Alongside Stonehenge, the passage graves of the Boyne Valley and the Carnac alignments, the Avebury henge is one of the pre-eminent megalithic monuments of the European Neolithic. Its 420 m diameter earthwork encloses the world’s largest stone circle. This in turn encloses two smaller yet still vast megalithic circles - each approximately 100m in diameter - and complex internal stone settings. Avenues of paired standing stones lead from two of its four entrances, together extending for approximately 3.5km and linking with other monumental constructions. Avebury sits within the centre of a landscape rich in later Neolithic monuments, including Silbury Hill and the West Kennet palisade enclosures. Reliable dates for the hypothesised construction phases at Avebury and other monuments in its environs remain scarce We can be confident that the main Avebury earthwork was created around 2500 BC, but this seals a primary earthen bank whose precise date is uncertain; there is similar ambiguity with regard to the dating of the Southern and Northern Inner Circles and the megaliths that they enclose. Secure knowledge of the monument’s chronology is essential, as it frames our understanding of how the henge and its megalithic settings came into being - whether through incremental development or as a single notionally planned entity.
In a paper published in the journal Antiquity on 10 April 2019, Mark Gillings of the School of Archaeology & Ancient History at the University of Leicester, and Joshua Pollard and Kristian Strutt of the Faculty of Arts and Humanities at the University of Southampton present the results of a recent (2017) excavation at Avebury, and its implications for our understanding of the construction of the Southern Inner Circle at the monument.
The Avebury monument. Gillings et al. (2019).
The earliest antiquarian records for the Southern Inner Circle comprise those made by John Aubrey and Walter Charleton in 1663, and William Stukeley’s plan and written narrative compiled between 1719 and 1724. Unfortunately there are few definitive areas of agreement between them. Charleton’s schematic plan, for example, depicts the Obelisk surrounded by a perfect circle of 13 megaliths. In contrast, Aubrey’s plan offers a more confused picture of the Southern Inner Circle’s settings. Aubrey mapped a portion of the Circle’s arc, within which he recorded four large stone positions and two smaller stone symbols annotated with the letter ‘Z’. To the north-east are three further stones, and Aubrey makes no mention of the Obelisk. By the time Stukeley began recording the site 56 years later, a combination of entropy and active destruction had taken its toll. The Obelisk had fallen, and much of the complexity in layout hinted at by Aubrey was gone. The presence of a single megalith standing in a somewhat anomalous location in the context of the Southern Inner Circle stones led Stukeley to propose the existence of a second concentric inner circle. Stukeley’s drawings show a stone of substantial size, comparable in basal dimension to the main Southern Inner Circle stones.
Stukeley’s Frontispiece - the single stone that had survived to the early eighteenth century is indicated by the arrow; it was subsequently destroyed. Gillings et al. (2019).
The Southern Inner Circle visible today is the product of a programme of excavation and reconstruction carried out by Alexander Keiller in 1939. Utilising a 50 ft (15.24 m) grid of squares subdivided into 25 ft (7.62 m) quarters, Keiller’s intention was to excavate areas not covered by village houses and gardens. The outbreak of the Second World War curtailed this operation, but not before a substantial area had been excavated, including the western arc and interior. Within the circle was the site of one of Avebury’s largest stones, the Obelisk, which had been recorded and so-named by the eighteenth century antiquary William Stukeley. During excavation, Keiller discovered an unexpected 30.8m-long line of stoneholes that had formerly held megaliths to the west of the Obelisk. His excavations also unearthed a series of medieval stone burial pits (cut along the same line) that contained distinctive reddish sarsens, which were much smaller than other Avebury megaliths; the maximum dimensions of these stones ranged from 1.3–2.4 m. Labelled by Keiller as the ‘Z-feature’, the presence of stoneholes perpendicular to the ends of the line hinted that these features may once have formed a rectangular setting. Keiller’s excavations also revealed the stonehole for a megalith (stone D) that did not appear to be part of either circle or the Z-feature, and a cluster of postholes, gullies and pits to the immediate north of the Obelisk. The Z-feature remains something of an enigma, though it has been suggested that if the excavated features were duplicated in reverse on the east side of the Obelisk, this megalithic component might resemble the stone kerb of an Early Neolithic long barrow.
The Southern Inner Circle showing recovered lithic densities. Gillings et al. (2019).
Critical re-evaluation of the Keiller excavation archive indicates that the excavated stoneholes were far too large for the Z-feature stones that Keiller re-erected into them. As a baseline, the excavated stoneholes of the main Southern Inner Circle ring range from 1.7–2.5 m in maximum length, and hold stones standing 2.74–4.15 m in height. With the exception of stonehole xii, which was genuinely intended for a small stone, the Z-feature stonehole dimensions fall comfortably within this range. Thus, these stoneholes originally held much larger stones - equivalent in size to those making up the Southern Inner Circle. This explains the difficulty Keiller had in matching Z-feature stones to stoneholes, and his decision to raise these megaliths above the bases of ‘their’ stoneholes, by between 0.15 and 0.40m, when re-erecting them.
Within the Southern Inner Circle, Keiller excavated two features, both of which he labelled ‘Natural Fissure (?)’, alongside a cluster of gullies, pits and postholes to the immediate north of the Obelisk. This cluster included a series of shallow hollows (maximum 2.7 × 1.8 m), which he interpreted as medieval marl pits. Of greater significance are the parallel lengths of gulley, which define a structure approximately 6.9 m wide and 6.8 m long—although the southern extent has been affected by the destruction of the Obelisk. Running between these gullies was a line of three oval pits or postholes, with hints of a shallow slot linking the westernmost two. A fourth such pit was located on the approximate central axis to the north. While Keiller was content to assign a prehistoric date to these pits/postholes, he was confident that the gullies formed part of a much later, open-ended structure, presumably medieval in date. This, he surmised, had been opportunistically built against the fallen bulk of the Obelisk, using the latter as an ersatz rear wall. While Keiller toyed with the idea of the structure being a pigsty, his supervisor, William Young, suggested that it may have been a cart shed. By the time that the fieldwork was formally published, these features had been reduced to the status of field boundary ditches.
The features excavated and interpreted by Keiller in 1939. The ‘1865 excavations’ refer to trenches dug in 1865 by Alfred Smith and William Cunnington on behalf of the Wiltshire Archaeological & Natural History Society. Gillings et al. (2019).
The medieval date assigned to the pits and structure can be questioned, as no medieval pottery was found within the gullies, and only three sheds were recovered from one of the pits. This is surprising, given the high density of twelfth- to fourteenth-century pottery recorded in the excavation archive that was recovered from the overlying soil (up to 100 shards per 25 ft/7.62 m square). The three shards of medieval pottery from the pit are probably intrusive, as Rabbit burrows were recorded in the vicinity. The pits may even be naturally formed features (e.g. tree-throw pits) of prehistoric date. It is the gully-defined structure, however, that takes on particular significance, once Keiller’s unsupported claim for a medieval origin is rejected.
Gillings et al. present lines of evidence which they argue suggest a prehistoric, and specifically Early Neolithic, date for the structure. Firstly its axis is parallel to the excavated line of the Z-feature stoneholes, and it occupies the geometric centre of the Southern Inner Circle, which is located just north of the Obelisk. Secondly it is associated with a localised spread of Neolithic worked flint and pottery, which is otherwise rare in the interior of the monument. Finally the plan of the structure bears a remarkable resemblance to those of smaller Early Neolithic houses from Britain and Ireland.
The spread of Neolithic artefactual material includes 346 pieces of worked flint from soil contexts in the area of the Southern Inner Circle, comprising 334 flakes, nine scrapers, a knife, a retouched flake and a polished axe. In addition, there are 138 worked flints from the Z-feature stoneholes and burial pits, from features associated with the Obelisk, and from the gullies. Amongst this material are two awls, a fabricator, a knife and a bifacially retouched flake. The associated debitage includes blades, narrow flakes and several thinning flakes. Such an assemblage is consistent with an Early Neolithic domestic site. The distribution of this material is particularly striking, as the greatest concentration of worked flint is focused on the gully-defined structure, with a lower density ‘halo’ of approximately 20m radius around this. This distribution compares to the artefact spreads around the Early Neolithic buildings at Hazleton North and Ascott-under-Wychwood.
Gillings et al. feel that the most expedient interpretation is that this structure is a Neolithic house. Keiller was correct to interpret the gullies as wall trenches, although, unfortunately, descriptions of fills and sections are lacking. Three of the prehistoric pits sit within the interior, central and perpendicular to the gullies. Their small diameter probably indicates that they are postholes for an internal division. The fourth pit is located at the end of the structure in a central, gable-end position. Taken together, they form a plan that has close parallels with several small post- and trench-constructed houses of the thirty-eighth to early thirty-seventh centuries cal BC from mainland Britain and Ireland. At close to 7 × 7 m, the Avebury structure falls comfortably within the size range. Close parallels include Fengate, Cambridgeshire, Ballintaggart 1 and 3,County Down, Newrath, County Kilkenny and Horton, Berkshire. The larger structure at White Horse Stone, Kent, was constructed within clear sight of a substantial sarsen spread, much as the Avebury building would have been. This would be the first such Early Neolithic house to be identified in Wessex.
The Early Neolithic house structure in the centre of the Southern Inner Circle and comparators. Gillings et al. (2019).
To investigate further the possible connection between the house and the excavated portion of the Z-feature, more data were required. Since Keiller’s excavations, fieldwork in the Southern Inner Circle has been limited to an inconclusive geophysical survey in 1989, alongside more ad hoc mapping of parch marks. More recent surveys elsewhere at Avebury have proven the efficacy of ground-penetrating radar and soil resistance survey for the detection of buried sarsens. Given the known presence of large (between 15 and 100 tonnes), buried sarsen stones at Avebury in close association with highly compacted stoneholes, Gillings et al. consider it surprising that no previous large-scale ground-penetrating radar surveys have been attempted in the Southern Inner Circle. This is despite the success of ground-penetrating radar in detecting buried megaliths on the Beckhampton Avenue.
In April 2017 Gillings et al. surveyed a 5670 m² area to the east of the areas excavated by Keiller. Soil-resistance survey was carried out using twin-probe and square arrays, and this was complemented by ground-penetrating radar. The resistance results display several anomalies indicative of former megaliths. These take the form of discrete high resistance anomalies (marking buried stones), moderately high responses (indicating either deeply buried stones or concentrations of stone debris) and lower-resistance features (destruction pits). Some of these had been previously identified as hollows by Keiller, and as generalised anomalies in the 1989 survey; several had not been identified at all. There are also south-east- to north-west- (and perpendicularly) aligned linear features corresponding to former boundaries - some of which are clearly visible in the field as earthworks - along with probable drainage features. Although not indicated on the interpretation plot, it is interesting to note that the interior of the Southern Inner Circle seems to be characterised by higher resistance. The south-west to north-east band of low resistance crossing the top third of the plot probably reflects the complex sequence of medieval and post-medieval boundary ditches that criss-cross this area.
Results of the soil-resistance survey carried out across the Southern Inner Circle with interpretation. Gillings et al. (2019).
Several clear anomalies are visible in the time-sliced ground-penetrating radar results, from the surface to a depth of 3.1 m. Alongside linear medieval property boundaries and the general ‘noise’ adjacent to the modern gardens, 16 stone-related and three other features have been identified. One of these is adjacent to a modern boundary and manifests as a zone of high resistance, as well as an amorphous ground-penetrating radar anomaly; it probably derives from medieval and/or post-medieval building activity. Another is the edge of Keiller’s 1939 excavation trench. Other anomalies, however, correspond to elements of Neolithic monumental architecture. These include a sub-circular feature evident in the ground-penetrating radar data at a depth between 0.5 and 0.9 m below the present surface, which appears to comprise a series of discrete, small circular anomalies that are probably postholes or pits, four buried sarsens associated with the continuation of the Z-feature setting, two destruction pits or debris relating to the continuation of the Z-feature setting, five substantial, deeply buried sarsens of the Southern Inner Circle, two probable destruction pits (low resistance) and the compressed bases of megalithic stoneholes (ground-penetrating radar reflection) of the Southern Inner Circle, probable destruction pits (low resistance) and compressed stone sockets (ground-penetrating radar reflection) relating to a pair of stones that form a linear alignment with tow other anomalies, a spread of large fragments of sarsen or packing stones, resulting from the destruction of a substantial Southern Inner Circle sarsen, and a possible stone position visible in the ground-penetrating radar data (depths 0.3–0.6 m), but partially masked by debris relating to the modern boundary.
Ground-penetrating radar interpretation combining anomalies identified in the sequential depth slices. In this figure, the level of re-inscription (i.e. over-drawing) acts as a direct proxy for the persistence of the features with depth. Gillings et al. (2019).
Six of the features mirror the position of the excavated Z-feature stoneholes. Taken together, they form a 30 × 30 m square megalithic setting that has been aligned to echo the principal axes of the house. The maximum dimensions of the ground-penetrating radar responses for the buried sarsens have been recorded as a proxy for the size of the buried stone, along with an estimate of the depth of the burial pit. Four of the anomalies fall at the upper end of the size range for the smaller Z-stones, while five are comparable in size to the main Southern Inner Circle megaliths. In all cases, the depth of burial is within the known range. Enough of this megalithic square had survived into the seventeenth and early eighteenth centuries for both Aubrey and Stukeley to record its remnants. This suggests that the constituent megaliths had not been dismantled or reconfigured in prehistory. The excavated sarsens and the unusually large stoneholes encountered by Keiller indicate a mixture of larger and smaller stones. If set in alternate fashion, the result would form a contrast between the grey of the larger sarsens and the distinctive orange-red of the surviving Z-stones. The megalithic square is a highly unusual monument in its own right, the closest parallel being the ‘cove’ inside site IV at Mount Pleasant in Dorset. At 36m², however, the latter is considerably smaller.
Additional newly identified features include the sub-circular anomaly seemingly cut by the Southern Inner Circle and two lines of stoneholes radiating from the centre. The former is reminiscent of a double concentric circular anomaly identified previously in the Northern Inner Circle.
Gillings et al.'s preferred structural sequence for these newly identified features begins with the putative house, followed by the erection of the Obelisk and the square stone setting, and then the construction of the Southern Inner Circle and associated lines. The circular anomaly may pre-date the Southern Inner Circle, but direct dating evidence is currently lacking. By analogy with other Early Neolithic structures, the putative house should date to the second quarter of the fourth millennium BC. Shards of Neolithic bowl and Peterborough Ware from stoneholes in the square setting are presumably residual, although some appear quite fresh. Perhaps this and the Obelisk were constructed in the late fourth or early third millennium BC - a period that might also have witnessed the erection of the Cove stones inside the Northern Inner Circle. The radiating lines form a final megalithic phase. They each have a different origin point and both appear to have been carefully keyed into stones of the square and Southern Inner Circle, implying that the former were already in place. Overall, the site may record activity spanning as much as 1500 years, from the Early Neolithic to the Early Bronze Age.
The newly revealed structural detail of the Southern Inner Circle. Gillings et al. (2019).
If Gillings et al.'s new interpretation of the structure within the Southern Inner Circle as an Early Neolithic house is correct, the implications for understanding Avebury’s origins are profound: the ancestry of one of Europe’s great megalithic monuments can be traced back to the monumentalisation of a relatively modest dwelling. This supports the view that fourth-millennium BC tombs and houses/halls played an active role in the creation and commemoration of foundational social groups. Eventually encased within the centre of the ‘deepest’ space of the henge, Gillings et al. hypothesise that it was the connections that this erstwhile building had with a significant, perhaps founder, lineage that led to it taking on a (mytho-)historic importance; and for the status of the site to move from the quotidian to the sacred. Avebury is not unique in this transformation from ‘mundane’ to monumental structure. The process is evidenced at several Neolithic monuments, such as in the construction of an earlier fourth-millennium BC chambered tomb over a former house at Hazleton North, Gloucestershire, and the later reworking of large free-standing buildings or halls into henges and timber and stone circles at Stenness, Orkney, and Coneybury, Wiltshire. What marks Avebury as exceptional is the heightened significance and long-term resonance of this act of ontological transformation.
The Early Neolithic house at Avebury would have lasted perhaps only a generation or two; the collapsed daub walls would probably have left a visible earthwork that was subsequently afforded careful respect. Later acts of pit digging and artifact deposition highlight the long-term memory work that could attend the visible traces of Early Neolithic houses. Similar activity has been observed in the deliberate digging and filling of later Neolithic pits containing Grooved Ware into the house/hall sites at Yarnton, Oxfordshire, White Horse Stone, Kent, and Littleour, Fife. A Middle Neolithic pit group was carefully dug between the traces of two early fourth-millennium BC houses at Llanfaethlu, Anglesey, while at Cat’s Water, Fengate, Cambridgeshire, pits containing Peterborough Ware were dug along the edge of a centuries-old house.
Since its unexpected discovery in 1939, the Z-feature at Avebury has presented an interpretative conundrum. Isobel Smith came close to Gillings et al.'s preferred explanation when she proposed a link with Early Neolithic funerary architecture, in that the settings within the Southern Inner Circle deliberately echo elements of a long barrow, with the Obelisk representing a burial deposit. Instead of a tomb, however, the Z-feature settings can now be considered to commemorate a form of domestic architecture. Neolithic house forms in Britain changed over time, from square and rectilinear to more oval and rounded later forms. It may be that an explicit link with concepts of the house and household was maintained at Avebury; the subsequent enclosure of the square megalithic setting and erstwhile house by the Southern Inner Circle may replicate - on a truly monumental scale - the square-in-circle format of later Neolithic houses and halls.
Finally, given the frequency with which Early Neolithic houses in Britain and Ireland occur in pairs or small groups, Gillings et al. suggest that we might expect there to be more early houses at Avebury. Indeed, the Cove that sits in the centre of the Northern Inner Circle, amidst a confusing array of un-investigated stone settings, may be a good candidate for the site of a second foundational
Follow Sciency Thoughts on Facebook.