Rectilinear Field Systems and Dispersed Settlement in South Somerset and Dorset by John Davey Abstract Geophysical survey coupled with trial excavation and cartographic analysis have revealed large but discrete regions in south Somerset and Dorset in which rectilinear field systems have been fossilised in the modern landscape. Excavation has shown these field systems to originate in the late prehistoric period and continue in use, albeit in a modified state, and with fluctuating intensity, into the modern era. This observation is of great significance for the early medieval period, because it suggests that there was relative stability in the rural economy from the Late Roman to the Late Saxon period. Land division and land holding may also have continued, until the establishment of manors and nucleated villages during the Late Saxon period. Further analysis of communications and dispersed settlement patterns associated with these field systems has given rise to a new model for the distribution of dispersed settlement and the subdivision of land units within areas of rectilinear field systems. Introduction C. C. Taylor was the first to recognise and highlight the fact that parish and tithing boundaries in Dorset frequently subdivide the landscape into parallel narrow units. These land units generally cut across the topographic grain from valley floor to ridge top. His cartographic and documentary analyses of tithings in south west Purbeck and the Piddle, Winterborne and Frome Valleys are cases in point (Taylor 2004, 52-62; 113-4). Fleming in his research into the Dartmoor Reaves, was the first to demonstrate that prehistoric land division could take the form of long co-axial field alignments (Fleming 1978). Keen noted that long field alignments, running parallel to the upper Yeo Valley in the Sherborne region of west Dorset, were aligned perpendicular to a Late Saxon herepath. This led him to suggest that they were of Middle Saxon origin (Keen 1984). Williamson developed the idea of rectilinear field alignments further and was the first to demonstrate at Scole-Dickelburgh that field systems of prehistoric origin can themselves be fossilised in the modern landscape (Williamson 1987). More recent work by Fleming (1989), Oosthuizen in west Cambridgeshire (2003) and the author in south east Somerset and north Dorset (Davey 2005, 63-83) have demonstrated the wider distribution and survival of such rectilinear or co-axial systems of land division. The above studies have shown that long co-axial alignments divided the land during the late prehistoric period, particularly in areas of gently sloping river valleys, cutting across the topographical grain. The units thus created were partly territorial and partly economic in function, ensuring that each social group were afforded a share of landscape resources from the valley floor meadows to the arable lower slopes and upper pastures and woodlands. Intensification of land use during the Late Roman period seems to have led to the formal subdivision of these larger units into individual fields whose boundaries have been continuously modified over the ensuing millenia, but still within the prevailing co-axial alignments. However, none of the above studies have addressed the distribution of settlements within these field systems, nor the implications for the early medieval rural economy. 85 Case Study 1: Horethorne Hundred, Somerset. Figure 1 shows the extent of the hundred of Horethorne, Somerset, as it was at Domesday. It is clear that the overall area of the hundred is approximately rectangular in shape, with a large chunk missing from the south east corner. This corresponds to a 7th century or earlier ecclesiastical estate at Sherborne in Dorset, which has been removed from Horethorne prior to AD 1086. The regular arrangement of parish boundaries within Horethorne hundred is also clear, particularly in the east where the parish boundaries of North Cheriton, Horsington, Abbas and Templecombe and Henstridge all have long alignments conforming to that of the hundred as a whole, that is in a ENE-WSW direction. These alignments run perpendicular to that of the Cale Valley, the western slopes of which are entirely occupied by the eastern half of Horethorne Hundred. The Cale Valley forms part of the northern arm of the wider Blackmore Vale. The watershed for the Cale runs approximately along the western boundaries of North Cheriton, Horsington, Abbas and Templecombe and Henstridge. On a wider scale, the long alignments of Horethorne Hundred run perpendicular to a series of geological strata of Jurassic age that dominate the landscape of south east Somerset and North Dorset. The oldest being the Yeovil Sands and Inferior Oolite in the west running through the Fuller’s Earth and Great Oolite series to the Cornbrash and Figure 1: Horethorne Hundred parish boundaries circa 1840 (After Dunning 1999, 73) 86 Figure 2: Horethorne Hundred field alignments showing selected dispersed settlements. Red dots represent nucleated settlements along the modern A357, other coloured dots represent parallel bands of dispersed settlement (After Davey 2005, 69) Oxford Clay in the East. These strata form a series of dips and scarps all running perpendicular to the long alignments of Horethorne Hundred. Furthermore, figure 2 shows that modern field boundaries also conform to the Horethorne long alignments, together with many of the lanes and roads. Of course it is difficult to date field boundaries depicted on modern or 19th century maps, without recourse to archaeological excavation. It would seem perfectly reasonable at first glance to assume that the field boundaries post-dated the formation of the parish boundaries in Horethorne. However, Williamson convincingly dated a similar field system at Scole-Dickelburgh to the late prehistoric period through the stratigraphic relationship between the field boundaries and the Roman Pye Road (Williamson, 1987). Similar stratigraphic relationships in Horethorne demonstrate that the field system actually predates the formation of, not only the parish boundaries, but also the Domesday County boundary between Somerset and Dorset. Figure 2 demonstrates that the alignments of these field boundaries continue into Sherborne. This raises the possibility that the early medieval ecclesiastical estate at Sherborne may once have formed part of a larger rectangular land unit from which Horethorne Hundred subsequently evolved. The stepped nature of the 1086 county boundary between the parishes of Oborne (Dorset) and Poyntington (Somerset) suggests that the rectilinear field system was already in place when these manorial boundaries were formed. The County Boundary later followed these pre-Domesday manorial boundaries. The field alignments in Sherborne depicted on figure 2 were 87 first recognised as significant by Keen (1984) who suggested they were of Saxon origin and planned by the monks of Sherborne. His reasoning was that; not only did the later Bristol Road cut across the pre-existing field boundaries, but that the field system was also aligned on the road from Sherborne to Marston Magna (The modern B3148). This road is recorded as a herepath in 10th century charters relating to Rimpton (Costen 1985) and forms the western boundary of Rimpton and Sandford Orcas parishes. Keen assumed that the field boundaries and the herepath were broadly contemporary and was further convinced by the available context for their instigation by the monks of Sherborne. However, this analysis is based on the assumption that the field boundaries did not extend beyond the Sherborne estate. Keen also assumed that the herepath was newly created in the Saxon period, rather than representing the Saxon designation of a preexisting route. Aston (1985, 146-8)) made an independent study of communication and settlement patterns in south Somerset, concluding that the alignment of the Sherborne to Marston Magna road may in fact be prehistoric in origin. This was a highly speculative conclusion however. Thus, from the available evidence, the Horethorne field system appeared to be Saxon or earlier in origin. Excavation was necessary in order to refine these conclusions. Englands, Charlton Horethorne. The opportunity to explore the date of the Horethorne field system arose during fieldwork associated with the South Cadbury Environs Project (SCEP) in 2003. Cooper (2004) first identified earthworks, accompanied by a scatter of Late Romano-British pottery, within fields known as Englands at the eastern extremity of Charlton Horethorne parish. They are located at the top of a high plateau of Forest Marble, which forms the watershed between the drainage basins of the River Parret in Somerset and the River Stour in Dorset. It is as far from modern settlement as it could be in this area, on the modern boundary between Charlton Horethorne, North Cheriton and Horsington parishes. The nearest settlement of any kind lies 1km to the NE at Clare Farm and the nucleated village of Charlton Horethorne lies 1.5km to the SW. The habitative nature of the name Englands, therefore, seemed to refer to the Romano-British site identified by Cooper. Furthermore, the Saxon origin of the name Englands implies that a settlement was still in existence on the site at the beginning of the Late Saxon period (Field 1972). Thus the site represents a potential example of continuity of settlement site from the Late Roman to Middle Saxon period. It was considered that a gradiometer survey of the site would be of considerable benefit. Gradiometer Survey and Excavation. The survey was carried out using a Bartington grad601/2 dual sensor fluxgate gradiometer, during July 2003 and the results are displayed in figure 3. It is clear from the survey that a considerable settlement existed on both sides of a double ditched trackway consisting of at least three buildings through at least two separate phases. Figure 3 represents a slight refinement of the previously published interpretation (Davey, 2004, 84; 2005, 48). The track itself is aligned on a section of the main road from Sherborne to Wincanton. This alignment in turn follows that of the rectilinear field system identified in Horethorne Hundred. The settlement is also laid out on 88 Figure 3: Englands, Charlton Horethorne: geophysical survey results and interpretation similar alignments suggesting that dating the settlement would provide a Terminus Post Quem for the field system. Further geophysical anomalies in the survey are directly aligned on the Charlton Horethorne parish boundary and modern field boundaries in the immediate vicinity. However, the morphology of the settlement, at least in its major phases, appears Romano-British rather than post-Roman or Saxon. Sample excavation was necessary, in order to demonstrate whether the settlement continued into the Saxon period and to date field alignments in the vicinity. The results of this excavation have been published elsewhere (Davey, 2004) and a brief summary only is presented here. Two trenches were excavated during early September 2003 with the intention of dating phases identified in the geophysical survey. They were deliberately located away from any potential house platforms; the sheer volume of artefacts recovered from such an excavation would have been disproportionate to the aims and available resources. Thus, Trench 1 was located at the junction of two linear anomalies, the first being associated with one of the main settlement phases and the other with a group of alignments running diagonally N-S across the settlement. Trench 2 was located at the intersection of the first anomaly in Trench 1 and another associated with a second major settlement phase. The excavation achieved its limited aims in that the phasing outlined in figure 3 was confirmed stratigraphically and broad date ranges could be assigned to the individual phases. The broader aims of identifying a post-Roman settlement site was not achieved, although this does not mean that the Late Roman settlement at Englands did not continue to be inhabited into the 5th and 6th centuries and the field-name evidence still supports the likelihood that it did. 89 Discussion The survey and excavation at Englands, Charlton Horethorne has provided an approximate date for the inception of the Horethorne Hundred field system sometime during the 1st to 2nd centuries AD. However, it must be noted that this provisional date can be refined by reference to observations that took place during the construction of a water pipeline from Bowden Reservoir to Penselwood (Newman et al. 2001). A number of Romano-British ditches were aligned perpendicular to the parish boundary between Horsington and Templecombe. They were distributed in two separate areas. To the east, a group of ditches with ‘u’ shaped profiles contained 3rd-5th century pottery. It is possible that these ditches could have silted up any time prior to the 10th century. The second group was situated close to an Anglo-Saxon cemetery excavated in 1992 (Newman, 1992). The ditches had ‘v’ shaped profiles and contained pottery dating to the 1st to 2nd centuries AD. Both groups of ditches were aligned with the Horethorne Hundred field system. However, a third site located between the two mentioned above, revealed a ditch aligned at a slight angle to the parish boundary and dating to the Late Iron Age. This is tantalising evidence that the Horethorne Hundred alignments may have originated in the late prehistoric period and their use was intensified, first in the early Romano-British period and again in the 3rd century. Furthermore, the Iron Age ditch is aligned close to that of a parish boundary. This lends weight to the theory that co-axial field systems developed first, as a number of parallel ranch or estate divisions that were gradually infilled by parallel rectangular fields (Wykes, 2003, 47). Further evidence to support the prehistoric origins of the Horethorne Hundred system comes from SCEP fieldwork directed by Tabor at Sigwells, Charlton Horethorne. Here, alignments associated with the Horethorne system were shown by excavation to have silted up in the Late Iron Age (Tabor (ed.) 2002, 57-67; 2004, 27-55). The field system itself forms only one part of an integrated rural economy. Evidence for further aspects of this economic system such as communications and settlement patterns can also be identified in Horethorne Hundred. Figure 4 shows the area in the immediate vicinity of the village of Charlton Horethorne. It can be seen that the double ditched track, revealed and dated to the Roman period at Englands, forms part of a wider system of communications within a dispersed settlement pattern. This communications and settlement pattern is integrated within the Horethorne Hundred rectilinear field system. Costen has shown, through his analysis of Late Saxon charters in Rimpton, that communications patterns are remarkably stable in SE Somerset. 100% of roads mentioned in the 10th century Rimpton charters are still in use today (Costen 1985). Because communications link settlements and fields, it is not difficult to see that the system of lanes and tracks would remain unchanged as long as the settlement pattern and field system also remained unaltered. It has been noted elsewhere that Charlton Horethorne is a rare example of a parish in SE Somerset where a single nucleated village has arisen at the expense of surrounding dispersed settlement (Davey 2005, 53). Figure 4 suggests a mechanism by which this may have occurred. Plot boundaries in the vicinity of Charlton Horethorne church suggest that a planned double row settlement once formed the core of the village. This planned settlement has been subsequently altered, but it is clear that originally the church yard and neighbouring plots extended further to the west and that a central spine road ran north-south through the middle of the plots. This north south alignment represents a departure from the prevailing alignments of Horethorne Hundred and is clearly superimposed upon and post dates them. Similar double row settlements have been dated to the Late Saxon period at Abbotsbury, Stalbridge and Sturminster Newton in Dorset (Davey 2008; 2009a; 2009b). It may be that Charlton Horethorne village, in its original form, represents a planned Late Saxon double row settlement planted on a newly built herepath from the royal estate centre, and known hub of herepaths, at Milborne Port. This 90 Figure 4: Communication patterns in Charlton Horethorne possible herepath may have linked Milborne Port with the Fosse Way and Bath to the north, via Blackford Hollow. The creation of the Late Saxon planned settlement at Charlton Horethorne seems to have been a deliberate and royally sponsored act; Charlton Horethorne formed part of the royal estate at Milborne Port at Domesday (Thorn & Thorn 1980). This planned nucleation was at the expense of the neighbouring dispersed settlements and it is likely to have been at this time that the Romano-British settlement of Englands finally went out of use. The disruption of the settlement pattern seems also to have led to the disruption of the communications pattern in Charlton Horethorne. Instead of the rectilinear patterns of lanes, a new radial pattern centred on the nucleated village emerged. This change was likely to have occurred over a protracted period of time. In fact it may have been as late as 1752-3 when the diagonal road was driven through the centre of the double row of house plots. It was at this time that the first turnpike act relating to Dorset was passed. This act instigated the Shaftesbury and Sherborne Turnpike Trust, concerned 91 primarily with the Great Western Post Road from London to Exeter and Plymouth. However, it also included provision for four shorter side roads of which the Sherborne to Wincanton road was one (Good 1966, 124). The pattern of settlement in Horethorne Hundred has also been plotted on figure 2. This depicts both nucleated villages and dispersed farmsteads or hamlets with a potentially early habitative name element. The distribution of these settlements falls into broad bands highlighted through colour coding. Perhaps the most obvious band is the line of nucleated settlements at the eastern end of the hundred lying along the modern A357 from North Cheriton to Stalbridge. These settlements lie along a spring line at the junction of the Cornbrash and the Oxford Clay. Dispersed settlements on either side of the nucleated band tend to lie away from the modern north-south routes but close to the east-west lanes aligned with the Horethore Hundred field system. The distribution of settlements within Horethorne Hundred is shown to be highly regular and evenly spaced within the rectilinear pattern. This suggests that the land units associated with them might also be regular and rectangular. These ideas will be discussed further below. Case Study 2: Christchurch Hundred, Dorset Introduction A settlement at Christchurch probably originated in the Middle Saxon period, if not before. The main archaeological evidence is the presence of a large Middle Saxon cemetery at Bargates and a Mid-Late Saxon oven at the northern end of the town, immediately south of the subsequent line of the Late Saxon burh bank (Jarvis 1983, 34-7; 104-134). The cemetery certainly suggests that the wider area of Christchurch was occupied by the late 6th to 7th century AD. The first documented mention of a settlement at Christchurch can be found in the AD 900 entry of the Anglo-Saxon chronicle in which a ‘ham’ or residence at Tweoxneam (Christchurch) is seized by the aetheling Aethelwold (Jarvis 1983, 9). This clearly suggests that Christchurch contained a royal residence by AD 900, and was likely to have functioned as an administrative, economic and religious centre for a large estate. This estate may have been similar in extent to the later Christchurch Hundred. Hase (1988) has suggested that Christchurch may have been the site of an early Minster church administering to a large parochia, possibly as early as the mid-8th century. Although there is no explicit dating evidence to substantiate this claim, other Minster churches in Hampshire have been dated to this period and there is no specific reason to suggest that Christchurch was founded any later. The marginal site at Christchurch may have resembled a low island rising from the surrounding marshes in the 8th century, a situation favoured by other early monasteries in South West Britain (Muchelney and Glastonbury for example). Keen has also proposed a large parochia for a pre-conquest Minster at Christchurch, including the later parishes of Christchurch, Holdenhurst, Ringwood, Sopley, Milton, Hordle, Milford, Lymington, Boldre, Brockenhurst, the larger part of the parish of Beaulieu, and parts of Harbridge and Ellingham. Apart from Beaulieu, Harbridge and perhaps Ringwood and Ellingham, the churches of all these parishes remained as dependent churches of Christchurch priory throughout the 12th century. Keen further suggests that the extent of a royal estate would have been identical with the parochia and that a significant population already existed at Christchurch by the early 9 th century (Keen, 1984, 227-8; 242). The juxtaposition of early Minster and royal residence is 92 repeated elsewhere (Cheddar, for example). The origins of this estate and its central place are not known. Christchurch (Twynam) is listed in the burghal hideage, clearly indicating that it was considered a town by the early 10th century. The modern name of Christchurch is relatively recent, being first recorded in 1177 and refers to the Priory, which was founded circa 1150 (Penn, 1980, 38). Early forms of the name include Chrischarche de Twenham and Christeschirche. Before the foundation of the Priory, the town had been known variously as; Tweoxneam, Tweonea (10th century) and Thuinham (11th century) (Page, ed. 1912, 83). This name is generally accepted to mean ‘place between the rivers’ derived from the Old English betweonan + ea (dative plural eam) (Penn, 1980, 38; Mills, 1991, 80). The topography of the region comprises the river valleys of the Stour and the Avon, which empty into Christchurch Harbour. The town itself occupies a narrow low gravel ridge between the two rivers and immediately north of the harbour. The lower courses of the rivers are low-lying and regularly flooded. Marshland still dominates this part of the region, notably at Stanpit marsh and Barlins. Either side of the central vale, the ground slopes gently up to the New Forest on the east and the former heaths, now developed for the urban area of Bournemouth, to the west. The Bournemouth heaths and the New Forest both represented vast tracts of unenclosed marginal land as late as the late 19th century incorporating large areas of common. However, between the river valley marshes and the unenclosed heaths and forests, is a narrow strip of cultivable land historically occupied by dispersed settlement and organised into tithings of Christchurch Hundred. Field Systems The hundred of Christchurch passed to the Malmesbury estate during the 18 th century and a map was produced depicting the various tithings and the land occupied by the estate tenants (figure 5). This map, held at Hampshire Record Office (HRO 9M73-139), provides a rare accurate depiction of land division in the region prior to the late 19th century parliamentary enclosure. A tracing of the map (figure 6) shows more clearly the late 18th century field boundaries and a number of interesting field patterns. These include oval enclosures at Street and Nea. The area now known as ‘Purewell’ was originally called ‘Street’. The settlement of La Stret formed part of the manor of Somerford and belonged to Christchurch Priory before AD 1150 (Page, 1912, 83-101). The settlement appears to be accompanied by or contained within a large curved enclosure, delineated in the modern townscape by the line of The Buttery, a ditch north of Scotts Hill Lane. This ditch may represent a primary boundary feature of the settlement serving to drain the low-lying marsh as well as enclose maximum area for minimum length of boundary. The curving shape of the ditch suggests that there had been few pre-existing restrictions to enclosure at that time. Field boundaries suggest that a curving drainage ditch was also constructed on the south side of the settlement. This has been modified through the cultivation of furlongs with reverse ‘s’ curves during the later medieval period. This and other reclamation works in the marsh to the south of the settlement may have been carried out under the auspices of the Priory. 93 94 Immediately east of Nea there is a large detached portion of Milton parish with boundaries that approximate to an oval. This contained a single large ornamental villa in 1796 but may have originated as a 1-hide unit enclosed from waste during the early medieval period. The two enclosures outlined above lie on the fringes of the band of cultivable land within former marsh and forest respectively. The most pertinent aspect of the map for this current paper however, is the depiction of rectilinear field alignments arranged into tithings running perpendicular to the corresponding river. That is, those on the east side of the hundred have long axes perpendicular to the river Avon, and those on the west run perpendicular to the river Stour. Furthermore, the regular land division only extends in the strip of land between the floodable marshes and the unenclosed heaths and forest. Parliamentary enclosures came late to the Christchurch region, bitterly opposed and resisted by the townspeople. The Portfield in Christchurch was not finally enclosed until the late 19th century. Other common fields at Stanpit and Street were also enclosed during the 19 th century. The 1796 map then, depicts a pre-enclosure landscape, demonstrating that the regular rectangular fields are not the result of parliamentary enclosures. Unfortunately, there is no further independent dating evidence for this field system. Nevertheless, the fact that tithing boundaries follow the same alignments suggests that they be medieval or earlier in origin. Similar field systems in Horethorne Hundred and East Anglia have a pre-Roman origin. It is purely conjectural, but in Christchurch, the proximity of Hengistbury Head might provide an economic context for pre-Roman land division in the region. Discussion The discovery of a previously unrecorded system of co-axial land division, fossilised in the modern landscape of the Christchurch region is significant. Further similar field systems can be identified in discrete areas throughout Somerset and Dorset running perpendicular to chalk and limestone scarps and valleys. In some respects the question of origins is irrelevant for the purposes of this paper. The Christchurch field system was in existence by the medieval period and it is the distribution of medieval dispersed settlement within it that will be considered here. Within the limited area covered by the 1796 map it is clear that the medieval settlements are regularly spaced, both running up the slope to the east and along the strike of the valley to the north. Thus we get columns of settlement sites; Stanpit - Bure- Nea; Street - SomerfordHubborne; and Staple Cross- Roeshott Farm. It is unclear as to whether settlements at Mudeford and Highcliffe existed during the medieval period. Similarly, on the west side of the region, the medieval settlements are also regularly spaced in pairs; Tuckton and Southbourne; Iford and Pokesdown. Furthermore, figure 7 (traced from the OS 1st edition 6 inch series maps) shows that this settlement pattern extended beyond the confines of the four tithings and continues up the respective river valleys to the north and west. It is also apparent that the lanes and tracks linking the settlements are regularly arranged into rectangular blocks. This is particularly clear on the eastern slopes of the Avon Valley. The larger nucleated villages lie in a string along the course of the rivers themselves. Sopley, Winkton, Burton, Wick, Tuckton, Iford and Holdenhurst. Behind these larger settlements are a series of regularly spaced 95 hamlets and farmsteads. Thus we have a three tiered settlement pattern; one row alongside the Figure 7: Settlement and Communications Patterns in the Christchurch region (from the OS 1 st edition 6 inch maps, 1870). river, the second at regular intervals in a row behind (The Bockhamptons, Waterditch Farm, Somerford and Bure) and a third group on the edge of the forest (Bransgore, Neacroft,Godwinscroft,Waterditch, Roeshot Farm, Hubborne and Nea). The Bockhamptons (North, Middle and South Bockhampton) form a particularly interesting group in that these dispersed settlements do not lie at nodes within the communications pattern, but rather along an isolated stretch of track running perpendicular to the river. These observations on settlement, field and communications patterns within a regular rectilinear arrangement cry out for an integrated model of explanation. Towards a new Model Figure 8 shows the generally accepted landscape model of dispersed settlement. The settlement itself lies at the centre of its land unit. Arable or the most intensely cultivated infields lie close to the settlement with unenclosed or less intensely cultivated pasture around it. Communications patterns connect the settlement with its regularly spaced neighbours in a radial pattern. This geographical model owes much to Christaller’s central place theory. However, Christaller’s model was chronologically static and assumed that settlement and associated land 96 Figure 8: A rural economic model (after Christaller) units occupied a previously blank canvas. In other words there was no system of settlement or land division preceding them, casting a residual influence on the pattern. This is clearly not the case for medieval dispersed settlement within a pre-existing system of Roman or late prehistoric land division and communication. A more appropriate model for the Horethorne and Christchurch Hundreds is forthcoming from geographical studies of the Dutch Polders described by Chisolm (1968, 62). Here vast tracts of flat land have been reclaimed from the sea over the last few hundred years or more. It was found that the most efficient way of dividing the land into farm units was in a rectilinear pattern similar to a co-axial field system. When considering the position of farmsteads within these land units a number of different options were available, illustrated in figure 9. In 9b the farms are placed at the centre of their land holdings. This would have made it easy to cultivate the land, but unfortunately the farm would have been further from the major communications routes making it more expensive to transport produce or obtain supplies. The farmhouse would also have been a long distance from its nearest neighbour should co-operation be necessary, at harvest or in a crisis. The second alternative (figure 9c) represents nucleated settlement. Here the advantages of communication and co-operation between neighbours are paramount. The main disadvantage is that each farmstead lies at such a distance from its cultivable land to render production in the farthest corners unsustainable. A successful rural economy is dependent on a balance between settlement, field system and communication. Figure 9d illustrates just such a balance, or compromise between the conflicting requirements of rural production, co-operation and distribution. On the one hand, the settlements are relatively close to the centre of their land unit, whilst on the other they are close to a linking track. Furthermore, the settlement is paired with a neighbour allowing for co-operation and pooling of resources in times of need. 97 Figure 9: A new model for the distribution of dispersed settlement within rectilinear field systems (after Chisholm 1968) Nucleated villages and markets tend to lie close to the major communication routes, which are represented by significant watercourses during the medieval period or, as in the case of Horethorne Hundred, a major road. A tithing, manor or parish extends back from the nucleated settlement, with paired dispersed farmsteads evenly spaced and strung out along a lane running perpendicularly back from the river. It is notable in the Bockhampton group of hamlets, that the settlement has been dragged away from the central point in their respective rectangular blocks, towards the nearest nucleated settlement. This is explained under the above model by the economic pull of the nearest market and major communication artery (The River Avon). This location provides for easier distribution of their produce. Conclusions The model outlined above is essentially an economic or geographic one. It seeks to explain the interrelationship between settlements, communications and Field systems in terms of their function and their position in the landscape. As such it is only one side of the coin, it does not explain land holding in terms of social relationships or historical events. Furthermore it is only relevant in a specific physical landscape context. Co-axial or rectilinear field systems that have survived into the modern landscape tend to be located on relatively gentle and consistent gradients, where a wide river valley, or series of parallel valleys, displaying broad consistency of slope over a considerable distance. The model also remains to be tested and refined, not only in the social sphere, but also within its narrow economic remit and for each major period in which rectilinear field systems are in use. Nevertheless, the model represents an important step forward 98 in understanding the economic processes that have contributed to the shaping of the landscape over the last two millenia. Work in Charlton Horethorne and the wider region of south east Somerset has shown that the survival of rectilinear communications patterns is dependent upon the survival of dispersed settlement. Where nucleated settlement arises in the Late Saxon period or later, then radial communications patterns tend to follow (Davey 2005, 102). Nevertheless, dispersed settlement patterns can be considered the norm over large parts of the country for the majority of the last two millennia. Patterns of production and redistribution of produce may vary however, both chronologically as well as spatially. Yet, in the Christchurch region, this variation may have been minimal. Hengistbury Head represented an important market or redistribution centre during the Late Iron Age and early Romano-British periods. By the Middle Saxon period this regional market had shifted to the royal caput of Christchurch. Both central places were situated on the edge of Christchurch harbour, close to the confluence of the rivers Avon and Stour. This suggests that not only has land division and dispersed settlement patterns remained stable in this region since the late prehistoric period, but the major routes of communication and distribution have also. Under these circumstances it is reasonable to suggest that the rural economic model outlined above may also have remained relevant since the Late Iron Age. In some respects, this model represents a refinement of that put forward by Williamson in order to explain the distribution of rectilinear land division. His suggestion that the land was divided in the late prehistoric period into territories, each with a share of lowland meadow, lower slope arable and upper valley pasture and forest, may still hold for the late prehistoric period (Williamson 2003, 40). It may have been as early as the Romano-British period however, that the land was further subdivided into individual and regularly arranged farmsteads for the purpose of intensification of production. Peterson has argued for Williamson’s co-axial alignments in south Norfolk being formed through a process of centuriation during the 1st century AD. He reaches this controversial conclusion through the computer-aided analysis of field boundaries (Peterson 2004). Centuriation was a method of planned land division in Roman Colonia of the 1st century AD. The land units thus divided tended to be square with sides of 10 or 20 acta (20 acta approximates to 710m). Very little evidence for square units of this size has ever been recovered in Britain, leading to suggestions that centuriation was not employed in Britannia. However, it may be worth noting that the lengths of the rectangular units suggested to the east of the River Avon in the Christchurch region vary from approximately 680m to 1230m, with those on the upper slopes being consistently closer to the 700m mark. These casual observations do not constitute good evidence for centuriation in the Christchurch region. Nevertheless, based on the excavated evidence from Horethorne Hundred, it is clear that the use of these co-axial systems were intensified and modified during the Romano-British period. It may be the pre-existence of such systems of late prehistoric land division that precluded the need impose a rigid system of centuriation in Britain during the 1st century AD. We have seen in Horethorne Hundred that Charlton Horethorne represents an isolated example of a parish in which a single nucleated village has arisen at the expense of dispersed settlements. Elsewhere in the hundred the regular pattern of dispersed settlements, linked by a rectilinear system of lanes and fields, has survived. One particular settlement at Holway lies within a rectangular 1-hide unit defined by parish boundaries. Bede mentions the hide in the 7th century as 99 being the unit of land required for the support a family. Charles-Edwards also maintains that it was the land required to maintain the status of a normal freeman, ploughed by 1 plough with origins earlier than the Saxon kings who used it as the basic unit of taxation (Charles-Edwards, 1972, 4-5). Barnwell (1996) goes as far as to suggest that the hide represents a continuation, under a different name, of the Roman centuria. The survival of such land units possibly dating from the creation of the rectilinear field system is plausible in an integrated rural economy in which field boundaries, settlement and communication patterns have also survived. It suggests a remarkable degree of continuity, not only of field system but land tenure and social organisation as well. Furthermore, the creation of a couple of oval land units at Street and Nea, demonstrates that the usefulness of the hide as the basic unit of land holding and taxation continued into the early medieval period. Because the well-drained land was already subdivided and cultivated however, new units were enclosed from the former waste on marginal marsh or forest and with curved boundaries reflecting the lack of pre-existing restrictions to enclosure. The rural economic model presented in this paper requires a certain level of agricultural specialisation for farmsteads during the Romano-British and medieval periods. This is because the broad range of resources distributed across the valley slope as a whole during the Late Iron Age were no longer available within subdivided 1-hide units. In very general terms, those on the lower slopes might specialise in arable or dairy with those on the upper slopes more dependent on pastoral regimes. Variations in the underlying geology and soils will also affect the agrarian regime within individual farms. This specialisation may be reflected in the names of the settlements east of the River Avon in Christchurch Hundred. Burton and Bure might suggest arable farms on the lower slopes; the Bockhamptons suggest proximity of woodland or game, boc meaning beech tree or bucca meaning male goat or deer (Gelling 2000, 222). Neacroft and Godwinscroft are suggestive of small upper slope farmsteads on the edge of common pasture. The remaining place names are largely topographical in origin. Perhaps of greater significance for the early medieval period is the insight these field systems provide into the rural economy during the Roman to medieval transition. Firstly, although it is not claimed that the field boundaries depicted on 18th and 19th century maps represent a direct continuity of specific fields from the Roman period, it is clear that there has not been a major upheaval in land use during that time. Large areas of farmland were not abandoned to woodland or unenclosed waste. There were not vast tracts of regenerated forest or virgin land into which the Saxons could insert their open fields and nucleated villages. There may have been a reduction in the intensity of rural production, from arable to a largely pastoral economy in many places. This would have been most effectively achieved through the continued use of already established field boundaries. The corralling of livestock would be facilitated in this way, enabling the farmer to control the consumption of pasture, only allowing his stock on to particular fields in rotation when the grass had regenerated. Small former arable fields linked by hedged lanes are ideal for this purpose. The study of the early medieval period has been plagued by culture-historical approaches in which archaeological evidence has been made to fit pre-existing notions concerning the ‘Dark Ages’ for which dubious and non-contemporary historical documents have provided the source material. The lack of good archaeological evidence, spanning the Roman and early medieval periods, has exacerbated the problem. The high visibility of Roman settlements contrasts starkly with elusive early medieval sites. Esmonde Cleary (2001) has called for new strategies to 100 overcome this material discontinuity. One such strategy is to examine sites or classes of monument that span the two periods. This paper has shown the potential for land division to bridge the gap between the Roman and early medieval periods. The evidence suggests that despite the reduction in durable material culture following the collapse of Roman administration in Britain, the rural economy continued in a very healthy condition. In fact it may have been the health of the rural economy that facilitated, in the late 10th century, the rapid rise of Alfredian Wessex, the construction of burhs, the maintenance of herepaths and the subsequent rise in sociopolitical complexity associated with the unification of England. References Griffiths D. (ed.) 1996 Anglo-Saxon Studies in Archaeology and History 9, Oxford University Committee for Archaeology. Good R. 1966 The Old Roads of Dorset. Bournemouth: Horace Cummin Hase, P.H. 1988 ‘Mother churches of Hampshire’, in Blair (ed), 45-66 Haslam J. (ed) 1984 Anglo Saxon Towns in Southern England. Chichester: Phillimore James S & Millett M (eds.) 2001 Britons and Romans: Advancing an Archaeological Agenda, CBA Res. Rep. 125. Jarvis, K. S. 1983 Excavations in Christchurch 19691980, Dorset Natural History and Archaeological Society, monograph series 5 Keen L. 1984 ‘The Towns of Dorset’, in Haslam J (ed), 203-47. Mills, A. D 1991 A Dictionary of English PlaceNames, Oxford University Press Morton, A 1992 Excavations at Hamwic: Volume 1. CBA Res. Rep. 84 Newman C 1992 ‘A Late Saxon Cemetery at Templecombe’, Proc Somerset Archaeol and Nat Hist Soc 136, 61-72. Newman C. Morris E.L. & Bonner D. 2001 ‘Iron Age and Romano-British Sites Along the Bowden Reservoir Link Pipeline, South-East Somerset’, Proc Somerset Archaeol and Nat Hist Soc 143 (for 1999), 127. Oosthuizen S. 2003 ‘The Roots of Common Fields: Linking Prehistoric and Medieval Field Systems in West Cambridgeshire’ in Landscapes 1, 40-64 Page, W 1912. Victoria County History of Hampshire and the Isle of Wight, London, Constable and Company Penn, K. J. 1980 Historic Towns in Dorset, Dorset Nat Hist and Archaeol Soc, monograph series 1 Peterson J. W. M. 2004 ‘Coaxial landscapes in South Norfolk and beyond: another view’, In The Annual: The Bulletin of the Norfolk Archaeological and Historical Research Group, 13, 61-72. Tabor R (ed). 2002 South Cadbury Environs Project Interim Fieldwork Report, 1998-2001, University of Bristol, Centre for the Historic Environment. Tabor R. (ed.) 2004 South Cadbury Environs Project Interim Fieldwork Report, 2002-2003, University of Bristol, Centre for the Historic Environment. Aston M. 1985 Interpreting the Landscape, Batsford, London. Barnwell P. S. 1996 ‘Hflaeta, ceorl, hid and scir: Celtic Roman or Germanic?’ in Griffiths (ed), 53-61. Charles Edwards T. M. 1972 ‘Kinship, Status and the Origin of the Hide’, Past and Present 56, 3-33. Chisolm M. 1968 Rural Settlement and Land Use. Hutchinson, London. Blair J. (ed) 1988 Minsters and parish churches: The local church in transformation 960 – 1200. Oxford University Committee for Archaeology Monograph 17 Cooper G. 2004 ‘Gradiometer Survey and Excavations at Englands, Charlton Horethorne 2003; Introduction’, in Tabor (ed) 2004, 75-6. Costen M. 1985 ‘Rimpton in Somerset, A Late Saxon Estate’, Southern History 6, 13-24. Davey J. E. 2004 ‘Gradiometer Survey and Excavations at Englands, Charlton Horethorne 2003’, in Tabor (ed) 2004, 77-84. Davey J. E. 2005 The Roman to Medieval Transition in the Region of South Cadbury Castle, Somerset. BAR British Series 399 Davey J. E. 2008 Abbotsbury Draft Historic Urban Characterisation Report. Dorset County Council and English Heritage Davey J. E. 2009a Stalbridge Draft Historic Urban Characterisation Report. Dorset County Council and English Heritage Davey J. E. 2009b Sturminster Newton Draft Historic Urban Characterisation Report. Dorset County Council and English Heritage Dunning R. W. (ed.) 1999 A History of the County of Somerset, vol.7, Institute of Historical Research, Oxford University Press. Esmonde Cleary S. 2001 ‘The Roman to Medieval Transition’, in James S & Millett M (eds.), 90-97. Field J. 1972. English Field Names. David & Charles, Newton Abbott. Fleming A. 1978 ‘The Prehistoric Landscape of Dartmoor. Part 1: South Dartmoor’, in Proc Prehist Soc 44, 97-123. Fleming A. 1989 ‘The Genesis of Co-axial Fields Systems’ in Van der Leeuw & Torrence (eds), 63-81 Gelling M. 2000 Place-Names in the Landscape. London; Phoenix Press 101 Williamson T. 2003 Shaping Medieval Landscapes; Settlement, Society and Environment. Windgather Press, Macclesfield. Wykes R. 2003 Prehistoric Agricultural Landscapes: The Limestone Region of the SW in the Counties of Somerset, West Wiltshire and Gloucestershire, Unpublished PhD thesis, University of Bristol. Taylor C. C. 2004 The Making of Dorset. paperback edition. Wimborne: Dovecote Press Thorn C. & Thorn F. (eds.) 1980 Domesday Book, Somerset, Chichester: Phillimore. Van der Leeuw S. E. & Torrance R. 1989 What’s New: a closer look at the process of innovation. One World Archaeology 14. London: Unwin Hyman Williamson T. 1987. ‘Early Co-axial Field Systems on the East Anglian Boulder Clays’, in Proc Prehist Soc 53, 9-31. 102
© Copyright 2026 Paperzz