Chapter Three
Stonehenge's Architecture and Landscape¹

Almost all twentieth-century considerations of Stonehenge have, perhaps understandably, ignored the fact that Stonehenge exists in, is related to, and is embedded in a landscape. The focus of attention has always been the stones themselves and the chronology and structural development of the monument. Thus Gowland (1902), Hawley (1921–1928), and Atkinson (1956) make no reference to the landscape setting of Stonehenge at all, and only Atkinson mentions and provides a map of monuments in its vicinity (ibid.: 146). The Royal Commission of Historical Monuments usefully puts Stonehenge into a wider spatial context in terms of an inventory of other sites in the Stonehenge ‘environs’ (RCHME 1979), while The Stonehenge Environs Project (Richards 1990) reports on the results of fieldwalking and excavations within a 33-km-square box centred on Stonehenge. However, in both of these studies the landscape contexts and interrelationships of monuments are not considered either from the perspective of Stonehenge or from anywhere else. The landscape, in both cases, is simply a more or less blank spatial field for analysis. Previous generations of archaeologists have diligently worked in the Stonehenge landscape while simultaneously ignoring it!

The first publication to actually start to seriously consider the landscape around Stonehenge was published little more than a decade ago (Cleal, Walker, and Montague 1995). In an excellent discussion, Mike Allen considers, in some detail, the geographical and topographical setting of the monument in relation to Bronze Age barrow cemeteries, and Julie Gardiner the view to it from the Avenue (Allen in Cleal, Walker, and Montague 1995: 34–40 and Gardiner in Cleal, Walker, and Montague 1995: 40), research that is discussed in some detail below. Elsewhere in the book, other monuments in more or less the same spatial box used by the RCHME (1979) and Richards (1990) are briefly considered in relation to various proposed phases of Stonehenge. The title of this book, Stonehenge in Its Landscape, promises a great deal, but the subtitle, Twentieth-Century Excavations, indicates what is, in fact, its main concern. Exon, Gaffney, Woodward, and Yorston (2000), in contrast, devote a short book to a discussion of the landscape around Stonehenge. However, their study is almost exclusively concerned with monument inter-visibility, combining primarily the use of GIS data with some ‘phenomenological’ fieldwork. Although they discuss the approach to Stonehenge along the Avenue and from elsewhere in some detail (see below), they do not consider Stonehenge itself, presumably because of Allen’s pre-existing work on the visual field from the monument itself. Both of these studies very usefully concern themselves with issues of monument visibility, providing important insights that inform the discussion here. But other aspects of the landscape around Stonehenge, principally the form and the topographic character of the hills and ridges, the river valleys and coombes, or dry valleys, are scarcely considered at all. Discussion of such landscape features around Stonehenge is confined by Allen to mentioning which near or more distant ridges or hills can be seen. Exon and associates throughout their book rarely consider any other aspect of the landscape beyond monument visibility and intervisibility. In both these studies, the Stonehenge landscape and its topography tend to be considered only in terms of a series of monuments that at various times are visible or not. In other words, ‘culture’ is writ large in these studies, but ‘nature’ has been virtually excluded. A much more holistic approach is adopted here, paying as much attention to the ‘natural environment’ of Stonehenge as to the positioning and visibility of monuments within it.

The first part of this chapter aims to address, and attempts to answer, one simple question: Why is Stonehenge located where it is in the landscape? Why here? Why this place? In all the voluminous literature on Stonehenge, this question never appears to have been directly addressed. In attempting to provide an answer to this question, we show the manner in which a consideration of the monument in its landscape context provides the basis for a novel interpretation of the architecture of Stonehenge itself and the locations of the Bronze Age barrow cemeteries around it, which forms the second part of the discussion.

In relation to the question raised above, this chapter presents a few of the preliminary results of a phenomenological landscape survey forming part of the Stonehenge Riverside Project (see Parker-Pearson et al. 2006). This survey involves the description and analysis of a 180-sq-km area of land with the henge monument of Durrington Walls at its centre (Figure 3.1). The area covered in this survey includes the entire landscape area covered in the ‘Stonehenge environs’ project (Richards 1990; Figure 3.2), that in Cleal and associates (1995), and the far wider area considered by Exon and colleagues (2000) (except to the south of their ‘enlarged study area’). It extends considerably farther to the east of the Avon and to the north in the Salisbury Plain army training ranges than does the Exon landscape study. Research has involved walking this entire landscape and studying in the field all known barrows and the locations of ring ditch sites recorded from aerial photographs.

FIGURE 3.1 Stonehenge Riverside Project landscape survey area showing some of the places mentioned in the text.

FIGURE 3.2 Landscape and barrow distribution in the vicinity of Stonehenge showing the Avon valley and coombe systems.

Stonehenge in Its Landscape

What is remarkable about the location of Stonehenge in its immediate landscape is that it appears to be absolutely unremarkable. Allen rightly notes that when looking toward Stonehenge from any direction, one sees that the location is undistinguished: ‘without the monument in place it would not easily be distinguished from the gently undulating surrounding countryside, and it cannot be said to form an obviously important landscape feature from any direction’ (Allen in Cleal, Walker, and Montague 1995: 37). The monument is located on virtually flat ground on a very gentle west-east slope that steepens markedly as it approaches the dry valley system of Stonehenge Bottom some 400 m distant to the east. Immediately to the north and the south of the enclosing bank and ditch, the land dips away toward shallow coombes running down to Stonehenge Bottom. The drop in height to the bottoms of these coombes is about 10 m in about 300 m to the north and 500 m to the south. To the west, the land rises by a similar amount. The area in the immediate vicinity of Stonehenge is ambiguously delimited. It is not located on a well-defined ridge or spur, of which there are many in the surrounding landscape. The land on which it is built is only 100 m high. There is absolutely no drama with regard to its location. The drama and theatrical power of the monument seem to derive entirely from the sheer size and height of the stones, and without these the place would long since have been forgotten.

In essence, Stonehenge confounds the perhaps all too contemporary expectation that such an impressive monument might be located elsewhere in the landscape, for example on the top of the Beacon Hill Ridge 7 km to the east or perhaps on the Sidbury Hill summit 12 km to the northeast or, nearer, on the Durrington/Larkhill ridge 2 km to the north (see Figure 3.1). However, monuments and barrows of any kind seldom occupy the very highest hill and ridge summits in the 180 sq km considered in the landscape survey, and even some more localised high points and ridges are often entirely avoided. Similarly, very few barrows are located in the ‘depths’ of this landscape, at or near the bottom of coombes or river valleys. The vast majority occur in intermediate locations, often on the mid-points of gently sloping ridges and spurs. The location of Stonehenge is thus quite typical of those occupied by the many and somewhat later Bronze Age barrow cemeteries in the area. It is absolutely ordinary in this respect. Perhaps this fact is not so surprising in the light of the location’s use as a major cremation cemetery in Phase 2, before the erection of the stones (ibid.: 115). In many respects, the location might be regarded as conforming to an expected norm. But although it conforms to the position of many later Bronze Age barrows, it was actually built 500–1,000 years earlier.

Allen (in Cleal, Walker, and Montague 1995) discusses Stonehenge in relation to a ‘visual envelope’ around it and considers both views out from Stonehenge and views into the monument in relation to a ‘foreground’, the nearest ground to the ditched enclosure, a ‘near horizon’ created by slight ridges, and a ‘far’ and a ‘distant’ horizon. Such horizons at different distances from the monument frequently merge, and, in practice, it is very difficult to distinguish them. Furthermore, even within parts of the immediate ‘visual envelope’ around Stonehenge, there are lower lying areas along Stonehenge Bottom, to the north and the east, that cannot be seen from the monument, nor is it visible from them. The interior of the visible field is thus more complex than that represented and gives a misleading impression that everything within it is visible (see Figure 3.3). Allen shows how important Bronze Age barrow cemeteries— principally those to the south on Normanton Down; to the east; those running along King Barrow ridge (the New and Old King Barrows); and the Cursus group of barrows to the northwest—run along the edges of his ‘near’ or ‘far’ horizons, indicating that they were deliberately located so as to be visible, running along the skyline, from Stonehenge itself.

Some, but by no means all, of these barrows are indeed monumental and dominant landmarks when seen from the perspective of the Stonehenge enclosure. Beyond this horizon barrows cannot be seen, but other topographic features are visible in the far distance, notably the Beacon Hill Ridge to the east and Rox Hill to the south.

This is a rolling chalk downlands landscape in which topographic distinctions are subtle. It has been, and still is, primarily shaped by the agency of water. Throughout the study area, these seven main topographic elements may be distinguished:

The Avon river valley, the only perennial water source.
The winterbourne river valleys of the Till and the Bourne River and the Nine Mile River to the west and the east.
The coombes or dry valley systems that run into these perennial or seasonal watercourses.
Well-defined and smoothly sloping ridges and spurs of various forms running between these valleys and coombes.
More rounded localised high points such as Rox Hill, Oatlands Hill, and Robin Hood’s Ball.
More amorphous and ambiguously defined sloping areas of slightly higher ground dissected by coombes.
The Beacon Hill Ridge, with a pronounced northern scarp slope and a much gentler and more irregular and dissected southern dip slope.

Stonehenge is located in a position in the landscape that can be classified as category 6 above. It is directly linked to the Avon by the ceremonial pathway of the Avenue. The Avon itself is directly or indirectly linked to all the other winterbournes and coombes in the study area or beyond it to the south. The Till is linked to it via the Wylye to the west; the Bourne joins it to the east, as does the Nine Mile River. All the coombe systems link in to the same overall

FIGURE 3.3 Visual envelope in the immediate vicinity of Stonehenge. The area bounded by the near horizon is unshaded.

dendritic system. Thus the Avon effectively articulates and joins the entire immediate and larger landscape around Stonehenge. The link created between Stonehenge and the Avon thus positions (see below) the monument at the centre of a localised world defined by water, the source of all life.

Stonehenge is also directly linked to the Avon by a ‘natural’ route: the course of the Stonehenge Bottom/Spring Bottom coombe system, across which the Avenue passes to the northeast. Looking out from the Stonehenge enclosure, one can see the line of Stonehenge Bottom quite clearly—in particular, the eastern side. This coombe system is by far the longest and most reticulated in the study area. It runs from Lake, on the Avon, for over 5 km, twisting and turning and branching to the west and the east (Figures 3. 1 and 3. 2). (The place name ‘Lake’ or ‘Lac’, adopted during the early Medieval period or before, signifies the presence of a large sheet of water distinct from the river itself.) Not only is it the longest coombe, it is also the most complex and is also unusual in taking a north-south course for much of its length. (Most other coombes run from the NW to the SE or the NE to the SW.) Its shorter western branches run to the south and the north of Stonehenge, whose immediate landscape is thus contained or enclosed on three sides, to the east, north, and south, by this coombe. By the river Avon, the ‘entrance’ to this coombe system is marked by a large and prominent barrow to the south situated high up on the edge of the coombe and by three further barrows (now ring ditch sites) to the north, a point also cogently noted by Exon and colleagues (2000: 91), who suggest that this barrow represents a portal into the Stonehenge landscape from the south. It is one of a very few in the entire landscape visible from the river Avon on a canoe journey down the river from the north to the south. It appears to mark a turning point toward Stonehenge and away from the river.

Morphological research has demonstrated that Stonehenge Bottom has virtually no colluvium within it, whereas thick colluvial deposits do occur in the coombe around which Durrington Walls was constructed (Richards 1990: 210–211). The reasons for this situation remain uncertain, but one of the possibilities is the removal of colluvium by running water. It is interesting to note that water has been observed by the present landowners flowing in Stonehenge Bottom south of the A303 road, and flooding has occurred at Lake near to the Avon; at times of heavy rainfall, there is often standing water. Stonehenge Bottom differs from other coombes and river valleys in the area in that it is neither truly a dry valley nor a seasonal winterbourne. Stonehenge is thus directly linked with both the only perennial source of water in the area, the Avon, and an exceptional coombe system of unpredictable character. In general, our knowledge of the Neolithic water table is inadequate, and water extraction has drastically reduced the water table, affecting river and stream levels throughout the area. The Nine Mile River, which the military started tapping in the early twentieth century is now completely dry in the summer for most of its course, as is the Bourne.

The river valleys and coombe systems both define and divide this landscape. Their courses delimit areas of higher ground and provide well-defined routes of movement through it. They can be conceptualised in terms of boundaries, transition points from the lowest to the highest ground, and as providing pathways to follow through the landscape. They are also the places where sarsen stones are typically exposed and ‘congregate’, as we know from the few dramatic sarsen-filled coombes that still exist (having survived quarrying) in the Marlborough Downs to the north of the Stonehenge landscape. The coombes, mythologically understood, give birth to sarsen stones. They may also give birth to water, either seasonally or unpredictably. The association of coombes with water in various ways would have been noticed by prehistoric populations, as would their resemblance to river valleys with water, such as the Avon. A problem that might have required a mythological explanation could have been: Why did these rivers of the past run dry?

Another important factor in the location of Stonehenge was its visual relationship to the Beacon Hill Ridge to the east and Sidbury Hill to the northeast. Both the Beacon Hill Ridge and Sidbury Hill punctuate the skyline in a distinctive manner in this landscape. They are, relatively speaking, ‘jagged’ compared with the rest of the Stonehenge Landscape, where the localised topography of the rises and ridges and coombe systems winding their way through the chalk downlands is either slight and indistinct, or if higher, rounded and smoothly rolling. These are by far the highest hills in the area, and indeed some of the highest in Wiltshire, with the Beacon Hill Ridge reaching a maximum height of 204 m at its western end and Sidbury Hill rising to 223 m.

The Beacon Hill Ridge (Figure 3.4) is by far the most dramatic in the study area. At the end of their landscape study, Exon and associates state that ‘we became overpowered by the influence of Beacon Hill. Lying toward the eastern margin of our study area, its high and jagged profile forms a visual focus for many monuments’ (Exon et al. 2000: 108). This is indeed the case. The ridge extends for about 4 km on an approximate southwest to northeast alignment. Stonehenge is located in the landscape so that most of the northern scarp slope of this ridge with its distinctive summit areas is visible. Had it been sited farther to the south, only the far western edge of the ridge would be visible, and the effect of seeing different summit areas would be lost. This ridge comprises five distinctive summit areas with lower ground in between; because of its orientation, most of this ridge can be seen from Stonehenge. Three of these summit areas (Jukes Brown 1905 notes only two) and Sidbury Hill have a thin but nevertheless distinctive capping of smooth and rounded flint and quartz pebbles in a clayey soil overlying the chalk, known geologically as the Reading Beds (Jukes Brown 1905: 40). These pebbles are round or oval in form, the largest being 5–6 cm in diameter, the smallest 2 cm. They are water-worn and perfectly smooth and rounded. They vary considerably in colour from white to black, to red, yellow, and brown (Figure 3.5). Their presence explains the unusual stepped form of the Beacon Hill Ridge contrasting with all other chalk ridges in the Stonehenge area, which have much more rounded and even contours, lacking distinctive and discrete summit areas.

FIGURE 3.4 Beacon Hill Ridge seen from the west.

Now, the final section of the Avenue, after it dramatically bends to turn and run up directly to Stonehenge, is orientated on a direct northeast line toward Sidbury Hill (the highest point in this landscape). The rising midsummer sun striking the Heel Stone before shining into the interior of Stonehenge emerges from behind Sidbury Hill in the distance, thus emphasising the symbolic significance of this pebble-capped summit (Figure 3.6). Today Sidbury Hill cannot be seen from Stonehenge, because trees and buildings on the Larkhill/ Durrington ridge to the northeast block the view. GIS-generated viewsheds produced by Mark Dover of the Stonehenge Riverside Project team show that the summit area of Sidbury Hill would probably have just been visible in the Neolithic if one had been standing on the western or northern sectors of the

Figure 3.5 Pebbles on the Beacon Hill Ridge.

FIGURE 3.6 Sidbury Hill seen looking out from the entrance to Woodhenge.

bank of the Stonehenge Phase 1 monument (assuming a relatively open and treeless landscape, as demonstrated by Allen 1997).

In view of the visual and symbolic significance of the Beacon Hill Ridge and Sidbury Hill, a number of the architectural features of Stonehenge itself in its final phase, seen today, can be understood in a new manner. First of all, the internal space framed by the trilithons and taller bluestones is orientated on the same NE-SW axis as the Beacon Hill Ridge. This emphasis on a NE-SW axis is shared with a number of other approximately contemporary later Neolithic monuments. The oval timber rings at Woodhenge (Cunningham 1929) are arranged on a NE-SW axis, and the single entrance faces to the northeast, as did the single entrance to the Coneybury henge Richards 1990: 123). From both these monuments, Sidbury Hill is visible today (see Figure 3. 6) if one looks out through the entrances, and the midsummer sun can be seen rising up from behind it.

The significance of the Beacon Hill Ridge may have been important both much earlier and before the construction of Stonehenge, and after the final phase of its construction. It is intriguing to note that the line of earlier Mesolithic pine timber posts discovered in the Stonehenge car park (Cleal, Walker, and Montague 1995: 43–47) is orientated toward it. The ridge is visible from almost all the c. 25 Neolithic long barrows and c. 450 round barrows in the study region. By far the greatest concentration of Bronze Age barrows in the study region flanks the Nine Mile River, a winterbourne stream running roughly NE-SW. This river arises to the east in the same part of the landscape as Sidbury Hill and flows along the foot of the north-facing scarp of the Beacon Hill Ridge. Its confluence with the Avon is just to the east of the Durrington Walls henge. By comparison, the Avon, Till, and Bourne rivers, which flow approximately north-south, have far fewer barrows and barrow cemeteries associated with them. Thus a general NE-SW axis appears to have become the auspicious directional axis in the entire landscape after the final phase of the construction of Stonehenge and throughout the early Bronze Age (see also Darvill 1997: 180–181). At Stonehenge this orientation is present during the initial erection of the bluestones around 2600 B.C.E.

The Approach from the Avenue

Stonehenge as a locale in the landscape cannot be understood simply in terms of constituting a fixed place, that is, in terms of its specific location. Part of its meaning and significance was created through the process of the experience of moving toward it in the right way, and from the most propitious direction following the path of the rising sun. At least in the final phase of the construction of the monument, we know this to have been by walking along the Avenue. This is by no means the shortest or easiest or ‘least cost’ route to Stonehenge, whether dragging bluestones along it or not, as Exon and associates (2000: 72) have shown. In brief, the approach involves ascending from the Avon to the top of the King Barrow ridge, from which Stonehenge can be seen for the first time from the east, descending into and across Stonehenge Bottom, where it disappears from sigh—and then a dramatic change of direction to approach the monument again when it is very near indeed. Here we analyse in detail the final part of this journey to the stones and into the interior of the monument.

As one approaches Stonehenge walking along the final part of the Avenue from the northeast, arrival at the monument takes the form of an ascending pathway that flattens off as one approaches and enters the sarsen ring. The internal arrangement of trilithons gradually disappears, becoming concealed by the lintel stones of the outer sarsen stone ring. It becomes visible again as trilithons only after one has finally entered the outer ring. The tallest and most impressive part of the monument thus goes out of sight while the outer ring of stones dominates the visual perspective. In effect, this external ring of stones, becoming more and more dominant and higher and higher in relation to a person approaching the monument, continues the ascending path of the Avenue in a most dramatic and outrageous way. Passing the Heel stone and the Slaughter stone, to the left, one has only limited glimpses into the interior of the monument. The details of its internal structures are almost entirely concealed from view. From the outside there is no obvious entrance into the sarsen ring, but rather a series of slots to pass through, which one might choose. The two stones through which one should pass remain unmarked. One is confronted with a massive structure of strong verticals and bold horizontals (Figure 3.7). The landscape beyond the monument to the southwest is entirely blocked out. The only other monuments visible in the landscape are a few of the most monumental and massive barrows at the eastern end of the Normanton Down group to the south. These also disappear from sight as one walks up to the stones. It is clear that anyone entering the monument for the first time in the correct way would need to be led, or provided with guidance, from someone with knowledge of the internal structure.

Inside the Stones

Passing through the outer circle of sarsen stones, one encounters a ring of bluestones, the two highest of which (stones 49 and 31: see Figure 3.8), concealed from the outside, flank the entrance way through this circle. Only after having passed through the outer sarsen ring can one see the horseshoe-shaped internal arrangement of trilithons and bluestones, as well as the outer circle of bluestones surrounding it. The concealment of this inner structural arrangement from the outside world and a view of almost all the blue stones from whatever direction one approaches the monument creates a crucial distinction between the internal and the external spaces of the monument, establishing a fundamental distinction between Stonehenge as viewed from the outside and as seen from the inside.

FIGURE 3.7 Stonehenge seen from the northeast as one approaches the monument along the final stretch of the Avenue.

The inner space of the monument is effectively graded, both by the increasing height of the sarsen trilithons and bluestones from front to back (or to the southwest) and the enclosing architecture of the horseshoe (Figure 3.9). The permeability of the outer sarsen ring thus contrasts with the terminal space of the horseshoe arrangements of stones beyond which one should not pass. There was only one correct way into the inner part of the monument and only one way out. Such an architectural arrangement of stones, it might be noted, is typical of Neolithic passage graves, which similarly have only one entrance and exit from the internal space of the structure and in which the internal arrangements of stones and corbelling rise to the back. All this suggests that the central interior space of Stonehenge was an unroofed temple constructed using the same general design principles as used in earlier megalithic tombs. Such an observation strengthens an interpretation that this monument was associated with the ancestral dead (see Parker-Pearson et al. 2006; Parker-Pearson and Ramilisonina 1998; Whittle 1997: 163).

FIGURE 3.8 Plan of Stonehenge showing the arrangements of sarsens and bluestones.

Although we acknowledge the graded nature of the central bluestone horseshoe, and the overpowering grandeur of the similarly graded encasing sarsen trilithons, we note that it was to the pale sandstone Altar Stone that this entire architectural edifice referred. Today, the Alter Stone lies little noticed, embedded in the turf and partially covered by the lintel (156) and fallen eastern upright (55) of the tallest sarsen trilithon. Indeed, its upper face is worn and polished through generations of visitors walking over its surface to gaze at the collapsed great sarsen trilithon. At one time, however, this stone stood proudly in a central position enclosed by the inner bluestone horseshoe (Atkinson 1956: 45) and providing a striking focal point (Stone 1924: 1). When erect, it stood to c. 4 m in height, and, although dwarfed by the great trilithon, it towered above the surrounding bluestones.

FIGURE 3.9 View across the central area of Stonehenge showing the grading of height of the bluestones in the inner ‘horseshoe’.

The hidden presence of the bluestones within the monument, situated both inside the outer sarsen ring and inside the trilithon setting, strongly suggests that the whole building project was designed to guard, shield, and conceal the exotic bluestones from the outside world. The bluestones were also of great antiquity, having formed the first stone architecture at Stonehenge (cf. Bradley 2000a: 94). Consequently, they may have needed to be surrounded by the sarsen stones to protect their magical powers and symbolic connotations.

Furthermore, there are important distinctions among the outer ring of bluestones, the internal horseshoe-shaped arrangement, and the central Altar Stone. All but two now-fallen stones (Nos. 36 and 42), which once formed lintels for trilithons, in the outer ring of bluestones, are unshaped and retain their natural forms and individual character. The size and shapes of these stones vary greatly, resembling those that may be observed on the Preseli mountains today. These stones are of mixed local origin but may all come from nearby sources at the eastern end of the Preseli mountains (Thorpe et al. 1991). Rhyolite, spotted and unspotted dolerite, and volcanic ash are all used. The inner bluestones are much taller; all are skilfully dressed and of spotted dolerite except for one (Atkinson 1956: 42). The uniformity of the material used for the stones in the inner horseshoe thus contrasts with the diversity of types of stone employed to construct the outer bluestone ring. Atkinson notes that ‘in every case where the upper part of the pillar survives intact, its top surface has been dressed flat and level. . . two pillars at least once terminated in a tenon’ (ibid.: 43). At least six, possibly seven, of these stones formed part of a previous structure that included at least two trilithons (ibid.: 44).

Of great consequence is that this megalithic architecture was of a form unlike that of any other stone monument in late Neolithic Britain. Its complexity is demonstrated not only by shaped components of trilithons but also by the presence of more complex forms of stone ‘joinery’. Bluestone 68 has the beautiful groove running down its western side. Atkinson identifies the presence of the broken bluestone stump 66 with the remains of a tongue in a corresponding position: ‘it may be accepted that at one time these stones stood side by side, the tongue of one fitting into the groove on the other’ (ibid.: 44). But, in employing at least two pairs of ‘tongue and groove’ jointed stones, the previous bluestone structure was of even greater complexity than envisaged by Atkinson. This complexity is revealed in J. F. S. Stone’s observation that of the remaining tongue and groove stones (bluetones 66 and 68) none actually fitted another (1953: 13). Hence, the bluestones that form the inner horseshoe were exotic not only in being derived from South Wales but also in being components of a unique and incredible megalithic monumental architecture.

Again, apart from its enhanced stature, the central Altar Stone stands out in its difference. While the inner bluestone horseshoe comprises the remnants of an earlier monument, presumably mainly formed of spotted and unspotted dolerite, the Altar Stone is a pale, fine-grained calcareous sandstone. Previously identified as originating from the Cosheston Beds that outcrop around Milford Haven (Atkinson 1956: 46; Thomas 1923: 244–245), the 6-tonne stone has now been recently suggested to derive from the Senni Beds, possibly from a more eastern location near the Brecon Beacons (Kellaway 2002: 59). A more cautious approach to provenance is adopted by Ixer and Turner (2006: 7), who suggest that the important issue is not the exact source location within the Senni Beds but rather that such stones outcrop in locations far removed from either the Preseli Hills or Milford Haven. In this respect, the nearly 5-m-long Altar Stone assumes even greater significance in being ‘exotic’ in comparison to the commonality of the Preseli dolerites of the inner horseshoe (and earlier bluestone monument).

Many of the bluestones forming the inner horseshoe were reshaped so as to resemble ground stone axe blades thrust into the ground with their blades facing down (Figure 3.10). None of the bluestones in the outer circle look like axes at all. These differences between dressed and undressed bluestones, taller and thinner stones, stones that resemble axes and those that do not are further accentuated by the contrast between the outer circular space formed by the bluestones and the inner oval space, open to the northeast.

FIGURE 3.10 Axe-blade-shaped bluestones forming the inner ‘horseshoe’.

The axe-shaped forms of the bluestones is particularly interesting to note in relation to the occurrence of copper axe engravings on some of the sarsens and the presence of functionally useless but symbolically powerful chalk axes deposited at Woodhenge (Pollard 1995: 149). No carvings are known on any of the bluestones themselves. These carvings occur on the outer faces of stones 3 and 4 and on the inner face of stone 53, one of the trilithons. Another may occur on stone 5, but its position is unknown (Lawson and Walker in Cleal, Walker, and Montague 1995: 30–32). These carvings all occur on the lower parts of the stones, with the lowest immediately above ground level. The majority resemble flanged axes of early Bronze Age date (Figure 3.11). All are unhafted axe blades with the blades pointing vertically up the stones. These, and the axe-blade-shaped bluestones, clearly indicate the continuing

FIGURE 3.11 Axe carvings on the inner face of stone 53 (source: Cleal, Walker, and Montague 1995: Fig. 20).

significance of axe symbolism from the Neolithic into the Bronze Age. However, there is a significant difference insofar as the bluestones are shaped in the form of Neolithic axe blades that are located at the inner core of the monument. The axe carvings on the sarsens thus indicate both symbolic continuity with the past, and difference, and only some, unlike the bluestone axes, are hidden within the inner sanctum of the monument. Those on the external faces of stones 3 and 4 would be dramatically illuminated and highlighted by the equinoxal sunrise and do not relate at all to movement toward the monument along the Avenue.

The shaped bluestones forming the inner arrangement are hidden within the horseshoe trilithon arrangement, also consisting of dressed stones with both the outer sarsen ring and the inner trilithons being furnished with lintels. The bluestones that never supported lintels form a permeable ring, alluding perhaps to a yet earlier bluestone circle never elaborated with trilithons. Its presence and that of the bluestones in the inner oval arrangement thus served to objectify the presence of earlier structures at Stonehenge and the past in its present and final form, in which the exotic bluestones once visible from the landscape and the outside became hidden inside. At the same time, it was only the more local sarsen stones that had lintels or were used for trilithons. They were clearly chosen for their brute monumentality, dwarfing a person, and their presence would clearly make any attempt to retain a bluestone trilithon structure appear like the work of lesser beings in comparison.

The trilithons forming the inner part of the structure were carefully chosen pairs of stones with capping lintels. The fact that these and all the other extant sarsen stones in the monument were dressed does not mean that their surfaces are smooth and uniform. Whittle has made the important observation that the surfaces and the dressing of the trilithon uprights are very different and differ between the external and internal faces from stone to stone (Whittle 1997: 155). Examining the internal broad faces of these stones seen from within the innermost oval space of the structure, one sees that there is a striking series of repetitive contrasts between each pair of stones. Many are riddled with hollows and holes and have a very uneven surface. In each of the surviving three pairs of stones still standing, one of the stones has a comparatively rough surface with many surface depressions, holes, and irregularities. The other, by contrast, is almost perfectly smooth and regular in form all over its surface. So, in each stone pair, one of the stones retains a surface, or parts of a surface, that uniquely individuates it, whereas the other is artificially shaped in such a manner as to remove all traces of its individual and original identity as a ‘natural’ or unworked stone (Figures 3.12 and 3.13). In each case, it is the monolith on the left-hand side of the pair that is smooth and regular in form and the one to the right that is much more irregular. This pattern of pairing stones with smooth and rough internal surfaces is likely to have been repeated in the cases of the two trilithons where today one of the stones (55 and 59) has collapsed with only the outer faces visible and both irregular.

FIGURE 3.12 Inner faces of stones 51 and 52 (left) and 53 and 54 (right).

This consistent contrast between comparatively smooth and comparatively rough broad faces of the stones seen from the inside is, however, not repeated when the same stones are seen from the outside. The external faces of stones 51 and 52 are both quite uniform and smooth. Stones 53 and 54 and 55 and 56 have external faces that are both smooth and rough. Stones 57 and 58 possess smooth external faces, whereas stones 59 and 60 both have rough external faces. So, while all combinations of smooth and rough or smooth and smooth or smooth and rough faces occur on the outside of the trilithon oval, a deliberate choice was made to choose stones with a rough and a smooth surface to erect on the inside, a deliberate pairing of stones with very different and contrasting surface characteristics creating an internal architectural space that was very different when seen from the inside. Here it is worthwhile noting that from a human perspective all the broad surfaces of the stones of the inner trilithon can be seen only when one is standing and looking around in the inner space. As one walks around and outside the same stones as the broad face of one comes into view, the previous stone disappears out of view. Thus a consistent pairing of stones with rough and smooth surfaces would not be likely to be appreciated or be so visually striking when seen from the outside.

FIGURE 3.13 Inner faces of stones 53 and 54.

Stone 54 and the fallen stone 55 (see Figure 3.8) in the arrangement of trilithons contrast significantly with all the others. The other stones are all grey in colour. These two stones are unusually brown. This strongly suggests that at least two different sources of sarsens were utilised to construct the inner arrangement of trilithons and that in two out of the five trilithons, including the highest of all, stones from these different sources were deliberately paired together. This replicates the use of different kinds of bluestones from different sources in the outer ring.

The inner arrangement of sarsen trilithons differs substantially from the outer ring of sarsens, not only in terms of their height and dimensions but also in terms of the gaps between the pairs of uprights through which nothing of the outside landscape can be seen. For a person standing in the central space of the monument, the outside world is completely screened off, and no barrows are visible. This outside world is only partially visible when one moves and looks through the gaps between the trilithons. Stonehenge, from the inside, is very much a monument that focuses attention on its internal architecture. Unlike with every other stone circle in Britain, the intention seems to have been to exclude the outside world. Although many of the locations of the numerous Bronze Age barrows in the surrounding landscape appear to have been deliberately chosen in relation to Stonehenge, they were not visible from the central part of the interior (Figure 3.14).

FIGURE 3.14 View out from the centre of Stonehenge looking east.

Only two massive bell barrows are visible when one moves around in the central space and looks out through the gaps between the trilithons: the bell barrow to the southwest of Stonehenge behind which the sun sets on the shortest day of the year (the so-called sunset barrow) and the Bush Barrow, with its fabulously rich grave goods (Ashbee 1960: 76-78) (see Figure 3.2). This strongly suggests that these two barrows were located in a very specific relationship to the central space of the monument following its construction in the form that we see today. The locations of many of the other Bronze Age barrows indicate that although a view to Stonehenge from them was important they were not located so as to be seen from the centre of the monument. In other words, views to Stonehenge from outside it and the surrounding landscape were far more significant than views of that landscape from the central space of the monument defined by the internal trilithons and bluestones. Thus part of the significance of Stonehenge in its final phase of construction was that it was deliberately designed so as to be seen from a distance rather than being a place from which to view the world beyond. There is often a substantial difference between the distance from which one can see looking out from the monument and from which one can see to it from the surrounding landscape. This difference is, of course, because the outer sarsen ring, and particularly the trilithons, are substantially taller than the height of an observer standing in the circle—in fact, more than three or four times the height of a person (6 m to over 7 m high). Thus it is possible to see the tops of the trilithons from some parts of Stonehenge Bottom to the east but not the bottom of this coombe from Stonehenge itself. Similarly, the tips of the trilithons of Stonehenge can be seen from the eastern end of the Winterbourne Stoke barrow cemetery to the east, but none of these barrows are visible from the monument. Stonehenge can be seen from Oatlands Hill 3 km to the southwest, but Oatlands Hill cannot be seen from Stonehenge. From the barrow cemetery at Durrington Down to the north, Stonehenge can be seen, but not vice versa (see further discussion of these landscape views into the monument below).

A more substantial view of the landscape beyond the monument is possible when one walks a circuit between the outer bluestone and sarsen rings. The sarsens, with their lintels, continually frame and break up this perception of the landscape. It has to be experienced in terms of a series of windows breaking up the continuity of the topographic forms of the ridges, groups of barrows, and the line of Stonehenge Bottom. By far the most dramatic view is to the east to the King Barrow ridge, forming the near horizon, and the Beacon Hill Ridge beyond, forming the distant horizon (Figure 3.15). Walking out from Stonehenge through the tallest bluestones in the outer ring and sarsens 30 and 1, one notes that this is the last gap between the five pairs of sarsens on the northeast side (stones 5–29) through which the Beacon Hill Ridge can be seen directly in front, when one is looking out.

On the western side, the view is curtailed by gently rising land to only about 250 m. To the southwest, the horizon is considerably longer, whereas to the south, Rox Hill, 3.5 km away, is on the distant horizon, but it is prominent today only because of the distinctive clump of trees on its summit. By far the most prominent Bronze Age barrows seen from the monument, apart from the Bush barrow and the Sunset barrow and the nearby bell barrow immediately to the east of it, are the six massive New King Barrows running along a ridge 1 km distant to the east.

FIGURE 3.15 New King Barrows seen from the Stonehenge enclosure looking east.

Architectural Order and the Ordering of the Landscape

Whittle notes that the stepped character of the sarsen settings is an important aspect of the architecture of Stonehenge (Whittle 1997: 150). He suggests that in some way this might be linked symbolically with a hierarchy of spirits or beings, the most powerful being high up and associated with the air. Going beyond this some more precise observations with regard to the stepped character of the stone settings can be made in relation to its landscape. There are five trilithons at Stonehenge, precisely matching the number of summits on the Beacon Hill Ridge. The Beacon Hill summits are graduated in height, with the highest at the southwest end. The tallest trilithon at Stonehenge is similarly located at the southwest end of the central space, thus suggesting a mimetic relationship between the orientation and graded height of the trilithons and the sequence of ridge summits. The inner horseshoe-shaped arrangement of blue-stones is similarly graduated in height to the southwest. So, the cultural form of the interior of the monument is the landscape in microcosm. Furthermore, the materials of the summit areas of the pebble-capped Beacon Hill Ridge and the Sidbury summit are alien to the area. Water-worn pebbles are found nowhere else in this landscape. In contrast to the pebbles on these hilltops, the stones encountered in the beds of the Avon, Nine Mile River, Bourne and Till rivers in the vicinity of Stonehenge and along Stonehenge Bottom itself are all jagged, angular, and irregular (Figure 3.16). So, pebbles do not occur in the river valleys directly associated with water, where we might perhaps most expect to find them, but on the very highest points in the landscape, where they might be least expected. This appears to be an inverted world!

The six New King Barrows on the nearby ridge to the east of Stonehenge stand out from all the others in the Stonehenge landscape in a number of important respects. They are all huge and monumental bowl barrows more or less equally spaced along the ridge top with significant gaps between each barrow. Nowhere else in the study area is such a large number of huge and regularly spaced barrows found in such close proximity. In other places, and in other barrow cemeteries, in the study region there are barrows of similar or even greater dimensions, but they occur only singly or in pairs, and their spacing is often irregular, or they may be conjoined as on Normanton Down and in the Cursus group. Clearly these barrows, which we know to have been built of stripped turves with a chalk cap obtained from digging the surrounding ditch (Cleal and Allen 1994), were constructed so as to be as prominent as possible from Stonehenge. Seen from Stonehenge, these six massive mounds punctuate the skyline, breaking up the otherwise smooth and rounded contours of the ridge in a manner that simply does not occur in relation to the barrows elsewhere running along the edge of its ‘visibility envelope’ (Figures 3.15 and 3.17). Our interpretation is that their relationship to the five summits of the Beacon Hill Ridge and to the summit of Sidbury Hill is again mimetic (six mounds and six summits). The monumental New King Barrows thus reiterate the symbolic significance of these pebble-capped hills to their east in relation to Stonehenge itself. These barrows have an inversed stratigraphy, chalk covering the soil, just as the presence of pebbles on the hill summits to the east is an inversion of a norm. Rather than beach pebbles being found low down by the sea, they are instead encountered far inland and next to the sky. The upside down King Barrows mimic the inversion of the wider world found on the ridgetop.

FIGURE 3.16 Gravels in the bed of the Nine Mile River.

FIGURE 3.17 One of the New King Barrows seen from the west.

Pebbles may have signified the sea and the connectedness of communities travelling by water and its buoyant potency. Pebbles from the summit areas of either the Beacon Hill Ridge and/or Sidbury Hill have been recorded from the recent excavations at Woodhenge in 2006 directed by Joshua Pollard. A substantial hollow was found directly underneath the bank of the late Neolithic henge on the southeast quadrant of the monument. This hollow was created by a fallen tree. In it, early Neolithic pottery (the remains of a carinated bowl) was found, together with bones and flint in the upper fill. These were directly associated with a deposit of pebbles brought from the Beacon Hill Ridge. At Stonehenge, Hawley records the presence of pebbles in two of the Y holes (Hawley 1925: 37–38); however, because they, unlike the sarsen and the bluestone chippings, were unlikely to have been of any interest to him, how many were left unrecorded remains uncertain. In this respect, Green remarks, in the context of a general review of stones found in the ‘Stonehenge layer’, that well-rounded flint pebbles occur at Stonehenge ‘over the whole period of its construction’ (Green 1997a: 5).

While the bluestones were an alien material from an exotic and distant source, the pebbles on the hill summits were an exotic local material. Excavations at Stonehenge have revealed that the entire interior of the monument was covered with sarsen and bluestone chippings. The bluestone chippings outnumber those of sarsen in a ratio of 1:3 (ibid). This is surprising in view of the fact that the dressing of the huge sarsen blocks would create much more waste material. It seems likely that during the construction of the final phase of the monument at least the bluestones were being dressed in situ, whereas the sarsen blocks were largely dressed away from the monument in the landscape and then brought to the site and erected. Or, alternatively, many bluestone chippings were collected to be deliberately deposited within the circle. Although it is very easy to appreciate the significance of the imported bluestones themselves, what is perhaps more surprising is the fact that bluestone mauls were brought from southwest Wales, too, further emphasising the magical significance and power of these stones.

An Unfinished Structure?

The existing arrangement of sarsens, with or without lintels, in the outer circle of Stonehenge covers only about three-quarters of the circumference of the circle. There are many stones absent on the southwest side, where the visual field from the monument is shortest and directly opposite the most significant axis of approach to Stonehenge along the Avenue. The outer sarsen ring of Stonehenge was, we think, never completed, and the reason may well be either that there simply were no stones of sufficient size to finish the building project or that a complete ring of sarsens with lintels was never intended or required on the southwest side of the monument where the horizon line is restricted and from which Stonehenge was never meant to be approached.

The internal trilithons, somewhat reduced in height, would have been sufficient in number to complete the perfect outer ring in the absence of any other stones of suitable size. Precisely where in the landscape surrounding Stonehenge the sarsens were obtained still remains a mystery, since today there are none of a similar size either in the immediate vicinity of Stonehenge or anywhere on the Marlborough Downs (cf. Ashbee 1988; also see Bowen and Smith 1977; Green 1997a: 5–7, 1997b: 260–263; Stone 1924: 44–57, 1926).

The idea of an external perfect sarsen ring was fully realised only on the northeast side of the monument facing toward the important approach from the Avenue. When one approaches from this direction, Stonehenge appears as relatively ‘open’. Seen from the southern side through the entrance through the outer bank and ditch, the visual perspective is totally different, with the interior oval space defined by the trilithons and the tall bluestones being completely concealed (Figure 3.18). A smaller monolith (stone 11) and an adjacent sarsen stone (No. 10) completely block any view into the inner space. This side of the circle acts as a screen, effectively blocking off movement into the circle itself from this direction. Stone 11 is both much shorter in height and significantly different in shape from the other sarsens in the outer ring (Figure 3.19). Although in the correct position to continue the outer ring on the southern side, it could never have supported a lintel. Atkinson suggests that the upper part may have been broken off and removed (Atkinson 1956: 24), but there is absolutely no evidence for this. Not only is this stone much shorter than the others, it is also significantly smaller in breadth and thickness. Hence, although there exist the collapsed upright (stone 12) and socket for missing stone 13 in the southwest, even if these once constituted a standing trilithon arrangement, they were never connected to the outer circuit of sarsens. This lack of conjoining stones reveals that Stonehenge was built in a piecemeal and probably different manner at the ‘rear’ of the monument.

FIGURE 3.18 View toward Stonehenge from the south.

Similarly, stone 16, again, in the correct position to continue the outer sarsen ring on the southwest circuit of Stonehenge, is completely anomalous in shape (Figure 3.20). Its sinuous form, thick base and sides, and tapering shape bear far more resemblance to a menhir, and its thin top is unlikely to have supported a lintel. Indeed, this stone is famous for its clear tool-marked surface (for example, Atkinson 1956: Fig. 8; Cleal, Walker, and Montague 1995, Plate 7.1); however, careful examination allows these marks to be re-interpreted as the results of extensive episodes of axe polishing subsequently pecked over. Overall, there is no evidence for a continuation of the outer sarsen circle beyond the socket for stone 13 in the southwest and the socket for stone 20 in the northwest.

In suggesting that the rear (southwest) area of the final Stonehenge monument was open and incorporated special and anomalous stones, we recall the initial bluestone architecture of Stonehenge. Here, too, a semi-circular arrangement was present, with an entrance having a NE-SW axis (Cleal, Walker, and Montague 1995: Fig. 80). Significantly, Cleal suggests that at this early time the focus of the semi-circular bluestone arrangement may have been the Altar Stone then standing in socket WA3639, C17 (ibid.: 188).