The Altar Stone — interactive 3D model

Stand alone webpage - https://icy-flower-e5bb.tim-daw.workers.dev/ - opens in new window. Built with Claude Fable 5

Rediscovered Stonehenge Aberdeen Monolith Link

Abstract

In 2024 the central Altar Stone of Stonehenge (Stone 80) was shown to be Old Red Sandstone from the Orcadian Basin of north-east Scotland, carried at least 750 km to Salisbury Plain. That result has revived an old intuition: that Stonehenge’s great recumbent slab and the recumbent stone circles (RSCs) of Aberdeenshire belong to a single, far-flung megalithic conversation. This paper tests that intuition against the most detailed record we have of an RSC under excavation — V. Gordon Childe’s final report on Old Keig and the engineering study H. E. Kilbride-Jones published alongside it, both in the Proceedings of the Society of Antiquaries of Scotland for 1934. The reports document deliberate quarrying and shaping of the monoliths, a standardised, asymmetric profile with a pointed base, and a lever-and-fulcrum method of erection that Kilbride-Jones explicitly carried south to the Stonehenge trilithons. Set beside the modern provenance work and the reading of the Altar Stone as a deliberately laid recumbent — an interpretation that goes back to Gordon Freeman’s alignment observations and that I developed in 2015 — these papers supply the concrete archaeological substance for comparison. I argue that the parallels are real and worth taking seriously, but that they are most safely read as evidence of a shared insular repertoire of practice and symbolism rather than of direct cultural transmission in either direction. I also show where the 1934 chronology, on which Kilbride-Jones built an explicit sequence, has since been overturned.

Introduction

On 14 August 2024, Clarke and colleagues reported in Nature that the Altar Stone at the heart of Stonehenge — Stone 80, a six-tonne slab of Old Red Sandstone — is not Welsh, as had been assumed for a century, but came from the Orcadian Basin of north-east Scotland, at least 750 km away and most plausibly moved by sea (Clarke et al. 2024). The finding did two things at once. It severed the Altar Stone from the Welsh bluestones with which it had long been bracketed, and it placed its geological birthplace in the one corner of Britain that possesses a stone-circle tradition built around a great horizontal stone: the recumbent stone circles of Aberdeenshire and its neighbours.

That coincidence is the subject of this paper. It is not a new thought that Stonehenge’s recumbent Altar Stone and the Scottish recumbents might be related; what is new is that we now have a geological reason to ask the question seriously. I have my own stake in it. Building on Gordon Freeman’s observation that the Altar Stone’s long axis is skewed to a secondary solstitial line — from midwinter sunrise to midsummer sunset (Freeman 2009; 2012) — I argued in 2015 that the stone was never a fallen upright but was laid flat deliberately, as one element of a designed and skewed central setting (the “Twisted Trilithon”; Daw 2015; 2025). Parker Pearson and colleagues (2024) have since reached the same conclusion about deliberate placement and gone further, proposing that the Altar Stone’s position — recumbent, in the south-western arc of the monument, at the foot of the two tallest uprights of the Great Trilithon — echoes the plan of a recumbent stone circle.

Are the echoes based in the actual monoliths? How a recumbent stone circle was actually built: how its great stones were quarried, shaped, raised and bedded. For that there is no better source than the excavation of Old Keig, the circle with the largest recumbent of them all, dug by Childe in 1932 and 1933. His final report appeared in the Proceedings of the Society of Antiquaries of Scotland for 1934 and — uniquely among excavation reports of its kind — it was accompanied in the same volume by a dedicated engineering study by his site assistant, H. E. Kilbride-Jones, who had watched the Old Keig monoliths come out of the ground and who explicitly carried his conclusions south to the trilithons of Stonehenge (Childe 1934; Kilbride-Jones 1934). These two papers, ninety years old and written long before anyone could have guessed where the Altar Stone came from, turn out to be the natural starting point for the comparison.

I will argue that the parallels are real and worth taking seriously, but that the honest reading of them is more cautious than the headlines. The most defensible conclusion is that Stonehenge and the Aberdeenshire circles drew on a shared insular repertoire — of deliberately shaped monoliths, of a clever method of erection, and of the symbolism of a great recumbent stone — rather than that one region taught the other. I will also show where the 1934 chronology, on which Kilbride-Jones built an explicit sequence from Stonehenge to Scotland, has since been turned on its head.

The recumbent stone circles of north-east Scotland

Recumbent stone circles are a regional speciality. Something over seventy survive in Aberdeenshire and the adjoining counties, with none of the type found elsewhere in Britain (Burl 2000; Hill 2021). The defining feature is a single massive stone laid on its side — the recumbent — set between the two tallest uprights of the ring (the “flankers”) in the south to south-west arc, with the remaining stones graded down in height towards the opposite side. Most enclose a low ring cairn, and excavation repeatedly finds cremated bone, broken pottery and white quartz within them. The recumbent and its flankers frame a section of the southern sky, and since the work of Burl and others the circles have been read as devices for watching the moon — the recumbent acting as a level “sill” over which the low southern moon could be seen to roll at its standstills (Burl 2000). Richard Bradley’s excavations at Tomnaverie and elsewhere have nuanced this picture: he found that the circles were often raised around pre-existing ring cairns, and questioned whether any precise lunar sightline was ever intended, with a general orientation towards the southern sky and midwinter sunset doing much of the work (Bradley 2005). The lunar association remains the dominant interpretation, but it is now held more cautiously than it once was.

Crucially for what follows, these are monuments of the late Neolithic and the very start of the Bronze Age. Modern dating — anchored by Bradley’s fieldwork and by the broader chronology of the associated ring cairns and the earlier Clava cairns — places their construction around 2500 BC, in the last few centuries of the third millennium (Bradley 2005; Hill 2021). That is a point to hold on to, because it is not what Childe and Kilbride-Jones believed in 1934.

Old Keig itself is the giant of the class. Its recumbent is the heaviest known, on the order of fifty tonnes — a single block whose transport and placement would have been a feat in its own right (Burl 2000; Hill 2021). It is precisely because the circle is so large, and was so carefully recorded, that Childe’s report repays close reading.

Old Keig as Childe found it

Childe stripped the south-eastern half of the circle to bare subsoil in 1933, having cut his first trench across it in 1932 (Childe 1933; 1934). He recovered the bases of the surviving uprights, traced the bank, and — most importantly — established the relationship between the stones, a “compaction layer” of trampled subsoil, and a central cairn.

The recumbent itself sits almost dead level, its underside “remarkably smooth,” about 1.1–1.2 ft above his datum. Strikingly, over most of its length it rests on neither bedrock nor undisturbed soil: east of his baseline the only solid support for the heavy eastern end is a single block of rock wedged tightly between the stone’s base and a ridge of bedrock (Childe 1934, 376–7). The great stone was, in effect, balanced and packed into place rather than founded on rock.

The flankers and the one surviving circle-pillar tell the same story of careful seating. The east flanker (Pe) stands 9½ ft high; about a foot above its lowest point it is 4 ft 6 in wide, and below that it tapers to a roughly triangular point lying about a foot back from the edge nearest the recumbent. Its base sits in a shallow groove quarried into the bedrock — only some 4 to 5 inches deep and about 3½ ft long, with only the inner edge clearly cut, because the bedrock shelved away on the outer side. The outer edge of the stone was then held by two wedge-stones driven in beneath it, and the inner side packed with large stones bedded on the compaction layer (Childe 1934, 377–8). The pillar P3, the only stone of the great circle still standing apart from the flankers, was set in a socket just nine inches deep and held by wedges, two of them placed obliquely so as to act as “skid-stones,” guiding the base down into the hole as the stone was raised (Childe 1934, 379). Childe noticed that all the pillars he examined were “more or less pointed at the base,” and referred his readers to Kilbride-Jones’s paper for the explanation (Childe 1934, 385).

What the circle enclosed was a burial. Towards the centre Childe found the disturbed remains of a ring cairn — a rough kerb of slabs about 15 ft in radius, set on the compaction layer — and beneath it a patch of subsoil baked brick-red by intense heat, cut by a small east–west trench that, in his words, “looks suspiciously like a grave.” Cremated human bone was spread through this area, together with fragments of several large coarse urns, which Childe suspected were containers for offerings rather than for the ashes themselves (Childe 1934, 386–8). The relationship of the cairn to the compaction layer let him reconstruct a sequence: the uprights were set up first, the trampling that produced the compaction layer happening during their erection, and only later was the cairn heaped over the burial — “just as a Pharaoh would build his pyramid before his death,” as he put it (Childe 1934, 388).

Then, the dating — and here the report shows its age. The only chronological evidence was the pottery: a coarse, flat-rimmed ware which Childe matched, on the one hand, to the fabric then assigned to “Iron Age A” in England and, on the other, to material associated with Late Bronze Age objects at Covesea, with a lignite armlet pointing the same way (Childe 1934, 389–91). He therefore placed Old Keig in a “Late Bronze Age” horizon. We now know this is far too late: flat-rimmed wares of this general kind were long-lived and undiagnostic, and the recumbent circles as a class belong to the late third millennium, some fifteen centuries or more earlier than Childe supposed (Bradley 2005). The excavation is excellent; the calendar attached to it is obsolete. That distinction matters for every comparison with Stonehenge that follows.

Illustrations from Kilbride-Jones

Deliberately shaped stones: Kilbride-Jones on the monoliths

Kilbride-Jones’s contribution begins with an observation that still carries weight: the stones at Old Keig were not rough boulders but were split from the living rock and shaped to a deliberate, repeated form. He measured three of them in detail (Table 1). Each is thicker towards the base and along one face; each has a pointed base, with the point set not in the middle but below one edge, so that the profile resembles a right-angled triangle; and in each the apex is carried vertically above the point of the base. The straight, thicker side and the pointed base together throw the centre of gravity towards one side and keep it low (Kilbride-Jones 1934, 83–8).

The finest of the three, the prostrate Pm, clinched the argument. Its four sides are comparatively straight, it is “a very regular trapezium,” and — most tellingly — its base bears careful tooling that Kilbride-Jones took as positive proof that the shape of the base was intentional, “the result of a preconceived idea and plan” (Kilbride-Jones 1934, 85). He thought the tooling had been done with a metal tool, probably bronze, and drew an explicit comparison with the dressing of the Stonehenge sarsens, which were worked with sarsen mauls (Kilbride-Jones 1934, 83, 85). The roughest stone, Px, reused as the footing of a field-dyke, shows the same design crudely executed; the form, he argued, was a standard, more or less well achieved according to the care taken.

Table 1. Principal dimensions of the three Old Keig monoliths recorded in 1934 (after Childe 1934 and Kilbride-Jones 1934). Figures are as given in the reports; the two authors’ measurements for Pe differ slightly.

Stone	Length / height	Max. breadth	Notes
Pe (east flanker; in situ)	9½ ft (Childe); 9 ft base–apex and 5 ft 11 in above turf (K-J)	4 ft 9 in (K-J)	Triangular pointed base; mass above the base; held by two outer wedges and inner packing
Pm (prostrate pillar)	7 ft 7 in (K-J); 7½ ft (Childe)	3 ft 8½ in (K-J)	“Finest example”; a regular trapezium; tooled, pointed base
Px (prostrate; reused in a dyke)	8 ft 6 in	3 ft 7 in	“Rudest” of the three; the same form, roughly executed
P3 (circle-pillar; in situ)	9½ ft	4½ ft at turf	Socket only 9 in deep; held by wedges and oblique skid-stones

 

A method of erection

Why shape a stone this way? Kilbride-Jones’s answer is the heart of his paper, and it is genuinely clever. He proposed that the shaped monoliths were not tipped upright from flat on the ground, nor levered up from outside or inside the ring, but were laid on their shorter, curved side; a triangular socket was cut to take the pointed base; and the stone was then rolled upright, pivoting on the point of its base as a fulcrum, with the push applied near the apex, as nearly as possible at right angles to the line from fulcrum to top (Kilbride-Jones 1934, 88–9).

The mechanics were worked out for him by an engineer, A. Regnauld, and printed as an appendix. For an idealised stone of even thickness weighing five tons, the initial push needed is about 1.575 tons; as the stone rises this force falls away rapidly, reaching zero once the stone has turned through 48°, at which point the centre of gravity passes over the fulcrum and the stone balances. For the remaining 42° to vertical the problem is reversed — the stone now tends to over-topple, and all that is required is a restraining rope around the apex (Kilbride-Jones 1934, 89, 94–6). Kilbride-Jones contrasted this with the conventional assumption of raising a symmetrical stone from flat on the ground, which on Regnauld’s figures would demand a push of about 2¼ tons sustained through almost the whole 90° arc. The shaped stone, raised his way, needs nearly a ton less force and roughly half the total work (Kilbride-Jones 1934, 89–90). The asymmetry is not decoration; it is mechanical advantage, designed in.

It is worth being clear about the status of these numbers. They are an idealised statics calculation for a five-ton stone, not a measurement of the Old Keig monoliths, which are larger and irregular; and they assume the push can be applied at the top in a particular direction — an assumption Regnauld himself flagged as questionable, since the relevant point can be well above head height (Kilbride-Jones 1934, 95–6). The value of the argument is qualitative, and it remains sound: a pointed, asymmetric, base-heavy stone is markedly easier to stand up than a symmetrical one, and the people who shaped these stones evidently knew it.

Childe’s report supplies the independent, practical check. His contractor, J. C. Milne of Bents, “experienced in handling large stones,” described essentially the same operation from the mason’s side: the stones brought up on log rollers with wooden levers, a socket dug, the base raised by leverage with packing stones slipped underneath after each small lift until it could slide into place, the oblique packers acting as skid-stones to guide the base down (Childe 1934, 384–5). The skid-stones Childe actually found in the sockets of Pe, P3 and the lesser uprights are physical confirmation that the bases were guided into prepared holes as the stones came up, exactly as a controlled, levered erection requires.

Kilbride-Jones takes the theory to Stonehenge

Kilbride-Jones did not stop at Old Keig. The second half of his paper applies the same analysis to Stonehenge, and it is here that the 1934 volume becomes directly relevant to the Altar Stone debate. He argued that the great sarsens were raised in the same way as the Scottish flankers, not tipped from horizontal as William Gowland and others had supposed. Stone 56, the tallest upright, has a pointed base resting against a wedge “in the same manner as the east flanking Stone Pe at Old Keig,” and Kilbride-Jones read Gowland’s puzzling cut-away in the chalk as the mark of a fulcrum that had slipped during raising (Gowland 1902, via Kilbride-Jones 1934, 90–2). He pointed to Lt.-Col. Hawley’s discovery of post-holes set against the broad faces of Stones 1, 7 and 30 as the footings of temporary timber buttresses — lines of saplings that steadied each stone as it was hauled past the vertical and adjusted sideways to receive its lintel, then cut off at ground level, which (he suggested) explained the decayed wood Hawley found in them. He even noted that several Stonehenge uprights (Stones 29 and 30, 6 and 7) taper to a point on both faces, a refinement that avoids the base striking the edge of the hole as the stone is brought up (Hawley 1921, via Kilbride-Jones 1934, 92–3).

Whatever one makes of the details, two things are notable. First, an experienced field observer in 1934, comparing the two monuments stone for stone, was struck by how much the engineering had in common. Second — and this is usually forgotten — Kilbride-Jones read the relationship in the opposite direction to the modern story. He believed Stonehenge “antedates the stone circles of the north-east of Scotland by several centuries,” and that the standardised, pointed-base form was an idea that appeared first, experimentally, at Stonehenge and only later became perfected and routine in Scotland (Kilbride-Jones 1934, 93–4). On his chronology, if anyone influenced anyone, Wessex influenced the north — the reverse of the implication many have since drawn from the Altar Stone’s Scottish origin.

That chronology cannot now stand. Stonehenge’s sarsen settings are dated to around 2500 BC — the sarsen circle to 2580–2475 cal BC (Parker Pearson et al. 2024) — which is essentially the same horizon as the floruit of the recumbent circles. Neither tradition is clearly the parent of the other. Kilbride-Jones’s “Stonehenge first” sequence dissolves; but so does any simple “Scotland first” replacement. On present dating the two were broadly contemporary, and the shared engineering is most economically explained as a common solution to a common problem, reached within a connected island.

The recumbent and the Altar Stone

This brings us back to the Altar Stone. It is a slab of Old Red Sandstone 4.9 m long, 1 m wide and 0.5 m thick — about six tonnes — lying flat in the south-western part of the centre of Stonehenge, immediately in front of the Great Trilithon and now pinned beneath that trilithon’s fallen upright (Stone 55) and lintel (Stone 156) (Atkinson 1956; Clarke et al. 2024; Parker Pearson et al. 2024).

Whether it was ever anything but recumbent has been debated for a century. Atkinson thought the bevelled dressing of one end showed it had once stood upright and later fallen. The contrary case rests on alignment. Aubrey Burl, in a careful discussion, calculated that the odds against the stone coming to lie on the solstitial alignment by a chance fall are about one in 165, and concluded that it was placed flat on purpose (Burl 2006). My own contribution (Daw 2015; 2025), building on Gordon Freeman’s work, was to notice that the Altar Stone is not merely lying along an alignment but shares a specific skew — about 80° from the main midsummer-sunrise axis — with the tallest trilithon (Stones 55 and 56) and the bluestones beside it, the whole group encoding a secondary solstitial axis from midwinter sunrise to midsummer sunset. A stone that participates in that designed geometry was not dropped there by accident. Parker Pearson and colleagues (2024) have independently come to the same view: the Altar Stone was set deliberately as a recumbent, and its solstice-aligned position in front of the Great Trilithon is the work of design, not collapse.

Once the deliberate placement is granted, the resemblance to the Scottish circles is hard to ignore, and Parker Pearson’s team make it explicit: a great stone, laid flat, in the south-western arc of the monument, at the foot of its two tallest uprights, “recalls the plans of recumbent stone circles of north-east Scotland” (Parker Pearson et al. 2024). The recumbent and its flankers at, say, Old Keig, and the Altar Stone beneath the towering Great Trilithon, occupy structurally similar places: the focal, southern, low-stone-between-tall-stones position around which the whole monument is organised.

It would be easy to overstate this, and rigour demands that the differences be set out as plainly as the likeness. A recumbent stone circle is a ring of graded uprights with two tall flankers immediately bracketing the recumbent; Stonehenge has no such graded ring, and no flankers in the Aberdeenshire sense — the Great Trilithon is a lintelled pair, an architectural form entirely absent from the Scottish tradition. The alignments differ in kind: the recumbents frame the low southern moon (or, on the cautious reading, the midwinter sun) across a level sill, whereas the Altar Stone lies along the solar solstitial axis of a monument famous for it. And the engineering similarities that so struck Kilbride-Jones — pointed bases, levered erection, timber buttresses — are, on his own argument, generic insular practice, the kind of thing competent stone-raisers anywhere in Britain might arrive at independently. The Altar Stone parallel is genuine and suggestive; it is not, by itself, proof that the builders of Stonehenge had a recumbent stone circle in mind.

What the provenance does and does not tell us

The geology is now the firmest part of the argument, and also the most often over-read. Clarke and colleagues (2024) fingerprinted the Altar Stone from the ages and chemistry of its detrital zircon, apatite and rutile grains, and matched it to the Old Red Sandstone of the Orcadian Basin — a match they tested against, and which excluded, the rest of Britain and Ireland. The implied journey is at least 750 km, and because hauling a six-tonne block that far overland across the grain of the country is so forbidding, transport by sea is the likeliest option (Clarke et al. 2024). A follow-up study, applying the same methods to Orcadian rocks and to the monoliths of the Stenness and Brodgar circles, ruled out Mainland Orkney specifically, pointing instead to the Scottish mainland (Bevins et al. 2024).

The Orcadian Basin is large. As Parker Pearson and colleagues (2024) note, it runs from Orkney and Shetland down to Inverness and eastward along the coast to Banff, Turriff and Rhynie. The last two names matter here: Turriff and Rhynie lie in the heart of recumbent-stone-circle country. So the provenance is at least consistent with a source in or near the region of the recumbent circles — but it does no more than that. It identifies a geological formation, not a quarry, still less a particular monument or a particular community of builders. The precise source remains unestablished, and the authors of the provenance study were careful to say so. To move from “the rock is from the Orcadian Basin” to “the Altar Stone was taken from an Aberdeenshire stone circle” is to add a great deal that the science does not supply.

What the provenance does establish is connection at a distance, and that is not trivial. A six-tonne stone did not travel several hundred miles by accident. Whether it came as raw rock, as a stone already standing (or lying) in a Scottish monument and later dismantled, or as a gift sealing some alliance — all of which have been suggested (Parker Pearson et al. 2024) — it testifies to sustained, organised, long-range contact between northern Scotland and Salisbury Plain in the centuries around 2500 BC. That contact is corroborated from a quite different direction by the close similarities between the timber house floors at Durrington Walls, beside Stonehenge, and those of Late Neolithic Orkney (Parker Pearson et al. 2024). The Altar Stone is one thread in a wider weave of northern connection.

Discussion: a shared repertoire, not a simple borrowing

Set side by side, the 1934 reports and the 2024 science point in the same broad direction without quite meeting. From Old Keig we learn that the builders of a recumbent circle deliberately shaped their stones, understood how to stand a heavy monolith up with the least effort, bedded and packed their stones with skid-stones and timber stays, and organised the whole monument around a recumbent stone and a cremation burial. From Stonehenge we learn that its central recumbent slab was placed deliberately, in a structurally comparable position, and that the rock for it came from the very region where recumbent circles are found. Kilbride-Jones, looking only at the stones, already thought the two traditions shared a method; Parker Pearson’s team, looking at position and provenance, think they shared an idea.

The cautious synthesis is that both are glimpses of a connected Neolithic Britain with a common stock of practices and symbols — the worked monolith, the economical erection, the meaningful recumbent stone — drawn upon differently in different places. That reading honours the genuine parallels without inventing a migration of architects from Aberdeenshire to Wessex (or, in Kilbride-Jones’s version, the reverse). It also fits the present chronology, in which the sarsen Stonehenge and the recumbent circles are broadly contemporary rather than parent and child.

Ninety years ago, two men crouched over the upturned base of a stone at Old Keig and saw in it the same problem the builders of Stonehenge had solved. They could not have known that the central stone of the monument they were comparing it to had itself come from their own corner of Scotland. That is the kind of coincidence that ought to make us look again — carefully — at what they wrote.

References

Atkinson, R. J. C. 1956. Stonehenge. London: Hamish Hamilton.

Bevins, R. E., Pearce, N. J. G., Hillier, S., Pirrie, D., Ixer, R. A., Andò, S., Barbarano, M., Power, M. & Turner, P. 2024. ‘Was the Stonehenge Altar Stone from Orkney? Investigating the mineralogy and geochemistry of Orcadian Old Red Sandstones and Neolithic circle monuments.’ Journal of Archaeological Science: Reports 58: 104738.

Bradley, R. 2005. The Moon and the Bonfire: An Investigation of Three Stone Circles in North-East Scotland. Edinburgh: Society of Antiquaries of Scotland.

Burl, A. 2000. The Stone Circles of Britain, Ireland and Brittany. New Haven & London: Yale University Press.

Burl, A. 2006. A Brief History of Stonehenge. London: Robinson.

Childe, V. G. 1933. ‘Trial Excavations at the Old Keig Stone Circle, Aberdeenshire.’ Proceedings of the Society of Antiquaries of Scotland 67: 37–53.

Childe, V. G. 1934. ‘Final Report on the Excavation of the Stone Circle at Old Keig, Aberdeenshire.’ Proceedings of the Society of Antiquaries of Scotland 68: 372–393.

Clarke, A. J. I., Kirkland, C. L., Bevins, R. E., Pearce, N. J. G., Glorie, S. & Ixer, R. A. 2024. ‘A Scottish provenance for the Altar Stone of Stonehenge.’ Nature 632 (8025): 570–575. doi:10.1038/s41586-024-07652-1.

Daw, T. 2015. ‘The “Twisted Trilithon”: a secondary solstitial axis at Stonehenge.’ Wiltshire Archaeological and Natural History Magazine 108. 15-24 

Daw, T. 2025. ‘The Geometry of the Horseshoe Stones at Stonehenge: Design, Alignment, and Cultural Significance.’ sarsen.org. doi:10.13140/RG.2.2.23505.65127.

Gowland, W. 1902. ‘Recent Excavations at Stonehenge.’ Archaeologia 58: 37–118.

Hawley, W. 1921. ‘Stonehenge: Interim Report on the Exploration.’ The Antiquaries Journal 1: 19–41.

Hill, J. 2021. The Recumbent Stone Circles of Aberdeenshire: Archaeology, Design, Astronomy and Methods. Newcastle upon Tyne: Cambridge Scholars Publishing.

Kilbride-Jones, H. E. 1934. ‘Stone Circles: A New Theory of the Erection of the Monoliths.’ Proceedings of the Society of Antiquaries of Scotland 68: 81–96.

Parker Pearson, M., Bevins, R., Bradley, R., Ixer, R., Pearce, N. & Richards, C. 2024. ‘Stonehenge and its Altar Stone: the significance of distant stone sources.’ Archaeology International 27 (1): 113–137. doi:10.14324/ai.27.1.13.

Tumuli Wiltunenses : a guide to the barrows on the plains of Stonehenge by Sir Richard Colt Hoare

 

Source:  https://hdl.handle.net/2027/uc1.a0003777299 from the HathiTrust

The Absence of Altar Stone Fragments in the Aubrey Holes at Stonehenge: Statistical Rarity or Evidence of a Late Arrival?

 

The Aubrey Hole Excavations

The Question

Why are there no identifiable fragments of the Altar Stone (Stone 80, a micaceous Old Red Sandstone) in the material recovered from the Aubrey Holes, despite the presence of representatives from virtually all other major and minor bluestone lithological groups? Is this absence meaningful—potentially indicating that the Altar Stone arrived later in Stonehenge’s construction sequence—or is it simply a statistical oddity resulting from the stone’s overall extreme rarity in the debitage record?

This question is worth asking because the Aubrey Holes represent one of the earliest structural features at Stonehenge (c. 3000 BC), and their fills provide a key window into the initial phases of bluestone activity. Understanding the timing and integration of individual stones, especially outliers like the Altar Stone, helps refine models of the monument’s multi-stage development and the logistics of long-distance stone transport.

Background and Why the Question Arises

The Aubrey Holes (56 pits forming a circle inside the bank and ditch) were excavated primarily by William Hawley and R.S. Newall in the 1920s, with a re-excavation of Aubrey Hole 7 in 2008. They have long been central to debates about whether they originally held bluestones or wooden posts.

Two important papers by Rob Ixer, Richard Bevins, and colleagues provide detailed modern petrographic re-analysis of historic collections:

• “And the first shall be last: the Aubrey Holes and their stones” (Ixer & Bevins, Wiltshire Archaeological & Natural History Magazine 114, 2021) examines ~46 thin sections from the 1920s Aubrey Hole material (primarily Holes 1–22 and 30) plus 2008 AH7 samples.
• “William Cunnington’s 1884 Stonehenge lithologies revisited” (Ixer, Bevins, Pirrie, Power & Pearce, WANHM 119, 2026) re-examines 33 thin sections from Cunnington’s late 19th-century surface and small excavation collections, providing a “pristine” pre-20th-century baseline.

Both papers update older material into the modern Ixer/Bevins bluestone classification scheme and confirm a restricted suite of Welsh-derived lithologies with no exotic glacial erratics.

The Evidence: Comparative Lithology

Table 1: Presence of Key Lithological Groups

Lithology Group	Aubrey Holes (2021)	Cunnington (2026)	Notes
Dolerites (spotted/unspotted)	Yes	Yes (incl. orthostats 32, 49, 61a)	Common in both.
Rhyolite Group C (Craig Rhos-y-felin)	Yes (dominant)	Yes	Ubiquitous.
Andesite/Volcanic Group A	Yes	Yes (type material from Stone 32c)	Cunnington defines the group.
Dacite/Volcanic Group B (SH38)	Yes (limited)	Yes (incl. from SH38)	Rare but present.
Rhyolite/Dacite Group D	Yes (single sample)	Yes (S57)	Both affirm genuine bluestone status.
Rhyolite Group E (SH48)	Yes	No	Absent in earlier collection.
Lower Palaeozoic Sandstone	Yes	Yes	Consistent.
Altar Stone (ORS sandstone)	No	Yes (S45)	Explicit absence in Aubrey Holes.
Greensand (packing)	Yes	Yes	Packing material.
Sarsen	Yes (common)	Absent in thin sections	Local.

The Aubrey Holes collection is comprehensive for bluestone groups except the Altar Stone, which is explicitly noted as absent.

Altar Stone Samples in Museum Collections

Verified or highly probable Altar Stone fragments are exceptionally rare across all museum holdings, reinforcing the pattern of scarcity. Many older labelled pieces have been re-evaluated and rejected through modern petrography, pXRF, and automated mineralogy.

Key examples include:

• Salisbury Museum sample 2010K 240 (also referenced as Wilts 277): Collected directly from the underside of the Altar Stone in 1844 by Mr Browne of Amesbury. This is the primary reference “type” sample, used extensively in recent provenancing (Scottish Orcadian Basin origin) and thin-section studies.
• Hawley/Newall 1920s excavated pieces (e.g., MS1–MS3; held in National Museum of Wales and related collections): Small fragments from excavations near Stone 1 or associated contexts; some authenticated via matching mineralogy/chemistry.
• Darvill & Wainwright 2008 excavation fragments (e.g., SH08 Context 1 FN196): Three probable pieces from within the Stonehenge Circle, authenticated by pXRF and petrography.
• Cunnington collection S45 (Wiltshire/Salisbury Museums): One sample from surface finds or small excavations in the Bluestone Circle area, identified in the 2026 paper.

Overall, Altar Stone debitage consists of tiny, lightweight pieces. No large concentrations exist, and many historic “Altar Stone” labels in museums have been disproven as other sandstones or tuffs.

Refined Provenance and Transport (Clarke et al. 2026)

New detrital zircon analysis strengthens a source in mainland northeast Scotland (Caithness/Sarclet area provides the closest match). Glacial modelling (Late Devensian British-Irish Ice Sheet) shows possible transport to Dogger Bank (~400 km from Stonehenge), but not directly to Salisbury Plain. Dogger Bank was exposed until ~8–7 ka BP, millennia before the Altar Stone’s likely emplacement at Stonehenge.

Clarke et al. float the idea that the Altar Stone may have had a convoluted route: glacial transport to Dogger Bank followed by later human hauling to Stonehenge. However, this would require an early arrival (pre-dating Doggerland flooding) or yet another undocumented stop along the route. There is no supporting evidence in the debitage record (including the Aubrey Holes or other early contexts) for such an early presence of Altar Stone material at or near Stonehenge.

Discussion: Rarity vs. Late Arrival

The Altar Stone is anomalous: it is the largest “bluestone” (~6 tonnes), recumbent, sourced from north-east Scotland (hundreds of km further than the Welsh bluestones), and possesses distinctive baryte cement. Its position (under a fallen sarsen from the Great Trilithon) and potential misalignment with the main bluestone phase invite questions about its integration timing.

Arguments for statistical oddity:

• Altar Stone debitage is demonstrably rare across all contexts, including museum collections and excavations. Even substantial samples (such as the Aubrey Holes material) can easily miss it by chance.
• Cunnington’s surface-focused collection captured one example, but later excavated pit fills (with different taphonomic and sampling biases, including Newall’s preferences) did not. The museum record shows the same low frequency everywhere.
• No Altar Stone in most debitage assemblages; its scarcity is a consistent pattern, not unique to the Aubrey Holes.

Arguments for possible late arrival:

• The Aubrey Holes are among the earliest features (~3000 BC). If the Altar Stone arrived later (possibly during sarsen phase rearrangements, c. 2500 BC or after), its fragments would not be expected in primary early fills.
• Its Scottish source suggests a separate long-distance transport event, potentially distinct from the main Welsh bluestone group.
• Its unique size, form, and position support the idea that it was handled differently or introduced at a different stage.

Conclusion

The absence of Altar Stone fragments in the Aubrey Holes is not strongly diagnostic of a late arrival, primarily because of the stone’s documented extreme rarity in the overall debitage record and museum collections. It is most parsimoniously explained as a statistical oddity given sampling limitations and the low frequency of Altar Stone material site-wide.

However, when combined with its anomalous Scottish provenance, large size, recumbent position, and potential chronological misalignment with the main bluestone phase, a later arrival or separate handling (including no support for a convoluted Doggerland route) remains a plausible and elegant explanation. The two papers together reinforce that Stonehenge’s “foreign” stones represent a limited, human-transported suite without glacial erratics, while highlighting the value of continued petrographic re-analysis of historic collections.

References

• Ixer, R.A. & Bevins, R.E. (2021). “And the first shall be last: the Aubrey Holes and their stones.” Wiltshire Archaeological & Natural History Magazine 114, 1–17.
• Ixer, R., Bevins, R., Pirrie, D., Power, M. & Pearce, N.J.G. (2026). “William Cunnington’s 1884 Stonehenge lithologies revisited.” Wiltshire Archaeological & Natural History Magazine 119, 1–19.
• Clarke, A. J. I., Veness, R. L. J., Kirkland, C. L., Clark, C. D., Gandy, N., Emery, A. et al. (2026) From Highlands to Henge: Refining the Provenance and Transport Pathways of Stonehenge's Altar Stone. Journal of Quaternary Science, 1–8. https://doi.org/10.1002/jqs.70080
• Bevins, R.E. et al. (2020–2025). Various papers on Altar Stone provenancing (e.g., Journal of Archaeological Science: Reports).
• Cleal, R.M.J., Walker, K.E. & Montague, R. (1995). Stonehenge in its Landscape: Twentieth-Century Excavations. English Heritage.

Stonehenge Rocks: Exotic But Not Erratic

This week the Altar Stone is back in the news. A new paper in the Journal of Quaternary Science (Clarke et al., 4 June 2026, “From Highlands to Henge”) refines its source within north-east Scotland and models ice-sheet flow to test whether a glacier, rather than people, could have carried a six-tonne slab some 700 km south. It is a good paper and a good story. But a much quieter article in the latest Wiltshire Archaeological and Natural History Magazine may have more to say about the question sitting underneath all the headlines: were the Stonehenge stones brought by people, or dumped by ice?

That paper is Ixer, Bevins, Pirrie, Power & Pearce, “William Cunnington’s 1884 Stonehenge lithologies revisited” (WANHM 119, 2026, pp. 1–19). Its starting point is an archival rediscovery: in August 2021, 33 of the thin sections William Cunnington had “cut for the microscope” in the 1880s turned up in Devizes Museum, with their hand specimens catalogued separately in Salisbury. Cunnington had described some 460 fragments in 1884 — the first serious classification of Stonehenge’s “foreign” stones — and these slides are the type and reference material behind that work.

Here is the point that is easy to miss. Cunnington collected the last of the loose debitage from the Stonehenge surface before the twentieth century got to work on the landscape — before the military camps, the roads and railways, the car parks and the clinker-gravel paths. His collection is, in the authors’ phrase, a pristine lithological baseline: a census of what was actually lying on Salisbury Plain before any modern contamination.

And it contains no surprises. Every major bluestone debitage group is present (bar one rhyolite), in the same proportions as every collection made since — dominated by dolerite, Andesite Group A and the Craig Rhos-y-Felin rhyolites, with smaller amounts of Lower Palaeozoic Sandstone and a little Altar Stone. What it does not contain is a single exotic glacial erratic. As the authors put it:

“The plethora of exotic erratic lithologies hoped for by others is totally absent.” — Ixer et al. 2026

That is the whole glacial-transport hypothesis, tested against thousands of fragments from the cleanest sample we are ever likely to have, and coming back empty. If ice had carried a random scatter of Welsh and other rock to Salisbury Plain, this is precisely the assemblage in which we should expect to see it. We don’t.

The paper is not only negative evidence. The rediscovered slides yield the first proper petrographic descriptions of two buried stumps, Stones 32c and 61a; additional material from Stones 32 and 38; the type section for Andesite Group A (from 32c); and confirmation that the rare Dacite Group D really is bluestone debitage rather than a stray. For anyone following the orthostat-by-orthostat work, that is a quietly substantial haul.

The contrast with later material makes the baseline point sharply. The genuinely exotic rocks in the Stonehenge landscape — analcime dolerite, “Markfieldite”, Carboniferous limestone — turn out to be twentieth-century roadstone from Midland and Mendip quarries: industrial litter, not Ice Age cargo. Without Cunnington’s clean snapshot it would be far harder to tell the two apart. With it, the “erratics” dissolve into modern gravel.

So while the Altar Stone makes the front pages — and, for what it’s worth, even the new Altar Stone paper weighs glacial against human transport, finds the glacial route hard to sustain on the modelling, and cites fresh zircon-and-apatite work that found no glacial signature in Salisbury Plain river sediment — the Cunnington study does something less glamorous and arguably more decisive. It establishes, from the last undisturbed evidence, that the stones people keep hoping the glaciers brought simply aren’t there. One spectacular stone from Scotland is a wonderful puzzle. A whole assemblage with no erratics in it is, I’d suggest, the more important result.

---

References
Ixer, R.A., Bevins, R.E., Pirrie, D., Power, M. & Pearce, N.J.G. 2026. William Cunnington’s 1884 Stonehenge lithologies revisited. Wiltshire Archaeological and Natural History Magazine 119, 1–19. https://www.academia.edu/167889244/Cunningtons_thin_sections_and_rocks_revisited
Clarke, A.J.I. et al. 2026. From Highlands to Henge: refining the provenance and transport pathways of Stonehenge’s Altar Stone. Journal of Quaternary Science. https://doi.org/10.1002/jqs.70080

From Highlands to Henge – No Glaciers Needed

The new Clarke et al. (2026)* paper in the Journal of Quaternary Science refines the Altar Stone’s likely source to Caithness in northeast Scotland. It tests whether ice could have helped move it south. The science itself is careful and measured: no direct glacial pathway to Stonehenge, major chronological and evidential problems with the Dogger Bank idea, and glacial pathways are merely models. Yet the media exploded with “glaciers moved it.”

The Headlines That Did the Damage

• “Stonehenge altar ‘travelled down on glacier from Scotland’” — The Times
• “Stonehenge altar may have travelled from Scotland by glacier” — The Telegraph
• “Stonehenge mystery: Altar stone moved by glacier and ancient rescue mission” — ITV News
• “Stonehenge altar stone may have travelled from Scotland via glacier” — PA / Daily Mail and many others

Media Spin in Action: Even the Daily Mail Played Both Sides

On 4 June 2026 the Daily Mail ran two contrasting stories on the same paper. The PA syndicated version used the clickbait glacial headline, while their own Science desk correctly led with:

“The grains of sand that solve Stonehenge mystery after 5,000 years: Scientists uncover new evidence key stone was moved hundreds of miles by HUMANS – and not glaciers.”

This shows how the same study can be framed in completely different ways depending on who writes the headline.

Why Journalists Love the Lazy Hook

“Glacier moves mysterious Stonehenge stone” is simple, dramatic, and mysterious. It generates clicks. The far more impressive reality — that Neolithic people organised the long-distance transport of a carefully dressed 6-tonne slab — requires context and nuance, so it loses out.

How It Becomes Folk Memory

In six months, most people who saw coverage of this paper will remember one “fact”: glaciers brought the Altar Stone. The careful qualifications, the lack of erratic features on the stone, the companion fingerprinting paper showing no glacial detritus on Salisbury Plain, and the authors’ own conclusion that substantial human transport was still required will largely be forgotten.

Where the Real Problem Lies

The fault lies mainly with the university press office (in this case likely Curtin University’s), which writes the initial press release and chooses the most attention-grabbing angle to maximise media pickup. Authors review these releases, but early-career researchers are often not in a strong position to push back against professional communicators whose job is to generate buzz. In this instance, the speculative Dogger Bank scenario was given too much prominence in the publicity materials, even though the paper itself is cautious. This is a systemic issue in modern academia: funding, citations, and reputation all reward sensational framing.

What Should Have Happened

The press release and abstract should have led strongly with the positive findings (refined Caithness provenance and reinforcement of Neolithic capability) and framed the glacial modelling as a ruled-out hypothesis rather than a highlighted “possibility.” They should also have explicitly highlighted the Altar Stone’s sharp, worked appearance as clear evidence against it being a typical glacial erratic.

Bottom line: The core science in this paper is useful. It tightens the provenance and reinforces that Neolithic people did the hard work — no glaciers needed. The media distortion, while predictable, is unhelpful and unnecessary. Let’s give credit where it is due: to the remarkable organisational ability of our prehistoric ancestors.

*Clarke, A. J. I., Veness, R. L. J., Kirkland, C. L., Clark, C. D., Gandy, N., Emery, A. et al. (2026) From Highlands to Henge: Refining the Provenance and Transport Pathways of Stonehenge's Altar Stone. Journal of Quaternary Science, 1–8. https://doi.org/10.1002/jqs.70080

First Look at Clarke et al. (2026) – From Highlands to Henge

A refined provenancing study with a new statistical “best match”, but it still feels like identifying one possible needle in a very large and varied haystack.

Clarke, A. J. I., Veness, R. L. J., Kirkland, C. L., Clark, C. D., Gandy, N., Emery, A. et al. (2026) From Highlands to Henge: Refining the Provenance and Transport Pathways of Stonehenge's Altar Stone. Journal of Quaternary Science, 1–8. https://doi.org/10.1002/jqs.70080

The paper uses Kolmogorov–Smirnov tests and MDS plots on the limited published detrital zircon datasets and finds the strongest match at Sarclet in Caithness (p = 0.96), with other northern mainland sites (Braemore, Kirtomy, Portskerra) also statistically compatible. More southerly Orcadian Basin and Grampian outlier samples fare worse. This is a genuine step forward from the 2024 basin-level result, yet the authors themselves stress that the Orcadian Basin is enormous (up to 10,000 km²) and lithologically heterogeneous, with only sparse published zircon coverage. So while Caithness now looks the most likely area among the data we have, it remains one plausible candidate among many unsampled possibilities.

On transport, the paper dismisses any direct Late Devensian glacial delivery to Salisbury Plain. Using time-integrated ice-flow modelling (Veness et al. 2025 method on BRITICE-CHRONO geometries), the favoured Caithness sources mostly disperse north and east; only specific sensitivity scenarios (or slightly more southerly seeds) allow material to reach as far as Dogger Bank.

The authors float an intriguing hybrid idea: glacial transport could have delivered the stone part-way to Dogger Bank (~400 km from Stonehenge), after which humans might have collected it and moved it the remaining distance, perhaps while the area was still accessible or as post-glacial sea-level rise began to inundate Doggerland. They note Dogger Bank was exposed dry land with moraines until roughly 8–7 ka BP.

However, several problems stand out even on a first reading:

• No supporting evidence on the ground (or seabed): No other Altar Stone-like erratics or matching detrital signatures have been reported from Dogger Bank or the surrounding North Sea floor in the right context. The companion Clarke & Kirkland (2026) fingerprinting study found no glacial detrital zircon–apatite signal from northeast Scotland in Salisbury Plain sediments at all.
• Timing difficulties: Any Dogger Bank staging post would require the stone to be removed and transported onward (or cached) across a gap of several millennia before erection at Stonehenge. The paper acknowledges this creates a complicated multi-stage history rather than a neat solution.
• Modelling limitations: Ice-flow reconstructions are models, not direct observations. They have known issues reconciling with the actual geological record of erratic distributions in places, and the present simulation has relatively coarse spatial (2.5 km) and temporal (1,000-year) resolution. The authors are open about the high sensitivity to exact starting position near the former ice divide and that only the Devensian can be modelled usefully here.

Conclusion Even in its most favourable reading, this paper does not rescue a primarily glacial explanation. It narrows the likely source area statistically while showing that direct glacial transport to Wessex is not supported, and that even a partial glacial assist to Dogger Bank still leaves substantial anthropogenic transport necessary — plus extra chronological and logistical complications. The absence of corroborating erratics or detrital evidence further weakens the hybrid staging-post scenario.

On balance, the work inadvertently strengthens the case that Neolithic people organised and executed the long-distance movement of this 6-tonne block, whether from Caithness directly or via some intermediate location. It is a careful, data-driven paper, but it does not move the glacial needle very far. More targeted zircon sampling from the Caithness area and higher-resolution modelling (or new seabed data) would be needed before any Dogger Bank idea could be taken seriously.