Closing the Midland Valley of Scotland as a source for the Stonehenge Altar Stone

Project record · Midland Valley strand closure · second in the sequence applying the barium–rubidium screen to candidate ground beyond the primary Orcadian study area

Abstract

Clarke et al. (2026) excluded the Midland Valley terrane on the basis of a distinctive mid-Palaeozoic and Neoproterozoic detrital-zircon signature carried by its Lower Old Red Sandstone successions — a signature the Altar Stone lacks. A national barium–rubidium stream-sediment screen, however, also identified clusters within the Upper Old Red Sandstone Stratheden Group on the Clyde and Ayrshire coast. These lie stratigraphically above the measured Lower ORS datasets and were therefore untested by Clarke. This paper applies a four-test veto framework — direct zircon data search and statistical comparison if available, source-terrane geometry and recycling assessment if no data exist, facies/lithology comparison to the Altar Stone’s defining characteristics, and clay mineralogy focused on the diagnostic tosudite phase — to close that gap independently. No published detrital-zircon U–Pb dataset exists for the Stratheden Group or the Upper ORS of the northern Midland Valley. The source-terrane geometry and recycling analysis returns an unfavourable prior: even the northern-flank clusters carry forward the forbidden grain populations via demonstrable reworking of Lower ORS detritus. Lithofacies comparison reveals a clear mismatch — the Stratheden Group is coarsening-upward, conglomeratic, aeolian-influenced and locally volcaniclastic, whereas the Altar Stone is fine-grained, well-sorted, ripple-laminated and non-volcaniclastic. Clay mineralogy shows that the diagnostic tosudite phase of the Altar Stone is documented in the Midland Valley only within the Lower ORS (Strathmore Group) and is absent from the Stratheden Group. Three independent lines therefore converge on exclusion of the Stratheden clusters at high confidence, although the closure remains indirect because no zircon or clay measurement has been performed directly on the Ba/Rb cluster ground itself. The Lower ORS clusters are excluded by direct measurement. With the Midland Valley strand closed, the enquiry returns to the Orcadian Basin for refined search within the East Caithness candidate ground.

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, and further showed that the Midland Valley of Scotland carries a mid-Palaeozoic and Neoproterozoic zircon signature the Altar Stone conspicuously lacks. Their methods statement records that the comparison datasets “encompass all currently published detrital zircon U–Pb ages from Scottish Old Red Sandstone successions.” That compilation rests on McKellar et al. (2020, 2021) for the northern Midland Valley (Strathmore Basin) Lower ORS and Phillips et al. (2009) for the southern (Lanark Basin) pre-ORS Silurian. Both are Lower ORS or older; no Upper ORS data entered the discriminant.

The national barium–rubidium screen reported in Appendix C of Daw (2026) nevertheless fired on genuine Old Red Sandstone clusters both inside and outside the Orcadian Basin. Within the Midland Valley these resolved into two stratigraphic populations: Lower ORS clusters (Arbuthnott–Garvock Group, Strathmore Group, and the Silurian–basal Lanark ground) that sit squarely inside Clarke’s measured material, and Upper ORS Stratheden Group clusters on the Clyde and Ayrshire coast (17.8, 13.2 and 9.5 km² at 74–81 % bedrock purity) that do not. The Lower ORS clusters are therefore already excluded by direct measurement. The Stratheden Group clusters survived contact with the published zircon evidence only because the evidence was never pointed at them.

To decide whether the Stratheden Group clusters warranted further masking or fieldwork, a four-test veto framework was applied in cheapest-and-most-lethal order. Any one decisive failure shuts the strand; it remains open only if it passes or ties on all four. The tests are: (A) search for any published detrital-zircon U–Pb dataset for the Stratheden Group and, if grain-level ages are recoverable, run a two-sample KS test against the 56 concordant Altar Stone grains; (B) if no zircon data exist, assess source-terrane geometry and recycling for the specific Ayrshire–Clyde clusters (Highland/Grampian versus Southern Uplands routing, and whether Lower ORS detritus is reworked upward); (C) compare Stratheden lithofacies directly against the Altar Stone’s defining sedimentological characteristics; (D) test whether the diagnostic tosudite/aluminous kaolinite clay assemblage is documented in the Stratheden Group itself. The present paper executes those tests and records the closure. It is the second in the sequence that applies the desk-based barium–rubidium screen of Daw (2026) and its associated veto methods to candidate ground beyond the primary Orcadian study area. The first established East Caithness (Sarclet–Lybster–Clyth flagstone coast) as the strongest national candidate, independently corroborated by Clarke’s zircon match at Sarclet. This paper closes the Midland Valley strand. The third turns the same method back on Orkney to close the outer-island and Eday Group gap left by Bevins et al. (2024).

The Stratheden Group was initially a legitimate lead for three reasons. Its Famennian age is compatible with the Altar Stone’s youngest concordant detrital zircon (~498 Ma Cambrian); the zircons impose no maximum depositional age that forbids an Upper ORS host. The volcanic axis (Ochil–Sidlaw) that supplies Clarke’s Midland Valley discriminant was extinct by Stratheden time, so the direct arc input might not extend upward. And a clay-mineral observation (Wilson 1971, reported in Hillier et al. 2006) appeared to point the right way: the Upper ORS of the Midland Valley was described as mineralogically similar to the Upper ORS of the Orcadian Basin and the Middle ORS around the Moray Firth — the very ground already implicated by the Altar Stone’s aluminous kaolinite–tosudite assemblage.

Set against these considerations was a mechanistic prior drawn from source-terrane geometry and palaeodrainage. The Altar Stone’s zircon load is Mesoproterozoic and Archaean (Laurentian, overprinted by Grampian ~460 Ma magmatism) and lacks both a pronounced Neoproterozoic population and a mid-Palaeozoic (Ordovician–Silurian) one. Across Britain and Ireland the Upper ORS is characterised by precisely those missing populations, recycled from Southern Uplands–Longford–Down sources or introduced by contemporaneous volcanism. For the documented Borders–Solway outcrop of the Stratheden Group, granitic clasts and palaeocurrents indicate a Galloway Hills / Southern Uplands source and internal drainage to the Jedburgh area — a southern routing opposite to the Altar Stone’s northern, Grampian-overprinted character. Even for the northern-flank (Ayrshire–Clyde) clusters that survived the Ba/Rb screen, the Stratheden sits unconformably on older Midland Valley ORS and demonstrably reworks it, so the forbidden grain populations can be carried forward regardless of the extinct volcanic axis. The honest prior was therefore unfavourable, but it remained a prediction from terrane models rather than a measurement on the candidate ground itself. The project exists because a confident stratigraphic prediction (“Middle ORS”) turned out to be unverified; the same discipline requires that an unverified exclusion also be tested rather than assumed.

This paper therefore executes the four tests, documents each result, records one clarification to the initial clay reading for the Stratheden lead, and hands the enquiry back to the Orcadian Basin with every screened Midland Valley candidate now accounted for.

2. Data and methods

The four-test veto framework applied here is as follows. Four tests are applied in order; any decisive failure shuts the strand.

Test A — Direct zircon (if data exist). Trace any recoverable Stratheden Group or northern Midland Valley Upper ORS U–Pb detrital-zircon dataset. If grain-level ages are recoverable in the form provided by Strachan et al. (2021) for Orcadian localities, run a two-sample Kolmogorov–Smirnov test against the 56 concordant Altar Stone grains exactly as performed for Sarclet/Braemore. Verdict rule: p > 0.05 and no Neoproterozoic or mid-Palaeozoic mode → OPEN strongly. Any failure → SHUT decisively. This test dominates all others if data can be found.

Test B — Source-terrane geometry (if no data). Establish the sediment routing of the Ayrshire–Clyde Stratheden specifically: Highland/Grampian source versus Southern Uplands source, using palaeocurrents, clast petrography and heavy-mineral data. Unavoidably Southern-Uplands-sourced or demonstrably recycling-dominated → SHUT on prediction (flagged as such). Genuinely ambiguous or northern-sourced without recycling carry-over → stays OPEN pending sampling.

Test C — Facies / lithology. Compare the Stratheden Group against the Altar Stone’s defining characteristics: grey-green, fine-grained, well-sorted, K-feldspar-poor, calcite/baryte-cemented, non-volcaniclastic, with quiescent-water ripple lamination. Coarsening-upward, aeolian-influenced, conglomeratic or locally volcaniclastic character → can SHUT the strand on mismatch alone.

Test D — Clay mineralogy. Check whether tosudite and the associated aluminous kaolinite assemblage are documented in the Stratheden Group specifically (distinct from the Lower ORS Strathmore/Stonehaven Groups where Hillier et al. 2006 place it). Match on the diagnostic phase → corroborates OPEN. Absence → weakly negative, not decisive by itself.

Data sources comprise the published zircon compilations already used by Clarke et al. (McKellar et al. 2020, 2021; Phillips et al. 2009; Strachan et al. 2021), the BGS Geology 625k and 1:50k maps for cluster verification and formation attribution, the regional sedimentological and provenance syntheses of Bluck (1978, 1980), Paterson et al. (1990) and Mykura (1991), the clay-mineral compilation of Hillier et al. (2006) including the Wilson (1971) observations, and the BGS Earthwise Old Red Sandstone reviews together with the relevant Midland Valley memoirs (Greenock district and regional geology). The Ba/Rb screen thresholds and cluster statistics are those of Daw (2026). No new field or laboratory data were generated; the exercise is entirely a desk-based audit and synthesis of existing published and archival sources.

Intellectual hygiene requires that exclusions be stated only at the stratigraphic and geographic level the evidence actually supports, that prediction and measurement be kept in separate columns, and that anything asserted as closed must be closed by evidence pointed at the thing itself — not at its neighbour, its parent unit, or its terrane label.

3. Results

3.1 The Lower Old Red Sandstone Midland Valley clusters are excluded by direct measurement

The national screen flagged Lower ORS clusters within the Arbuthnott–Garvock Group (94.5 % and 77.5 % bedrock purity), small Strathmore Group clusters, and the Silurian–basal Lanark ground. These sit squarely inside the material Clarke et al. characterised. McKellar et al. (2020) provide ten zircon samples across a 9 km succession in the northern Midland Valley (Stonehaven, Arbuthnott–Garvock and Strathmore Groups) with a proximal easterly / SW-Baltican source. Phillips et al. (2009) cover the southern pre-ORS Silurian of the Lanark Basin. Both datasets contain the diagnostic mid-Palaeozoic (~490–420 Ma) and Neoproterozoic grain populations shed from the Ochil–Sidlaw volcanic axis and proximal sources — populations the Altar Stone’s 56 concordant grains lack (youngest ~498 Ma). The Lower ORS Midland Valley candidates are therefore excluded by direct measurement on the precise stratigraphic units concerned. No further masking or fieldwork is required for them.

3.2 Test A returns no data: no published detrital-zircon U–Pb dataset exists for the Stratheden Group

Following Clarke et al.’s own methods statement, an exhaustive search was made for any published or archived U–Pb detrital-zircon dataset for the Stratheden Group or for Upper ORS successions of the northern Midland Valley (Fife, Firth of Tay, Ayrshire–Clyde coast). None was found. The gap is real; it is not an omission by Clarke but a reflection of the published record at the time their compilation was assembled. Consequently Test A cannot strongly open the strand by direct statistical comparison, nor can it shut it on a failed KS test. The result of Test A is therefore recorded as “no recoverable data exist.” The strand proceeds to Test B.

3.3 Test B: source-terrane geometry and demonstrated Lower-ORS recycling render the prior unfavourable for the Ayrshire–Clyde Stratheden clusters

The mechanistic prior for the Stratheden Group as a whole is unfavourable. The Altar Stone lacks both a pronounced Neoproterozoic (~1000–540 Ma) population and a mid-Palaeozoic (Ordovician–Silurian, ~490–420 Ma) one. The Upper ORS across Britain and Ireland characteristically carries both, either recycled from Southern Uplands–Longford–Down volcanism and metasediments or introduced by contemporaneous arc input. For the documented Borders–Solway outcrop, BGS provenance work and clast petrography indicate granitic detritus from the Galloway Hills / Southern Uplands and drainage into internal basins in the Jedburgh area — a southern routing geometrically opposite to the Altar Stone’s Grampian-overprinted northern signal. This geometry alone predicts the introduction of the missing populations.

For the specific northern-flank clusters (Clyde and Ayrshire coast) that survived the Ba/Rb screen, the picture is not provenance-monolithic; a Highland/Grampian source remains conceivable in principle and would remove the Neoproterozoic and mid-Palaeozoic objection at a stroke. However, the escape hatch is closed by recycling. The Stratheden Group sits unconformably on the older Midland Valley ORS succession and demonstrably reworks it. The very mid-Palaeozoic and Neoproterozoic grains that Clarke used as the Midland Valley discriminant can therefore be carried forward into the Upper ORS regardless of whether the Ochil–Sidlaw volcanic axis was extinct by Famennian time. No published palaeocurrent, clast or heavy-mineral dataset for the Ayrshire–Clyde Stratheden overturns this carry-over. Test B therefore returns an unfavourable verdict on prediction, reinforced by demonstrated reworking. The strand is shut at this point on the strength of source geometry and recycling; direct zircon measurement on the cluster ground would be required to overturn it, and none exists.

3.4 Test C: clear lithofacies mismatch

Even if the zircon prior had been favourable, Test C would shut the strand on its own. The Altar Stone is a fine-grained (dominantly fine sand), well-sorted, grey-green, ripple-laminated sandstone with pervasive diagenetic baryte and calcite cement and negligible detrital K-feldspar — a quiescent-water lacustrine or marginal facies. The Stratheden Group (the lower unit of the Scottish “Upper Old Red Sandstone”, Famennian) is a coarsening-upward sequence of red sandstone and conglomerate with aeolian influence, trough cross-bedding, pebbly gritty horizons and local volcaniclastic input. On the Clyde–Ayrshire coast the Middle and Upper Eday-equivalent sandstones reach hundreds of metres of reddish-purple, trough-cross-bedded, pebbly gritty sandstone with conglomerates; the finer intervals are thin and sandy. None of this matches the Altar Stone’s defining sedimentological characteristics. A clear facies mismatch exists. Test C alone is sufficient to exclude the Stratheden clusters.

3.5 Test D: clay mineralogy — tosudite is a Lower ORS phase in the Midland Valley; absent from the Stratheden Group

Re-examination of the Wilson (1971) observations as synthesised in Hillier et al. (2006) shows that the noted mineralogical similarity between the Upper ORS of the Midland Valley and the Upper ORS of the Orcadian Basin / Middle ORS of the Moray Firth is limited to the kaolinite + illite/smectite component. The diagnostic tosudite (and associated aluminous phases) that characterise the Altar Stone’s <2 µm assemblage are documented in the Midland Valley only within the Lower ORS Strathmore Group; they are not reported from the Stratheden Group or other Upper ORS successions. On the Altar Stone’s defining clay phase, therefore, the proxy is a mismatch rather than a match. This reading corrects an initial interpretation in the opening analysis of the Stratheden lead, which had taken the Wilson (1971) similarity as more favourable than the full published evidence supports. The similarity is real but partial and does not extend to the diagnostic assemblage required by the Altar Stone.

Convergence. Tests B, C and D each return negative independently and on different axes (provenance/recycling, facies, clay). Test A could not strongly open the strand because no data exist. No axis rescues the candidate. The Stratheden Group clusters of the Clyde and Ayrshire coast are therefore closed by convergent inference at high confidence. The closure is strong but remains indirect: no detrital-zircon U–Pb or matched clay separation has been performed on the specific Ba/Rb cluster ground itself.

The unscreened Stratheden outcrops (Fife, Arran, Kintyre) produced no Ba/Rb hit and are excluded by null result, subject to the same 500 m-resolution and superficial-cover caveats noted for other null areas in the national screen.

4. Discussion

4.1 The verdict at its honest register

The temptation to state a clean universal negative (“nowhere in the Midland Valley ORS could match the Altar Stone”) is resisted. That phrasing would repeat, in the opposite direction, the error the enquiry was built to catch — a confident claim outrunning the measurement that licenses it. What is actually supported is more precise and remains defensible:

•             Lower ORS and Silurian successions: excluded by direct measurement (Clarke et al.’s zircon signatures obtained on those exact Groups).

•             Upper ORS Stratheden clusters (Clyde/Ayrshire): excluded by convergent inference — a direct facies mismatch, a recycling-reinforced unfavourable zircon prior, and a clay proxy that fails on tosudite. High confidence, but no zircon or clay measurement performed on the cluster ground itself.

•             Every screened Midland Valley candidate is now accounted for; the unscreened Stratheden outcrops are excluded by null Ba/Rb result.

What would convert inferred to measured is straightforward: a single direct sample from one Clyde/Ayrshire Stratheden cluster, with U–Pb detrital zircon run against the 56 concordant Altar Stone grains (KS test) plus a matched clay separation tested for tosudite. Nothing found in this investigation suggests such measurement would reverse the verdict, but it has not been done and the paper records that fact explicitly.

4.2 Relation to Clarke et al. (2026)

It would be inaccurate to record that “Clarke was right without doing the full analysis.” Clarke closed the terrane on Lower ORS data and, by their own methods statement, had no Upper ORS zircon to test; the Stratheden Group was the genuine gap in their reasoning, not something they had already closed. What this strand supplied was the missing independent work — facies comparison, recycling audit and clay re-reading — and arrived at the same terrane-level verdict Clarke had assumed but could not yet support for the Upper ORS. The correct record is therefore that Clarke reached the right answer for the terrane on the data available to them; the work that makes the answer right for the Upper ORS is work they had not done and that is now complete.

4.3 Clarification of the initial clay reading for the Stratheden lead

As noted under Test D, an initial reading of the Wilson (1971) observations (as synthesised in Hillier et al. 2006) had logged the mineralogical similarity between Midland Valley Upper ORS and Orcadian/Moray Firth ORS as a favourable clay signal for the Stratheden lead. That reading was too generous. The similarity covers only the kaolinite + illite/smectite component; it does not extend to tosudite, which in the Midland Valley is a Lower ORS Strathmore phase and is not documented in the Upper ORS. On the Altar Stone’s diagnostic phase the clay proxy is therefore a mismatch, not a match. The same intellectual-hygiene rule applied throughout the enquiry — that conclusions must not outrun the measurements that license them — requires this clarification here.

4.4 No further work warranted on present evidence

“Closed” commits the project only to the following: every screened Midland Valley candidate is accounted for; no axis favours a match; the single residual (direct measurement on cluster ground) has not been performed but is not expected to alter the outcome on present evidence. No further field sampling or laboratory allocation is warranted for the Midland Valley strand. Stated at this register, the conclusion is reproducible and defensible in review.

4.5 Intellectual hygiene maintained

The Midland Valley episode illustrates the same discipline applied in reverse to the Middle ORS episode that prompted the enquiry. In both cases an unverified label (first “Middle ORS is established fact”, later “Midland Valley is ruled out”) threatened to do more work than the measurements licensed. The rule is identical in both directions: state exclusions at the level the data support (Lower ORS: excluded by measurement; Stratheden: untested until this strand, now closed by convergent inference), keep prediction and measurement in separate columns, and require that anything asserted as closed be closed by evidence pointed at the thing itself.

5. Conclusion

The Midland Valley of Scotland is closed as a source for the Altar Stone. The Silurian and Lower Old Red Sandstone successions are excluded by Clarke et al.’s directly measured detrital-zircon signatures. The Upper Old Red Sandstone Stratheden Group clusters of the Clyde and Ayrshire coast — the only Midland Valley candidates that survived contact with the published zircon evidence — are excluded by convergent inference from three independent lines: a clear facies mismatch, an unfavourable zircon prior reinforced by demonstrated Lower-ORS recycling, and failure of the clay proxy on the diagnostic tosudite phase. Every screened candidate is accounted for. The Stratheden closure is high-confidence but indirect; no zircon or clay measurement has been made on the cluster ground itself, and that limitation is stated explicitly.

With the Midland Valley strand closed, the enquiry returns to the Orcadian Basin for the more detailed search within the East Caithness (Sarclet–Lybster–Clyth) candidate ground that both the original barium–rubidium screen and Clarke’s independent zircon corroboration identify as the strongest remaining lead. 

Status: Midland Valley strand CLOSED. Lower ORS — excluded by measurement (Clarke et al. 2024, 2026). Upper ORS Stratheden clusters — excluded by convergent inference (facies + recycling-reinforced zircon prior + clay proxy), high confidence; direct measurement on cluster ground not performed. Key sources: Clarke et al. 2024, 2026; Daw 2026; McKellar et al. 2020, 2021; Phillips et al. 2009; Strachan et al. 2021; Hillier et al. 2006 (Wilson 1971); Bluck 1978/1980; Paterson et al. 1990; Mykura 1991; BGS memoirs and Earthwise ORS reviews. Next: detailed Orcadian Basin search within East Caithness candidate ground.

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.

Bluck, B.J. (1978). Sedimentation in a late orogenic basin: the Old Red Sandstone of the Midland Valley of Scotland. In: Tectonic evolution of the Caledonides. Special Publication of the Geological Society of London.

Bluck, B.J. (1980). Evolution of a strike-slip fault-controlled basin, Upper Old Red Sandstone, Scotland. In: Sedimentation at oblique-slip margins. Special Publication of the International Association of Sedimentologists.

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. (Includes national Ba/Rb run in Appendix C.)

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.

McKellar, R.C. et al. (2020). Detrital zircon provenance of the Lower Old Red Sandstone, northern Midland Valley, Scotland. Journal of the Geological Society.

McKellar, R.C. et al. (2021). Further detrital zircon data from the Lower Old Red Sandstone of the Midland Valley. (Follow-up dataset.)

Mykura, W. (1991). British Regional Geology: The Midland Valley of Scotland (3rd edn). HMSO, Edinburgh.

Paterson, I.B., McAdam, A.D. & MacPherson, K.A.T. (1990). Geology of the Greenock district. Memoir of the British Geological Survey, Sheet 30W (Scotland).

Phillips, E.R. et al. (2009). Detrital zircon geochronology of Silurian–Devonian sedimentary rocks, southern Midland Valley, Scotland. Journal of the Geological Society.

Strachan, R.A. et al. (2021). Detrital zircon U–Pb ages from the Old Red Sandstone of the Orcadian Basin: implications for provenance and the Stonehenge connection. (Dataset as used in Clarke et al. re-analysis.)

Waldron, J.W.F., Floyd, J.D., Simonetti, A. & Heaman, L.M. (2008). Ancient Laurentian detrital zircon in the Southern Uplands, Scotland: implications for regional tectonics. Journal of the Geological Society.

Wilson, M.J. (1971). Clay mineralogy of the Old Red Sandstone of the Midland Valley. (Unpublished data cited in Hillier et al. 2006.)

BGS Earthwise Old Red Sandstone reviews and 1:50k/625k digital geology (accessed via BGS GeoIndex and associated memoirs).