PERIGLACIAL GEOGRAPHY AND HOLOCENE LAND SURFACE

While Charles Darwin’s work, probably on Stones 8, 9 and 12 or 14, emphasised how the soil profile was raised by earthworm activity leaving the stones partially buried (Darwin Reference Darwin1881, 154–6; Field & Pearson Reference Field and Pearson2010, 43–4), Richard Atkinson’s (Reference Atkinson1957) investigation into worms and weathering indicated that the land surface at Stonehenge had been truncated quite considerably as a result of dissipation of the chalk sub-soil over some 4000 years. Groube and Bowden (Reference Groube and Bowden1982, 16–18), however, subsequently demonstrated that such truncation across the chalk downland was in great part a result of recent cultivation. Despite the anticipated truncation of the chalk outside the enclosure, the new surveys provide some indication of the state of the former land surface. The GPR reveals a series of north-east trending, gently inclined anomalies recorded at an approximate depth of 0.9 m downwards (Fig. 5). These reflections begin at too great a depth to relate to recent intervention and cannot be the signature of surface trackways (for trackways see Bowden et al. Reference Bowden, Barber, Field and Soutarin press: Field & Pearson Reference Field and Pearson2011). All share a similar inclination, dipping along a north-west to south-east strike on a comparable orientation to the periglacial features recently identified along the Avenue (Parker Pearson Reference Parker Pearson2012, 242–4). Similar anomalies have also been recorded recently in the GPR data from the wider WHS landscape collected by the Ludwig Boltzmann Institute (Baldwin Reference Baldwin2010; Gaffney et al. Reference Gaffney, Gaffney, Neubauer, Baldwin, Chapman, Garwood, Moulden, Sparrow, Bates, Löcker, Hinterleitner, Trinks, Nau, Zitz, Floery, Verhoeven and Doneus2012), and it seems most likely that they represent a geological interface, perhaps concentrated flint or marl seams embedded within the underlying chalk. It may be relevant that the Andover soils that cover the area are said to have a tendency to occur over ‘striped soil patterns’ (Soil Survey 1983).

Fig. 5 Stonehenge: a selection of GPR amplitude time slices; the arrows mark the periglacial features recorded at approximate depths of between 0.9 m and 1.78 m. ©English Heritage.

It is evident from excavation section drawings that the surface of the chalk at Stonehenge is not level and may rise in the south-eastern part of the stone settings. The chalk revealed in Gowland’s trench adjacent to Stone 55 rose markedly from one end of the trench to the other (Gowland Reference Gowland1902, 52, fig. 12), while Hawley (Reference Hawley1921, 22) noted that the natural chalk rose higher behind Stone 7. Here the surface expression of an ephemeral mound, 0.25 m high, survives on the surface as a topographic feature (Fig. 2). This measures some 15 m wide with a basal plinth that appears to be surmounted by a secondary tump and its west side is demarcated and emphasised by an irregular but c. 5 m wide shallow hollow or depression, no more than 0.25 m deep that is best depicted on the contour plot (Fig. 6). In the north-east, a rebate in the 102.25 m contour that the stones in the sarsen circuit cut across, provides emphasis. The mound was the subject of extensive enquiries to establish whether it is of recent or historical date, but photographs of Gowland’s site caravan parked upon it during the excavations of 1901 indicate that it was not the result of 20th century excavation spoil, or other disturbance, while a number of artists (and others) who produced sketches, watercolours, etc, of Stonehenge during the 18th and 19th centuries depicted its uneven or undulating surface. Additionally, the course of the Y and Z Holes appear to respect the low mound (see below) and it would therefore appear to pre-date them. Crucially, the natural interface between pits visible in the cutting made by Darvill and Wainwright (Reference Darvill and Wainwright2009, fig. 7; T. Darvill pers. comm.) indicates that the chalk lay close to the surface here and we therefore tentatively conclude that the mound is a surface expression of the geology rather than a structure of cultural origin.

Fig. 6 Stonehenge: the mound within the sarsen circle: terrain model and contours from field survey. ©English Heritage.

Fig. 7 Stonehenge: high Amplitude GPR reflectors. Numbers refer to anomalies mentioned in the text and taken from the sequence in Linford et a.l 2012. ©English Heritage.

Why such a feature occurs at this point is not clear, given that, as noted above, the chalk is highly susceptible to weathering and it is unusual for a natural mound to remain on the Wiltshire chalk (Newall’s mound set to the north-east along the Avenue 20 m south of the elbow, while set over a solution hollow, lies at a former field edge and may be related to the boundary – see Field et al. Reference Field, Bowden and Soutar2012). Elsewhere, earthen structures are known to retard weathering at, for example, Horslip long barrow (Ashbee et al. Reference Ashbee, Smith and Evans1979, 211) and it might be considered that the Stonehenge Layer contributed here. While of considerable thickness (Darvill & Wainwright Reference Darvill and Wainwright2009), however, that is more extensive across the area and not confined to the area of the mound. The only two locally occurring weather retardant materials are flint and sarsen and, since local flint is only thinly distributed here, in the absence of alternatives sarsen may have to be considered as the preserving agent. The circumstances determining natural sarsen emplacement have unfortunately seen little research and, Darwin’s observations notwithstanding, the degree to which they may rest on the surface or sink into it is unknown: both can be observed on Fyfield Down and additionally a link with solution hollows is recorded (Clark et al. Reference Clark, Lewin and Small1967; Bowen & Smith Reference Bowen and Smith1977). Several authorities have suggested that the sarsens were sourced from the area of Stonehenge or somewhere closely adjacent (see below). Darvill et al. (Reference Darvill, Marshall, Parker Pearson and Wainwright2012, 1029) have recently suggested that some stone was probably present on site at the outset; notably the massive cone-shaped Heelstone as it has been considered to have been extremely difficult to move (Johnson Reference Johnson2008, 121), an observation supported by measurements taken from the laser scan data which demonstrates that it is the heaviest stone on site (Abbott & Anderson Whymark Reference Abbott and Anderson-Whymark2012; Field et al. Reference Field, Linford, Anderson-Whymark, Barber, Bowden, Linford and Toppingforthcoming).

EARLY FEATURES

Ground Penetrating Radar has revealed the presence of several significant anomalies that lie in an area to the north-west of the Stonehenge enclosure ditch (Linford et al. Reference Linford, Linford and Payne2012) and which are sealed at an estimated depth of between 1.3 m and 1.6 m in a location beyond earlier earth resistance coverage and in an area too disturbed to produce any useful data by magnetic survey (Fig. 7, gpr 63, 64). Although speculative, some comparison of these can be made with the size, spacing, and orientation of the Mesolithic pits excavated in the former visitor centre car park to the north (Cleal et al. Reference Cleal, Walker and Montague1995, 42–7), but as with other anomalies greater certainty will only come with direct intervention.

Amesbury 10a and the pre-barrow features beneath Amesbury 6–9

Elsewhere within the ‘triangle’ there are other features that potentially pre-date the Stonehenge enclosure and/or the stone settings, or are contemporary with them. One of these is a mound (Amesbury 10a – parish numbers were first allocated by Goddard (1913) and added to by Grinsell (Reference Grinsell1957)) identified by Sir Richard Colt Hoare and depicted on Philip Crocker’s plan (Hoare Reference Hoare1812, 128). Subsequently catalogued as a long barrow, it remains visible as a slightly oval earthwork (Fig. 8), 24×20 m and just 0.2 m high, located c. 250 m to the west of Stonehenge. No side ditches are visible and none was recorded by geophysical survey in 1993–4 (Payne Reference Payne1995, 498) or from the air by the National Mapping Programme. Hoare cut a trench into it but his intervention resulted in ‘no discovery’, meaning that a burial deposit was not discovered. Such an outcome would not be unusual for a long barrow but, in this case, the existing form of the mound appears as a simple ovoid. It was not depicted alongside the other barrows of this group on the First or Second Editions of the OS 25 inch map published in 1877 and 1901 respectively, and Goddard (1913–Reference Goddard14, 165) simply catalogued it as a ‘small long barrow opened by Hoare’ rather than recording contemporary field observations. Soon after, Maud Cunnington (Reference Cunnington1914, 408) noted that it was ‘practically ploughed out’, which is not surprising given the extensive cultivation in the area (above). The nature of this feature can now only be resolved by excavation.

Fig. 8 Earthwork survey of barrows in the south-west of the ‘triangle’ showing Amesbury 4–10a. ©English Heritage. Height Data: Licensed to English Heritage for PGA, through Next Perspectives™.

To the south-west of the Stonehenge enclosure lies a group of seven further barrows (Fig. 8: Amesbury 4–10) of which little is known beyond some ambiguous antiquarian accounts (Stukeley Reference Stukeley1740, 45; Hoare Reference Hoare1812, 127–8; and see Field & Pearson Reference Field and Pearson2011). Like Amesbury 10a, these have suffered from agricultural activity but also the dumping of rubbish from the aerodrome in late 1917–early 1918 which may have obscured archaeological detail (Memo, Charles Peers to Lionel Earle, 22 January 1918 – The National Archives (TNA) WORK 14/214). Despite this, they still survive as earthworks (the mounds themselves are likely to be of Early Bronze Age date and are reported below), but geophysical survey in 1993 (Payne Reference Payne1995, 497) revealed that several masked earlier activity. Recent enhancement of the fieldwork data, using higher density caesium magnetometer survey, together with GPR (Linford et al. Reference Linford, Linford and Payne2012), has provided greater clarity and extra detail of these earlier phases (Fig. 9).

Beneath Amesbury 6

Within the peripheral circular ditch of the mound of Amesbury 6 lies an oval ditch which, being sealed by the later mound, is indicative of an earlier phase of construction activity (Figs 9 & 10). This inner ditch appears to be formed of conjoined pits with causeways or entrances in the north-west and south-east and given these gaps it is likely to indicate the former presence of a henge-like monument. Disturbance to the north-east ditch may be the result of Hoare’s (Reference Hoare1812, 128) trenching. The space between the two ditches may mark the position of an external bank which could actually be sealed within the surviving mound matrix. Comparison might be made with the similar response revealed recently during the survey of Amesbury 50 (Gaffney et al. Reference Gaffney, Gaffney, Neubauer, Baldwin, Chapman, Garwood, Moulden, Sparrow, Bates, Löcker, Hinterleitner, Trinks, Nau, Zitz, Floery, Verhoeven and Doneus2012; and see Amadio & Bishop Reference Amadio and Bishop2010, 8–9, 22–4; Linford et al. Reference Linford, Linford and Payne2012; Darvill et al. Reference Darvill, Lueth, Rassman, Fischer and Winkelmann2013), but more particularly with the pit setting at Site 1 at Dorchester on Thames (Atkinson et al. Reference Atkinson, Piggott and Sanders1951, fig 4; Whittle et al. Reference Whittle1992), although there an entrance gap was only present in the north-west. There, however, a linked series of pits appears to have formed an encircling ditch associated with Mortlake Ware. In certain respects comparison can also be made with the excavated oval pit circuit at Monkton-Up-Wimborne in Dorset where more widely spaced pits enclosed an area c. 35 m across with a wide entrance gap at either end (Green Reference Green2000, 77–8: French et al. Reference French, Lewis, Allen, Green, Scaife and Gardiner2007, 115). Finds within the upper filling of those pits included a kite-shaped arrowhead and Peterborough Ware potsherds indicating a Middle Neolithic date for its later phase of use and it is worth recalling the surface finds of Peterborough Ware along with chisel and oblique arrowheads found beyond the ‘triangle’ to the west during fieldwork during the 1980s (Richards Reference Richards1990, 35, figs 26 & 158) and chisel and oblique arrowheads found there in a recent intervention (Parker Pearson et al. Reference Parker Pearson, Allen, Bayer, Casswell, Chan, French, Garwood, Nunn, Pitts, Pollard, Pullen, Richards, Richards, Robinson, Rylett, Shaw, Teather and Thomas2008, 109, 114). Whether the Amesbury 6 configuration originated during the same period cannot be ascertained by non-invasive survey but, like Amesbury 7, this site may well pre-date – or be contemporary with – the Stonehenge enclosure earthwork.

Fig. 9 Caesium magnetometer survey, 2010, over Amesbury 5–10 superimposed over the previous Fluxgate gradiometer survey of the Triangle, 1994. Inset: Amesbury 5–9. Caesium magnetometer plot (left) and GPR data (right) from between 6.0 ns and 7.2 ns (0.3 – 0.36 m). ©English Heritage

Fig. 10 Amesbury 6: left – detail from the 2010 caesium magnetometer plot; right – combined new earthwork survey and interpretation of geophysical anomalies. ©English Heritage.

Beneath Amesbury 9

The new caesium magnetometer survey has revealed additional detail beneath Amesbury 9 (Figs 9 & 11), clarifying some anomalies that were only hinted at in the previous fluxgate data. The site now appears to be defined by an irregular and segmented ovoid ditch, sharply angled in the south. The southern arc of this is broadly continuous, although highly variable in anomaly width (<1–2.5 m) and magnetic character with a distinct corner at its southernmost point leaving the south-east facing section almost straight. However, the northern arc is composed of four discrete pit-type responses with each anomaly ~2.5 m in diameter. Three of these pits form a line, separated from each other by a distance similar to their diameter. The fourth is positioned to the east and has the weakest peak field strength. Between these is a discontinuous and indistinct narrow anomaly less than a metre wide, which might suggest the existence of a conjoining ditch or palisade.

Fig. 11 Amesbury 9: left – detail from the 2010 caesium magnetometer plot; right – combined new earthwork and interpretation of geophysical anomalies. ©English Heritage.

Enclosed by this outer circuit is a second oval of 22–24 post-pit anomalies, visible in both the magnetic and GPR datasets, each between 0.5 m and 1 m in diameter, separated by gaps of 1–2 m. It is difficult to determine whether gaps exist in the north-west and south-east due to lack of clarity amid a complex of features. To the south-west there is evidence of a third concentric alignment comprising at least four post-pit anomalies, each ~0.5 m in diameter, although there is no sign of others to suggest that they were part of a more complete circuit. The GPR data provides a similar level of detail and largely confirms the series of small, post-pit anomalies immediately inside the ditch circuit. The limited vertical extent of these anomalies may well be indicative of a response from a thin, compacted basal layer, rather than the full profile of the pit. Compared to the magnetic data the more pronounced post-pit anomalies to the north-east generally correspond to low amplitude GPR responses, and the more subtle response to the south-west with mainly high amplitude reflectors. A scatter of high and low amplitude GPR anomalies within the central area of the site are difficult to interpret and may result from Early Bronze Age use, or possibly, be due to more recent antiquarian intervention.

Assuming the ditch and pits to be broadly contemporary it seems unlikely that ditch spoil was placed within the enclosure; taken together the new survey evidence suggests that, as Payne (Reference Payne1995, 498) inferred, Amesbury 9 might have been a henge-like monument not unlike the smaller of the Wyke Down henges. The similarity to site 1 at Dorchester on Thames (Atkinson et al. Reference Atkinson, Piggott and Sanders1951, fig 4: Whittle et al. Reference Whittle1992) is striking. The latter is likely to represent several phases with the pits containing Abingdon Ware, while the ditch is associated with Mortlake Ware. While its precise chronological position and phasing will only be established by excavation, the proximity of Amesbury 9 to Stonehenge means that the development of ceremonial activities on Stonehenge Down can now be considered in a fresh light.

None of these structures has been dated independently and of course some caution is required as it is evident that there has been remodelling and re-use up to the point when they were mounded over. Nevertheless, the parallels here appear to be with monuments that comprise henge-like affinities and/or fall neatly within Kinnes’s (Reference Kinnes1979) Neolithic enclosed cemetery/ring-ditch continuum. In this respect it is worth recalling that the pit/segmentally cut Stonehenge enclosure ditch was also assigned to this category, although more recently considered to fall into a class of ‘formative’ henges (Harding Reference Harding2003, 13; Burrow Reference Burrow2010). It is clear that far from being isolated, Stonehenge may have been a component part in a ceremonial and funerary area within its immediate environs, with origins and traditions potentially traceable to the earlier Neolithic. Initially at least the focus of activities may have been around one or more of these other monuments.

North and South ‘Barrows’

Ever since Richard Atkinson recorded the presence of bluestone chips in its bank, and Thom and Atkinson cut a small trench to investigate the expected position of the stone-hole for Station Stone 94 (Cleal et al. Reference Cleal, Walker and Montague1995, 273, 276), ‘North Barrow’ has been considered to be a relatively late feature in the Stonehenge sequence. Cleal and colleagues assigned it to Phase 2 on account of similarities with the ditch surrounding the Heelstone and to the evidence from the ‘South Barrow’. Careful study of the earthworks, however, reveals that the ‘North Barrow’ does not overlie the enclosure bank to complete a circuit (Fig. 12), and the early aerial photographs taken by Lt Sharpe from a balloon in 1906 (Fig. 13) strengthen the observation that the ‘North Barrow’ earthworks probably pre-existed the enclosure earthworks, for the enclosure bank can clearly be seen as continuous and unbroken. Whether this sequence can be proven is now difficult to ascertain from the disturbed surface evidence alone, for a succession of trackways and pathways have cut through and destroyed the greater part of the interface. Comparison of Sharp’s photograph with a second, vertical, example taken on the same date (Fig. 14) is interesting. This appears to show a low bank with an internal ditch, the latter distinctly circular and visible clearly as an arc of more prominent grass growth among the more generally parched vegetation. The earth resistance survey also indicates that the ditch is interrupted by the enclosure bank (Fig. 15), although the GPR (Fig. 7) is unclear about the relationship. In addition, the configuration of the outer scarp of the enclosure bank unusually – and in contrast to the rest of the circular enclosure – incorporates two straight lengths of bank (Fig. 12) which may have been influenced by the presence of an earlier circular structure. Earlier accounts support this observation. Philip Crocker’s survey carried out for Sir Richard Hoare also depicts the earthworks in this way (Hoare Reference Hoare1812, facing p. 143) and Hoare commented that the ‘North Barrow’ was probably constructed before the enclosure ditch was dug (Wiltshire Museum Cunnington MSS: Hoare Reference Hoare1812, 145). Similarly, William Long concluded that the Stonehenge enclosure ditch ‘cuts through the ring of a low barrow on the northwest side’ (Long Reference Long1876, 54). Considering these observations, it might be tentatively suggested that the ‘North Barrow’ is an earlier feature.

Fig. 12 Stonehenge: the new earthwork plan showing the remaining portion of the ‘north barrow’ bank and ditch which does not cross the enclosure bank. ©English Heritage.

Fig. 13 Stonehenge: oblique view by 2nd Lt P.H. Sharpe RE, taken from a Royal Engineers reconnaissance balloon in the summer of 1906. From Archaeologia 1907. Reproduced by permission of the Society of Antiquities of London.

Fig. 14 Stonehenge: original 1906 print of a vertical view by 2nd Lt P.H. Sharpe RE, one of five aerial views taken by him in the summer of 1906. Reproduced by permission of the Society of Antiquities of London.

Fig. 15 Stonehenge: earth resistance plot, 1994. ©English Heritage.

As a discrete field monument, the remains of the ‘North Barrow’ earthwork today comprise a ditch within an encircling bank, with an apparently level interior, and it might therefore fall into the category of small ring-ditch. However, it is noteworthy that the earliest accounts of both ‘North and ‘South Barrows’ describe them rather differently. Stukeley recorded the existence of ‘two remarkable cavities’. His illustration (1740, 16, table 9) shows the southern cavity drawn with vertical hatching surrounded by a ditch and evidently overlain or curtailed by the enclosure bank. Stukeley’s contemporary, John Wood, who surveyed the site in 1747 and published his field drawing of the stones along with survey measurements, left his station rods in position so that others could check them. His observations therefore might be considered with some seriousness. Like Stukeley, he commented on the circles (Wood Reference Wood1747, 44, 50), describing them as cavities rather than barrows, ‘one with a single Bank of Earth about it; and the other with a double bank separated by a ditch’. They were ‘compleat Works of themselves’, the depression measuring nearly 5 m in diameter and the feature as a whole about 12 m across, ‘each Pit is about Sixteen Feet [c. 4.9 m] diameter, and the banks that surround them increase the Diameter of each Work to about forty Feet [c. 12.2 m]’. The northern cavity was ‘further enlarged by a Ditch and a Second bank of earth’. Almost 25 years later, John Smith’s description (Reference Smith1771, 52) supports these accounts, referring to the features as two ‘circular holes encompassed with the earth that was thrown out of them; but they are now almost effaced by time’. If we accept these descriptions it implies that these earthworks have seen considerable modification at some later date.

It was Hoare (Reference Hoare1812, 144) who first described them as barrows, following Cunnington’s description of the northern example as a ‘small Druid [i.e. disc] Barrow’ (Wiltshire Museum Cunnington MSS Book 1, 13 letter to Hoare 18 December 1806) and noting that they were ‘only slightly elevated above the surface’. The attribution was commented on by Hawley (Reference Hawley1928, 174) who referred to the southern example as a ‘so called barrow’ (1923, fig. 1) and demonstrated that it was anything but; Cleal et al. (Reference Cleal, Walker and Montague1995, 26) concurred that they are not burial mounds.

The evidence from excavation brings further complexity and, to a degree, contrasts with that of field survey. Cunnington encountered a deposit of burnt bones in the northern ‘barrow’ and it has since been presumed that he may have discovered deposits associated with Aubrey Hole 46 which, if the spacing of others is projected, might be expected to lie on or beneath the central platform of the feature (Newall Reference Newall1929, 82; Cleal et al. Reference Cleal, Walker and Montague1995, 96). Atkinson cut a small trench across the ditch and part of the bank at a point where it almost met the enclosure bank but unfortunately did not extend it to test the relationship between the ‘barrow’ and the bank, nor that of the predicted Aubrey Hole 47. The ‘barrow’ ditch proved to have a V-shaped profile and among items recovered from secondary silts were sherds of Iron Age and Romano-British pottery, while bluestone fragments were found in both the bank and ditch (Cleal et al. Reference Cleal, Walker and Montague1995, 276–80). A small trench partially excavated into the north-west quadrant of the central platform by Thom and Atkinson in 1978 (ibid., 273) demonstrated that the stone-hole encountered (Station Stone 94) was not central to the earthwork and was sealed by a considerable covering (0.6 m) of chalk lumps and humus enriched soil. That stone-hole is not the same thing as the 5 m diameter depression noted by Wood and others, but it might be presumed that herein lies the solution. However, scrutiny of the trench section drawing makes things clearer, for it would appear that the hollow had been filled and levelled with over 0.5 m of soil sometime between the observation of Smith in Reference Smith1771 and descriptions by Hoare in 1810. The presence of three projected Aubrey holes, two in the ditch and one on the platform introduces additional complexity, particularly if they originally held stones or posts, which would then have needed to be dismantled if the ditch were cut at a later date.

The South ‘Barrow’ is rather different and the field survey supports the observation of Hawley that it is later than the enclosure bank, although it must be stressed that surface observation here is of Hawley’s reconstructed feature and it is by no means clear to what extent the ditch now visible on the surface represents an original earthwork, so this interpretation has to be treated with caution. Lt Sharpe’s 1906 aerial photographs demonstrate that prior to excavation the proud profiled earthwork was not visible (Figs 13 & 14). The surface evidence is currently of a surrounding ditch that forms a D-shaped plan and contrasts with the profile of the North ‘Barrow’; it also differs in that there is no external bank. On excavation Hawley (Reference Hawley1922, 48) depicted the ditch as V-shaped in profile (but see his pl. viii, figs 1 & 2 which may indicate a recut). The level floor encountered consisted of a ‘yellow substance resembling chalk beaten fine and mixed with clay …’ up to 0.15 m thick and parts of the surrounding ditch was ‘covered with the same kind of compo as the floor’. Hawley considered that the feature was more likely to be a hut than a barrow. While the purpose and date of the ‘South Barrow’ remains difficult to determine, the recent discovery of Neolithic houses at Durrington Walls henge which had floors not dissimilar to those recorded by Hawley, has allowed Parker Pearson to re-appraise the evidence and conclude that this may indeed have been a building platform (Parker Pearson et al. Reference Parker Pearson, Cleal, Marshall, Needham, Pollard, Richards, Ruggles, Sheridan, Thomas, Tilley, Welham, Chamberlain, Chenery, Evans, Knusel, Linford, Martin, Montgomery, Payne and Richards2007; Reference Parker Pearson, Chamberlain, Jay, Marshall, Pollard, Richards, Tilley and Welham2009, 33–4; 2012, 310). If the ‘South Barrow’ was a building platform it would make sense of Hawley’s description of it cutting into the bank.

Scrutiny of the trench profile (Cleal et al. Reference Cleal, Walker and Montague1995, fig. 165) reveals that, like the ‘North Barrow’, this was covered by at least 0.5 m of soil and, consequently, the hollow described by antiquarians appears to have been filled and levelled at some point. Hawley did encounter a hollow although the size of it does not match the descriptions of the earlier observers. Initially he considered it to be the result of an earlier excavation (Hawley Reference Hawley1922, 48) – presumably that of Cunnington who, according to Hoare (Reference Hoare1812), found nothing – and later considered that it was a stone-hole (i.e. Station Stone 92; Hawley Reference Hawley1923, 15). Unfortunately the relationship of the stone-hole to the platform remains unclear, but there is an assumption that the stone-hole must have been dug through the platform otherwise it would not have been visible to the early observers. Similarly the floor material that lay over the ditch is curious, for it implies that the cutting of the ditch preceded not only the floor but potentially the stone-hole.

There is some additional complexity adjacent to the ‘South Barrow’ suggested by the GPR, comprising the presence of a slightly arced linear response extending to the north and an amorphous, sub-circular anomaly immediately to the west (Fig. 7). The previous earth resistance data show a diffuse area of increased background response that correlates with the GPR anomaly to the west and a discrete high-resistance anomaly immediately to the north, although the latter was subsequently tested and found to be an area damaged by rabbit burrowing (Payne Reference Payne1995).

In more recent times interpretation of these features has suggested that they are the product of excavations for two of the ‘the four stations’ – i.e. Station Stones 91–4. E. Herbert Stone considered that the ‘two mounds are sites which were at one time occupied by two stones which, with the pair remaining, formed a symmetrical group of four … the earth forming the mounds … would simply represent the spoil thrown out by excavations when the stones were removed’ (Reference Stone1924, 114, 116). It is evident, however, that in each case there are a considerable number of problems and several phases of activity to account for.

THE STONEHENGE ENCLOSURE

The Stonehenge enclosure survives as a well-defined earthwork comprising two banks separated by a ditch that describes a circle 102 m in diameter when measured between the internal lips. Much of the outer bank has been reduced by cultivation but was evidently largely extant in the 18th century as recorded by Wood (Reference Wood1747, 79). It is well preserved in the north where the proximity of the road may have discouraged tilling. Here it is of significant proportions and need not be envisaged as a counterscarp (Hawley Reference Hawley1923, 14; 1928, 52) despite the prevalence of this terminology (Cleal et al. Reference Cleal, Walker and Montague1995, 24). Geophysical survey in 1994 and again recently, emphasised that the width of the outer matched that of the internal bank and it was clearly once of substance (Linford et al. Reference Linford, Linford and Payne2012; Payne Reference Payne1995, 501). Its presence has certain implications in terms of the interpretation of the construction sequence, for notwithstanding the comments above, it is conceivable that the enclosure may have been defined initially by the outer bank, i.e. the standard ‘henge’ configuration, the inner bank being added later. Given the form of contemporary comparable monuments (Burrow Reference Burrow2010, fig. 11.2) that interpretation need not be pursued unduly, although it is eminently feasible, or even likely, that two or more phases of activity are represented by the enclosure earthworks and the recently recognised recut visible in Hawley’s section drawings of the ditch (Parker Pearson et al. Reference Parker Pearson, Chamberlain, Jay, Marshall, Pollard, Richards, Tilley and Welham2009, 29–30, pl. 18, fig. 2; 1923) serves to emphasise the point.

The form of the enclosure provides some evidence for this, as its character differs from west to east (Fig. 12). Both banks and the ditch are of wider proportions in the west than in the east. The western arc is relatively broad and flat-bottomed compared to that in the east (where it has, admittedly, been excavated leaving a different surface appearance). Similarly, the inner bank in the west is more substantial than that in the east which can be clearly seen in the digital terrain model (Fig. 16). This difference is also visible on the resistivity plot (Fig. 15). The changing character of the ditch was noted by Hawley who commented on the pit-like nature of the excavated chalk in the south, compared to the great segment that he had excavated in the east. Apart from a small trench excavated by Atkinson at a point where the trackway breaks through (Cleal et al. Reference Cleal, Walker and Montague1995), the western part of the ditch has not been excavated and it is unclear to what extent deposits and the presence of cultural material, particularly cremations, might differ from those in the east.

Fig. 16 Stonehenge: digital terrain model of earthworks set within a lidar background (outer area). ©English Heritage. (Lidar data ©Environment Agency December 2001).

While the eastern half of the enclosure has undergone excavation it is not this that is responsible for the differences in scale. It is, however, possible that it might have resulted from an episode of variable ditch re-cutting that overcut the ditch sides in the west thus producing a slightly asymmetrical plan. The GPR survey depicts a complex response over the unexcavated western circuit of the ditch and bank (Fig. 7), perhaps indicating a degree of internal structure that is not readily observed on the surface; this may, in part, represent sectional or phased construction of the bank. A series of pit-like features are also visible, which, given the cremations recovered during the excavation of the eastern half of the enclosure, could indicate that the earthwork was used in a similar fashion here. It may also be that there were cosmological considerations that influenced the differential use of the western and eastern perimeters of the enclosure (Pollard & Ruggles Reference Pollard and Ruggles2001, 83).

Undulations on the present ditch base indicate that further causeways or segments exist in the unexcavated portion of the ditch similar to those recorded in the east (Fig. 12). In two cases these undulations lie opposite breaks in the inner bank which could result from traffic erosion, although there is no evidence of such damage continuing either side of the enclosure perimeter. The causewayed or pitted nature of the excavated ditch has led to a suggestion that the site falls within the causewayed enclosure tradition, albeit late in the sequence (Cleal et al. Reference Cleal, Walker and Montague1995, 113). The ditch and bank is now estimated to have been constructed around 2990–2755 cal bc (95% probability) and probably 2955–2830 cal bc (68% probability: Marshall et al. Reference Marshall, Darvill, Parker Pearson and Wainwright2012: Darvill et al. Reference Darvill, Marshall, Parker Pearson and Wainwright2012) and the site has been compared with the similarly sized circular enclosure at Flagstones, Dorchester. Enclosures of this nature have recently been termed ‘Formative henges’ (Harding Reference Harding2003) and viewed as part of the continuum alongside the early henges at Llandegai in north Wales, and more loosely with the Priddy Circles (Burrow Reference Burrow2010). Little is known of the Great Circle at Stanton Drew which, at 136 m diameter, is rather larger, yet its variation in ditch width coupled with the complexity of concentric internal settings and an Avenue – this time in stone – provides one of the closest parallels to Stonehenge (David et al. Reference David, Cole, Horsley, Linford and Martin2004).

Additionally, it is worth noting the two straight sections in the north-west part of the enclosure circuit which are particularly noticeable in the alignment of the outer bank. As observed above, this may be a response to the presence of a pre-existing feature at that point, but it is also worth bearing in mind that the admittedly larger enclosure at Avebury is similarly angular on one side, in that case the west (McOmish et al. Reference McOmish, Riley, Field and Lewis2005, 21). A further pre-enclosure feature might have influenced the sinuous indentation in the east, 15 m south of Station Stone 91, and in this context it is worth noting the proximity of the oval vegetation mark alongside this point on Lt Sharpe’s aerial photograph (Fig. 14) that, if not a fungus ring, might also point to an anomaly here.

Gaps in the west and north-west cut through the enclosure earthworks although as noted by others only those breaks in the north-east and south can be considered to be prehistoric (Cleal et al. Reference Cleal, Walker and Montague1995, 109–11). At 11 m between ditch terminals, the break in the north-east is a wide entrance gap, much wider than is necessary for access of single individuals or animals and may have been built that wide to allow several individuals to pass through it at once, or to focus upon a view. For a considerable period, however, any such use would have been obstructed by the various arrangements of posts or stones that were suggested by Hawley’s (Reference Hawley1924) excavations. The subtleties of ephemeral surface undulations in this area are rather too amorphous to interpret easily, but a slight rise in ground level suggests that the outer bank may once have continued across the entrance gap. This shallow camber is visible in Hawley’s photographs (Hawley Reference Hawley1924, figs 1 & 2); he made no comment upon it and instead referred to the manner in which this area had been eroded by tracks. It is worth noting that the Aubrey holes similarly continue across the north-eastern entrance. Unfortunately, neither the digital terrain model nor the laser scan help with this issue and the ambiguity of the depiction must result from effects of excavation trenching and heavy traffic (Hawley Reference Hawley1925, 23).

The present survey indicates that each of the Avenue ditches terminates about 3 m from the enclosure ditch and Hawley’s excavation results leave little room to doubt this. In effect they butt against the outer bank of the enclosure. There is thus no evidence that the Avenue banks and ditches continue into the interior of the enclosure and the mismatch of enclosure entrance and the Avenue at their junction, i.e. the width of the enclosure entrance is narrower than that of the Avenue, has been noted before. Addressing the question of the structural mismatch, Atkinson considered that several metres of ditch at the south-east terminal had been backfilled to match the width of the Avenue in order to allow unhindered passage into the enclosure. He suggested that the eastern part of the enclosure ditch excavated by Hawley ‘appears to have been filled up deliberately, when it was already partly silted, in order to enlarge the entrance to match the width of the Avenue’ (Atkinson Reference Atkinson1979, 70). The interpretation appears to assume that the ditch at this point had been filled level with the surface so that it was possible to pass across unimpeded and in this he was perhaps influenced by R. H. Cunnington’s (Reference Cunnington1935, 86) earlier account of the site. Hawley certainly referred to a deposit of clean white chalk ‘which had been brought from elsewhere and cast into the ditch’ but he did not appear to think that the ditch was backfilled to the top such that it would allow unimpeded access from the Avenue. Hawley backfilled his excavations according to a plan agreed with Charles Peers. Less fill was put back in than had been taken out, in order to give the ditch a deeper appearance. That this is so can clearly be seen on the earthwork plan where a 12 m length is depicted as deeper than the rest. Spare spoil was evidently not added to the bank but mainly used to level the interior, and especially the former tracks crossing it (Letters from Col Hawley to Charles Peers, 8 April 1920 & 29 August 1921, TNA WORK 14/2463). Hawley made no mention of whether the bank at this point was missing or denuded, for Atkinson’s suggestion demands that it would need to have been cast back into the ditch in antiquity. That cannot have been the case, however, for it survives perfectly well as an earthwork, nor has the entrance been modified in order to allow passage and consequently both ditch and bank still presented an obstacle. Lt Sharpe’s 1906 aerial photographs (Figs 13 & 14) indicate quite clearly that the ditch was of this form, i.e. of an earthwork that impinged on the line of the Avenue, prior to Hawley’s excavations. In the 1906 oblique aerial photographs the dark line of the ditch and the lighter traces of the internal bank can be seen continuing across the line of the Avenue for several metres before terminating. Additionally, both bank and ditch are visible on a vertical photograph taken in 1922 before Hawley had begun excavation of the terminal (English Heritage Archives SU1242/14 CCC8651/73). It is noteworthy that Cleal et al. (Reference Cleal, Walker and Montague1995, 139) refuted Atkinson’s interpretation which relied heavily on the presence of bluestone in backfilled layers and they pointed out that this material was in fact derived from disturbance by a later burial. The present survey thus supports those views, and it can be concluded that the enclosure ditch entrance terminal earthwork was not originally backfilled to match the Avenue.

The simplest explanation for the mismatch is of two separate and self-contained construction episodes: enclosure construction followed by the Avenue. Given the prolonged chronological gap between the date of the enclosure and that of Avenue construction and the potentially complete change in purpose over time, there is no need for symmetry, as by the second phase of use the enclosure may have been a silted, grass covered and ancient earthwork. Cleal et al. (Reference Cleal, Walker and Montague1995, 170 fig. 70; see also Newall Reference Newall1929, 83–4, 88) indicate that there was a change in orientation between these phases of activity. Phases 1 and 2 of the monument were based on an axis that bisected the original entrance. In contrast, in Phase 3, which incorporated the Avenue construction, the axis bisected the stone settings 1 and 30. Both axes pass to the west of the Heelstone. As Cleal et al. (Reference Cleal, Walker and Montague1995, 140) point out, the two components are not conceptually related and are not reliant on each other to form part of a coherent design. If the full width of the Avenue was planned with procession or access to the stone settings as its main purpose it might indeed be expected to have matched the width of the enclosure entrance or at the very least approach it in a symmetrical manner. However, if the alignment was a solar one and its objective related to viewing or some other ritual or spiritual purpose there is no need for it to ‘fit’. It is worth bearing in mind that this is a circumstance matched at both the Great Circle at Stanton Drew and Avebury. In the former case, the stone avenue aligned east-north-east is at considerable variance with the gap in the enclosure ditch (David et al. Reference David, Cole, Horsley, Linford and Martin2004); in the latter the West Kennet Avenue leads away from the southern entrance before an uncertain realignment, while the Beckhampton Avenue is offset from the western entrance in a not dissimilar manner to that at Stonehenge (Gillings et al. Reference Gillings, Pollard and Wheatley2008, 122–3).

The slight angle change in the southernmost Avenue ditch towards the fence alongside the A344 road detected by surface survey can also be observed on the aerial photograph taken in 1906 (Fig. 13). It can be interpreted in two ways. As noted above, it may result from a need to respect the ditch around the Heelstone, which would imply that the Avenue is later than the Heelstone ditch. Alternatively, if construction of the Avenue ascended the slope from Stonehenge Bottom, perhaps aligned on some visible marker, the Avenue ditch may have required a minor re-alignment on to the enclosure entrance. In the latter case it could have been better executed, and it can be noted that the other side of the Avenue is fairly straight. It is curious that the realigned Avenue bank is also the one that fails to meet the appropriate enclosure terminal. GPR survey along a section of the A344 revealed a group of tentative, diffuse anomalies between 0.32 m and 0.80 m from the current surface in a location where the Avenue is known to cross beneath the road. The result of excavations here will be reported separately (Wessex Archaeology in prep.).

A glance at the GPR plot (Fig. 7) reveals that a multitude of currently unexplained anomalies exist within the enclosure. Aubrey Holes are most apparent in the near surface data to the south-west of gpr 23. Of particular interest is an arc of ditch with a diameter of c. 5 m set close to the inner lip of the bank in the west (gpr 41). Within it pits or posts appear to be arranged and the feature partially encloses an incomplete square or right-angled structure. This need not, of course, be of prehistoric date and it is possible that this relates to an obscure linear feature, perhaps a path or track outside the ditch, shown on the 1906 aerial photograph of the site (Capper Reference Capper1907). Like many of the GPR anomalies it needs testing. A further, complex sub-rectangular anomaly (gpr 43) with dimensions of 7×5 m found between 0.6 m and 1.08 m deep immediately to the south of Station Stone 93 is also of interest. Again, this anomaly does not appear to correlate with any known intervention or produce any surface expression recorded by either the earthwork survey or the historic aerial photography. It can be seen as forming a linear alignment with gpr 44, 45, and 46. There is no evidence for a trackway here, although it should be noted that there is hollowing in the enclosure bank at either end. Just to the west of the stone settings a large, deep and amorphous anomaly, gpr 60, with a diffuse response remains unexplained, though its position relevant to the stones is interesting. It is too large for a stone-hole, but could invite comparisons with the enormous hole encountered by Gowland to the north of Stone 56 (Parker Pearson et al. Reference Parker Pearson, Cleal, Marshall, Needham, Pollard, Richards, Ruggles, Sheridan, Thomas, Tilley, Welham, Chamberlain, Chenery, Evans, Knusel, Linford, Martin, Montgomery, Payne and Richards2007, 624–6; 2012, 131–2; Field & Pearson Reference Field and Pearson2010, 67; Darvill et al. Reference Darvill, Marshall, Parker Pearson and Wainwright2012, 1025).

THE BARROWS

Significant additions to the former ceremonial sites situated to the west of the stones were evidently made during the Early Bronze Age. Burials were added in some, though as far as it is possible to tell from the antiquarian literature, not all cases; mounds were added, while several other mounds were built on fresh sites. Despite the cultivation reported by Long (Reference Long1876) and reports that they had been levelled, all of these survive as earthworks (Figs 3, 8, & 17). Although unusual in respect of the number of pre-barrow features (as discussed above), when compared to others within the World Heritage Site, there appears to be little outstanding about these mounds in terms of size or complexity and, aside perhaps from Amesbury 11, they do not appear to enhance the monumentality of Stonehenge in any way. Some of this may of course be a result of the cultivation that took place during the 18th–19th centuries. Assessing the impact that the military presence had on these mounds is difficult to evaluate (Barber Reference Barber2014b) and the boundary between the Stonehenge Aerodrome and the monuments seems to have been ill defined. As early as 22 January 1918, Charles Peers, Chief Inspector of Ancient Monuments, was complaining about their treatment: ‘Immediately to the West of the Stones is a group of barrows, several of which have lately been covered over with rubbish dumped on them …The young officer in charge of the Aerodrome had no idea what a barrow was, when my informant asked him why a position where barrows were so numerous had been chosen’ (TNA WORK 14/214). The informant was Hawley, and a sketch plan produced by Charles Peers on the basis of Hawley’s information appears to show the dumping to have been concentrated in the area of barrows Amesbury 5, 7, and 10.

Fig. 17 Stonehenge ‘triangle’: model of the relative surface gradients, derived from the laser scan of the ground surface and shown against a lidar background. ©English Heritage. Lidar © Environment Agency (December 2001).

THE Y AND Z HOLES

In contrast to the Aubrey Holes, the Y and Z Holes survive as visible surface earthworks within the Stonehenge enclosure (see below). Given that they were first discovered following the removal of the topsoil by Hawley (Reference Hawley1925, 27–8), it might be thought that their appearance as earthworks is the result of slumped backfilling. However, it is clear that they were present in this form as they can be observed on Lt Sharpe’s aerial photographs taken in 1906 (Capper Reference Capper1907) and are particularly clear on a RAF vertical photograph (English Heritage Archives CCC 8561/173 SU1424/14) taken in February 1922 before his excavations. In any case, further depressions were observed corresponding to Y Holes 25–28 where no excavation had taken place. The neat circuits around the sarsens depicted on earlier plans are not confirmed and instead there is some irregularity, suggesting a slightly angular rather than perfectly circular arrangement (see also the GPR response in Fig. 7), although damage in the north-west has obscured crucial detail and it is impossible to pursue this point. Oval rather than circular (excavated examples tended to be sub-rectangular), the Y Holes form a complete circuit except for that portion in the west where the earthworks have been obliterated by the visitor pathway. An additional depression was noted in the south-east as part of the circuit but it lies a little out of alignment with the others. The geophysical surveys produced similar results. The GPR data response varies from very near surface reflections due to the shallow depressions marking the location of these features, to deeper lying pit-type anomalies, a variation that may be linked to the excavation history of the individual holes around the circuit.

A similar circuit of near-surface GPR anomalies corresponds with the location of the excavated Z Holes. A trench-like disturbance recorded by the earthwork survey between the prone sarsens 9 and 12 obscures the identification of any Z Holes in proximity to Stone 14. This is the probable area of Charles Darwin’s work but it also corresponds with the presumed location of the seismic array that was inserted in the 1960s as a security system to detect nocturnal intruders (Linford et al. Reference Linford, Linford and Payne2012, 6). Some additional pit-type anomalies are found on the presumed Z Hole circuit and these appear to correlate with discrete low resistance responses (Fig. 15). The Z Holes are particularly well represented as earthworks on the north-east side of the monument where, like the Y Holes at this point, many mirror the stones of the extant sarsen circle. There is a problem in the area of Z Hole 8. Atkinson did not encounter this pit during excavation and was confident that it did not exist (Atkinson Reference Atkinson1979, 34); however, two depressions were encountered on the surface a little outside the predicted circuit and beyond Atkinson’s trench. Their signature was also encountered by the geophysical team in 1994 where a second anomaly was referred to as 8a (Payne Reference Payne1995, 503).

Immediately within each circuit of depressions is an extremely low bank identified as both a surface expression and GPR response. As with the depressions this feature is not precisely circular in plan but has a series of sinuous bulges and relatively sharp angles. The nature of these banks cannot at present be determined. It is entirely feasible that they are recent or the product of activity during the historic period. A trench was dug in which floodlight cables were placed in 1968 but these avoided the Y Holes. A similar trench cut in the same year for geophone cables does, at least in part, correspond with the position of the Z Hole bank (Cleal et al. Reference Cleal, Walker and Montague1995, 12, 563, fig. 289). However, slight traces of the inner bank can be seen in Lt Sharpe’s photograph and the outer on enlargements of the 1922 air photograph mentioned above; therefore the idea that they may represent spoil from Hawley’s excavation of the Y and Z Holes can be discounted as they are clearly of some antiquity and could instead represent the original soil dug out of the Holes.

Hawley’s excavation method involved removing the topsoil straight down to the chalk bedrock, allowing him to more easily observe features cut into the chalk by their darker, humic content. Consequently, it is likely that the presence of cultural material above the chalk was lost. Hawley’s plan (Cleal et al. Reference Cleal, Walker and Montague1995, 14, fig. 8) suggests that, having initially encountered certain Y Holes by chance, he was then able to discover the remainder by probing or by recognising the earthworks and then targeted them with small trenches, each of which avoided the internal banks. The banks, therefore, remain unexplored and may be substantially intact. Other features may also have been overlooked, as a series of parch marks observed in 2013 after the survey programme had been completed, were observed to be set radially in relation to the Y and Z holes and to form a part of a circuit between them (Banton et al. Reference Banton, Bowden, Daw, Grady and Soutarforthcoming).

‘CELTIC’ FIELDS

It is noteworthy that the primary silts from Y Hole 16 are now considered to represent a windblown deposit resulting from an area of cultivated fields (Cleal et al. Reference Cleal, Walker and Montague1995, 260) and radiocarbon dating of antlers at the base of the feature indicate that this occurred around 1640–1520 cal bc at a time when ‘Celtic’ fields were being established across Salisbury Plain (McOmish et al. Reference McOmish, Field and Brown2002, 51–6). With prevailing south-westerlies the deposits may derive from a ‘Celtic’ field system plotted from aerial photographs as part of the National Mapping Programme which is present a little to the west of Stonehenge, just beyond the ‘triangle’ field boundary (Fig. 18). It is partially co-axial and may have aggregated and accumulated through several stages of activity, its orientation being changed in the process by incorporating what appears to be a trackway; the overall configuration suggests an intensity of use and the potential for settlement to be present amongst the complex. Based on the template established for such fields in the area (McOmish et al. Reference McOmish, Field and Brown2002, 18–20), the primary phase, at least, is likely to be Middle Bronze Age in date and it is noteworthy that recent excavations (Parker Pearson et al. Reference Parker Pearson, Allen, Bayer, Casswell, Chan, French, Garwood, Nunn, Pitts, Pollard, Pullen, Richards, Richards, Robinson, Rylett, Shaw, Teather and Thomas2008) have revealed the presence of a curving ditch with a distinct terminal, along with pits and post-holes associated with Middle Bronze Age material. It may also be that some of the nearby round barrows, one of which (probably Amesbury 3) produced the ‘Stonehenge Urn’ (Cunnington MSS, Wiltshire Museum; Hoare Reference Hoare1812, 126; R. H. Cunnington Reference Cunnington1975, 164), are associated with the settlement here.

Fig. 18 Stonehenge Down: archaeology plotted from aerial photographs and lidar showing the Palisade Ditch and ‘Celtic’ fields. ©English Heritage.

LINEAR DITCH

A linear feature known as the Palisade Ditch recorded by the 1993–4 geophysical survey (Payne Reference Payne1995, 497) cuts across the north-west corner of the ‘triangle’ and is visible beneath two phases of cultivation on the laser scan (Fig. 17) and as a shallow earthwork, its narrow scarp barely 10–20 cm high and facing north-west. Aerial survey also records its presence (Fig. 18) and it is clear that it cuts across the system of ‘Celtic’ fields to the west, confirming the sequence suggested by the remote sensing techniques. Recent excavations have demonstrated that this feature is of Middle Bronze Age date or slightly earlier (Parker Pearson et al. Reference Parker Pearson, Allen, Bayer, Casswell, Chan, French, Garwood, Nunn, Pitts, Pollard, Pullen, Richards, Richards, Robinson, Rylett, Shaw, Teather and Thomas2008). Its curvilinear route is unusual for a Bronze Age linear ditch, although this might be explained by its use of the course of an earlier palisade (Parker Pearson et al. Reference Parker Pearson, Allen, Bayer, Casswell, Chan, French, Garwood, Nunn, Pitts, Pollard, Pullen, Richards, Richards, Robinson, Rylett, Shaw, Teather and Thomas2008). It may be significant that a number of other linear ditches appear to focus on the higher point of Stonehenge Down (Bowden et al. Reference Bowden, Barber, Field and Soutarin press).