The six-tonne Altar Stone at the heart of Stonehenge came from the far north of Scotland rather than south-west Wales as previously thought, new analysis has found. The discovery shows the construction of Stonehenge was a far greater collaborative effort than scientists realised. It also means that the ancient monument, near Salisbury in south-west England, was built with stones from all parts of Great Britain. The findings suggest Neolithic Britain was a far more connected and advanced society than earlier evidence indicated.

The distance between Stonehenge and the far north of Scotland is about 700km (434 miles). The research was led by a Welsh PhD student, Anthony Clarke, now working at Curtin University in Western Australia. Such is the importance of the discovery that it has been published in one of the world’s leading scientific journals, Nature, which is an enormous achievement for an apprentice researcher. But it is a bittersweet moment for the young Welshman, who was born in Pembrokeshire, where the Altar Stone was until now thought to have come from. “I don’t think I’ll be forgiven by people back home,” he joked to BBC News. “It will be a great loss for Wales!”

But Mr Clarke points out that the remaining stones in the central horseshoe, which are known as bluestones, are from Wales and the larger stones in the outer circle are from England. “We’ve got to give the Scots something!” he said. “But on a serious note, Stonehenge seems to be this great British endeavour involving all the different people from all over the island," he said.

The bluestones at Stonehenge were identified as coming from the Preseli hills in Pembrokeshire in 1923, by Welsh geologist Henry Herbert Thomas. The central Altar Stone was made of a different rock but always assumed to have come from the same area, until 20 years ago when scientists first began to question its origins.

Last year, researchers including Prof Nick Pearce from Aberystwyth in Wales, of all places, concluded that the Altar Stone could not have come from Wales. But its origin had remained a mystery, until now. “It blew our socks off when we discovered it was from north-east Scotland,” Prof Pearce, who was also involved in the current discovery, told BBC News. “It was a shock to say the least. Coming from that distance, more than 700km, was remarkable. “The Neolithic people must have been pretty well connected, far more connected than people give them credit for. They must have been very well organised”.

The breakthrough was made by the team at Curtin university who analysed the chemical composition of fragments of rock that had fallen off the Altar Stone and dated them. The composition and date are unique to rocks from different parts of the world, rather like a fingerprint. The Australian team had access to one of the most comprehensive global rock fingerprint databases and found the best match was from the Orcadian Basin, which includes the Caithness, Orkney, and Moray Firth regions of north-eastern Scotland.

Construction at Stonehenge began 5,000 years ago, with changes and additions over the next two millennia. Most of the bluestones are believed to have been the first stones erected at the site. Dr Robert Ixer, from University College London, who was also involved in the study, described the result as “shocking”.

“The work prompts two important questions: how was the Altar Stone transported from the very north of Scotland, a distance of more than 700 kilometres, to Stonehenge, and, more intriguing, why?”

The distance is the longest recorded journey for any stone used in a monument at that period and Prof Peace says that the next mystery to solve is how it got there. “There are obvious physical barriers to transporting by land, and an equally daunting journey if going by sea. “These findings will have huge ramifications for understanding communities in Neolithic times, their levels of connectivity and their transport systems”.

The new research will be pored over by archaeologists working for English Heritage, which looks after Stonehenge, according to one of the monument’s senior curators, Heather Sebire. ”This discovery certainly implies that there were great social connections in Britain at the time,” she told BBC News. “It is phenomenal that the people of the time brought such a large stone all this way. They must have had a compelling reason to do it. “They had a sophisticated and developed society and so they probably had a spiritual side, just like we do“.

Source: https://www.bbc.com/news/articles/c207lqdn755o

Understanding the provenance of megaliths used in the Neolithic stone circle at Stonehenge, southern England, gives insight into the culture and connectivity of prehistoric Britain. The source of the Altar Stone, the central recumbent sandstone megalith, has remained unknown, with recent work discounting an Anglo-Welsh Basin origin1,2. Here we present the age and chemistry of detrital zircon, apatite and rutile grains from within fragments of the Altar Stone. The detrital zircon load largely comprises Mesoproterozoic and Archaean sources, whereas rutile and apatite are dominated by a mid-Ordovician source. The ages of these grains indicate derivation from an ultimate Laurentian crystalline source region that was overprinted by Grampian (around 460 million years ago) magmatism. Detrital age comparisons to sedimentary packages throughout Britain and Ireland reveal a remarkable similarity to the Old Red Sandstone of the Orcadian Basin in northeast Scotland. Such a provenance implies that the Altar Stone, a 6 tonne shaped block, was sourced at least 750 km from its current location. The difficulty of long-distance overland transport of such massive cargo from Scotland, navigating topographic barriers, suggests that it was transported by sea. Such routing demonstrates a high level of societal organization with intra-Britain transport during the Neolithic period. 

Stonehenge, the Neolithic standing stone circle located on the Salisbury Plain in Wiltshire, England, offers valuable insight into prehistoric Britain. Construction at Stonehenge began as early as 3000 bc, with subsequent modifications during the following two millennia3,4. The megaliths of Stonehenge are divided into two major categories: sarsen stones and bluestones (Fig. 1a). The larger sarsens comprise duricrust silcrete predominantly sourced from the West Woods, Marlborough, approximately 25 km north of Stonehenge5,6. Bluestone, the generic term for rocks considered exotic to the local area, includes volcanic tuff, rhyolite, dolerite and sandstone lithologies4 (Fig. 1a). Some lithologies are linked with Neolithic quarrying sites in the Mynydd Preseli area of west Wales7,8. An unnamed Lower Palaeozoic sandstone, associated with the west Wales area on the basis of acritarch fossils9, is present only as widely disseminated debitage at Stonehenge and possibly as buried stumps (Stones 40g and 42c).

The central megalith of Stonehenge, the Altar Stone (Stone 80), is the largest of the bluestones, measuring 4.9 × 1.0 × 0.5 m, and is a recumbent stone (Fig. 1b), weighing 6 t and composed of pale green micaceous sandstone with distinctive mineralogy1,2,10 (containing baryte, calcite and clay minerals, with a notable absence of K-feldspar) (Fig. 2).

Previous petrographic work on the Altar Stone has implied an association to the Old Red Sandstone10,11,12 (ORS). The ORS is a late Silurian to Devonian sedimentary rock assemblage that crops out widely throughout Great Britain and Ireland (Extended Data Fig. 1). ORS lithologies are dominated by terrestrial siliciclastic sedimentary rocks deposited in continental fluvial, lacustrine and aeolian environments13. Each ORS basin reflects local subsidence and sediment infill and thus contains proximal crystalline signatures13,14.

Constraining the provenance of the Altar Stone could give insights into the connectivity of Neolithic people who left no written record15. When the Altar Stone arrived at Stonehenge is uncertain; however, it may have been placed within the central trilithon horseshoe during the second construction phase around 2620–2480 bc3. Whether the Altar Stone once stood upright as an approximately 4 m high megalith is unclear15; nevertheless, the current arrangement has Stones 55b and 156 from the collapsed Great Trilithon resting atop the prone and broken Altar Stone (Fig. 1b).

An early proposed source for the Altar Stone from Mill Bay, Pembrokeshire (Cosheston Subgroup of the Anglo-Welsh ORS Basin), close to the Mynydd Preseli source of the doleritic and rhyolitic bluestones, strongly influenced the notion of a sea transport route via the Bristol Channel12. However, inconsistencies in petrography and detrital zircon ages between the Altar Stone and the Cosheston Subgroup have ruled this source out1,11. Nonetheless, a source from elsewhere in the ORS of the Anglo-Welsh Basin was still considered likely, with an inferred collection and overland transport of the Altar Stone en route to Stonehenge from the Mynydd Preseli1. However, a source from the Senni Formation (Cosheston Subgroup) is inconsistent with geochemical and petrographic data, which shows that the Anglo-Welsh Basin is highly unlikely to be the source2. Thus, the ultimate provenance of the Altar Stone had remained an open question.

Studies of detrital mineral grains are widely deployed to address questions throughout the Earth sciences and have utility in archaeological investigations16,17. Sedimentary rocks commonly contain a detrital component derived from a crystalline igneous basement, which may reflect a simple or complex history of erosion, transport and deposition cycles. This detrital cargo can fingerprint a sedimentary rock and its hinterland. More detailed insights become evident when a multi-mineral strategy is implemented, which benefits from the varying degrees of robustness to sedimentary transportation in the different minerals18,19,20.

Here, we present in situ U–Pb, Lu–Hf and trace element isotopic data for zircon, apatite and rutile from two fragments of the Altar Stone collected at Stonehenge: MS3 and 2010K.24021,22. In addition, we present comparative apatite U–Pb dates for the Orcadian Basin from Caithness and Orkney. We utilize statistical tools (Fig. 3) to compare the obtained detrital mineral ages and chemistry (Supplementary Information 1–3) to crystalline terranes and ORS successions across Great Britain, Ireland and Europe (Fig. 4 and Extended Data Fig. 1).

Laurentian basement signatures

The crystalline basement terranes of Great Britain and Ireland, from north to south, are Laurentia, Ganderia, Megumia and East Avalonia (Fig. 4a and Extended Data Fig. 1). Cadomia-Armorica is south of the Rheic Suture and encompasses basement rocks in western Europe, including northern France and Spain. East Avalonia, Megumia and Ganderia are partly separated by the Menai Strait Fault System (Fig. 4a). Each terrane has discrete age components, which have imparted palaeogeographic information into overlying sedimentary basins13,14,23. Laurentia was a palaeocontinent that collided with Baltica and Avalonia (a peri-Gondwanan microcontinent) during the early Palaeozoic Caledonian Orogeny to form Laurussia14,24. West Avalonia is a terrane that includes parts of eastern Canada and comprised the western margin of Avalonia (Extended Data Fig. 1).

Statistical comparisons, using a Kolmogorov–Smirnov test, between zircon ages from the Laurentian crystalline basement and the Altar Stone indicate that at a 95% confidence level, no distinction in provenance is evident between Altar Stone detrital zircon U–Pb ages and those from the Laurentian basement. That is, we cannot reject the null hypothesis that both samples are from the same underlying age distribution (Kolmogorov–Smirnov test: P > 0.05) (Fig. 3a).

Detrital zircon age components, defined by concordant analyses from at least 4 grains in the Altar Stone, include maxima at 1,047, 1,091, 1,577, 1,663 and 1,790 Ma (Extended Data Fig. 2), corresponding to known tectonomagmatic events and sources within Laurentia and Baltica, including the Grenville (1,095–980 Ma), Labrador (1,690–1,590 Ma), Gothian (1,660–1,520 Ma) and Svecokarellian (1,920–1,770 Ma) orogenies25.

Laurentian terranes are crystalline lithologies north of the Iapetus Suture Zone (which marks the collision zone between Laurentia and Avalonia) and include the Southern Uplands, Midland Valley, Grampian, Northern Highlands and Hebridean Terranes (Fig. 4a). Together, these terranes preserve a Proterozoic to Archaean record of zircon production24, distinct from the southern Gondwanan-derived terranes of Britain20,26 (Fig. 4a and Extended Data Fig. 3).

Age data from Altar Stone rutile grains also point towards an ultimate Laurentian source with several discrete age components (Extended Data Fig. 4 and Supplementary Information 1). Group 2 rutile U–Pb analyses from the Altar Stone include Proterozoic ages from 1,724 to 591 Ma, with 3 grains constituting an age peak at 1,607 Ma, overlapping with Laurentian magmatism, including the Labrador and Pinwarian (1,690–1,380 Ma) orogenies24. Southern terranes in Britain are not characterized by a large Laurentian (Mesoproterozoic) crystalline age component25 (Fig. 4b and Extended Data Fig. 3). Instead, terranes south of the Iapetus Suture are defined by Neoproterozoic to early Palaeozoic components, with a minor component from around two billion years ago (Figs. 3b and 4b).

U–Pb analyses of apatite from the Altar Stone define two distinct age groupings. Group 2 apatite U–Pb analyses define a lower intercept age of 1,018 ± 24 Ma (n = 9) (Extended Data Fig. 5), which overlaps, within uncertainty, to a zircon age component at 1,047 Ma, consistent with a Grenville source25. Apatite Lu–Hf dates at 1,496 and 1,151 Ma also imply distinct Laurentian sources25 (Fig. 4b, Extended Data Fig. 6 and Supplementary Information 2). Ultimately, the presence of Grenvillian apatite in the Altar Stone suggests direct derivation from the Laurentian basement, given the lability of apatite during prolonged chemical weathering20,27.

Grampian Terrane detrital grains

Apatite and rutile U–Pb analyses from the Altar Stone are dominated by regressions from common Pb that yield lower intercepts of 462 ± 4 Ma (n = 108) and 451 ± 8 Ma (n = 83), respectively (Extended Data Figs. 4 and 5). A single concordant zircon analysis also yields an early Palaeozoic age of 498 ± 17 Ma. Hence, with uncertainty from both lower intercepts, Group 1 apatite and rutile analyses demonstrate a mid-Ordovician (443–466 Ma) age component in the Altar Stone. These mid-Ordovician ages are confirmed by in situ apatite Lu–Hf analyses, which define a lower intercept of 470 ± 29 Ma (n = 16) (Extended Data Fig. 6 and Supplementary Information 2).

Throughout the Altar Stone are sub-planar 100–200-µm bands of concentrated heavy resistive minerals. These resistive minerals are interpreted to be magmatic in origin, given internal textures (oscillatory zonation), lack of mineral overgrowths (in all dated minerals) (Fig. 2) and the igneous apatite trace element signatures27 (Extended Data Fig. 7 and Supplementary Information 3). Moreover, there is a general absence of detrital metamorphic zircon grains, further supporting a magmatic origin for these grains.

The most appropriate source region for such mid-Ordovician grains within Laurentian basement is the Grampian Terrane of northeast Scotland (Fig. 4a). Situated between the Great Glen Fault to the north and the Highland Boundary Fault to the south, the terrane comprises Neoproterozoic to Lower Palaeozoic metasediments termed the Dalradian Supergroup28, which are intruded by a compositionally diverse suite of early Palaeozoic granitoids and gabbros (Fig. 4a). The 466–443 Ma age component from Group 1 apatite and rutile U–Pb analyses overlaps with the terminal stages of Grampian magmatism and subsequent granite pluton emplacement north of the Highland Boundary Fault28 (Fig. 4a).

Geochemical classification plots for the Altar Stone apatite imply a compositionally diverse source, much like the lithological diversity within the Grampian Terrane28, with 61% of apatite classified as coming from felsic sources, 35% from mafic sources and 4% from alkaline sources (Extended Data Fig. 7 and Supplementary Information 3). Specifically, igneous rocks within the Grampian Terrane are largely granitoids, thus accounting for the predominance of felsic-classified apatite grains29. We posit that the dominant supply of detritus from 466–443 Ma came from the numerous similarly aged granitoids formed on the Laurentian margin28, which are present in both the Northern Highlands and the Grampian Terranes28 (Fig. 4a). The alkaline to calc-alkaline suites in these terranes are volumetrically small, consistent with the scarcity of alkaline apatite grains within the Altar Stone (Extended Data Fig. 7). Indeed, the Glen Dessary syenite at 447 ± 3 Ma is the only age-appropriate felsic-alkaline pluton in the Northern Highlands Terrane30.

The Stacey and Kramers31 model of terrestrial Pb isotopic evolution predicts a 207Pb/206Pb isotopic ratio (207Pb/206Pbi) of 0.8601 for 465 Ma continental crust. Mid-Ordovician regressions through Group 1 apatite and rutile U–Pb analyses yield upper intercepts for 207Pb/206Pbi of 0.8603 ± 0.0033 and 0.8564 ± 0.0014, respectively (Extended Data Figs. 4 and 5 and Supplementary Information 1). The similarity between apatite and rutile 207Pb/206Pbi implies they were sourced from the same Mid-Ordovician magmatic fluids. Ultimately, the calculated 207Pb/206Pbi value is consistent with the older (Laurentian) crust north of the Iapetus Suture in Britain32 (Fig. 4a).

Orcadian Basin ORS

The detrital zircon age spectra confirm petrographic associations between the Altar Stone and the ORS. Furthermore, the Altar Stone cannot be a New Red Sandstone (NRS) lithology of Permo-Triassic age. The NRS, deposited from around 280–240 Ma, unconformably overlies the ORS14. NRS, such as that within the Wessex Basin (Extended Data Fig. 1), has characteristic detrital zircon age components, including Carboniferous to Permian zircon grains, which are not present in the Altar Stone1,23,26,33,34 (Extended Data Fig. 3).

An ORS classification for the Altar Stone provides the basis for further interpretation of provenance (Extended Data Figs. 1 and 8), given that the ORS crops out in distinct areas of Great Britain and Ireland, including the Anglo-Welsh border and south Wales, the Midland Valley and northeast Scotland, reflecting former Palaeozoic depocentres14 (Fig. 4a).

Previously reported detrital zircon ages and petrography show that ORS outcrops of the Anglo-Welsh Basin in the Cosheston Subgroup1 and Senni Formation2 are unlikely to be the sources of the Altar Stone (Fig. 4a). ORS within the Anglo-Welsh Basin is characterized by mid-Palaeozoic zircon age maxima and minor Proterozoic components (Fig. 4a). Ultimately, the detrital zircon age spectra of the Altar Stone are statistically distinct from the Anglo-Welsh Basin (Fig. 3a). In addition, the ORS outcrops of southwest England (that is, south of the Variscan front), including north Devon and Cornwall (Cornubian Basin) (Fig. 4a), show characteristic facies, including marine sedimentary structures and fossils along with a metamorphic fabric13,26, inconsistent with the unmetamorphosed, terrestrial facies of the Altar Stone1,11.

Another ORS succession with published age data for comparison is the Dingle Peninsula Basin, southwest Ireland. However, the presence of late Silurian (430–420 Ma) and Devonian (400–350 Ma) apatite, zircon and muscovite from the Dingle Peninsula ORS discount a source for the Altar Stone from southern Ireland20. The conspicuous absence of apatite grains of less than 450 Ma in age in the Altar Stone precludes the input of Late Caledonian magmatic grains to the source sediment of the Altar Stone and demonstrates that the ORS of the Altar Stone was deposited prior to or distally from areas of Late Caledonian magmatism, unlike the ORS of the Dingle Peninsula20. Notably, no distinction in provenance between the Anglo-Welsh Basin and the Dingle Peninsula ORS is evident (Kolmogorov–Smirnov test: P > 0.05), suggesting that ORS basins south of the Iapetus Suture are relatively more homogenous in terms of their detrital zircon age components (Fig. 4a).

In Scotland, ORS predominantly crops out in the Midland Valley and Orcadian Basins (Fig. 4a). The Midland Valley Basin is bound between the Highland Boundary Fault and the Iapetus Suture and is located within the Midland Valley and Southern Uplands Terranes. Throughout Midland Valley ORS stratigraphy, detrital zircon age spectra broadly show a bimodal age distribution between Lower Palaeozoic and Mesoproterozoic components35,36 (Extended Data Fig. 3). Indeed, throughout 9 km of ORS stratigraphy in the Midland Valley Basin and across the Sothern Uplands Fault, no major changes in provenance are recognized36 (Fig. 4a). Devonian zircon, including grains as young as 402 ± 5 Ma from the northern ORS in the Midland Valley Basin36, further differentiates this basin from the Altar Stone (Fig. 3a and Extended Data Fig. 3). The scarcity of Archaean to late Palaeoproterozoic zircon grains within the Midland Valley ORS shows that the Laurentian basement was not a dominant detrital source for those rocks35. Instead, ORS of the Midland Valley is primarily defined by zircon from 475 Ma interpreted to represent the detrital remnants of Ordovician volcanism within the Midland Valley Terrane, with only minor and periodic input from Caledonian plutonism35.

The Orcadian Basin of northeast Scotland, within the Grampian and Northern Highlands terranes, contains a thick package of mostly Mid-Devonian ORS, around 4 km thick in Caithness and up to around 8 km thick in Shetland14 (Fig. 4a). The detrital zircon age spectra from Orcadian Basin ORS provides the closest match to the Altar Stone detrital ages25 (Fig. 3 and Extended Data Fig. 8). A Kolmogorov–Smirnov test on age spectra from the Altar Stone and the Orcadian Basin fails to reject the null hypothesis that they are derived from the same underlying distribution (Kolmogorov–Smirnov test: P > 0.05) (Fig. 3a). To the north, ORS on the Svalbard archipelago formed on Laurentian and Baltican basement rocks37. Similar Kolmogorov–Smirnov test results, where each detrital zircon dataset is statistically indistinguishable, are obtained for ORS from Svalbard, the Orcadian Basin and the Altar Stone.

Apatite U–Pb age components from Orcadian Basin samples from Spittal, Caithness (AQ1) and Cruaday, Orkney (CQ1) (Fig. 4a) match those from the Altar Stone. Group 2 apatite from the Altar Stone at 1,018 ± 24 Ma is coeval with a Grenvillian age from Spittal at 1,013 ± 35 Ma. Early Palaeozoic apatite components at 473 ± 25 Ma and 466 ± 6 Ma, from Caithness and Orkney, respectively (Extended Data Fig. 5 and Supplementary Information 1), are also identical, within uncertainty, to Altar Stone Group 1 (462 ± 4 Ma) apatite U–Pb analyses and a Lu–Hf component at 470 ± 28 Ma supporting a provenance from the Orcadian Basin for the Altar Stone (Extended Data Fig. 6 and Supplementary Information 2).

During the Palaeozoic, the Orcadian Basin was situated between Laurentia and Baltica on the Laurussian palaeocontinent14. Correlations between detrital zircon age components imply that both Laurentia and Baltica supplied sediment into the Orcadian Basin25,36. Detrital grains from more than 900 Ma within the Altar Stone are consistent with sediment recycling from intermediary Neoproterozoic supracrustal successions (for example, Dalradian Supergroup) within the Grampian Terrane but also from the Särv and Sparagmite successions of Baltica25,36. At around 470 Ma, the Grampian Terrane began to denude28. Subsequently, first-cycle detritus, such as that represented by Group 1 apatite and rutile, was shed towards the Orcadian Basin from the southeast25.

Thus, the resistive mineral cargo in the Altar Stone represents a complex mix of first and multi-cycle grains from multiple sources. Regardless of total input from Baltica versus Laurentia into the Orcadian Basin, crystalline terranes north of the Iapetus Suture (Fig. 4a) have distinct age components that match the Altar Stone in contrast to Gondwanan-derived terranes to the south.

The Altar Stone and Neolithic Britain

Isotopic data for detrital zircon and rutile (U–Pb) and apatite (U–Pb, Lu–Hf and trace elements) indicate that the Altar Stone of Stonehenge has a provenance from the ORS in the Orcadian Basin of northeast Scotland (Fig. 4a). Given this detrital mineral provenance, the Altar Stone cannot have been sourced from southern Britain (that is, south of the Iapetus Suture) (Fig. 4a), including the Anglo-Welsh Basin1,2.

Some postulate a glacial transport mechanism for the Mynydd Preseli (Fig. 4a) bluestones to Salisbury Plain38,39. However, such transport for the Altar Stone is difficult to reconcile with ice-sheet reconstructions that show a northwards movement of glaciers (and erratics) from the Grampian Mountains towards the Orcadian Basin during the Last Glacial Maximum and, indeed, previous Pleistocene glaciations40,41. Moreover, there is little evidence of extensive glacial deposition in central southern Britain40, nor are Scottish glacial erratics found at Stonehenge42. Sr and Pb isotopic signatures from animal and human remains from henges on Salisbury Plain demonstrate the mobility of Neolithic people within Britain32,43,44,45. Furthermore, shared architectural elements and rock art motifs between Neolithic monuments in Orkney, northern Britain, and Ireland point towards the long-distance movement of people and construction materials46,47.

Thus, we posit that the Altar Stone was anthropogenically transported to Stonehenge from northeast Scotland, consistent with evidence of Neolithic inhabitation in this region48,49. Whereas the igneous bluestones were brought around 225 km from the Mynydd Preseli to Stonehenge50 (Fig. 4a), a Scottish provenance for the Altar Stone demands a transport distance of at least 750 km (Fig. 4a). Nonetheless, even with assistance from beasts of burden51, rivers and topographical barriers, including the Grampians, Southern Uplands and the Pennines, along with the heavily forested landscape of prehistoric Britain52, would have posed formidable obstacles for overland megalith transportation.

At around 5000 bc, Neolithic people introduced the common vole (Microtus arvalis) from continental Europe to Orkney, consistent with the long-distance marine transport of cattle and goods53. A Neolithic marine trade network of quarried stone tools is found throughout Britain, Ireland and continental Europe54. For example, a saddle quern, a large stone grinding tool, was discovered in Dorset and determined to have a provenance in central Normandy55, implying the shipping of stone cargo over open water during the Neolithic. Furthermore, the river transport of shaped sandstone blocks in Britain is known from at least around 1500 bc (Hanson Log Boat)56. In Britain and Ireland, sea levels approached present-day heights from around 4000 bc57, and although coastlines have shifted, the geography of Britain and Ireland would have permitted sea routes southward from the Orcadian Basin towards southern England (Fig. 4a). A Scottish provenance for the Altar Stone implies Neolithic transport spanning the length of Great Britain.

Methods

Overview

This work analysed two 30-µm polished thin sections of the Altar Stone (MS3 and 2010K.240) and two sections of ORS from northeast Scotland (Supplementary Information 4). CQ1 is from Cruaday, Orkney (59° 04′ 34.2″ N, 3° 18′ 54.6″ W), and AQ1 is from near Spittal, Caithness (58° 28′ 13.8″ N, 3° 27′ 33.6″ W). Conventional optical microscopy (transmitted and reflected light) and automated mineralogy via a TESCAN Integrated Mineral Analyser gave insights into texture and mineralogy and guided spot placement during LA-ICP–MS analysis. A CLARA field emission scanning electron microscope was used for textural characterization of individual minerals (zircon, apatite and rutile) through high-resolution micrometre-scale imaging under both back-scatter electron and cathodoluminescence. The Altar Stone is a fine-grained and well-sorted sandstone with a mean grain size diameter of ≤300 µm. Quartz grains are sub-rounded and monocrystalline. Feldspars are variably altered to fine-grained white mica. MS3 and 2010K.240 have a weakly developed planar fabric and non-planar heavy mineral laminae approximately 100–200 µm thick. Resistive heavy mineral bands are dominated by zircon, rutile, and apatite, with grains typically 10–40 µm wide. The rock is mainly cemented by carbonate, with localized areas of barite and quartz cement. A detailed account of Altar Stone petrography is provided in refs. 1,59.

Zircon isotopic analysis

Zircon U–Pb methods

Two zircon U–Pb analysis sessions were completed at the GeoHistory facility in the John De Laeter Centre (JdLC), Curtin University, Australia. Ablations within zircon grains were created using an excimer laser RESOlution LE193 nm ArF with a Laurin Technic S155 cell. Isotopic data was collected with an Agilent 8900 triple quadrupole mass spectrometer, with high-purity Ar as the plasma carrier gas (flow rate 1.l min−1). An on-sample energy of ~2.3–2.7 J cm−2 with a 5–7 Hz repetition rate was used to ablate minerals for 30–40 s (with 25–60 s of background capture). Two cleaning pulses preceded analyses, and ultra-high-purity He (0.68 ml min−1) and N2 (2.8 ml min−1) were used to flush the sample cell. A block of reference mineral was analysed following 15–20 unknowns. The small, highly rounded target grains of the Altar Stone (usually <30 µm in width) necessitated using a spot size diameter of ~24 µm for all ablations. Isotopic data was reduced using Iolite 460 with the U-Pb Geochronology data reduction scheme, followed by additional calculation and plotting via IsoplotR61. The primary matrix-matched reference zircon62 used to correct instrumental drift and mass fractionation was GJ-1, 601.95 ± 0.40 Ma. Secondary reference zircon included Plešovice63, 337.13 ± 0.37 Ma, 9150064, 1,063.78 ± 0.65 Ma, OG165 3,465.4 ± 0.6 Ma and Maniitsoq66 3,008.7 ± 0.6 Ma. Weighted mean U–Pb ages for secondary reference materials were within 2σ uncertainty of reported values (Supplementary Information 5).

Zircon U–Pb results

Across two LA-ICP–MS sessions, 83 U–Pb measurements were obtained on as many zircon grains; 41 were concordant (≤10% discordant), where discordance is defined using the concordia log distance (%) approach67. We report single-spot (grain) concordia ages, which have numerous benefits over conventional U–Pb/Pb–Pb ages, including providing an objective measure of discordance that is directly coupled to age and avoids the arbitrary switch between 206Pb/238U and 207Pb/206Pb. Furthermore, given the spread in ages (Early Palaeozoic to Archaean), concordia ages provide optimum use of both U–Pb/Pb–Pb ratios, offering greater precision over 206Pb/238U or 207Pb/206Pb ages alone.

Given that no direct sampling of the Altar Stone is permitted, we are limited in the amount of material available for destructive analysis, such as LA-ICP–MS. We collate our zircon age data with the U–Pb analyses1 of FN593 (another fragment of the Altar Stone), filtered using the same concordia log distance (%) discordance filter67. The total concordant analyses used in this work is thus 56 over 3 thin sections, each showing no discernible provenance differences. Zircon concordia ages span from 498 to 2,812 Ma. Age maxima (peak) were calculated after Gehrels68, and peak ages defined by ≥4 grains include 1,047, 1,091, 1,577, 1,663 and 1,790 Ma.

For 56 concordant ages from 56 grains at >95% certainty, the largest unmissed fraction is calculated at 9% of the entire uniform detrital population69. In any case, the most prevalent and hence provenance important components will be sampled for any number of analyses69. We analysed all zircon grains within the spatial limit of the technique in the thin sections70. We used in situ thin-section analysis, which can mitigate against contamination and sampling biases in detrital studies71. Adding apatite (U–Pb and Lu–Hf) and rutile (U–Pb) analyses bolsters our confidence in provenance interpretations as these minerals will respond dissimilarly during transport.

Comparative zircon datasets

Zircon U–Pb compilations of the basement terranes of Britain and Ireland were sourced from refs. 20,26. ORS detrital zircon datasets used for comparison include isotopic data from the Dingle Peninsula Basin20, Anglo-Welsh Basin72, Midland Valley Basin35, Svalbard ORS37 and Orcadian Basin25. NRS zircon U–Pb ages were sourced from the Wessex Basin33. Comparative datasets were filtered for discordance as per our definition above20,26. Kernel density estimates for age populations were created within IsoplotR61 using a kernel and histogram bandwidth of 50 Ma.

A two-sample Kolmogorov–Smirnov statistical test was implemented to compare the compiled zircon age datasets with the Altar Stone (Supplementary Information 6). This two-sided test compares the maximum probability difference between two cumulative density age functions, evaluating the null hypothesis that both age spectra are drawn from the same distribution based on a critical value dependent on the number of analyses and a chosen confidence level.

The number of zircon ages within the comparative datasets used varies from the Altar Stone (n = 56) to Laurentia (n = 2,469). Therefore, to address the degree of dependence on sample n, we also implemented a Monte Carlo resampling (1,000 times) procedure for the Kolmogorov–Smirnov test, including the uncertainty on each age determination to recalculate P values and standard deviations (Supplementary Information 7), based on the resampled distribution of each sample. The results from Kolmogorov–Smirnov tests, using Monte Carlo resampling (and multidimensional analysis), taking uncertainty due to sample n into account, also support the interpretation that at >95% certainty, no distinction in provenance can be made between the Altar Stone zircon age dataset (n = 56) and those from the Orcadian Basin (n = 212), Svalbard ORS (n = 619) and the Laurentian basement (Supplementary Information 7).

MDS plots for zircon datasets were created using the MATLAB script of ref. 58. Here, we adopted a bootstrap resampling (>1,000 times) with Procrustes rotation of Kolmogorov–Smirnov values, which outputs uncertainty ellipses at a 95% confidence level (Fig. 3a). In MDS plots, stress is a goodness of fit indicator between dissimilarities in the datasets and distances on the MDS plot. Stress values below 0.15 are desirable58. For the MDS plot in Fig. 3a, the value is 0.043, which indicates an “excellent” fit58.

Rutile isotopic analysis

Rutile U–Pb methods

One rutile U–Pb analysis session was completed at the GeoHistory facility in the JdLC, Curtin University, Australia. Rutile grains were ablated (24 µm) using a Resonetics RESOlution M-50A-LR sampling system, using a Compex 102 excimer laser, and measured using an Agilent 8900 triple quadrupole mass analyser. The analytical parameters included an on-sample energy of 2.7 J cm−2, a repetition rate of 7 Hz for a total analysis time of 45 s, and 60 s of background data capture. The sample chamber was purged with ultrahigh purity He at a flow rate of 0.68 l min−1 and N2 at 2.8 ml min−1.

U–Pb data for rutile analyses was reduced against the R-10 rutile primary reference material73 (1,091 ± 4 Ma). The secondary reference material used to monitor the accuracy of U–Pb ratios was R-19 rutile. The mean weighted 238U/206Pb age obtained for R-19 was 491 ± 10 (mean squared weighted deviation (MSWD) = 0.87, p(χ2) = 0.57) within uncertainty of the accepted age74 of 489.5 ± 0.9 Ma.

Rutile grains with negligible Th concentrations can be corrected for common Pb using a 208Pb correction74. Previously used thresholds for Th content have included75,76 Th/U < 0.1 or a Th concentration >5% U. However, Th/U ratios for rutile from MS3 are typically > 1; thus, a 208Pb correction is not applicable. Instead, we use a 207-based common Pb correction31 to account for the presence of common Pb. Rutile isotopic data was reduced within Iolite 460 using the U–Pb Geochronology reduction scheme and IsoplotR61.

Rutile U–Pb Results

Ninety-two rutile U–Pb analyses were obtained in a U–Pb single session, which defined two coherent age groupings on a Tera–Wasserburg plot.

Group 1 constitutes 83 U–Pb rutile analyses, forming a well-defined mixing array on a Tera-Wasserburg plot between common and radiogenic Pb components. This array yields an upper intercept of 207Pb/ 206Pbi = 0.8563 ± 0.0014. The lower intercept implies an age of 451 ± 8 Ma. The scatter about the line (MSWD = 2.7) is interpreted to reflect the variable passage of rutile of diverse grain sizes through the radiogenic Pb closure temperature at ~600 °C during and after magmatic crystallization77.

Group 2 comprises 9 grains, with 207Pb corrected 238U/206Pb ages ranging from 591–1,724 Ma. Three grains from Group 2 define an age peak68 at 1,607 Ma. Given the spread in U–Pb ages, we interpret these Proterozoic grains to represent detrital rutile derived from various sources.

Apatite isotopic analysis

Apatite U–Pb methods

Two apatite U–Pb LA-ICP–MS analysis sessions were conducted at the GeoHistory facility in the JdLC, Curtin University, Australia. For both sessions, ablations were created using a RESOlution 193 nm excimer laser ablation system connected to an Agilent 8900 ICP–MS with a RESOlution LE193 nm ArF and a Laurin Technic S155 cell ICP–MS. Other analytical details include a fluence of 2 J cm2 and a 5 Hz repetition rate. For the Altar Stone section (MS3) and the Orcadian Basin samples (Supplementary Information 4), 24- and 20-µm spot sizes were used, respectively.

The matrix-matched primary reference material used for apatite U–Pb analyses was the Madagascar apatite (MAD-1)78. A range of secondary reference apatite was analysed, including FC-179 (Duluth Complex) with an age of 1,099.1 ± 0.6 Ma, Mount McClure80,81 526 ± 2.1 Ma, Otter Lake82 913 ± 7 Ma and Durango 31.44 ± 0.1883 Ma. Anchored regressions (through reported 207Pb/206Pbi values) for secondary reference material yielded lower intercept ages within 2σ uncertainty of reported values (Supplementary Information 8).

Altar Stone apatite U–Pb results

This first session of apatite U–Pb of MS3 from the Altar Stone yielded 117 analyses. On a Tera–Wasserburg plot, these analyses form two discordant mixing arrays between common and radiogenic Pb components with distinct lower intercepts.

The array from Group 2 apatite, comprised of 9 analyses, yields a lower intercept equivalent to an age of 1,018 ± 24 Ma (MSWD = 1.4) with an upper intercept 207Pb/ 206Pbi = 0.8910 ± 0.0251. The f207% (the percentage of common Pb estimated using the 207Pb method) of apatite analyses in Group 2 ranges from 16.66–88.8%, with a mean of 55.76%.

Group 1 apatite is defined by 108 analyses yielding a lower intercept of 462 ± 4 Ma (MSWD = 2.4) with an upper intercept 207Pb/ 206Pbi = 0.8603 ± 0.0033. The f207% of apatite analyses in Group 1 range from 10.14–99.91%, with a mean of 78.65%. The slight over-dispersion of the apatite regression line may reflect some variation in Pb closure temperature in these crystals84.

Orcadian basin apatite U–Pb results

The second apatite U–Pb session yielded 138 analyses from samples CQ1 and AQ1. These data form three discordant mixing arrays between radiogenic and common Pb components on a Tera–Wasserburg plot.

An unanchored regression through Group 1 apatite (n = 14) from the Cruaday sample (CQ1) yields a lower intercept of 473 ± 25 Ma (MSWD = 1.8) with an upper intercept of 207Pb/ 206Pbi = 0.8497 ± 0.0128. The f207% spans 38–99%, with a mean value of 85%.

Group 1 from the Spittal sample (AQ1), comprised of 109 analyses, yields a lower intercept equal to 466 ± 6 Ma (MSWD = 1.2). The upper 207Pb/ 206Pbi is equal to 0.8745 ± 0.0038. f207% values for this group range from 6–99%, with a mean value of 83%. A regression through Group 2 analyses (n = 17) from the Spittal sample yields a lower intercept of 1,013 ± 35 Ma (MSWD = 1) and an upper intercept 207Pb/ 206Pbi of 0.9038 ± 0.0101. f207% values span 25–99%, with a mean of 76%. Combined U–Pb analyses from Groups 1 from CQ1 and AQ1 (n = 123) yield a lower intercept equivalent to 466 ± 6 Ma (MSWD = 1.4) and an upper intercept 207Pb/ 206Pbi of 0.8726 ± 0.0036, which is presented beneath the Orcadian Basin kernel density estimate in Fig. 4b.

Apatite Lu–Hf methods

Apatite grains were dated in thin-section by the in situ Lu–Hf method at the University of Adelaide, using a RESOlution-LR 193 nm excimer laser ablation system, coupled to an Agilent 8900 ICP–MS/MS85,86. A gas mixture of NH3 in He was used in the mass spectrometer reaction cell to promote high-order Hf reaction products, while equivalent Lu and Yb reaction products were negligible. The mass-shifted (+82 amu) reaction products of 176+82Hf and 178+82Hf reached the highest sensitivity of the measurable range and were analysed free from isobaric interferences. 177Hf was calculated from 178Hf, assuming natural abundances. 175Lu was measured on mass as a proxy85 for 176Lu. Laser ablation was conducted with a laser beam of 43 µm at 7.5 Hz repetition rate and a fluency of approximately 3.5 J cm−2. The analysed isotopes (with dwell times in ms between brackets) are 27Al (2), 43Ca (2), 57Fe (2), 88Sr (2), 89+85Y (2), 90+83Zr (2), 140+15Ce (2), 146Nd (2), 147Sm (2), 172Yb (5), 175Lu (10), 175+82Lu (50), 176+82Hf (200) and 178+82Hf (150). Isotopes with short dwell times (<10 ms) were measured to confirm apatite chemistry and to monitor for inclusions. 175+82Lu was monitored for interferences on 176+82Hf.

Relevant isotope ratios were calculated in LADR87 using NIST 610 as the primary reference material88. Subsequently, reference apatite OD-30678 (1,597 ± 7 Ma) was used to correct the Lu–Hf isotope ratios for matrix-induced fractionation86,89. Reference apatites Bamble-1 (1,597 ± 5 Ma), HR-1 (344 ± 2 Ma) and Wallaroo (1,574 ± 6 Ma) were monitored for accuracy verification85,86,90. Measured Lu–Hf dates of 1,098 ± 7 Ma, 346.0 ± 3.7 Ma and 1,575 ± 12 Ma, respectively, are in agreement with published values. All reference materials have negligible initial Hf, and weighted mean Lu–Hf dates were calculated in IsoplotR61 directly from the (matrix-corrected) 176Hf/176Lu ratios.

For the Altar Stone apatites, which have variable 177Hf/176Hf compositions, single-grain Lu–Hf dates were calculated by anchoring isochrons to an initial 177Hf/176Hf composition90 of 3.55 ± 0.05, which spans the entire range of initial 177Hf/176Hf ratios of the terrestrial reservoir (for example, ref. 91). The reported uncertainties for the single-grain Lu–Hf dates are presented as 95% confidence intervals, and dates are displayed on a kernel density estimate plot.

Apatite Lu–Hf results

Forty-five apatite Lu–Hf analyses were obtained from 2010K.240. Those with radiogenic Lu ingrowth or lacking common Hf gave Lu–Hf ages, defining four coherent isochrons and age groups.

Group 1, defined by 16 grains, yields a Lu–Hf isochron with a lower intercept of 470 ± 28 Ma (MSWD = 0.16, p(χ2) = 1). A second isochron through 5 analyses (Group 2) constitutes a lower intercept equivalent to 604 ± 38 Ma (MSWD = 0.14, p(χ2) = 0.94). Twelve apatite Lu–Hf analyses define Group 3 with a lower intercept of 1,123 ± 42 Ma (MSWD = 0.75, p(χ2) = 0.68). Three grains constitute the oldest grouping, Group 4 at 1,526 ± 186 Ma (MSWD = 0.014, p(χ2) = 0.91).

Apatite trace elements methods

A separate session of apatite trace element analysis was undertaken. Instrumentation and analytical set-up were identical to that described in 4.1. NIST 610 glass was the primary reference material for apatite trace element analyses. 43Ca was used as the internal reference isotope, assuming an apatite Ca concentration of 40 wt%. Secondary reference materials included NIST 612 and the BHVO−2g glasses92. Elemental abundances for secondary reference material were generally within 5–10% of accepted values. Apatite trace element data was examined using the Geochemical Data Toolkit93.

Apatite trace elements results

One hundred and thirty-six apatite trace element analyses were obtained from as many grains. Geochemical classification schemes for apatite were used29, and three compositional groupings (felsic, mafic-intermediate, and alkaline) were defined.

Felsic-classified apatite grains (n = 83 (61% of analyses)) are defined by La/Nd of <0.6 and (La + Ce + Pr)/ΣREE (rare earth elements) of <0.5. The median values of felsic grains show a flat to slightly negative gradient on the chondrite-normalized REE plot from light to heavy REEs94. Felsic apatite’s median europium anomaly (Eu/Eu*) is 0.59, a moderately negative signature.

Mafic-intermediate apatite29 (n = 48 (35% of grains)) are defined by (La + Ce + Pr)/ΣREE of 0.5–0.7 and a La/Nd of 0.5–1.5. In addition, apatite grains of this group typically exhibit a chondrite-normalized Ce/Yb of >5 and ΣREEs up to 1.25 wt%. Apatite grains classified as mafic-intermediate show a negative gradient on a chondrite-normalized REE plot from light to heavy REEs. The apatite grains of this group generally show the most enrichment in REEs compared to chondrite94. The median europium (Eu/Eu*) of mafic-intermediate apatite is 0.62, a moderately negative anomaly.

Lastly, alkaline apatite grains29 (n = 5 (4% of analyses)) are characterized by La/Nd > 1.5 and a (La + Ce + Pr)/ΣREE > 0.8. The median europium anomaly of this group is 0.45. This grouping also shows elevated chondrite-normalized Ce/Yb of >10 and >0.5 wt% for the ΣREEs.

Source: https://www.nature.com/articles/s41586-024-07652-1