A monumental mystery: Unpicking the evolution of Arthur’s Stone

On a soaring green ridge overlooking Herefordshire’s Golden Valley, an enormous slab of stone perches precariously on nine much smaller uprights. The feat of engineering ingenuity involved in raising the capstone – which weighs an estimated 25 tonnes – on to its supports and keeping it from crashing to the ground, all without the help of modern technology, seems astonishing, but that is exactly what was accomplished c.6,000 years ago. Known as Arthur’s Stone, the Neolithic monument has long been linked to legend and literature, associated with Arthurian giant- slaying escapades and inspiring the Stone Table on which the lion Aslan was sacrificed in C S Lewis’ The Chronicles of Narnia. Until recently, however, it had seen little modern archaeological research. Now, investigations led by the Universities of Manchester and Cardiff (permitted by English Heritage, in whose care the stones lie, and by Historic England) have revealed that the site’s real-life story is no less compelling than its imaginative associations.

Arthur’s Stone was traditionally thought to be an unusually northern outlier of a category of monuments known as the Cotswold-Severn Group. These are Neolithic long cairns – stone-lined communal burial spaces covered by mounds of stone and earth – that share distinctive features such as a trapezoidal form, curving banks or ‘horns’ embracing a forecourt at the wider end of the cairn, and sometimes a large stone known as a ‘false portal’, which looks like a doorway but is completely non-functional. Around 200 examples are known in western Britain, predominantly in the Cotswolds and South Wales (see CA 395) but, so far uniquely, Arthur’s Stone preserves the remains of a rather different, earlier form of monument at its heart.

It transpires that the tomb had, in fact, had an earlier incarnation as a dolmen – a single burial chamber formed from upright stones supporting a large horizontal slab (see CA 286 and CA 390) – whose entrance faced north, away from the rolling valley landscape. Later, however, this monument was covered over with soil and stones to transform it into a long cairn, and the new construction was apparently oriented in the opposite direction, with its horns and blind portal pointing to the south (though more work is required to understand fully the details of this change). The recent investigations (undertaken as part of the wider Beneath Hay Bluff Project, which has been exploring the early prehistory of south-west Herefordshire since 2010) set out to unpick this evolution – and excavations just beyond the upstanding remains have revealed that the monumentalised area is even older and was once much more extensive than above-ground evidence suggests.

Immediately to the south of the stones, in a field just outside the fenced area of the scheduled monument, the project team have found what is thought to be the earliest structure yet seen on the site: the remains of a long mound of stacked turf, enclosed by a timber palisade in a continuous slot. This mound appears to have pointed south-east, towards a neighbouring and highly significant Neolithic site on Dorstone Hill, a little more than 1km away. Intriguingly, it is strikingly similar in construction to the central of the three earth and stone mounds that have been found at that site (see box on p.28), which also comprised a stack of turves held in place by a modest encircling palisade.

Contemporary with the turf mound was an avenue of at least ten pairs of posts (their southern limit has not yet been found) running towards the mound in two parallel lines. A redesign then followed, replacing these wooden uprights with a smaller number of stones whose sockets yielded fragments of Peterborough ware. The presence of middle Neolithic pottery sherds indicates that the addition of the stones was a later step – albeit still unusually early for the erection of a stone row, project co-director Keith Ray notes, with the caveat that these features are generally currently poorly dated – and the team do not know of another example in Britain where an avenue of timber posts was replaced by one of standing stones.

The precise dates of, and relationships between, all of these different phases are not yet clear, nor is it yet fully understood how (or if) the revetted mound related to the dolmen. Today, the turf feature stands clear of the stone-built monument, but it may have been cut away during the construction of the long cairn’s façade; the team now know that it had once extended at least as far north-west as the passageway examined this summer (of which, more below). While their sequence is still being unpicked, though, all of these traces testify to this picturesque spot being acknowledged as somehow special for a long time, and that this perception was commemorated in turf and timber before a more permanent stone monument arose nearby.

Tracing transformation

As for Arthur’s Stone itself, the last two years of excavations in particular have helped to illuminate both its dolmen and cairn phases – and have revealed that some features previously thought to be contemporary in date actually reflect the monument evolving over time.

In 2022, a long trench opened just below the cairn’s ‘horns’ exposed an expanse of beautifully worked, neatly ordered drystone walling. This suggested that the long cairn’s covering had not been carelessly heaped on top of the burial chamber: rather, it had been very carefully constructed – and that was not all. A series of thin, flat stones had been deliberately placed vertically along the outer face of the wall. Were they purely decorative in function, or might there have been a deeper significance behind their positioning? This year’s excavations have added a wealth of detail to this picture: another trench was opened to the west of its predecessor, wrapping around the side of the upstanding remains in hope of catching the outer edge of the cairn to determine its full extent. The team were not disappointed in their efforts: the 2023 trench did indeed locate the cairn’s long south-west-facing side, and also allowed archaeologists to explore its make-up in detail.

Once again the team uncovered finely worked courses of drystone walling and again it was double-skinned, with two narrow walls built directly against each other. Keith suggests that these do not represent successive phases of construction, but that they were built at the same time. The inner of the two is more structural in form, while the outer is finer and more decorative, like that found the previous year. Significantly, in this area too, a single layer of small, thin, stones were found against the outer face of the outermost wall. This was not random tumble, but clearly deliberate in intent, as each stone had been seemingly set upright against the wall and then ‘pinned’ in place with an extra layer of stone revetment. Could these flat stones have been brought by visitors to the tomb, and left as a sign of reverence? Or might their addition have come at the end of the cairn’s life, and represent some kind of ‘deconsecration’ or closing gesture? This last interpretation might be suggested by the fact that some of the little uprights appear to go across the entrance of a passageway entering the monument from the north-west, potentially blocking access to the tomb (whether practically or symbolically).

This passage was initially thought to be contemporary with the original stone chamber, as it connects with two large upright stones marking the entrance to the dolmen. However – as project co-director Julian Thomas highlighted during my visit – it is now thought to have been a later addition, as the edge-set large stone slabs are of a more laminated lithology that is very different to those at the mouth of the dolmen, which are in turn much more weathered. Both sets of stones are thought to have been quarried locally, Julian added, possibly just reflecting different layers of the bedrock, but they nonetheless indicate that the two components of the monument represent different phases.

The passageway itself has proven to be a rather curious construction: its course is not straight, and while most of its walls are formed from thin slabs, one section is edged not with orthostats, but with a short stretch of drystone. The team wonder whether this might hint at an originally T-shaped corridor, part of which was subsequently blocked up. The passageway ultimately led to a small chamber formed in part by the entrance stones of the dolmen, but after a period of time both of these elements were sealed by stony deposits and the entrance blocked up. Within this space, the researchers were delighted to discover undisturbed floor deposits; as the orthostats have been exposed to view for centuries, the team had feared finding a ruinous interior, its contents shovelled away by 18th- and 19th-century antiquarian activity. Instead, hopes are high that analysis of material from the passage will shed more light on the tomb’s use and environment. These investigations have already recovered a number of deposits of disarticulated human bone – a rare survival in the area’s very acidic soil.

At other long cairns, where local conditions are more conducive to the survival of human remains, analysis of their contents has revealed that they were used to house generations of a community’s dead: indeed, at Hazleton North in Gloucestershire, a recent genetic study was able to reconstruct a five-generation family tree (CA 384). Arthur’s Stone may not be able to produce this level of detail, but the bone fragments discovered there this year could provide tantalising echoes of its long-finished funerary function. They are thought to reflect the partial remains of multiple individuals, who were most likely brought to the monument as fleshed cadavers – then, after decomposition had taken place, the bones were rearranged, mixed together, and placed in discrete piles.

Excavations around the monument’s entrance and within the passageway and chamber have also recovered Neolithic pottery and stone tools, as well as more ‘exotic’ items. These include a piece of worked rock crystal, a material that was used extensively at Dorstone Hill too (CA 391), though – as Dr Nick Overton, one of the project’s Associate Directors, notes in a recent article for the Cambridge Archaeological Journal – it has not previously been recognised as occurring widely on Neolithic sites in Britain. Nick has led the project’s research into rock crystal, and it is believed that the Arthur’s Stone example may have come from Snowdonia or Cornwall. Another key find was a piece of black, glass-like pitchstone which only occurs in the natural volcanic dykes of the Isle of Arran. This material is frequently found on Neolithic sites in Scotland, but this is thought to be the furthest south that it has been seen in Britain. Such discoveries hint at the dynamic and far-reaching exchange networks that these early farming communities were able to tap into.

Evidence of origins

Around to the east of the monument, the stones are embraced by a substantial curving bank, shaped almost like half a doughnut. It had previously been thought that this was a modern feature, representing grassed-over upcast from work on the road running beside the site – however, this year’s investigations have revealed its much earlier origins. Unlike the neat drystone walling of the cairn, the make-up of this mound comprised much larger, more cobbly pieces that were less ordered, though still deliberately arranged in layers with facing stones at each level. It is now thought that the ‘doughnut’ was associated with the dolmen phase, built soon after the capstone chamber, as excavation revealed that its end butted up against one of the two uprights marking the chamber’s entrance. These may have framed a way through the ‘doughnut’, which perhaps once completely encircled the monument (with the capstone always visible, the team suggest) before the western portion was later robbed away to create the half-moon seen today. If this is the case, to the team’s knowledge it would be the first example of a dolmen with a bank and entrance like this in Britain, though there are possible parallels in Denmark.

The passageway is then thought to have been added as the site evolved into a long cairn, and the original monument was subsumed within the large trapezoidal mound bounded by drystone walling. How long, though, was the long cairn? As well as the excavation focused on the upstanding stones, this year two more trenches were opened in a field on the other side of the modern road. The closest of these was designed to see if the northern tail of the cairn extended into this space, and at the time of my visit a number of potentially promising concentrations of stone were emerging from the soil – though it was not at that stage clear whether they did represent part of the Neolithic monument, or something like a later building platform.

Another aspect of the monument’s construction was under investigation further up the field. The dolmen’s mighty capstone measures some 13ft by 7ft, and it would have been a considerable undertaking to bring it up to the ridge where the dolmen was being built – unless it was sourced rather more locally. A short distance upslope from the road, about 100m from Arthur’s Stone, the team had noticed a distinctive, big depression in the ground. Could this isolated dip represent the remains of an ancient quarry – one large enough to have produced the capstone? Sure enough, excavation of the depression has revealed what does appear to be quarrying evidence, its base littered with broken stone, and this is believed to be prehistoric in date – though it is more likely to be the source of stone used in the later cairn phase.

What, then, can we understand about the meaning behind the monument’s metamorphosis? Within the horseshoe of Cotswold-Severn long cairns known in this region, Arthur’s Stone stands out as the only dolmen among the group. This earlier form of tomb tends to be concentrated more towards Wales; why was an isolated example built out to the east? From the construction of the Dorstone Hill halls that were later burnt and turned into barrows, to the outlying dolmen, the people who lived in this part of the Golden Valley during the Neolithic period clearly did things differently to their neighbours – at least at first. Their monuments may have been intended as highly visible statements of ancestral claims to the surrounding land – but, thanks to evidence from the recent excavations, we can now see that their distinctive sites were ultimately firmly drawn into the Cotswold-Severn tradition, with their designs changing dramatically to conform to this different kind of construction, and their once highly localised practices replaced by more regional ones. Whether this was driven by conquest, conversion to new religious or political beliefs, or a coming together of communities cannot be known – but today, as our understanding of it expands, Arthur’s Stone stands as a symbol of this monumental change.

A ceremonial landscape 

Arthur’s Stone is not the first Neolithic site in the immediate area to be investigated by the Beneath Hay Bluff Project. Between 2011 and 2019, their excavations shed illuminating light on a near neighbour at Dorstone Hill (CA 285 and ‘Context’ in CA 321). There, three long mounds were found to overlie the remains of once-imposing wooden halls that had been built in a row, gable end to gable end. These structures had met a dramatic end, with their timbers mainly burnt in situ – and, after the fire, the charred remains were covered over with earthen barrows, transforming gathering places for the living into houses for the dead. A further mound, in line with the others, covered an earlier post-framed burial chamber. This mound was ditched and cremated human bone had been placed in its lowest fills.

Recent Bayesian statistical modelling (published in Antiquity and subsequently summarised in the ‘News’ section of CA 402) has revealed that the halls were broadly contemporary in date, spanning a period of c.3870-3710 BC and representing some of the earliest Neolithic activity known in the West Midlands. Despite their impressive scale, however, they had not operated for long before they were destroyed and replaced by mounds, between c.3770 and 3705 BC. Still within the 4th millennium, shafts were then dug down into the eastern and central mounds, descending as far as the layer of burnt material that they covered. The bases of these shafts were paved, and on these surfaces were placed the cremated remains of cattle and people. 

The meaning behind these enigmatic interventions remains obscure, but there was one final act to come: a series of stone-lined cists were later inserted into the northern side (and between) all the mounds. After this was done, a single stone façade facing north over the Wye Valley was constructed to link all three mounds, creating a massive (and, so far, unique) linear feature more than 100m long.

With the discovery, too, of a Neolithic causewayed enclosure just to the south-east (a particularly early example, dating to c.3770-3640 BC) and Arthur’s Stone thought to have been built c.3700 BC, it appears that this part of the valley was once a busy ceremonial landscape peppered with highly visible monuments that continued to be visited and venerated for many years. For a more detailed exploration of Neolithic perceptions of ancestral landscapes and lineage history, see Keith Ray and Julian Thomas, Neolithic Britain: the transformation of social worlds (OUP, 2018 and 2020).

Source:
• Keith Ray is Honorary Professor in Archaeology at Cardiff University.
• Win Scutt is Senior Properties Curator at English Heritage.
• Julian Thomas is Professor of Archaeology at the University of Manchester.

Further information:
For more information about Arthur’s Stone, visit the website http://www.english heritage.org.uk/visit/places/arthurs-stone.
This summer, the same field school (a partnership of the Universities of Cardiff and Manchester, Herefordshire Archaeology, and the Institute for Field Research, California) was also carrying out excavations at nearby Snodhill Castle. I visited this dig, too, during my Herefordshire trip (it was led for the Snodhill Castle Trust by Tim Hoverd of Herefordshire Council, one of the Associate Directors of the Beneath Hay Bluff Project); watch out for a feature on their findings in a future issue of CA.

All Images: Adam Stanford, unless otherwise stated