Chapter FourRound Barrows and Cross Dykes as Landscape Metaphors
Chapter FourOn the northern edge of Cranborne Chase (Figures 1.1 and 4.1), there are two striking and dramatic chalk ridges separated by the Ebble valley, termed here the Ebble-Nadder ridge, to the north, and the Ox-Drove ridge, to the south. Overall, this landscape comprises four very different topographic worlds: (1) the flat and relatively undifferentiated lowlands to the north and the west of the ridges; (2) the winding Ebble valley that divides the two ridges; (3) the ridgetops themselves (narrow and irregular with striking and often panoramic views); and (4) the secret and interiorised world of the coombes (dry river valleys). The study that follows is a detailed account of the locations of round barrows and spur and cross-ridge dykes in the landscape. It is based on fieldwork undertaken over a period of eighteen months and has involved repeated and regular visits to the places discussed, as well as walks between them and up and over the ridge, along its steep northern escarpment and the coombes that cut into it. The interpretative framework put forward could never have arisen without this personal physical experience and knowledge of place; creating this framework would be absolutely impossible just using a map. The
FIGURE 4.1 Location of the study area in southwest Wiltshire and northeast Dorset and the Ebble-Nadder and Ox-Drove ridges, showing some of the places mentioned in the text.
landscape itself and the barrows and dykes in it exerted their own agency, or effects, in my experience and perception of them. I would like to claim that they both influenced and constrained what became possible to write. In this sense, they are the mute co-authors of this and the other chapters in this book.
I argue that both the barrows and the dykes acted as material metaphors for the wider landscape. In other words, their locations were significant not just in themselves, as markers of specific places in the landscape, but also in the manner in which they were dialectically related both to their immediate and more distant surroundings in the landscape as a whole.
The Ebble-Nadder ridge, forming the northern edge of Cranborne Chase in southwest Wiltshire, is a dramatic stretch of chalk downland bounded by the river Nadder and the undulating greensands and clays of the Vale of Wardour to the north and the more narrow and incised valley of the river Ebble to the south (Figure 4.1). The ridge forms a bold scarp on its northern side, an unbroken barrier extending from Hoop Side Hill (181 m) in the east to Whitesheet Hill in the west (242 m), a distance of 14 km (Figure 4.2). The ridge gradually descends in height from west to east along its length, and the land dips gently away from the ridgetop to the south and the Ebble valley. The crest of the entire ridge is narrow, usually only about 180 to 270 m wide. Along it, usually just to the south of the very highest ground, runs the former Shaftesbury to Salisbury turnpike road. Arable land is now characteristic along the ridgetop but with the steep scarp slopes remaining unploughed and under pasture. Below the northern scarps, small woods or copses, sometimes called Ivers, still remain. The southern side of the ridgetop is broken up with a series of steep-sided coombes running into it from the south, dissecting the otherwise fairly gentle slopes running down to the Ebble valley. Altogether two long barrows, fifty certain or probable round barrows, and sixteen cross-ridge and spur dykes are recorded along the ridge as a whole.
This ridge to the south of the Ebble valley, running approximately east to west, is far less regular in form. Along part of it runs the ancient Ox-Drove ridgeway from Woodminton Down in the east to Win Green in the west. The ridge, but not the ridgeway, continues to Melbury Beacon and Fontmell Down in the west, a stretch of 14 km (Figure 4.3). Along the ridge there are
Figure 4.2 Ebble-Nadder ridge showing the distribution of long barrows, round barrows, cross-ridge and spur dykes, and places mentioned in the text.
Figure 4.3 Ox-Drove ridge showing the distribution of long barrows, round barrows, cross-ridge and spur dykes, and places mentioned in the text.
two long barrows, at least 53 certain or probable round barrows and at least twelve well or partially preserved cross dykes running dramatically for greater or shorter distances across the chalk downland. Win Green (277 m) is the highest point and most significant landmark on the ridgeway and indeed the highest point on Cranborne Chase. Melbury Beacon, a dramatic rounded hill marking the terminal point of the chalk ridge before the heavy clay lowlands of the Blackmore Vale at 263 m, is the second-highest point; the impressive spur at Winkelbury (260 m) just to the east of Win Green is only slightly lower. To the east and south of Win Green the land gently falls away along the ridgeway and off it to the south. To the north there is a very steep and indented scarp falling away to the valley of the Ebble. The top of the ridge-way itself is almost flat or rises and falls only gently from west to east and is quite narrow, about 250 m or less. There is an enormous contrast between the dramatic northern escarpment and the manner in which the land dips away almost imperceptibly to the south. However, these gentle southern slopes are cut into at regular intervals by long steep-sided and meandering coombes. None manage to cut across the chalk ridge, but some to the south and east of Win Green come close to doing so. At the eastern end of the ridgeway, the slope down from the summit is dissected by far shallower valleys, creating a far simpler and less bold relief.
The northern escarpment is far from uniform, having a series of short fretted spurs jutting out from it, Winkelbury Hill being the most significant and well-defined. A series of rounded concave hollows, all unique in form, cut into the scarp, creating bowl-like forms or more irregularly shaped declivities. The drop in relief may be up to 100 m and the slopes truly precipitous. Their dizzying steepness, together with views from the ridgetop down the scarp slopes and across the lowlands below, are quite breathtaking.
The deeply incised coombes are equally dramatic but in an utterly different way. The eye tends to follow and run along their contours and down into their hidden depths. The flat base of these coombes may be as little as 5 or 10 metres with their side profiles being wider or narrower and more or less meandering and contorted. At the western end of the ridge, Melbury Beacon (colour plate 2) and Fontmell Down are the two most striking topographic features on the western escarpment of Cranborne Chase. Here two narrow coombes do cut into the escarpment edge in a manner equivalent to those to its south but run from west to east. Melbury Bottom meanders for 4 km through the chalk and almost all the way up to Win Green. Longcombe Bottom is deeply incised, flanking the southern side of Fontmell Down. There are no equivalents on the northern scarp edge, which lacks coombes, except at the far eastern end, where they terminate the scarp edge, reducing it to a series of steep-sided rounded hills. East of Longcombe Bottom a series of coombes cut into the chalk ridge from the south and southeast. The most complex narrow and deep series of meandering coombes occurs immediately to the south and west of Win Green and between it and Winkelbury Hill, almost, but not quite, cutting right through the ridgetop. These coombes meander through the chalk for 3–4 km or more, steep sided and V-shaped in profile, with numerous arms and side branches. East of Winkelbury Hill, the coombes to the south of the scarp are no longer dramatic but form shallower valleys situated much farther away from the ridgetop.
The remainder of this chapter is divided into two main sections. The first section discusses the relationship between long barrows and round barrows and the escarpment edges, ridgetops and coombes. Section II extends the analysis to cross-ridge and spur dykes. The concluding paragraphs pull the observations together to provide a new interpretation of their relationship both to one another and to the landscape as a whole.
Table 4.1 The dimensions and elevation of the long barrows along the Ebble-Nadder and Ox-Drove ridges.
There are four recorded long barrows on the two ridges (Table 4.1). All are of standard form, with side ditches running parallel to the mound whose highest and/or broadest end faces to the east or southeast. On each ridge there is a smaller barrow situated toward the middle, or the eastern end, and a much larger and higher barrow, twice as big, situated at or toward the western end of the ridge. The length and the height of the mounds directly relate to their elevation in the landscape. The barrows on the Ebble-Nadder ridge are placed 4.7 km apart; those on the Ox-Drove ridge, 9 km apart, and none are intervisible. The locations of these monuments and their relationship to the landscape are highly individualised. They do not form a coherent group except in terms of their shared morphology.
The small barrow on Buxbury Hill is located across the neck of the only true spur to jut north from the Ebble-Nadder ridge (Figure 4.4). It is situated on a gentle slope rising up to the top of the ridgeway, and hence visibility is restricted due south. It is more extensive along the ridge top about 4 km to the east as far as Chiselbury Hill and about 2 km to the west as far as Swallowcliffe Down. To the north one looks out across the lowlands of the plain of Wardour. The southwest end of the mound is orientated toward Win Green, the highest point on the Ox-Drove ridge to the south some 8.2 km distant. The barrow is precisely located so as to afford a window to this hill otherwise hidden along this part of the ridge. The barrow marks the first point at which Win Green can be seen along the Ebble-Nadder ridge when approached from the north or east, and this would appear to be the significance of its location. It references a distant and very significant hill on the Ox-Drove ridge.
FIGURE 4.4 The long barrow on Buxbury Hill.
The much larger barrow on White Sheet Hill is situated on the highest part of the Ebble-Nadder ridge at its far western end. From here there are panoramic views along the entire Ox-Drove ridge to the south, north across the plain of Wardour, and east along the ridgetop. The long axis of the barrow duplicates that of the ridgetop in mimetic fashion. The southwest end of this barrow is orientated toward another very significant hill on the Ox-Drove ridge, Melbury Beacon, the second-highest point and the hill that marks the western end of the ridge. It is behind this hill that the sun sets on the midwinter solstice from the barrow. Although this barrow cannot be seen from off the ridge to its north, its western end is just visible on the ridgetop from a considerable distance away when seen from the lowlands to the west. Because of the manner in which the ridge itself gently rises from east to west, the eastern end of the barrow comes into view only about 300 m away when approached from the east. It is located to be highly visible in the landscape from the south and the west. The barrow is situated only 250 m to the north of the head of Berwick Coombe, but its presence is virtually hidden from the barrow site itself—nor can one look down to the bottom of the plummeting scarp slopes of the ridge a short distance to the north and west. Similarly, no coombes or scarp slope edges are visible from the Buxbury Hill barrow. Both barrows are instead related through their directional orientation to the two most prominent and significant points on the Ox-Drove ridge, which in both cases are to the southwest. The visibility of both along the ridgetop is restricted and limited, and they are invisible from off the ridge to the north despite being located relatively near to the scarp edge. Both mark significant points along the Ebble-Nadder ridge—the highest western end and the only true northern spur.
The locations of the two barrows along the Ox-Drove ridge, by contrast, appear to be much more ambiguous and muted. The small barrow in Vernditch Chase is located on the mid-point of a gentle north-south slope with extensive vistas to the south across Cranborne Chase but restricted visibility up to the ridgetop to the north. Its west-east long axis respects the axis of the ridge itself and runs parallel to the course of a small west-east coombe, Chickengrove Bottom, 750 m to the north. This coombe is unusual, because it is one of only two to cut into the Ox-Drove ridge in a west-east direction, and it is situated at its eastern end but is not visible from the barrow, and its long axis does not relate to any significant landscape feature.
The barrow on Ashcombe Down contrasts considerably. Situated just beneath the top of the ridge, it straddles a north-south slope only 1 km to the southwest of the Win Green summit, which is visible from it. Approached from the south, it marks the point at which Win Green can first be seen in the distance. The long axis of the mound is not, however, orientated so as to reference Win Green but to the very head of Berwick Coombe 2.2 km to the northeast. In terms of the local topography, this is a very significant point, indeed, because the head of Berwick coombe almost cuts right across the top of the ridge and is the only coombe to do so. It falls short of cutting completely through the ridge top by only a few metres. There are other prominent coombes cutting into the ridge in the vicinity of this barrow—Melbury Coombe only 250 m to the northwest and Quarry Bottom and its branches to the south and west—and the barrow is roughly equidistant from the heads of Melbury Coombe and the western branch of Quarry Bottom. Melbury Coombe up-slope cuts deeply and unusually into the ridge running west to east but is invisible from the barrow itself. From the barrow site, one can look along and down to the bottom of the western branch of Quarry Bottom to the southeast. The location and the orientation of the long axis of this barrow seem to be intimately connected to the coombes in the barrow’s vicinity rather than to significant points along the ridgetop itself, and this barrow is the only one of the four long barrows from which coombes and their bottoms are both visible and directly referenced in terms of its landscape setting.
SUMMARY A number of significant themes can be drawn from the discussion of the long barrow locations:
These barrows thus on the one hand relate strongly to their immediate localities and on the other to much more distant worlds—although not to one another. The only long barrow from which other long barrows are visible is that on Ashcombe Down. It is intervisible with two long barrows associated with the cursus in central Cranborne Chase, Thickthorn Down, and Gussage Cow Down, some 10 km away to the southeast and forming a very different Neolithic landscape and world (see the discussion in Tilley 1994: 147ff).
The character of these great chalk ridges changed dramatically during the early Bronze Age. Numerous round barrows were constructed along the ridgetops and their spurs. The landscape became filled by barrows. Fifty certain or probable barrows were constructed along the Ebble-Nadder ridge and approximately the same number along the Ox-Drove ridge (Figures 4.2 and 4.3). Distances between individual barrows or barrow groups along both ridges rarely exceed 1 km and are often less. They are not closely related to the locations of the long barrows, none of which have round barrows in their immediate vicinity. The nearest round barrows to the White Sheet Hill long barrow are about 250 m distant. They are up to 500 m distant from the Sutton Down and Ashmore long barrows and more than 1 km distant from the Vernditch Chase long barrow. It is as if these places were being deliberately avoided, contrasting with the situation elsewhere in southern England, where round barrows are quite frequently aligned in the landscape in relation to the earlier long barrows, notable examples being in the great barrow concentrations along the South Dorset Ridgeway (see Chapter 5) and around Stonehenge (Chapter 3).
ROUND BARROWS AT THE EASTERN END OF THE EBBLE-NADDER RIDGE In this section I discuss in detail the locations of twenty four round barrows and four dykes toward the eastern end of this great ridge. Here (see Figure 4.5) a reticulated series of branching coombes cut deeply into the ridge from the south. Three V-shaped coombes with narrow flat bottoms, only about 5–10 m or so wide, meander through the chalk approximately north-south. That to the west and that to the east fork at their terminal ends. The central coombe is more linear and regular in form, lacking a bifurcated end and it cuts farthest into the ridge to the north. There is thus a rough symmetry in their form, with the forking coombes to the east and west of the central simpler straight coombe. The deepest and widest of these coombes is that to the west, which, from its southern end, at first runs approximately west to east before swinging round to the northwest and then to the north. These three coombes run roughly parallel to one another, north-south, for 1.3 km before joining where the coombe bottom becomes comparatively wide (see Figure 4.6), about 200 m across. Here it runs east to west before curving round to run north-south and then north-west to south-east, where another narrow coombe joins the system before it eventually opens out into the Ebble valley. These three parallel coombes are unique along the Ebble-Nadder ridge, giving them an added significance. Elsewhere along the ridge the coombes are more widely separated and usually constitute a single irregular series of bifurcating and meandering dry valleys.
The walking distance, following the overall course of the coombe system, between the river Ebble and the terminal ends of the coombes is about 5.6 km. The journey from the river leads from a comparatively wide and open river valley, up to the narrow twisting and turning of the coombes into valleys that become at first successively narrower, deeper, and steeper. They then become more and more narrow and shallow toward the points at which they terminate on the ridge. To the east of this there are a further series of coombes, but these are considerably shallower and far less distinctive.
Figure 4.5 Distribution of round barrows and dykes at the eastern end of the Ebble-Nadder ridge in relation to topographic features.
On the northern side of the Ebble-Nadder ridge there are only two coombes along the entire 14-km stretch of the northern escarpment that cuts into it (see Figures 4.3 and 4.5). Both of these occur in the area under discussion, at the far eastern end, just before the ridge itself dips away to the east and is lost altogether. The most easterly coombe is rather shallowly incised and is V-shaped. It lacks any clearly defined flat bottom and meanders into the ridge from the northeast. Just 250 m to the west, Punch Bowl Bottom (Figure 4.10) is utterly different and absolutely unique in form in a number of respects. It cuts into the ridge from the Nadder valley at first running south and then swinging around to the southwest.
Figure 4.6 (Top) Locations of barrows 6–13 on the Ebble-Nadder ridge at the point at which three coombes join. (Bottom) Photo taken from the top of Barrow 10 looking east along the coombe bottom. The spur end on which Barrows 6–9 are located is to the left of the photo above the white track.
It is bold, very steep sided, and has a comparatively wide tongue-shaped flat bottom that widens out, rather than narrows, which is otherwise always the norm, toward the coombe end. It converges with a long spur running north out from the ridge. This leaves only a narrow sliver of high land between the coombe end and the northern escarpment edge to the west. This dramatic convergence of coombe, steep escarpment edge, and jutting spur is an elemental clash of distinctive topographic forms that are separated everywhere else along the ridge.
Overall, five topographically distinctive features of the landscape at the eastern end of the Ebble-Nadder ridge can be defined:
These are shown in Figure 4.5.
ROUND BARROWS LOCATIONS Approximately twenty-four round barrows are known from the eastern end of the Ebble-Nadder ridge. These are all round barrows of simple form and generally small in size, 7–15 m in diameter. Eight have traces of a surrounding ditch. Some, with pits in their tops, have obviously been plundered. None have any known recorded artefacts or excavation records. Eleven, or nearly 50%, have now been destroyed by ploughing, and their location is known only from aerial photographs.
In this section I describe the individual relationships of these barrows to the major topographic features of the landscape noted above based on observations taken from the barrows themselves, or from former barrow sites, in the case of those destroyed.
Bishopstone (Figure 4.5: 1) A barrow, now destroyed, was situated on the southern tip of a gentle spur sloping south. It was situated on flat ground on the eastern side of the spur top just above the point at which the land dips at first gently, then more steeply, south to the Ebble valley. From the barrow site there are surprisingly extensive views to the west and east along the Ebble valley. To the north the visual field is over 2 km to the ridgetop and extends south for over 1 km to the hills marking the other side of the Ebble valley.
Broad Chalke (Figure 4.5: 2–3) To the west, two additional probable barrow sites (2–3) are recorded on the mid-points of gentle southern slopes running down to the Ebble valley. That farthest west is only about 100 m north of a spring line on the eastern side of a coombe system running down into the Ebble valley. Views are extensive along the Ebble valley to the east and west, limited by rising ground to the north and across the valley to the south. The eastern barrow is on a gentle southeast slope, again overlooking the Ebble valley but with restricted views to the north because of rising land. These barrow sites are not intervisible, also not with the site at Bishopstone.
Poor Patch, Stoke Down (Figure 4.5: 4–5) Here, two small adjoining barrows in a north-south alignment are situated above the shoulder of the scarp slope on gently sloping ground, dropping down to the west, about 400 m to the south of the head of a coombe that meanders south to join the valley linking the Hydon Hill and Little Down coombe systems. The bottom of the coombe is not visible from immediately below the barrows to the west, although one can look along the course of the coombe to its head to the north and along it for a short stretch to the south. Views to the east are very restricted by the gently rising ground of the spur top. To the west they are limited by another rising spur. No other barrows are visible.
Hydon Hill/Little Down (Figure 4.5: 6–13; Figure 4.6) A dispersed group of eight round barrows are located around the area at which three coombes merge. These are all relatively small, between 7 and 15 m in diameter and less than 0.8 m high. At least four appear to have had a surrounding ditch. Four barrows cluster at each of the southern ends of the two spurs, Hydon Hill and Little Down, separating the three coombes. In each case the southernmost barrow is located low down the scarp slope, falling away to the coombe bottom. One of these, Barrow 6, is located exceptionally low down in the landscape, only about 30 m to the north of the very bottom of the coombe system. Three others, 7, 8, and 9, less than 100 m to its north, are situated on much more steeply sloping ground below the shoulder of the scarp, thus making them invisible from the spur top to the north but effectively skylined and prominent from the coombe bottom below to the south. From all these four barrows, which are intervisible, one can look down to the bottom of the coombe to the south, but any view to the north is blocked out by the steep slope. They were thus positioned so as to be seen from the bottom of the coombe and occur on the terminal point of a narrow spur less than 200 m wide separating parallel coombes. Because of the manner in which the coombes curve, meander, and branch, it is impossible either to see up to the end of any of them or beyond their terminal points up to the top of the chalk ridge to the north.
The western group of barrows is structured slightly differently in the landscape. Barrow 10 is intervisible with Barrows 6–9, but among the others closer to it, it can be seen only from Barrow 11 situated about 100 m to its north, exactly on the shoulder of the scarp. Barrows 12 and 13 are situated higher up the slope, above the shoulder of the scarp on the far southern end of the spur of Hydon hill, and are invisible. Although Barrow 10 is situated only 250 m due west of Barrow 6, it is situated far higher up the slope of its respective spur, approximately halfway up the slope from the base of the coombe to the scarp shoulder above. This is the only barrow in the western group (Nos. 10–13) that can be seen from the coombe bottom below to the south or from which the coombe bottom itself is visible immediately below it.
The barrows are thus structured in relation to the landscape in the following way:
Walking between the barrows, from east to west and from the lowest to the highest, one moves from the coombe bottom to mid-points on the scarp slope to the shoulder of the scarp to the higher flat spur tops beyond. Taken together, and in relationship to one another, the barrows thus mark all the significant transition points in the landscape between the coombe bottom and the ridge spurs to the north and beyond.
Burcombe Ridgetop Barrows (Figure 4.5: Nos.14–16; 22) Stretched out along the ridgetop there are an additional six barrows. Four of these are situated on the very highest points to the east and the west, the distance between them being 2.5 km. These barrows on the flat ridge summit, all but one now destroyed, would all have been intervisible along the ridge top. Barrows 14–16 mark the limits of the visual field looking west along the ridge top from Barrow 22, and Barrow 22 similarly marks the limits of the field of vision looking east from Barrows 14–16. None of the coombes cutting into the ridge from the south or north are visible from them. Similarly, the presence of a steep escarpment edge to the north is hidden from the barrow locations, but views are extensive in this direction.
Burcombe Ivers Barrows (Figure 4.5: 17–18) Two additional barrows, now destroyed, were also situated near to the top of the ridge but below the flat ridge summit on gently sloping ground a little distance below it, but well above the sharp shoulder of the scarp. Each barrow is situated near to the south of a distinct gully in the northern escarpment edge that runs due west-east at this point. Views to the south are restricted by the rising ridgetop. They are extensive off the ridge to the north. The scarp slope beyond to the north can be seen from these barrow sites but not its bottom immediately below them. To the west, the visual field is again limited by the rising land of the ridgetop, and the barrows would not be intervisible with the summit barrows (Nos. 14–16) in this direction. These two barrows are situated in a transitional zone between the flat ridge summit and the shoulder of the scarp slope. They, and the other ridgetop barrows, are all highly visible from off the ridgetop in the Nadder valley below to the north.
Hoop Side (Figure 4.5: 19–21) Here a unique cluster of three adjoining barrows are situated just below the shoulder of the scarp where it drops away precipitously to the north. These are aligned in a staggered west-east row on sloping ground. There is a distinct gradation in size. The largest and highest barrow, about16 m in diameter and 2.6 m high, is situated at the eastern end of the group highest up the slope; the smallest, 1.3 m high and 10 m in diameter, is situated lowest down the slope at the western end. Visibility to the south is very restricted by the rising land of the ridgetop. To the north, it is extensive off the ridge; to the east it is limited again by rising ground. To the west, it extends to the Burcombe summit top barrows (Nos. 14–16) that would be skylined in this direction. The three Hoop Side barrows are situated at the point of the escarpment edge just where it begins to swing round to the north to form the spur cut into by the Punch Bowl coombe to the east. From the two westerly barrows, the base of the northern scarp slope immediately below is visible; from the third higher barrow to the east it is not. The barrow on the flat summit area (No. 22) only 400 m to the east is invisible. This contrasts with the view 2 km west to the summit barrows (Nos. 14–16), which would have been visible in this direction. From the barrows the Burcombe Punch Bowl barrow (No. 23) is also visible 500 m away to the northeast and another on the end of the north running spur below 1.5 km to the north (No. 24).
Burcombe Punch Bowl (Figure 4.5: 23; Figure 4.7) This barrow is dramatically situated at the head of Burcombe Punch Bowl coombe (see Figure 4.10). It is situated on land gently sloping to the south well above the shoulder of scarp slopes to the west and north. This is the largest and most impressive surviving barrow in the area, 18–20 m in diameter and 2 m high. From the barrow, one can look down the lower part of the coombe. However, the base of the coombe immediately below the barrow is concealed by the slope. The summit barrow just 300 m to the north (No. 22) is concealed by rising ground, while all those to the west along the ridgetop are visible, as is the barrow marking the end of the spur below to the south (No. 24).
Figure 4.7 The round barrow marking the head of Punch Bowl Bottom, the largest barrow at the eastern end of the Ebble-Nadder ridge.
Burcombe Spur (Figure 4.5: 24) This barrow is situated on flat land on the far northern end of a spur on the western side before the land dips down sharply to the Nadder valley. From the barrow site (now destroyed), there are extensive views to the west and east along the Nadder valley, to the north up to the top of the Grovely ridge beyond and south up the ridgetop to the Punch Bowl barrow. To the southwest, the Burcombe summit top barrows (Nos. 14–16) would have been sky-lined. This most northerly barrow and its relationship to the topography is almost a mirror image of the most southerly barrow (No. 1) on the low spur above Bishopstone.
SUMMARY A number of points can be drawn from these brief descriptions of the twenty-four barrow locations:
ROUND BARROWS ON THE CENTRAL AND WESTERN END OF THE RIDGE Here there are an additional twenty-six certain or probable barrows. As a result of massive arable destruction since the 1940s, only six of these (23%) survive today as visible monuments. They are all relatively small bowl barrows, some with traces of surrounding ditches varying in diameter from 10–20 m, and none is more than 1 m high. Most appear to be of soil and turf construction. A few smaller examples are mounds comprising mainly flints. Eleven (42%) are single monuments without others in the immediate vicinity. Eight barrows (31%) appear paired with distances of 1–200 m between them (Figure 4.2: 49–50; 46–45; 47–48; 34–35). An additional two barrows (Figure 4.2: 37–38) were directly adjacent to each other in a west-east alignment (Clay 1926a: 434). There is only one cluster of five barrows (Figure 4.2: 39–43) on Swallowcliffe Down.
Of the single monuments, six occupy the ridgetop, three spurs running out from it, and one is situated directly at the head of a narrow coombe (Figure 4.2: No. 29). The barrow pairs similarly occupy the ridgetops and spurs, and the largest concentration of barrows occurs in a rough arc around the head of one of the most dramatic coombes along this part of the ridge. Coombes are partly visible from fourteen (54%) of the barrow sites, from which there are views along or across parts of them either looking up or across the coombes in the case of those barrows situated on spurs or down and along them in the case of the barrows on the ridgetop. However, from only a couple of barrow locations (Nos. 29 and 32) can the bottom of the coombe be seen immediately below the barrow site. Thus the relationship with the coombes is more distanced and generalised than that encountered along the eastern end of the ridge. What differs here is the major cluster of barrows around the coombe head on Swallowcliffe Down, emphasising its significance in a direct way, and the location of No. 29 at the very terminal end of the coombe whose sides run up to and terminate at the barrow itself. This barrow does not overlook the deep coombe below. It is placed at the point at which a shallow coombe terminates fading away into the chalk.
The ridgetop barrows are all situated either on flat or gently sloping ground well above the shoulder of the northern escarpment edge, which is not visible from them, and it is not possible to see down to the bottom of the scarp slope from any of them. Almost all are sited on the flat ridgetop itself or on gently sloping ground dipping away from it. This, together with the presence of large numbers of barrows on southern spurs with coombes on either side, indicates the importance of the southern side of the ridge as opposed to its northern escarpment edge, and only two barrows (Figure 4.2, Nos. 45 and 33) would have been visible from off the ridge to the north. They are sited instead so as to be visible from the top of neighbouring Ox-Drove ridge to the south.
The barrows on the southern spurs (with the exception of No. 32) all occupy central areas of these spurs rather than their sides, hence the lack of direct views down into the coombes below them. Coombes occur to the west and east of them, defining the edges of the spurs that they occupy, but the relationship is not an intimate one. None occur at the bottoms of these coombes or mark points at which they turn or join.
Patterns of intervisibility between these barrows are shown on Figure 4.8. Again, the pattern is very similar to that encountered along the eastern end of the ridge. Some barrows in close proximity to others on the ridge top are not intervisible. but there may be views of distant barrows along it. Barrows situated on southern spurs often have restricted views up to the ridgetop because of rising ground but have very extensive views south to the Ebble valley and up to the Ox-Drove ridge.
A number of these barrows were excavated by Clay in the 1920s, but the results are meager and not all that well reported. Barrow 26 contained the partial remains of a primary, possibly crouched skeleton of a young man lying on the left-hand side with the head facing southwest and the legs northeast in an oak coffin made from a hollowed and split tree trunk in a grave orientated NE-SW cut into the chalk. A large red deer antler was found in front of the
FIGURE 4.8 Intervisibility of round barrows in the central part and at the western end of the Ebble-Nadder ridge.
skull and another, possibly a pick, by the hips; a calcined flint was also found. The orientation of the coffin and the skeleton was thus the same as the ridge. The skeletal remains consisted unusually of frontal and temporal bones, a molar, a 12th rib, and a pisiform. This skeleton appears to have been disturbed and disarticulated either before or after being placed in the coffin (Clay 1926a: 102). In Barrow 30, Clay found remains of a cremation in a late Bronze Age urn, possibly remains of a secondary burial. In Barrow 33, Clay discovered traces of a cremation with an urn and fragmentary skeletal remains of a child and a woman. Beaker sherds were discovered nearby, and the finds may represent the remains of a primary Beaker inhumation with a secondary cremation urn burial (ibid.: 250; Grinsell 1957: 192). In Barrows 37 and 38, Clay inferred the presence of primary contracted adult Beaker burials. A pit was cut into the chalk in Barrow 37. Barrow 38 had no burial pit but remains of a crouched adult skeleton with the head to the west and the feet to the east. Remains of a collared urn were found in the ditch, possibly indicating a secondary burial (Clay 1926a: 432–434). Barrows 40 and 42 contained remains of a primary cremation burials in a central cist (ibid.: 435). Barrow 44 contained pieces of bone and Bronze Age sherds (ibid.: 432).
THE OX-DROVE RIDGE There are fifty-three certain or probable round barrows along the Ox-Drove ridge, that is, an almost identical number to those along the Ebble-Nadder ridge. Their distribution is similarly virtually continuous along the ridgetop, with short distances usually of 1 km or less separating individual barrows or barrow groups (Figure 4.3). Just less than half survive today. All except one, a saucer barrow, are standard bowl barrows, some with surrounding ditches and occasionally berms and banks. Dimensions vary from smaller examples 5–6 m in diameter to larger ones 16–20 m. Their height rarely exceeds 2 m and is usually considerably less, 0.5–1 m, even in the cases of barrows that have never been ploughed or disturbed. Most are discrete rather than monumental landscape markers. Twenty-one (40%) occur on their own, but often distances between them amount to no more than 200–300 metres. Eight occur in pairs (15%). There is one group of three barrows, one group of four conjoined barrows, and two clusters of six barrows. The groups and pairs of barrows are all aligned exactly or approximately west-east or northwest to southeast or northeast to southwest following the general alignment of the ridgetop, except the large group of six on Winkelbury Hill, which instead are dispersed along the axis of this north-south spur, the only prominent northern spur to occur along the ridgetop. Almost all are situated in close proximity to the northern scarp of the ridgetop. The maximum distance between these barrows and the northern scarp edge is 250 m, and the majority are considerably closer. The feature that is most remarkable about their distribution is that so few are situated at any distance from the scarp edge to the south. Only three are sited on southern spurs (Figure 4.3: 21–22 and 30) and an additional pair (Figure 4.3: 51–52) on the extreme western spur of Fontmell Down where the ridge ends. Their location thus contrasts with many of the barrows along the Ebble-Nadder ridge, which quite frequently occupy southern spurs. This locational difference may be related to a fundamental difference in topographic boundaries. Although the Ebble-Nadder ridge is well defined to the south by the valley of the river Ebble toward which the spurs and coombes run, the Ox-Drove ridge has no similar topographic boundary on its southern side. The land here dips gently away to the south, except in places where it is broken up by deep and dramatic coombes and spurs at the western end, toward the rolling downland of central areas of Cranborne Chase. There is no southern edge to the ridge. For a distance of 6 km or more from the northern ridge scarp edge to the south, there are no long barrows and only a few isolated round barrows until the next major concentration occurs associated with the Dorset cursus: a cultural rather than a natural topographic marker of the downland landscape. Thus the barrows along the Ox-Drove ridge are intimately related, in this general sense, to the presence of the dramatic scarp edge that defines it to the north.
The majority are localised in relation to the highest points along the ridge summit. The group of six barrows on Marleycombe Hill at the eastern end of the ridge occupy the centre of a localised summit 200 m high, which is almost a hill being cut into and defined to the west and east by deep coombes. Similarly, the three barrows on Trow Down occupy the summit area, rising here to 243 m. The group of six on Winkelbury Hill occupy the highest part of this northern spur. Win Green, the highest point of all, is surmounted by a barrow, and to the west barrows are strung out along the top of Charlton Down and Breeze Hill, where the ridgetop sweeps round to run toward the southwest. Most of these ridgetop barrows are situated on flat or only gently sloping ground well above the shoulder of the northern scarp. It is not possible to look down to the bottom of the scarp slope to the north, and from some (for example, Figure 4.3: 15–16; 18–20) even the presence of such a dramatic slope is hidden, because they are set well back from it. From these barrows, there are extensive views to the north across the Ebble valley and up to the top of the Ebble-Nadder ridge. They were placed to see, and be seen, from the barrows occupying the ridgetop and southern spurs of the Ebble-Nadder ridge. They were not located to be seen from the sloping ground to the south, and intervisibility between individual barrows and barrow groups along the ridgetop itself does not appear to have been as significant (see Figure 4.9). As along the Ebble-Nadder ridge, some nearby barrows are not intervisible, whereas more distant ones along the ridgetop may be. The primary visual reference points were to the barrows on the neighbouring ridge rather than to
FIGURE 4.9 Intervisibility of round barrows along the Ox-Drove ridge.
one another along the same ridge, which appear to have been of secondary significance. It was possible to see not only the barrows on the Ebble-Nadder ridge but also the deep coombes cutting into it and breaking it up. The mystery of these coombes may have been of deep significance. When one walks along the Ox-Drove ridgetop between the barrow groups and looks north to the Ebble-Nadder ridge, the character and forms of these coombes constantly change. At some points, one can see directly up to their heads. Where they twist and turn, the coombe ends may appear as huge hollows or bowls in the side of the ridge rather than sinuous dry valleys. In days with scattered clouds and sunlight, the coombes and‚ hollows’ are sometimes darkened, while the rest of the ridge is bathed in sunlight, or vice versa. They are thus mobile and ever changing in character rather than static fixed forms. They change in tandem with the movements of the body and the character of the light.
Along the northern scarp edge, there is one unique and unusual group of barrows on Woodminton Down. These, unlike all the other barrow groups, are conjoined in a west-east line rather than being separate, closely spaced mounds. Conjoined mounds occur elsewhere only among a couple of the group of six, which are arranged in two parallel rows on Marleycombe Hill. Unlike the other barrows, those on Woodminton Down are placed well below the shoulder of the scarp, about halfway down the slope, and from these barrows the scarp bottom immediately below them is visible. From a few of the other barrows one may be able to see some way down the scarp slope but not to its very bottom. These barrows thus reference the scarp base in a manner similar to the group at Hoop side at the eastern end of the Ebble-Nadder ridge and would not themselves be visible from it.
From thirty-five of the barrows, coombes are visible cutting into the Ox-Drove ridge (66%), but usually this is a distant rather than intimate relationship, since one can rarely see down into or along the coombes themselves. Only the upper edges of one or the other of their scarp slopes can be seen in the distance. Coombe bottoms are visible from only seven of the barrows (13%). As along the Ebble-Nadder ridge, some barrows are, however, intimately related to the presence of the coombes in various ways. These are all located toward the western end of the ridge, where the coombes dramatically cut into it from the south and west, and it is only here where barrows occupy southern spurs, as mentioned above.
Barrow 31 (see Figure 4.3) is located immediately below the shoulder of the scarp slope plunging dramatically down to the bottom of Berwick Coombe to the west on the eastern side of the coombe, only 250 m from its head. While the land drops down sharply to the west, it also dips from north-south, which means that this barrow is invisible except when approached from either the east or the north, but it can be seen from up to 2 km away to the west and about 1 km to the south. It is intervisible with the long barrow on Ashmore Down, whose northwest orientational axis references the head of the same coombe that, as discussed above, uniquely almost breaches through the ridgetop. From the barrow site, one looks down into the depths of the heavily wooded coombe below.
Barrow 33 is located at the very head of a western branch of the same coombe system with views down and along the coombe bottom. Just to its east and above the head of the same coombe, another barrow on Win Green occupies the highest point of the ridge. It is thus associated with the spiritual power associated with the height of the ridge and that related to the depths of the coombes. Barrow 35 (now destroyed) was sited at the head of Melbury Coombe, cutting into the ridgetop and splitting it into two. Barrows 45 and 46 are situated near to the head of a northern side branch of Melbury Coombe. Barrow 49 is placed in a similar position to Barrow 31 just below the shoulder of the steep scarp slope plunging down into Melbury Coombe. This barrow, until reached, is invisible from the ridgetop from the south. From it, one can look along the coombe bottom and up toward the barrows to the north at the southern end of Breeze Hill. This is one of the largest barrows along the ridge, 17 m in diameter and 1 m high, with traces of an external ditch and outer bank. These barrows referencing the coombes and relating to their depths replicate the pattern of relationships that occur along central parts of the Ebble-Nadder ridge and particularly toward its eastern end, but with the difference here that no barrows are found actually in the coombe bottoms or at various points down the scarp slopes, way beyond their shoulders, dropping to them.
Excavation of the six barrows on Marleycombe Hill by Clay revealed some interesting evidence about the continuity of barrow placement along the ridgetop. Here there are two parallel lines of barrows (Figure 4.3: 2–7). The northernmost three form a staggered line, with the largest and most easterly barrow situated slightly to the south of the other two. These are the barrows closest to the northern scarp edge situated on a gentle south-north slope and are probably earlier in date, with Beaker material being recorded from the old turf lines below the barrows (Clay 1927b: 548–551). The presence of three humeri in the largest barrow indicates two inhumation burials (later destroyed by barrow diggers). Barrows 2 and 3 covered large central cairns of heaped flints. Three pits in barrow 2 probably contained a primary Beaker inhumation burial and two secondary urn cremation burials. Clay discovered no primary burial in barrow 3 but a secondary deposit of an inverted barrel-shaped urn that contained no ashes or bones. The southern group, on the skyline when seen from those below, to the south, are in a dead straight line and are of later Bronze Age date. They contained primary and secondary cremations in Deverel-Rimbury urns (ibid.: 551–556).
The four barrows also excavated by Clay on Woodminton Down also seem to date back to the earlier Bronze Age, with secondary burial taking place in the later Bronze Age. In Barrow 1, farthest to the east, Clay discovered a central heap of flints mixed with turf, which he concluded may have originally have contained a primary inhumation burial. Twenty-one cremations in urns were discovered in the northwest segment of this barrow (according to Clay’s plan, stated as being in the southwest in his text) (Clay 1926b: Plate IV). These globular and barrel-shaped urns had been dug into the barrow surface and did not penetrate into the old ground surface below it, indicating their intrusive nature. Flat stone slabs of non-local material, sandstone or Purbeck stone, were placed over the tops of the urns and would have been visible on the barrow surface. Barrow 2 appears to have been built later with a primary cremation in an urn in a pit at the centre and three secondary cremation burials in urns. Barrow 3 probably contained a primary inhumation, which was later disturbed, and barrow four a primary inurned cremation (ibid.: 313–315). One possible interpretation is that there were originally two barrows here constructed in the earlier Bronze Age, with two more barrows being built immediately to the west of each of them—all four becoming conjoined—followed by another period of cremation burials in the barrow mounds.
Situated on the ridgetop above the four barrows on Woodminton Down, Clay excavated an additional barrow (Figure 4.3: 14). This one is less than 500 m away but out of sight of the four below and situated well above the shoulder of the scarp slope, rather than halfway down it. It contained an inverted urn with the cremated bones of a woman and a bronze awl. The urn rested on the ground surface rather than being buried in a pit (ibid.: 322).
Apart from these barrows, the only barrows for which there are any excavation records are from the isolated barrow on Barrow Hill situated just to the north of the ridge (Figure 17) and from those along the northern spur of Winkelbury Hill. On Barrow Hill, Clay found a primary cremation in a Late Bronze Age barrel urn with charcoal, ashes, and burnt bones in and around it. A small hole dug into the old ground surface contained the urn. (ibid.: 325–326). In one of the barrows on Winkelbury Hill excavated by Pitt Rivers, a Late Bronze age urn was found filled and surrounded by flint chippings but without any evidence of a cremation. In two others, primary earlier Bronze Age inhumation burials were discovered, one together with pot sherds and a bronze awl (Pitt Rivers 1888: 258–259).
All this evidence points toward a strong continuity in the use of the same places for barrow construction and burial along the ridge top, and well down the scarp in the case of the Woodminton Down barrows from the early Bronze Age onward. Such continuity in barrow location from Beaker burials to later Bronze Age burials is also evident from Clay’s excavations of the barrows along the Ebble-Nadder ridge, discussed above. Once barrows had been built, they attracted other barrows, and when barrows ceased to be built, they were reused over and over again as places of secondary burial, presencing the dead forever in the landscape and connecting past and present and future.
This area of Cranborne Chase has been a classic area for the study of cross dykes from their first recognition as a particular class of monuments in the landscape. In total there are twenty-eight dykes on these two ridges (see Tables 4.2–4.7 and Figures 4.2 and 4.3). In the following section, I first discuss their general distribution, then briefly review the literature on cross dykes and linear ditch systems and then consider their landscape locations in detail.
On the chalk downlands of southern England, cross-ridge and spur dykes frequently link coombes (dry valleys) to escarpment edges or, less frequently, coombes to coombes or cut across spurs between opposing scarp slopes. They are principally concentrated in particular areas, sometimes together with other types of linear ditch systems, on the Wessex chalk downlands of Dorset, Wiltshire, and Hampshire, along the chalk ridges of the South Downs of southern England and the Yorkshire Wolds, and along the unusual limestone Tabular Hills of northeast Yorkshire. Even within specific regions, they are both very unevenly distributed and peculiarly concentrated. For example, within south and west Wiltshire twenty-eight are concentrated along two short chalk ridges forming the northern edge of Cranborne Chase. Two additional concentrations occur on White Sheet Hill near to Mere and around Cold Kitchen Hill, Kingston Deverill, over 12 km distant to the north. By contrast, none occur along the topographically indistinct Grovely Ridge in-between, and they are absent over almost all the rest of Cranborne Chase to the south and on Salisbury Plain to the north and east, where extensive linear ditch systems occur. In central Dorset, eleven cross dykes are densely concentrated around Lyscombe Bottom in an area of chalk ridge extending for only a few kilometres, and ten more are located along a short stretch of chalk escarpment, bordering the Blackmore Vale, northeast of Ibberton. By contrast, in south Dorset only two examples occur along the entire 23-km-long great chalk ridge running from Lulworth Cove in the west to the Ballard Down in the east. Similarly, along the 16-km-long south Dorset Ridgeway, near to Dorchester, with its major linear Bronze Age barrow cemeteries, only a few are recorded.
Their overall distribution immediately suggests three things: (1) that cross dykes are usually clustered together as more or less integrated systems
Table 4.2 Forms of cross-ridge and spur dykes in the study region.
| Map No. | Type | Form | Ditch | Approx Orientation | |
| Ebble-Nadder Ridge (Fig. 4.2) | |||||
| 1 | Spur | Univallate | Upslope to south | West-East | |
| 2 | Spur | Univallate | Upslope to south | West-East | |
| 3 | Spur | Bivallate | Medial | West-East | |
| 4 | Cross ridge | Bivallate | Medial | North-South | |
| 5 | Cross ridge? | Bivallate | Medial | North-South | |
| 7 | Spur | Univallate | Upslope to south | West-East | |
| 8 | Cross ridge | Bivallate | Medial | North-South | |
| 9 | Cross ridge | Bivallate | Medial | West-East | |
| 10 | Cross ridge | Univallate | To west | Northwest-Southeast | |
| 11 | Spur | Multiple | Medial | West-East | |
| 12 | Spur | Univallate | To north | West-East | |
| 13 | Spur | Bivallate | Medial | West-East | |
| 14 | Cross ridge | Bivallate | Medial | North-South | |
| 15 | Cross ridge? | Univallate | To west | North-South | |
| 16 | Spur | Univallate | Upslope to north | West-East | |
| Ox-Drove Ridge (Fig. 4.3) | |||||
| 1 | Cross ridge | Bivallate | Medial | North-South | |
| 2 | Cross ridge | Bivallate | Medial | Northwest-Southeast | |
| 3 | Cross ridge? | Univallate | To east | North-South | |
| 4 | Spur | Univallate | Upslope to south | West-East | |
| 5 | Spur | Univallate | To north in dip | West-East | |
| 6 | Spur | Multiple | Upslope to south | West-East | |
| 7 | Cross ridge | Bivallate | Medial | Northwest-Southeast | |
| 8 | Cross ridge | Univallate | To east | North-South | |
| 9 | Spur | Univallate | To north | Northwest-Southeast | |
| 10 | Spur | Bivallate | Medial | Northwest-Southeast | |
| 11 | Cross ridge | Univallate | To south | Northwest-Southeast | |
| 12 | Spur | Univallate | Upslope to west | North-South | |
Table 4.3 The approximate present-day dimensions of the cross-ridge and spur dykes (metres) in best preserved sections.
| Map No. | Length | Overall width | Ht. of bank | Depth of ditch | |
| Ebble-Nadder Ridge (Fig. 4.2) | |||||
| 1 | 110 | 12 | 0.6 | 0.6 | |
| 2 | 220 | 12 | 0.6 | 0.9 | |
| 3 | 900? | ? | ? | ? | |
| 4 | 650? | 11 | 0.15 | 1.2 | |
| 5 | 720 | 15 | 0.4 | 0.2 | |
| 6 | 500? | 13 | 0.3 | 0.6 | |
| 7 | 200 | 12 | 1.2 | 0.9 | |
| 8 | 500 | 14 | 0.6 | 0.6 | |
| 9 | 400 | 15 | 0.6 | 0.6 | |
| 10 | 500? | ? | ? | ? | |
| 11 | 250? | 19 | 0.4 | 0.6 | |
| 12 | 210 | 8 | 0.3 | 0.6 | |
| 13 | 190 | 14 | 0.6 | 0.6 | |
| 14 | 430 | 21 | 1.2 | 1.5 | |
| 15 | 160 | 9 | 0.3 | 0.3 | |
| 16 | 390 | 9 | 0.6 | 0.6 | |
| Ox-Drove Ridge (Fig. 4.3) | |||||
| 1 | 700 | 14 | 0.8 | 0.6 | |
| 2 | ? | 12 | 0.6 | 0.6 | |
| 3 | ? | ? | ? | 0.6 | |
| 4 | 170 | 11 | 1.0 | 0.8 | |
| 5 | 250 | 14 | 1.1 | 1.1 | |
| 6 | 120 | 18 | 0.9 | 1.4 | |
| 7 | 370 | 19 | 1.4 | 2.5 | |
| 8 | 530 | 18 | 0.7 | 0.6 | |
| 9 | 440 | 14 | 0.9 | 0.6 | |
| 10 | 280 | 18 | 1.0 | 1.0 | |
| 11 | 720 | 11 | 0.4 | 0.7 | |
| 12 | 270 | 12 | 1.2 | 0.6 | |
of earthworks on ridgetops; (2) they occur only in areas where differences in topography are very marked and distinctive: they are found in very special and dramatic landscapes; (3) their high frequency in southwest Wiltshire and northeast Dorset and certain areas of central Dorset is paralleled only in a few cases elsewhere on the Chalk downlands of southern England. This in itself suggests, perhaps, that there must have been something especially significant
Table 4.4 Characteristics of the courses of the cross-ridge and spur dykes.
| Map No. | Straight | Curved | 'Meandering' | Directional shift in orientation | |
| Ebble-Nadder Ridge (Fig. 4.2) | |||||
| 1 | + | - | - | - | |
| 2 | - | + | - | - | |
| 3 | - | - | + | - | |
| 4 | - | - | + | - | |
| 5 | - | - | ? | + | |
| 6 | - | - | ? | - | |
| 7 | - | + | - | - | |
| 8 | + | - | - | - | |
| 9 | + | ? | ? | ? | |
| 10 | ? | ? | ? | ? | |
| 11 | - | + | - | - | |
| 12 | + | - | - | - | |
| 13 | + | - | - | - | |
| 14 | + | - | - | - | |
| 15 | ? | ? | ? | ? | |
| 16 | - | + | - | - | |
| Ox-Drove Ridge (Fig. 4.3) | |||||
| 1 | - | - | + | - | |
| 2 | ? | ? | ? | ? | |
| 3 | ? | ? | ? | ? | |
| 4 | + | - | - | - | |
| 5 | - | - | - | + | |
| 6 | +? | - | - | - | |
| 7 | + | - | - | - | |
| 8 | - | - | - | - | |
| 9 | + | - | - | - | |
| 10 | - | - | - | + | |
| 11 | - | - | + | - | |
| 12 | + | - | - | - | |
about the particular character of the landscapes in which they occurred that stimulated the embellishment of the ridgetops with such earthworks.
A characteristic feature of both the earlier and more recent literature on linear bank and ditch systems in general and cross-dykes and spur dykes, in particular, is a concentration on describing the forms of the monuments themselves and discussing their relationship to others. The specific relationship of these earthworks to the topography has been hardly discussed at all. What has barely been mentioned is the form and character of the ridges and spurs that are cut across by these bank and ditch systems. Are they distinctive in any way? Similarly, the form and character of the coombe systems linked to escarpment edges by cross dykes have not been discussed. Are they irrelevant? Would any
Table 4.5 The relationship of the cross-ridge and spur dykes to escarpment edges, coombes, and ridge slopes.
Table 4.6 Relationships of dyke ends to scarp edges of coombe and escarpment sides.
coombe or part of an escarpment edge be equally suitable for the addition of a cross dyke? Or might their shapes and forms and other characteristics be important in influencing their siting? Similarly, specific structural relationships between the dykes and the scarps and coombes have not been addressed in the literature. The dykes are invariably described simply as running to the ‘head of a coombe’, but this is an oversimplification covering a wide variety of structural relationships, as the account below hopes to demonstrate, both in terms of where the dyke is located in relation to the coombe or escarpment edge and how far down the scarp, or coombe edge, it continues.
This chapter shows that a detailed consideration of the relationship between cross dykes and their topographic settings leads directly to a different kind of interpretation of their meaning and significance than has been commonplace. In some of the older literature, dykes were somewhat romantically referred to as ‘wandering’ or ‘travelling’ earthworks because of the manner in which they weave their way across and through the landscape. To provide a novel interpretation of their meaning and significance, in the conclusions to this chapter I take up and explore further this metaphor of earthworks that in some sense ‘travel’.
Table 4.7 The relationship of cross dykes and spur dykes to escarpment edges.
| Map No. | Relationship to Escarpment Edge | ||||
| Ebble-Nadder Ridge (Fig. 4.2) | |||||
| 1 | On curving NE section of scarp | ||||
| 2 | On curving NE section of scarp | ||||
| 3 | Change of direction of scarp from NE- SW to W-E? | ||||
| 4 | Middle of straight W-E section of scarp jutting out to north | ||||
| 5 | Middle of indented straight section of scarp running NE-SW | ||||
| 6 | Middle of straight W-E section of scarp jutting out to north | ||||
| 7 | Middle of straight N-S scarp on East, Curving NE- SW scarp on West | ||||
| 8 | Change of direction of scarp from W-E to N-S | ||||
| 9 | On curving N-S section of scarp | ||||
| 10 | Change of direction of scarp from N-S to W-E | ||||
| 11 | Not applicable | ||||
| 12 | Not applicable | ||||
| 13 | Not applicable | ||||
| 14 | Slight change of direction of scarp From NE-SW to ENE-WNW | ||||
| 15 | Not applicable | ||||
| 16 | Middle of straight section of N-S scarp | ||||
| Ox-Drove Ridge (Fig. 4.3) | |||||
| 1 | Head of hollow in escarpment edge | ||||
| 2 | Head of hollow in escarpment edge | ||||
| 3 | Head of hollow in escarpment edge | ||||
| 4 | Not applicable | ||||
| 5 | Not applicable | ||||
| 6 | Side of hollow in escarpment edge to east, curving to west | ||||
| 7 | Head of hollow in escarpment edge | ||||
| 8 | Not applicable | ||||
| 9 | Middle of straight NE-SW section of escarpment edge | ||||
| 10 | Middle of straight NE-SW section of escarpment edge | ||||
| 11 | Not applicable | ||||
| 12 | Head of hollow to head of hollow in escarpment edge | ||||
It is extremely difficult to picture the purpose of these dykes. . . . They seem crazy meaningless things—monuments of apparently purposeless energy. (Curwen 1951: 100)
Curwen’s evident frustration with being able to understand cross dykes at all has probably been shared by most archaeologists, including me, attempting to interpret cross-ridge and spur dykes and, more generally, linear ditch systems. Cross dykes and spur dykes run across ridges and spurs from scarp to scarp, and their length is more or less determined by the width of the ridge or spur that is crossed. It is interesting, in this respect, to note that these earthworks are exceptional in the literature insofar as they are actually defined by their relationship to the local topography.
Excavations and careful fieldwork have demonstrated that most cross dykes and linear ditch systems are of late Bronze Age to early or final Iron Age (for example, Bradley, Entwistle, and Raymond 1994; Ford 1982; Fowler 1964; Rahtz 1990; Spratt 1982; 1989; Stone 1934; Wacher 1957). In some areas, where they occur, there is a clear association between clusters of cross dykes and hillforts, with the former perhaps serving as outer earthworks. In other areas, this direct association is much weaker or absent altogether. Some dykes are associated with later Iron Age and Romano-British settlement complexes. Some seem to have continued in use over a considerable period until the early Roman period. In a general review, Bradley and colleagues (1994) have noted the relative dearth of settlement evidence on the chalk downlands of southern England in the Late Bronze Age apart from metalwork. However, there is ample evidence for early Bronze Age activity in the form of barrows. Middle Bronze Age settlement appears to be limited to certain specific areas such as central Cranborne Chase, and the chalk uplands as a whole were occupied on a limited scale. The construction of the cross dykes may thus be associated with a new intensive occupation of the downlands during the final Bronze Age and early Iron Age.
Colt Hoare was one of the earliest antiquarians to systematically record linear ditch systems and cross dykes. He suggests that ‘some were designed for boundaries, and others for lines of communication between the British villages’ (Colt Hoare 1812: 244). He distinguishes between two classes:
Colt Hoare effectively set the intellectual agenda for much of the research well into the twentieth century. For Cranborne Chase in southwest Wiltshire, he records most of the cross dykes and ditch systems known today along the Ebble-Nadder and the Ox-Drove ridgeways (ibid.: map of Fovant Station VIII and Hindon Station IX, some of which have since been totally or partially destroyed). Working almost one hundred years later, Sumner (1913) provides the earliest detailed descriptions and plans of the cross dykes, again working in Cranborne Chase. He suggests that their purpose was military in nature. Those at Burcombe and on Buxbury Hill were to ‘guard against an enemy coming down the ridge toward the valley’, because the banks are positioned on the northern side (Sumner 1913: 63). The dyke on Buxbury Hill is suggested to be far-flung outer defences of the Castle Ditches hillfort near to Tisbury, 2.5 km away to the northwest (ibid.: 64). One of the dykes on Swallowcliffe Down is suggested to be a defence connected with the Iron Age settlement. Others are argued to be later in date ‘thrown up by the Romano-British as a barrier to stop the oncoming West Saxon’ proceeding along the ridgeways from the east (ibid.: 64). The dyke crossing White Sheet Hill is instead interpreted as a boundary ditch or cattle stop rather than having a defensive purpose. Sumner thus provides two different explanations—the more massive cross dykes were for defensive purposes; the smaller ones served as boundary markers—and he dates these types to two different periods.
Curwen and Curwen (1917) were the first to identify and describe cross dykes on the Sussex Downs, terming them ‘covered ways’, the same term originally used by Colt Hoare (1812: 244) for those in south Wiltshire. Cross dykes are defined as consisting of either a single ditch with a bank on each side or a series of such banks and ditches running parallel with each other, passing from scarp to scarp. Of the sixteen they discuss, fourteen run directly across the chalk ridge, two across outlying spurs. Curwen and Curwen thus distinguish between cross dykes that cut across the main chalk ridge and ‘spur dykes’ that cut across spurs of higher land projecting out from the main ridge. They note that although ‘some of the earthworks keep a direct course, others bend for a reason unconnected with the surface of the ground’ (Curwen and Curwen 1917: 66). They suggest that most may date to the late Bronze Age, and they list a number of suggested interpretations:
Curwen and Curwen point out that most of these earthworks would have provided a very inefficient means of defence as compared with a single bank and ditch with one high bank thrown up on the inner side and that such an explanation cannot account for bends and deviations in most instances. Since they end on steep scarps, rather than continue down them, that the dykes might be tribal boundaries is also questionable (ibid.: 68). As barriers to prevent cattle straying they would have been fairly useless, because they do not form enclosures, and the more natural place for cattle enclosures to be found would be in the valleys. The authors conclude, following Colt Hoare, that the dykes must have acted as covered ways or roads of communication: ‘the fact that there is a bank on both sides suggests a desire to screen the ditch that lies between them, as if the fosse was the most important element and the centre of activity’ (ibid.: 69). That these earthworks were used as ways of communication across the downland ridges is suggested by the existence of tracks leading up the sides of the escarpment to some of them. The authors point out from excavation evidence that the earthworks were carefully and purposefully dug with even and regular banks on both sides, with the floors of the ditches cut into the chalk. They could not be the result of a haphazard cleaning of puddled mud from a track. The narrowness of the floor of the ditches, usually between 0.5–1 m, suggested the single-file passage of people and animals along them (ibid.: 72). These bank and ditch systems, the authors also note, cut across the downland ridges in areas presumed to be free of woodland between heavily wooded escarpments and coombes. They would thus provide protection and concealment in open areas where it would be most needed. Why travellers along the ditches should need to hide ‘cannot be answered satisfactorily’ (ibid.: 75).
But directly in contradiction to this interpretation of the dykes as track-ways, Curwen and Curwen also note that ‘the extraordinary steepness of the slopes on or near which the Covered Ways of Sussex and Dorset terminate, and the further fact that they so often pass over the highest portions of the Downs rather than across the dips, tell against the theory that the earthworks were thrown up as Ways’ (ibid.: 74). However, they suggest, the cross-dyke builders were tough and hardy people, ‘exceedingly muscular without the physical limitations of ourselves’, so they would not necessarily have taken the easiest solution for the siting of a road!
Clay (1927a) distinguished between hollow ways, or sunken roads, and cattle ways. The former are bounded by slight banks and lead in a slanting direction along the easiest gradient to the ridgetops, sometimes leading to inhabited settlements, possibly Saxon in origin (ibid.: 61). Cattle ways are ‘earthworks consisting of a ditch between two banks that usually run a perfectly straight course and connect the heads of two coombes by passing over the dividing ridge of down’ (ibid.). In relation to the dykes on Cranborne Chase, Clay argues that these ways are grouped within areas closely connected with early Iron Age habitations. He incorrectly claims that the numbers of such cattle ways ‘coincides with the number of opposing coombes that can be connected up; thus they are found together where there are numerous coombes and widely separated where coombes are scanty. The extent of an area is roughly four miles, the width of an area being the width of the downland’ (ibid.). He argues that they had no defensive purpose, could not be boundaries of tribal areas because of the close proximity of many. The presence of smooth faces and hard trodden floors, revealed by sections, suggested that they were worn smooth by their use as cattle ways along which cattle were driven in single file between grazing grounds, preventing the cattle from running over and damaging the crops growing in fields covering the chalk ridges (ibid.: 64).
Williams-Freeman (1932) distinguishes between ‘univallate’ cross dykes with a single bank and ditch and ‘bivallate’ forms with a single ditch between two banks, either of which may be large (over 3 m high from the crest of the bank to the bottom of the ditch), medium (between 3 m and 1 m in height), or small (ibid.: 24). Such cross-ridge dykes may be single, double, treble, or a multiple series of univallate or bivallate forms or form groups of similar or different types:
Their essential characteristics are well marked: their ends rest on the steepest slopes often at the heads of coombes, or upon large or small patches of thick impenetrable wood, and they all cross an old track, often the main ridgeway where one or two spurs with their secondary ridgeways have converged upon it. The track may pass through a simple gap—in no case defended—which may or may not be the original way through; or it may pass the end of the cross dyke; in some cases uncomfortably near the steep edge of the scarp. There can be no doubt that the position of nearly all cross dykes is eminently suitable for obstruction of the road. (ibid.: 25)
Williams-Freeman further suggests that single univallate forms were ‘defensive’, or, if small, merely ‘obstructive’ or ‘protective’ in function. Larger ones may have been designed to hold up travellers going along the ridges and for the purpose of demanding tolls for free passage. Single bivallate forms, he suggests, cannot have been roads intended for the passage of either humans or cattle because of their position between the steepest scarps. Instead, they may have functioned as seasonal or overnight cattle pens. In the cases of those examples with multiple ditches and banks, they might have been used for collecting and sorting the cattle, in which case they would have been provided with wattle fences, gates, bars, and posts. (ibid.: 33–34)
In a later paper, again discussing dykes on the South Downs, Curwen (1951) distinguished between (1) cross-ridge and (2) spur dykes. The former is described as an earthwork running across the downland between two opposing valleys. The latter crosses a spur between the edges of two valleys. Cross-ridge dykes may be univallate, bivallate, or single, multiple or spaced. Spur dykes are univallate with the bank on the downhill side. These can occur singly or as two or more parallel earthworks. Curwen still regarded the cross-ridge dykes as being sunken drove ways connecting pasture on both sides of a ridge of downland. By contrast, he understood the cross-spur dykes as ‘barriers’ or ‘toll bars’, since, he argued, it would be unnecessary to build a dyke to connect converging valleys.
According to Curwen, the only possible clue to the use of any of the dykes is the fact that paths or terraces are occasionally found emerging from the ends of their ditches. However, that such dykes might be actually constructed to screen, contain, or conceal a path running over a hill from one valley to another, he admitted, seemed crazy (ibid.: 101). He notes that the majority of the cross dykes run over the main chalk ridge, probably covered with scrub at the time of their construction, and away from areas with fields usually located on the spurs off the main ridge. So, for want of any other alternative explanation, he continues to regard them as being cattle (or even pig!) ways. Spur dykes did not act as covered ways but instead were intended ‘to control traffic ascending the escarpment by means of terraced tracks climbing the flanks of the spurs’ (ibid.: 107). In other words, they acted as barriers protecting the spurs and diverted traffic around them. But why this might be required is not explained.
Thus much early twentieth-century research came to be dominated by the idea that these earthworks were roads rather than boundaries or military works. They were primarily associated with controlling and managing the movement of livestock.
A more neutral and ‘scientific’ generic terminology of ‘linear ditches’ or ‘linear earthworks’, sometimes used to describe and link cross dykes to other forms of bank and ditch systems in the landscape, becomes commonplace in the literature only from the 1940s onward, and these terms were invented only in the early part of the twentieth century (Crawford 1953: 107). In much subsequent discussion, the study and understanding of cross dykes becomes linked to that of more extensive linear ditch systems, usually located off the steep chalk ridges. The relationship of some of these linear ditch systems with hillforts led to an alternative view that they may have acted as ‘ranch boundaries’ involving the large-scale enclosure of livestock during the early Iron Age (Bowen 1978; Crawford 1953: 107–111; Piggott 1942; see discussion in Bradley, Entwistle, and Raymond 1994).
Fowler (1964) provides the most recent discussion of cross dykes in Cranborne Chase, reviewing the evidence for the Ebble-Nadder ridge. His argument is essentially ‘transitional’ in character between the older and the newer literature on the subject in so far as he suggests that some of the dykes functioned as territorial boundaries; others were for controlling the movements of people and animals. He argues that the bivallate cross-ridge dykes are primarily land boundaries, not tracks. However, he does not rule out a secondary purpose for them as protective earthworks or tracks. He argues that univallate dykes are in some cases parts of track systems. In other cases, they diverted and controlled tracks leading to and from the ridge (ibid.: 46). He notes that all the dykes that cross from one side of the ridge to another, apart from three on White Sheet Hill at the western end, which is effectively a spur, are bivallate in form. All but one of the dykes crossing spurs are univallate and associated with tracks or ‘terraced ways’. Fowler notes that the south ends of the six bivallate cross-ridge dykes are related to the end of a coombe and that the dykes therefore occur at points where the ridgetop is narrowest. However, not every coombe that runs up to the foot of the ridge has a cross dyke associated with it. Therefore, Fowler argues, that there must be an ‘artificial’ factor determining the position of these dykes, namely, distance (ibid.: 48). He argues that the bivallate dykes divide the ridgetop into distinct units of land and thus constitute boundaries between them, suggesting that the entire ridgetop was divided into four major units, each with subdivisions, giving a total of six or seven discrete downland units.
Although Fowler suggests that the bivallate earthworks had a single function as land divisions, he has much more difficulty in interpreting the univallate dykes and suggests a variety of purposes. Three possible examples at the west end of the ridge (two of which do not appear to be dykes at all) form part of a track system with ‘terrace ways’ issuing out of the ends of the ditch. These trackways, as elsewhere, on the ridge run obliquely up to these dykes. None connect with the bivallate dykes whose ends, where they continue, drop straight down the scarps. Three others, cutting across spurs, are suggested to have had the function of diverting and controlling traffic to and from the ridgetop. The others remain, to Fowler, inexplicable (ibid.: 50).
Influenced by Fowler’s work on Cranborne Chase, in a new study of dykes on the South Downs, Bradley (1971) called into question Curwen’s general distinction between cross-ridge and spur dykes, pointing out that several sites show earthworks of both kinds. He questions the functional association of cross dykes with pastoralism and spur dykes with blocking human traffic (ibid.: 9). Curwen’s argument suggests that the valleys should have been used as grazing land and the intervening ridges as arable or wasteland, but Bradley points out that there is no evidence for this. The supposed downland valleys, which should be pasture, nearly all contain field lynchets. On the South Downs, there are at least twelve cross dykes that do not communicate between valleys. In seven cases, they are linked to hillforts or enclosures, and an additional five turn through a distinctive double bend and are interrupted by an entrance (ibid.: 9).
Few trackways emerge from the end of Curwen’s ‘covered ways’, and most are probably not contemporary. Bradley goes on to point out that the toll bar theory is also unsatisfactory, since trackways flanking the ends of these dykes may well have been used after the dykes were constructed: ‘this establishes a sequence of events rather than a motive’ (ibid.: 10). Bradley instead argues for an affinity between the cross dykes on the chalk ridges and linear ‘ranch boundary’ ditches found elsewhere. Boundary ditches occur in plateau conditions, for example, on Salisbury Plain and in more low-lying areas of Hampshire and Berkshire, whereas cross dykes occur where the ground is steeper: ‘tracts of unploughed land on the ridges now seem to be divided not by roads but by land boundaries, which are linking the heads of valleys simply from economy of effort. These are the ranch boundaries of a hilly terrain’ (ibid.: 11). Bradley’s interpretation of the pattern is effectively the reverse of Curwen’s. The ridges of the South Downs probably acted as areas of pasture, while the land below was cultivated. The ridges are those areas where the soil is thinnest and most unsuitable for cultivation. He argues that enclosures associated with some of the Sussex dykes were pastoral in nature and associated with stock raising. Some early Iron Age ‘hillforts’—often with demonstrably slight ‘defences’ and few traces of domestic activity within them—may also have functioned as pastoral enclosures. Bradley suggests that the pattern on the South Downs suggests ‘an enclosed landscape of independent and self-sufficient communities’ (ibid.: 14).
From the 1970s onward, cross-ridge and spur dykes, as a separate category of earthworks, have been little discussed, and most of the literature has concentrated on linear ditch systems whose distribution is far more extensive. Either explicitly or implicitly, cross-ridge and spur dykes have been unhelpfully assimilated into a wider category of linear monuments or linear ditches. In tandem with the dominance of functionalist or ‘processual’ explanations in archaeology, both cross dykes in particular and other linear earthworks in general become regarded as having a primary economic significance. They are now variously described as being used to define or bound tracts of arable land, divide arable land from pasture, enclose tracts of uncultivated land, or act as territorial boundaries. Some particularly long and/or more massive linear bank and ditch systems are understood as major sociopolitical or tribal boundaries (see Bowen 1978, 1990; Bradley 1978; Ford 1982; Fowler 1981, 1983; Spratt 1982).
Fowler refers to Ebble-Nadder ridge of Cranborne Chase in terms of settlements existing in a context of ditches, field systems, and access networks, although lacking in cemeteries: ‘the landscape appears to be thoroughly under control while intensively exploited’ (Fowler 1983: 63). The cross dykes subdivide the land into agricultural units and acted to control cattle:
Cross-ridge dykes, traversing ridges and cutting off spurs in Sussex, Hampshire, and south Wiltshire seem to be best explained in terms of both territorial divisions and control of stock movement, in particular to keep livestock off arable fields. . . . Most seem to be related to the filling up of the landscape. . . . Sometimes they appear to block off one area from another; in many other cases they wind for considerable distances across country, and could have served as either or both barriers or trackways. (ibid.: 192–193)
Following Bradley (1971), Fowler argues that many of the large univallate enclosures in southern England (hillforts) may have originated as cattle enclosures. He singles out four different situations, all related to stock management and potentially leading to the local preeminence of a local hilltop:
Bowen similarly argues that large areas of the landscape were subject to orderly arrangement in association with large-scale pastoral management or ‘ranching’ (Bowen 1978: 120). Some linear ditches were ways permitting passage through arable fields to be channelled. Others were built over and not around fields, putting them out of use (ibid.: 122). Longer ones are suggested to be major land boundaries. He points out that some Banjo enclosures of the later pre-Roman Iron Age almost always have long ditches extending away from their entrances, making large enclosures directly related to the banjo.
In his extended discussion of linear bank and ditch systems in southern and central parts of Cranborne Chase, where it should be noted, cross-ridge and spur dykes do not occur, Bowen defines three main types:
Some of the linears are closely associated with local settlements and fields. A few with looped ends may suggest containment or penning (Bowen 1990: 11). Bowen argues that ‘the fields traversed were almost certainly put out of use at the time [or]. . . their arable use must have been seen as less important than the function performed by the new linear’ (ibid.: 12). The major spinal linears are not closely connected with the hillforts and appear to represent an organisation of the landscape predating them (ibid.: 13).
This massive change in emphasis on land use from arable to pasture, associated with the construction of cross dykes on the chalk ridges and linear ditch systems elsewhere, has been recently challenged by Bradley and associates (1994), who did not find much evidence in their study region to show that the ‘boundary’ works cut across existing fields. The first linear ditches on Salisbury Plain enclosed a pattern of open settlements, not tracts of empty grassland. There is no indication that the Late Bronze Age economies differed substantially from those in the Middle Bronze Age, and it is only in the early to middle Iron Age that linear ditches seem to define areas of grassland—but this definition occurred in the context of an intensively farmed arable landscape (Bradley, Entwistle, and Raymond 1994: 150). There were clearly distinctive phases in the ditch system.
Molluscan and stratigraphic evidence for the Wessex chalk downlands suggests a common history of woodland clearance giving rise to largely open grassland conditions during the early Bronze Age (for example, Allen 2000; Entwistle and Bowden 1991; Evans 1972). The landscape of Salisbury Plain also seems to have been largely cleared of woodland by the time the linear ditches were established in the Late Bronze Age (Entwistle 1994). The evidence for Berkshire suggests an open grassland environment immediately after the earthworks were built, with later land use being more varied (Ford 1982: 17).
For Berkshire, Ford suggests that linear ditches define a series of valley-based territories whose boundaries follow the ridges overlooking the steepest ground. For the Tabular Hills of northeast Yorkshire, Spratt has argued that major territories are defined initially by a series of round barrows located along watersheds, later supplemented by linear earthworks. Major early dykes run off the northern scarp toward the valleys, and together with scarps and watercourses divide the land into ‘estates’. These he suggests date to around 1000 b.c.e. and are associated with, but are later in date, than round barrows. Each ‘estate’ integrated a number of different ecological zones with upland grazing, lowland fields, access to water, and meadow grazing (Spratt 1989). Cross-ridge dykes, some of which may date to the end of the first millennium b.c.e. subdivide the ‘estates’ (ibid.: 12). Dykes are absent where there are valleys suitable for boundaries without the subdivision of dykes or where dykes have been converted into tracks and roads. Many of the dykes run from the northern scarp into the heads, or along the sides of valleys opening southward to the Vale of Pickering. In some areas where the valleys are dry, the dykes run along the crests of the valley sides, so there were no boundaries along valley bottoms. This situation is paralleled in some parts of the Berkshire Downs, where long linears are more or less aligned along ridges or the edges of steep slopes (Ford 1982). It has been suggested by both Ford and Spratt that in some cases cross dykes represent continuations of, or link to, ‘natural’ boundaries in the landscape such as coombes, valleys, and streams across higher land. Thus when natural boundaries were not present, cultural ones might be created instead. This work more or less represents the limits of the discussion of the significance of the natural topography in the entire literature on the subject.
Cunliffe (1990) draws attention, as does Bradley (1971) and Fowler (1983), among others, to the association of some linear ditches and cross dykes with hillforts. He suggests that the overall pattern in Hampshire is one in which early hillforts constructed in the sixth and fifth centuries b.c.e. were frequently preceded by earlier phases of enclosure related to existing systems of linear boundaries (ibid.: 329). Certain focal points located on systems of liner boundaries were chosen for enhancement by the creation of enclosures in the Late Bronze Age, and some later were developed into hillforts. He distinguishes between hillforts placed at the end of major linears (for example, Danebury) and those at nodes where major spinal linears were joined by subsiduary linears (for instance, Quarley Hill). Cunliffe develops a narrative for the period in which around 1000–800 B.C.E. a major programme of land division was begun involving the digging of a massive system of linear ditches across the chalk landscape. The land was divided and controlled in a radically new way. Enclosures associated with the linears were probably related to communal stock raising and control. From 800–550 B.C.E., strongly defended forts were constructed often on ridge ends and associated with artefacts, suggesting an elite occupation along with enclosures on the linear ditch systems and the emergence of new settlements, some defined by ditches at previously unoccupied locations. In a final stage between 550 and 350 B.C.E., some of the focal locations chosen for enclosure were heavily fortified in a new phase of construction (ibid.: 333–334). Cunliffe interprets this process as involving a radical re-organisation in land ownership from communal to ‘private’ ownership. The early linear ditch systems were primarily associated with stock control with the animals held or run in common along with a need to define territories. The colonisation of focal points was an attempt to establish authority over the land by individuals and/or lineage groups. Prestige moved from the acquisition and control of rare goods such as gold, amber, faience, and so on to control of the land and its productive capacity. Bradley (1994) points out that Cunliffe’s argument is in fact not based on a consideration of the boundary system itself but on its assumed relationship to ‘defended’ sites (hillforts) that Cunliffe regards primarily as being high-status settlements. This clearly runs counter to the interpretation that some hillforts were only temporarily occupied or had other probable functions associated with cattle management. On Salisbury Plain, the settlements associated with the linear ditches are open sites showing no evidence of hierarchy (ibid.: 150).
In their discussion of linear ditches on Salisbury Plain, Bradley, Entwistle, and Raymond (1994) argue that the use of these earthworks took place over such a long period of time that no single interpretation for their function is possible. They demonstrate these earthworks’ establishment in the Late Bronze Age in an area with a long history of settlement. The construction of the ditches, they argue, formalised land boundaries that may have already have been present from the middle Bronze Age or earlier. Larger territorial divisions evolved before the foundation of hillforts, reflecting the emergence of the social and economic structures out of which hillforts and ditched enclosures developed (ibid.: 137). They suggest that in the final stages of this process much of the ditch systems seems to have become redundant, and during the Iron Age some ditch systems were incorporated into a closely integrated system of organised fields.
Despite Bradley and associates stressing the need for multiple interpretations of the ditch systems on Salisbury Plain, however, this position only appears to be the case, in the accounts that they give, through time. At the period of their initial establishment, all ditch systems appear to have been attributed the same function: they acted as boundaries. Bradley and associates argue that the linear ditches defined areas of settlement, including pasture and arable fields. These ditches emphasised the alignment of the main ridges and watersheds, as did the distribution of earlier Bronze Age round barrows clustering on high ground and along watersheds. Some linear ditches are aligned on prominent barrows or barrow cemeteries (ibid.: 141). These, Bradley and colleagues argue, may have been part of an already existing but far less formalised territorial system that became supplemented and extended by the physical presence of a continuous boundary. Such a boundary had sociopolitical significance. It was not a barrier to movement or cultural interaction and was not related to the creation of large areas of grazing ground or an economy with a greater emphasis on pasture. The boundary earthworks enclosed a pattern of large open settlements rather than empty areas of grassland. It is only later in the early to middle Iron Age that parts of the boundary system define pasture areas, but only in an intensively farmed arable landscape (ibid.: 150).
Two very significant general points seem to arise from this review. First, there is a strong tendency in the literature to try to seek a single explanation, or set of explanations, for all cross dykes and linear ditch systems wherever they might be found in England. Instead, with Bradley and colleagues (1994), I think there is a need to acknowledge that cross dykes and spur dykes are very different types of monuments and do not really fit into a broader category of linear monuments at all. Both they and linear ditch systems meant different things in different areas and at different times. There can be no overall explanation. What these monuments meant, I argue, depends very much on their relationship to the surrounding landscape. Consequently, what happened on Salisbury Plain is not necessarily relevant to anywhere else. We need highly specific interpretations.
Second, cross dykes have always been understood in the most general sense as either dividing the land, acting as a marker or a barrier, or funneling movement across it from one place to another as specific kinds of pathways. If the latter was the case, where did these pathways lead; what did they link together, and why?
A few of the dykes in the study area of northern Cranborne Chase have been excavated. Four bivallate dykes were sectioned by Clay (1927) along the Ebble-Nadder ridge. On the floor of the ditch of the dyke running diagonally across Swallowcliffe Down, he recovered a fragment of pottery of early Iron Age date (ibid.: 63; Figure 4.2: No. 9). He concludes that the dyke was constructed at the same time as the nearby Iron Age palisaded settlement. This dyke was also either earlier than or contemporary with ‘Celtic’ fields immediately to the southwest. Two short dykes, one to the north and one to the south, are linked with the early Iron Age hillfort of Chiselbury (Figure 4.2: No. 6). From surface inspection, because of plough damage and disturbance, it is impossible to tell whether they are contemporary with this dyke or perhaps earlier in date with the eastern rampart of the hillfort following the line of an earlier single cross-ridge dyke. Fowler suggests that this dyke (or dykes) is either contemporary with or later than the hillfort on the basis that the original line of the dyke would have had a considerable bend or kink in it if it had been built earlier (Fowler 1964: 53). However, bends and kinks are not an uncommon feature of cross dykes (see Table 4.4). Fowler’s excavation of a section across the dyke running across Buxbury Hill provided no direct dating evidence, but the presence of sherds of ‘Romano-British’ type in an upper level of the ditch section late in the sequence of ditch deposits provides a terminus ante quem, perhaps again suggesting an early Iron Age date (Fowler 1965: 49).
Along the Ox-Drove ridge, the bivallate cross dyke to the west of Win Green is clearly cut through by the Roman road and is earlier in date. Sections cut through Great Ditch Banks and Middle Chase Ditch at the far eastern end of the ridge have dated them to the late pre-Roman Iron Age just before the conquest (Rahtz 1990). But these are not true cross dykes, because they do not extend to the ridge scarp to the north. However, they do seem to form a coherent system, being approximately equidistantly spaced, with a cross dyke immediately to their west that has not been excavated.
The dykes are quite consistent in size. The univallate dykes are about 10 m–12 m in overall width, and bivallate dykes are 12 m–15 m wide. The largest bivallate dyke on the Ebble-Nadder ridge is 22 m in overall width (Figure 4.2: No. 14). Their length is highly variable and partly determined by the relief. None have any original breaks unless the old turnpike roads running along the Ebble-Nadder and Ox-Drove ridges pass through an original break in every case, which seems unlikely. Both these ridges do not appear to have been used as long ridgetop trackways at the time when the dykes were constructed and used. Movement seems rather to have been up and over the ridges.
Today banks are about 0.5 m–1.2 high, and ditches are of a similar depth (see Table 4.3), but most are considerably denuded or entirely destroyed in long sections where they cross the ridgetops. Sometimes they survive only as ‘tails’ on the steep scarps, which are impossible to plough. Overall widths range from 10 m to 20 m. The best preserved examples to the east of White Sheet hill on the Ebble-Nadder ridge (Figure 4.2: No. 14) and on the western side of Win Green on the Ox-Drove ridge (Figure 4.3: No. 7) are still fairly formidable monuments with a drop between the top of the banks and the bottom of the ditches of about 3 m and very real barriers to movement along the ridgetops.
One important distinction between the bivallate and the univallate dykes is in their profile and dimensions. All the four bivallate dykes on the Ebble-Nadder ridge sectioned by Clay (Chiselbury Figure 4.2: No. 6; Row Ditch Fig 4.2: No. 8; and at Swallowcliffe Figure 4.2: No. 9 and No. 13) were less than 1.5 m deep and had a distinctive V profile (Clay 1927: 62–63). Fowler’s section through the univallate cross dyke running across the spur of Buxbury Hill (Figure 4.2: No. 7) showed that the ditch was 5.4 m wide at the mouth, 1.8 m deep, with a 1.2-m-wide flat bottom (Fowler 1965: 49). Excavations at Great Ditch Banks, a univallate bank and ditch system at the eastern end of the Ox-Drove ridgeway, revealed that it had a ditch as much as 3.5 m deep with a V-shaped profile with a bank on one side that might originally have been 3 m or more high, giving a total in excess of 7 m, which would have been a formidable barrier (Rahtz 1990: 11). The nearby Middle Chase Ditch was of similar profile and dimensions (ibid.: 22). Hence it would be unwise to conclude, as Fowler seems to suggest (Fowler 1965), that the ditches of univallate cross dykes had flat bottoms and were more massive. In the present discussion, what seems to be more significant is that the V- and U-shaped profiles of the ditches are mirrored in the V- and U-shaped profiles of the coombe ‘ditches’, some with distinctive but always narrow flat bottoms, others without. The bivallate cross dykes, which most closely resemble the coombes, almost always run between the heads of the coombes, thus continuing their lines over the ridgetop to the escarpment edges. By contrast, the univallate cross dykes, which with only a single bank and ditch do not resemble the forms of the coombes, link escarpment edge to escarpment edge or cut across spurs or at right angles to the coombes rather than continuing their lines over the ridgetops. Both their form and positioning in the landscape contrast with the coombes rather than mirroring them. There is only one exception to this general distinction, where a bivallate cross dyke cuts across a spur on Swallowcliffe Down, linking two coombes.
DYKES ALONG THE EBBLE-NADDER RIDGE Burcombe (Figure 4.2: Nos. 1–3: Figure 4.10) The first two dykes at the eastern end of the ridge run parallel to each other, only 58 m apart, cutting across a low spur projecting from the northern escarpment. The northern dyke is 64 m long with a 1-m-high bank situated on the northern downhill side. The southern dyke is 160 m long and similarly has a single bank situated on the downhill side of the ditch. The dykes are situated on the mid-point of a fairly gentle slope with the land rising above them. The shorter northern dyke is almost straight, terminating at its western and eastern ends at the lip of much steeper slopes cut into by trackways and holloways. The southern dyke is markedly more curved and terminates farther down the lip of the slope at its western end.
FIGURE 4.10 Punch Bowl Bottom Coombe at the eastern end of the Ebble-Nadder ridge. Dykes Nos. 1 and 2 (see Figure 4.2) cut across the spur to the right of the coombe end below the clump of trees, which marks the position of a large round barrow terminating where the scarp slope to the coombe bottom becomes precipitous.
Both dykes cut across a north-running spur. This spur is dramatically defined by a tongue-shaped coombe, Punch Bowl Bottom, widening rather than narrowing at its end (the latter being characteristic) on the eastern side and by the escarpment edge, which swings round to the north on the western side (see Figures 4.2 and 4.5). This narrows the spur at first, hence the much shorter length of the northern dyke, before it widens out and flattens to form a low plateau to the south of the river Nadder. From the end of the southern dyke, which terminates below the lip of the very steep slope running down into the coombe, the base of the coombe below is visible. The northern dyke ends farther up the slope, and the base of the coombe immediately below it is not visible from its end. Punch Bowl Bottom, together with another slighter and much narrower coombe just to the east, is the only coombe to cut into the northern escarpment of the Ebble-Nadder ridge. All other coombes cut into the ridge on the southern side.
The eastern ends of both dykes are visible along almost the entire length of Punch Bowl Bottom, from where it begins just to the south of the river Nadder, and they appear to have been positioned so as to be seen in the distance while one was moving along the course of the coombe toward them. The southern dyke runs down to the very head of the coombe, so as to give an impression of being a natural continuation of it, whereas the northern dyke is positioned on the western (right) side of it. Both dykes link the head of the coombe with the escarpment edge to the west. From their western terminal ends, one looks out across the plain below and along the impressive and unbroken line of the northern escarpment. From the western end of the southern dyke, one can look down to the base of the escarpment below. This is not the case from the northern dyke, which terminates higher up the slope.
About 300 m to the south of these two cross dykes, there was probably a third much longer dyke running west-east for about 750 m across the crest of the ridge. This earthwork was marked by Colt Hoare (1812) on his Station VIII map, but he did not describe it. The western end began on or near to the shoulder of the scarp but does not appear to have run down it just to the south of three round barrows unusually sited on a marked incline just above the steep scarp slope. There is no trace of this earthwork on the ground now, but Fowler was able to note its presence in the early 1960s (Fowler 1964: 54). It then ran across the ridge summit just to the south of another round barrow on the highest point and then descended, turning somewhat to the north to terminate at the head of a shallow meandering coombe cutting into the scarp adjacent to the coombe to which Burcombe 1 and 2 are linked at their eastern ends. Here there are slight visible traces in woodland that may be the remains of this dyke. The two adjacent coombes are not visible from their heads at the scarp shoulders; neither are the three cross dykes, except at their western ends. The two coombes are markedly different in form. That to the east is sinuous, shallow, and meandering, that to the west much wider, deeper, and flat bottomed and far more striking topographically. The southern end of the eastern coombe, into which Burcombe 3 probably ran, narrows to effectively the dimensions of a dyke itself so that the eastern end of Burcombe 3 would have created the impression of the coombe itself continuing on and out of sight across the hilltop to the west. In this case, the experiential effect of the dyke simply extended the coombe onward.
Compton Hut (Figure 4.2: No. 4) This cross dyke is shown by Colt Hoare (Station VIII map) as running in a meandering line between the head of Hut Bottom coombe to the south and the escarpment edge to the north. Today it is obliterated apart from a short length on the northern scarp where it runs down into Burcombe Ivers wood. Here it is cut diagonally by a hollow way running northwest to southeast up the slope. The preserved section is univallate with a bank on the eastern side, although Sumner (1913: 63) records the dyke as being bivallate in form. The bank runs out, and the ditch continues well down the steep escarpment beyond the point at which one can see the base below. The ditch line is virtually indistinguishable from a natural depression or gully in the escarpment edge. The ditch line’s precise relationship with the head of Hut Bottom Coombe cannot be verified, but at its northern end the coombe is shallow, straight-sided, and narrow (Figure 4.11). Again, as with Burcombe 3, the end of this coombe and its dimensions resemble a bivallate cross dyke, and the dyke would have run down a gentle incline to join it.
Figure 4.11 Northern end of Hut Bottom Coombe on the Ebble-Nadder ridge. The southern end of dyke No. 4 terminated at the head of this coombe, running over the ridgetop beyond.
Compton Ivers (Figure 4.2: No. 5) Two short stretches of a dyke remain today that originally cut right across the ridgeway, linking the northern escarpment edge with a narrow coombe to the south. The northern end runs steeply down over the lip of the escarpment to terminate about halfway down the slope and way beyond the point at which one can first see the base of the scarp below. This position, of course, would make the dyke highly visible from off and below the escarpment edge to the north. From this end, there are extensive views across the plain below. The southern end is unusual in that it at first runs diagonally down the side of a coombe on its western side before swinging round to the east to terminate just to the south of the head of the coombe at its side and almost down to the coombe bottom (Figure 4.12). This is a narrow, meandering, steep sided V-shaped coombe, thus contrasting significantly with Punch Bowl Bottom.
Chiselbury (Figure 4.2: No. 6) Immediately to the north and the south of Chiselbury hillfort two dykes extend out. The southern dyke runs to the head of a coombe; the northern dyke extends down the northern escarpment edge (Figure 4.13). These dykes may originally have been linked together under the line now followed by the eastern rampart and ditch of the hillfort or, alternatively, as Fowler suggests (1964: 53), have been contemporary with it, drawing a continuous line across the ridge. The northern end extends some way over the lip and down the precipitous slope of the escarpment, thus making it highly visible from the plain below to the north. At the southern end, the banks run out, but the ditch extends almost to the bottom of a very steep slope at the head of a relatively large and regular V-shaped coombe, its axis continuing the N-S orientational axis of the coombe (Figure 4.14).
Buxbury Hill (Figure 4.2: No. 7) Buxbury Hill is the only real prominent and significant steep-sided spur defined by steep escarpment edges on both sides (rather than a coombe scarp and escarpment edge as at Burcombe)
Figure 4.12 View south from the head of the coombe down to which the Compton Ivers cross dyke (Figure 4. 2: No. 5) runs. The dyke is visible to the right of the photograph, running down almost to the base of the coombe.
FIGURE 4.13 View of the northern escarpment edge of the Ebble-Nadder ridge seen from the north. The rampart of the Chiselbury hillfort is skylined, and dyke No. 6 can be seen dropping over the escarpment edge to the left of the photograph, terminating just above the line of the Fovant military badges.
FIGURE 4.14 View south from the southern end of dyke No. 6 down the coombe south of Chiselbury.
jutting out along the northern escarpment of the Ebble-Nadder ridge. Other parts of the scarp, such as Chiselbury, appear to be spurs when seen from the eastern side only, an impression created by the manner in which the ridge itself is orientated NE-SW. From the west, they appear merely as rounded jutting protrusions of the ridge rather than true spurs. The hill is cut across by a curved single bank and ditch at almost its lowest and narrowest point, with the land rising and widening toward the north to the end of the spur and also rising along the ridge to the south. The bank is on the south (downhill) side. From the western end, one can see down to the base of the scarp, and this end is orientated so as to look out along the line of the escarpment edge to the west. Much of the eastern end is much mutilated by a (later?) field system. The ditch continues some way down and just over the lip of the precipitous slope of the escarpment edge to the point at which the base is visible.
Row Ditch, Sutton Ivers (Figure 4.2: No. 8) This dyke, just to the southeast of the Buxbury spur, cuts across the ridge joining the northern escarpment to the head of a coombe to the south. The southern end of the dyke continues the north-south orientational axis of the coombe, which has a very narrow bottom and is steep-sided. Unusually, the dyke runs down the precipitous slope almost to the very bottom of the coombe (Figure 4.15). It thus appears to be a ‘natural’ continuation of the coombe. The dyke’s northern end runs over the lip of the scarp and roughly a third of the way down the precipitous slope before terminating, way beyond the point at which one can first see the base of the scarp (Figure 4.16). It runs to the northern scarp edge just before it turns to run out to the north forming the spur of Buxbury Hill.
Swallowcliffe Down (Figure 4.2: Nos. 9–13) Here the greatest concentration of cross dykes occurs on the ridgeway. Two dykes traverse the ridgetop. Another three cut across spurs jutting out from it between coombes to the south. The only comparable situation where this occurs in the study area is on Berwick Down along the Ox-Drove ridgeway. The longest of these dykes (No. 9) is bivallate. It cuts across the ridge, at right angles to it, running roughly west to east for c. 360 m. The western end descends over the shoulder of the scarp and terminates just below the point at which it descends precipitously, from which the bottom of the scarp is visible below (Figure 4.17). The eastern end similarly terminates on the lip of an extremely steep scarp on the side of the head of a deep, wide-topped and shallow bottomed coombe (Figure 4.18) at the very head of which, on the ridgetop, the late Iron Age Swallowcliffe Down open settlement is situated (Clay 1925, 1927). The dyke terminates at precisely the point at which the base of the coombe is visible directly below it. Before this point, the coombe base, running away farther to the south, is visible. The terminal is clearly visible from the base of the coombe.
FIGURE 4.15 View toward the head of the coombe at which the southern end of the Row Ditch cross dyke (Figure 4.2. No. 8) terminates. This dyke runs down the scarp slope almost to the bottom of the coombe.
FIGURE 4.16 Northern end of the Row Ditch (Figure 4.2. No. 8) cross dyke seen from spur dyke No. 7 crossing Buxbury Hill to the northwest.
FIGURE 4.17 Western end of the Swallowcliffe Down (Figure 4.2. No. 9) cross dyke running down and just over the lip of the northern escarpment edge of the Ebble-Nadder ridge.
FIGURE 4.18 View down the coombe, looking toward the south, from the eastern terminal end of cross dyke 9 on the Ebble-Nadder ridge.
No. 10 is univallate with the bank to the east. It runs across the ridge from the NW-SE. Its northern terminal ends on the steep scarp below the shoulder, and one can see the scarp bottom below. To the south, Colt Hoare (1812: Station IX map) depicts it running to the side of the head of a coombe, but there is no trace of this today.
No. 11 is unusual in that here a bivallate and a univallate dyke run side by side to the head of a wide, deep, V-sided coombe with a narrow flat bottom. They run down the slope and terminate just at the point at which the scarp becomes very steep. The bivallate dyke, situated immediately to the north of the univallate dyke, ends at precisely the point at which one can see directly down into the coombe bottom below it. The univallate dyke terminates a little farther up the slope, and from its end the coombe bottom immediately below is not visible. Both dykes are visible from the base of the coombe to the south and continue its line up and over the ridgetop, where all traces have been obliterated. The western terminal of one of these dykes is visible on the side of another coombe to the west, where it runs just over the shoulder of the steep scarp to the point at which the base of the coombe is visible below.
No. 12, a short distance to the south, is a univallate dyke with its bank on the south side. It runs west-east across a spur connecting two coombes. Its western end runs just over the shoulder of the scarp, and from it the coombe bottom below is visible. Opposite it, on the other side of the coombe, No. 13 continues the line across another spur to another coombe. The eastern end terminates just short of the shoulder of the scarp. From it the coombe bottom is not visible, nor is this dyke visible from the bottom of the coombe.
No. 13 is a bivallate dyke crossing a spur and linking two coombes. Its eastern end terminates at the side of a coombe opposite No. 12, well past the point at which one can see the coombe bottom on precipitous slopes (Figure 4.19). Both it and No. 12 can be seen from the bottom of the coombe below. The western end terminates at the side of the head of another coombe. The banks stop, but the ditch continues over the shoulder of the slope to terminate on the steep scarp from which the coombe bottom below is visible running into a natural gully that continues the ditch line down to the coombe bottom.
Half-Mile Ditch (Figure 4.2: No. 14) A few hundred meters to the east of the highest point of White Sheet Hill there is another dyke, which is the largest of all the cross dykes on this ridge. It runs north to south at right angles to the ridge connecting the northern escarpment edge, with a wide coombe to the south. The dyke runs down the precipitous escarpment edges at both ends some considerable way. The southern end running down the head of the coombe is particularly dramatic (Figure 4.20). Here there are views across a wide coombe, the Ox-Drove ridge to the south, and
FIGURE 4.19 Western terminal end of cross dyke 9 crossing over the top of Swallowcliffe Down on the Ebble-Nadder ridge and terminating just beyond the shoulder of the scarp.
FIGURE 4.20 Looking down Norrington coombe from the southern end of Half Mile Ditch (dyke 14) on the Ebble-Nadder ridge.
to Winkelbury Hill. Walking along the line of the ditch, one sees that the entire landscape along the ridge to the west and east is entirely blotted out. From the eastern end, Castle Ditches and Castle Ring hillforts are visible and roughly equidistant from a line drawn outward across the landscape from the end of the dyke.
Berwick Down (Figure 4.2: No. 15) The length and form of the dyke at the head of Berwick Coombe remain uncertain. This dyke occurs just below the highest point on White Sheet Hill, and its southern end terminates dramatically at the very head of Berwick Coombe, extending some way down the precipitous escarpment edge. This cross dyke is aligned with the head of the coombe rather than joining one side of it, and from it one can see down to the base of the coombe.
White Sheet Hill (Figure 4.2: No. 16) White Sheet Hill forms the higher western end of the Ebble-Nadder ridge. Here the escarpment edge swings prominently to form a spur running approximately north-south before terminating 500 m to the northwest of Berwick St. John. To the west, the greensand plain extends toward Shaftesbury, while Berwick Coombe cutting in to the ridge from the south forms its eastern side. Across the middle of this spur runs a prominent univallate cross dyke, with its ditch to the north, linking the escarpment edge with the western side of Berwick Coombe. The eastern end is mutilated but appears to run down the gentle higher parts of the scarp. The base of the coombe is not visible. The western end terminates as a ditch high up the slope, and again the base of the escarpment is not visible. Neither end of this dyke ends dramatically, and Berwick Coombe itself is shallow and wide and very different in character from some of the coombes at the western end of the Ox-Drove ridge to the south.
DISCUSSION Of the sixteen cross dykes, five cut across spurs but in very different ways. The dyke on White Sheet Hill cuts straight across the most prominent southern spur of the entire ridge, dividing it in two. That on Buxbury Hill cuts and marks off the only well-defined northern spur of the ridge. The three Burcombe dykes link the only two coombes to cut into the northern escarpment, with the scarp farther to the west. All these dykes, with the exception of the southernmost at Burcombe, are univallate. The greatest concentration of dykes is those on Swallowcliffe Down. Three cut across spurs to the south and link two parallel coombes, and two other link coombes to the northern escarpment. Their concentration here seems to be directly related to the presence of the Swallowcliffe Down Iron Age settlement. Dyke No. 9, unusually orientated northwest to southeast, seems to run in this direction across the ridgetop in order to avoid cutting through the settlement area adjoining the left-hand or western side of the head of the coombe obliquely. None of these dykes, except possibly No. 10, continues directly the line of the coombe over the ridgetop. All the other major bivallate cross dykes (14, 8, 6, 5, and 4) do so. No. 5 is unusual in running down the side of the head of the coombe to the left (west). None, apart from No. 8 appears to have had a particularly straight course, but instead they meander, a little like the courses of many of the coombes themselves.
From White Sheet Hill to Burcombe, ten coombes and interlinked coombe systems cut into the chalk ridge from the south, subdividing the southern slopes of the ridge into a series of spurs. As noted, the coombes running in to the ridges from the south have a hidden character, and when one walks along the centre of the ridge one is scarcely aware of their presence and certainly not of their depth and extent, form, and character. This becomes apparent only when one is standing on the coombe lips. Each of these coombes is distinctive, with its own individual character. The two most westerly of these, Berwick Coombe and the coombe to the west of Norrington (Figure 4.20), are relatively broad, simple in form, and open, being as much as 750 m wide from shoulder to shoulder on either side and with a fall of height of about 100 m from the ridgetop to the coombe bottom. They are comparatively short. The other coombes to the east are considerably more complex, with numerous meandering side branches, and they tend to become progressively shallower, narrower, and longer from west to east. They provide natural and ready-made divisions of the chalk downlands. Two of these coombes, in particular, stand out from the others— the one immediately to the south of Swallowcliffe Down (Figure 4.18) and the coombe to the south of Chiselbury (Figure 4.14). These two coombes are distinguished by the sheer steepness of their sides and their regularity in form. Both have almost flat bottoms about 10–15 m in diameter and are about 300 m wide from shoulder to shoulder, with a drop in height of about 100 m from the ridgetop to the lowest part of the coombe floor. The line of Swallowcliffe coombe curves away gently, in an almost perfectly smooth line from the northwest to the southeast. The northern part of Chiselbury coombe is straighter and more V-shaped in side profile. These two particularly dramatic coombes come closest to cutting right through the ridge to the northern scarp. It is perhaps not surprising that the only two known settlements to occur along the ridgetop, the open settlement on Swallowcliffe Down and the enclosure, or ‘hillfort’, at Chiselbury, are situated immediately above the heads of these coombes and are associated with cross dykes.
Only particular coombe systems and branches of coombes have cross dykes at, or toward, their ends. Three out of the ten coombes have no cross dykes. These are all relatively narrow and shallow and end farthest away from the northern scarp. What determines the spacing of the dykes appears to be that they should be approximately the same distance apart, together with their relationship with a coombe or a coombe system of distinctive form and character.
The relationship of these cross and spur dykes to the escarpment edge is summarised in Table 4.7. The majority occur either in places where the scarp edge changes in direction or in the middle of more or less straight sections of the scarp, but there appears to be no overall pattern to this relationship, and the relationship of particular dykes to particular coombes appears to be of far greater significance. There is a much stronger association between the locations of the dykes and topographic features of the escarpment edge along the Ox-Drove ridge (see below). The majority of the dykes are visible from the land below to the north, because they extend beyond the shoulder of the escarpment and part of the way or well down the slope (see Table 4.6). The pair of cross-spur dykes running across Burcombe Hill cannot be seen from the north because of a projecting spur. However, they are clearly visible from the NW and the NE. From the NE, the end of No. 3 would have been visible, as are the three round barrows situated close to its probable western end today. They were meant to be seen and when newly constructed would have stood out as gleaming white. A problem here, though, is to what extent the scarp slope was heavily wooded. Some areas, generally the lower slopes, of the scarp are still heavily wooded today. We might expect woodland coverage to be far greater in the past, but building the dykes would have required removing the trees, and their removal would still be likely to stand out as gaps on the upper slopes.
South Down (Figure 4.3: No. 1) On South Down, remnants of what was once a very substantial bivallate dyke cross the ridgeway in a meandering NE-SW course. The dyke runs from the head of a well-defined hollow in the northern escarpment, continues over the top of the ridge, and ends on a gentle slope to the south. Its northern end joins the hollow on its left-hand (western) side rather than at its centre and runs down the escarpment edge to the point at which it becomes extremely precipitous and the base of the slope is visible.
Pincombe Down (Figure 4.3: No. 2) Two kilometres west, another bivallate dyke cuts across the ridge in a similar manner, terminating on gently sloping land to the south and at the head of a steep hollow in the escarpment to the north. Again the northern end runs down the escarpment edge, whose base is visible from the end of the dyke. Here the dyke is in the approximate centre of the head of the hollow. Another 1 km to the west, another dyke (Figure 4.3: No. 3) runs down to the head of a hollow to the north, terminating on precipitous slopes, with its southern end terminating on gently sloping land to the east. The relationship of these three dykes to the ridge and the landscape is more or less the same. A very different situation occurs on Berwick Down an additional 2 km to the west.
Berwick Down (Figure 4.3: Nos. 4 and 5) Two univallate cross dykes, situated only 300 m apart, straddle the ridge of Berwick Down. Immediately to the south of the southern dyke, an Iron Age and Romano-British settlement occupies the highest part of the ridge, bounded on the western side by Ashcombe Bottom and by Malacombe Bottom on the eastern side (see Colour Plate 2). These are two narrow steep-sided coombes between which both cross dykes run roughly E-W across the ridge, connecting them. The southern dyke can be traced today only for about 80 m but can be seen terminating on the lip of the precipitous slopes of Malacombe bottom, an additional 90 m to the east. This is a particularly dramatic narrow and deep-sided, V- shaped coombe with a flat bottom little wider than a track (see Figure 4.21 and Colour Plate 2). The western end of the dyke extending toward Ashcombe Coombe ends much less dramatically on the gentle upper slopes leading down to the coombe, which is less narrow and confined in form.
The northern dyke (Figures 4.21 and 4.22) contrasts considerably insofar as its eastern end (now mutilated) terminates at the head of the coombe on its western side, and its western end extends much farther down the slope into Ashcombe Bottom, below the point at which it becomes precipitous. Here it is possible to see down into the very bottom of the coombes at both ends of the dyke, which is possible only at the eastern end of the southern cross dyke.
FIGURE 4.21 Looking along Malacombe Bottom from the eastern terminal end of cross dyke 5 on the Ox-Drove ridge.
Both dykes are intervisible and invert each other in form, the southern dyke having its ditch on the southern side and the northern dyke having its ditch to its north. The northern dyke cuts across a saddle of lower ground at the lowest point on Berwick Down, whereas the southern dyke traverses land gently rising to its south. When one walks along the best preserved stretches of the ditch of the northern dyke, one notes that the landscape is completely blotted out except at its terminal ends, where views are directed down toward the interiorised and secret world of the coombes (Figure 4.21).
FIGURE 4.22 Eastern terminal of cross dyke 5 on the Ox-Drove ridge on the left (west) side of the head of Malacombe Bottom.
Win Green (Figure 4.3: Nos. 6 and 7) On the eastern side of Win Green, 2 km to the west of the cross dykes on Berwick Down, another two dykes with a markedly different relationship to the topography occur. Win Green is the highest point of Cranborne Chase. Two dykes traverse the escarpment edge, the northernmost laterally and parallel with the main line of the ridge, the southernmost running between a deeply incised coombe to the south, across the main line of the ridge and dipping down the northern side of the escarpment. This is the most dramatic and better preserved of the two today.
The northernmost dyke, orientated approximately NE-SW, is badly mutilated at both ends, but it may originally have terminated shortly below the lip of a steep slope on the eastern side running down to a deep hollow in the northern escarpment edge NW of Win Green. From this point, it runs across a low spur extending to the NW of the main escarpment line to possibly terminate toward the end of another hollow on the western side. It consists of twin ditches with downhill banks and simultaneously links two hollows in the northern escarpment and cuts off a low spur running out from it. Situated approximately halfway down the hill slope, its positioning in the landscape is rather unusual, as is its form. It is asymmetrically sited, its eastern end pointing down toward the terminal point of the coombe, its western end terminating some way down the side of the coombe. The bottom of the escarpment edge was probably visible at the original eastern terminus; however, this was not the case at the western terminus, now variously destroyed by a chalk pit, the course of the Roman and modern road, and a sunken trackway.
The southern cross dyke is well preserved and in some respects inverts the features of the northern one. It consists of two banks with a medial ditch and runs approximately NW-SE, although the line of the dyke is markedly curved toward both ends. This dyke runs across the ridge and drops steeply down the slopes of the ridge at both ends to terminate around the 225 m contour at both ends, a drop of 35 m from the ridgetop. The dyke terminates at precisely the points at which the bottom of the escarpment edge and coombe to the north and south, respectively, become visible for the first time, strongly suggesting that a view to the bottom of both was important. This dyke links the end of a deep, narrow coombe, Quarry Bottom, to the end of a large hollow in the northern escarpment bounded by the ridge of Charlton Down to the west and the spur across the northern Win Green cross-dyke runs (Figure 4.23).
From the terminal point on the northern escarpment edge, there are wide and panoramic views across the plains below toward the ridge of high land to the northeast of Shaftesbury occupied by the Castle Rings hillfort 5 km away—to and from which we can suggest a probable line of movement to and from the escarpment edge. The view from the southeast end of the dyke is very different, being dominated by the sinuous lines of the coombes breaking up this part of Cranborne Chase, with only limited views to the distance. Both the NW and SE ends terminate next to well-defined ‘natural’ dykes or gullies breaking up the chalk escarpments and terminate on their sides toward the top on the west. In many respects, these natural dykes afford much easier paths of movement up and down the chalk escarpment than does the course of the cross dyke, which is extremely steep and difficult to climb at both ends. From the southeast end, no other prehistoric monuments are visible, whereas from the northwest end, the other Win Green cross dyke can be seen and possibly two round barrows in the distance.
When one walks along the ditch from one end of the dyke to the other, one notes that the wider landscape on both sides is dramatically blotted out. All that is visible is the landscape at the terminal ends. So, the dyke channels vision, and one cannot see across the chalk ridge at all. It is only at the terminal points of the dyke that the vistas widen out at all, and the contrast between the wide, expansive, and open views at the northwest end and the enclosed interiorised world of coombes at the southeast end could not be more marked, and one must, of course, pass over the top of the ridge before the landscape at either end becomes visible.
Figure 4.23 Northern terminal end of the Win Green South (cross dyke No. 7) on the Ox-Drove ridge running to a gully on the northern escarpment edge.
The whole of the northern dyke is visible from long distances away to the north of the chalk escarpment and when new would have created a dramatic sinuous line across it. Only the very top of the southern dyke where it crosses over the ridge is visible from off the scarp to the north, the rest being concealed by the spur. It is clearly visible from Castle Rings hillfort off the ridge to the northwest.
Hatt’s Barn (Figure 4.3: No. 8) The dyke at Hatt’s Barn, consisting of a ditch with one or more banks on its western side, runs in a staggered course from the edge of a deep meandering coombe to the north (Figure 4.24) to the side of the end of a very shallow coombe at its southeast end. This shallow coombe in turn runs down into the depths of Boyne Bottom, which runs roughly parallel with part of the southern course of the dyke. This change in orientation of this dyke from NNW-SSE to NNE to SSW may be explicable as an alteration to an original intention for the dyke to run down a steepening slope to end at the side of Boyne Bottom, and it is of interest to note that the coombe that the dyke links to resembles a wide shallow dyke in many respects, so that the dyke in effect becomes a continuation of the natural ‘dyke’ running down to the coombe.
FIGURE 4.24 View along the coombe cutting into the chalk ridge from the west, at which the northern end of the Hatt’s Barn cross dyke (No. 8) terminates along the middle of its course to the right of the photograph.
Fontmell Down (Figure 4.3: Nos. 9–11) Another 2 km to the west is found the first of two dykes cutting across the ridge of Fontmell Down, a dramatic terminal spur on the line of the chalk escarpment to the west. These two dykes are only 250 m apart but of markedly different character. The easternmost dyke forms a virtually straight line running obliquely across the spur from WNW to ESE. It consists of a single bank with the ditch on the northern side. The southern end runs to the middle of the upper slopes of Longcombe Bottom, a dramatic narrow coombe to the south, and the northern end terminates on the far gentler slopes of the northern escarpment. Here there are extensive views across the Blackmore Vale. At neither the southern or the northern end is it possible to see down to the bottom of the scarp edge.
The eastern dyke is bivallate in form. Its northern end runs to the middle of the scarp edge to the point at which one can look down to the base of the scarp. Here the slopes are quite gentle (Figure 4.25). By contrast, the southern end of the dyke terminates some way down the side of the precipitous slopes leading down to the head of Longcombe Bottom considerably below the point at which one can see down to the base of the slope (Figure 4.26). From the northern end, the view is out across the lowland plain to Melbury
FIGURE 4.25 Northern end of the Fontmell Down east cross dyke (No. 10) on the Ox-Drove ridge.
FIGURE 4.26 Southern end of the Fontmell Down east cross dyke on the Ox-Drove ridge.
Beacon beyond with its dramatic cross dyke clearly visible. From the southern end, the view seems to be directed toward the depths of the coombe bottom. When one walks along well-preserved sections of the ditch, one’s view of the wider landscape is blotted out except toward the terminal ends. This dyke changes direction in the middle as it crosses over the top of the ridge from running NW-SE to ESE-WSW. Walking north along the ditch, one notes that this change in direction coincides with the point at which Melbury Beacon can first be seen in the distance.
To the south of these dykes, another dyke known as the Tennerley Ditch runs NW-SE from the side of Longcombe Bottom, across the brow of the chalk escarpment, to end somewhat indeterminately on a gentle slope to the south. Its northern end contrasts by ending some considerable way down the lip of the coombe looking down to the depths of the coombe. Were it not for woods, it would be intervisible with the southern end of both the dykes crossing Fontmell Down on the other side of the coombe.
Melbury Beacon (Figure 4.3: No. 12) The final dyke is 1.5 km to the north of those crossing Fontmell Down. Running NNE to SSW, it cuts dramatically across a ridge of land rising to the west joining Melbury Beacon to Compton Down. Univallate in form, it constitutes a ditch with a bank on the uphill side. In places where the ditch is well preserved, an observer’s view of the landscape on both sides of it is again blotted out, with the observer’s vision being directed along the course of the dyke. The eastern end stops abruptly some way down the lip of an extremely steep slope dropping down to the bottom of Dukum hollow, a deeply incised declivity in the escarpment edge (Figure 4.27). At this end of the dyke, one’s eye is directed down to the base of the escarpment and across the plain below. The dyke ends at precisely the point down the slope at which the base of the escarpment can be seen. The western end runs much farther down the slope of the hill. It is much less steep on this side, and it stops just to the northwest of a small natural ‘dyke’ or gully in the escarpment edge, at the point of which its base is visible (Figure 4.28). This linkage of the dyke to a ‘natural’ dyke in the chalk scarp is reminiscent of the southern terminus of the Hatt’s Barn cross dyke. The dykes both cut across the chalk downlands and are closely related to pre-existing features.
Discussion The cross dykes along the Ox-Drove ridge are clearly a diverse set of structures varying considerably in original length from examples 300 m or less in length to others, such as Tennerby Ditch, running for over 700 m. Four are bivallate with a medial ditch flanked by banks, seven appear to be univallate, and to the northwest of Win Green one badly mutilated example has two parallel banks and ditches. The earthworks are located at regular 2-km intervals along the ridge but do not divide up the ridgetop in any readily comprehensible pattern. If their purpose was simply to divide up the ridgetop into blocks of land of roughly equal size, we might expect those on Berwick Down to run north-south rather than west to east and cut across the ridgeway rather than cutting across a spur to the south between two coombes. The bivallate
FIGURE 4.27 Northern end of the Melbury Beacon cross dyke (No. 12) on the Ox-Drove ridge dropping over the ridge top and down toward the declivity in the escarpment edge of Dukum Bottom.
FIGURE 4.28 Southern end of the Melbury Beacon cross dyke (No. 12) on the Ox-Drove ridge terminating just above and beside a distinctive gully in the escarpment edge.
examples do not appear to be more important or strategically sited than those with a single bank and ditch, and this distinction may be unimportant in terms of their purpose.
In terms of their topographic locations, the following three cross-dyke groups can be distinguished:
In virtually all cases it seems to have been of great importance that an observer be able to see from the end of the dykes down to the very base of the escarpment edge or the bottom of the coombes below. The escarpment or coombe bottom is visible immediately below the dyke end from at least one end of eleven out of twelve of the dykes. In four cases (4, 7, 9, 12), these features are visible from both ends. For this to be the case requires sometimes that the dyke descend over the lip of the escarpment and some way down the escarpment edge. Some, resembling a slide, descend the steep lip of the escarpment or coombe for anything up to 50 m or more in an exaggerated fashion, and far beyond the point at which the base first becomes visible. Others, such as the northern end of Win Green South (No. 7) or the eastern end of the dyke cutting across Melbury Beacon (No. 12), terminate more or less exactly at the point at which the base of the coombe or escarpment first becomes visible when one is walking along them. All this suggests the importance of one’s vision becoming directed downward at the end of the dykes rather than simply to look out and across the wider landscape beyond. These dykes make use of and serve to emphasise the precipitous slopes to which they are intimately related. Only one dyke (No. 7) plunges down precipitous slopes at both ends. This dyke also has another unique feature. Both ends terminate adjacent to natural ‘dykes’ or linear depressions cutting into the chalk escarpment. This characteristic occurs at one end of two other dykes (Nos. 8 and 12). In these cases, a metaphorical or analogical relationship between coombes and dykes appears to be made explicit: that is, that they resemble one another in many respects. This resemblance enables one to suggest that some of the cross dykes were considered to be continuations of the coombes and vice versa and that the linkage created between dyke and coombe was of deep symbolic significance. In this respect, it is interesting to note the sinuous and meandering nature of most of the dykes and their often sudden change in direction and orientation. Straight dykes such as the western of the two cutting across the spur of Fontmell Down appear to be the exception rather than the rule.
The manner in which the dykes relate to the microtopography of the escarpment edge and that of the coombes is also of interest. In relation to the coombes, four of the cross dykes (4, 7, 8, and 9) terminate at the head of a coombe. In three cases, they end at the side of the coombe, and in only one case (No. 7) does the dyke continue along the line or directional orientation of the coombe. Other dykes (10, 5, 11) terminate along the side of the coombe edge some distance from its end. When the dykes terminate on escarpment edges that are indented or scalloped, they are much more likely to run to the head of these edges (Nos. 1, 2, 3, 5, 7, 8, 9, 12)—but rarely in a symmetrical fashion, that is, in their middle but generally on one side or the other.
ROADS AND TRACKS All the paths and tracks on the sides of the ridges have one thing in common, irrespective of their antiquity: they run obliquely up and down the scarps. In some cases, they cut diagonally across the lower ends of the cross dykes. There are no cases where trackways lead up to the bases of the bivallate cross dykes, suggesting their possible use as ‘cattle ways’. These dykes plunge directly down the sides of the scarps and coombes, making their use as regularly used roadways highly unlikely, if not downright impossible. Trackways do pass the ends, or cut through the spur dykes at Burcombe, on Buxbury Hill, and on White Sheet Hill along the Ebble-Nadder ridge, and on Win Green and Melbury Hill along the Ox-Drove ridge, but the association cannot be described, since these trackways leading to the dykes and the fact that some pass around the end of the dykes may be purely fortuitous—that is, these trackways were later in date and avoided having to pass over the dykes. The argument that the spur dykes were constructed purely to divert traffic along trackways away from the spurs seems again unlikely.
HILLFORTS AND SETTLEMENTS (FIGURES 4.1, 4.2, 4.3) Along both the Ebble-Nadder ridge and the Ox-Drove ridges there are two enclosed settlements or univallate ‘hillforts’—respectively, Chiselbury and Winkelbury. Chiselbury was built directly to the north of one of the deepest and most dramatic coombes to cut into the Ebble-Nadder ridge, Winkelbury on the only true northern spur of the Ebble-Nadder ridge. There are no Bronze Age barrows in the immediate vicinity of Chiselbury, which occupies one of the few places along the ridgetop without them. By contrast, Winkelbury Hill has one of the largest clusters of barrows along the Ox-Drove ridge. Here the hillfort was constructed below them on quite steeply sloping ground at the northern end of the spur rather than around its flat top with its ramparts slung low around the western, eastern, and northern spur slopes. When one walks from the south to the north through the middle of the enclosure, neither the western or eastern banks are visible, providing uninterrupted views up and down the Ebble valley. For much of the way, the northern rampart is similarly invisible. So, one has the impression of being in an unenclosed space open to the world beyond. By contrast, when one stands in the middle of the Chiselbury enclosure, which slopes much more gently from south to north, the view west along the ridge and south off the ridge is completely blocked by the bank, and it is partially blocked to the east. It appears completely open only when one looks north.
Standing on the northern bank of Chiselbury, one can look out across the plain of Wardour but not down to the bottom of the scarp slope to the north, because the bank is set some way back from it. Again, this situation contrasts with Winkelbury, where the bottom of the scarp slope is visible along much of the western side. Chiselbury is a circular enclosure that is designed to cut out much of the world beyond; Winkelbury opens itself to the depths surrounding it. The single entrance to Chiselbury facing southeast is not orientated toward any significant landmark. It just looks out across the flat ridgetop, which runs southeast at this point. By contrast, the staggered southwest entrance of Winkelbury is orientated toward Win Green and its summit barrow is the highest point along the Ox-Drove ridge.
The earthwork associated with Chiselbury is a cross-ridge dyke linking the dramatic coombe to its south, with the northern scarp of the ridge the enclosure itself being set between the two. At Winkelbury, there is also a possible outwork or unfinished cross-ridge dyke situated 500 m to the south and out of sight from the hillfort. This is a bank up to 6 m wide and 150 m long running from east-west across the eastern side of the spur top. It has a staggered entrance way through it that, as Pitt-Rivers noted, resembles the southern entrance of the hillfort (Pitt Rivers 1888: 236). Just as Winkelbury remained unfinished, this earthwork may be an unfinished spur dyke originally intended to run right across the spur from scarp slope to scarp slope. It is in the area of land between this bank and the hillfort that the Bronze Age barrows are situated.
Two further hillforts, Castle Ditches, Tisbury, and Castle Rings, near to Shaftesbury, are located on hills a short distance away to the north (see Figure 4.1). Castle Ditches is by far the most massive and elaborate, with multiple ramparts defining its southeast entrance. These hillforts may be linked downland and lowland settlements. From Castle Ditches, the entire Ebble-Nadder ridge is visible from its western end at Whitesheet Hill as far east as Chiselbury, which marks the limit of the visual field to the east. None of the cross dykes are visible from it. From Castle Rings, the visual field is from Melbury Beacon in the west, marking the end of the Ox-Drove ridge, to Winkelbury and beyond. Part of the Ebble-Nadder ridge is also visible as far as Ansty Down. It is interesting that the visual field of both these hillforts off the ridges more or less ends with the two hillforts, to the east, situated on the ridges. Neither of the hillforts on the ridges is intervisible nor are the hillforts situated to the north off the ridges. From Castle Rings, the cross dykes on Win Green and White Sheet Hill are visible. The earliest ramparts of Winkelbury Hill, situated on the summit, are particularly prominent from Castle Rings and resemble a particularly massive cross dyke cutting across the hilltop. These pairs of hillforts may have been interconnected by patterns of upland and lowland movement incorporating both upland pasture and lowland grazing.
Both Castle Rings and Winkelbury have outworks, to the west and south respectively, that run across the ridge. Those at Winkelbury run across the spur on which the hillfort is situated from west to east about 500 m to the south. Those at Winkelbury may have been intended to be a true cross dyke, but like the hillfort itself it may be unfinished (Feacham 1971). To the east, a bank extends to the shoulder of the scarp but does not run down it. To the west, there is no trace of the bank beyond the top of the hill. This earthwork was described by Pitt-Rivers as an outer defence of the hillfort. He noted that it has a staggered entrance mirroring that of the main southern hillfort banks and ditches to the north across the spur.
The Chiselbury enclosure has been dug into twice but produced no evidence of occupation (Colt Hoare 1812; Crawford and Keiller 1928: 76). One or two cross dykes are directly linked to it. Excavations at Winkelbury hill-fort (Pitt Rivers 1888) do not provide much indication of anything other than temporary occupation, and the hillfort itself is unfinished. The hillfort occupies the only really prominent northern spur along the ridge jutting out from the escarpment line.
The Swallowcliffe Down settlement (Clay 1925, 1927) is situated on the ridge top at the head of a particularly dramatic and deep coombe. There is a dramatic view from this place down the coombe and across the Ebble valley to Winkelbury and along the ridgetop east to Chiselbury and west to White Sheet Hill. A ditch surrounds the southern part of the settlement area and curves round to the north terminating by and just avoiding a Bronze Age barrow to its east, which therefore still seems to have been a significant landscape marker to be both used and avoided. Extending over about 1.5 ha, it is bounded on the north, northwest, and northeast by the steep escarpment of the downs and on the south, southwest, and southeast by a semicircular ditch. There is about 150 m of level downland between the northern extent of the c. 100 storage pits and areas with hearths, excavated by Clay, and the northern scarp. Post holes suggest that the enclosure may have been palisaded. To the south and separated from it by the Ridgeway track is a semicircular ‘amphitheatre’ of uncertain use, open to the north, where an original entrance to a possibly circular enclosure may have been. The settlement area is located less than 200 m to the north of a dramatic coombe (see Figure 4.18) and seems to be directly associated with the coombe and a cluster of five cross dykes (30% of those on the ridge) linked to it and parallel coombes to the west and east. Clay reports that at the head of the coombe are signs of a dam and a catchment pond, probably the water supply for the village (Clay 1925: 59). Clay considers that the cross dyke flanking the settlement area to the southwest (Figure 2: No. 9) was constructed when the settlement was extant (Clay 1927: 51), and indeed this explains its unusual orientation in relation to the ridge. This dyke is flanked by ‘Celtic’ fields to its southwest that were earlier or contemporary with it. Field lynchets also adjoin the village area to the north and northwest.
Only three other settlements are known along the Ebble-Nadder ridge on Prescombe Down and Fyfield Bavant Down (Clay 1924; Fowler 1964: 56). These all occur on the tops of spurs running south from the ridge and at a distance of 1 km from the ridgetop and are approximately contemporary with the Swallowcliffe Down settlement (Figure 4.1). None are associated with cross dykes. The spurs are separated by comparatively small and shallow coombes that lack cross dykes linking them to the northern escarpment. No other enclosed or unenclosed settlements are known, despite extensive aerial photography of an area most of which has been ploughed.
Along the Ox-Drove ridgeway, dense settlement areas are known from transcription of aerial photographs, at the far eastern end of the ridge. Here the cross dykes and linear ditches clearly form one element within a complex of settlements, enclosures, and associated field systems and drove ways dated to the late Iron Age and continuing in use during the Roman period (Corney 1990; Rahtz 1990). The cross dyke (Figure 4.3: No. 1) at the eastern end of a series of linear ditches and banks runs from the northern escarpment up and over the slope, terminating to the south in an area with extensive fields but lacking any clearly defined enclosures. The two other dykes (Figure 4.3: Nos. 2 and 3) along this eastern part of the ridgeway may have terminated in a similar manner in areas with fields and enclosures on land gently sloping to the south. Their course and extent across the ridge are uncertain, and none seem to be linked to coombes to the south, which are relatively shallow and insignificant along this part of the ridge.
Elsewhere, late Iron Age settlements are known from Berwick Down (Wainwright 1968) and Rotherley Down (Pitt Rivers 1888). On Berwick Down there are at least two Iron Age settlements. The smaller, to the south, excavated by Waninwright, is of late Iron Age date, founded perhaps a generation before the Roman invasion of c.e. 43. It consisted of a kite-shaped enclosure containing one round house, four granaries, and thirty-four pits and working hollows. This was a small farmstead, perhaps occupied by a single family. The settlement area to the north is much more extensive and comparable in size to that of Rotherley. Here Pitt Rivers’s excavations revealed a complex of enclosures, storage pits, granaries, working hollows, foundations of a corn-drying furnace, one or two house foundations, and a small rectangular building.
Both settlements are situated on spurs projecting south from the ridge bounded by deeply incised coombes to the west and the east in an area with the deepest and most complex coombe systems in the study area. Both settlements, which are intervisible, are situated almost opposite each other about 750 m apart on the southern ends of spurs gently sloping to the south, 1 km from the ridgetop. The occurrence of two contemporary settlements on southern spurs, separated by an intervening coombe, mirrors the location of the Fyfield Bavant settlements on the Ebble-Nadder ridge. The Berwick Down and Rotherley settlements are located only 2 km to the south of Winkelbury hill-fort, which is due north of the Rotherley Down settlement, and both occupy the same area of higher ground. Two univallate cross dykes cut across Berwick Down to the north of the settlement area and seem to bound it off. By contrast, none are known on Rotherley Down.
FIELD SYSTEMS The relationship of the dykes to the pattern of ancient fields in the study area is difficult to assess, since there has been such widespread arable destruction of the chalk ridges. For the Ebble-Nadder ridge, Fowler’s map (1964) shows that with only a few exceptions the distribution of cross dykes and field systems are separate. The only known association between field systems and a bivallate cross dyke is on Swallowcliffe Down. Celtic fields are largely confined to scarps south of the ridge but almost certainly continued to run across the spurs between coombes. Aerial photographs show that the bivallate and univallate dykes to the south of the Swallowcliffe Down settlement (Figure 4.2: No. 11) may originally have curved in their course to form some form of irregular enclosure (Fowler 1964: 52). Celtic fields on the west and north scarps of the spur, across which the Buxbury univallate cross dyke runs (Figure 4.2: No. 7), suggest a close association, and the dyke may mark the southern limits of these fields that did not extend across the main ridgetop. Along the Ox-Drove ridge, aerial photography has shown the presence of complex field systems across and to the south of the ridgeway (Corney 1990: Figure 4.3). Elsewhere they occur on scarps and spurs, as along the Ebble-Nadder ridge. Interestingly, the only evidence of settlement we have along the Ox-Drove ridge is on southern spurs running away from the ridgetop, the enclosed site of Winkelbury Hill being the exception. A similar pattern occurs along the Ebble-Nadder ridge, the two exceptions being the ridgetop settlements of Swallowcliffe and Chiselbury hillfort. The former is surrounded by fields, the latter is not.
The picture that seems to emerge is that the ridgetops across which the cross dykes run may have been largely open and used predominantly for pasture, whereas lower spurs and ridges were settled with a dense pattern of arable fields. The cross-ridge dykes were therefore built across areas of open downland used for pasture and for the most part separate from areas with settlement and arable land. The blocks of downland defined by these dykes along the ridgetop were highly variable in size: a few to no more than 500 metres wide across the ridgetop and between 1.1 and 3.5 km long along the Ebble-Nadder ridge and 0.75 and 4 km along the Ox-Drove ridge. On the Ebble-Nadder ridge, the longest block includes Chiselbury hillfort,which is at the far eastern end. Again along the Ox-Drove ridge, the longest block includes Winkelbury hillfort, which is situated toward the eastern end.
The clustering of many of the round barrows in relation to the coombes indicates the significance of these places. Other round barrows also seem to be related to significant points along the northern escarpment edge— places where it is indented by gullies or changes direction. The construction of the cross-ridge and spur dykes appears to involve both a continued recognition of the significance of these places and that of some of the earlier Bronze Age barrows. Both the cross-ridge and spur dykes link coombes and escarpment edges. Significantly, it is the central coombe of the Hydon Hill/ Little Down system on the Ebble-Nadder ridge, around the southern end of which the barrows cluster at the point at which these coombes join, that is later linked to the escarpment edge to the north by the Compton Hut dyke D, which effectively continues the line of the coombe up and over the ridge and down the other side. Both of the only two coombes to cut into the northern escarpment edge are linked by dykes to the northern scarp. The course of the longer dyke, Burcombe C, is obviously related to the existence of pre-existing barrows in the landscape. Burcombe A and B cut across and mark out the same spur as the Punch Bowl barrow (No. 23). They run up to the coombe head, whose importance is already marked by the barrow. The southern dyke, situated only about 30 m north of the barrow, continues down the precipitous scarp slopes leading down to the coombe bottom and the bottom of the escarpment edge, ending about where the bottom becomes visible below. The Compton Hut dyke D continues far down the precipitous slope of the northern scarp well beyond the point at which the base of the scarp slope becomes visible. All the dykes link with the heads of coombes, either continuing their lines across the landscape or establishing a change of direction.
The characteristics of the chalk ridges and spurs and the coombes are strikingly different. These aspects of the chalk landscape, together with the bold and indented escarpment edges, give it its special qualities and character. Some of the contrasts are summarised here:
| Ridgetops | Coombes |
| Wide views | Restricted views |
| Exterior | Interior |
| Windswept | Sheltered |
| Light | Shade |
| Dry | Wet |
| Treeless | Wooded |
| Looking down | Looking up |
| Visible | Secret/Hidden |
| Sound dulled | Sound amplified (echoes) |
The interiorised worlds of the coombes, each with its own individual qualities and character, are utterly different from the ridgetops and spurs that separate the coombes. The coombes wend and wind their way, join and bifurcate, open out and close in on themselves as they pass through the chalk. They have their own microtopographies, climate, and vegetation. They are hidden places, visible at all only from short distances away. All are invisible from central areas of the ridge and spur tops. These are, by contrast, relatively undifferentiated and uniform in character. It is only the escarpment edge that differentiates different parts of the ridgetops. The coombes amplify sound and have different qualities of light and shade. They invite one to follow and explore their courses, both dividing the landscape and establishing different natural paths of movement up and through it. The world of the ridgetop is, by contrast, a big wide and open landscape, a macro world of the extensive vista as opposed to the small enfolded world of the coombe. Atmospheric effects and temperature inversions may fill the coombes with mist and cloud or alternatively blanket the ridgetops above. When the mists fill the coombes, they are magically transformed into lakes. In exceptionally wet periods, water may begin to appear in the base of the coombes only to sink away again after a few hours or days.
Some of the barrows obviously mark out places of especial significance along the courses of the coombes—places where they join, or open out, and places where they end, perhaps conceived as doorways to a world below. These low, wet, mysterious, and hidden incisions in the landscape, with their inner depths, were probably associated with particular spirits, mythical forces, and the underworld. Such places could be conceived as dangerous; hence from some barrow sites one looks across rather than down into the depths of the coombe. The association of other barrows with transitional places in the landscape on the way down to the coombe bottoms may be indicative metaphorically of the passage from life to death, the sky and the heavens to a watery underworld; hence their siting on the shoulder of slopes, below the shoulder, halfway down the slope, and so on There is an important changing visual perspective in all this. From only a few barrows, the coombe bottom or the escarpment edge immediately below them is visible: from them the depths can be seen. Because the barrows are sited progressively higher up in the landscape, this visual perspective of looking directly down into a different world below becomes successively diminished. Instead, one has only partial views along or across coombe bottoms and escarpment edges. From the ridgetop summit barrows, such a view of the landscape is entirely removed. Here one is in contact only with the sky. The relationship of other barrows to gullies and places where scarp slopes change direction indicate the symbolic significance of these aspects of the topography, too, as places of transition, perhaps again of a metaphoric journey from life to death, high to low. The barrows on the flat ridgetop summits, at the very highest points in the landscape, must obviously be associated with the sky, and they emphasise that height as well as watery depths was of great ritual significance.
It can thus be suggested that the entire barrow distribution, when considered as a whole, networks or links every distinctive topographic element in the landscape into a coherent whole with possible cosmological significance in terms of a life journey. It also obviously marks out the entire landscape and lays claim to it. The patterns of movement of people from coombe to ridgetop to escarpment edge would always be marked out by these barrows. Now, monuments to the dead can, of course, also be used to highlight significant differences between the status and the power of those in the world of the living. Each coombe is unique in various ways. Some are strong and dramatic incisions in the landscape; others are weak. Those who could symbolically control the ‘strong’ coombes and their spirit powers could enhance their authority in the world of the living. It is then, perhaps, not so surprising that the largest barrow at the eastern end of the Ebble-Nadder ridge (No. 23) is directly associated with the most prominent coombe, Punch Bowl Bottom, and is located at its head. Patterns of intervisibility between barrows might be related to social connections between particular lineage groups and coombes. A barrow could be linked to a particular coombe in such a way without having to be sited near it.
The detailed study of dykes has shown that in virtually all cases it seems to have been of great importance that an observer be able to see from the ends of the dykes down to the very base of the escarpment edge or the bottom of the coombes immediately below. For this to be the case requires sometimes that the dyke descend over the shoulder of the escarpment and some way down the escarpment edge. Some, resembling a slide, descend down the steep slope of the escarpment or coombe for anything up to 50 m or more in an exaggerated fashion, and far beyond the point at which the base first becomes visible below the dyke end. Others terminate more or less exactly at the point at which the base of the coombe or escarpment first becomes visible when one walks along them. As mentioned earlier, all this suggests the importance of one’s vision being directed downward at the end of the dykes rather than simply to look out and across the wider landscape beyond.
A direct metaphorical relationship between coombes and dykes appears to be made explicit in the case of bivallate forms with a medial ditch in that the forms of the dykes and the coombes closely resemble one another in many respects. This enables one to suggest that some of the cross-ridge dykes were considered to be continuations of the coombes and vice versa, and that the linkage created between dyke and coombe was of deep symbolic significance. The dykes as artifical coombes continued the lines of the coombes up into the sky and over the ridgetop and down the northern scarps to the lowland bottoms beyond. In this respect, it is interesting to note the sinuous and meandering nature of many of the dykes and their often sudden change in direction and orientation—and in this respect again they resemble the coombes. Straight dykes appear to be the exception rather than the rule. If the dykes were coombes in the sky, what implications would this have for our understanding of them?
It may have been that their function of linking both the coombes and the lowlands was of equal or of greater importance to the manner in which they divided the ridges. Those dykes that cut across the ridgetops ipso facto divided them, but this division cannot necessarily be assumed to be their primary purpose. The alternative argument is that they served to link and network the topographies of coombes and lowlands. The dykes then were expressions of sociocultural norms investing the landscape with meaning, a coding of space in relation to socially significant ridges, spurs, and coombes. Their construction, use, and meaning may have been part of a ritual practice seeking to maintain harmonious relationships with a complex pantheon of invisible beings and forces associated with wet places such as the coombes, river valleys, and the lowlands, significant places for the deposition of votive deposits from the Bronze Age onward (Bradley 1990, 2000). Thus the dykes were ritual works, perhaps processional routes or travelling ways, either in reality or in the imagination, between the coombes to the south and the lowlands to the north, from one coombe to another and from scarp to scarp. They were part of the manner in which cultural meanings in the landscape became materialised. As such, they may have formed part of a ritualised order of space, time, and movement linked to the seasons, the significance of different cardinal directions, height and elevation, the juxtaposition of ridges and valleys, rivers, spurs, and significant hills. Particular hills and spurs were clearly marked out and emphasised by the spur dykes, whereas the cross-ridge dykes linked significant coombes, each with its own particular identities and associations, with the lowland. Yet other dykes joined one coombe to another. In sum, they connected important elements of the topography into a reticulated system, improving on what nature had already done. They thus completed the link that ‘nature,’ or the ancestral forces, had not made between the coombes and the lowlands beyond.
I have argued that both the long and the round barrows were located to create connections and establish relationships both between themselves and other barrows and to refer to, or connect, significant ‘natural’ places in the landscape far beyond their specific location. In other words, the significance of the location of a barrow in one place was linked to that of another in a quite different place. The location of one barrow was understood in terms of that of another. They thus both marked specific places as meaningful and simultaneously acted as material metaphors for the wider landscape as a whole. And so they also served to codify important topographic features of the landscape both in relation to themselves and through their links to other barrows in different places in it. Through the process of constructing round barrows in different places, people networked that landscape into a coherent whole. Through these connections, metaphorically an individual barrow became the wider landscape, and in turn the landscape was the barrow location. So people made themselves and their social relations and constructed their identities in relation to both the specificity of place and the totality of the wider landscape conceived as a network of relationships among different places within it.
It was the relationships among the barrows and in turn their relationships with their landscape settings that empowered people to identify with the landscape as a whole rather than just specific places (individual barrow locations) within it. The construction of dykes represented an alternative way of thinking through, understanding, and relating to landscape. Interconnections among different places in the landscape previously marked out as significant through the scattered individual locations of groups of round barrows in it now became physically joined in the form of one large and continuous monument sweeping across it. Linkages that had previously been only conceptually implicit in the overall patterning of the individually very different locations of round barrows were now made explicit and objectified in a material form through the process of dyke construction. What had previously been a nonmaterial resource in which the social and cosmological significance of the contrasts in the different landscape locations of barrows had to be connected through experience and talk was now made materially explicit through the network of dykes inscribed across it.
Although we have no direct evidence for Bronze Age settlement anywhere along either the Ebble-Nadder ridge or the Ox-Drove ridge, the early and later Iron Age enclosures and associated field systems discussed above may indicate a much more intensive and permanent pattern of occupation and use than is suggested by the earlier Bronze Age barrow distribution.
The construction of the dykes may therefore have related to an increasing social and political need to physically control and lay claim to the land itself and the material and symbolic resources that it provided. The dykes’ morphology and direct relationships to the coombes would effectively serve to naturalise them in the landscape. They might be perceived to be more a part of an order of nature, rather than an order of culture, and therefore had added social and political power when there was a desire to control the land. The argument here is that to control the land involved physically networking different elements of the topography of the ridge as opposed to dividing it up and erecting boundaries across it.
The cross-ridge and cross-spur dykes are clearly not closely linked to the distribution of Bronze Age round barrows, which seems to rule out any possibility that they might have been related to any possible marking out of ‘territories’ or land divisions supposedly marked out by these barrows, which has sometimes been claimed. Nor do they seem to be closely related to trackways up or down the scarps and into the coombes. They make improbable boundaries for field systems, and they do not appear to cut across fields, putting them out of use. The direct link with hillforts in the case of Chiselbury is much stronger, and those on Swallowcliffe Down and Berwick Down are clearly associated with settlement areas that were either earlier or that were associated with their construction and use. Now these two settlements and the Chiselbury hillfort, the Fyfield Down, Rotherley, and Prescombe Down settlements are all closely associated with coombes. They are situated either on spurs between coombes or at the head of particularly dramatic coombes. Coombes seem therefore to be a common factor relating to both the locations of the settlements and many of the dykes.
Although the coombes separated the settlements on the spurs, some dykes joined the coombes. Here I wish to re-emphasise that it was their function of linking both the coombes and the lowlands that was of equal or of greater importance to the manner in which they divided the ridges. Those dykes that cut across the ridges ipso facto divided them, but this cannot be assumed to be their primary purpose, which may have been instead to link and network the lowland topographies of coombes and plains and the Ebble valley between the Ebble-Nadder and Ox-Drove ridges. The major bivallate cross dykes in profile most resemble coombes, and, in effect, they are artificial coombes that continue the lines of the coombes up into the sky and over the ridgetop and down the northern scarps to the coombe bottoms. The bivallate forms clearly bore a close analogical or metaphorical relation with the coombes, cultural constructions resembling natural forms. Their form and relationship to the coombes would effectively have served to naturalise them in the landscape. They would have been perceived to be part of an order of nature, rather than an order of culture, and therefore they had added power.
The dykes were expressions of sociocultural norms investing the landscape with meaning, a coding of space in relation to socially significant hills, ridges, spurs, and coombes. Their construction, use, and meaning may have been part of a ritual practice seeking to maintain harmonious relationships with a complex pantheon of invisible beings and forces associated with wet places such as the coombes, river valleys, and the lowlands, significant places for the deposition of votive deposits from the Bronze Age onward (Bradley 1990, 2000). Thus the dykes were ritual works not ‘practical’ tracks or cattleways. Instead, they were perhaps processional routes, travelling ways, between the coombes to the south and the lowlands to the north, from one coombe to another and from scarp to scarp. They were part of the manner in which cultural meanings in the landscape became materialised. As such, they may have formed part of a ritualised order of space, time, and movement linked to the seasons, the significance of different cardinal directions, height and elevation, the juxtaposition of ridges and valleys, rivers, spurs, and significant hills. Particular hills and spurs were also clearly marked out and emphasised by the spur dykes, whereas the cross-ridge dykes linked significant coombes, each with its own particular identities and associations, with the lowland, while yet other dykes joined one coombe to another. It was clearly the heads or terminal points of the coombe systems that were of particular significance, more so than the character of the escarpment edges to which the dykes run, in many cases, and the very head of the coombe, if it ran into the ridges from the southern side, and its west or left, rather than its east or right side, was emphasised by the positioning of the dyke (see Table 4.8 and Figure 4. 29).
Table 4.8 The relationship of cross dykes to coombes (see Fig. 4.23).
| Map No. | Relationship of Dyke to Coombe | Characteristics of Coombe | |||
| Ebble-Nadder Ridge (Fig. 4.2) | |||||
| 1 | Left (west) side of head of coombe | Deep, regular sides, wide, flat bottom | |||
| 2 | Left (west) side of head of coombe | Deep, regular sides, wide, flat bottom | |||
| 3 | Head of coombe? | Shallow, meandering, narrow, V-shaped | |||
| 4 | Head of coombe? | Shallow, regular, narrow, V-shaped | |||
| 5 | Left (west) side of head of coombe | Shallow, meandering, narrow, V-shaped | |||
| 6 | Head of coombe | Deep, regular sides, wide, flat bottom | |||
| 7 | Not applicable | ||||
| 8 | Head of coombe | Deep, regular sides, wide, V-shaped | |||
| 9 | Left (west) side of head of coombe | Very deep, regular sides, wide, flat bottom | |||
| 10 | Head of coombe? | Deep, regular sides, wide, flat bottom | |||
| 11 | Head of coombe | Deep, narrow, branching at end, flat bottom | |||
| 12 | Left (west) side of head of coombe to side of coombe | Deep, narrow, branching at end, flat bottom to very deep, regular sides, wide, flat bottom | |||
| 13 | Side of coombe to right (east) side of head of coombe | Very deep, regular sides, wide, flat bottom to deep, regular sides, wide, flat bottom | |||
| 14 | Head of coombe | Wide and open | |||
| 15 | Head of coombe | Narrow steep-sided, V-shaped hollow | |||
| 16 | Left (west) side of coombe | Wide and open | |||
| Ox- Drove Ridge (Fig. 4.3) | |||||
| 1 | Not applicable | ||||
| 2 | Not applicable | ||||
| 3 | Not applicable | ||||
| 4 | Left (west) side of coombe to right (east) side of coombe | Narrow, meandering, very deep, V-shaped | |||
| 5 | Head of coombe to right (east) side of coombe | Narrow, meandering, very deep, V-shaped to narrow, meandering, very deep, V-shaped | |||
| 6 | Not applicable | ||||
| 7 | Head of coombe | Narrow, meandering, very deep, V-shaped | |||
| 8 | Right (south) side of coombe to right (north) side of head of coombe | Narrow, meandering with spurs, very deep, V-shaped to very shallow, flat-bottomed side branch of deeper meandering coombe | |||
| 9 | Left (west) side of coombe | Regular sides, wide, very deep, flat bottom | |||
| 10 | Left (west) side of head of coombe | Regular sides, wide, very deep, flat bottom | |||
| 11 | Right (east) side of coombe | Regular sides, wide, very deep, flat bottom | |||
| 12 | Not applicable | ||||
The death of the sun in the western sky in itself might suggest, of course, a basic cosmological association between water, the west or left, and rituals concerned with death and the underworld with which the coombes and the dykes may have been particularly associated. There is an obvious precedent for the bivallate dykes in the form of the Neolithic cursus monuments, but
FIGURE 4.29 Sketch plans of the relationship of the terminal ends of dykes along the Ebble-Nadder and Ox-Drove ridges in relationship to coombes.
although the cursus monuments may cut across both ridges and valleys, as is the case for the Dorset cursus (Tilley 1994), the cross-ridge dykes appear to be much more subtly keyed and linked to the landscape. In sum, they connected important elements of the topography into a reticulated system, improving on what nature had already done. They completed the unfinished work of the Gods.
