Just over half a century ago, roadworks to widen the A303 near Amesbury revealed intriguing evidence of Neolithic artistry about 1.2 miles from Stonehenge. At the bottom of a pit that also contained animal bones, an antler pick, and fragments of Grooved Ware pottery were two square plaques carved from chalk, each decorated with complex geometric patterns that are thought to represent deliberate designs. At the time of the road scheme in 1968, precise radiocarbon-dating technology was not available, but more recent analysis of the animal bone from the pit has placed its contents in the earlier 3rd millennium BC. Since their discovery, these plaques (studied by permission of The Salisbury Museum and now included in the permanent displays at the Stonehenge Visitor Centre) have been joined by two more examples recovered from the Stonehenge landscape, all excavated within three miles of each other.

One of these was found at Butterfield Down in 1990. It too was discovered in a pit, albeit one dated to the Romano-British period. Nevertheless, the markings on this rectangular/roughly trapezoidal piece of chalk – on one side, a series of vertical lines contained with a narrow border; on the other, two horizontal bands filled with vertical lines; and further grooves scored into the edges – suggest that it is of similar date to the Amesbury plaques. The most recent addition to the group, discovered in 2017, is a small fragment of a plaque excavated at Bulford. It came from a Neolithic pit – one of an arc of around 50 – and was accompanied by other artefacts including chisel arrowheads, animal bones, and Grooved Ware pottery. This collection has been interpreted as a refuse deposit, which the animal bones suggest was dumped in c.2950 BC. Interestingly, the wider site also produced some 22 other worked chalk objects, including cups and balls.Chalk was a popular material for expressive creations during the Neolithic period. It is easily carved and has a surface that can be smoothed and reworked with little effort compared to harder types of stone. Where chalk was locally available, it was a convenient and versatile medium for Neolithic crafters. Numerous other examples of decorated chalk pieces are known from this period, from a block marked with scored lines that was recovered from the enclosure ditch at Maiden Castle, Dorset, to a block with hatched markings from Durrington Walls. There are much more elaborate artefacts, too, not least the enigmatic Folkton Drums, a trio of chalk cylinders that were found in a North Yorkshire grave, nestled against the skeleton of a Neolithic child. The purpose of the Drums is unknown, though they are elaborately decorated, wound about with ornate patterns of grooves, stylised faces, chevrons, and triangles. Some of the geometric elements have been compared to decoration on Grooved Ware pottery, and to the designs found incised on some of the stones at the Ness of Brodgar in Orkney (CA 335).
A suggested sequence
Since their discovery, the Amesbury and Butterfield Down plaques have been recorded using monochrome photographs and hand-drawn illustrations (the Bulford find has no formal report yet). Elements of their decoration have proven difficult to reconstruct due to surface erosion – but now advances in photographic technology have allowed far more details to be revealed in a recent study by Wessex Archaeology.
The initiative, co-authored by Bob Davis, Dr Phil Harding, and Dr Matt Leivers, and published in the Proceedings of the Prehistoric Society (see ‘Further reading’), used reflectance transformation imaging to explore the plaques’ surfaces in detail. This photographic technique combines multiple images that have been taken with light projected from different angles to produce a variety of highlights and shadows, revealing surface detail, shape, and colour invisible to the naked eye. As a non-invasive technique, RTI is also well suited to fragile artefacts, and has been used to great effect on the Folkton Drums, revealing evidence of a previously unknown, erased, face on one of them, suggesting that their designs had evolved over time (CA 345). It has also helped greatly with analysis of the Norse runes carved on the walls of Maeshowe chambered tomb in Orkney.
When applied to the chalk plaques, the study authors write, this method has revealed an intriguing range of artistic abilities behind their designs. Some more ephemeral elements were apparently rapidly executed sketches, while others suggest more deliberate, even mathematical, planning and composition. The detailed images produced during the study have enabled the researchers to tease out a possible order in which each element of the designs was created – and have even revealed some possible insights into the inspiration behind the motifs, suggesting that they might not all be entirely abstract in intention.
Take, for example, Plaque 1 (the larger of the two Amesbury finds), whose markings appear particularly logical and systematic in their execution. The researchers’ reconstructed sequence for this design begins with a number of faint horizontal lines that probably reflect an initial preparation of the chalk surface. After this, lightly cut lines forming fairly poorly executed chevrons of irregular depths were added – markings that were previously undocumented, revealed only by the RTI – together with a vertical ‘arrow’-shaped mark. The top layer comprised a narrow border filled with short slashes, which surrounded a central panel. The top half of this space is filled by a distinctive, meandering ‘Greek key’ design, below which are a series of horizontal slashed lines, and then another, inverted ‘Greek key’ mirroring the one above. The reverse is much simpler, bearing two lozenges.
Abstract images?
It was initially thought that the horizontal slashes dividing the two ‘Greek keys’ were fairly random, but RTI revealed a notable similarity of orientation and depth in their lines. Moreover, none stray beyond the border, suggesting they were created with more care and intent than first suspected. As a result, the study team wonders if these motifs might not represent an abstract scheme. In their paper, they describe suggestions that the upper ‘Greek key’ could represent some kind of object known to the Neolithic artists, possibly some form of stone or wooden portal placed on the ground and reflected in water, with the horizontal lines depicting some kind of river, and the inverted motif a reflection in water.

‘Elements of the design bear an undeniable similarity to the Stonehenge trilithons, and it is tempting to suggest that this is what the plaque represents: Stonehenge reflected in its water-filled ditch’, the authors write – then immediately acknowledge that radiocarbon dating makes such an interpretation impossible and that the ditch once contained water is equally improbable. Plaque 1 is considerably older than the phase of Stonehenge’s construction when the sarsen trilithons were erected – however, the team suggests, another possible link could be made to the smaller bluestones. Some have traces of tenons, suggesting that they too once bore lintels, and the researchers wonder whether some of these stones might have originally formed parts of other nearby circles – perhaps at the henge at west Amesbury, at the end of the Avenue, where they might be reflected in the Avon.
While such interpretations are purely speculative, there is one other aspect of the plaque that the team thinks could be more figurative than abstract. RTI has revealed the design of the plaque’s border in much greater detail, and it has been suggested that the fill within these lines might be intended to represent some kind of twisted cord or rope.
Details of designs
Plaque 2, the other Amesbury find, is markedly different to, though no less detailed than, its counterpart. Its surface is carved with a narrow border filled with chevrons, surrounding a busy central panel containing two rows of chevrons and three of lozenges. RTI indicates that, after the surface was prepared, this border was created first, and that the central panel had been carefully divided into thirds before the decorations were cut, suggesting painstaking attention to the detail of the design.

Plaque 3, from Butterfield Down, is rather simpler in its design, though its surface is heavily abraded, possibly due to its later Romano-British use. Unlike the two Amesbury plaques, which have an apparent ‘main’ design and a simpler reverse, this object is decorated with equal emphasis on both sides. One is marked with a border in which a series of near-parallel lines divide a central panel into ten vertical columns. The other is more heavily damaged, making it more difficult to discern the design, but it appears to have a border surrounding three panels. The outer two are filled with vertical lines, creating what the team describes as a ‘ladder’ effect, while the central panel has been left plain.
The most recent find, Plaque 4, from Bulford, is much more ephemeral in its decoration than the well-constructed and complex designs on the others. As only a small portion survives (if the original was a similar size to the other three, perhaps 25-30% of the plaque remains), it is impossible to say what its complete scheme looked like. However, incised lines can still be seen on both sides of the fragment, creating very basic patterns. There are no borders or central panels in this case, but two oblique parallel lines cut with shorter ones create a ladder effect that is overlain with two deep and apparently unrelated lines. On the reverse, lightly cut lines are overlain with two narrow leaf shapes cut more deeply and with more purpose. The team believes that these markings are conscious creations, but it is less clear than with the other plaques whether they were intended as a structured design or as a casual doodle or a ‘sketch book’ setting down multiple ideas.
As for what the markings on all four plaques might mean, their geometric patterns do fit into the wider repertoire of Neolithic imagery, with similar designs known from carved stones and pottery originating in Orkney. Chevrons, lozenges, and ladder motifs appear in artistic endeavours spanning the middle Neolithic period to the Early Bronze Age. It is interesting, however, that the chalk plaques mostly involve straight lines, with none of the spirals, cup marks, and circles of contemporary rock art. Representative art in the British Neolithic is very rare, with abstract imagery dominating known designs, although a small number of figurines from Orkney, and the faces on the Folkton Drums, show that artists of this period could execute more literal depictions when they wanted to. The more ‘figurative’ elements of Plaque 1 revealed in the study are therefore particularly noteworthy, illustrating how new technology can shed transformative light on our understanding of Neolithic creativity.
Artistic experiments
The level of detail provided by the RTI study has helped the team to speculate on the type of tools used to carve the plaques. After analysing the cut marks, two test plaques were produced so that the researchers could experiment with flint and antler tools (a flake and a piercer, and a spatula and a point respectively), making different marks on the chalk. They found that all of the tools worked efficiently to incise the objects’ surfaces, but that the flint flake produced results closer to those on the Neolithic originals, creating light, narrow, but clear lines with steeply angled side walls – as seen in Plaque 1’s ‘Greek key’ designs and the ‘ladder’ motif seen on Plaque 4.

Another experiment set out to explore what might have caused the surface damage on the plaques, and the impact this could have had on their designs. To investigate, one of the test pieces was placed outdoors from early February to late April, and rephotographed using RTI after two days and two months. Despite being exposed to frost, wind, and rain, weathering seems to have had little impact on the plaque’s surface or the clarity of its design. This suggests that the artefacts were robust, durable objects that could have had prolonged use – perhaps so enduring that they achieved a second life as objects of fascination for much later observers, as was the case for Plaque 3, which had apparently attracted Romano-British attention.
When the team tried carrying the plaque around in a cloth bag for two weeks, however, this had a much more dramatic effect, polishing and rounding sharp edges and raised points even in this short period. The designs remained visible, however, suggesting that even in their worn state the motifs seen today were probably close to the original imagery.
Further reading
B Davis, P Harding, and M Leivers (2021) ‘Reflectance transformation imaging (RTI) investigation of engraved chalk plaques from the Stonehenge region’, Proceedings of the Prehistoric Society pp.1-28. Available free at https://doi.org/10.1017/ppr.2021.13.
IMAGES: Courtesy of The Salisbury Museum and Wessex Archaeology.