Dropping the Central Trilithon: a physics toy

 

Dropping the Central Trilithon: a physics toy for Stonehenge's oldest cold case

Open it here

Nobody recorded the tallest stones at Stonehenge coming down. Stone 55 of the Central Trilithon — the pair that once framed the winter-solstice sunset, and which most books will insist on calling the "Great" Trilithon (a friend who knows these stones far better than I do assures me the name properly belongs elsewhere, and I've learned not to argue with him) — fell and snapped in two at some unrecorded moment in antiquity, taking lintel 156 with it. By the time anyone drew the monument reliably, the wreckage was already lying where it lies today: 55a, 55b and the lintel sprawled across the recumbent Altar Stone, pinning it to the ground. Its partner, stone 56, leaned on alone until it was hauled upright and set in concrete in 1901.

That wreckage poses a question archaeologists still argue about: was the Altar Stone standing when the trilithon fell on it — or was it already lying flat? The Altar Stone has never been fully excavated, partly because the fallen sarsen pieces sit on top of it. So the resting positions of four stones are nearly all the evidence there is.

Which makes it a perfect problem for a physics sandbox. If every hypothesis has to end with the stones where they actually are, you can throw hypotheses at a rigid-body simulator all day and see which ones survive.

The toy

The simulation below (or [here], if the embed misbehaves) is a 2D cross-section along the solstice axis, built on the Matter.js physics engine at true scale — the uprights are 6.7 m proud of the ground, the lintel a metre-thick slab perched at over 7 m. Stone 56 is fixed, as the evidence demands. Everything about stone 55 is a slider:

• how far it leaned before letting go, and how sharp the subsidence "kick" was;
• how deep its socket was — 56's was measured at a famous 2.4 m in 1901; 55's is thought to have been embarrassingly shallow, which is presumably why it's the one that fell;
• how strong the sarsen was (how hard an impact snaps it in two);
• how tenaciously the mortise-and-tenon joints dragged the lintel along before letting go;
• and, in the standing scenario, how firmly the Altar Stone was bedded.

Dashed outlines mark where the pieces lie today. After every run, a verdict card checks the outcome against them, and a "trench notebook" narrates the collapse: the base kicking out of its socket, the tenons shearing, the fracture, the strike. Three preset buttons reproduce the runs discussed below, so you don't have to take my word for any of it.

What the model says

The lazy expectation is that one scenario matches and the other doesn't. That's not what happens, and the way it fails is the interesting part.

Fall onto an already-recumbent Altar Stone, and everything works on nearly the first try. With middle-of-the-road parameters, stone 55 tips, its upper end strikes the raised edge of the flat slab at around 4 m/s, and it snaps over that edge like a bar over a fulcrum — the butt end flat at about 2 m, the upper half flat at about 5 m across the altar, the lintel coming to rest on top of the pile. That is, uncannily, the arrangement in the ground.

Fall onto a standing Altar Stone, and the knock-over itself is easy — almost too easy. A ~30-tonne slab sweeping through its arc fells a half-metre-thick pillar without much argument. But here's the result I didn't see coming: the collision absorbs so much of the fall's energy that stone 55 tends to land in one piece. The standing altar acts as a crumple zone. And an unbroken 55 contradicts the most solid fact we have — the real stone broke. To get the full observed sequence in the standing scenario (altar felled flat and 55 snapped), I have to dial the sarsen strength down to a conveniently flawed stone. Push the other way — bed the altar firmly — and you get a third outcome the ground flatly rules out: a permanent stalemate, with 55 propped against a still-standing Altar Stone like a failed game of dominoes.

So the model can't prove which history happened; with enough slider-turning, both scenarios reproduce today's arrangement, and that underdetermination is honestly the deepest lesson in it. But the two scenarios are not on equal footing:

Falling onto a recumbent Altar Stone breaks stone 55 for free — the slab's edge is exactly the anvil the fracture needs. Felling a standing Altar Stone instead cushions the fall, and the break has to be bought with an extra assumption about a weak stone.

Call it an argument from parsimony, delivered by a physics engine. It lines up with where much of the archaeology has been drifting anyway: many researchers suspect the Altar Stone was placed recumbent by design, a threshold rather than a pillar. The simulation adds a small, mechanical voice to that side of the debate: the flat-altar story needs nothing special to be true.

The Tenon Problem and an Insight

A real difficulty is that the mortise-and-tenon joints at Stonehenge were remarkably tenacious. Elsewhere on the monument we can see the evidence: stones 6 and 7 of the outer circle, one leaning out and the other in, still carried their twisted lintel until a scaffold was put up in 1881; observers at the time noted that without the joints that section would have collapsed long before. Stone 56 of the Central Trilithon itself stood for centuries at a lean of something like 12–15° after its partner and lintel had already fallen. A simple model that treats the joints as springs, or even as a clean kinematic coupling that releases after a few degrees of lean, cannot reproduce this. The joints were “sticky”; they held under conditions that look precarious to us.

What made the Central Trilithon different was almost certainly not a sudden failure of the tenons at the top, but the progressive failure of the shallow socket at the base of stone 55 — the heel being pushed or washed out. Once that footing gave way, the upright was free to rotate past the point of no return while the still-engaged joints dragged the lintel with it. Our 2-D simulations are limited because they cannot properly separate those two mechanisms: a base that can soften and release, and an upper joint that remains tenacious until the geometry finally forces it apart.

Even so, the simplified model yields two clear archaeological implications. First, trilithons fail from the base, not from the top; it is the heel coming out that allows the structure to fall. Second, the observed breakage of stone 55 is far more consistent with an already-recumbent Altar Stone acting as an impact surface than with a still-standing one. More elaborate modelling of the joints can wait; these two results are already visible.

Caveats, cheerfully admitted

This is a toy, and it wears its assumptions on its sleeve. It's 2D, so both uprights live in the same cross-section and the lintel's three-dimensional tumble is collapsed onto a plane. Fracture is a threshold on impact velocity at the contact point, not real crack mechanics; the tenon coupling — genuinely unknowable — is a parametrized "how long did the joints hang on" knob rather than simulated joinery. The stones are rigid rectangles on level ground. None of the slider ranges pretend to precision; they pretend to span the plausible, which is all a hypothesis machine needs.

The right way to read it: not "this is how it happened," but "here is the space of ways it could have happened, and notice which corner of that space doesn't need a coincidence."

Go push the stones over yourself. Start with the "already recumbent" preset, then try to make the standing scenario work without weakening the stone. That failure is the most informative thing in the whole toy.

---

Built with Matter.js. The trilithon geometry follows the standard published dimensions; socket depths per Gowland's 1901 excavation of stone 56. Nomenclature corrected under expert duress. All remaining errors of physics and prehistory are mine.

The A303 Drive-By Guide to Stonehenge

For the motorist doing 4 mph past a World Heritage Site

You are stuck in traffic on the A303. You know this because you are reading a blog post on your phone, which I trust means you are the passenger. Somewhere ahead, beyond a Eurobox towing a caravan, is Cornwall. Somewhere to your right, rather sooner, is Stonehenge.

The official advice is to visit the visitor centre, park a mile and a half from the stones, and take a shuttle bus. This is excellent advice for people with a spare half-day and a healthy tolerance for other people's children. You, however, have a holiday cottage to get to, and the geometry of the situation is entirely in your favour: the A303 passes within about 165 metres of the monument, considerably closer than the visitor centre does, and your current speed is ideal for detailed observation. What English Heritage sells as a premium experience, the Highways Agency provides free at the point of delivery.

So here is what you are actually looking at, in order, westbound from Amesbury. Eastbound readers should hold the guide upside down.

Countess Roundabout

The jam traditionally begins here, so consider it the ticket barrier. As you crawl off the roundabout and up the hill, the wooded lump immediately on your left is Vespasian's Camp — an Iron Age hillfort that has nothing whatever to do with Vespasian, who never went near it. The name is antiquarian enthusiasm, not history. At the foot of the hill, out of sight by the springs, is Blick Mead, a Mesolithic camp where people were feasting on aurochs thousands of years before anyone at Stonehenge lifted anything heavier than an opinion. You cannot see it. Very few people can. But you have now driven past the oldest bit of the story, which is more than most visitors manage.

The King Barrows

As the road climbs, look right at the clumps of beech trees along the ridge. Those trees are planted on the New King Barrows — a cemetery of Bronze Age round barrows, each one the burial mound of somebody who mattered around 4,000 years ago. The beeches are an eighteenth-century landscaping decision, which tells you the barrows have been furniture in a gentleman's view for longer than the United States has existed.

Stonehenge Bottom

The road then drops into a dry valley. Somewhere in the grass here, invisible from the car, the Stonehenge Avenue crosses it— the earthwork processional route that once linked the monument to the River Avon. Prehistoric people approached Stonehenge along it on foot, in ceremony. You crossed as you climbed the hill in an MPV, which is at least a novel form of procession.

The Main Event

And there they are, on your right, on the skyline and then broadside on. A few points the shuttle bus commentary won't make:

You are looking at the back. Stonehenge has a front — the north-east façade, with its neat run of lintelled sarsens, faces the midsummer sunrise and, conveniently for English Heritage, the paying customers. The A303 shows you the south and south-west side, which is the ruined side. Roughly half the monument is missing from this arc: stones fallen, broken up, carted off over the centuries for bridges, buildings and ballast. Whether that sector was ever properly finished at all was a live question until a dry summer in 2013, when parchmarks revealed the buried holes of missing circle stones — spotted, as it happens, because a hosepipe didn't reach far enough. The discoverers may be known to the management of this blog.

The tall one. The single upright towering above the rest is Stone 56, the surviving half of the Great Trilithon — the largest stones on the site. Its partner fell over long ago and lies broken. Stone 56 itself was leaning at an alarming angle until 1901, when Professor William Gowland winched it upright and set it in concrete. So the most imposing thing you can see from the road is, strictly speaking, an Edwardian restoration of a prehistoric monument. It is not the only one; a fair amount of what stands today was straightened and concreted in the twentieth century. This information is best deployed loudly at the summer solstice.

What you can't see. The Heel Stone, the famous outlier on the solstice axis, is on the far side of the monument by the line of the old A344 — closed in 2013 so that visitors could enjoy the stones without the sound of traffic. The sound of traffic was then supplied entirely by you.

The Left-Hand Side

While the driver gawps right, passengers should look left at the ridge to the south. That skyline is Normanton Down, one of the richest Bronze Age cemeteries in Europe. Among those mounds is Bush Barrow, whose occupant was buried with a sheet-gold lozenge of astonishing craftsmanship, now in the Wiltshire Museum in Devizes. Devizes has parking. Just saying.

Longbarrow Crossroads and Release

At the roundabout by Winterbourne Stoke you pass another barrow cemetery, anchored by a genuine Neolithic long barrow — several centuries older than anything standing at Stonehenge. Nobody photographs it. It has borne this with dignity for five and a half thousand years, and shortly afterwards the dual carriageway resumes and so, briefly, does your holiday.

A Note on Why You Are Stationary

You may reasonably ask why a single-carriageway trunk road still runs 165 metres from Britain's most famous monument. The answer is that for thirty years governments proposed to put the road in a tunnel, and in 2024 a government decided not to, after spending some £179 million on the scheme. In March 2026 the planning consent was formally revoked, so the not-building of the tunnel is now official and complete. You are, in a sense, driving through the result of the most expensive decision to change nothing in the history of British heritage. Do slow down and take it in. You will anyway.

---

Drivers: eyes forward, hands at ten and two. The stones have waited four and a half thousand years; they will still be there next jam.

Enhancing LIDAR with AI

 A simple test to see how effective copying Lidar images into AI engines could enhance them. Click any to enlarge them.

The prompt used was:

"Enhance this lidar view to bring out every detail you can, recolour in the yellow- brown spectrum"

The original from the EA LIDAR Composite Viewer

Grok

Gemini

ChatGPT (on second attempt after crash)

Meta AI - after second prompt asking it to increase contrast, it offered more options. 

The Altar Stone Source - The Algorithm Result

 

"A transparent, multi-proxy desk screen—barium–rubidium stream-sediment geochemistry (a proxy for the Altar Stone’s diagnostic baryte cement and K-feldspar deficit), thermal-maturity mapping, clay mineralogy, sedimentary facies, and detrital-zircon provenance—has been applied iteratively across the UK Old Red Sandstone (ORS). It has narrowed a national-scale problem, step by step and at the appropriate level of evidence, to a single field-accessible target. The screen and Clarke et al.’s (2024, 2026) detrital-zircon data agree that the source lies on the East Caithness coast; but the prime barium cluster (Sarclet–Lybster–Clyth) sits in the high-maturity zone mapped by Hillier & Marshall (1992), where vitrinite reflectance of 3–6% R₀ has driven the delicate expandable and aluminous clays of the Altar Stone past preservation.
Two independent lines then refine, rather than defeat, the result. The clay evidence (Hillier & Clayton 1989; Hillier et al. 2006) shows the Altar Stone’s tosudite–kaolinite–dioctahedral-chlorite assemblage to be an aluminous sandstone diagenetic pathway. While classically expressed in the UK Lower ORS, work on offshore Middle Devonian blocks (e.g., the Clair Group) demonstrates that this signature is strictly facies- and fluid-controlled rather than stratigraphically restricted, meaning it can be seamlessly accommodated within porous sandstone bodies encased in lower-maturity segments of the Caithness Flagstone Group.
One place is identified where such a rock could occur:"

I've included a photo of a megalithic monument that is there, but I need to just check it again before revealing the site.

The main problem is that the lack of data means that unsampled gaps in the record don't get filtered out, and that small areas are filtered out by being smeared in with the surrounding geology. Other data may highlight such areas, and suggest better screening. So such a desktop screening exercise can only suggest places worth further investigation, with a rock hammer.

The Myth of the Myth: Thomas Was Just Reviewing and Rejecting Judd — With Evidence

 

Claims have been circulating that the idea of human transport for Stonehenge’s bluestones was “invented” by geologist Herbert H. Thomas around 1920–1923. The story goes that, in the aftermath of the First World War, there was a national need for a feel-good narrative about heroic, highly skilled ancestors — and Thomas obligingly supplied one, while ignoring or suppressing the glacial transport views supposedly held by his fellow geologists.

This version is itself a myth.

The debate did not begin with Thomas. In 1902, geologist J.W. Judd published his thoughts on the foreign stones at Stonehenge. Judd proposed they were glacial erratics — boulders carried by ice and left on Salisbury Plain. He was struck by the variety of rock types and especially by the abundance of bluestone fragments around the monument. This, he argued, suggested the stones had been worked and dressed on site from pre-existing glacial deposits rather than being laboriously brought from afar. Judd acknowledged challenges with the known limits of glaciation but suggested earlier, more extensive ice action could explain it.

H.H. Thomas’s 1923 paper (“The Source of the Stones of Stonehenge”, Antiquaries Journal) did what scientists are supposed to do: he reviewed the existing hypothesis, applied new petrographic analysis, and tested it against the geological evidence available at the time.

Thomas’s key points were straightforward:

• He matched many of the bluestones — particularly the distinctive spotted dolerites — to specific outcrops in the Preseli Hills of Pembrokeshire.
• He assessed the glacial transport idea and found it implausible on geological grounds. There was no convincing evidence of the extensive glacial drift, boulder trains, or ice-scratched surfaces that would be expected if a glacier had carried large stones all the way to Wiltshire. In his view, the ice front did not extend far enough east or in the right way.
• Therefore, the only reasonable explanation for the stones’ presence at Stonehenge was deliberate human transport by Neolithic builders.

Thomas did not invent human transport as a patriotic fable. He arrived at it by examining Judd’s hypothesis and finding the geological evidence against it stronger. This is normal scientific process — propose, test, refine or reject on the basis of data — not myth-making or morale-boosting.

Note on the 1921 discussion: In the discussion following Hawley’s interim excavation report (published 1921), two contributors — Mr. Dale and Rev. G.H. Engleheart — supported Judd’s glacial erratics view and mentioned possible striations on fragments. Neither was a professional geologist (Dale appears as a contributor to antiquarian/geological discussions of the period; Engleheart was a clergyman and local antiquary). Thomas, as Petrographer to the Geological Survey, was the qualified geological voice engaging directly with Judd’s ideas. Claims of Thomas ignoring a consensus of “fellow geologists” do not hold up.

The notion of a post-WWI patriotic conspiracy or deliberate ignoring of colleagues doesn’t hold up. Thomas engaged directly with Judd’s arguments in a scholarly journal. His work was published in a scholarly journal and focused on facts from rock samples and field geology, not morale-boosting narratives.

Of course, science moves on. Later researchers like Kellaway (1971) revisited glacial possibilities with new data, and the debate continues today with advanced geochemical fingerprinting, LiDAR, and field studies. Recent work has strengthened the case for human transport while refining exact source locations (e.g., Carn Goedog, Craig Rhos-y-felin).

But the 1923 paper was not myth-making. It was Thomas doing the unglamorous work of reviewing a prior hypothesis (Judd’s) and rejecting it on evidence.

The real myth is the one that turns a careful piece of geological reasoning into a conspiracy of patriotic invention. The evidence shows something much more ordinary — and more interesting: a scientist looking at the rocks and following where they led.

Lithofacies review of the Nairn and southern Moray Firth Devonian Old Red Sandstone

Abstract

The national barium–rubidium screen and subsequent Orcadian prioritisation (Daw 2026) identified East Caithness (the Sarclet–Lybster–Clyth flagstone coast) as the strongest candidate ground for the Altar Stone, independently corroborated by the detrital-zircon match of Clarke et al. (2024, 2026). The only other coastal belt within the Orcadian Basin that returned any screen signal, and merited a stratigraphic check, is the Nairn–Findhorn–Elgin margin on the southern shore of the Moray Firth. A lithofacies audit of the published BGS mapping, memoirs and Geological Conservation Review accounts confirms that this belt is a marginal expression of the same Lake Orcadie system: the Middle Old Red Sandstone here (Inverness and Black Isle Sandstone groups) is predominantly fluvial red sandstone deposited along the lake margin, with subordinate lacustrine flag intervals, and the Upper Old Red Sandstone (Nairn Sandstone Formation and the Elgin beds) is a mixed fluvial–lacustrine, red-to-grey calcareous, cornstone-bearing succession. Its fine lacustrine flag intervals are the same facies family as the Caithness flagstones and carry the same Achanarras-assemblage fish. That last point is decisive in an unexpected direction: because the East Caithness lead is itself built on the fish-bearing quiet-water flag facies, the presence of fish beds cannot be used to exclude the Moray margin. We therefore correct the provisional “too many fossils” dismissal rather than endorse it. What genuinely separates the Moray margin from East Caithness is not the fine facies but its marginality and oxidation, the subordinate and heterogeneous development of its flags (poor monolith potential), and above all its weak, poorly bedrock-verified barium signal — the proxy for the Altar Stone’s diagnostic baryte cement — against the strongest national signal at East Caithness. The diagnostic baryte–tosudite assemblage is untested at Nairn, as it is still under validation at East Caithness. The Nairn strand is therefore recorded as deprioritised and untested rather than eliminated on lithofacies. The Permian–Triassic New Red Sandstone of the coast (Hopeman and Burghead sandstones) is excluded on age and requires no facies argument.

1.  Why the Moray margin was the remaining Orcadian candidate to check

Clarke et al. (2024, 2026) place the Altar Stone’s detrital-mineral source in the Orcadian Basin and exclude the Midland Valley, the Anglo-Welsh Basin and Mainland Orkney. The barium–rubidium stream-sediment screen of Daw (2026) ranked East Caithness — the Sarclet–Lybster–Clyth flagstone coast — as the strongest national hit, with independent zircon corroboration at Sarclet. Within the Orcadian Basin the only other coastal belt that returned any screen signal was the southern Moray Firth margin around Nairn, Findhorn and Elgin, although the signal was weak and, as Section 5 notes, poorly bedrock-verified. This belt exposes Middle and Upper Devonian rocks laid down in the southern, marginal part of Lake Orcadie. It was provisionally set aside on the informal observation that its fine facies “had too many fossils.” This note tests that dismissal formally, and finds it must be replaced: the fossil criterion does not survive scrutiny, but the belt is nonetheless a low-prior, untested candidate for other, sounder reasons.

2.  Data and methods

The audit uses published sources only: BGS 1:50 000 and 1:625 000 digital mapping; the BGS Earthwise accounts for the Devonian of the Grampian Highlands and the Northern Highlands; the BGS memoir for Fortrose and eastern Inverness (Sheet 84W); and the Geological Conservation Review volume on the Old Red Sandstone of Great Britain (including the Tynet Burn and related Moray fish-bed sites). No new field data were collected. Lithofacies were evaluated against the Altar Stone criteria established in the screening and mineralogical work (Bevins et al. 2024; Clarke et al. 2024, 2026; Daw 2026): fine- to very fine-grained, well-sorted sandstone; ripple or planar lamination indicating quiescent water; grey-green colour; negligible detrital K-feldspar; pervasive baryte (with calcite) cement; and a tosudite / aluminous-kaolinite clay assemblage. Two method limits are carried throughout: stream-sediment values are not rock values, and a barium anomaly is a proxy for baryte cement, not a measurement of it; and formation identity is not facies identity.

3.  Lithofacies of the Nairn–Moray Devonian succession

3.1  Middle Old Red Sandstone — the marginal facies of Lake Orcadie

Around the southern Moray Firth the Middle ORS is represented by the Inverness and Black Isle Sandstone groups, which BGS characterises as predominantly fluvial red sandstone successions deposited along the lake margins — the marginal counterpart of the deep-water Caithness Flagstone Group at the centre of the basin. Within this dominantly marginal, sandier and more oxidised succession, subordinate lacustrine flag intervals occur — the Inshes Flagstone, Nairnside and Hillhead sandstones — as grey and purple flaggy micaceous sandstones and dark calcareous flags with laminated shaly mudstones and limestone nodules. These finer intervals carry the Achanarras (and post-Achanarras Eifelian) fish assemblage, the same faunal marker that defines the lacustrine flagstones of Caithness. The Hillhead Sandstone, with its post-Achanarras fish, is unconformably overlain in the Ardersier–Cawdor area by the Nairn Sandstone Formation (Section 3.2).

3.2  Upper Old Red Sandstone — Nairn Sandstone Formation and the Elgin beds

The Nairn Sandstone Formation, the oldest Upper ORS unit of the district, comprises an irregular basal reddish conglomerate overlain by red, grey and yellow calcareous cross-bedded and flaggy sandstones with thin conglomerate beds and soft limestone-bearing mudstones; it is a mixed fluvial–lacustrine sequence carrying a Givetian fish assemblage. In the Findhorn area desiccated mudstones (clay galls) and a calcrete horizon (the Cothall Limestone) are recorded. The overlying Whitemire, Alves and Scaat Craig beds are grey-to-reddish siliceous pebbly sandstones and fine conglomerates with marly intervals and cornstone (calcrete) palaeosols, and the Upper ORS subdivisions here are themselves defined on six successive fossil-fish assemblages (Asterolepis, Psammolepis, Bothriolepis, Holoptychius and others). The succession records shallow-water and periodically dry-bed conditions, with rapid lateral facies changes and local overstep along the margins of fault-bounded sub-basins. Fine flaggy and shaly intervals are present but are subordinate, laterally impersistent, and interbedded with pebbly and pedogenically modified beds.

3.3  The New Red Sandstone of the coast is not Devonian

The Hopeman and Burghead sandstones of the Hopeman–Burghead–Lossiemouth coast are post-Devonian and rest unconformably on the Upper ORS. They are not a single unit and are not both aeolian: the Hopeman Sandstone Formation is a Late Permian to Early Triassic aeolian dune sandstone (the ‘Elgin Reptile’ beds, with Chelichnus trackways), while the overlying Burghead Sandstone Formation is Triassic and fluvial (waterlain), part of the New Red Sandstone. Their historical confusion with the Old Red Sandstone — the very controversy that the 1851 discovery of Leptopleuron (Telerpeton) at Spynie brought to a head — is a caution rather than a candidacy: on age alone neither is relevant to the Altar Stone, and no facies argument is required to set them aside.

4.  Comparison with the Altar Stone criteria

Assessed against the Altar Stone benchmark, the Moray margin divides into points of genuine similarity and points of genuine difference — with the single criterion the provisional dismissal relied upon, fossil content, belonging to neither.

●        Grain size, sorting and structures. The fine, well-sorted, ripple- and planar-laminated intervals of the Moray succession are the lacustrine flag intervals of Section 3.1–3.2. Texturally these are comparable to the East Caithness flagstones — they are the same quiet-water facies — but here they are subordinate to marginal fluvial red sandstone and are laterally impersistent.

●        Colour and composition. The succession is dominated by red, grey and yellow calcareous sandstones with cornstone palaeosols; a clean, grey-green, K-feldspar-poor sandstone of the Altar Stone type is not specifically reported. This oxidised, pedogenically modified, marginal character is a real point of difference from the reduced, deep-water grey flagstone facies — though the grey lacustrine flags of Section 3.1 show the reduced facies is locally present.

●        Monolith potential. The fine intervals are thin, impersistent and interbedded with coarser and nodule-rich beds — a poorer prospect for a coherent monolith of the required dimensions than the thick, laterally persistent Caithness flag sequences.

●        Diagnostic diagenesis (baryte, tosudite). No published clay or cement data exist for the Nairn–Moray flags. On the screen, the belt returned only a weak and poorly bedrock-verified barium signal — the proxy for the Altar Stone’s pervasive baryte cement — in contrast to the strong, well-verified signal at East Caithness. This is the most substantive point against the belt, and it is a proxy, not a measurement.

5.  Why the fossil criterion does not discriminate

The provisional dismissal rested on the observation that every fine lacustrine facies in the Moray belt is fish-bearing. That observation is correct but non-discriminating, because the East Caithness ground on which the whole enquiry rests is itself the fish-bearing flag system. The Lower Caithness Flagstone Group is built from the Clyth and Lybster subgroups above the Sarclet Group, with the Achanarras Fish Bed within it; the quiet-water fish-bed facies of the Clyth and Lybster subgroups are locally carbonate-rich, approaching dolomitic limestone; and the basin’s fish beds carry bituminous residues from oil generation. The Achanarras assemblage that marks the Moray fish beds is the same marker found at the Niandt Limestone of east Caithness, the Sandwick Fish Bed of Orkney and the Cromarty and Edderton fish beds of Easter Ross. The fish beds are the correlatable quiet-water phase of a single lacustrine system, not a property that distinguishes one part of it from another.

It follows that fossil content cannot exclude the Moray margin without also excluding the preferred East Caithness lead — an argument that proves too much. The Altar Stone is barren of macrofossils simply because it derives from the barren sandstone phase of a depositional cycle rather than from the thin fish-bed phase at the cycle base; both phases are present in every cycle, in Caithness and in the Moray margin alike, and a monolith is by definition drawn from the sandy phase. If anything, the presence of the Achanarras-assemblage flags on the Moray margin is evidence that the correct lacustrine facies family is developed there — a point of similarity, not the decisive difference the provisional dismissal took it to be. The criterion is therefore withdrawn.

6.  Assessment: deprioritised and untested, not eliminated

The honest position is narrower than a lithofacies closure but is sufficient for the enquiry’s purposes. The Nairn–southern Moray Firth belt is a marginal, more oxidised, more heterogeneous expression of the same Lake Orcadie lacustrine system that reaches its deep-water optimum in East Caithness. Its fine lacustrine flags are the right facies family and cannot be excluded on facies or fossils; what places the belt well below East Caithness is the combination of a marginal and oxidised overall character, subordinate and impersistent flag development with poor monolith potential, and — most substantively — a weak, poorly bedrock-verified barium signal where the Altar Stone’s defining baryte cement should produce a strong one. The diagnostic baryte–tosudite assemblage is untested here, exactly as it remains under active validation at East Caithness.

The belt is therefore recorded as deprioritised and untested, not eliminated. What would resolve it is the same test that will confirm or refute East Caithness: direct sampling of a fine grey lacustrine flag interval — clay XRD for tosudite, modal K-feldspar, baryte-cement habit and rock geochemistry — benchmarked like-for-like against the Altar Stone. On present evidence the prior is low and the enquiry’s effort is better spent on the stronger East Caithness lead; but recording the Moray margin as a proven facies exclusion would overstate the evidence, and would rest on an argument that also excludes the lead.

7.  Conclusion and implications for the enquiry

A systematic lithofacies audit of the Nairn–Findhorn–Elgin margin does not reproduce the provisional “too many fossils” dismissal; it replaces it. The fine lacustrine flag intervals of the belt are the same fish-bearing quiet-water facies as the East Caithness flagstones, so fossil content cannot discriminate between them. The belt is instead deprioritised on its marginal and oxidised character, its subordinate and impersistent flag development, and its weak barium signal, with the diagnostic clay and cement assemblage untested. This distinguishes the Moray margin, within the enquiry’s ledger, from the Midland Valley and Orkney: those were screened out; the Moray margin is present but deprioritised and untested. With that distinction stated honestly, the enquiry’s signal remains concentrated on the East Caithness flagstone coast — the only ground where a strong, well-verified barium signal, a compatible and thickly developed flag facies, and independent detrital-mineral geochronology converge, and where the diagnostic baryte–tosudite assemblage is under active validation. The next phase is detailed target refinement within that East Caithness fairway.

Status

Nairn / southern Moray Firth Devonian ORS strand: DEPRIORITISED and untested — a marginal expression of the same lacustrine system as East Caithness; the fine flag facies is present and shares the Achanarras fish assemblage, so it is not a facies or fossil exclusion; deprioritised on marginality, poor monolith potential and a weak barium signal, with the diagnostic baryte–tosudite clays untested. Not eliminated. New Red Sandstone (Hopeman aeolian, Burghead fluvial) excluded on age. Enquiry focus remains on East Caithness (Sarclet–Lybster–Clyth) refinement.

Selected references

Bevins, R.E. et al. (2024). Was the Stonehenge Altar Stone from Orkney? Journal of Archaeological Science: Reports, 58, 104738.

British Geological Survey. Devonian, Grampian Highlands; Middle Old Red Sandstone, Northern Highlands of Scotland; Bedrock Geology UK North — the Old Red Sandstone Supergroup. BGS Earthwise.

British Geological Survey. Fortrose and eastern Inverness (Sheet 84W), memoir for the 1:50 000 geological map.

Clarke, A.J.I. et al. (2024). A Scottish provenance for the Altar Stone of Stonehenge. Nature.

Clarke, A.J.I. et al. (2026). From Highlands to Henge. Journal of Quaternary Science.

Daw, T. (2026). The Stonehenge Altar Stone: Screening the Orcadian Basin. sarsen.org.

Dineley, D.L. & Metcalf, S.J. (1999). Fossil Fishes of Great Britain. Geological Conservation Review Series 16 (incl. Tynet Burn).

Trewin, N.H. & Thirlwall, M.F. (2002). The Old Red Sandstone of Scotland (in The Geology of Scotland, ed. Trewin).

Screening Orkney

Re-examining Orkney: a barium–rubidium and lithofacies screen of the Eday Group and outer islands as a candidate source for the Stonehenge Altar Stone

Third in a sequence applying the Screening the Orcadian Basin method (Daw 2026) to candidate ground beyond the primary study area

sarsen.org Altar Stone Sourcing Enquiry · working paper · 12 July 2026

Abstract

Bevins et al. (2024) excluded Mainland Orkney as the source of the Stonehenge Altar Stone on the basis of a limited set of Stromness- and Rousay-Flagstone field samples, leaving the outer islands and the Eday Group untested. We re-examine Orkney comprehensively, combining the barium–rubidium stream-sediment screen of Daw (2026) with a lithofacies analysis of the Eday Group and a re-reading of the one published Eday clay-mineral datum, and we test the specific fault-controlled high-barium corridors on Sanday that a parallel desktop analysis proposed as candidate ground. Orkney returns essentially no signal on the primary screen: a 0.03% barium-floor hit rate against 10.8% for the mainland Orcadian outcrop — a single cell, at Yesnaby, independently attributable to vein baryte. The proposed Sanday corridors fall below both the barium floor and the ratio threshold and do not register. On lithofacies, Sanday and the coarse Eday Sandstone are disqualified: the only Altar-Stone-compatible facies, the Eday Flags, thins northward from about 150 m at Deerness and South Ronaldsay to roughly 10 m of flaggy sediment on Sanday and the Calf of Eday, where the succession is dominated by coarse, pebbly, cross-bedded sandstone. The sole facies-plausible residual — the Eday Flags of Deerness and northern South Ronaldsay — is not a screen hit, shows no elevated barium consistent with the Altar Stone’s diagnostic baryte cement, lacks tosudite (absent from every Orkney sample measured), and remains gated on an untested Eday-Group detrital-zircon comparison. We conclude that the screening process does not identify Orkney as a candidate source. The result is independent of, and consistent with, the Bevins mineralogical exclusion, and extends it to the outer-island gap that direct sampling had not reached.

1.  Introduction

Clarke et al. (2024, 2026) established, from detrital-zircon and apatite/rutile U–Pb geochronology, that the Altar Stone’s detritus derives from the Orcadian Basin of northern Scotland. Bevins et al. (2024) then tested Mainland Orkney directly, using portable XRF, automated SEM-EDS mineralogy, and clay X-ray diffraction on field samples of the Stromness and Rousay Flagstone formations, and concluded those units do not match the Altar Stone — principally on their abundant detrital K-feldspar, the absence of the pervasive diagenetic baryte cement that characterises the stone, and the absence of its diagnostic tosudite clay. That is a well-evidenced exclusion of the sampled units. It does not, by itself, speak to the Orkney ground that was not sampled: the outer islands, and in particular the Eday Group, which crops out most fully on Eday and Sanday and only marginally on the Mainland localities Bevins examined.

This paper closes that gap. It is the third in a sequence that applies the desk-based barium–rubidium screen of Daw (2026) — developed to rank the Orcadian Basin, and extended nationally in that paper’s Appendix C — to candidate ground the primary study set aside. The first established East Caithness (near Sarclet, and the Lybster–Clyth flagstone coast) as the strongest candidate, independently corroborated by the Clarke zircon match. The second closed the Midland Valley of Scotland. Here the method is turned back on Orkney, prompted by a parallel desktop analysis that proposed a set of fault-controlled, high-barium corridors on Sanday — in the Upper and Middle Eday Sandstone, near the North Scapa Fault and the Cata Sand system — as a previously untested Orkney target. We assess those corridors, and Orkney more generally, on the same two axes the method rests on: stream-sediment geochemistry, and lithofacies.

2.  Data and methods

The geochemical screen is that of Daw (2026): a basin-relative composite condition on the BGS G-BASE 500 m kriged stream-sediment grids — a barium floor (1025 ppm) combined with a barium/rubidium ratio threshold (the 95th percentile, ≈ 14) — with each surviving cell bedrock-verified by point-in-polygon join against the BGS Geology 625k map. Rubidium substitutes for potassium and so tracks K-feldspar and mica; the ratio isolates the Altar Stone’s distinctive combination of high barium (baryte cement) with a deficit of K-feldspar. Two limits of the method are load-bearing here and are stated at the outset. First, stream-sediment values are not rock values, and cannot be compared directly against the Altar Stone’s measured rock geochemistry; the screen is calibrated against stream-sediment thresholds only. Second, formation identity is not facies identity: a cell on genuine Old Red Sandstone bedrock may still be the wrong lithofacies, so a geochemical hit is a necessary but not sufficient condition and must be read together with the sedimentology.

The lithofacies analysis draws on the regional survey of Mykura (1976), after the sedimentological work of Fannin (1970) and Ridgway (1974), which maps the Eday Group and its internal thickness variation along the 58 km north–south outcrop. The clay-mineral comparison uses Table 4 of Bevins et al. (2024); we re-read the single Eday Group entry (sample 5514) directly from the published table image, because a flattened transcription of that row misassigns its values (Section 4.4).

3.  Results

3.1  Orkney is effectively silent on the primary screen

On the barium floor alone, the Orkney archipelago returns a 0.03% hit rate against valid grid cells — a single cell, at Yesnaby, itself independently attributable in the literature to vein-hosted baryte rather than diagenetic cement — against 10.8% for the mainland Orcadian outcrop (Caithness, Sutherland, Moray, Black Isle) and 2.8% for Shetland: a roughly 350-fold contrast between Orkney and the mainland basin. Under the full composite condition, no Orkney cluster survives. The Deerness area of East Mainland and the island of South Ronaldsay — which, as Section 3.3 shows, carry the only Altar-Stone-relevant facies — both fall within this near-zero population. There is no Orkney screen hit to rank.

3.2  The proposed Sanday corridors fall below threshold

The parallel analysis reported its strongest Sanday barium pixels at roughly 670–675 ppm with barium/rubidium ratios of about 8.5–8.9. Both figures sit below the screen’s thresholds — the barium floor of 1025 ppm and the ratio threshold of about 14 — by a clear margin. On the method’s own terms these are not anomalies; they are ordinary background, and they do not register. The corridors were interpreted as fault-controlled secondary baryte along the North Scapa Fault trend, which is the same vein-baryte association that makes the single Yesnaby cell a documented false positive rather than a candidate. We note one mitigating caveat, developed in Section 4.3: Orkney is extensively covered by blown sand and till, and its G-BASE coverage is sparser than the mainland, so a stream-sediment null over Orkney is softer evidence than a null over open mainland ground.

3.3  Lithofacies: the fine facies has pinched out in the north

The Eday Group is roughly 1,000 m of dominantly fluvial sandstone (Lower, Middle and Upper Eday Sandstone), with red marls and two finer intervals; of its formations, only the Eday Flags contain the lacustrine, finely laminated, grey ‘quiescent-water’ facies that resembles the fine, well-sorted, ripple-laminated Altar Stone. Mykura (1976) maps the thickness of the Eday Flags along the whole outcrop, and the gradient is monotonic: thick in the south, effectively gone in the north.

 

The proposed Sanday corridors sit in the bottom row, and in the coarse Middle/Upper Eday Sandstone rather than the Eday Flags. On Sanday the Middle Eday Sandstone alone reaches some 400 m of reddish-purple, trough-cross-bedded, pebbly gritty sandstone, with conglomerates at Hegglie Ber — a fundamental lithofacies mismatch to the Altar Stone. The one fine interval on the island, at roughly 10 m, is too thin and too sandy to source a coherent monolith of the required dimensions, and is untested. Sanday, the Calf of Eday, and the fault-proximal Eday ground (where, along the North Scapa Fault, the Flags horizon passes entirely into sandstone) are therefore closed on facies.

3.4  The Eday Flags residual is not screen-supported

The lithofacies analysis leaves a single facies-plausible residual: the Eday Flags where they are thick and best developed, at Deerness and northern South Ronaldsay. This ground is not, however, a product of the screen — it carries no elevated barium (Section 3.1). Since the Altar Stone’s defining diagnostic feature is its pervasive baryte cement, which is precisely what generates a high-barium signal, the absence of any barium anomaly over the fairway is at best neutral and arguably mildly counter-indicative: it suggests the baryte cement is not developed in these Eday Flags. Bevins et al. (2024) found no diagenetic baryte in any Orkney sample, consistent with that reading.

3.5  Clay mineralogy: a partial match, missing the diagnostic phase

The only published Eday Group clay analysis is Bevins et al. (2024) sample 5514, from an undivided Eday Group exposure at Bu 1 on southern Mainland — not from the Sanday corridors, and not from the Deerness/South Ronaldsay fairway. Read directly from the table, its <2 µm assemblage is about 10% illite, 49% R1-ordered mixed-layer illite/smectite (expandability ~25%), and 41% kaolinite, with no tosudite and no dioctahedral chlorite. Against the Altar Stone (illite 14–19%, dioctahedral chlorite 12–13%, tosudite 15–21%, R1 I/S 26–33%, kaolinite 16–25%), this is a partial match: it shares the kaolinite and the R1-ordered illite/smectite — the closest any Orkney sample comes to the stone — but lacks the tosudite and dioctahedral chlorite that are the Altar Stone’s diagnostic phases. Tosudite is absent from every Orkney sample Bevins et al. measured.

4.  Discussion

4.1  Closing Orkney on the right grounds

It is worth being explicit about why Orkney closes, because the parallel analysis reached a superficially similar conclusion by arguments that do not hold. It compared a Sanday stream-sediment barium/rubidium ratio (≈8.6) against an Altar Stone rock-level ratio (≈105), an invalid cross-medium comparison that would equally ‘exclude’ the corroborated East Caithness lead (stream-sediment ratio 18.2); it invoked a basin-wide thermal ceiling to argue tosudite could not survive, an argument that would also exclude the Caithness flagstones, where tosudite is likewise unreported; and it rested partly on a clay transcription that misassigned sample 5514’s values (Section 4.4). None of those arguments is needed, and none is sound. The defensible closure is simpler and independent of them: Orkney produces no screen signal (Section 3.1–3.2), and the coarse Eday facies that the barium corridors actually sample is the wrong rock, while the only compatible facies has pinched out in the north (Section 3.3).

4.2  The residual, and the gate that remains open

Intellectual honesty requires that the Deerness/South Ronaldsay Eday Flags be recorded as a residual rather than a closed case. It is the one Orkney ground that is the right facies and carries a partial clay match. But it is a facies-and-clay inference the geochemistry does not support, and it is gated on a detrital-zircon question that has not been answered. The Eday Group is stratigraphically higher than the Stromness/Rousay flags on which Clarke’s Orcadian match was established, and it was fed from the south and south-west carrying rhyolitic and volcanic detritus, with contemporaneous Middle-Devonian volcanism at the base of the Eday Flags. Its provenance therefore cannot be assumed to reproduce the Altar Stone’s signature merely because it lies within the basin — the same reasoning that governed the Midland Valley Upper Old Red Sandstone. A young or distinct volcanic zircon population, if present, would be inconsistent with the Altar Stone’s youngest concordant grain of ~498 Ma. Resolving whether any Eday Group detrital-zircon dataset exists, and how it compares, is the one desk check that could either close this residual outright or promote it to a field target; until then it is untested, not eliminated.

4.3  The drift-cover caveat

A stream-sediment null over Orkney is weaker evidence than the same null over open mainland ground. Orkney carries extensive blown sand and till — Sanday especially, with the Cata Sand system — and its G-BASE sampling is correspondingly sparser, so the grid may in places be sampling superficial cover rather than bedrock. This is a genuine reason not to treat the Orkney null as a hard exclusion. It is not, however, a reason to read the null as encouragement: the absence of a barium signal remains consistent with the mineralogical absence of baryte cement that Bevins et al. measured directly, and the lithofacies closure of Sanday (Section 3.3) does not depend on the geochemistry at all.

4.4  Relation to Bevins et al. (2024)

This re-examination is independent of the Bevins exclusion — it uses different data (national stream-sediment geochemistry and regional lithofacies rather than local field mineralogy) — and reaches a consistent result by a different route. Where Bevins et al. excluded the sampled Mainland flagstone units on mineralogy, the present screen finds no Orkney signal at all, and the lithofacies analysis accounts for the outer-island Eday Group that direct sampling had not reached. The two lines together close the outer-island gap that the original Orkney rejection had left open, with the single, explicitly flagged residual of Section 4.2.

5.  Conclusion

Applied comprehensively to Orkney — including the previously untested outer islands and the specific fault-controlled corridors proposed on Sanday — the screening process does not identify a candidate source for the Altar Stone. Orkney is effectively silent on the barium–rubidium screen (a single vein-baryte cell against a 350-fold-higher mainland hit rate), the Sanday corridors fall below threshold, and Sanday together with the coarse Eday Sandstone is disqualified on lithofacies, the only Altar-Stone-compatible facies having thinned to a marginal remnant in the north. The sole residual, the Eday Flags of Deerness and northern South Ronaldsay, is facies-plausible but unsupported by the geochemistry, without the Altar Stone’s baryte cement or tosudite, and gated on an untested Eday-Group zircon comparison; it is recorded as untested rather than eliminated. This is a negative result, and a useful one: it removes Orkney from contention on the method’s own terms and, with the Anglo-Welsh Basin and the Midland Valley, leaves the enquiry’s signal concentrated where the screen, the lithofacies, and the independent zircon evidence agree — the East Caithness coast of the mainland Orcadian Basin.

Selected references

Bevins, R.E. et al. (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.

Clarke, A.J.I. et al. (2024). A Scottish provenance for the Altar Stone of Stonehenge. Nature.

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.

Daw, T. (2026). The Stonehenge Altar Stone: Screening the Orcadian Basin. sarsen.org / repository.

Fannin, N.G.T. (1970). The sedimentary environment of the Old Red Sandstone of western Orkney. PhD thesis, University of Reading (unpublished).

Hillier, S., Wilson, M.J. & Merriman, R.J. (2006). Clay mineralogy of the Old Red Sandstone and Devonian sedimentary rocks of Wales, Scotland and England. Clay Minerals, 41, 433–471.

Mykura, W. (1976). British Regional Geology: Orkney and Shetland. HMSO, Edinburgh.

Ridgway, J.M. (1974). The Sedimentology and Palaeogeography of the Eday Group, Middle Old Red Sandstone, Orkney. PhD thesis, University of London (unpublished).