Champion and Woodland? Landscape Evolution beyond the Central Zone in Greater East Anglia
Champion and Woodland? Landscape Evolution beyond the Central Zone in Greater East Anglia
Abstract and Keywords
This chapter considers a region that is termed ‘greater East Anglia’, embracing Norfolk, Suffolk, Essex, and Cambridgeshire, which straddles the eastern edge of England's central zone characterized by villages and open fields. To the south of the Gipping and Lark valleys in Suffolk (i.e. southern Suffolk and Essex) there was a considerable degree of continuity between the Roman and medieval periods with no evidence for a major restructuring of the landscape. To the north, there was a significant change in how the landscape was exploited with a nucleation of settlement and intensification of agriculture around the eighth century. This emergence of villages—‐which is probably part of the same phenomenon seen in the East Midlands—‐was, however, short‐lived, and a greater degree of dispersion soon emerged in most areas.
The preceding chapters examined the south‐western limit of England's central zone and found that while the upland areas of the Blackdown and Quantock Hills appear to mark the limit to which villages and Midland‐style common fields spread across Somerset, topography and soils do not explain all of the local and regional variation within that county's landscape character. Indeed, across the Blackdown and Quantock Hills, the countryside in Dumnonia appears to have undergone a profound period of change around the eighth century, several centuries before the creation of villages and common fields in Somerset. It also appears that the south‐west peninsula had seen a significantly different cultural landscape compared to areas further east for many centuries and this deep‐rooted distinctiveness may have been a significant factor in its divergent landscape development during the ‘long eighth century’. We can now turn to another case study of regional variation in landscape development at the margins of England's central zone in what is referred to here as ‘greater East Anglia’ (Fig. 5.1).
This study area embraces the counties of Norfolk and Suffolk (which together form East Anglia in the modern sense), along with Essex and the eastern part of Cambridgeshire. It straddles the boundary between Roberts and Wrathmell's (2000, fig. 10) ‘central’ and ‘south‐eastern’ provinces, Rackham's ‘planned’ and ‘ancient’ landscapes, and Leland's ‘champion’ and ‘woodland’ countryside. This region has benefited from a number of recent seminal studies including Christopher Taylor's (2002) examination of the eastern limit of villages and common fields in Cambridgeshire, Tom Williamson's (2003; 2006a; 2007) discussion of the development of the medieval countryside across East Anglia and the East Midlands, and the results of Edward Martin and Max Satchell's (forthcoming; Martin 2007) East Anglian Fields Project. The region benefits from a virtually continuous and relatively well‐dated ceramic sequence, and has a long history of fieldwalking (made possible (p.139)
THE PHYSICAL AND CULTURAL REGIONS OF ‘GREATER EAST ANGLIA’
In contrast to the dramatic variations in topography seen across Somerset and the south‐west of England, greater East Anglia is a region of more subtle—although nonetheless significant—differences in relief, geology, and soils (Fig. 5.1). The region is dominated by a Boulder Clay plateau that extends from the northern coast of Norfolk, through Suffolk, and into central Essex. To the west there is a chalk escarpment that forms the northern extension of the Chiltern Hills and underlies the ‘East Anglian Heights’, beyond which lies the lowlands of Hertfordshire, Cambridgeshire, and the vast wetland expanse of Fenland. The chalk itself only outcrops on the scarp‐face and in valley sides, elsewhere being capped by superficial drift deposits, notably clay with flints and brickearths on the Chilterns themselves, extensive areas of sand known as Breckland on the Suffolk–Norfolk border, and Arthur Young's (1804, 10–12) ‘Good Sand’ region to the north. To the south of the Boulder Clay, in southern Essex, there are extensive areas of heavy London Clay, while the eastern, coastal, districts of Norfolk, Suffolk, and Essex are dominated by a series of estuaries that dissect areas of lighter loams and sandy gravels. Across the whole of greater East Anglia the topography is relatively flat, although the extensive Boulder Clay plateau is dissected by river valleys that have lighter, better‐drained soils compared to the interfluvial areas and watersheds.
Culturally, the north‐west corner of modern Essex around Great Chesterford—beyond the chalk escarpment—appears to have looked towards northern Hertfordshire and Cambridgeshire. Artefact distributions suggest that in the Iron Age and Roman periods it lay within the Catuvellaunian tribal area/civitas, and in the fifth to seventh centuries the Early Saxon cemeteries around Great Chesterford share their greatest affinities with the Cambridge region (Fig. 5.2; Evison 1994, 46–518, fig. 10; Baker 2006b, fig. 3). To the east, the Gipping and Lark valleys—or the interfluvial area to the south that (p.141)
There is also a marked difference in the landscapes of northern and southern East Anglia in the early medieval period, with far greater evidence for Anglian settlement in the north than there is for actual Saxon immigration in the south, which is largely restricted to coastal districts of Essex and the Chelmer valley (Brown 1981; Evison 1981; Hines 1984; West 1998). Although Parker Pearson et al. (1993) suggest that the famous early seventh‐century cemetery at Sutton Hoo, in south‐east Suffolk, may have been the burial ground of an East Saxon king, and the material culture both there and at the nearby cemetery of Snape shows some affinities with that in Essex (Filmer‐Sankey and Pestel 2001, 265), Bede states that Rendlesham between the two was the East Anglian king's country seat (Farmer 1990, 180; although Blair (2005, 66–90) has argued that such villae regialis may not have been such permanent features of the landscape as has often been assumed). Other evidence also firmly points to the Gipping–Lark boundary as marking the division between East Anglia and the East Saxons. Series S sceattas, for example, were probably the royal coinage of the East Saxons produced as an expression of independence from their political overlords in Mercia: coins of Mercia were copied, but the king's head was replaced by a sphinx, perhaps recalling the past classical glories of Colchester (Metcalf 1993, 21; 2001, 46; Rippon 1996b, 117; Morris 2005, figs. 9.5–6). Their occurrence is largely restricted to modern Essex, Hertfordshire, and London, a distribution that is almost mutually exclusive to that of Ipswich Ware, whose widespread use was restricted to East Anglia (Fig. 5.2D: Blinkhorn 1999). Despite the proximity of Essex and south‐west Suffolk to the production site at Ipswich, very little of this material moved (p.143) south. Only two sites—Wicken Bonhunt (Wade 1980) and Barking (Redknap 1991)—have produced more than a handful of sherds, and of the remaining sites there is a strong coastal relationship, a pattern that the recent find of a near complete vessel from Althorne Creek, on the Crouch estuary, reinforces (Rippon 1996b, 117; Cotter 2000, 26; Walker 2001).
This distribution of Ipswich Ware suggests that the Gipping–Lark line was therefore the boundary between the ‘East Sexena’ and ‘East Engle’ of the Tribal Hidage: whilst in other periods the distribution of pottery is largely a reflection of trading links, this was a period before a money‐based market economy when exchange would have been more socially embedded. The distribution of eighth‐century ‘productive sites’ in East Anglia could also be significant. These possible centres for trade and exchange occur across south‐east England, often in coastal and estuarine locations, particularly around the fringes of East Anglia: with the exception of Caistor St Edmunds, most of these sites occur around the margins of the main area of Ipswich Ware, and the location of Brandon in the Lark Valley, and Barham and Coddenham in the Gipping Valley, upstream from the emporium at Ipswich, reinforces the impression of these valleys being a zone of liminal interaction. It was only around the tenth century, when the English shires were created, that the boundary between Essex and Suffolk was established along the Stour valley to the south.
A NOTE ON TERMINOLOGY
Across East Anglia the traditional terminology of ‘Early Saxon’ (AD 410–700), ‘Middle Saxon’ (AD 700–850), and ‘Late Saxon’ (AD 850–1066) is used for both periods and material culture (i.e. we read of ‘Early Saxon pottery’ and the ‘Early Saxon period’). This confusion of cultural identity and period terminology is unfortunate and potentially misleading as an ‘Early Saxon’ site (in terms of a settlement occupied during the fifth to seventh centuries) may actually have been inhabited by the direct descendants of the native Romano‐British population rather than immigrants of ‘Germanic’ origin. The arguments over who is buried in fifth‐ to seventh‐century Anglo‐Saxon cemeteries is now well rehearsed, but whether one takes a minimalist or maximalist view of the extent of the Anglo‐Saxon migration, it is clear that there was at least some settlement of people of continental origin in fifth‐century greater East Anglia, and that they brought with them new forms of burial rite (cremation and the deposition of grave goods), buildings (sunken‐featured buildings or Grubenhäuser), and material culture (e.g. Schlickung‐treated pottery with a distinctive surface treatment involving the application of a coarse slip, and which is largely restricted to the fifth century: Hamerow 1993, 31–5). In other cases, however, ‘Germanic’ material culture could easily have been used by (p.144) descendants of the native Romano‐British population. For example, assertions such as ‘excavated evidence shows that the first wave of Saxons settled on Roman villa sites at Little Oakley . . . and at Rivenhall’ (Tyler 1996, 110) are not necessarily true: both sites have indeed produced sherds of fifth‐ to sixth‐century pottery with clear continental affinities but there are no distinctively ‘Germanic’ Grubenhäuser or burials with grave goods. It is impossible to say whether this pottery was used by first‐generation immigrants from the continent and their descendants, or the native British occupants of the site who no longer had access to pottery produced in the Romano‐British tradition. Unfortunately, the terms ‘Early Saxon’, ‘Middle Saxon’, and ‘Late Saxon’ are so well embedded within the literature of Norfolk, Suffolk, and Essex it is impossible to avoid their use here, but the terminology must be used with great care.
A CHARACTERIZATION OF THE HISTORIC LANDSCAPE
Based on Roberts and Wrathmell's (2002, fig. 1.4) analysis of nineteenth‐century settlement, it would appear at first sight that the regionally significant difference in landscape character either side of the Gipping–Lark valleys had disappeared over the course of the medieval period (Fig. 5.3A). Their characterization divides the historic landscape of greater East Anglia along a roughly north‐east to south‐west line with predominantly dispersed settlement patterns to the east (their ‘Anglia’ sub‐province) and more nucleated patterns in the west (their ‘Wash’ sub‐province). They suggest, however, that the latter lay outside their central zone that extended only as far as western/central Cambridgeshire and north‐west Essex (the ‘East Midlands’ sub‐province). This boundary of the central zone broadly concurs with Gray's (1915, frontispiece) mapping of the two‐ and three‐field systems (Fig. 5.3B), and more recent work has confirmed that the scale, structure, and management of the common field systems in Norfolk and Suffolk were indeed different from those of the Midlands (see below). There was, however, also a significant difference in the character of the field systems within northern East Anglia. To the east, the open fields were enclosed earlier and in a piecemeal fashion compared to the west, with very little common field surviving into the era of Parliamentary Enclosure. By the sixteenth century these central and eastern districts of Norfolk and Suffolk were predominantly pastoral, used mostly for dairying and pig rearing, while the western areas, with their light sandy soils, were characterized by mixed sheep‐corn farming (Thirsk 1967, 40–9).
This east–west difference in the character of the medieval landscape of northern East Anglia is evident in the work of Roberts and Wrathmell (Fig. 5.3A), (p.145) although Rackham (1986a, fig. 1.3) is wrong to place western Norfolk and Suffolk in the region of England's champion landscape (his area of ‘planned countryside’: Fig. 5.3C) as this was not a landscape of Midland‐style villages and two‐ and three‐field systems. It is certainly true, however, that the boundary of the central zone in northern East Anglia is not as clear as it is to the south, where the edge of landscapes characterized by strongly nucleated settlement and regularly arranged common fields corresponds with the Chiltern escarpment, to the east of which settlement patterns were highly dispersed and common fields were of more limited extent (Roden 1973, 345–55; Lewis et al. 1997). This difference in the clarity of the central province's eastern boundary—the south being clearer than the north—suggests that there may have been a significant difference in landscape character within greater East Anglia, and this can be seen in Williamson's (2006a, fig. 3.12) simple but effective division of nineteenth‐century field boundary patterns into those indicative of planned enclosure, earlier piecemeal enclosure, and irregular field boundary patterns suggestive of direct enclosure from ‘waste’ (Fig. 5.3D): the boundary between the last two types runs along the Gipping–Lark valleys.
Another indication of a significant north–south variation in medieval landscape character is the distribution of deserted medieval villages (DMVs). The correlation between known DMVs—as mapped by Beresford and Hurst (1971, fig. 13)—and Roberts and Wrathmell's (2000, fig. 21) ‘central province’ is generally strong but with the exception of two main areas. The first is eastern Wessex, where a large number of DMVs lie outside their ‘central province’, although here a strong case can be made for its inclusion as it had a pattern of predominantly nucleated (albeit linear) villages in the nineteenth century, very low levels of dispersed settlement also reflected in the scarcity of place names with the element ‘green’, and extensive regularly arranged two‐ and three‐field systems in the medieval period (e.g. Roberts and Wrathmell 2000, fig. 23; 2002, figs. 2.10–11; Dyer 2001, 118; Hinton 2005). This is significant as the second area that has an abundance of deserted medieval villages outside the ‘central province’ is Norfolk and northern Suffolk. We have to take great care in assuming that Beresford and Hurst's database of ‘deserted medieval villages’ is an accurate reflection of the degree of settlement nucleation in the medieval period as, for example, some of what are called ‘villages’ may have been no more than hamlets, while other abandoned villages may lie undiscovered, but as it stands it is another indication that the northern part of East Anglia had a significantly different landscape character in the medieval period compared to the area to the south. Indeed, an overview of the results from fieldwalking and large‐scale development‐led excavations suggests that the southern part of East Anglia never did have nucleated villages (see below).
The most comprehensive analysis of landscape character in this region has been the East Anglian Fields Project that studied both the physical fabric of the historic landscape and documentary sources that describe how the fieldscapes were managed (Fig. 5.4; Martin 2007; Martin and Satchell forthcoming). In the south and west of Cambridgeshire they have confirmed that Midland‐style, regularly arranged two‐ and three‐field systems predominated (‘Type 1 common field’). In the east of Cambridgeshire, along with north‐west Suffolk and western Norfolk, common fields on a similar scale to those of the Midlands frequently survived into the period of Parliamentary Enclosure, although they were managed in a less regular way (‘Type 2a common field’). There were some communal cropping and folding arrangements, and the strips of individual tenements were spread fairly evenly across the fields, but there was little ridge and furrow and larger numbers of smaller fields than seen in the Midlands. In eastern Norfolk and coastal districts of Suffolk the open fields tended to be even more numerous but less extensive (i.e. covering a smaller percentage of the agricultural land within a community), the remaining areas being held in severalty, as closes (‘Type 2b common field’). These small open fields were subject to earlier enclosure, complex/flexible cropping arrangements (the ‘shift’ system), have less evidence for communal cropping and folding, and a marked tendency for the strips of individual tenements to be concentrated close to the homestead of the tenement rather than right across two or three great fields (and see Postgate 1973; Hesse 1997; Amor 2006). In the far south of Norfolk and north‐eastern Suffolk the evidence for open fields is less and they were mostly enclosed by the sixteenth century (‘Type 3 common field’; and see Dyer 2007). They covered less than half the agricultural land in a parish, and there is little evidence for communal cropping and folding. Landholdings were (p.148) concentrated in those fields nearest to the farmstead to which they belonged rather than being evenly distributed across all the fields.
To the south of the Gipping–Lark valleys the medieval field systems were very different from Martin's ‘block holdings’. Gray (1915, 387) suggested that ‘the early field system of few English counties is so difficult to describe as that of Essex’. The fieldscape of this area, and indeed south‐west Suffolk, is actually fairly simple in that it appears to have been dominated by what were always enclosed fields or closes. There were some open fields but they were very limited in extent, and were subdivided between just a few tenements that were located in the numerous hamlets that characterized the dispersed settlement patterns. Roden (1973, 355–63) could find no explanation for this difference in field systems within greater East Anglia in terms of the strength of manorial authority, inheritance patterns, or the extent of woodland and pasture. Martin (2007) suggests two possible causes. The first relates to topography: south of the Gipping–Lark valleys the landscape is more undulating and so has more sloping ground, whereas to the north there are greater areas of flatter, more poorly drained, land and he argues that common fields in the north were a way of sharing out the more limited areas of good land. This accords with Williamson's (2003; 2006a, 28) contention that the character of different field systems is primarily a function of the way in which the natural environment determines what farming practices are most effective, with the nature of settlement patterns following on from that (nucleated villages being appropriate for strongly communal forms of landscape management, and more dispersed patterns for smaller‐scale open field farming and tenements held in severalty). He therefore suggests that the differences in field systems seen either side of the Gipping–Lark line reflect ‘in large measure, aspects of soils and topography, for north of the Gipping the clay plateau was only infrequently cut by major valleys, whereas to the south it was much more dissected’ (Williamson 2002, 84). Rather than a long‐term cultural boundary between peoples, Williamson (2006a) also suggests that the sustained differences in material culture in these two regions are due to the ways in which topography has influenced the orientation of economic systems.
Martin also argues, however, for a significant cultural factor, in that the area of common fields in the north appears to correspond with the area of greater Scandinavian influence reflected in place names, language, and the proportion of the rural population that were freemen. The majority of Anglo‐Scandinavian metalwork is also concentrated to the north of the Gipping and Lark valleys. He therefore suggests that the ‘block’ field systems of Essex and southern Suffolk are the oldest in our greater East Anglian region and that to the north landscapes such as these were restructured following the Scandinavian (p.149) settlement, when land was redistributed amongst peasant farmers. The antiquity of these ‘block field’ systems of southern greater East Anglia is also inherent in the traditional views of how Midland‐style villages and common fields impinged upon the region, spreading from the Midlands but stopping short of the South‐East, leaving its existing landscape to develop in its own way (Fig. 1.3). In explaining this long‐term boundary in landscape character, we therefore have a fascinating set of opposing hypotheses that include the rival of some very traditional views, and as such make this an exciting time to be placing this area in a wider context.
It would appear, therefore, that in the context of this case study there are three crucial issues that we must address. First, does the landscape of southern East Anglia represent a slowly evolving, but essentially Romano‐British countryside, unaffected by the later transformations seen in areas such as the Midlands? Secondly, when did the distinctive landscape character of northern East Anglia emerge? And thirdly, what is the date when villages and common fields were created at the eastern edge of the central zone in Cambridgeshire, and why did this transformation of the landscape not spread any further into East Anglia? The approach to answering questions such as these in the South‐West—which is aceramic and poorly documented in the early medieval period—was to see how far the late prehistoric/Romano‐British landscape continued in use during the early medieval period, and how far back we can trace the origins of the historic landscape of today. In greater East Anglia the same approach is adopted, although in this case a near continuous ceramic sequence gives us far better chronological definition.
THE ORIGINS OF THE HISTORIC LANDSCAPE IN SOUTHERN EAST ANGLIA
‘Persistence, Continuity and Change’: The Trinovantian Civitas and the Kingdom of the East Saxons
The extensive surveys and excavations that have been carried out particularly in Essex make this an ideal region within which to study the Roman to medieval transition, and Morris (2005; 2006) has recently reviewed ‘persistence, continuity and change’ in the Essex landscape. The story that is emerging is that while a number of sites appear to have been abandoned during the third century, around 70 per cent of first‐century AD settlements were still occupied in the fourth century. This mid Roman period of change included the replacement of small fields with larger enclosures at a number of sites, and may have been due to the emergence of larger agricultural estates or a change in the emphasis of agricultural production from sheep to cattle, seen in the (p.150) food consumed in towns such as Colchester, Chelmsford, Great Dunmow, and Kelvedon, and at rural sites such as the villa at Wendens Ambo (Luff 1993, 128–38; Going 1996, 104). In the most recent overview of the Essex landscape in the Romano‐British period, Going (1996, 104) suggests the fourth century was one of decline, although the circulation of albeit small amounts of both late fourth‐century pottery (including shell‐tempered ware and late Oxfordshire colour‐coat types) and coins (up to Honorius and Arcadius) does suggest that many settlements were still occupied at least until the last decade of the fourth century (Drury and Rodwell 1980, 71; Lavender 1997; Garwood 1998; Germany 2003a; Baker 2006b, 61–8, 78–80; Morris 2006, fig. 7.37). There is also a surprising amount of evidence for the demolition of major buildings in the late fourth century, suggesting that the material was being reused elsewhere (Morris 2005, 38–9; 2006, 164–8).
Bryn Morris (2005; 2006) has also found that across Essex 34 per cent of Romano‐British settlements occupied in the fourth century have produced small amounts of what in Essex is called ‘Early Saxon’ (i.e. fifth‐ to seventh‐century) pottery. Crucially, these Romano‐British sites with fifth‐ to seventh‐century pottery are distributed right across Essex, having a far broader distribution than distinctively Germanic sites associated with burials with grave goods, Grubenhäuser, and Schlickung‐treated pottery that are largely restricted to the far south and east of the county, and around Great Chesterford in the far north‐east (Fig. 5.5: Jones 1980; updated in Morris 2006, figs. 9.13–14, 9.19). Drawing conclusions from such distributional patterns is fraught with danger, although it is noticeable that recent finds (e.g. Lavender 1998b; Reidy and Maynard 2000) confirm this pattern: the lack of distinctively ‘Germanic’ settlement in central Essex, first noted by Wheeler (1935), still holds true. Morris (2005, fig. 14) has similarly shown that this coastal bias is genuine by plotting the location of Early Saxon Grubenhäuser and burials against a background showing the density of archaeological activity between 1990 and 2001 (Fig. 5.5). The results are startling, as while concentrations of Early Saxon sites in the Thurrock, Southend, and Blackwater estuary areas do correspond to areas that have seen large amounts of archaeological investigation, other areas that have seen as much or even more survey and excavation such as Chelmsford, Greater London, and the Lea Valley have little or no evidence for these distinctively Germanic sites (with the exception of a small cluster of findspots at Harlow: Tyler 1996, 108, fig. 1; Germany 2008). ‘Early Saxon’ burials also concentrate in the far north‐west of Essex, and in the far south and east, a pattern confirmed by recent discoveries (Tyler 1996; Ennis 2005; Baker 2006b). Baker (2006a) also notes the broadly coastal distribution of hām place names, which are thought to relate to the earlier phases of Anglo‐Saxon migration and colonization. (p.151)
Beyond the Edge of the Excavation: Fieldwalking in the Wider Landscape
The results of excavations on settlement and burial sites have been summarized very briefly above, as this material is relatively well known (e.g. Tyers 1996; Baker 2006b; Morris 2006). What has received less attention is the wider landscape, which forms the core of this study. The analysis of the results of fieldwalking surveys allows the trends seen in these county‐wide distribution maps to be explored in far greater detail. On the Boulder Clay plateau of north‐west Essex, near Saffron Walden, Williamson (1984; 1986) found 1.3 late Romano‐British settlements per square kilometre. These were mostly located on the lighter soils of the valley sides but also on the heavier soils of the interfluvial areas, although pottery assemblages from the latter suggest that they were of lower status or shorter‐lived. Relatively dense manure scatters around settlements on the interfluvial areas, however, suggest extensive cultivation, with more lightly manured ‘outfields’ beyond, and relatively little woodland. In the early medieval period settlement was restricted to the lighter soils of the valley sides, with many Romano‐British sites there still occupied into the tenth to twelfth centuries, when Saxon‐Norman pottery was in use, suggesting there (p.152) was no major dislocation of the settlement pattern in the first millennium AD (as described later, this is a crucial difference from northern East Anglia where around the eighth century there was a marked nucleation of settlement into villages that went on to acquire parish churches). Other large‐scale fieldwalking surveys on the Boulder Clay plateau have been carried out at Stanstead Airport and its associated infrastructure (including the A120 trunk road). A light scatter of Romano‐British settlements was uncovered and although none was associated with extensive ditched field systems—suggesting large areas of pasture—the evidence of a corn drier and crop‐processing debris at Duckend Farm indicates there was some arable cultivation (Havis and Brooks 2004a; 2004b, 536–7). Far fewer sites yielded fifth‐ to seventh‐century pottery, and bearing in mind the scarcity and very friable nature of the pottery in this period, this may also suggest that the heavier clayland soils saw a decline in settlement (Havis and Brooks 2004b, 341, 346–7).
A review of fieldwalking projects by the Essex County Council Archaeological Unit in Essex reveals that across 1,865 ha examined there was an average density of one Romano‐British site per 38 ha, compared to one ‘Saxon’ period site per 266 ha. The latter figure in part reflects the poor preservation of what is generally very friable early medieval pottery: in some cases sites that when fieldwalked produced a single sherd were found to have clear evidence for occupation when excavated (Medlycott and Germany 1994, 17; Medlycott 2005). Subsequent fieldwalking projects have done little to change this picture, with fairly widespread Romano‐British settlement and manuring, and a reduction in the number of sites producing early medieval material (e.g. Lavender 1997; 2004; Guttman 2000; Medlycott 2000; Brown and Germany 2002; Foreman and Maynard 2002). At Crondon Park near Stock, for example, of three Romano‐British settlements within the survey area just one produced ‘Early Saxon’ sherds (Germany 2001).
Excavations are also increasingly confirming that many Romano‐British settlements were still occupied into the fifth and sixth centuries. In many cases the evidence is ephemeral, such as occasional features containing ‘Early Saxon’ pottery (e.g. Coggeshall: Isserlin 1995, 96; Great Waltham: Tyler and Wickenden 1996; Ship Lane, Aveley: Foreman and Mynard 2002), or simply a scatter of residual artefacts (e.g. New Source Works in Castle Hedingham: Lavender 1996, 22). At North Shoebury (Fig. 5.6) and Buildings Farm in Great Dunmow, a small number of fifth‐century sherds were recovered from the upper fills of late Romano‐British ditches suggesting that they remained open and were still being manured (Wymer and Brown 1995, 46; Lavender 1997, 81). In other cases it is less clear‐cut whether there was actually continuity in occupation, such as the Grubenhaus at Chadwell St Mary which lies close to a Roman villa but was associated with only sixth‐century pottery (Lavender 1998b). While some Romano‐British and early medieval settlements were long‐lived, others were occupied relatively briefly, suggesting a landscape that (p.153) was constantly evolving (the same is true of the evidence from field systems: see below). At Frog Hall Farm in Fingringhoe, for example, a scatter of sixth‐ to seventh‐century pottery was unrelated to any Romano‐British or later settlement (Brooks 2002).
The Evolution of Fieldscapes in Southern East Anglia
So far the discussion has been largely about settlement, with a picture emerging of a dispersed Romano‐British pattern that focused on the river valleys but extended up onto the interfluvial areas, with a contraction of the areas occupied in the early medieval period but no widespread landscape desertion. Another aspect of the countryside that is crucial to our understanding of the origins of the historic landscape is that of the field systems, although this has to date been a rather neglected topic. The lighter soils (ideal for cropmarks) and extensive development‐led excavations particularly in the east and south of Essex mean, however, that there is now a large corpus of data, although unfortunately there has been less work on the Boulder Clays of south‐west Suffolk (Carr 1991: a trawl through the annual ‘Archaeology in Suffolk’ summaries in the Proceedings of the Suffolk Institute of Archaeology and History shows that the bias in fieldwork in favour of the north and east of the county remains). A review of this evidence across Essex is, for the first time, showing some fascinating trends.
The Fluidity of Iron Age/Romano‐British Field Systems
One consistent picture that is emerging from large‐scale excavations is that late prehistoric and Romano‐British field systems in southern East Anglia were usually relatively localized and short‐lived, being maintained for a few centuries, abandoned, and then replaced by a new enclosure system often on a different orientation. One of the clearest examples is at North Shoebury (Fig. 5.6) where a rectangular enclosure and fragmentary traces of a ditched field system were laid out on a NNW–SSE orientation in the Middle Bronze Age (c.1500–1000 BC), being replaced in the Late Bronze Age (c.1000–600 BC) by a field system oriented NW–SE that continued in use into the Early Iron Age (until c.300 BC). This area was then abandoned and a new north–south oriented field system laid out to the west in what had presumably been an area of open pasture. The orientation of this Middle–Late Iron Age system was maintained into the Roman period when the area of fields expanded eastwards, although in the third century there was a further remodelling. Several late Roman ditches must have remained open for a while, as they contained sherds of ‘Early Saxon’ pottery, while a small fifth‐century cremation and inhumation cemetery also indicates occupation in the area. Other examples of field systems in Essex that underwent frequent remodelling include Buildings Farm near Great Dunmow (Lavender 1997), Chigborough (Wallis and Waughman 1998, 76–98), (p.154) (p.155)
Around the third century AD, Going (1996, 101) has noted a general tendency for small enclosures to be replaced by larger fields, and this has since been seen at Buildings Farm in Great Dunmow (Lavender 1997), Founton Hall near Little Oakley (Barford 2002), Mill Hill in Braintree (Humphrey 2002), Monument Borrow Pit in Rawreth (Dale et al. 2005), Ship Lane in Aveley (Foreman and Maynard 2002), and Strood Hall (near Great Dunmow) and the Rayne Roundabout (Timby et al. 2007). It is also worth stressing that there clearly was not a continuous fieldscape in Roman Essex, with a number of large‐scale excavations failing to produce evidence for field ditches of this date (e.g. the eastern area excavated at North Shoebury: Fig. 5.6). Along the new A130, for example, Iron Age field systems were present at Sandon Brook, Windmill Hill, Shotgate Farm, Doublegate Lane, and Dollymans Farm, but all went out of use, and there were no Romano‐British replacements (Dale et al. 2005). The same was seen at Stifford Clays in Thurrock during the construction of the new A13 (Wilkinson 1988, 23), Slough House Farm to the north of the Blackwater estuary (Wallis and Waughman 1998, 41), and St Osyth near Clacton‐on‐Sea (Germany 2003b). At Great Holts Farm, part of an extensive mid Roman enclosure system has a very clear boundary (ditch 198) with what was presumably open pasture to the south (Fig. 5.7; Germany 2003a).
Fieldscapes of Continuity
There is, however, a less consistent picture emerging with regard to the relationship between Romano‐British field systems and the historic landscape. Great Holts Farm in Boreham typifies the situation, in that there appears to be both continuity and discontinuity within the same site (Fig. 5.7; Germany 2003a). To the south of what was an extensive area excavation, the Roman field system was presumably abandoned, as the historic landscape overlies it unconformably. In contrast, to the north, elements of the late Roman enclosure system are on the same orientation as the historic landscape, with one boundary having been re‐dug in the medieval period (ditches 63 and 59). A few sherds of ‘Early Saxon’ pottery, including some with Schlickung‐treated surfaces, suggest occupation into the fifth century, and although the sixth to ninth centuries are not represented within the excavation, the tendency for early medieval settlements to shift around means that occupation of this period could easily lie in adjacent fields. Building 440 is dated to the tenth to thirteenth centuries while Great Holts Farm immediately to the east is documented from the late thirteenth century (Reaney 1935, 241).
Potential continuity in field systems is also seen elsewhere. At the New Source Works in Castle Hedingham and Buildings Farm near Great Dunmow the latest Roman‐period field systems are on the same orientation as the historic landscape (Lavender 1996; 1997). The same similarity in orientation between late Romano‐British and medieval features is also seen at East View Close in Radwinter (Havis 2001), Palmer's School in Grays (Rodwell 1983), Shillingstone Field in Great Sampford (Garwood 1998), William Edwards School in Grays (Lavender 1998a), Stebbing Green (Bedwin and Bedwin 1999), (p.158) and Wendens Ambo (Hodder 1982). A mid Roman ditch at Stebbingford must have survived as an earthwork into the twelfth century, as it was incorporated into the layout of a farmstead (Medlycott 1996, 100). At Bishops Park College in Clacton‐on‐Sea the Roman field system was re‐excavated in the medieval period, implying either continuity in its use, or, if abandoned, that it survived as a set of earthworks (Letch 2005). At Founton Hall near Little Oakley a late Roman field system was on the same orientation as the historic landscape, with fourth‐century ditches containing fifth‐ to sixth‐century pottery in their upper fills; significantly, the underlying Iron Age/early Romano‐British landscape was on a different orientation, giving a date around the third century for the creation of the framework of today's historic landscape (Barford 2002). At Chignall St James, the enclosure surrounding a palatial Roman villa must have survived into the medieval period in order to influence the course taken by a trackway, while a Roman or later field system to the south perpetuated the line of a late Roman enclosure system (Clarke 1998, 65). In south‐west Suffolk, the latest Romano‐British landscape at Preston St Mary appears to share alignments with the historic landscape (Martin et al. 1996, 482).
The ‘Coaxial’ Landscapes
There have been various claims that planned landscapes of Roman or pre‐Roman date survived in the landscape of ‘greater East Anglia’. Some have tried to identify Roman centuriation, although the modest ideas of Haverfield (1920), based on two straight, parallel roads west of Colchester, and the more ambitious claims of Coles (1939) that such planning can be identified across north and west Essex, have been rightly rejected (Dilke 1971, 191–5). In a number of places across East Anglia it has been suggested that the historic landscape has a ‘coaxial’ layout based around a series of sinuous but roughly parallel boundaries that appear to be prehistoric in origin as they are overlain by Roman roads (e.g. Little Waltham in central Essex: Fig. 5.8; for general discussions see Williamson 1987; 1998; Rippon 1991; Hinton 1997; Martin 1999a, 52–8). There has been much discussion of the date and character of these landscapes, with Wilkinson (1988) and Rippon (1991) suggesting that some of the southern Essex examples are medieval in date, and Hinton (1997) questioning the date and character of the East Anglian examples (and see Williamson 1998 for a response). In the most recent review, Martin and Satchell (forthcoming) suggest that rather than extensive, planned landscapes, these coaxial systems actually consist of a series of smaller blocks, and that while those around Scole do indeed appear to incorporate late prehistoric boundaries, those around South Elmham St Michael are probably early medieval in date.
Although just outside this region, work at Broxbourne and Wormley, in eastern Hertfordshire, has used survey and excavation to establish a Romano‐British date for a very similar coaxial landscape (Bryant et al. 2005), but unfortunately the dating evidence from Essex is at best fragmentary. In the Slade Valley, at Saffron Walden, in the far north‐west of the county, Bassett (1982, 7–9, 14, 30) suggests that a Late Iron Age ditch was on the same orientation as the coaxial landscape there, and that a minor Roman settlement was laid around one of its major elements. The place name Walden (Waela‐denu) means ‘valley of the Britons/serfs’, which implies the survival of a native population in the area which lends support to the possibility that a Romano‐British landscape of fields and roads could have remained in use throughout the early medieval period (Reaney 1935, 537; Bassett 1982, 10). A cemetery of perhaps 200 inhumations excavated in 1830 and 1876 appears to date from the Romano‐British and early medieval periods (Bassett 1982, 9–12). Some burials were associated with third‐ to fourth‐century grave goods and oriented north to south, although a majority were unaccompanied and oriented east to west. None was associated with Anglo‐Saxon artefacts although there was a scatter of sand and grass‐tempered pottery (broadly (p.160) fifth to ninth century) from the site; distinctively ‘Anglo‐Saxon’ decorated pottery was noticeably absent. A Scandinavian necklace from one of the graves probably dates from the ninth century (Bassett 1982, 9–15, 80). Overall, although based on antiquarian excavations, Saffron Walden suggests the survival of a native Romano‐British community at the very margins of the area around Great Chesterford and southern Cambridgeshire that saw extensive Anglo‐Saxon immigration, supporting the evidence for continuity from Williamson's nearby fieldwalking surveys on the Boulder Clays to the south (see above).
Around Braintree, Chelmsford, and Witham other coaxial landscapes appear to be cut by Roman roads (Drury 1976; Drury and Rodwell 1980, 62; Rodwell 1993, 58–9), as is also the case at Little Waltham in the Chelmer Valley to the north (Fig. 5.8; Drury 1978, 134–6). Here, along with the rather more fragmentary and less convincing system at Rivenhall in the Cressing Brook Valley, excavations have established that elements of the historic landscape date back to the Late Iron Age (Rodwell and Rodwell 1986, 68, figs. 8, 10, and 50). Further tentative support for the hypothesis that these coaxial landscapes are Late Iron Age in origin is the way that they are best preserved on the lighter soils of river valleys, which surveys and excavations across Essex have shown were the most intensively occupied, including in the early medieval period, as opposed to the heavier soils of the interfluvial areas that tended to have been less densely populated or even abandoned in the early medieval period.
In the south of Essex there are also areas of historic landscape with apparently planned layouts but which differ significantly from the long, sinuous, coaxial systems in having a more rectilinear layout (Rippon 1991; Short 2006, 113, fig. 5.6). To the south of Wickford, the historic landscape is structured around a series of long, straight boundaries that appear to have been laid out according to Roman surveying methods (Fig. 5.9). One of these boundaries proved to be a Roman road upon excavation (Rodwell 1966), although Rippon (1991, 57) was cautious in ascribing a date to the landscape suggesting that ‘maybe this is very fragmented centuriation, or more likely, an example of “a land assignation made in multiples of actus by someone with at least a vague notion of Roman surveying” (Dilke 1971:193, discussing Ripe in Sussex)’. Recent excavations within this landscape at Monument Borrow Pit, along the line of the new A130 to the east of Great Fanton Hall in North Benfleet, however, have revealed a field system dating to the Late Iron Age to early Roman period, which included a trackway whose line was perpetuated in the mid to late Roman period when a ditch was dug along its length: all of these excavated features conform to the orientation of the historic landscape, suggesting a late pre‐Roman date for its origin (Fig. 5.10; Dale et al. 2005).
Further west, but also on the London Clay, the historic landscape also has a rectilinear layout, but rather than a single planned network of roads and fields, this landscape can be broken down into a series of discrete blocks laid out around roughly parallel boundaries that may have originated as trackways linking coastal districts with the interior (Rippon 1991), the same process that may have occurred south of Dartmoor at Bittaford (Fig. 4.8). In various locations these areas of landscape are clearly medieval in date, as they overlie Romano‐British enclosure systems that are on a different alignment. To the east, in the Rochford Hundred near Southend‐on‐Sea, another carefully planned landscape has a radial pattern that is on a different orientation from, and so clearly post‐dates, the late Romano‐British field system, with ‘Early Saxon’ pottery in its upper fills, excavated at North Shoebury (Figs. 5.6, 5.9, and 5.11). Further Roman‐British ditches on a different orientation from the historic landscape have been excavated at Great Wakering, to the north (Rippon 1991), and South Shoebury to the south (Mattinson 2005).1
A third area of planned landscape in southern Essex covers the Dengie peninsula (Drury and Rodwell 1978). This is a remarkable system of almost square blocks of fields and roads that clearly pre‐dates the late eleventh‐ or early twelfth‐century church at Asheldham which overlay one of the major axial elements of the landscape. A ditch perpendicular to this axial boundary had a series of ten recuts, most of which were aceramic but the last containing (p.163)
Fieldscapes of Discontinuity
In contrast to this evidence for the survival of Romano‐British field systems in some places, in other cases they appear to have been abandoned and were on a different orientation from the medieval landscape. At Gun Hill in West Tilbury, Woodham Walter, Woodside Industrial Park in Birchanger, Howells Farm near Heybridge, and Elm Park in Ardleigh, the Romano‐British field systems were abandoned in the second century (Drury and Rodwell 1973; Buckley and Hedges 1987; Medlycott 1994; Wallis and Waughman 1998, 117; Brooks 2001). At Lofts Farm, Slough House Farm, and Chigborough, all to the north of the Blackwater estuary, an extensive cropmark complex of trackways and field systems, dated through excavation to the Roman period, was on a different orientation from a historic landscape that certainly existed by the twelfth century, when an enclosure was dug conforming to the latter's orientation (Maldon Archaeological Group n.d.; Wallis and Waughman 1998, 41–53, 76–98). These sites are all on light sands and gravel, but the same is seen on heavier clay soils. At Coggeshall (Isserlin 1995), Curry Hill in Rettendon (Dale et al. 2005), Downhouse Farm (Dale et al. 2005), Rayne Roundabout and Strood Hall along the A120 between Stanstead Airport and Braintree (Timby et al. 2007), Ivy Chimneys in Witham (Turner 1999), and Newman's End in Matching (Guttman 2000), late Roman ditches were similarly on a different orientation from the historic landscape. At Preston St Mary, in southern Suffolk, the medieval field system also lies unconformably over the late Romano‐British landscape (Martin et al. 1996).
At the Rayne Bypass (Smoothy 1989) late Roman ditches were also on a different orientation from the overlying medieval field systems, and these late Roman ditches had sherds of ‘Early Saxon’ pottery in their upper fills. This was also seen at North Shoebury (see above) where there was no evidence from the sixth to tenth centuries, although around the eleventh or twelfth centuries a rectangular enclosure was constructed to the east of the parish church (the present structure of which dates from the early thirteenth century: RCHME 1923, 101). This enclosure, and a number of other nearby features, (p.165)
A number of key conclusions have emerged from this overview of the development of late prehistoric and Romano‐British field systems in southern East Anglia, and their relationship to those of the medieval period. First, field systems in many areas were relatively discrete and surrounded by areas of unenclosed land, especially by the fourth century: there was no continuous fieldscape until the creation of today's historic landscape. Secondly, in the later prehistoric and Roman periods field systems were relatively unstable, with cycles of creation, abandonment, and re‐creation every few hundred years: the roughly ten centuries of comparative stability since the creation of (p.166) the framework of today's historic landscape is quite unprecedented. Thirdly, there are many examples of Romano‐British field systems that went out of use (although often after the fifth century) and are unrelated to the medieval landscape that lies unconformably on top, implying a discontinuity in landscape exploitation, although these tend to occur in what can be regarded as physically marginal areas in terms of arable‐based agriculture, notably very light soils of the Thames and Blackwater gravel terraces, and heavy soils of the Boulder Clay interfluvial plateaux (confirming the evidence from fieldwalking surveys that these areas were abandoned: Williamson 1986). Even here, however, palaeoenvironmental evidence does not indicate a woodland regeneration, suggesting at least continued grazing (Havis and Brooks 2004a; 2004b). Fourthly, however, there are also many cases of Romano‐British and medieval field systems that appear to show continuity. Sometimes elements of both continuity and discontinuity are seen within a single site such as Great Holts (see above) as well as cropmark complexes such as Ardleigh (Fig. 5.13; Brown 1999, 180–4, fig. 4) and Lofts Farm, near Maldon (Wallis and Waughman 1998, fig. 130). Both on individual sites, and across Essex as a whole, the overall picture appears to be one of partial survival of Romano‐British landscapes into the medieval period, and partial discontinuity that suggests a decrease in the intensity with which the landscape was exploited, but not its abandonment. There is also some patterning in the topographical location of these landscapes of discontinuity which tends to occur most frequently in areas of soil that were less favourable for arable‐based farming, although this was not always the case: the early medieval site at Mucking, for example, is on a gravel terrace some distance from areas of medieval settlement, whereas North Shoebury lies on the fertile brickearths of south‐east Essex that are amongst the most fertile soils in the county. What both these examples of discontinuity have in common, however, is that the Romano‐British/early medieval landscapes were replaced by planned field systems: the open fields of Mucking (laid out sometime after the early eighth century), and radial block fields, apparently always held in severalty, of the Rochford Hundred (datable to the sixth to tenth centuries).
This leaves one crucial question: does the close relationship between some Romano‐British and medieval field systems imply that they have remained in constant use, or might the earlier fields have been abandoned but preserved as earthworks that were then reused? Observations by the author, and a sequence of cartographic sources showing modern field systems that have gone out of use on the Benfleet Downs in south‐east Essex, show that previously cleared land is enveloped with brambles within a few years, closely followed by the invasion of hawthorn scrub, making the area impassable within around ten years. After about thirty years the area is covered in light woodland. The question is, if medieval farmers had then cleared such woodland and scrub, would significant earthworks from the earlier field systems have survived, and if so (p.167)
(p.168) The Palaeoenvironmental Evidence
So far discussion has focused on the sometimes close physical relationship between the Romano‐British and medieval landscapes, but a crucial question that has emerged is whether this necessarily means that the landscape was occupied and used continuously. Even into the 1990s, the traditional view that the early medieval period saw an extensive woodland regeneration can be found in the literature for the history and archaeology of this region (e.g. Brooks 1992, 49), while Oliver Rackham (1976, 57) even suggested that ‘village and hamlet names ending in ‐ley are strikingly concentrated on the Huntingdon and Suffolk borders [of Cambridgeshire], suggesting that here the early Anglo‐Saxons found and colonized tracts of wildwood that the Romans had left’; his figure 9 shows a similar cluster of ‐ley names on the Cambridgeshire/Hertfordshire/Essex border which by implication were similarly carved out of wildwood. Archaeological surveys in a number of places have also revealed cases of Roman material from areas that in the medieval period were wooded, but these tend to be on areas of poorly drained clays on the interfluvial plateaux (e.g. the Boulder Clays of north‐west Essex: Williamson 1984, 228; the Rayleigh Hills in south‐east Essex: Rippon 1999, 23).
Fortunately there is now increasing palaeoenvironmental evidence with which to explore the extent of any post‐Roman woodland regeneration. A national overview of tree ring dating does suggest that a number of timbers used in medieval structures started to grow around the fifth century and several examples are from Essex (Barking Abbey, Mersea Island, and Slough House Farm: Tyers et al. 1994, 18–21, fig. 3.5). This does not, however, necessarily suggest the widespread regeneration of woodland over what had been agricultural land, as these timbers may simply be from hedgerows that were no longer being laid, or from former coppiced woodland that was left unmanaged. Palaeoenvironmental sequences provide a better indication of land‐use history. The only possible evidence for post‐Roman woodland regeneration is from the Mar Dyke in Thurrock, where there is a slight increase in tree pollen at the very top of a sequence that, although not radiocarbon dated, may be early medieval (Wilkinson 1988, 109–14, fig. 98).
The majority of the evidence, however, suggests that the early medieval landscape in Essex remained largely open with no woodland regeneration. In the Crouch estuary, a peat layer with radiocarbon dates from its upper and lower surfaces of 1610+/−70 BP (cal. AD 380–540) and 1380+/−80 BP (cal. AD 604–81) produced very low levels of tree pollen, suggesting an open landscape (Crouch site 9: Wilkinson and Murphy 1995, 49). At Chigborough and Slough House Farm, north of the Blackwater estuary, samples from a series of seventh‐century features show a less wooded (in fact almost treeless) landscape compared to the Late Iron Age/early Roman period, and (p.169) with far greater cereal cultivation (Wallis and Waughman 1998, 172–204). At Stanstead Airport, on the Boulder Clay plateau in north‐west Essex, pollen and plant macrofossils from the ‘BRS’ palaeochannel suggest a largely open landscape in the fifth to sixth centuries with some cereal cultivation in the vicinity; the levels of microscopic charcoal similarly indicate human activity. Around the late sixth or seventh centuries the palaeochannel started to dry out and there was an increase in cereal cultivation (Havis and Brooks 2004b, 350–4). At the Sandon Culvert site, in the mid Essex Chelmer Valley, plant macrofossils from a sequence of channel fill that accumulated between the Roman period (1770+/−70 BP, cal. AD 117–413) and around the twelfth century (869+/−70BP, cal. AD 1029–1265) suggests an open landscape throughout, with relatively little woodland (Murphy 1994, 25–6). None of these sequences shows a significant change in land‐use around the eighth century.
Historians have often regarded Essex as a relatively well‐wooded county in the eleventh century, contributing to a view that its landscape of dispersed settlement was the result of relatively late colonization (e.g. Roden 1973). Oliver Rackham (1980) suggests that 20 per cent of Essex was wooded in 1086, compared to 15 per cent in the rest of England, although he acknowledges that the measure used in the Domesday survey of Essex—‘woodland for x swine’—is rather imprecise. Locating this woodland is also especially difficult due to the practice of enclaving, whereby some of a manor's resources was held in detached parcels some distance away. Round (1903) has shown how this explains another curiosity of the Essex Domesday—the ‘pasture for x sheep’, which is recorded for some inland manors but actually lay in the coastal marshes—while Rackham (1986b, fig. 14) and Rippon (2000a, fig. 69) have shown the extent to which medieval woodland was similarly enclaved. Bryn Morris's (2005, fig. 16; 2006, figs. 9.8–10) careful mapping of woodland‐indicative place names against the topography also shows that it was not evenly distributed across the landscape as some earlier, often smaller‐scale, distribution maps suggest (Fig. 5.14; cf. Darby 1952, fig. 61; Roberts and Wrathmell 2000, fig. 24).
The Value of Place‐Name Evidence
Baker (2006b) has recently reviewed the place‐name evidence in Essex, Hertfordshire, Middlesex, Buckinghamshire, and southern Cambridgeshire for the survival of a native British population, and the extent of immigration from the continent. It is claimed that there are place names containing pre‐English elements, although they occur both in central Essex and Hertfordshire, where archaeological evidence suggests little fifth‐ or sixth‐century Germanic settlement, and in the far south and east of Essex where the opposite is true (Baker 2006b, fig. 4.1). The interpretation of these place names also seems far from (p.170)
(p.171) Where are the ‘Middle Saxon’ Settlements?
So far the picture that is emerging is one of gradual change and evolution in the Essex landscape in the Roman and earliest medieval (fifth to seventh centuries) periods, with continuity in some areas, discontinuity in others, but no widespread transformation of the landscape and certainly no widespread woodland regeneration. Throughout this period settlements and field systems were established, modified, and abandoned, with some being longer lasting than others. The landscape appears to have been in a constant but gradual state of evolution, albeit with certain periods that may have experienced greater change than others, such as the mid Roman period when some landscapes may have been restructured, and the late fourth or fifth century, when there appears to have been a contraction of settlement from areas such as the heavy interfluvial clays of the Boulder Clay plateau.
It has been suggested that across southern England significant numbers of excavated ‘Early Saxon’ settlements were abandoned around the seventh century (e.g. Mucking in Essex, West Stow in Suffolk, Bishopstone in Sussex, Chalton in Hampshire), which led to the model of a ‘Middle Saxon shuffle’ whereby Early Saxon sites on lighter soils were abandoned in favour of richer soils in the valleys around the late seventh century (Arnold and Wardle 1981; Moreland 2000, 86–7). One problem with this idea is that most of the major excavations of ‘Early Saxon’ settlements are in locations that extensive field survey is suggesting are not typical of that period, such as the high gravel terrace at Mucking. Hamerow (1991; 2002, 121–4) has also argued that, as many settlements have not been completely investigated, their final phases may lie beyond the edge of the excavations, and Mucking can also be reinterpreted in this way (Fig. 5.12). The eastwards migration of the settlement was clearly demonstrated by Hamerow (1993), and if we place Mucking in an even wider context we see that from the fifth to the seventh centuries the focus of occupation shifted at least 1.2 km south‐west to north‐east, this distance being calculated from the centre of fifth‐century occupation that includes the Linford Quarry site to the west of the main Mucking excavation (Barton 1962), as far north as the ‘North Ring’ site excavated in 1978 separately from the main Mucking campaign where the occupation is dated to the late seventh century (Bond 1988, 20, 45–51). Just c.900 m further west lies the parish church, the earliest fabric of which is (?)twelfth or thirteenth century (RCHME 1923, 94). This raises the possibility that rather than there being a sudden dislocation of settlement, whereby it shifted from the gravel terrace to the lower‐lying site occupied by the parish church—the ‘Middle Saxon shuffle’ model—it actually ended up there by continued, gradual, migration. Also significant is that a few sherds of Ipswich Ware were found from the far east of the site which if (p.172) Blinkhorn's (1999) redating is correct takes the occupation of the settlement into the eighth century (and see Hamerow 1993, 22). The discovery of two early eighth‐century sceattas from the area immediately beyond the edge of the excavations (the precise location is not known) supports the idea that the settlement at Mucking continued to be occupied until at least that date but continued to shift its location (Helena Hamerow and Michael Metcalf pers. comm.). Another significant feature of this site is that the early medieval settlement appears to have been replaced by an open field system whose strips and furlong boundaries were still preserved within the historic landscape when it was first mapped in 1846 (Fig. 5.12; ERO D/P 108/27/2; Clark 1993, 22). After several centuries of migration, maybe there was a transformation of the landscape associated with the creation of the open fields sometime after the early eighth century. The open fields did not last long, however, as Walton's Hall and Sutton's Farm, which appear to lie in the southern part of the now enclosed open field, are recorded as early as 1199 and 1220 respectively (Reaney 1935, 164).
So where are the eighth‐century settlements in Essex? There have been a number of important excavations of what in this region are called ‘Middle Saxon’ sites, although these mostly relate to the higher echelons of society, notably the high‐status site at Wicken Bonhunt (Wade 1980) that Rippon (1996b, 121) has suggested was a probable villa regalis, Christian communities at Barking (MacGowan 1987; Redknap 1991; Hull 2002), Nazingbury (Huggins 1978; Bascombe 1987), and Waltham Abbey (Huggins 1988), and a possible eighth‐ to early ninth‐century coastal trading site near Barking Abbey (Hull 2002). Unfortunately, these sites tell us little about the wider rural landscape. A re‐examination of a number of ‘Early Saxon’ sites, however, suggests that there may have been a closer relationship between the Romano‐British and medieval landscapes than was previously thought.
In a number of cases, ‘Early Saxon’ settlements and cemeteries have been discovered during the excavation of earlier enclosures that were visible as cropmarks.3 Understandably, these mostly rescue excavations focused on major cropmark complexes, but, if the early medieval settlements had migrated to the extent of that at Mucking, then they would soon have drifted beyond the areas that were later to be excavated. At the Orsett Cock, for example, three Grubenhäuser to the east of the cropmark enclosure were associated with mid‐fifth‐ to mid‐sixth‐century pottery, including distinctively early Schlickung‐treated ware (Milton 1987, 30–1), whereas three Grubenhäuser to the west, within the old enclosure, were sixth century in date (Carter 1998, 102). Did the settlement continue to drift further west, beyond the edge of the excavations? (p.173) And what are the origins of Barrington's Farm immediately to the north, first documented in 1548, and possibly named after the Barrington family recorded in Orsett in 1482 (Reaney 1935, 166)? Another example of early medieval settlement migration potentially ending up at an extant farmstead could be Little Oakley, where small amounts of eighth‐ or ninth‐century ‘Middle Saxon’ pottery were found halfway between a Romano‐British to late fifth‐century settlement and the Domesday manor of Founton Hall (Barford 2002, 164).
The problem that ‘Middle Saxon’ settlements may have drifted away from archaeologically very visible sites, such as enclosures and villas, and so have simply not been identified, is compounded by the very limited material culture that appears to have been used on most lower‐status sites in this period and the ephemeral traces that the timber buildings have left. On the gravel terraces north of the Blackwater estuary, for example, what has been described as a ‘boat‐shaped’ building (structure 38) constructed with earth‐fast posts was not associated with any material culture, although parallels for the building are most common from the eighth to tenth centuries (Wallis and Waughman 1998, 98, 106–8). A group of eight seventh‐ to eighth‐century loom weights have been found as packing in a post hole of a rectangular building just 110 m to the north (Tyler 1986). Another aceramic rectangular timber building built from earth‐fast posts has been excavated at Takeley, but here a radiocarbon date of 1245+/−35 BP (cal. AD 670–880) establishes a date around the eighth century (Timby et al. 2007, 152–6). It lies 100 m to the north of the parish church and on a different orientation from the historic landscape. At Bishops Park College in Clacton‐on‐Sea, midden deposits containing mostly seventh‐ to eighth‐century grass‐tempered pottery and a single sherd of Ipswich Ware were found in the slumped upper fill of a largely silted‐up Late Bronze Age ditch, but the only other features certainly dating to this period were a small number of pits. A series of post holes, however, formed the plan of what is described as a ‘bow‐sided’ building for which the only dating was a single fragment of (? residual) Roman brick but the plan is in keeping with an early medieval date (Letch 2005). Another ‘bow‐sided’ building constructed of earth‐fast posts has also been excavated at Downhouse Farm in West Hanningfield where the few sherds of fifth‐ to sixth‐century pottery were ‘insufficient to provide conclusive dating’ (Dale et al. 2005). Early to Middle Saxon pottery has also been recovered from Roxwell Quarry, 1.5 km south of the Chignall St James villa (Bennett 2000, 220). Along with the early eighth‐century occupation now recognized at Mucking it appears that, unlike in northern East Anglia (see below), ‘Middle Saxon’ occupation in the south of our region was found scattered across the landscape rather than nucleated into villages that went on to become parish centres.
(p.174) The Origins of the Later Medieval Settlement Pattern
So how old is the historic landscape in southern East Anglia? Documentary sources, place‐name evidence, and dendrochronological dating of standing structures all show that the highly characteristic dispersed settlement pattern of later medieval Essex and south‐west Suffolk was in existence by the thirteenth to fourteenth centuries: just as we have seen on the National Trust's Holnicote Estate near Porlock in West Somerset (Fig. 2.7), standing medieval buildings in Essex are spread across the landscape rather than only being found in a small number of discrete villages (e.g. Stenning 1996; 2003; Ryan 2000; Andrews 2004). Fieldwalking and major development‐led excavations are also confirming that there are no deserted medieval villages hidden in this landscape: the medieval settlements being found are all isolated farmsteads scattered throughout ditched field systems. At Roxwell Quarry in Chignall St James, for example, on the Boulder Clay of central Essex, excavations have revealed a small farmstead of the tenth to twelfth centuries whose enclosure ditches conformed to the historic landscape (Brooks 1992). At Stebbingford and Blatches, in the Stebbing Valley also on the Boulder Clay, isolated farmsteads have similarly been excavated dating to the mid twelfth to mid fourteenth and mid thirteenth to fourteenth centuries respectively (Medlycott 1996; Timby et al. 2007, 161–6). At Slough House Farm and Chigborough, on the gravel terraces north of the Blackwater estuary, two small farmsteads date from the twelfth to early thirteenth and twelfth to fourteenth centuries respectively (Wallis and Waughman 1998, 53, 98). On the London Clays in southern Essex work along the line of the new A130 revealed small areas of medieval occupation at Ashdale Bridge near Battlesbridge (eleventh to sixteenth century), Downhouse Farm near East Hanningfield (eleventh to thirteenth century), Old Barn Lane near West Hanningfield (mid thirteenth to mid fourteenth century), and Dollymans Farm near Rawreth (eleventh to fourteenth century) (Dale et al. 2005). None of these developments—which have revealed large numbers of deserted isolated farmsteads—has revealed a deserted medieval village.
So how far back can we trace the origins of this later medieval settlement pattern? While it is not uncommon to find ‘Late Saxon’ material from still‐occupied medieval settlements (Rippon 1996b, 124) there is growing evidence, often from small‐scale work, that at least some were also occupied before the tenth century. The ‘Middle Saxon’ occupation at Asheldham is well known (Drury and Rodwell 1978, 137). At Great Wakering, late Roman and early medieval occupation associated with ‘Early Saxon’ and ‘Middle Saxon’ pottery has been recorded immediately east of the medieval parish church (Bennett 2001, 260; Medlycott 2003). At Chadwell St Mary a sixth‐century Grubenhäus has been excavated just 120 m from the church (Lavender 1998), while in Chipping Ongar ‘Saxon’ pottery was found the same distance from its parish (p.175) church (Gilman and Bennett 1995, 242). ‘Early Saxon’ sherds have been found next to the church in Tollesbury (Bennett 2002, 410). At Great Waltham late Roman, ‘Early, (?)Middle and Late Saxon’ pottery has been found 300 m from the church (Tyler and Wickenden 1996), while ‘Saxon’ pottery has been found adjacent to St Peter and St Paul's churchyard in St Osyth (Bennett 2000, 214). It is not just parish churches that have evidence for early medieval occupation in their vicinity. Mid Roman to ‘early/mid Saxon’ material was recorded at Thorpe Hall in Thorpe le Soken, while ‘Early Saxon’ sherds have been found at Cuton Hall in Chelmsford close to the excavated Late Saxon settlement at Springfield Lyons (Bennett 2002, 393). Limited excavations adjacent to Heybridge Hall revealed Late Iron Age to early Roman occupation along with ‘Early‐Middle and Late Saxon’ pottery (Bennett 1999, 220). Late Roman and ‘Early Saxon’ occupation has also been recorded immediately south of Downhouse Farm in West Hanningfield (Dale et al. 2005). In south‐west Suffolk, small amounts of residual Iron Age, Romano‐British, and ‘Early and Late Saxon’ pottery from excavations at St Botolph's church in Burton suggest broad continuity in the occupation of that site (Murray 2005).
The overall theme that is emerging from this overview of settlement history is the same as that for field systems: some continuity, some discontinuity, but an overall picture of gradual evolution rather than sudden transformation. As population grew, and more and more land was taken into field systems, there must have been an increasing number of fixed elements within the landscape: the roads that linked settlements, the boundaries around increasingly scarce resources such as woodland and common grazing, and indeed parish churches that emerged around the tenth or eleventh centuries (which in some places did form a—but never the only—focus of settlement within a parish). There is no evidence of a major transformation of the countryside in the tenth century, as appears to have been seen in central Somerset (see Chapters 2 and 3), or around the eighth century as we will soon discover happened in the northern part of East Anglia: in Essex and south‐west Suffolk landscape evolution happened at a far slower pace.
Churches and Burial in the Landscape
A further aspect of the origins of the historic landscape is that of its churches. Essex has numerous examples of medieval churches lying on or close to Roman villas, a relationship that has seen much discussion nationally (e.g. Morris and Roxan 1980; Morris 1989; Bell 1998; 2005). Of the 88 certain, probable, and possible villas in Essex, 21 (23 per cent) are overlain by or adjacent to a manorial hall, church, or church–hall complex, and of the nine of these villas that have good dating evidence all were still occupied in the late fourth century (Morris 2006, 242, appendix 9.3). It is possible that this relationship simply represents the desire of the church builders to be close to a source of stone, as (p.176) in Essex, at least, the evidence for timber phases pre‐dating the present stone churches is slim (see below). While it is also possible that the location of these churches may simply have been due to the ‘sense of history’ of these places (Morris and Roxan 1980, 82), we cannot rule out some form of continuity of these locations as central places for the surrounding territories.
So what of the origin of Essex's churches? Archaeological and documentary evidence from across England suggests that burials within minster churches started in the eighth century, as the earlier trend for kings to be buried within churches trickled down to the lower levels of society (Blair 2005, 228–9). It is generally thought, however, that churchyard burial for the whole population started several centuries later, perhaps around the tenth century when the new phenomenon of manorial churches forced people to bring their dead to these central locations, leading to the abandonment of scattered rural cemeteries (Blair 1994, 72–3; Hadley 2000; 2002; Hamerow 2002, 123): ‘the tenth century may have been the first time when English people at large were told where they had to be buried’ (Blair 2005, 463). This timescale for churchyard burial is based largely on documentary sources, such as a law code of Æthelstan of c.930 that refers to burial within ‘a hallowed graveyard’, although it receives some support from archaeological evidence. At well‐known sites such as Wharram Percy, in Yorkshire (Bell and Beresford 1987), and Raunds Furnells in Northamptonshire (Boddington 1996), excavations have shown that churchyard burial did indeed begin around the tenth century, but at Yarnton in Oxfordshire, it appears that people started burying their dead in small family groups near to their homes in the Middle Saxon period before the creation of centralized, communal, churchyards (Hey 2004).
So what of the situation in Essex? There are only nineteen parish churches in Essex with some pre‐Conquest fabric (Taylor and Taylor 1965; Taylor 1978, 1076), and the county's Domesday survey contains just 36 places with churches and/or priests, compared to 345 in Suffolk (Darby 1952, 249–51). Along with other documentary and archaeological evidence, Rodwell and Rodwell (1977, 92) suggest that there is some evidence for 86 pre‐Conquest churches in the county. There is, however, almost no evidence for churches before the tenth or eleventh centuries.4 Rodwell and Rodwell (1986) have claimed that ‘Structure 1’ adjacent to the Roman villa at Rivenhall was a Middle Saxon church, although only the corner of an undiagnostic rectilinear foundation trench was actually excavated (see Reece 1986, Millett 1987, and the discussion of (p.177) Asheldham below, for similar over‐interpretations of much the same evidence; the Rodwells' putative eastern boundary ditch of their ‘Middle Saxon’ cemetery, F.58, also failed to show up in the recent excavations: Clarke 2004, 70, fig. 22). The attribution of ‘Structure 1’ to the ‘Middle Saxon period’ was on the basis that it cut a number of graves, one of which (G.326) is radiocarbon dated to cal. AD 789–980,5 while another grave (G.298) dated cal. AD 1166–1265 cut through it (Rodwell and Rodwell 1993, 104). Further excavations and radiocarbon dating, however, overwhelmingly point to a tenth‐century and later date for this cemetery, while ‘Structure 1’ has been redated to the eleventh or twelfth century (Clarke 2004, 69–70). The Rodwells' ‘Structure 1’ lay 30 m to the north‐west of the medieval parish church, excavations at the east end of which revealed a series of fragmentary traces of east–west oriented gullies that they interpret as a timber church (‘Building 7’). Quite how a church can be reconstructed from this disparate series of gullies, not all of which are parallel, and several of which have butt ends, must be regarded as questionable, and the dating evidence is simply that they post‐date Roman features and pre‐date the fourteenth‐century extension to the chancel. The dating for the first phase of the stone church at Rivenhall is a little firmer, as it post‐dates robbed‐out footings of the Roman villa, containing pottery of the ninth to eleventh century, and is associated with radiocarbon‐dated burials of cal. AD 998–1153 and cal. AD 990–1163 (Rodwell and Rodwell 1986, 90–1; 1993, 104).
Other archaeological evidence for pre‐Conquest churches in southern East Anglia suggests that there is sometimes evidence for structures and/or burials that pre‐date the often twelfth‐century standing structures, although none can be pushed back further than the tenth century. Drury and Rodwell (1978) claimed that excavations around the medieval parish church at Asheldham church revealed a timber church that pre‐dated the present stone structure, although this ‘church’ was in fact simply a gully (F.151) that formed the corner of a rectangular building, adjacent to which were two graves that were sealed by a buried soil containing tenth‐ to eleventh‐century pottery. The gully appeared to date to the tenth or eleventh century but failed to show up in later work (Andrews and Smoothy 1990). At Little Oakley, two post holes and a grave that stratigraphically pre‐dated the twelfth‐century stone church could relate to some timber predecessor, although there is no evidence for their date (Corbishley 1984, 21). At All Saints Cressing, excavations show that the present structure had a stone predecessor, with a possible timber structure beneath that (Hope 1984), while St Mary's West Bergholt also has a range of ephemeral features interpreted as an early timber church but which are of unknown date or character (Turner 1984). Work over the past decade by the Essex County Council Archaeology Section has failed to reveal any other (p.178)
The Later Medieval Field Systems of Southern East Anglia
Finally, we must consider the character of the medieval field systems in Essex. The far north‐west of the county had the greatest amount of open field, much of which survived until the era of Parliamentary Enclosure (e.g. Thaxted: Newton 1960; Saffron Walden: Cromarty 1966). At Langley, for example, a relatively rare pre‐enclosure map of 1851 shows the strips and furlongs within an extensive open field system, although aspects of its management distinguish it from the Midland‐style common fields, such as the clustered, as opposed to scattered, distribution of strips belonging to individual tenements (Fig. 5.15: ERO Q/RDc 41B; Williamson 1986; Hunter 1999, 40). Langley also had relatively large areas of anciently enclosed land and woodland, and lacked a single compact village, having a more dispersed settlement pattern of scattered hamlets. In western Essex (and Hertfordshire) open fields also appear to have been quite numerous, although they were similarly less extensive than in true ‘champion’ countryside, and not organized in such a regular fashion (Roden 1973). Each of the several hamlets within a parish would have its own open field, and the holdings of individual tenants were in relatively close proximity rather than spread across the whole field system (e.g. Roydon: Fig. 5.16; Medlycott 2004). Although mostly enclosed before the eighteenth century, some small parcels of open field in the west of Essex survived to be recorded on Tithe maps (e.g. Buckhurst Hill in Chigwell: Erith 1948).
Beyond the far north and west of the county, open fields appear to have been very limited in both their frequency and individual extent. The historic landscape, as mapped in the nineteenth century, contains very few traces of former open field, and the relatively large numbers of sixteenth‐century and later estate maps are equally devoid of strip fields. In Cressing, for example, the northern part of the parish was occupied by a series of isolated farmsteads that in the nineteenth century were associated with compact blocks of land and with a predominance of ‘croft’ field names and nothing in the field boundary patterns to suggest the former existence of open field. In the south of the parish, occupied by the Knights Templar, there were a series of large documented ‘fields’ (e.g. North Field, covering c.200 m by 500 m), although their single ownership precludes them having been communal open fields (cf. Hunter 1993a; 1995; 2003, 15–19).
But what if there had been common fields in Essex that were enclosed far earlier than in areas such as the Midlands, with the longer period of time between enclosure and the earliest cartographic sources meaning that there was more opportunity for continued evolution to alter the form of the immediate post‐enclosure closes making their origins as former common field more difficult to recognize? While this possibility cannot be ruled out, medieval documentary sources also fail to reveal evidence for anything other than a scatter of small‐scale and highly irregular open fields within a landscape dominated by closes. In the large Thames‐side parish of Havering, for example, (p.181) there is ‘no clear indication of open‐field agriculture . . . prior to 1251 and none at all thereafter’ (McIntosh 1986, 93). Around Colchester and in the Lexden Hundred there was ‘an unsystematic arrangement of fields large and small, some subdivided and some not, in which holdings were usually made up of compact blocks of land rather than strips’ and weakly developed rights of common pasturage on uncropped arable (Britnell 1988, 159; Cooper 2001). In Witham, Rivenhall, Great Waltham, and High Easter the twelfth‐ to fifteenth‐century landscape was largely enclosed, and although there are references to some parcels within subdivided fields, ‘newly enclosed’ land, and widespread use of a three‐course rotation of crops, there is no evidence for communal cropping or pasturing in common on uncropped arable land. The land of individual tenements was in compact blocks, and nineteenth‐century field names give no indication of extensive former common field (Poos 1983, 191–2; 1991, 51–6; Britnell 1983; Rodwell and Rodwell 1993, 134–73; Pitchforth 2001).
In contrast to areas such as the East Midlands and the northern part of East Anglia (see below), the landscape history of Essex and south‐west Suffolk has been relatively neglected. It is a landscape of dispersed medieval settlement and predominantly enclosed fields. In places there appears to be a degree of continuity from Roman to medieval, while elsewhere landscapes of the two periods bear little relationship to each other. This appears to have been a countryside of gradual evolution, with periods of expansion and contraction, and occasional local reorganizations of settlement patterns and field systems, against a backdrop of long‐term continuous exploitation. Around the fifth century there may have been a decline in occupation on the heavier clays of the interfluvial areas, while settlements may have shifted away from some of the gravel terraces around the eighth century, but even in these areas the landscape may not have been abandoned entirely as there is no palaeoenvironmental evidence for the widespread woodland regeneration that would have happened if grazing had ceased entirely. There are several carefully planned landscapes that do suggest localized replanning of the landscape—most notably in the Dengie and Rochford hundreds—and in the case of the latter at least, this can be dated to between the sixth and tenth centuries. There are some open fields, but they were not widespread. The landscape of southern East Anglia clearly did not experience a major transformation like that seen in England's central zone, which led to the creation of villages and common fields: of the regions we have looked at so far, this one sits alongside the far west of Cornwall in being characterized by gradual evolution rather than sudden transformation.
(p.182) THE ORIGINS OF THE HISTORIC LANDSCAPE IN NORTHERN EAST ANGLIA
The landscape of northern East Anglia—Norfolk and the north and east of Suffolk—is a particularly interesting one to study, as it has seen a series of extensive and well‐published parish‐based fieldwalking surveys,6 and a long history of metal detectorists reporting their finds. Compared to Essex there have, however, been fewer large‐scale excavations, and this affects any discussion of how the landscape of this region developed: we have a clear picture about how settlement patterns developed across the landscape, but less information on the detailed evolution of individual sites.
The Icenian Civitas and Kingdom of the East Angles
In the Roman period this region became the civitas centred on Venta Icenorum (Caistor‐by‐Norwich), but compared to other parts of lowland Roman Britain the area saw the development of relatively few villas (Fig. 5.2B; Gurney 1994; Plouviez 1999). This is not just a product of differential visibility, as extensive fieldwalking across the region, such as in the Deben Valley in south‐east Suffolk, has failed to reveal any new villa‐type settlements (Newman 1992, 29). The distribution of other known rural settlements shows a consistent pattern across the region, with Romano‐British sites concentrating in the river valleys but also spread across the flatter interfluvial areas.7 The size of pottery scatters suggests a predominance of farmsteads and small hamlets. In Barton Bendish, for example, there was an average of 0.74 sites per square kilometre with each site associated with what appear to have been extensively manured areas (Fig. 5.17); in Hales, Heckingham, and Lodden the figure is 0.89 sites per square kilometre, and Fransham 0.85 (Rogerson et al. 1997, 13–17). In the clay districts of the Deben Valley in Suffolk, fieldwalking revealed nearly one site per square kilometre, although the density on the lighter soils of the Sandlings was half this (Newman 1992, 30).
The analysis of coin finds from eastern Suffolk suggests that a marked decline in circulation had set in by the third or fourth quarter of the fourth century, although whether this reflects depopulation or simply a decline in the (p.183) market economy is unclear (Newman 1992, 31; Plouviez 1995, 74): there is a desperate need for more radiocarbon dating of the latest phases of occupation on all Romano‐British sites in this region to see when their occupation really did cease. In the fifth century there was continuity in some areas, notably in the sandy regions and the lighter soils of the valley sides, and discontinuity elsewhere, especially the flatter and more poorly drained interfluvial areas. In Norfolk, Witton is typical. An extensive Romano‐British settlement on the high ground near Whitton Hall produced no ‘Early Saxon’ pottery, while on the lower slopes near Common Farm to the east large amounts of residual third‐ to fourth‐century pottery were recovered from the excavated fifth‐ to sixth‐century settlement associated with several Grubenhäuser, suggesting late Romano‐British occupation in the vicinity. The site was abandoned before the use of Ipswich Ware, whose distribution is concentrated around the parish church (Lawson 1983, 44–5). Surveys in Hales, Heckingham, and Lodden, as well as Barton Bendish, similarly show that the scattered late Romano‐British settlements were mostly deserted in the fifth century, apart from a small number of sites associated with ‘Early Saxon’ pottery on the lighter soils of the valley sides; none of these fifth‐ to seventh‐century sites is associated with Ipswich Ware (Davison 1990, 16, 66; Rogerson et al. 1997). In Mannington, Wickmere, and Wolterton there was similarly a marked decrease in the density of settlement, with one ‘early Saxon’ site compared to seven of the Romano‐British period, while in the Fransham and the Wymondham parishes, where extensive metal detecting has complemented fieldwalking, the heavier interfluvial areas were similarly deserted, although settlement continued in the valleys (Williamson 1993, 58; Rogerson 2005). A similar picture is emerging in Suffolk where the heavy clays were abandoned, while settlements continued in use on the lighter sandy soils into the sixth or seventh centuries. Only a few of these sites have a few Ipswich Ware sherds, suggesting a dislocation of the settlement pattern perhaps towards the end of the seventh or early eighth century (Warner 1987, 9; Newman 1992, 32; 2005b, 481–3; Martin et al. 1986, 232–4; 1992, 378; 1994, 206–7; 1995, 344).
Palaeoenvironmental evidence suggests a very similar picture. A series of pollen sequences all show broad continuity in land‐use from the late Roman through to the early medieval period, with a largely open, pastoral landscape and some arable on the lighter soils: there may have been some contraction in the extent of cultivation, but this was replaced by grassland, and there was very limited or no woodland regeneration.8 An indication that the heavier soils of (p.184) (p.185)
As was discussed for southern East Anglia, the results of such surveys enable us to discuss the development of settlement patterns in broad spatial terms, but not who was actually living in those settlements: we have to remember that some of the fourth‐century settlements that lack fifth‐ to seventh‐century ‘Early Anglo‐Saxon’ pottery could have been occupied by the aceramic descendants of the native British population. In contrast to southern East Anglia, however, the area north of the Gipping–Lark valleys does have a large number of fifth‐ to seventh‐century cemeteries (Fig. 5.2C), suggesting a substantial Anglian settlement from the second or third quarters of the fifth century (Scull 1992, 8). The extent to which a native population survived and how they interacted with the newcomers is unclear, but in a number of cases it is noticeable how ‘Early Saxon’ settlements were located at the periphery of late Romano‐British sites, as seen in the relative locations of West Stow and Icklingham, and also at Burgh Castle, Hacheston, Long Melford, Pakenham, and Venta Icenorum (Williamson 1993, 67; Plouviez 1995, 78–9; Blagg et al. 2004). It may have been either that the immigrants were drawn to these agriculturally productive areas but wanted to avoid living amongst the ruins of abandoned Roman sites, or that an aceramic native British population survived and the newcomers were forced to settle at the periphery of what were still functioning territorial centres.
A ‘Middle Saxon Shuffle’? Reorganization and Intensification in the Settlement Patterns of Northern East Anglia During the Long Eighth Century
Across northern East Anglia, fieldwalking and metal detecting surveys show a major dislocation in the settlement pattern around the eighth century, with the dispersed scatter of Romano‐British and ‘Early Saxon’ settlements being abandoned in favour of a series of nucleated villages that were later associated with parish churches. Archaeological work in what is now farmland (i.e. outside deserted and still‐occupied medieval settlements) is also consistently revealing a dispersed settlement pattern in the Roman and earliest medieval periods (the fifth to seventh or eighth centuries),9 while work within deserted medieval villages and still‐occupied medieval settlements is increasingly (p.187) producing evidence for occupation associated with ‘Middle and Late Saxon’ pottery.10 Although the fieldwalking surveys suggest that most of the dispersed settlement pattern associated with ‘Early Saxon’ pottery was abandoned before Ipswich Ware came into circulation, which is now dated by Blinkhorn (1999) to the early eighth century, a number of earlier sites in northern Suffolk have produced small amounts of this ‘Middle Saxon’ pottery such as Lakenheath (Jo Caruth pers. comm.) and Bloodmoor Hill in Carlton Colville (Tipper et al. forthcoming). West Stow has also produced Ipswich Ware (West 1985, 137), and although the excavated areas of the nearby site at Lackford Bridge and Melford Meadows near Thetford (Dallas 1993; Mudd 2002) lacked Ipswich Ware, considering the shifting nature of early medieval settlement, occupation of this period could lie in adjacent unexcavated areas. In the Deben Valley some, but not all, of the ‘Early Saxon’ sites produced a few sherds of Ipswich Ware, once again suggesting a dislocation of the settlement pattern around the early eighth century (Newman 1992, 32; 2005b, 481–3). The abandonment of the excavated areas of settlement at Kilverstone also gives us a terminus post quem for the laying out of the common fields over the same area (Fig. 5.18; Davison 1988, 18–32; Garrow et al. 2006).
The nucleation of settlement in the ‘Middle Saxon’ period is also clearly evident in the major published fieldwalking surveys of Norfolk parishes such as Barton Bendish (Fig. 5.17), and in northern and eastern Suffolk (e.g. Deben Valley: Martin et al. 1995, 344; Newman 1992; 2005; the South Elmhams: Martin et al. 2002, 213; Sudbourne: Martin et al. 1992, 378; Westleton: Martin et al. 1994, 208). These pottery scatters are sometimes relatively extensive and are suggestive of more than single farmsteads. Occupation associated with Ipswich Ware has also been confirmed by excavations at now isolated churches such as Iken and Barham in Suffolk (West and Scarfe 1984; Martin et al. 1982, 159). Across northern East Anglia a high proportion of these settlements associated with ‘Middle Saxon’ pottery are associated with hām place names which is in keeping with the view that this was the habitative name most favoured in the earliest years of English name giving (Gelling 1992, 55): in Launditch Hundred, for example, six out of eleven villages producing Ipswich Ware have hām names, and all but one of the hām names that have been surveyed have produced Ipswich Ware (Wade‐Martins 1980, 84–5). In contrast, of the six ‐ton place names none has produced Middle Saxon pottery (Wade‐Martins 1980, 85; Gelling 1992, 56–7). It is noticeable that there are relatively few hām names in South‐West Suffolk, providing yet another (p.188) example of how this region has more in common with Essex than the rest of Suffolk and Norfolk (Martin 1999b). It is also during the period when Ipswich Ware was in use that the heavier claylands of northern East Anglia started to be recolonized (Newman 1992, 34).
It is likely that the locations chosen for these ‘Middle Saxon’ villages were among the many isolated farmsteads that were spread across the landscape in what were to become medieval parishes. Some deserted medieval villages that have been fieldwalked have produced significant amounts of late Roman pottery suggestive of settlements in the vicinity (e.g. Kilverstone), although the very small amounts of material from other sites are more difficult to interpret, and could simply have been from manure scatters (e.g. Beachamwell, Egmere, Rougham, Roundham, and Whissonsett: Cushion et al. 1982, 50, 86; Davison 1988, 10–11, 34, 63; Mellor 2004, 6; Trimble 2006). This was certainly the case at the hamlet of Cotes in Great Palgrave as documentary sources and fieldwalking both suggest that this hamlet was only established in the thirteenth century (Davison 1982). In the Launditch Hundred Survey, on the relatively high Boulder Clay plateau of Norfolk's central watershed, only four out of the twelve villages associated with Ipswich Ware scatters have produced Roman material, although ‘Early Saxon’ pottery was entirely absent (Wade‐Martins 1980, 82–4; Rogerson et al. 1997, 44). This lack of pre‐‘Middle Saxon’ material is also seen elsewhere: relatively extensive excavations at Barton Bendish, for example, produced just a single sherd of first‐century AD pottery (Rogerson et al. 1997, 59).
Across northern East Anglia there are just a few Ipswich Ware scatters away from the village cores and these are probably secondary ‘daughter settlements’, in peripheral areas of parishes, in that they are just a few sherds amongst what are predominantly ‘Late Saxon’ scatters (Newman 2005b, 483). The formation of these secondary settlements marks the beginning of a trend towards the dispersion of settlement with the migration of farmsteads to the edges of the large commons that occupied areas of heavier soil in the interfluvial areas (Warner 1987, 17–18; Newman 2005b, 483). From around the eleventh century the migration of settlement away from the ‘Middle to Late Saxon’ villages towards nearby commons and greens became widespread, and many parish churches were eventually left isolated: for much of northern East Anglia, the era of villages was short‐lived.
Interestingly, this same pattern—of a small number of relatively compact villages associated with ‘Middle and Late Saxon’ pottery, followed by the expansion and migration of settlement along droveways towards areas of common land—is also seen in the extensive reclaimed wetlands of the Norfolk Marshland, in the far west of the county. As a reclaimed wetland that saw extensive post‐Roman flooding, this was a ‘cleaned slate’ upon which the medieval landscape was created without an antecedent cultural landscape to affect its character (as was also seen in Somerset: see Chapters 2 and 3), and (p.189) the similarity of landscape development here compared to the dryland areas of Norfolk suggests that nucleated villages were the way that Norfolk society chose to structure its landscapes in the ‘long eighth century’ (Silvester 1988; 1993; Rippon 2000a, 208–11; Crowson et al. 2004).
The Romano‐British and Early Medieval Field Systems of Northern East Anglia
Another issue to consider is the relationship of the Romano‐British fieldscape to the historic landscape. As in southern East Anglia, excavations are suggesting that some Romano‐British field systems remained in use into the fifth century, although they invariably go out of use sometime before the creation of the historic landscape of today. At Lakenheath, Spong Hill, Barking, Melford Meadows, and Broome, for example, late Romano‐British ditches were still open in the fifth century, as they contained sherds of ‘Early Saxon’ pottery; in many cases these ditches were respected during the construction of mid‐ to late fifth‐century Grubenhäuser, but they were then abandoned and an entirely new pattern of medieval fields created at some unknown date (Martin et al. 1993, 217–18; Rickett 1995; Boulter 2002a; 2002b; Mudd 2002, 52–69; Robertson 2003). At Kilverstone, in the Norfolk Breckland, in contrast, the Romano‐British landscape appears to have been deserted before the area was reoccupied in the sixth century although this settlement was in turn abandoned in the seventh century, being replaced at an unknown date by the common fields of Kilverstone village (Fig. 5.18; Garrow et al. 2006). At Bloodmoor Hill in Carlton Colville, near Lowestoft in Suffolk, an extensive first‐ to second‐century Romano‐British field system was similarly abandoned before the site was reoccupied in the fifth century, and this settlement was in turn deserted in the early eighth century, with the medieval pattern of fields lying unconformably over these earlier landscapes (Dickens et al. 2006). One site where a Romano‐British field system is on a similar orientation to the historic landscape is Grange Farm, Snetterton, in the Norfolk Breckland, although this may be a coincidence: the way in which a Grubenhaus cuts one of the Romano‐British ditches suggests that the earlier field system had all but disappeared, while a ditch associated with tenth‐ to eleventh‐century pottery was on a completely different orientation from both that and the overlying medieval landscape (Robertson 2004).
This review of the relationship between excavated Romano‐British field systems and the historic landscape suggests that there is far less evidence for continuity than in southern East Anglia (see above), and this appears to be confirmed by the evidence from cropmarks (although these are mostly restricted to the lighter soils in the east of Norfolk and Suffolk). Here, in places such as Witton, a fairly extensive series of cropmarks that excavation has dated to the Romano‐British and ‘Early Saxon’ periods are on a different orientation from the medieval field system that evolved into the historic landscape (p.190)
One argument in favour of the survival of Romano‐British field systems was put forward by Williamson (1987; 1993) and Warner (1996, 45–53), who have identified a series of apparently planned coaxial landscapes across the (p.191) heavier claylands of northern East Anglia that they suggest ignore the natural topography (being ‘terrain oblivious’), and must date to the late prehistoric period as they appear to be cut by Roman roads (as at Little Waltham in Essex: Fig. 5.8). Such an extensive survival of field systems would imply a significant degree of continuity in land‐use (for which there is some palaeoenvironmental evidence: see above), and the maintenance of field boundaries. Although the late prehistoric date for these coaxial landscapes was rejected by Hinton (1997), and reiterated by Williamson (1998; 2003, 40–3), unfortunately there is still no undisputed dating evidence, although at North Creake, in Burnham Sutton (north Norfolk), it appears that a common field system was created within a framework of roughly parallel boundaries inherited from the Roman period (Percival and Williamson 2005).
In common with southern greater East Anglia, a final characteristic of the Romano‐British and ‘Early Saxon’ landscapes in the north is their fluidity, with sites being periodically occupied, abandoned, and reoccupied (e.g. Thornham and Wighton: Gregory and Gurney 1986). Romano‐British field systems were equally transitory, with localized ditched enclosure systems created, left to silt up, and recreated, often on slightly different orientations and alignments (e.g. Melford Meadows in Brettenham: Mudd 2002, figs. 7, 13, and 14; Spong Hill: Rickett 1995, figs. 38, 44, and 47). The mobility of ‘Early Saxon’ settlement has long been recognized with West Stow a classic example (West 1985), to which fieldwalked settlements such as Illington and recently excavated sites at Melford Meadows, Bloodmoor Hill, and Lakenheath can be added (Davison et al. 1993; Mudd 2002; Tipper et al. forthcoming; Jo Caruth pers. comm.). The overall impression is that away from the major Roman roads there was little in the way of a fixed physical framework within the landscape, and that many settlements and field systems had a lifespan of just a few generations before being abandoned: the stability of some medieval villages in northern East Anglia, occupied from around the eighth century to the present day, is a real change from earlier times.
Changing Land‐Use Around the Eighth Century
Palaeoenvironmental sequences are also suggesting that the period around the eighth century was one of change, most notably an expansion in agricultural production. The well‐dated pollen sequence from the Oakley palaeochannel at Scole suggests that there was a marked intensification in cereal cultivation in the fifth century and no change in the extent of woodland or scrub in the post‐Roman period; the area of pasture either contracted or was grazed more heavily. Around the eighth century (based on a calibrated radiocarbon date of AD 670–820) there was then a further agricultural intensification, with an increase in cereal pollen, the emergence of viticulture, and the cultivation of hemp, which was also seen at Diss and Old Buckenham (see below) (Wiltshire (p.192) forthcoming). At Micklemere there was also a marked increase in cereal pollen dated 1290+/−100 BP (cal. AD 588–972) at the same time as there was a high influx of mineral sediment, implying increased soil erosion in the catchment (Murphy 1994, 29). It is possible that other pollen sequences from northern East Anglia that show a period of agricultural intensification in the early medieval period date to this ‘long eighth century’, although the dating evidence is not as good: at Old Buckenham Mere there is a decline in oak woodland dated ‘c.800 AD’ (Godwin 1968, 102), while at Diss Mere (Peglar et al. 1989), Old Buckenham Mere (Godwin 1968), and Sea Mere (Sims 1978) there is a marked increase in the pollen of Secale cereale (rye), Avena/Triticum type (wheat), Hordeum type (barley), and Cannabis type (hemp) at around ‘1500 BP’ (around the sixth century). At Hockham Mere a similar expansion in cultivation is dated to ‘around 1300 BP’ (around the ninth century) (Sims 1978, 57; Bennett 1983a; 1983b).
Burials and Churches in the Landscape
One view of these changes in the rural landscape, including the expansion of settlement onto heavier soils, is that it reflects the need to increase agricultural production to support the population of newly emerging coastal trading settlements such as Ipswich (Moreland 2000; Hamerow 2002, 123). Indeed, the recent identification of large numbers of ‘productive sites’—a term used by numismatists to describe locations yielding relatively large numbers of early medieval coins—suggests that there was a growing non‐agriculturally productive sector of society, as trade and exchange increased across the region, reflected in the widespread distribution of manufactured goods such as Ipswich Ware. So was the nucleation of settlement into villages simply a product of the need to increase agricultural efficiency?
The ‘long eighth century’ was also the period when there was a profound change in burial practice, with the abandonment of ‘early Anglo‐Saxon’ cemeteries and the cessation of burial with (datable) grave goods (Geake 1997). While many fifth‐ to seventh‐century cemeteries were close to settlements but not within them (e.g. West Stow, Suffolk: West 1985; Flixton, Suffolk: Boulter 2003), others such as Spong Hill were so large that they must have served a considerable area, and it has been argued that they were replaced by small, Christian ‘final‐phase’ cemeteries associated with individual settlements (e.g. Faull 1976). Although Boddington (1990) has questioned this model, suggesting that these eighth‐century burial grounds were part of the constant process of the creation and desertion of cemeteries (and see Geake 1997), we are still left with the problem of where people were buried, between the demise of the ‘early Anglo‐Saxon’ cemeteries in the seventh or early eighth centuries, and the appearance of large numbers of parish churches in the landscape, which is traditionally dated to around the tenth century (see above): so when did burial (p.193) grounds change from being located beyond the edges of settlements to being situated within them? In the context of this study, this is crucial to understanding the relationship between parish churches and the villages associated with Ipswich Ware that occur around them in northern East Anglia: did settlement nucleate around a church, or were churches added to existing villages?
North Elmham shows how ecclesiastical sites became a focus for ‘Middle Saxon’ settlement, although as an episcopal centre it is hardly typical (Wade‐Martins 1980). There is, however, growing evidence that some eighth‐ and ninth‐century settlements, in East Anglia at least, had churches. Rik Hoggett (pers. comm.) has observed that where a scatter of pottery around a medieval church contains only Ipswich Ware and no ‘Late Saxon’ material (the settlement having shifted its location to a nearby green), it suggests that the church was contemporary with the ‘Middle Saxon’ settlement (e.g. Mileham: Wade‐Martins 1980, 47–8). Unfortunately, we will never be able to excavate large numbers of churchyards, as the majority are still in use, although at All Saints in Barton Bendish a cemetery that was stratigraphically beneath the first stone church itself cut a buried soil containing tenth‐century pottery (Rogerson et al. 1987). At Iken and Framingham, burials have also been recorded that pre‐date the stone church (West and Scarfe 1984; Rogerson et al. 1987, 81), while at Whissonsett a series of graves are associated with Middle Saxon pottery (Mellor 2004; Trimble 2006). Burial also appears to have occurred within eighth‐century settlements at Ipswich (Scull and Bayliss 1999; Scull 2001) and Bloodmoor Hill in Carlton Colville (Tipper et al. forthcoming), both in the east of Suffolk, along with Gamlingay in Cambridgeshire (Murray and McDonald 2005). A ‘Middle to Late Saxon’ cemetery has also recently been excavated close to the parish church at Caister‐on‐Sea in Norfolk (Albone 2001). At Sedgeford, in north‐west Norfolk, a cemetery of 184 east–west oriented graves without grave goods c.350 m to the south‐east of the medieval parish church was sealed by an eighth‐ to ninth‐century occupation horizon, and one of the graves was radiocarbon dated to cal. AD 600–760. It is unclear, however, whether this was an ‘ordinary’ rural settlement, as there have been a number of finds including two styli that suggest this may be a ‘productive site’ of a type that some have claimed have monastic associations, such as Brandon and Burrow Hill in Suffolk (Fenwick 1984; Faulkner 2000, 125; Cabot et al. 2004; Pestel 2005, 31–64).
The Romano‐British and earliest medieval (fifth‐ to seventh‐century) landscape in northern East Anglia appears to have been of broadly the same character as that to the south (with the exception of there having been fewer villas): there was a predominantly dispersed settlement pattern, with isolated farmsteads and small hamlets associated with localized field systems. Settlement was (p.194) concentrated on the lighter soils of the valleys but also spread up onto the heavier clays of the interfluvial areas. There was considerable fluidity in a countryside that was constantly evolving, and the fifth century appears to have seen a period of settlement contraction from the heavier soils, but around the eighth century there was a more profound change. The dispersed settlement pattern and localized field systems were replaced by a series of nucleated villages that were presumably associated with the open field systems that survived into the late medieval period or later. These open fields, along with the network of roads that linked the villages and the churches that were added to them around the tenth century, provided a degree of stability to the landscape that it had not previously experienced, although in some areas the migration of settlement to the greens and commons left the churches to stand in splendid isolation. Around the eighth century palaeoenvironmental sequences also show a period of agricultural intensification, also not seen in southern East Anglia.
THE ORIGINS OF VILLAGES AND COMMON FIELDS AT THE EASTERN EDGE OF THE CENTRAL ZONE IN CAMBRIDGESHIRE
To complete this overview we should turn to Cambridgeshire, which sees the eastern limit of the central zone as defined by Roberts and Wrathmell (2000) and discussed by Taylor (2002). Fieldwalking in Cambridgeshire has confirmed that a dispersed Romano‐British settlement pattern continued into the early medieval period (e.g. Caxton: Oosthuizen 2005, 169; Whittlesford: Taylor 1989, 217–18). At Cardinal Distribution Park, in Godmanchester, for example, a Romano‐British settlement was occupied throughout the fifth to seventh centuries, with just six sherds of Ipswich Ware and the absence of Maxey Ware suggesting it was abandoned by the early eighth century (Gibson and Murray 2003). In common with areas such as Northamptonshire, this scatter of farmsteads was at some stage swept away and replaced with nucleated villages, and excavations at Gamlingay show that the common fields there were laid out over a settlement associated with ‘Early to Middle Saxon’ pottery (including Maxey and Ipswich wares) that was abandoned by the mid ninth century (‘Late Saxon’ pottery being absent: Murray and McDonald 2005). At Haslingfield, there was a scatter of settlements associated with ‘Early and Middle Saxon’ pottery, some of which were deserted when a large oval‐shaped green was created to the north of a possibly planned village next to the church (Oosthuizen 2002, 75–6). Research by Oosthuizen (2005; 2006) in the Bourne Valley has revealed an extensive common field system whose long furlongs cut across several parishes, a hundred boundary, and the boundary between Comberton and Toft that is recorded in a charter of 975, suggesting they were in place before the early tenth century. They also cut unconformably (p.195) across several Romano‐British settlements, and as such large‐scale landscape planning is regarded as unlikely in the fifth to seventh centuries, Oosthuizen suggests they were laid out in the eighth or ninth centuries in order to increase agricultural production.
Roberts (1987, 49–51) and Taylor (1989; 2002) have provided us with a series of fascinating discussions of village plans and the possible ways that they may have evolved over time, but perhaps the most significant advance has been recent archaeological work within the historic cores of still‐occupied villages. In Taylor's (1989) discussion of Whittlesford, for example, the crucial piece of evidence in the context of this study is that the village has produced ‘Late Saxon’ Thetford Ware, as has also been the case at Houghton and Wyton (Lewis 2005; forthcoming). Several excavations within extant villages, however, are suggesting an origin in the eighth century. One of the most extensively excavated sites is at West Fen Road in Ely where an extensive village‐like settlement—extending some 500 m north to south—was laid out in the early eighth century (Mortimer et al. 2005, 4, 144–8). The orientation of the tenement boundaries is the same as the furlong boundaries within the common field system that covered Ely in the medieval period (Hall 1996, 40), suggesting that the whole landscape was laid out at the same time. Even if this was a somewhat untypical settlement, associated with the nearby monastery founded in c.673, it still shows that the concept of planned, nucleated settlements existed in this region at that time. In Cottenham, a cluster of settlement enclosures developed during the eighth century which were then abandoned as the focus of occupation shifted to the south where a radially arranged series of tenement plots developed (Mortimer 2000). At Hinxton, a small settlement, probably a single farmstead, was occupied between the late sixth/early seventh and eighth centuries after which it was replaced by a series of timber halls set within enclosures suggestive of a series of adjacent farmsteads which also appear to represent the beginnings of the village; in the eleventh century they were replaced by the present planned settlement (Taylor et al. 1994; Taylor 2002, 55–6). At Cherry Hinton, extensive excavations have revealed a spread of ‘Middle Saxon’ pottery and several ditches, which were replaced in the late ninth or early tenth century by a large, possibly manorial, enclosure associated with a church and cemetery containing 670 east to west oriented inhumations (Cessford 2005). Various excavations within Chesterton have similarly revealed residual ‘Middle Saxon’ pottery, with stratified occupation deposits from the ‘Late Saxon’ period associated with a series of ditched enclosures (Cessford 2004). ‘Middle Saxon’ pottery has also been recovered from the villages at Fordham, Whaddon, and Willingham (Oosthuizen 1993; Mortimer 2000, 20; Taylor 2002, 63).
Taken altogether, this recent evidence from Cambridgeshire is suggesting that the ‘Middle Saxon’ period saw a crucial change in the landscape with the start of settlement nucleation. Often, the earliest ‘Middle Saxon’ occupation (p.196) is simply represented by a scatter of residual pottery and just a handful of features, suggesting either small‐scale occupation or that the focus of activity lay outside the excavated area: indeed, a degree of settlement mobility is seen at both Cherry Hinton and Cottenham. As Mortimer (2000) suggests, this putative eighth‐century date for the origins of villages is a departure from the prevailing view that nucleation happened around the mid ninth century at the interface between the periods when ‘Middle Saxon’ and ‘Late Saxon’ pottery was in use, although some villages were undoubtedly created at this later date. Indeed, all of these recent excavations within still‐occupied medieval villages show how the settlements continued to evolve with expansion, reorganization, and desertion of certain areas (and see Oosthuizen 1993; 2002; Taylor 2004).
DISCUSSION: REGIONAL VARIATION IN CONTINUITY, REORGANIZATION, AND DIVERGENT DEVELOPMENT
This discussion of the landscape in greater East Anglia has confirmed that it lay outside England's central zone of champion countryside: this was not a region of compact nucleated villages surrounded by large common fields. It is also clear, however, that this was not a landscape carved out of woodland in the later first millennium AD: palaeoenvironmental sequences consistently show there was an extensively cleared landscape by the Roman period and that there was no significant woodland regeneration in the post‐Roman period. There was, however, a significant difference within this region along the line of the Gipping–Lark valleys or watersheds that corresponds to a long‐lasting socio‐political division between the tribal area and civitas of the Iceni and kingdom of the East Angles to the north (in Norfolk and northern/eastern Suffolk), and the tribal area and civitas of the Trinovantes and the kingdom of the East Saxons to the south.
Williamson (2002, 84; 2006a, 42) suggests that the long‐term division of this region along the Gipping–Lark line is probably due to continuity in the systems of contact and exchange that were determined by the configuration of coast and rivers, rather than by any direct survival of political entities. This is, however, a rather physically deterministic explanation, and if it was the case then one would surely expect a far stronger east–west division within the landscape of greater East Anglia, along the lines of the East Anglian Heights and the central watershed in Norfolk, with society to the east focused on the estuaries flowing into the North Sea, and society to the west looking towards the Fenland Basin. Such a boundary is evident in landscape character to a certain extent—with more nucleated settlement and regularly arranged two‐ and three‐field systems to the west, and more dispersed settlement and smaller‐scale open fields in the east (cf. Figs. 5.1 and 5.3)—but it was the north and south divide, along the (p.197) Gipping–Lark valleys, that was far more significant. It is suggested here that, as we have seen in the South‐West, there are certain places within the British landscape where different societies have taken certain natural features as the boundaries between themselves, of which the Blackdown and Quantock Hills in the South‐West and the Gipping and Lark valleys in East Anglia are two examples.
This boundary was particularly important during the ‘long eighth century’, when many of the character‐defining features of the historic landscape were created. Before that time, settlement patterns across the region were predominantly dispersed, with a scatter of farmsteads and hamlet‐sized settlements spread right across the landscape, although with greater densities in the river valleys. There was greater Romanization in the south, as reflected in the higher density of villas, but society in the north was still clearly stratified as reflected in the accumulation of wealth deposited in hoards. This dispersed settlement pattern was associated with relatively discrete and unstable fieldscapes that were periodically deserted and then reordered. In the fourth and fifth centuries there are some signs of economic decline and a decrease in the intensity with which some areas were exploited, although few if any areas appear to have been completely abandoned: if grazing had ceased across large areas there would very soon have been a woodland regeneration and this is not evident in the palaeoenvironmental record. During the earliest medieval period (the fifth to seventh centuries), settlement certainly appears to have retreated from the heavier interfluvial areas, although once again some form of agriculture must have continued to prevent the spread of scrub and woodland.
Fluidity and mobility were still key characteristics of this period, and in southern East Anglia this remained the case throughout the medieval period, albeit within an increasingly fixed framework of roads and field boundaries in what was becoming a crowded landscape. There is growing evidence here for a scatter of ‘Middle Saxon’ and later settlements spread across the landscape, both beneath still‐occupied places and what are now agricultural fields. Palaeoenvironmental sequences point to broad continuity in land‐use as a backdrop for the localized creation and desertion of settlement. There were some open fields, perhaps an innovation brought in from the northern part of East Anglia or the Midlands, but the field systems were predominantly enclosed: this remained a physical landscape designed by and for the individual rather than the community.
Until there is a major programme of excavation to test the date of the coaxial and other planned landscapes, their real significance cannot be appreciated. Small‐scale work at sites such as Little Waltham and Saffron Walden appears to support an Iron Age date for at least some of the coaxial patterns, although in Dengie all we can say is that its rectilinear layout is pre‐tenth century. The radial landscape in Rochford Hundred is clearly post‐fifth century and pre‐tenth century in date. The different dates of these planned landscapes make (p.198) an important general point about the landscape in southern East Anglia: it is a complex palimpsest that has been constantly evolving.
In northern East Anglia, however, there was a profound change in landscape character around the eighth century. The dispersed settlement pattern was abandoned as farmsteads coalesced into a series of nucleated settlements that were in essence villages, and also provides the most obvious context for the emergence of common fields. By the eleventh century these settlements had parish churches, most of which probably originated before the tenth century. This transformation of the landscape during the ‘long eighth century’ does not appear to be much different from that seen in the East Midlands, although in East Anglia it was much shorter‐lived. Even within the period when Ipswich Ware was in use, we start to see secondary settlements being created, and by the eleventh century fieldwalking and excavation show a widespread tendency for settlement to drift away from the early villages, and their parish churches, towards greens and commons that so characterized the interfluvial areas.
So how can we explain this marked difference in landscape character between northern and southern East Anglia? Williamson (2003) and Martin (2007) both note differences in topography and soils, with the clayland in the south, for example, being more frequently dissected by river valleys. What to modern farmers might seem relatively minor differences in the properties of soil may indeed have been more significant in the past, given the nature of medieval technology and farming practices. Williamson (2006a, 56) has also argued that the greater abundance of meadow in the Midlands, compared to East Anglia, promoted a more communal approach to managing the landscape as hay making requires good weather, abundant labour, and careful timing of its collection. This, he argues, means that large areas of meadow encourage the growth of large nucleated villages ‘in which the workforce could be quickly assembled’, whereas in places where meadow occurred in smaller, more scattered areas, settlement would be more dispersed. Williamson argues that southern greater East Anglia (to the south of the Gipping–Lark valleys) had less meadow than the Midlands, and this is certainly in keeping with the former having more dispersed settlement than the latter, but the hypothesis falls down when it comes to northern East Anglia. This region has more nucleated settlement compared to southern East Anglia yet it has less meadow, as the more widely spaced valleys tend to be filled with peaty soils. In fact, it is very difficult to see any difference in the physical environment that can explain why landscape and society in the north and south of East Anglia were so different, or why the Gipping–Lark valleys were such a persistent boundary between peoples and the ways that they structured and managed the landscape.
Instead of being dictated by the natural environment, is it not possible that the landscape of northern East Anglia was different from that of the south because of deep‐rooted differences in society that are clearly evidenced in the (p.199) Iron Age, Roman, and earliest medieval periods, and which appear to have persisted as social and economic networks (though it must be acknowledged that these could have been influenced by topography)? Just a few examples must suffice. One regionally distinctive aspect of the ecclesiastical landscape in the north was the large numbers of circular church towers: there were 185 to the north of the Gipping–Lark valleys, and just six in Essex (Williamson 2006a, 91). It has been argued that this design was adopted due to the lack of good freestone in the county necessary for building corners, but this cannot have been the reason (Heywood 1988). First, Essex and southern Suffolk are similarly deficient in this material yet have overwhelmingly square church towers, and secondly there were several sources of freestone in the west of Norfolk used in church building including both circular and square towers (Allen 2004a). Clearly, the decision of these communities to build circular church towers had nothing to do with the nature of local building supplies. Another oddity of the ecclesiastical landscape to the north of the Gipping–Lark line is that there are seventeen examples of two or three churches being located in the same churchyard; there is just one in Essex (Williamson 2006a, fig. 4.12). This cannot, in any way, be due to regional variation in the physical landscape and must be due to differences in the structure of society, such as the relatively large proportions of freemen in the north who may have collectively founded a new church next to an existing one as an expression of their independence from lordly control (Williamson 2006a, 89): in the Domesday survey this area certainly had a very high proportion of freemen in the population compared to the rest of the country (Darby 1977, fig. 20).
There are even differences either side of the Gipping–Lark line in vernacular architecture, for example in how roofs were supported, and in the presence of ‘Wealden’ styles houses in both towns and the countryside in southern greater East Anglia but not the north (Alston 1999; Colman 1999; Colman and Barnard 1999; Haward and Aitkins 1999). During the sixteenth century, when open halls were floored over and chimneystacks replaced hearths, in south‐west Suffolk and Essex they tended to be inserted against the cross‐passage, while in north‐east Suffolk and Norfolk they were usually placed between the hall and the inner room (Williamson 2006a, 100). And one last example: north of the Gipping–Lark line ceramic pantile roofs were found on over 30 per cent of post‐medieval buildings (in Norfolk and north‐west Suffolk the figure is over 60 per cent), whereas in Essex and southern Suffolk they occurred on less than 5 per cent of buildings (Williamson 2006a, 104). It is remarkable how persistent and ever present the Gipping–Lark boundary is in so many facets of the landscape of greater East Anglia: soils will have affected farming practice, which may indeed account for some of the differences in how the countryside was structured, but they did not determine how house plans and roof structures were designed. In an interesting revisiting of old ethnic and migration‐based explanations for landscape change, Martin (2007) has also noted that (p.200) place‐name and linguistic evidence suggests a greater Scandinavian influence in northern East Anglia compared to the south, but this may simply have been another layer added to a pattern of regional difference that was already firmly established. The nucleation of settlement around the future site of parish churches—or perhaps the early churches themselves—clearly happened around the eighth century, before any Scandinavian influence, and in seeking possible causal factors we may have to look no further than the wider economic and social changes of the ‘long eighth century’ that will be discussed in Chapter 6.
(1) Note that contrary to an earlier assumption (Rippon 1991) that the D‐shaped enclosure towards the apex of this radially planned landscape was a Danish fort that the Anglo‐Saxon Chronicle records was constructed at Shoebury in 894 (Swanton 1996, 87), subsequent excavations have shown that this earthwork is Middle Iron Age in date (Gifford and Partners 1999; Mattinson 2005).
(2) Based on its associations at sites such as Othona (Bradwell‐on‐Sea), Nazingbury, Wicken Bonhunt, Heybridge, Waltham Abbey, Mucking, Barking Abbey, London, and Colchester (Rodwell 1976; Huggins 1978; Wade 1980; Drury and Wickenden 1982; Huggins 1988; Hamerow 1993; Redknap 1991; Vince and Jenner 1991, 48; Cotter 2000, 23–6).
(3) e.g. Ardale School (Wilkinson 1988, 42–57); Gun Hill in Tilbury (Drury and Rodwell 1973); Frog Hall Farm in Fingringhoe (Brooks 2002); Orsett Cock (Carter 1998); Orsett Causewayed Enclosure (Hedges and Buckley 1985); Springfield Lyons (Tyler and Major 2005).
(4) It has been suggested that two possibly abutting structures within the rural settlement at Springfield Lyons near Chelmsford may have formed the nave/chancel and tower of a church (Tyler and Major 2005, 127–8, 193), although the evidence is not convincing. The two structures, while aligned approximately east–west, are slightly offset, and although the ‘nave’ appears to have contained an inhumation without grave goods abutting its north wall the stratigraphic relationship is unclear and this burial may equally relate to an earlier 5th‐ to 6th‐century cemetery of which several inhumations with grave goods in the vicinity are on the same orientation (Tyler and Major 2005, 18). Pottery from the Springfield building(s) also dates to the 10th/11th centuries, not the Middle Saxon period.
(6) Fieldwalking surveys: Norfolk: Barton Bendish and Caldecot: Rogerson et al. 1997; Egmere and Roundham: Cushion et al. 1982; Godwick: Davison 2003b; Hales, Heckingham, and Lodden: Davison 1990; Hargham: Davison 1995; Illington: Davison et al. 1993; Kilverstone and Rougham: Davison 1988; Langford: Davison 2001; Launditch Hundred: Wade‐Martins 1980; West Acre: Davison 1996; 2003a; Witton: Lawson 1983. Suffolk: Deben Valley: Newman 1992; East Suffolk Claylands: Warner 1987; Menham, Metfield, and South Elmham: Moore et al. 1988, 58–9; Walsham le Willows: West and McLaughlin 1998.
(7) Examples of distribution maps of known settlements include the Burn and Stiffkey valleys (Gregory and Gurney 1986, fig. 10), the Blackwater and Wensum valleys (Rickett 1995, fig. 139), the Black Bourn and Lark valleys (West 1985, fig. 304), and the Little Ouse Valley (Mudd 2002, fig. 2).
(8) Caudle Heath (Wiltshire 1999), Diss Mere (Peglar et al. 1989), Hockham Mere (Godwin and Tallantire 1951; Bennett 1983a; 1983b; Sims 1978), Micklemere (Murphy 1996, 29–31), Old Buckingham Mere (Godwin 1968), Scole (Wiltshire forthcoming), and Seamere (Sims 1978). In such a flat landscape, the catchment of these meres is likely to have been from a radius of about 10–20 km (Jacobson and Bradshaw 1981).
(9) Archaeological work outside extant villages: the A12 roundabout in Hopton‐on‐Sea (Gurney and Penn 2000, 528), Bowthorpe in Cotessey (Trimble 2001; Gurney and Penn 2005, 753), Broome (Robertson 2003), Flixton Park Quarry in Flixton (Boulter 2003), Gallows Hill in Barking (Boulter 2002a; 2002b), Grange Farm and Wash Lane in Snetterton (Gurney and Penn 2001, 175; Robertson 2004), Melford Meadows in Brettenham (Mudd 2002), Priory Farm in Aldeby (Gurney and Penn 2000, 522), RAF Lakenheath (Caruth 2002a; 2002b), and Yarmouth Road in Broome (Gurney and Penn 2002, 164).
(10) Archaeological work within extant Norfolk villages: Mileham: Gurney and Penn 1998, 201; Burnham Market: Gurney and Penn 1999, 370; 2000, 524; Sedgeford: Faulkner 2000, 126; Cabot et al. 2004; South Walsham: Albone 2000, 27; Gurney and Penn 2001, 724; Brettenham: Gurney and Penn 2005, 752, 762; Whissonsett: Mellor 2004; Gurney and Penn 2005, 752, 762; Trimble 2006. Archaeological work within extant Suffolk villages: Grundisburgh: Martin et al. 1993, 95–6; 1994, 213; Haughley: Martin et al. 2000, 523; South Elmham St Margaret: Martin et al. 2002, 213.