Prehistoric Linear Earthworks Reconsidered

PAUL C. TUBB PREHISTORIC LINEAR EARTHWORKS RECONSIDERED

Summary: Extensive fieldwork on the Pewsey Downs of Wiltshire in central southern England has challenged prior classifications and interpretations of linear earthworks. A novel classification of linear earthworks is offered, and their place in the social structure and subsidence system of first millennium BC prehistoric society considered. It is suggested that linear earthworks were complex, long-term structures that changed meaning over time and that simple explanations of their nature, chronology and meaning are insufficient. Consideration is also given to the role of linear earthworks in the creation of special places, particularly the partition of burial mounds from the everyday landscape, and the place of linear earthworks in the origin of Iron Age hillforts.

INTRODUCTION Linear earthworks, normally composed of at least one ditch and bank, represent the largest archaeological structures built in Britain between the completion of Neolithic henges and the construction of Iron Age hillforts. Found on the chalk uplands of England: the Yorkshire Wolds, the North Wessex Downs, the South Downs, Cranborne Chase, and Salisbury Plain, most authorities date their construction to the Late Bronze Age (c.1100 to c.850BC), later, and intrusive into, the extensive field systems covering much of the southern chalklands.

Their

archaeological significance lies in the number and density of linear earthworks in these places, their complexity and diversity of form, the scale of construction, and the range of uses their builders put them to: linear earthworks represent a highly adaptable technology. In one small part of Salisbury Plain covering some 138km², there are 43km of linear earthworks (Williams-Freeman 1934, 232), the product of centuries of construction and alteration, but their nature, purpose, and meaning are poorly understood and too often defined in terms that have changed little since the

first decades of the Twentieth Century.

A robust understanding of the nature,

context and meanings of these earthworks could potentially reveal much about late prehistoric society and subsistence in Britain. This paper originated in extensive fieldwork undertaken as part of the writer’s PhD research into the extent and nature of Late Bronze Age/Early Iron Age (LBA/EIA) (850-650BC) activity in the Vale of Pewsey, Wiltshire, in central southern England (Tubb 2009; Tubb 2011a; Tubb 2011b).

Evidence from this period is

typified by so-called midden or black-earth sites, formed by complex processes of feasting, curation of cultural material, husbandry of large numbers of livestock, and settlement.

Fieldwork identified 40 black-earth sites in and around the Vale of

Pewsey, indicating that exceptional activity took place in the area during the LBA/EIA transition (op. cit.).

Many sites were located on the Lower Chalk fringes of the Vale

and were associated with field systems and linear earthworks. These integrated linear/field systems were oriented perpendicular to the axis of the Vale suggesting some process for the equitable division of resources in the Vale and on the adjacent Downs had taken place. Fieldwork indicated that existing interpretations of linear ditch-and-bank earthworks were insufficient, and their complexity necessitated both an extensive survey of the earthworks in the study area, and a re-examination of the historiography of the topic.

It became clear that linear earthworks were a

sophisticated, flexible technology that performed a variety of roles including the definition of spaces, the facilitation of movement through those spaces, and conspicuous display.

This complexity was mirrored by a wide range of

constructional forms, requiring the formulation of a novel classification scheme.

PAST FIELDWORK Colt Hoare (1811, 1821) considered many linears to be associated with movement or control of livestock. Sumner (1913), the Curwens (1918) and Clay (1927) likewise saw linears, especially cross-dykes, as devices for the herding of livestock.

Excavations and fieldwork on linear earthworks by JFS Stone (1937),

Collin Bowen (1978) and others, in the years to either side of the Second World War,

interpreted these earthworks as ranch boundaries, the interpretation still found on the National Monument Record maintained by English Heritage. The entry for the Old Ditch linear in Tilshead (NMR SU 04 NW 25) on Salisbury Plain states ‘This linear earthwork is clearly a ranch boundary’. After the Second World War, the networks of linear earthworks found on the chalklands of central southern England were termed Wessex Ditches. Bowen (1990) undertook an extensive survey of the archaeological remains of Bokerley Dyke and its environs, transcribing and analysing extensive stretches of linear earthwork on the Hampshire/Dorset border. Bradley et al. (1994) excavated a series of sections across linear earthworks in the vicinity of Sidbury Hill, Wiltshire, in an attempt to understand the relationship between linear ditches and the extensive traces of Middle and Late Bronze Age activity found in the area. Based loosely on the ceramic evidence, it was argued that the linears delineated a series of territories. Subsequent to the fieldwork of Bradley et al., Llobera (1996) used GIS and digital terrain modelling to analyse the relationship between the linear earthworks around Sidbury, and was unable to substantiate Bradley et al.’s interpretation. Giles (2000) re-evaluated the extensive linear earthworks of the Yorkshire Wolds, challenging a number of concepts, including the idea that linear earthworks represented geo-political or territorial boundaries.

THE DATING OF LINEARS As Bowen (1990, 12-13) noted, the dating evidence for linears is ambiguous. Crawford (1924) argued, based on aerial photographic evidence, that the construction of these earthworks slighted earlier field systems, implying that the subsistence economy of chalkland areas shifted from an arable to pastoral basis at some point before the Iron Age. The composite linear (q.v.) on All Cannings Down (NMR SU 06 NE 33) has field systems situated on either side of a straight section located on a spur to the north of Tan Hill. This section terminates in a right-angled turn to the north-west incorporating a round barrow (NMR SU 06 NE 19) into the linear. The linear, west of this sharp deviation, forms the northern edge of one of those field systems, a relationship that strongly suggests the linear and the field

system are contemporary, or the field system postdates the construction of the linear, and challenging the view that field systems predate linears and tend to be slighted by them (McOmish et al. 2002, 53 & 61). McOmish et al.’s interpretation contrasts with Bowen’s earlier work on Cranborne Chase where he states (1990, 12) that a number of linears do integrate into field systems. The results of the Wessex Linears Ditch Project showed that linear earthworks found on Salisbury Plain had long histories incorporating sporadic phases of re-cutting and remodelling dating from the Middle Bronze Age through to the Middle Iron Age (Bradley et al. 1994, 149-50). These events suggest that, to contemporary groups, the meaning of the structure changed over time.

Some linears, of course, originate in the Historic

period: the substantial Sub-Roman earthwork, the Wansdyke, cuts through and incorporates a number of prehistoric linears to the north of Tan Hill (q.v.).

THE MORPHOLOGY OF LINEARS Evidence from the Pewsey Downs has indicated a marked variety in earthwork morphology. On Salisbury Plain, McOmish et al. (2002, 58) identified three principal forms of earthwork: lengths of ditch with or without an accompanying single bank; ‘parallel embanked linear ditches with a medial bank’ and multiple ditched and banked linears. This classification is based on Bowen’s volume on Bokerley Dyke and environs (Bowen 1990, 10), although it has been modified. It became clear from fieldwork that this tripartite division was not sufficient to explain the variety of linears encountered.

The following types of earthwork are found on the Pewsey Downs (See Appendix 1): Cross-dykes These are restricted to locations where a topography of very steep scarp slopes and narrow spurs occurs. This earthwork type is consequently absent from the relatively gentle slopes found on the northern edge of Salisbury Plain but present in considerable numbers on the steep, high scarp slopes of the Pewsey Downs. This form of earthwork has attracted the attention of antiquarians and archaeologists for more than two centuries. Clay (1927, 61) termed these cattle ways, defining

them as a ditch between two banks running a perfectly straight course and joining the heads of two coombs by crossing the intervening ridge of downland in order to drive cattle from one to the other. Erroneously, he cited the observations of Colt Hoare (1811, 1821), whose covered ways were interpreted as the means for moving livestock. Colt Hoare interpreted cross-dykes as defensive works, applying the term covered way to linears of some length and found in places untypical for cross-dykes. Heywood Sumner, studying the cross-dykes on Fore Top, Fontmell Down, Dorset (Sumner 1913, 67) reached the same conclusion as Colt Hoare. Clay’s work was prompted by the findings of the Curwens in Sussex (Curwen & Curwen 1918), who had studied 16 linears, comprising either a single ditch with a bank on both sides or composed of a series of banks and ditches running roughly parallel to each other, and wrongly applied Colt Hoare’s collective term to them. Williams-Freeman (1932) identified two cross-dyke forms: a small number of large, probably defensive, earthworks and a more numerous class of ‘small’ and ‘medium’ cross-dykes (ibid., 25).

He refuted the Curwens’ interpretation, arguing many smaller cross-dykes

formed cattle compounds (ibid., 32). Recently, Tilley (2010, 182-6) has argued that cross-dykes linked coombs, codifying space and emphasising topographical places significant to the groups that constructed them. He has suggested they may form processional ways, although the steepness of associated slopes, and the placing of the ditch terminals of those dykes studied on the Pewsey Downs, makes this interpretation unlikely.

The cross-dykes located on the Pewsey Downs share two features: they have a single ditch of varying dimensions and at least one of the ditch terminals is found on the brink of, or part way down, a steep scarp slope. In many instances, both terminals are located on steep slopes. The massive earthwork on the northern slope of Martinsell Hill forms a cross-dyke writ large, with its western terminal on the brink of a precipitous slope above Rainscombe and its eastern end terminating above an equally steep slope at Clench. The size of cross-dyke banks varies, some being very slight, others substantial. There is also variation in the number of banks with some double and others single. The position of the single or largest bank varies,

with some placed on the uphill side of the ditch and others down-slope. This variety suggests that whilst certain constructional and topographical characteristics were common, the purpose and meaning of these earthworks varied considerably.

Spinal Linears This term is used in a different sense to that of Bowen (1990, 11), who used it to describe very long linears, often more than 5 km in length. A spinal linear in this study represents a section of earthwork, normally more than 500m in length, and composed of a ditch and bank, the morphology of which remains unchanged over that distance. The morphological definition of this linear type is critical because other forms of linear are of similar length but differ in their structure, e.g. field system edges. Spinal linears separate field systems or areas of different land use but also represent longitudinal processes e.g. walking to or from a settlement or driving cattle up from the Vale floor to the Downs. These are complex earthworks and often form part of a larger structure: the earthwork running north to south across All Cannings Down incorporates a settlement and apparently had a further linear earthwork added to its northern terminal at some stage (Tubb forthcoming a).

Field system edges This form of linear is novel in its definition. Composed solely of a lynchet, these earthworks are found at the top edge of a hill-slope. The lynchet appears not to be formed by initial digging and subsequent plough action (Bowen 1961, 15), but by deliberate, large-scale digging-out of the slope to form a substantial scarp following the contour line, with spoil being dumped on top of the scarp to increase its height. The results of this process are best seen on the south-western section of the field system edge on the northern slope of Golden Ball Hill, Alton, where a substantial lynchet scarp was dug along the top of a steep slope. Uphill of the scarp, a flat area some 4m wide, seemingly composed of spoil from the digging of the scarp, extends the length of the lynchet forming the edge of a prehistoric field system. Similar structures exist on Draycott Hill, Easton Hill, and at Burbage Down where the parish boundary follows the route of the lynchet.

Many of these earthworks are associated with dewponds. The earthwork on the southern edge of Draycott Hill has dewponds at both eastern and western terminals, and another close to the deviation in its course adjacent to an inconspicuous round barrow (NGR SU 1311 6385) (Tubb forthcoming a).

The

lynchet of the field system edge on the southern slope of Golden Ball Hill originates next to the Pit Farm dewpond and another dewpond is located at its north-eastern terminal. Both the Burbage Down and Easton Hill East, Easton Royal, field system edges run down slope as they extend south and terminate close to Falstone Pond, a long established water source.

These lynchets suggest an integrated mixed agricultural subsistence system at some point in the later prehistoric period, probably contemporaneous with the LBA and IA settlement activity on Golden Ball Hill, Gopher Wood and Easton Hill. The topography of the lynchets suggests field systems on the hilltops, at times adjacent to settlements, and grazing in the valleys.

The lynchets in the dry valley

approaching Golden Ball Hill and the Pit Farm dew pond from the north-east form a funnel and a similar arrangement is present to the east of Easton Hill. In both instances a dewpond is located at the narrowest point between the linears.

Composite linears Composite linears equate to Bowen’s spinal linears (1990, 11) in that that they cross long distances and deviate over their course. How such lengthy, rambling structures came into existence is difficult to reconstruct without excavation, and as Bradley et al.’s (1994) work showed, excavation does not always render a clear picture of events. These large, complex earthworks represent the final stage of a long process of accretion, with some composite linears originating as shorter, isolated structures that were eventually extended and connected as their meaning and purpose changed or was renegotiated over time. These large earthworks were significant in the prehistoric period, as the work of McOmish et al. (2002, 56-66) and Giles (2007, 106-14) shows, but there is insufficient detailed fieldwork to extend our

understanding of these complex structures.

Too often, theories are developed

simply on the basis of aerial photographic evidence alone (Kirkham 2005). If we accept that composite linears had long, complex histories, assertions that they represent the boundaries of ‘territories’ (ibid., 149) are inadequate.

Composite

linears form the sides of field systems on the Pewsey Downs, for example, the linear running along the northern edge of the summit of Golden Ball Hill forms a boundary in the form of a lynchet to field systems at its south-western and north eastern ends. In that sense, linears comprise edges and create defined spaces, but whether those edges ever represented anything more than an indication of change in land-use is debatable.

Horizontal linears These anomalous linears have no identifiable comparables in the Vale and its environs and seem to be restricted to the scarp slope of Tan Hill and surroundings. Six examples found at or near Tan Hill divide into two groups: those associated with linears and transverse tracks on the eastern slopes, and two isolated examples found on the southern slope. These linears may act as attention-focussing devices: the white chalk of the earthworks contrasting starkly with the turf of the hillside. The three located at the junction of Tan Hill and Milk Hill are located to either side of what appears to be a very long-established track from Stanton St. Bernard to East Kennett, whilst the two below the summit of Tan Hill are situated on the steepest slopes of the hill.

Simple bank and ditch linears These most basic of earthworks, a ditch and bank, often placed on the downhill side of the ditch, and commonly no more than 4m in total width, are frequently overlooked.

The course of such earthworks can mark the boundary

between parishes, as in the case on the Giant’s Grave spur at Martinsell Hill. In uncultivated areas, or in ancient woodland such as Withy Copse or Gopher Wood, this type of linear abounds, suggesting this form had a long history of currency, making any attempt at dating very difficult.

THE PURPOSE OF LINEARS Boundaries Linears on the Pewsey Downs and elsewhere tend to be described as boundaries (Hawkes 1939, 142-51; Bowen 1990, 11; McOmish et al. 2002, 64-5). Explicitly associated with this notion of linears as boundaries, is the concept of ‘territory’: where major linears define the limits of individual geo-political entities (Bradley et al. 1994, Chap. 7; McOmish et al. 2002, 65). Recent re-assessments of linear earthworks (Giles 2000; 2007) have rejected the notion of boundaries in any social or territorial sense (Giles 2007, 108) during the Early Iron Age. Giles (op. cit.) argues that, in East Yorkshire at this time, there is no evidence of settlement hierarchy and, therefore, the concept of linears as territorial boundaries becomes problematical to sustain.

Rather than existing within an established settlement

hierarchy, groups on the Yorkshire Wolds moved between enclosed and unenclosed sites at different times of the year. Wickstead (2007), in her analysis of the networks of Bronze Age reaves on Dartmoor, criticised archaeologists’ tendency to perceive evidence of territorialisation in the past.

She argued that this stemmed from a

modern preoccupation with nationalism. Wickstead (2007, 247) suggested that, in prehistory, the idea of tenure in relationships with land is more appropriate than to consider simply the division into territory and ownership blocks.

Although the territorial interpretation of linear earthworks has been refuted, linears did frequently form edges to parcels of land; they formed functional rather than political boundaries. Linears define field systems at Golden Ball Hill, Draycott Hill and Tan Hill East. The Golden Ball Hill and Draycott Hill group are deeply cut lynchets following the contour of the slope and may have been stock-proof field boundaries. The provision of a hedge on the top of the lynchet would be necessary and, although evidence of prehistoric hedging is very uncommon (Pollard 1996, 108), it seems likely many linears had such a barrier. Personal observation has shown that livestock are reluctant to negotiate steep slopes, preferring to be driven along the earthwork. These field system edges, however, were not hermetic; there

is no evidence of any permanent boundary running at right angles connecting the two linears on Golden Ball Hill, although a hedge, fence, palisade or abatis established some three millennia ago would leave no surviving trace on the land surface. Field system edges were intended not only to keep livestock out of the fields at certain times but also to enclose them within the fields at others. The open ‘ends’ of the field systems resonate strongly with Wickstead’s (2007) emphasis on the Dartmoor Reaves as flexible and inclusive. The size and nature of these field edge earthworks served to indicate changes in land use and meaning but not territoriality.

It seems likely that the groups active in the Vale had established a system of land utilisation that was, at least partly, equitable, through precedent, negotiation, conflict or a combination of all three. Godelier (1978, 400) argued that concepts of property need not be exclusive, incorporating ideas of access, utilisation, control and transfer into norms and customs accepted by a number of adjacent groups. If linear earthworks served as a metaphor for the social relations current at their construction, and later restructurings (Giles 2007, 110-1) solidified those relations, then these relations would have, in part, determined the way individuals perceived and utilised the landscape. Linears provided a framework, reinforced by the presence of other monuments and natural phenomena, for the daily, seasonal and annual cycle of activities.

Cross-dykes as barriers Conversely, cross-dykes are linears with impermeable characteristics that created enclosed areas. These earthworks were constructed to bar, either physically, symbolically or both, or, at least, inhibit access. A small minority of crossdykes possess an entrance; the Giant’s Grave cross-dyke is pierced by a narrow entrance, facilitating access to the spur end settlement site (Tubb 2009). Crossdykes vary greatly in size from the insubstantial but lengthy earthwork on the south-

western spur of Milk Hill, to the large, imposing structure on the eastern flank of Tan Hill. All seem to be built with the express intention of preventing access to, or at least defining, a specific place.

The deliberate siting of cross-dyke terminals on very steep slopes is significant. It is likely, because of subsequent weathering, that the slopes below the scarp edge were even steeper at the time of the construction of the cross-dyke, making it difficult to circumvent the terminals. Additionally, a number of these terminals are located where hills change direction, often the steepest section of slope. This practical difficulty in getting around the ends of linears may have transformed into a cultural taboo and cross-dykes became symbolic, as well as physical, barriers.

If the presence of a cross-dyke on a spur represented a special enclosed place, the combination of cross-dykes and precipitous slopes flanking an entire hill created an unusually large space. Martinsell, Milk Hill and Tan Hill witnessed this process; at Martinsell, the large cross-dyke on the northern slope denied access to the hill from the north, whilst the steep slopes to east and west performed the same function. Access to the large, flat summit area was achieved by the transverse track up the southern slope of Martinsell emerging at the Hassocks black-earth site (Tubb 2009). The placing of the Giant’s Grave cross-dyke close to the end of the spur was intended to prevent access to the hilltop from the WSW and the significance of the boundary was, at a later point, reversed, with the space behind the cross-dyke when viewed from the Giant’s Grave ridge, becoming important as a settlement.

The topography of Milk Hill, the highest point in Wiltshire at 295m OD, like Martinsell and Tan Hill, is a combination of steep slopes and more gently rising spurs that provide access to the summit. This large, flat, roughly rectangular area of some 25 hectares is fringed by a series of prehistoric settlement sites associated with dewponds (Tubb 2009; Tubb 2011b; Tubb Forthcoming a). The gentle slopes to the south-east, south-west and north-west are blocked by linear earthworks, having at

least one terminal placed on the brink of a precipitous slope. Access to the summit was possible from the Furze Hill dry valley running south from East Kennett, a valley flanked by two composite linears. The southern linear incorporates two in-turning entrances situated in shallow re-entrants on the northern slope of Milk Hill.

The composite linear on the northern slope of Golden Ball Hill, with its very deliberate siting and design, can be seen as a developmental step preceding the full enclosure of sites to form hilltop enclosures. Its builders did not need to construct a complete enclosure: the site was already emphasised by the pattern of converging lynchets to the north-east, a small cross-dyke, field system edges to the south-west and the scarp slope to the south.

At Tan Hill, the broad, slightly domed, summit is surrounded by a combination of steep slopes and carefully placed cross-dykes.

Access to the summit was

achieved by the path running past the All Cannings Cross black-earth site up the saddle between Rybury and Tan Hill, taking the route formalised by the digging of a simple bank and ditch earthwork. What is striking at Martinsell, Milk Hill, Golden Ball Hill and Tan Hill, is the deliberate manipulation of the natural environment combined with judicious use of earthworks to orchestrate access to the summit.

The

sophisticated manipulation of local topography and deliberately sited linear earthworks (some with entrances) created de facto hilltop enclosures where access to the summit/enclosure interior was carefully orchestrated.

Linear earthworks and burial mounds There are two main traditions of burial mound in the study area: Neolithic long mounds (Kitchen Barrow, Bishop’s Cannings; Adam’s Grave, Alton; and Giant’s Grave, Milton Lilbourne are the only three long barrows close to the Vale) and Bronze Age round barrows. Both Adam’s Grave and Giant’s Grave comprise long, trapezoidal earthen mounds with flanking quarry ditches and date from the early Neolithic (Barker 1985). Round barrows are burial mounds associated with the Early and Middle Bronze Age. Peters (2000) argued for the existence of two traditions of

barrow building on Salisbury Plain: conspicuous barrows, large earthen mounds often built in prominent places in the landscape and relatively short-lived in their currency in the Early Bronze Age, and; inconspicuous barrows, a form that was current well into the Middle Bronze Age and comprised small burial mounds often constructed on the fringes of contemporary settlements and field systems. Barrows still formed the focus of mortuary practice in the Middle Bronze Age in the form of secondary burials and adjacent cremation cemeteries but, by the Late Bronze Age, had ceased to be the focus of funerary activity.

Linear earthworks are associated with barrows in two ways: cross-dykes and barrows are found adjacent to each other in a number of instances or alternatively, linear earthworks either incorporate long or round barrows in their course or are aligned on them.

The association of cross-dykes with barrows has received

attention from archaeologists previously. Tilley applied phenomenological theory to the round barrows and cross-dykes of the Nadder-Ebble ridge in southern Wiltshire (Tilley 2004), an area previously investigated by Fowler (1964).

Tilley’s initial

conclusions emphasising the intervisibility and relatedness of round barrows scattered over a landscape and arguing that the location of barrows ‘served to codify important topographic features of the landscape’ (2004, 198), have been reworked and modified in his most recent paper on the topic (Tilley 2010). He argues that cross-dykes are found in groups forming inter-related networks of earthworks (ibid., 101) defining special places but, contradicting his earlier argument (Tilley 2004), were not related to the distribution of barrows on the Ebble-Nadder ridge. The distribution of barrows and cross-dykes in south-west Wiltshire and north-east Dorset does suggest, however, that some dykes are strongly associated with barrows, for example, Winklebury Hill, White Sheet Hill and Burcombe Down (Tilley 2010, Figs 4.2 & 4.3).

These earthworks create spaces around

barrows in a

manner similar to the cross-dykes on the eastern spur of Tan Hill, which form two loosely defined enclosures around the three round barrows located on the saddle of the hill. On the western spur of Tan Hill, five round barrows are enclosed on three sides by a series of cross-dykes. This association parallels the arrangement of

linears on Snail Down (McOmish et al. 2002, Fig. 2.23, 43), where the main barrow cemetery is similarly enclosed on three sides. In both instances the barrow group is ‘open’ to the east and south-east. Whether there was any intention to fully enclose the barrow cemetery is debatable: the presence of the linears was sufficient to denote a special place. Cross dykes tend to be situated on the narrowest part of a spur, set apart from the associated round barrow.

This ‘setting apart’ may be

indicative of the cross-dyke builders’ desire to create a cordon around the earlier burial monument.

Many writers have noted the association between linears other than crossdykes and barrows (Sumner 1913, 76; Field 2001, 60; McOmish et al. 2002, 64; Kirkham 2005, 152-3; Giles 2007, 113; Sharples 2010, 28-32,34). Fieldwork on the Pewsey Downs identified two forms of association between these earthworks: the course of some linear earthworks is aligned on barrows and, occasionally, barrows are incorporated into linear earthworks. The Wilcot 5 linear is aligned between the Gopher Wood dewpond and an inconspicuous barrow some 800m west.

The

courses of the composite linears on All Cannings Down are sighted on round barrows and one earthwork executes a right-angled turn to incorporate a round barrow (NMR SE 06 NE 33). The twin linears on the western spur of Milk Hill run to the east and west respectively of a disc barrow (NMR SU 16 SW 6). Giles (2007) viewed this incorporation of barrows as an attempt to ‘involve the past in the present’ changing the perception of that past and renegotiating movement and access to particular spaces in an increasingly pressured First Millennium BC environment and society (ibid., 111-4). Giles’ analysis is contradicted by evidence from Wiltshire:

the placing of

linears in relation to barrows on Tan Hill, Snail Down and the Nadder-Ebble ridge point to an intent to separate burial mounds from quotidian life. This suggests that by the Late Bronze Age burial mounds were associated with meanings and activities outside the discourses of everyday life, and had become places of myth, superstition and fear. The Stonehenge palisade ditch, a similar structure at Avebury (Pollard

pers. comm.), and other linears in the Stonehenge environs (Sharples 2010, 32) appear to reflect a desire to create cordons around the stone circles and barrow cemeteries. The association with barrows suggests that linear earthworks, by the act of incorporation or sighting, were invested with some of the mythic associations of the burial mounds, perhaps as an act of spiritual consolidation of their significance in the contemporary landscape, and to emphasise their special nature and meaning. In these actions to separate round barrows from the everyday landscape, a developing sense of taboo associated with burial mounds and their associations can be detected in a manner not seen in earlier periods.

The partitioning of barrows and other earlier monuments dates to the Late Bronze and thereafter. Round barrows fell out of use as the foci for cremation burials at the end of the Middle Bronze Age. Brück (1995) noted the presence of fragmented human remains on a number of Late Bronze Age settlement sites, and fragmentary human bone dating from this period has been found with votive deposits in watery places and on boundaries.

This phenomenon continues through the

LBA/EIA transition with partial human remains being recorded at Potterne (Lawson 2000, 95-101) and East Chisenbury (McOmish et al. 2010, 65-6). The presence of fragmented human bone, especially the dominance of skull fragments and long bones, has been interpreted as the result of excarnation and other post-mortem processes and suggests neither inhumation nor cremation burials took place in this period and that Late Bronze Age beliefs about the body and individual were markedly different from those of previous periods (Brück 1995). This change in mortuary practice may be the root cause of the desire to separate burial mounds from the surrounding environment, disassociating past practices from contemporary society as part of a changing mythic landscape.

Contemporary perceptions of

barrows altered, as funerary practices changed, and the partitioning off of burial mounds suggests that they had become places to be avoided. The construction of absolute or de facto boundaries around barrows indicates a desire to minimise contact with these places.

Droveways and Trackways The presence of hollow-ways along sections of the All Cannings Down linears indicates these earthworks served as trackways into the historic period.

Some

linears may have been partially intended for that purpose from inception.

Field

system edges follow the contour between the hilltop and slopes of Golden Ball Hill, Milk Hill/Furze Hill, Draycott Hill and Easton Hill. All gently descend onto lower ground following the the diminishing relief of the upland and many terminate at dewponds. This association, combined with the typical topographical location of this form of linear, is indicative of subsistence practice in the late prehistoric period with the steeper slopes of the chalk downs being utilised as grazing. Giles (2007, 111) notes that cattle require considerably more water than sheep and given the generous provision of dew ponds adjacent to these earthworks, it seems reasonable to assume that a sizable proportion of the grazing livestock were cattle. Although the faunal assemblage from Potterne (Lawson 2000) indicate the proportion of cattle to sheep diminished over time (ibid., 108), a trend also noted by Hambleton (1999, 87) across central southern Britain, there were significant numbers of cattle throughout the occupation of the site.

Perhaps cattle grazed the slopes and

bottoms of the dry valleys (at this period still probably containing spring-fed streams) and were driven to the local water source or adjacent settlement for milking or byreing on a daily basis. Steeply cut lynchet edges prevented cattle from straying onto the hilltop fields and facilitated cattle drives. Eventually cattle would become habituated to this diurnal cycle (Giles 2007, 109) and perform this twice daily ritual with little input from cowherds.

Other linears functioned as parts of long distance trackways. A ditch and bank earthwork leads up the saddle between Rybury and Tan Hill from the All Cannings Cross black-earth site, across the summit of Tan Hill and along the crest of Allington Down, intersecting a number of places, mostly settlements, significant in the LBA/EIA transition. At Gopher Wood linears are positioned in such a way as to emphasise the narrow defile and track located between them leading from the floor

of the Vale of Pewsey to the Kennet valley at Lockeridge. These twin earthworks simultaneously enclose and signpost the defile, at an ambiguous point in the landscape.

The LBA/EIA transition is characterised as a period following the

collapse of long-distance networks (Sharples 2007, 176), however, people and herds probably moved over moderate distances on a regular basis and may have periodically travelled considerable distances within the region.

Attention-focussing devices Linear earthworks were constructed to be visually obtrusive within the landscape. McOmish et al. (2002, 64) note that many linears on Salisbury Plain are placed in positions of prominence. In the study area, many linears were placed on sites where elevation maximised the visual impact of the earthwork; in short, elevation and the resulting visibility mattered to the builders of linear earthworks. Deliberation went into the siting and execution of these earthworks; cross-dykes, for example, are placed where they visually dominate their surroundings. The crossdyke on the northern slope of Martinsell is visible to observers as far away as the scarp crest of the Marlborough Downs some 12.5 km to the north.

These

earthworks also had visual impact when viewed from the side: the cross section of the large cross-dyke to the east of Tan Hill is obvious as a ‘notch’ from Fyfield over 7 km to the north-east.

The construction of lynchets along slopes, for example at Golden Ball Hill, created a visible notch, as well as a linear that was obvious from below.

The

lynchets draw the eye of the observer to the earthwork and the area of land around it. At Golden Ball Hill the low domed hilltop and the activities or structures present on the summit were obstructed by the earthwork.

This combination of drawing

attention to the hilltop, but simultaneously obstructing any view of it, reaches its apogee in the construction of the central section of the composite linear on the northern slope. Here, following a line along the break in slope established by field system edge lynchets, a ditch was dug and a bank thrown up on the uphill side of it.

An entrance, emphasised by the convergence of minor earthworks on the slope below it, was incorporated into the linear at the point where the hilltop rose immediately behind the linear, thereby obstructing any view of the interior of the space through the entrance gap. The earthwork is extremely obvious from the valley bottom and slopes to the north and was intended to draw attention to the

site

present on Golden Ball Hill (N Sharples pers. comm.).

THE MEANING OF LINEARS Linears as corporate monuments Recent discussion of linears has emphasised their corporate nature; they have been interpreted as the products of communal labour. Giles (2007, 110) sees linear earthworks as gang-dug structures and McOmish et al. (2002, 64) argue that the construction and maintenance of linears demanded considerable labour.

Wickstead (2007) has criticised the emphasis archaeologists place on the role of labour increasing attachment to land. Based on Boserup’s (1965) intensification theory, and Locke’s labour theory of property, this approach emphasizes a narrow category of ‘productive’ work excluding many activities important in the formation in the concept of tenure. emphasises

the

way

As an alternative, Wickstead (ibid., Chap. 4 & 247) agricultural

systems

can

incorporate

intensification,

extensification and diversification in an innovative and flexible system.

By

incorporating the ‘non-productive’ activities into this system, Wickstead argues that the formation of the Dartmoor reaves was important in the process of social reproduction, rather than creation of territory through labour. A significant proportion of linears could have been dug by a small number of labourers, perhaps no more than two individuals. A modern analogy is illustrative: for over two years the writer witnessed two dry-stone wallers rebuilding part of the northern boundary of Badminton Park Estate, South Gloucestershire. By the time they had completed

their task, more than 2.2km of wall had been rebuilt. An experienced waller can construct between 3m and 5m of wall per day: the task is complex involving selection and trimming of stones and adherence to a specific profile for the structure. Walling is far more complicated and time consuming than the digging of a simple ditch and dump bank, suggesting many of the linear earthworks with a small to medium profile are probably the work of a small number of people.

These

earthworks may be communal in the sense that ditch-digging requires very little skill and therefore the task could be performed by all, but suggesting that many linears reflect any more than the labour of a handful of individuals may be misleading. This mode of construction may seem monotonous, strenuous, unrewarding, and inefficient, but that is true for the overwhelming majority of tasks undertaken in the late prehistoric period. Perhaps even some large cross-dykes were constructed in this manner and their final size is the result of successive construction events. The stepped, unfinished ditch of Giant’s Grave may be indicative of a limited labour force and the adoption of spit-digging techniques to extract chalk for the construction of the bank.

The tendency of archaeologists to perceive linear earthworks as single-phase constructions, leads them to interpret these structures as the products of communal endeavour. However as excavation (Bradley et al. 1994, 149-50), and this study have shown, it seems that many linears only acquired their final form after a series of constructional events. It is difficult to assess the length of this process, but evidence from the Sidbury North double linear (op. cit.) suggests a series of phases over some 500 years or more. Giles (2007, 106) is right to emphasise the re-dug nature of many linears: what we see today is the eroded final phase of earthworks with long constructional histories.

The evidence suggests that linears accumulated and

changed over lengthy periods of time, perhaps centuries, making any argument that these earthworks represent a communal effort throughout their currency difficult to sustain. Linear earthworks represent the end product of a long process of change and accretion.

Monuments or earthworks? The recognition of only the final phase of a linear earthwork also affects archaeologists’ perception of the nature of these structures. Many linear earthworks were only perceived as monuments in their own right by subsequent generations, such as the Anglo-Saxons, long after their final phase of construction. Given that the origins of these earthworks are found in a period when there is very little emphasis on monumentality, the builders of the first phases of these linears were not intent on constructing monuments in the sense of, say, a hillfort. The subsequent process of accretion, re-cutting and remodelling over centuries, gave these earthworks a form that archaeologists have considered monumental. The final monumental phase of the Sidbury North linear (Bradley et al. 1994, 120) only occurs in the Middle Iron Age and is strongly associated with the construction or alteration of Sidbury hillfort to its final form. The builders of the hillfort monumentalised an existing linear earthwork, mimicking the multiple ditches and banks of the enclosure, symbolically drawing the landscape to the north of the hillfort into its environs.

Within the long currency of linear earthworks, there is evidence of a tendency towards monumentality over time.

The Giant’s Grave cross dyke; the massive

cross-dyke located to the north-west of the LBA/EIA transition site at Hengistbury Head, Dorset (Bushe-Fox 1915) and the central section of the linear on the north slope of Golden Ball Hill, were clearly constructed as monuments. At Martinsell and Tan Hill, the first steps towards the construction of large hilltop enclosures or hillforts in the Early and Middle Iron Ages were being taken. These de facto enclosures are found in close proximity to large, significant, black-earth sites dating from the LBA/EIA transition (Tubb 2009). These represent the concatenation of a number of more communally based, status bearing, competitive activities based around feasting, middening and deposition of metalwork that developed in and around the Vale of Pewsey in the period. The construction of hilltop enclosures or hillforts supplanted the accumulation of cultural material that formed what we now perceive as black-earth sites.

Field systems and linears The accepted view is that linears postdate field systems (Crawford 1924) and tend to slight them, the implication being that the field systems fall out of use when the linear is constructed (Bowen 1990, 12). Bowen et al. (1978), however, were of the opinion that field systems continued in use after the construction of linear ditches, in this case those located on the eastern edge of Cranborne Chase: “Its ditches frequently cross ‘Celtic’ fields in such a way as to indicate that they were however temporarily (emphasis mine), put out of use” (ibid., 149). At times the intrusive nature of linear construction is overstated, for instance Cunliffe’s (2005, 63) interpretation of the relationship between field systems and linears to the north of Sidbury Hill (McOmish et al. 2002, Fig 3.6). Bradley et al. (1994) concluded that many of the ‘Celtic’ fields post-date linears (ibid., 150) although the dating evidence to support this assertion was not robust. The double linear running north from Sidbury cuts through very few field boundaries and, if we accept the developing concept of taboo in association with round barrows in the LBA/EIA transition, linears to the east and west of the Sidbury Double created a cordon around the barrows found on Snail Down and Haxton Down. The extremities of these linears are oriented on the axis of the field systems to minimise the number of fields bisected. The linear approaching Sidbury from the east does cut through a field system but it takes a line that minimises the fragmentation of individual fields. Aerial photographic evidence fails to establish the chronological relationship between field systems and linears, and from Crawford (1924) onwards, generalising assumptions based on uncritical analysis have tended to reinforce interpretations of land use in the LBA/EIA transition that are not appropriate to every circumstance.

The presence of large linears extending unconformably across field systems, may well indicate fields that have either become redundant, or represent a change in the agricultural regime for that locality manifested by a change in status or role for the fields. The construction of linears across field systems may also represent a desire to facilitate the rapid movement of livestock. These two purposes need not be mutually exclusive: if we assume an agricultural cycle where some fields were left

fallow, others ploughed, and a third group provided either high quality grazing or forage, there would be a need to move livestock onto the field system.

If

assumptions about the presence of hedges and banks between these small fields are correct, the movement of livestock would cause degradation to the physical field system structure.

This situation was perhaps initially acceptable but, given the

marked increase in the quantity of livestock, especially sheep, in the LBA/EIA transition and after, some of these large linears represent a response to the need to move larger herds and flocks. On the Salisbury Plain Training Area (McOmish et al. 2002, Fig i.1) and around Danebury (Palmer 1984), many linears have at least one terminal located in an area devoid of field systems, perhaps representing grazing.

In the study area, there are relatively few unconformable linears crossing field systems.

Rather, linears seem to follow the main axis of the field systems

suggesting an integrated system of land use, a pattern noted in the Test Valley around Danebury (Palmer 1984, 67). This suggests that the development of field systems in the Vale and environs was far less piecemeal than those found on Salisbury Plain and the Marlborough Downs.

CONCLUSION It is clear from this study that linear earthworks are far more complex than has previously been thought. They represent a sophisticated, flexible technology that was used to support a range of intentions. Their morphology is complex and diverse and many earthworks are the product of long histories of addition and alteration. It is no longer sufficient to assert that they supplant field systems and that they represent the product of communal labour.

In particular the assertion that they form the

boundaries of late prehistoric territories is debatable. Accordingly their place in the prehistoric landscape of southern Britain must be reassessed. Linear earthworks provide archaeologists with evidence of the changing nature of subsistence in the later prehistoric period towards a complex and responsive mixed agricultural base that facilitated the development of black-earth sites with their abundant midden deposits in the LBA/EIA transition. They also represent longitudinal processes in the

landscape, forming routeways through a densely settled and utilised environment. The association of linears with burial mounds and earlier monuments suggests the meaning of these sites had changed substantially to one where they were excluded from normality. In the creation of de facto enclosures we see the one of the origins of the Iron Age hillfort.

Linear earthworks represent a complex technology of

boundary-making, not in any political sense, but in subtle and nuanced ways that have been lost.

Their remodelling and accretion indicates their longevity and

potency in the prehistoric landscape over at least half a millennium, and their importance to the societies that maintained their currency.

Acknowledgements Grateful thanks to Josh Pollard, Mike McQueen, and Kathy Bragg for commenting on earlier versions of this paper and to Tony Roberts and Mike McQueen consecutively for their invaluable help over seven years of winter fieldwork.

Dept. of Archaeology & Anthropology University of Bristol 43 Woodland Road BRISTOL BS8 1UU

Appendix 1: Table of linear morphology

Type

Description

Cross-dykes

Single ditch running across spur with terminals frequently placed on scarp slopes. At least one associated bank, size and position of which varies

Spinal linears

Field system edge

Composite linear

Ditch and bank earthworks that separate field systems

Large lynchet scarp often situated at break of slope on edge of hill summit

Lengthy structures often incorporating linear earthworks of different morphologies

Example/Plan

Martinsell North linear

All Cannings Down

Alton 2 & 3, Wilcot 8

Alton 2

Simple ditch and bank linears

Composed of a single ditch and bank, often no less than 4.0m

Horizontal linears

Simple ditch and bank linears that run horizontally across the face of steep hill slopes

REFERENCES

BARKER, C. T. 1985:

The Long Mounds of the Avebury Region.

Wilts.

Archaeological and Natural History Magazine 79, 7-38. BOSERUP, E. 1965: The Conditions of Agricultural Growth: The economics of agrarian change under population pressure (London). BOWEN, H. C. 1961: Ancient Fields (London). BOWEN, H. C. 1978: ”Celtic” fields and “ranch” boundaries in Wessex. In Limbery, S. & Evans, J. G. (eds.), The Effect of Man on the Landscape: the Lowland Zone (London: CBA Research Report 21). 115-22. BOWEN, H. C. 1990: The Archaeology of Bokerley Dyke (London). BRADLEY, J., ENTWHISTLE, R. & RAYMOND, F. 1994: Prehistoric Land Divisions on Salisbury Plain (London). BRÜCK, J. 1995: A place for the dead: the role of human remains in Late Bronze Age Britain. Procs. Prehistoric Society 61, 245-277. BUSHE-FOX, J.P. 1915: Excavations at Hengistbury Head, Hampshire, in 1911-12 (Oxford). CLAY, R. C. C. 1927: Some Prehistoric Ways, Antiquity 1, 54-65. COLT HOARE, R. C. 1811: The Ancient History of Wiltshire, Vol. 1 (London). COLT HOARE, R. C. 1821: The Ancient History of Wiltshire, Vol. 2 (London). CRAWFORD, O. G. S. 1924: Air Survey and Archaeology (London). CUNLIFFE, B. 2005: Iron Age Communities. (4th Ed). (London). CURWEN, E. & CURWEN, E. C. 1918:

Covered Ways on the Sussex Downs,

Sussex Archaeological Collections 59, 35-75. FIELD, D. 2001: Place and memory in Bronze Age Wessex in Brück J. (ed.), Bronze Age Landscapes: Tradition and Transformation (Oxford) 57-64. FOWLER, P. J. 1964:

Cross ridge dykes on the Ebble-Nadder ridge, Wilts.

Archaeological and Natural History Magazine 59, 46-57.

GILES, M. 2000: Open-weave, Close-knit: Archaeologies of Identity in the Later Prehistoric Landscapes on the Yorkshire Chalk. Unpubd. PhD thesis, University of Sheffield. GILES, M. 2007: Refiguring rights in the Early Iron Age landscapes of East Yorkshire. In Haselgrove, C. & Pope R. (eds.) The earlier Iron Age in Britain and the near continent (Oxford) 103-18. GODELIER, M. 1978: Territory and property in primitive society, Social Science Information 17, 399-426. HAMBLETON, E. 1999:

Animal Husbandry regimes in Iron Age Britain: a

comparative study of faunal assemblages from British Iron Age Sites (Oxford: BAR Brit. Ser. 282). HAWKES, C. F. C. 1939: The Excavations at Quarley Hill. 1938, Procs. Hants Field Club & Archaeol Soc 14, 136-94. KIRKHAM, G. 2005: Prehistoric linear ditches on the Marlborough Downs. In Brown, G., Field, D. & McOmish, D. (eds.): The Avebury Landscape (Oxford) 149-55. LAWSON, A. J. 2000:

Potterne 1982-5: Animal Husbandry in Later Prehistoric

Wiltshire. Wessex Archaeology Report 17 (Salisbury). LLOBERA, M. 1996:

Exploring the topography of mind: GIS, social space and

archaeology. Antiquity 70, 612-22. MCOMISH, D. 1996: East Chisenbury: ritual and rubbish at the British Bronze AgeEarly Iron Age Transition. Antiquity 70, 68-76. MCOMISH, D., FIELD, D. & BROWN, G. 2002: The Field Archaeology of the Salisbury Plain Training Area (Swindon). PALMER, R. 1984:

Danebury: An Iron Age Hillfort in Hampshire.

An aerial

photographic interpretation of its environs (London). PETERS, F. 2000:

Two Traditions of Bronze Age Burial in the Stonehenge

Landscape. Oxford Journal of Archaeology 19(4), 343-358. POLLARD, J. 1996: Iron Age riverside pit alignments at St Ives, Cambridgeshire, Procs. Prehistoric Society 62, 93-115.

SHARPLES, N. 2007:

Building communities and creating identities in the first

millennium BC in C. Haselgrove & R. Pope (eds.), The earlier Iron Age in Britain and the near continent (Oxford), 174-184. SHARPLES, N. 2010: Social Relations in Later Prehistory (Oxford). STONE, J. F. S. 1937: An enclosure on Boscombe Down East, Wilts. Archaeological and Natural History Magazine 47, 466-489. SUMNER. H. 1913: The Ancient Earthworks of Cranborne Chase (London). TILLEY, C. 2004: Round Barrows and Dykes as Landscape Metaphors. Cambridge Archaeological Journal 14:2, 185-203. TILLEY, C. 2010: Interpreting Landscapes. (California). TUBB, P. C. 2009: The Late Bronze Age-Early Iron Age Transition in the Vale of Pewsey, Wiltshire. Unpubd. PhD dissertation. University of Bristol. TUBB, P. C. 2011a: The East Chisenbury Midden in its regional context. Wilts. Archaeological and Natural History Magazine 104, 44-61. TUBB, P. C., 2011b:. The Late Bronze Age-Early Iron Age Transition in the Vale of Pewsey, Wiltshire ( Oxford: BAR Brit. Ser. 543). TUBB, P. C. Forthcoming a: A late prehistoric landscape on the Pewsey Downs. Wilts. Archaeological and Natural History Magazine 106. TUBB, P.C. Forthcoming b: “A tangled, impenetrable mass”? The prehistory of the Vale of Pewsey Procs. Prehistoric Society. WICKSTEAD, H. 2007: Land Division and Identity in Later Prehistoric Dartmoor, South-West Britain: Translocating Tenure. Unpubd. PhD dissertation. University College London. WILLIAMS-FREEMAN, J. P. 1932: Cross-dykes, Antiquity 6, 24-34. WILLIAMS-FREEMAN, J. P. 1934: Review of Celtic Earthworks of Salisbury Plain, Antiquity 8, 231-2.