Clarke, Leary and Kirkland have published a detrital zircon–apatite provenance study of the Devil’s Arrows (Boroughbridge, North Yorkshire) in Proceedings of the Royal Society A. Open access: doi:10.1098/rspa.2026.0504.
The stones are Millstone Grit standing on Triassic Sherwood Sandstone; the nearest grit outcrop is about 5 km west, so the megaliths were moved — by people, by ice, or both. The paper fingerprints the three pillars to Brimham Rocks (~18 km west) rather than Plumpton Rocks, and states that regional ice-flow reconstructions exclude glacial transport from that source.
This updates an earlier sarsen.org note that followed secondary sources to Plumpton. That claim is superseded.
Grit on the wrong bedrock
Lithology already said the Arrows are not local. Sherwood Sandstone is red–brown, quartzose, Triassic fluvial–aeolian stuff; the pillars are grey, coarse, lithic Namurian Millstone Grit. Zircon agrees: two-sample KS against published Sherwood spectra gives D = 0.17, p = 2.9×10⁻⁴. The paper concludes the megaliths were not sourced locally.
Heights: south 6.86 m, centre 6.7 m, north 5.5 m. The paper specifies these as Britain’s tallest surviving standing stone row, with the southern stone the second-tallest British megalith after Rudston and the central stone the third. Erection date remains Late Neolithic to Early Bronze Age by convention — the paper states the “precise date remains unknown” for want of recent excavation. The U–Pb ages date the mineral grains, not the erection of the row.
People or ice?
Thorpe and Williams-Thorpe (1991) floated glacial erratics as an alternative to long-distance human haul — for the Arrows, grit moved south from the Northallerton area and dumped near Boroughbridge, then erected. Burl pushed back the same year. Plumpton Rocks (~13 km), with lookalike fluting, became the usual human source in the literature and in the NHLE entry. Brimham (~18 km west) was already on the menu as another grit tor field.
The paper’s contribution is isotopic fingerprinting combined with ice-flow-geometry analysis, drawing on the existing BRITICE-CHRONO and Veness et al. reconstructions, not new excavation.
What they actually sampled
Scheduled Monument Consent, granted 23 October 2024. Sampling used a peel method: 20 cm of adhesive tape pressed for five minutes onto dry, inconspicuous, lichen-free faces of the stones; grains released by dissolving the glue in turpentine, then separated by magnetic and heavy-liquid methods and mounted for LA-ICP-MS at Curtin’s GeoHistory Facility. Comparative Brimham and Plumpton samples were ~1 kg rock, cut and disaggregated by pulsed electric discharge (SelFrag) and sieved to <425 µm — a different preparation route from the peel method used on the standing stones, though a valid basis for comparing age spectra. Peel sampling recovers only near-surface grains; it is not a thin-section of the core, a clast count, or a sample from a demonstrated quarry scar.
Zircon is the workhorse. Stone 1 yielded 67 concordant ages, Stone 2 64, Stone 3 105 (n = 236 pooled). Pairwise KS and Kuiper tests between the three stones all give p > 0.05 (their Table 1): the spectra are statistically indistinguishable at 95% confidence, so the paper pools them into one composite dataset. High p means the test fails to reject “same distribution,” not that the spectra are identical — but the three pillars share one fingerprint family, not three unrelated erratics from three different outcrops.
Against Brimham (n = 75): KS D = 0.08, p = 0.77; Kuiper V = 0.121, p = 0.73. Against Plumpton (n = 76): KS D = 0.29, p = 6.7×10⁻⁵. Brimham and the Arrows share the same polymodal structure — dominant Caledonian (~460–400 Ma), Mesoproterozoic and Palaeoproterozoic components, minor Archean contributions. Plumpton, though the same Namurian Upper Plumpton Grit interval as Brimham, carries a different mix, dominated by more Silurian zircon. The paper attributes this to lateral heterogeneity already documented within the Pennine Basin’s fluvial channels.
MDS residual. In multidimensional scaling of Pennine Basin datasets (stress = 0.11), Brimham plots closest to the Arrows; Plumpton sits further off, pulled by its Silurian-dominated zircon. The least-dissimilar dataset with an overlapping uncertainty ellipse is a Langsettian sandstone at Binchester Crags, ~70 km north — the same broader northerly Pennine River signature. The paper is explicit that this places Brimham as the best match among the outcrops sampled, not as a unique source: distance, morphology and the identified local candidates still do the narrowing.
Apatite is secondary support. Most Arrow grains are reworked phosphatic bioapatite (low U, high Th/U, common-Pb-rich), not primary igneous provenance clocks; the paper reads their Pb-isotope signature as later diagenetic overprinting, not a source-rock age. A smaller detrital igneous subset does match Brimham’s apatite populations (~480 Ma, ~1300 Ma, ~2600 Ma) better than Plumpton’s.
Battle Cross and Peg Bridge masonry were also sampled. Battle Cross zircon carries the same dominant Caledonian and minor Neoarchean components as the Arrows; Peg Bridge zircon falls within the Arrows’ Neo- to Mesoproterozoic population. No grain from either source is inconsistent with the Millstone Grit fingerprint — support for the antiquarian record that both were built from a lost fourth Arrow.
Once it’s Brimham, ice has a geometry problem
The old erratic hypothesis had grit moving south from Northallerton and being dumped near Boroughbridge. Brimham sits west, on elevated Pennine ground outside that trajectory, so a Brimham source changes the transport question from southbound to eastbound.
Late Devensian ice-flow reconstructions for the Vale of York have ice advancing northwest to southeast, carrying material down-valley south — “rather than eastward from the Pennine uplands” (Catt 2007; Clark et al. 2022; Gibbard & Clark 2011; Veness et al. 2025). Brimham lies outside that principal trajectory; the paper calls the entrainment and eastward transport of >25-tonne blocks from there “improbable.”
The paper doesn’t treat the regional ice map as a fixed transport route: erratic journeys can be multistage as ice divides shift, and Veness et al.’s models are spatially coarse — they “cannot resolve or demonstrate a specific transport pathway” over the ~18 km from Brimham to Boroughbridge. What the paper argues is narrower: given a Brimham source, an eastward glacial haul is a poor fit to the reconstructed ice geometry, and all three stones sharing one outcrop-like spectrum is the wrong pattern for a grab-bag of glacial erratics.
| Claim | Strength |
|---|---|
| Local Sherwood source | Ruled out |
| Plumpton as best tested grit match | Ruled out among sampled candidates |
| Brimham as best tested match | Strong |
| Unique Brimham among all Pennine channel sands | Not proven (Binchester ellipse; unsampled twins) |
| Principal Late Devensian flow carrying Brimham grit east to Boroughbridge | Contradicted by regional reconstructions |
| Some exotic multistage glacial path for these exact blocks | Not fully excluded by coarse models; called improbable |
| Human haul from Brimham | Least-bad remaining explanation once provenance + ice geometry are stacked |
Not a proof that ice never moved grit in Yorkshire — the Vale has documented erratic trains of other rocks. A case that these three pillars, matching this one western tor field, do not look like ordinary Vale-of-York erratics.
The paper places the Arrows alongside other British and Irish monuments where stone was sourced from specific, often distant, locations rather than the nearest convenient outcrop: the Altar Stone traced to the Orcadian Basin (~750 km), the Preseli bluestones (~225 km), the sarsens forming Stonehenge’s trilithons (~30 km), the stone circles of Brodgar and Stenness, and the passage tombs of Newgrange and Knowth. At ~18 km, the Arrows sit at the modest end of that scale — which, if anything, makes deliberate human transport the easier case to argue, not the harder one.
What this still isn’t
No Brimham quarry face with tool marks or debitage. Provenance chemistry is not quarry archaeology.
No new erection date.
No resolution between practical and symbolic motive. The paper notes that favourable joints and bedding planes at Brimham would have reduced the effort needed to detach elongate blocks, and sits this alongside the site’s cup-and-ring rock art and its tor-like landscape as parallel, not competing, explanations for why this source was chosen.
Sources
Clarke A, Leary J, Kirkland C. 2026. Deliberate prehistoric sourcing of the Devil’s Arrows, Britain’s tallest stone row. Proc. R. Soc. A 482: 20260504. https://doi.org/10.1098/rspa.2026.0504
Thorpe RS, Williams-Thorpe O. 1991. The myth of long-distance megalith transport. Antiquity 65: 64–73. Burl A. 1991. Megalithic myth or man the mover? Antiquity 65: 297–298.
Ice context as cited by Clarke et al.: Catt 2007; Clark et al. 2022; Gibbard & Clark 2011; Veness et al. 2025.
Clarke AJI, Kirkland CL, Bevins RE, Pearce NJG, Glorie S, Ixer RA. 2024. A Scottish provenance for the Altar Stone of Stonehenge. Nature 632: 570–575.
Prior sarsen.org note (Plumpton; superseded): https://www.sarsen.org/2024/09/the-long-flight-of-devils-arrows.html
© Tim Daw / sarsen.org · CC BY-SA 4.0
No comments:
Post a Comment
Comments welcome on fresh posts - you just need a Google account to do so.