preservation. Large quantities of hawthorn and elderberry at Dragonby perhaps reflect seasonal gathering, though the Waterdale site at Doncaster produced a range of fruit and nuts, including imported and comparatively rare foods, such as fig, grape, date, walnut and stone pine (Davies 2013). At Frenchgate (Site DG) in Doncaster, a secondcentury A.D. pit included 1400 apple seeds, estimated to have been from 150–300 apples, alongside 23 plum stones, and a few seeds of sloe, elderberry and grape. The high number of apple seeds was interpreted by the archaeobotanist as possible waste from cider production, though the 06/09/2016 17:16:10 280 THE RURAL SETTLEMENT OF ROMAN BRITAIN feature was also considered to have been a cess pit (Williams 1986, 198). Although this is an outstanding example, the more frequent recovery of horticultural foods, such as fruits and pulses, appears to distinguish roadside settlements from farmsteads. Otherwise, the presence of most plant taxa is similar between the two types of site, with differences potentially lying in the scale of processing and consumption rather than the range of foods eaten. In general, there is little difference between the main settlement types in terms of the presence of different plant taxa. One of the few differences between enclosed and complex farmsteads is that a much greater proportion of the former have produced emmer wheat (42 per cent), and, as mentioned previously, this appears to have been more common on late Iron Age sites. Only six villa sites have records of plant remains, and their poor representation is further hampered by small samples. The exception is found at Ingleby Barwick where a wide range of plant taxa included freethreshing wheat, oats, flax and Celtic bean, while the weed seed assemblage suggested that the heavier clay soils were being tilled (Huntley 2013b). REGION SUMMARY As we have seen throughout this chapter, the North-East was a diverse region with a range of landscapes that fostered quite different patterns of settlement and land use. These reflect localised, cultural traditions which persevered and developed from the Iron Age, and it may be tempting to associate this intra-regional diversity with the late Iron Age tribal groups that are thought to have held political power across the region – the Brigantes, the Corieltauvi, and the Parisi. Caution must be exercised here, however, since our knowledge of these groups is based upon sketchy historical records written from a Roman perspective (see Ch. 12 for discussion of ‘tribes’). We have very little idea as to how far the ‘everyday’ activities of the rural population were a reflection of their political identity, if at all. Instead, it may be more productive to look beyond the tribal history of the region and assess the diversity of the archaeological record in its own right with regard to farming activity, trade and exchange, and the adoption of ‘new’ building forms and material culture. The Roman conquest of this region was a gradual process, taking place over a number of decades in the second half of the first century A.D. (Bidwell and Hodgson 2009). Excavated evidence has shown that existing settlements were quickly swept aside in advance of road or fort construction. Yet, the impact of the Roman military on the wider landscape is less obvious. Many late Iron Romanch7.indd 280 Age field systems appear to have been maintained into the second century, retaining their form and orientation, and it is not until the late Roman period that more significant changes occurred. The presence of the military appears to have affected patterns of farming, industry, and trade, over a long period of time. The establishment of fortresses and coloniae at Lincoln and York may have been responsible for stimulating the expansion of rural settlement in the Vale of York and Mowbray and the Lincolnshire Coversands and Clays during the early Roman period. However, this pattern is quite different in landscapes where Roman influence was minimal. In Holderness, settlement numbers declined after the late Iron Age, while in the Yorkshire Wolds the distinctive ladder settlements seen across the chalk appear to have continued from the late Iron Age into the Roman period, with little evidence for disruption or change (Stoertz 1997, 52–4). The increasing frequency of cattle bones over time may have reflected their economic importance to the towns, roadside and military settlements. Certainly, sites with higher frequencies of cattle tend to be located in landscape zones through which the major roads ran. In contrast, a greater emphasis on sheep and goats in the Yorkshire Wolds and on Holderness may be tied to more traditional forms of pastoral farming. A telling feature of the ladder settlements in the Wolds is their association with the local topography, as seen at Heslerton and Burton Fleming, with their long droveways and enclosure complexes, indicating that long-distance movement of livestock may have been important. The cattle bone evidence from Ferry Fryston is also highly suggestive of seasonal transhumance of people and animals, while the lack of Roman material culture implies the presence of communities who were not fully engaged with Roman systems of trade and exchange. The lack of economic integration between settlements on the road network and outlying farmsteads is partly borne out in the patterns of material culture consumption. That roadside settlements were centres where agricultural surplus could be gathered, processed, and redistributed, is indicated by the presence of corndryers and millstones, while textile equipment is noticeably more common than on other types of site, perhaps pointing to the commercial importance of wool. The presence of mules on sites on the road network is also significant for the transportation of goods and produce, while imported and locally cultivated, horticultural plant foods indicate that the diets of some inhabitants of roadside settlements were relatively diverse. Some farmsteads and villas also clearly benefited from 06/09/2016 17:16:10 THE NORTH-EAST the Roman economic system, such as at Faverdale and Heslington East, where higher numbers of coins and other finds have been noted. In addition, some imported pottery types were more widely adopted than others on farmsteads, which perhaps suggests that new forms of food preparation and dining filtered into the countryside, but on a limited and varied scale. Major financial investment can be seen in the architecture of some of the larger villas, such as at Scampton and Beadlam, while more recent excavations have improved our understanding of the landscape context of some villas. Investigations at Ingleby Barwick and Welton Wold have revealed evidence for wider ‘estates’ in which these villas may have operated. Romanch7.indd 281 281 Overall, the North-East is a region that lies on the edge of the ‘core zone’ of Roman Britain, connecting the wealthy, arable-based communities of the Central Belt to the military-controlled, uplands of the north. The establishment of roads, towns and forts brought with it new opportunities for the local population. Yet the region has a distinctive character, with forms of settlement not seen elsewhere in Britain, and there is a strong level of continuity from the late Iron Age. We get a sense of conservatism prevailing in the countryside, though the adoption of new Roman fashions varied considerably between different settlements; this is an aspect that will be examined further in volumes 2 and 3 of this series. 06/09/2016 17:16:10