This section started life simply reporting my ‘visit out of interest’. I had collected Anglian erratics on Gower that clearly had been transported by glacier ice on its way ‘towards Stonehenge’ from Pembrokeshire (see previous section). So I thought it would be interesting to see what is there. The report, however, developed into something of an essay, from trying to find out and write down what really is known about the Stonehenge ‘bluestones’. I found a jumble of confusing claims, counterclaims and manifest nonsense, with a good measure of misleading media hype thrown in. Heaven only knows what truly interested people in general think about how the bluestones got to Stonehenge; most people I talk to innocently subscribe to “remarkable human transport”, because that is what the BBC and several newspapers have sold, repeatedly. To save the reader from any suspense here, perhaps, cutting away some nonsense and applying Occam’s Razor (i.e., simplest tends to be best), I come out in favour of “mostly transport by ice”.
(I apologise for the unavoidable use of some technical terms, which I try to minimise or explain; I have favoured illustrations instead of text where possible. Reader please be aware that this is work in progress – substantially incomplete and liable to change! I am open to constructive criticism and discussion of my ideas – see the section on feedback – and would hope to be properly acknowledged if they are recycled.)
A visit to see the Stonehenge ‘bluestones’
September 1st 2025. With wonderful early morning light and a brief respite from persistent thick cloud and thundery showers, an ‘English Heritage Inner Circle Experience’ afforded a close look at Stonehenge and in particular its enigmatic bluestones. (‘Bluestones’ is a bucket term for the rocks there that aren’t the huge sarsen sandstones.) Main findings: it is worth the cost to be free to roam amongst the stones before the hordes arrive to view them from ‘outside’, most of the the bluestones are familiar friends from Pembrokeshire, and the common ‘spotted’ dolerite bluestone is not a rock type found amongst Gower erratics.
Figure 1. The site is dominated by the huge ‘sarsens’ of tough sandstone while the bluestones form a circle and a horse-shoe pattern within the megalith periphery. Anthony Johnson’s map gives the conventional stone numbering, locating, e.g., 80 as the only exposed non-sarsen sandstone. This one is known as the Altar Stone and is of Scottish origin, although, like the bluestones, its ‘path’ to Stonehenge is debatable (see below).
Stonehenge bluestones (Figs 2, 3 and 4) are mostly of spotted and unspotted dolerites, averaging more than 1 tonne with several perhaps up to about 3.5 tonnes. These occur with a few significantly smaller stones, of rhyolite, foliated rhyolitic tuff (silicic ignimbrite), and volcanic breccia. All these stones derive from rock outcrops in Pembrokeshire (e.g., Bevins & Ixer 2018), most notably the dolerites from the Preseli hills (SN14 32). The broken slab of pale green sandstone known as the Altar Stone weighs about 6 tonnes. Rock fragments matching the monumental stones are abundant in the vicinity of Stonehenge and those nearby have been interpreted to indicate manual ‘dressing’ (shaping) at the construction site. Although there are no ‘dressed’ rhyolite stones, some analysed fragments match and are interpreted to derive ultimately from a rhyolitic tuff outcrop at Craig Rhosyfelin (SN1167 3615), on the north flank of the Preseli hills. The so-called Newall Boulder excavated at Stonehenge and the buried stump of Stone 32d share this apparent origin (Bevins et al. 2023). Sandstone fragments also occur, similar to buried stones (40g and 42c), and on fossil evidence these originate elsewhere in north Pembrokeshire (Ixer et al. 2017, 2020; Bevins et al. 2020). Evidence of shaping of the dolerites, which are exceptionally tough rocks, is scarce, although small fragments of both spotted and non-spotted varieties have been found (Bevins et al. 2023). Sockets in dolerite lintel Stone 150 were hand made, while a groove in Stone 68 seems enigmatic, not clearly shaped by hand. Quite a few of the Stonehenge dolerites closely resemble the polyhedral (flat-faced and sharp-edged) joint columns and blocks that occur naturally amidst and fallen around their source outcrops in the eastern Preseli hills, but many of the bluestones are to an extent rounded, with some unquestionably of abraded round shapes like classic erratics (Fig. 4). As noted by many authors (e.g., Thorpe et al. 1991), the bluestone assemblage in its entirety is a mix that one would expect to have been deposited at the limit of an ice sheet, where large rocks had been transported on top (supra-glacially) with smaller more rounded stones having travelled within or at the base of it.
Figure 2. Stones 150 and 43 are of the common ‘spotted’ dolerite, while 62 is one of the only three non-spotted dolerites (the others are fallen Stones 44 and 45), with the cut ring showing where a core was taken for analysis. Above the core site the dolerite is freshly broken revealing some clean rock; elsewhere the surfaces are widely obscured by centuries of weathering and by diverse lichens. Stone 62 has been found to be compositionally similar to the eastern Preseli natural dolerite outcrop of Garn Ddu Fach (SN151 332). It is geochemically unlike a group of non-spotted stones 7 km west of there that was inferred to be part of an early monumental site and incorrectly interpreted to be its source for transport to Stonehenge (see below).
Figure 3. Broadly speaking*, Pembrokeshire igneous rocks are compositionally of two main types, dominated by the relatively silica-poor basalts, dolerites (as found here) and gabbros, with less abundant relatively silica-rich rhyolites and dacites. Stone 48 is a well-rounded erratic of rhyolite, with 46 being the only other exposed rhyolite (Stone 32d, a broken stump, is buried) but in this case with an indistinct foliation (left to right here). Stone 38 (not shown) is a poorly exposed, rounded, silicic volcanic erratic, and Stone 40 is another silicic rock, apparently a volcanic breccia and reported as dacite. *(There are intermediate compositions, but volumetrically they are far less abundant).
Figure 4. Twenty-four of the known forty-three bluestones at Stonehenge. These are sufficiently well exposed to show the range of original shapes, from polyhedral cooling-contraction-jointed columns of dolerite (top left) through to well-rounded erratics (bottom right), including the two rhyolites. Stone 150, perhaps originally a lintel bridging across upright stones, shows a hollow made to receive a locating boss (red arrow). A groove in the side of Stone 68 (white arrow) is of unknown origin. The photos were cropped from Simon Banton’s ‘The Stones of Stonehenge’ http://www.stonesofstonehenge.org.uk licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Sources of the Stonehenge ‘bluestones’
Spotted dolerite outcrops form most of the highest and main peaks in the eastern Preseli hills: Carn Menyn – Carn Gyfrwy at 365-350 m OD (Fig. 4), Carn Breseb at 320 m, Carn Goedog at 300 m (Fig. 5), and Cerrigmarchogion at 360-410 m. The outcrops occur in an area of roughly 4 km2 and probably represent a single intrusion, although the ground between them is partially covered by superficial debris. Thorough mineralogical and geochemical analyses by Richard Bevins and his colleagues effectively ‘fingerprint’ these outcrops as the unique source of the spotted dolerite bluestones at Stonehenge, which has led those favouring human transport to seek and infer one or more quarries here from which the stone was deliberately carried away.
Figure 5. A typical craggy summit outcrop of spotted dolerite at Carn Menyn, eastern Preseli (SN141 325). The rock naturally breaks into slabby monoliths with flat surfaces along original joints and rounded shoulders and sides where ‘weathered’ over millennia, conceivably for more than 450,000 years. The white spots are original plagioclase crystals or crystal clusters that became hydrothermally altered to clumps of fine-grained secondary minerals, which, in turn, were recrystallised during regional burial and heating. (See explanatory note below.) This rock type occurs locally as erratics and is the most common of the Stonehenge ‘bluestones’. It has, however, not been found as erratics on Gower or farther east in old Glamorganshire, and it is this ‘absence’ that has underpinned the inference of human transport.
Figure 6. Outcrop of spotted dolerite at Carn Goedog (SN128 332). The slope abounds with naturally formed and fallen slabby and polyhedral (multi-faceted) columnar, weathered monoliths. Dolerite intrusions like these in the Preseli hills are common widely in western and northern Britain where their upland outcrops invariably break, weather and collapse in similar fashion, without any human intervention.
Spot the difference
What is ‘spotted dolerite’? Igneous rocks form from molten material, known as magma, which can either be erupted via volcanoes at the Earth’s surface, where it tends to cool quickly, or be intruded at depth to cool more slowly. Magmas may initially partly crystallise so they contain individual free crystals (phenocrysts) or clusters of them (glomerocrysts). Preseli dolerites represent magmas intruded as sheets beneath erupted lavas of similar composition (Fishguard Volcanic Group) and within what had been sea-floor sediments, at the time probably retaining some water. Some of the magmas contained cm-scale free crystals of plagioclase, others did not. In the case of the developments of ‘spots’, it appears that late-stage volatiles and heat from the magma caused hot-water circulation (i.e., hydrothermal) that affected the cooling and crystallisation of the intrusion. Here fresh large crystals of plagioclase in the cooled magma became replaced by fine-grained secondary minerals. Later, during deeper burial with regional heating and increased pressure, owing to ‘mountain building’, the minerals recrystallised finally to form the spots we see today (Bevins et al. 2021). The Preseli spotted dolerite outcrops only occur in one small area of the eastern Preseli hills, forming the highest summits and conceivably belonging to a single intrusion now partially masked by superficial deposits.
Figure 7. A. Perfectly fresh plagioclase crystals (phenocrysts) with sharply defined edges in a fine-grained lava matrix (collected from a Chilean volcano). B. Rather dirty-looking recrystallised plagioclase crystal cluster in the Limeslade dolerite erratic on Gower (previous section Fig. 8), considered to be typical of crystal-bearing but non-spotted dolerites derived from Pembrokeshire. C and D. Examples of Preseli spotted dolerites showing the fuzzy outlines characteristic of an early phase of hydrothermal alteration. Scale bars are 1 cm. The distinguishing spots can be relatively sparse, as few as one per square metre, so small-diameter samples could easily miss them. (Photos C and D courtesy of Richard Bevins).
Bluestones transport
In the following sections I outline some of the propositions regarding the transport histories of the Stonehenge bluestones. There is robust geochemical evidence that the bluestones, except the Altar Stone, were derived from rock outcrops in Pembrokeshire. Similarly, there is no doubt that the Pembrokeshire source region of igneous rocks was over-ridden by glacial ice, much of it clearly part of the Irish Sea Ice Stream that also advanced over Gower and parts of old Glamorganshire some 450,000 years ago. I have no certain proof of how the bluestones arrived at Stonehenge, whether by ice, by human power, or perhaps by some hybrid of agents, but I do offer a simple model as a plausible explanation. I am certain, however, that the ‘debate’ has involved somewhat alarming instances of serious subjectivity and media hype unlikely to promote the lay person’s trust in ‘expert opinions and methods’, probably to the contrary.
Figure 8. Top. Map showing the penetration of the Anglian (450,000 years BP) Irish Sea Ice Stream into what is now the Bristol Channel and, in yellow, the probable ice-transport path based on unequivocal occurrences of Pembrokeshire igneous erratics on Gower and farther east as far as Flat Holm (see previous section). The unequivocal ‘Till with boulders, glacial sand and gravels’, in north Somerset, relates to the vicinity of Kenn and Court Hill described in detail by Hawkins & Kellaway (1971) after a field meeting when there were ongoing engineering excavations. They reported investigations related to construction of the M5 motorway where it crosses Court Hill (ST436 722). Here was found a deep sand and gravel-filled channel cutting across the limestone ridge, with laterally associated till on the north flank at up to 85 m OD, some 60 m above the present valley floor (SSSI now heavily vegetated; visited 17th October 2026). The ‘contentious reach’, which is between the proven ice penetration and Stonehenge, is an area where bluestone fragments are widely dispersed and interpreted by some as glacial debris and by others as human-scattered detritus. Bottom. Overland and sea routes proposed as alternative ways that humans transported bluestones from the Preseli sources to Stonehenge (modified from English Heritage Stonehenge by Julian Richards, p. 8, and the English Heritage Stonehenge website). An additional suggested sea route that circumnavigates the entire southwest peninsula, around Lands End, seems sensibly ignored. The blue arrow shows the likely path of bluestones transported high within the ice or on top of the ice (supra-glacially) from source outcrops forming one or more nunataks near the summits of the Preseli hills (see below).
Glacial progress
There is little point in reviewing all of the claims and counterclaims made in the previous century regarding the transport of ‘bluestones’ to Stonehenge (initially Judd in 1902 by ice, then Thomas in 1923 by humans). Kellaway (1971) argued strongly for glacial transport, citing fair evidence from erratics and from till containing striated boulders closely associated with periglacial deposits. He added interpretations of geomorphology, especially of the Mendip Hills, which perhaps was problematic, and he subsequently changed his original views. Scourse (1997) provides a usefully comprehensive review of all of this before an analytical discussion of the glacial transport hypotheses. The latter analysis comes out not to favour ice, seemingly because there is no clear evidence of glacial transport all the way to the vicinity of Stonehenge. He discusses the paucity of material evidence on Salisbury Plain, such as in unequivocal erratics, till or substrate scour marks, and a lack of evidence for ice having surmounted and somehow modified the Mendip Hills or the Wiltshire Downs escarpment and Chalk strata ‘on the way’. But, irrespective of arguable points, e.g., marginalising evidence from north Somerset, these perceived problems all go away if glacier ice, at its relatively thin and wasting maximum reach, really did deliver large robust rocks near enough to warrant their collection and removal to the ceremonial site. Here, now, this last remark may look like a rather surprising special plead, but, when all of the claims for human transport are stripped away for being without any foundation, ice is the only obvious agent, even if the last reach of it is not obviously or unequivocally recorded. Largely because of the media hype, which persistently favours the “extraordinary human endeavour”, ice transport currently gets practically no mention (e.g., the BBC Countryfile on Stonehenge of 14th December 2025 and the Channel 5 “Stonehenge: Secrets of the new stone” of 3rd March 2026), but that is misleading. Ice transport does warrant the light of reasonable consideration.
What follows here in no way settles the transport dispute. For the foreseeable future both sides will be able to claim a “dearth of direct evidence”, but what follows offers an extremely simple explanation seemingly overlooked before. It is glaciologically sound and makes it easy to shelve human transport, with all its difficult-to-imagine and impossible-to-disprove elements, and its once-claimed ‘evidences’ that are now debunked (e.g., Parker Pearson et al. 2010; John et al. 2015; Bevins et al. 2022). Occam’s Razor springs to mind. Later we will recall some of the media’s misdirected presentations of the human transport view as if it was well supported by archaeological or scientific facts. This has been a bizarre distraction at least, if not damaging.
Little bits of background. What follows has mostly been said before, but in making a fair case (hopefully) it seems worthy of coherent review here to set the scene. Ironically, in the course of writing this, a new paper was published that attempts to discount ice while it may in fact provide evidence best explained by ice having encroached Salisbury Plain. However, first we need to set the scene…
It is important to consider that if ice delivered rocks to within reach of Stonehenge 450,000 years ago then those stranded there and exposed to the elements for so long would be thoroughly weathered. Less robust types would probably be rendered to small fragments (‘chippings’), small weathered stones or soil in the interim, e.g., the 77 pieces of ‘granodiorite’ totalling 22 kg excavated at West Kennet, Avebury (Ixer et al. 2025). That long interval here involved several extremely cold episodes (glacial-periglacial; wooly mammoth times) and relatively warm episodes (interglacial; hippopotamus times).
Periglacial freeze-thaw shattering (thermoclastic fragmentation) typically produces angular stone chips and part of the Stonehenge transport controversy concerns whether or not the widely extensive occurrence of many thousands (>6,500) of bluestone and other fragment types on Salisbury Plain (e.g., Ixer & Bevins 2010) registers glacial delivery with subsequent periglacial material degradation. There has been a strong tendency amongst supporters of human transport to interpret all exotic rock fragments (not all of monolith stone types) as reflecting ‘dressing’ and accidental distribution by human hands, and thus to refer to them as ‘debitage’. This might be justified for monolith rock types in the vicinity of the monuments, but the interpretation potentially involves circular reasoning. As carefully detailed by Thorpe et al. (Table 2, 1991), there are bluestone and many other rock fragments widely scattered, e.g., 30 km north around Avebury and 3 km south near Lake. Together with pieces of known erratic types in Barrow Amesbury 39, a burial site that pre-dates erection of the large stones of Stonehenge, these all together suggest a natural distribution and render associated claims of human scattering as special pleading. Amongst the thousands of fragments there are non-bluestone rock types, such as basaltic tuff or ‘variable argillaceous lithic tuff’ (e.g. Ixer & Bevins 2010, 2016), which occurs in Pembrokeshire but is not known amongst the megaliths and would not have made robust monumental stones. And, in almost half a million years of big variations of temperature, precipitation and vegetation, any original post-Anglian land surface is likely to have been widely reworked and chemically or physically broken down. Thus the “clear evidence of glacial transport” may be more than should be expected. Further, from my own experience of mapping glacial deposits on Gower, one should not simply dismiss normally practiced field clearance of stones that superficially looks like original absence. Centuries of land clearance on Gower has led to apparently clean fields, with erratics dumped in hollows or forming field boundaries, and significant clearance over millennia is to be expected on the Wiltshire Downs.
Relatively recent excavations around Stonehenge (Allen et al. 2016; Leivers 2021) have revealed an extensive tract of monumental ground – “Sacred landscape” – and, for our particular interest, also an early bluestone henge site from which the stones apparently were removed to Stonehenge. The excavations found bluestone chips as well as buried areas of “periglacial frost-heave (cryoturbation) features primarily created by in situ freeze-thaw alteration of the chalk…”. These frost features are well described and known to be preserved widely in the gentle chalk slopes of Wiltshire. So we know of typical periglacial conditions here and that the bluestones seemingly were moved to be reused at the present site, as if they were not readily available for new supply from elsewhere.
If the stones were glacially delivered, it would explain why experts on Stonehenge archaeology have commented that the site was never fully completed although repeatedly restructured (Late Neolithic ca. 2950 BC to Early Bronze Age ca. 2,100 BC; Parker Pearson 2023). Based on the irregular gaps (Fig. 1) and the use of rather unremarkable and mixed types of small and rounded stones, it has been inferred that the source of substantial dolerite bluestone megaliths must have been “used up”, so that a full pattern of pillars was never completed. There are 43 stones known today; filling the gaps hypothetically yields 80.
One might guess that any actual quarry selection and transport by humans would have produced a more uniform or regular collection, and not as it seems to be, i.e., exactly what one would expect of a natural distribution, not all of robust dolerite and of varying shapes from columns to rounded irregular weathered forms (Fig. 4). However, we are reminded that we should not presume to know the Neolithic (2,500 BC) mind set… But, not sensibly to be dismissed, the huge effort, the very real physical risks, and the considerable time that would be taken in transporting so many (at least 43) unremarkable and heavy rocks so far inexorably seems contrary to our modern common sense. Repeated re-use of the seemingly insufficient bluestones over many centuries does not suggest their derivation from known accessible quarries or pre-existing monuments. Further, the argument that bluestone erratic material should exist in river gravels if it was ever delivered by ice is spurious, because there is no need to claim an original bluestone-containing cover reworked and eroded off Salisbury Plain. All that is needed is the gathering together of a few prominent, weathering-resistant, glacially deposited stones across old and heavily altered periglacial terrain from somewhere near enough to warrant the removal. Ironically, as we shall see, ice probably did deliver exotic material…
Conversely, in the human-transport view, Stonehenge would be the only known ritual celebration site where numerous (at least 43) pieces of non-local and not especially remarkable material, up to 3.5 tonnes in weight, were carried several hundreds of kilometres (overland some 300 km / 186 mi and by water 435 km / 270 mi). There is no known field record of this anywhere (e.g., Kellaway 1971; Thorpe & Williams-Thorpe 1991). We do know, however, that prominent, weather-resisting stones lying around within a largish area – perhaps 10s of kilometres away – definitely were commonly collected, brought together and carefully erected. Some of the Stonehenge sarsen stones, a few over 30 tonnes, are thought to have been collected from about 24 km away (Field et al. 2015; Harding et al. 2024; Daw 2025). In the Preseli area spotted dolerite stones were only used where they occurred locally, near to where they are found today, and there exists no evidence of them being especially revered. Motivation for the “stupendous feat” in human transport has always been a problematic weakness in the case, earlier attributing fantastic reasons like magical powers or sonic properties to the stones, or later mistakenly claiming reverence for them in sites of previous circles that then acted as sources for removal and transport (e.g., Parker Pearson et al. 2021). Fantastic claims, including inference of active quarrying to produce the stones at rock outcrops that are actually typical of natural jointing, weathering and collapse, are now, with sensible geology and geomorphology, and with robust geochemical evidence, thoroughly debunked. So, no quarries and no uprooting of former monuments (Bevins et al. 2022; John et al. 2015; John 2025).
It doesn’t really matter. Perhaps most notorious in the ‘debate’, and illustrating how confusing it can be if one is trying to understand what actually is known, is the apparent finding of a well-rounded spotted dolerite boulder (80 cm diameter, ~340 kg) built into the structure of Boles Barrow (a burial mound) 19 km west-northwest of Stonehenge. It was found by William Cunnington who excavated the barrow in 1801 and apparently it occurred together with other boulders of sandstone akin to the Stonehenge sarcens. He recorded [in Wm. Cunnington’s handwriting]: “I discover among them the Blue hard Stone also, ye same to some of the upright Stones in ye inner Circle at Stonehenge”. Contemporary letters note that sarcen boulders are common in the vicinity of the burial mound and also, from a sample flake of the “Blue hard Stone” sent to HH Thomas of the Geological Survey (he of human-transport conviction), that “There is no doubt at all that the specimen … is of the spotted Prescelly type and identical with the spotted Blue stones of Stonehenge.” (Cunnington 1924). Critically, the barrow’s construction predates Stonehenge by some 1000 years so that this boulder seems sensibly interpreted as a naturally rounded erratic and thus evidence of ice transport at least to near there. However, the range of convoluted and contradictory claims made through time, not helped by the boulder being collected and moved as a garden ornament more than once, and also the suggestion that it was stolen from Stonehenge for a garden ornament (subsequently refuted), is astonishing and mind-numbing. (For a fair reflection of this, search Boles Boulder in Brian John’s blog https://brian-mountainman.blogspot.com/ and see Burl (1999) and the Cunnington (1924) paper at https://archive.org/details/wiltshirearchaeo421922192/page/n483/mode/2up).
Similarly, there has been considerable ‘debate’ as to whether the so-called Newall Boulder, of rhyolite (22x15x10 cm) excavated at Stonehenge in 1924 and known to originate on the north flank of the Preseli range (Craig Rhosyfelin), is a weathered glacial erratic or a weathered fragment from a human-transported stone (Bevins et al. 2023, 2025; John 2024). Significantly, it was in a collection of small rock fragments that included spotted dolerite, dolerite, andesite, rhyolite and sandstone, and even 100 years ago these were of disputed origin. Newall (1959) thought they were glacial erratics. The abundant fragmental rhyolitic debris around Stonehenge that also derives from Craig Rhosyfelin and is taken to be waste (debitage) from stone reworking (Bevins et al. 2012), also suffers from unknowable transport to the site. However, these contentious and ambiguous cases are not materially relevant to any human- versus ice-transport judgement. Similarly irrelevant is the inferred human transport of the Altar Stone, a (broken) sandstone slab interpreted to have come from northeast Scotland. As explained below, this stone is more likely to have come from western Scotland, and, like the Ailsa Craig erratics, delivered via the Irish Sea Ice Stream. The ‘evidence’ seems always ambiguous, so perhaps a really simple model explanation could be favoured… But first a simplified excursion into the fantastic technological realm of precise age determination that is now being used to ‘fingerprint’ rocks and their original geological setting.
Zircon-age ‘fingerprinting’
Zircon (ZrSiO4) is an extraordinarily robust and fairly common mineral, the original crystallisation of which readily accommodates trace amounts of the elements uranium and thorium (U and Th) along with its zirconium (Zr), but not lead (Pb). Without going into details, it is the long-living radioactive decay of isotopes of U and Th and the associated accumulation of ‘daughter-element’ isotopes, especially of Pb, that make zircon especially suitable for dating its crystallisation. Because zircon does not accommodate Pb during the crystal growth, the abundance of Pb in zircon (assuming no loss) is taken as a measure of the lifetime of its production within the mineral from U; hence we talk of U-Pb dating.
It is well known now that zircon crystals, which typically are tiny (sub-mm-scale), may retain cores with element isotope contents that can reflect formation billions of years ago (e.g., >4,000 million years), and that subsequent crystal growth events may form rims around the cores with different ages, in some instances more than once. Sophisticated (and expensive) microanalysis can, with extreme care, measure the isotope compositions of parts of zircons thus yielding their growth or recrystallisation histories. Because zircon crystallises or dissolves only at the high temperatures of molten rock (magma) or those deep in the Earth’s crust, ideally the age data should reflect magmatic events or high-temperature recrystallisation (high-grade metamorphism) typically associated with major crust-modifying events, as in volcanic activity and mountain building.
Growth and dissolution of zircons depends on the composition of its host, presence of ‘water’ (dissociated), pressure and temperature, typically in the range of about 600-900oC and up to 1100oC. When magmatic or deep crustal rocks are uplifted and eroded at the Earth’s surface the contained zircons are typically liberated and become free ‘detrital’ zircons. Any population of such tiny grains may become reworked as sediments and widely dispersed. A collection of detrital zircons, therefore, may include grains with variably differing ages, and aggregated age data thus may form a spectrum of ages (Figs 9 and 10).
The ~25 km-thick crust of the Earth beneath northern Scotland, broadly north of Glasgow, belonged to a continent named Laurentia, which, via plate-tectonic movements and continental collision, parked itself adjacent to crustal segments to the south that had utterly different evolutionary histories. Magmas and metamorphic rocks formed within Laurentia thus carry zircons with age spectra different to more southerly counterparts. The zircons released from rocks uplifted and eroded to form sediments in northwest Scotland are thus also distinctive for their zircon-age spectra (Fig. 9), with the peaks in them giving the ‘fingerprint’ that is recognisably of Laurentian crust. To clarify the jargon: for ‘spectra’ and ‘kernel density estimate plot’ (a statistical-cum-graphical treatment), simply read ‘pattern’.
Figure 9. Kernel density estimate plot of 2,469 published detrital zircon ages from the British part of Laurentia (from Stevens & Baykal (2021) who used the age data of Cawood et al. 2003, 2012; Friend et al. 2003; Kirkland et al. 2008; Waldron et al. 2008, 2014; McAteer et al. 2010; Strachan et al. 2013; Johnson et al. 2016). The names above the main peaks in the plot are those given in Britain to the major episodes of crustal upheaval in mountain building and production of magmas when zircons would grow to record the associated high temperatures and thus the formation ages. The anorthosites are interpreted as formed from magma intruded about 1,877 million years ago, which in turn had inherited zircons dated at about 2,491 million years (*Mason et al. 2004; Mason 2015).
Figure 10. Simplified plots from Stevens & Baykal (2021 and references therein) graphically showing how the Laurentian continental crust has a fundamentally different history of formation and a distinctive zircon-age ‘fingerprint’ to match. The three lower plots are from southern Britain with different crust.
The far-fetched Altar Stone
This broken 6 tonne sandstone slab has been considered as one of the bluestones since it was first thought to come from Pembrokeshire along with the others. Cutting through a great deal of painstaking work that was done to locate a source, finally, in 2024, its Scottish origin was announced (Clarke et al. 2024; but see also Bevins et al. 2024). There was considerable media attention, primarily centred then on the colossal apparent transport distance from north-easternmost Scotland to Stonehenge. Human transport by sea was favoured rather than such a great distance over land. The Scottish origin is not disputed here, as it was reasonably founded on the similarity of age spectra found in zircons of the Altar Stone and those in Scottish sandstones that occur on both sides of the Great Glen Fault (see below). However, the colossal distance of some 750 km for human transport, far greater than ever contemplated previously, is fairly simply dismissed as improbable in the light of overlooked geology (see below).
Most recently it has been speculated that Devensian ice could have carried the Altar Stone from northeast Scotland to the (then) landmass of Dogger Bank, currently in the middle of the North Sea (Clarke et al. 2026), but even then there remains a major (400 km) human-transport issue. Although modelling of possible erratic transport by Devensian ice provides an interesting perspective well founded in field observations, it has no bearing on the more extensive and less well preserved Anglian glacial record, which the human-transport ‘community’ incline to ignore. Unsurprisingly the media jumped on this North Sea glacial transport suggestion, so now human communities of ‘Doggerland’ are added in the archaeologist-conjured mix of nation-wide social cooperation and party going centred about Stonehenge.
The figures and their captions below report material facts and, despite obvious uncertainties of interpretation, hopefully tell a simple and plausible story that warrants recognition that the most likely transport of the Altar Stone and Preseli bluestones to the vicinity of Stonehenge was by Anglian ice. The utilised zircon-age fingerprinting is explained briefly above.
Figure 11. These are illustrations from my studies in the Grampian Highlands of Scotland together with Quentin Crowley and PhD student Joanne Neilson (Neilson 2008; Neilson et al. 2009). A shows the peak of Clach Leathad (1099 m OD; viewed from Rannoch Moor), which is composed of a beautiful monzogranite that was intruded 418 million years ago (B). Our work was to determine the history of volcanoes and the intrusions of granite (s.l.) into them, within the Grampian Terrane of Scotland. The study area is within the red box in C. Our rock samples yielded zircon crystals (D) from which ages of intrusion were determined by Uranium-Lead (U-Pb) dating. The zircon crystals, which are extraordinarily robust and persistent through geological time, were found commonly to contain ancient cores within them that had been picked up (inherited) from the deep crustal source where the molten material originated (see also Oliver et al. 2008). These inherited crystal cores confound the analyses of recorded dates of eruption and intrusion of magmas and have to be excluded if possible, as they preserve their original chemistry and older ages. However, we know that rapid crustal uplift and erosion of the volcanoes and intrusions here caused these zircons to become free detrital grains in the associated sediments. These formed sedimentary rock layers (E) akin to those that form the Altar Stone, but here they occur in the southwest Grampian Highlands. The views are flaggy brown sandstones and red siltstones on the island Kerrera, within the study area. These are interbedded with conglomerates derived by erosion of the ‘granites’ (sl) and associated volcanics in the Grampian Terrane. The right-hand image is of the ‘fish bed succession’ from which 3 samples were collected for detrital zircons, which, as anticipated, were found to contain strong signals of Grenvillian and Laxfordian inheritance just like the Altar Stone (Brookfield et al. 2024). (By the way: the igneous rocks are notably rich in barium (Ba; Neilson et al. 2009), which would tally with the distinctive abundance of barite in the Altar Stone matrix.) In essence, ‘our igneous zircons’ and those detrital zircons sampled on Kerrera contain the same ‘fingerprints’ (age spectra involving cores inherited from deep Laurentian crust of Scotland) as those that Clarke et al. (2024, figure 4) used to invoke the northeastern Scottish source for the Altar Stone. While those authors made comparisons with other crustal terranes of Britain, they did not consider sediments with similar age spectra that are widely exposed in the same terrane in southwestern Scotland, e.g., on Kerrera.
Figure 12. Map showing the area (yellow overlay) in which sediments derived from erosion of the mapped Highland igneous rocks can be expected to have formed sedimentary rocks containing zircons bearing the same compositional fingerprint. It includes the Orcadian Basin sedimentary rocks in the far northeast, which Clarke et al. (2024) claimed as the “Altar Stone Source”, but it also shows the area of our detailed study to the southwest, in the red box, within which the purple ellipse locates sedimentary rocks known to be similarly derived and amongst fair possibilities for a southwestern source of the Altar Stone. Here we use the evidence from the Scottish erratics found on Gower that record transport southwards across potential Alter Stone source areas. (See previous section ‘The Scottish connection’ figures 13 and 14.) Known Gower erratic sources: the red triangle marks Ailsa Craig, the blue hexagon marks the main Hebridean source of garnet-amphibolite gneiss and anorthosite, the blue diamond is the general vicinity of the Dalradian schist, and the green circle marks the Sgùrr of Eigg source of pitchstone.
Figure 13. The presence on Gower of erratics from western Scotland north of the distinctive-zircon-bearing sandstones provides an extension to the transport route that significantly simplifies the Altar Stone saga. We know Scottish erratics reached Gower where they occur together with Pembrokeshire-derived stones, which also occur as far east as Flat Holm. It is unknowable precisely how transport from the Hebridean Terrane to join the Anglian Irish Sea Ice Stream occurred. However, it seems most probable that earlier glaciations carried the Hebridean rocks broadly southwestwards and interglacial sea level rises (marine transgressions) and storms drove such debris back eastwards. This latter mechanism accounts for granite pebbles from Lundy Island, 50 km southwest of Gower, being driven ashore on Gower (Jenkins et al. 1985). Certainly, material with a ‘Laurentian’ origin did reach Gower and hence conceivably farther east. Simplification of the entire transport debate warrants consideration in the face of zero evidence for human involvement.
Finally on this topic, it should be clear now where this is leading (Fig. 13). Sandstones with zircons characteristic of Laurentian crust occur in southwest Scotland. Anglian ice known to form the Irish Sea Ice Stream has brought Ailsa Craig and also Hebridean and Grampian rocks to Gower and so it would seem far simpler to invoke glacial transport of the Altar Stone from southwest Scotland to South Wales, and then on, with the rest, … ‘towards Stonehenge’. With apologies to the Doggerlanders, Occam’s Razor is just fine!
Zircon irony
A recent paper on detrital zircon and apatite grains in 4 samples of river sediments from drainages on Salisbury Plain, widely around Stonehenge (Clarke and Kirkland 2026), finds age spectra consistent with the Laurentian basement of northern Britain, just like the Altar Stone. It argues that the zircons were derived from mass wasting of originally overlying strata (Thanet Formation 59.2-56 Ma old), and that the paucity of more southerly-derived material indicates that ice transport was not involved. The arguments are tortuous, subjective, and closed to ice. They reported that 85% of their zircon ages reflect Laurentian origins, with only 8% of Phanerozoic age indicative of more southerly sources, including one of Ordovician age (~464+/-16 Ma) that would be consistent with Pembrokeshire bluestone provenance*.
*(As an aside, it is worth considering what the authors actually said in reference to this one Ordovician signature; it underscores the loaded and prejudiced nature of this poor science and misleading writing : “The extreme rarity of this signature underscores the negligible contribution of southwest Welsh, or specifically Preseli, crystalline rock input to the detrital cargo of Salisbury Plain. If ice of sufficient thickness and extent had transported tens of large, multi-tonne bluestone erratics from west Wales to Salisbury Plain, it would be expected to leave a pronounced Darriwilian (Ordovician) zircon signature in the fine detrital fraction; its near absence instead indicates that large-scale glacial transport did not occur.” There are several subjective-cum-illogical remarks in this. The “extreme rarity” is within their collection and unproven as generally true, because the sampling was only in the broad vicinity around Stonehenge. The “specifically Preseli” is loaded and unjustified, as if Preseli could be expected to make a special contribution. And on the nature and extent of the ice … “If ice of sufficient thickness etc., …” is a ‘straw man’ of overstatement and expectation. “… expected to leave a pronounced… zircon signature” is an unwarranted supposition given the samples taken, while “near absence” … “indicates large-scale glacial transport did not occur” is absolute nonsense. Are we to suppose that the single Ordovician zircon reflects small-scale glacial transport, or that it fell from a human conveyor?)
With fine irony and on the contrary, Clarke and Kirkland (op. cit.) may have provided evidence in favour of ice transport. Their analysed minerals are from drainages around Stonehenge and do not include sites farther west where ice could have dumped collectable bluestones. However, it was interesting to me to find that igneous erratics from the potential southerly sources of the English Lake District and Snowdonia have not been found on Gower, while Scottish rocks liable to carry Laurentian basement signatures do appear to dominate (see previous chapter). Now, Clarke and Kirkland (2026) used the zircon-age spectra that had been compiled from numerous authors by Stevens and Baykal (2021) in determining the sources of the Thanet Formation they had sampled in Kent, at Pegwell Bay (P in Fig. 14 below). Stevens and Baykal’s figure 1 shows the sediment pathways inferred as contributing zircons to their Kentish samples. These pathways, reproduced precisely here in Figure 14, include those of Scottish zircons possibly added to the Thanet Formation widely in southeast England via longshore drift. However, the figure (Fig. 14) also shows that it is highly unlikely that zircons with Laurentian (Scottish-only) ages could form some 85% in the populations sampled by Clarke and Kirkland (2026) in the vicinity of Stonehenge (Fig. 14). Thus, without exceptional special pleading, the Thanet Formation mass-wasting argument is flawed, while the predominantly Scottish signature with the one Ordovician age is more probably a strong indicator of Anglian ice having reached Salisbury Plain via Pembrokeshire.
Figure 14. Black arrows are from Stevens and Baykal (figure 1, 2021) showing “possible transport pathways of the Thanet Formation sediments” to account for the zircons they recovered at Pegwell Bay (P). This ‘palaeogeography’ is based on Gibbard and Lewin (2003, 2016) and shows that it is highly unlikely that Scottish-Laurentian zircons would dominate populations collected from Salisbury Plain around Stonehenge (S) if they were derived from the Thanet Formation as suggested by Clarke and Kirkland (2026). Blue diamond BS marks the location of sampled Early Devonian Brownstone rock that has yielded Laurentian zircons (Waldron et al. 2025), but this source would not form the populations reported. We know that Scottish Highland erratics are (surprisingly) common on Gower (G) and the fine-sediment fraction carried by the Anglian Irish Sea Ice Stream across Gower inevitably would have carried Laurentian zircons too…
Now it remains to account for the lack of spotted dolerite erratics in the trail from southwest Wales to Salisbury Plain. This lack has underpinned the human-transport case since it was first highlighted and interpreted by HH Thomas in 1923.
A revised glacial transport model
Scourse (1997) discusses at some length the glacial transport favoured and explained by Thorpe et al. (1991). They had advocated transport of the dolerite bluestones as free boulders on top of the ice (supra-glacially) from the Preseli summits, having been collected there from nunatak(s) and with fines lost through meltwater drainage. That notion was deemed theoretically reasonable by Scourse but seemingly fell at the hurdle of ‘no clear evidence of transport all the way to the vicinity of Stonehenge.’ Thorpe et al. also noted this as an apparent problem.
We know that glacial till as well as glacial sand and gravel occur at up to 85 m OD in north Somerset (Fig. 15; Hawkins & Kellaway 1971; Gilbertson & Hawkins 1978) and that there is some evidence in nearby hill-top drift that ice reached ≥120 m OD (Colbourne et al. 1974), so ice transport most of the way seems feasible. And physically, irrespective of the basal topography passed over by an ice sheet, ice will flow according to the slope of the top surface. At Preseli the inferred source nunatak(s) are at about 300 m OD, while between the Mendip Hills and the Cotswolds elevations widely are less than 100 m (Fig. 15). The Bristol Channel lobe of the Irish Sea Ice, at its maximum reach and within the sheet’s ablation zone (not added to by snowfall), would have thinned eastwards to some extent before completely wasting away.
It is recognised now (e.g., Lee & Robertson 2025) that what once might have been recorded as a glacial limit is commonly better referred to as the ‘limit of preserved evidence for glaciation’, because, especially with ancient glaciations, the evidence is commonly obscured or obliterated. The oft-repeated claims of there being no evidence of ice having encroached Salisbury Plain are extremely flimsy in the face of the material found widely and in abundance there (see above; e.g. the large amounts of various exotic fragments subjectively dismissed as ‘debitage’, including non-bluestone types, those found in burial mounds and in their soil, and the Newall and Boles boulders).
In the same vein, it is a common mistake to interpret “ice-marginal glacial sand and gravel” as necessarily deposited at or beyond the maximum reach of an ice sheet. They will be deposited also throughout ice-sheet retreat (recession), in some instances with the outwash streams and solifluction erasing evidence of the full-advance extent. This has happened on Gower… Erasure similarly applies where ground once covered by ice tends to be reworked under periglacial conditions of frost heave and material shattering before ultimately becoming vegetated and bioturbated.
Figure 15. Physical context of the contentious reach west of Stonehenge showing the range over which huge stones could be moved and the location of Boles Barrow, where the unequivocal erratic of spotted dolerite (80 cm diameter and ~340 kg) is recorded as having been found as part of a burial mound that was built some 1000 years before Stonehenge. Perhaps not so contentious after all? (Maps: https://www.openstreetmap.org/copyright).
A really simple explanation
There is not a shred of evidence that humans transported any ‘bluestones’ all the way from Pembrokeshire to Stonehenge. The human-transport view initiated with HH Thomas (1923) and thereafter persistently centred on the lack of spotted dolerites in the erratic trail eastwards beyond Pembrokeshire, plus the seeming absence of evidence of ice movement onto Salisbury Plain. The claimed bluestone quarries and previous-monument sources have been disproved and the bizarre embellishment concerning the transport by cattle similarly is unwarranted. The Altar Stone transport from Scotland most sensibly also did not involve humans either, and it is more readily understood if derived via a western route (Fig. 13). We know rocks from western Scotland were transported in the Irish Sea Ice Stream of Anglian times and that ice did transport rocks from Pembrokeshire towards Stonehenge. It certainly penetrated at least as far east as what is now Flat Holm in the Bristol Channel (Fig. 8) and probably also Court Hill in north Somerset (Fig. 15). Thus, in the absence of any evidence to the contrary and with ice reaching so close, surely all this begs attempting to explain the selective ice transport and the possible reach of the ice.
The fact that most Stonehenge bluestone megaliths come from the high flanks and summits of the Preseli hills speaks volumes regarding their transport. The characteristic spotted dolerite is unique to these near-summit outcrops and open spaces between them are smoothly rounded by the erosive passage of ice (Fig. 16 top). Any visit to the eastern Preseli summit area outcrops reveals abundant loose rocks with joint-fracture faces defining blocks, slabs and columns of the same general scale as the Stonehenge dolerite megaliths (Fig. 16 bottom). In contrast, most of the Pembrokeshire erratics on Gower and farther east clearly come from low-ground sources, e.g., Ramsey, Skomer, and St David’s, before these became coastal.
Figure 16. Top. Glacially eroded and smoothed slabby outcrops occur between the craggy outcrops of the Preseli dolerites, recording the north-to-south passage of ice (right to left here). Bottom. Outcrop of spotted dolerite at Carn Goedog (SN128 332). The slope abounds with naturally formed and fallen slabby and polyhedral (multi-faceted) columnar, weathered monoliths.
Here it is hypothesised that the Preseli summits at the climax of glaciation were nunataks from which bluestone megaliths were collected (Figure 17). From there they were transported, on or near the ice top surface, to be deposited on Salisbury Plain, near enough to the vicinity of Stonehenge to warrant human collection and their first arrangement.
Figure 17. Illustration of the specific collection and transport of erratics from a nunatak (inset is in Alaska, photo provided by National Park Service). The source outcrops naturally yield blocks, slabs and columns with flat-angled facets (angular polyhedral shapes), some of which arrived little-modified somewhere on or near Salisbury Plain. The size-range of the dolerite blocks is characteristic of the cooling-contraction jointing that formed them; there are no bus-sized unfractured blocks and so never any huge dolerite erratics.
Key to this hypothetical model is that collection from Preseli nunataks and delivery to Salisbury Plain occurred during a glacial maximum (Fig. 18 top), and that during the subsequent recession and ice thinning no more spotted dolerites were collected and thus none were deposited in the erratic trail back to the source (Fig. 18 bottom).
Figure 18 (obviously not to scale). Model following the inference by Thorpe et al. (1991) that the spotted dolerite bluestones were transported on or near the ice surface (supra-glacially) from the Preseli summit source when it was a nunatak (i.e., an outcrop of rock standing proud of a surrounding ice sheet). This collection from the nunatak(s) could only occur when the ice was sufficiently thick to reach up there. Throughout the glacial activity erratics picked up from lowland sources become rounded and reduced downstream in the base of the ice (sub-glacially). During recession, however, ice-sheet thinning halted the collection of spotted dolerite from the Preseli nunatak(s), explaining why it is absent in the erratic populations found back towards Pembrokeshire. This model allows delivery of large bluestone erratics close enough to be collected and taken to build the various versions of arrangements near or at Stonehenge. The model also can account for the apparent insufficiency of ‘decent’ stone to finish the circle(s) and the need then to use the various lesser stones, such as the rounded erratics of rhyolite.
The problem of knowing how far the ice delivery of spotted dolerites penetrated towards Stonehenge remains ‘difficult’ and possibly warrants further research. Perhaps so many people have taken it for granted that transport was by humans that there has been little recent focus on the glacial-periglacial geomorphology. But, if one takes away all of the now-disproved ‘evidences’ of human quarrying and transport, then much of the abundant so-called ‘debitage’ so widely distributed across Salisbury Plain, together with the odd contentious occurrences of small stones and boulders, all naturally become glacial in origin. This is entirely consistent with the interpretation (above) that the Scottish zircons recovered from streams around Stonehenge strongly reflect the ice stream that originated in western Scotland and crossed Pembrokeshire and Gower on the way to Salisbury Plain.
As previously admitted, this view will not settle the transport dispute, because of a “dearth of direct evidence”, but it certainly is about time that transport by ice gained its warranted recognition as by far the more likely on available evidence.
Headless chickens
From sublime to ridiculous. What follows is a brief review of some of the extraordinary headline claims made about Stonehenge, particularly involving Pembrokeshire and the bluestones.
“Recent excavations at two … bluestone sources – one for rhyolite and one for spotted dolerite – have identified evidence of megalith quarrying around 3000 BC” (Parker Pearson et al. 2017, 2020).
Fair enough, since Preseli outcrops are the long-recognised sources of Stonehenge dolerite megaliths, one would expect dedicated Stonehenge archaeologists to go and see them. However, nothing about those virtually pristine igneous rock outcrops could lead a field geologist to consider them anything other than naturally formed, including their aprons of fragmental debris. There are hundreds of such dolerite outcrops in the high grounds of British hills. They typically show jointing-related break up and separation, involving freeze-thaw mechanical processes and weathering for tens of thousands to hundreds of thousands of years, and with probable ice-related movement, plucking and scattering blocks. The extreme toughness of the rock is reflected in the fact that many of our high summits and prominent crags are of dolerite. Most of the Preseli tops are dolerite, while classic examples form the summits of Cader Idris (south Snowdonia), Moel Siabod (central Snowdonia) and Tal y Fan (north Snowdonia), and so on across northern England (e.g., the Whin Sill beneath Hadrian’s Wall), Northern Ireland (e.g., Fair Head, on the Antrim coast) and widely in western Scotland. To ‘inform’ the BBC-viewing public that quarrying by humans broke out the pillars ready for transport at the very summit of a perfectly natural outcrop, is ridiculous (Fig. 19).
Figure 19. The 2021 BBC Lost Circle Revealed commentary in part was delivered at the site of the yellow arrow above the outcrop at Carn Goedog. There are hundreds of perfectly natural outcrops of jointed dolerite like this dotted throughout upland Britain, some quite nearby too, and of course also abroad. Surely many thousands of years of freeze-thaw mechanical break-up cannot be overlooked? From this nonsense, ignoring simple natural processes and geology and by implication suggesting ancient quarry sites all over the country, there apparently can have been no scientific-editorial oversight and advice for the archaeologist(s). Gross mis-pleading like this, on a television channel that prides itself in distinguishing the truth, surely undermines simple enquiry and learning, undermines the credibility of actual experts, and, given the n-fold repetition of the BBC programme, is effectively misleading a host of people. Thankfully numerous sensible people have taken issue with the programme and the ‘quarries’ have been debunked and abandoned amongst aficionados, although the BBC continues the promotion. Also thankfully, most of the ground excavation has been restored to nature.
The Craig Rhosyfelin outcrop also featured as a bluestone megalith quarry, in this case of rhyolite (foliated rhyolitic tuff), despite there being no known megaliths from there, just a buried stump and a few lumps and many chips recovered at Stonehenge. The provenancing is probably robust (disputed by some), for its geochemistry and rock texture, but again, despite extensive and time-consuming excavations, no quarrying could be proved. Despite objections made regarding the ‘quarries’ (e.g., John et al. 2015), there was no retraction while the focus shifted to removal and transport to Stonehenge of stones from pre-existing monuments (Parker Pearson et al. 2021). In the infamous BBC ‘Stonehenge: The Lost Circle Revealed’ (2021) a long search eventually ‘found’ the site of a dismantled circle at Waun Mawn that was claimed to have sourced Stonehenge stones. This was ‘proved’ for viewers by the imprint of a removed stone at Waun Mawn that fitted a stone at Stonehenge “like a key in a lock”, with computer-graphic confirmation. The stone that was supposed to fit is Stone 62 (see Figs 1, 2 and 4). The interview based on this ‘clinching discovery’ invoked a revolutionary new view of collaboration between ancient societies. The trouble is, however, that robust data show that the Waun Mawn stones came from a local source (Bevins et al. 2022) unlike anything at Stonehenge, while Stone 62 came from the eastern Preseli hills not far from other known natural sources. So the whole programme, with its intense and dramatic revelations of quarrying and removal from a former stone circle, proved to be spectacularly wrong. One might say that hindsight is a wonderful thing, but contemporary expert advice was always available and ignored, and the media show on what is known to be a topic of wide interest was an information disaster.
The nonsense carried on. After the Altar Stone saga in 2024, in 2025, just this last year, there came another remarkable gem of mis-information. British Geological Survey Press (20/8/2025): “Scientists uncover secrets of Stonehenge’s mysterious cattle” and next day BBC News: “New analysis of a Neolithic cow’s tooth found at Stonehenge shows the animal likely came from Wales, reinforcing theories that cattle helped transport the enormous stones”. And “You can tell that the animal had been grazing on Palaeozoic rock, typical of those found in Wales, particularly in and around where bluestones are found… We’re beginning to see so many connections between Wales and Stonehenge. Not only is it the closest rock, but there are other links as well.” The italics above are mine; nothing supports these contentions, which are false and are sheer hype.
The tooth of the cow that had been buried at Stonehenge did yield geochemical evidence of part of its life and whereabouts, primarily in analyses of element-isotopes of carbon, strontium and lead (Evans et al. 2025). We learned from carbon and strontium (C and Sr) that the 6 months or so of the cow’s existence that was analysed involved a migration and shift of forage while the lead (Pb) was characteristic of ancient rocks (Lower Palaeozoic) widely exposed across Wales. The rider “particularly … where bluestones are found” cannot be justified*. And the Upper Carboniferous Sandstones of the South Wales and Forest of Dean Coalfields have yielded fabulous monoliths that are impressively incorporated in major structures of the Industrial Revolution, and they are significantly closer to Stonehenge. *(A similar claim regarding strontium isotope analyses of cremated human remains from Stonehenge (Snoeck et al. 2018) places analyses that could originate widely across Wales as “supporting links with west Wales”…”the source of the bluestones…”, only because human involvement in stone transport from Pembrokeshire was accepted.)
The claim that the cow-tooth findings add to confirmation of the “theory” that cattle were involved in transport of the megaliths is udderly ridiculous. Apart from confusing theory and hypothesis in this media hype, the unjustified claims were disturbing as they came from the Press Office of the British Geological Survey. No sane geologist would support them. Actually, the original scientific report of the cow-tooth findings, in contrast to the hype, was quite reserved, acknowledging limitations to interpretations that should be borne in mind. What is it then, really, that causes decent science to be so compromised in the media? There certainly is cavalier ignorance on the part of media producers, whose driver seems to be promotion of viewing figures or sales… Honesty about the issues might have been more interesting and more stimulating of genuine interest… Some of the media nonsense is arrogant and insulting.
References
Allen MJ, Chan B, Cleal R, French C & 10 others (2016) Stonehenge’s Avenue and ‘Bluestonehenge’. Antiquity 90, 352, 991-1008.
Bevins RE, Donnelly R (1992) The Storrie Erratic Collection: a reappraisal of the status of the Pencoed ‘Older Drift’ and its significance for the Pleistocene of South Wales. Proceedings of the Geologists’ Association 103, 129-142.
Bevins R, Ixer R (2018) Retracing the footsteps of H.H. Thomas: a review of his Stonehenge bluestone provenancing study. Antiquity 92 363, 788-802.
Bevins R, Ixer R, Pearce N, Scourse J, Daw T (2023) Lithological description and provenancing of a collection of bluestones from excavations at Stonehenge by William Hawley in 1924 with implications for the human versus ice transport debate of the monument’s bluestone megaliths. Geoarchaeology 38, 677-829.
Bevins RE, Ixer RA, Webb PC, Watson JS (2012) Provenancing the rhyolitic and dacitic components of the Stonehenge landscape bluestone lithology: new petrographical and geochemical evidence. Journal of Archaeological Science 39, 1005-1019.
Bevins RE, Pearce NJG, Hillier S, Pirrie D, Ixer RA & 4 others (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.
Bevins RE, Pearce NJG, Ixer RA, Scourse J, Daw T, Parker Pearson M & 5 others (2025) The enigmatic ‘Newall boulder’ excavated at Stonehenge in 1924: New data and correcting the record. Journal of Archaeological Science Reports 66, 105303.
Bevins RE, Pearce NJG, Pearson MP, Ixer RA (2022) Identification of the source of dolerites used at the Waun Mawn stone circle in the Mynydd Preseli, west Wales and implications for the proposed link with Stonehenge. Journal of Archaeological Science: Reports 45, 103556.
Bevins RE, Pirrie D, Ixer RA, O’Brien H, Parker Pearson M, Power MR, Shail RK (2020) Constraining the provenance of the Stonehenge ‘Altar Stone’: evidence from automated mineralogy and U-Pb zircon age dating. Journal of Archaeological Science 120, 105188.
Bögli A (1964) Mischungskorrosion: Ein Beitrag zum Verkarstungsproblem. Erdkunde 18, 83-92.
Bögli A (1980) Karst Hydrology and Physical Speleology, 284pp, Springer-Verlag: Berlin.
British Geological Survey (2002) Worms Head. England and Wales Sheet 246. Solid and Drift Geology. 1:50 000 Provisional Series. (Keyworth, Nottingham: British Geological Survey).
Brown DJ, Bell BR (2013) The emplacement of a large, chemically zoned, rheomorphic, lava-like ignimbrite: the Sgurr of Eigg Pitchstone, NW Scotland. Journal of the Geological Society 170, 753-767.
Burl A (1999) Essay extract from Glaciers and bluestones of Wales. Yale. British Archaeology 45, June.
Cawood PA, Nemcin AA, Smith M, Loewy S (2003) Source of the Dalradian Supergroup constrained by U-Pb dating of detrital zircons and implications for the East Laurentian margin. Journal of the Geological Society 160, 231-246.
Cawood PA, Hawkesworth CJ, Dhuime B (2012) Detrital zircon record and tectonic setting. Geology 40, 875-878.
Clarke AJI, Kirkland C (2026) Detrital zircon-apatite fingerprinting challenges glacial transport of Stonehenge’s megaliths. Nature Communications Earth & Environment doi: 10.1038/s43247-025-03105-3.
Clarke AJI, Kirkland CL, Bevins RE, Pearce NJG & 2 others (2024) A Scottish provenance for the Altar Stone of Stonehenge. Nature 632, 570-575.
Clarke AJI, Veness RLJ, Kirkland CL, Clark CD & 5 others (2026) From Highlands to Henge: refining the provenance and transport pathways of Stone Henge’s Altar Stone. Journal of Quaternary Science 1-8, https://doi.org/10.1002/jqs.70080.
Colbourne GJ, Gilbertson DD, Hawkins AB (1974). Temporary drift exposures on the Failand ridge. Proceedings of the Bristol Naturalists’ Society, 33, 91-97.
Collings-Wells P (2025) Ogof Ffynnon Ddu – a photographic exploration of the deepest cave in the UK (with 34+ contributors), WhiteSpringBooks and Gomer Press, 170 pp.
Cunnington BH (1924) The “Blue Stone” from Boles Barrow. The Wiltshire Archaeological & Natural History Magazine CXL, Vol. XLII, 431-437.
Daw T (2025) The origins of Stonehenge’s sarsen stones: A comprehensive review of provenancing studies. doi:10.13140/RG.2.2.28682.1706
Evans J, Madgwick R, Pashley V & 4 others (2025) Sequential multi-isotope sampling through a Bos taurus tooth from Stonehenge, to assess comparative sources and incorporation times of strontium and lead. Journal of Archaeological Science 180, 106269.
Field D, Anderson-Whymark H, Linford N & 15 others (2015) Analytical Surveys of Stonehenge and its Environs, 2009-2013: Part 2 – the Stones. Proceedings of the Prehistoric Society 81, 125-148.
Friend CRL, Strachan RA, Kinny PD, Watt GR (2003) Provenance of the Moine Supergroup of NW Scotland: Evidence from geochronology of detrital and inherited zircons from (mete)sedimentary rocks, granites and migmatites. Journal of the Geological Society 160, 247-257.
Gibbard PL, Lewin J (2003) The history of major rivers of southern Britain during the Tertiary. Journal of the Geological Society 160, 829-845.
Gibbard PL, Lewin J (2016) Filling the North Sea Basin: Cenozoic sediment sources and river styles. Geologica Belgica 19, 201-217.
Gilbertson DD, Hawkins AB (1978) The col-gully and glacial deposits at Court Hill, Clevedon, near Bristol, England. Journal of Glaciology 20, 173-188.
Harding P, Nash DJ, Ciborowski TJR & 2 others (2024) Earliest Movement of Sarsen Into the Stonehenge Landscape: New Insights from Geochemical and Visibility Analysis of the Cuckoo Stone and Tor Stone. Proceedings of the Prehistoric Society 90, 229-251.
Harvey JC, Morgan R, Webley DP (1967) Tooth Cave, Ilston, Gower. An Early Bronze Age occupation. Archaeology and Art, 22, 277-283. (author: i.e., Ilston parish not cwm)
Hawkins AB, Kellaway GA (1971) Field Meeting at Bristol and Bath. Proceedings of the Geologists Association 82, 267-291.
Ixer RA, Bevins RE (2010) The petrography, affinity and provenance of lithics from the Cursus Field, Stonehenge. Wiltshire Archaeological & Natural History Magazine 103, 1-15.
Ixer RA, Bevins RE (2016) Volcanic Group A debitage: its description and distribution within the Stonehenge landscape. Wiltshire Archaeological & Natural History Magazine 109, 1-14.
Ixer R, Bevins R (2017) The bluestones of Stonehenge. Geology Today 33, 180-184.
Ixer R, Bevins R, Pearce N, Pirrie D & 4 others (2025) Exotic granodiorite lithics from Structure 5 at West Kennet, Avebury World Heritage Site, Wiltshire, UK. Wiltshire Archaeological & Natural History Magazine 118, 1-18.
Ixer R, Bevins R, Pirrie D (2020) Provenancing the stones. Mapping the Stonehenge bluestones with mineralogy. Current Archaeology 366, 34-41.
Ixer RA, Turner P, Molyneux S, Bevins RE (2017) The petrography, geological age and distribution of the Lower Palaeozoic Sandstone debitage from the Stonehenge Landscape. Wiltshire Archaeological & Natural History Magazine 110, 1-16.
Jehu TJ (1904) The glacial deposits of Northern Pembrokeshire. Transactions of the Royal Society of Edinburgh 41, 53-87.
Jenkins DG, Beckinsale RD, Bowen DQ, Evans JA & 3 others (1985) The origin of granite erratics in the Pleistocene Patella beach, Gower, South Wales. Geological Magazine 122, 297-302.
Jenkins DG, Boulter MC, Ramsay ATS (1995) The Flimston Clay, Pembrokeshire, Wales: a probably late Oligocene lacustrine deposit. Journal of Micropalaeontology 14, p. 66.
John BS (2024) A bluestone boulder at Stonehenge: implications for the glacial transport theory. E&G Quaternary Science Journal 73, 117–134.
John B (2025) Carn Goedog on Mynydd Preseli was not the site of a Bluestone Megalith Quarry. Archaeology in Wales 63, 5-18.
John B, Elis-Gruffydd D, Downes J (2015) Observations on the supposed “Neolithic bluestone quarry” at Craig Rhosyfelin, Pembrokeshire. Archaeology in Wales 54, 139-148.
Johnson TE, Kirkland CL, Reddy SM, Evans NJ, McDonald BJ (2016) The source of Dalradian detritus in the Buchan Block, NE Scotland: Application of new tools to detrital datasets. Journal of the Geological Society 173, 773-782.
Kellaway GA (1971) Glaciation and the stones of Stonehenge. Nature 233, 30-35.
Kirkland CL, Strachan RA, Prave AR (2008) Detrital zircon signature of the Moine Supergroup, Scotland: Contrasts and comparisons with other Neoproterozoic successions within the circum-North Atlantic region. Precambrian Research 163, 332-350.
Kokelaar BP (2024) Gower’s inland flooding: Legacy of the last glaciation of Wales. Gower Journal 75, 6-11.
Kokelaar BP, Bevins RE, Roach RA (1985) Submarine silicic volcanism and associated sedimentary and tectonic processes, Ramsey Island, SW Wales. Journal of the Geological Society, London 142, 591-613.
Kokelaar BP, Howells MF, Bevins RE, Roach RA (1984) Volcanic and associated sedimentary and tectonic processes in the Ordovician marginal basin of Wales: a field guide. In Kokelaar BP, Howells MF (eds) Marginal Basin Geology. Spec. Pub. Geol. Soc. London, 16, 291-322.
Lee JR, Robertson S (2025) Refining the known extent of major onshore Quaternary glaciation in the UK – Types of evidence, nomenclature and uncertainty. Proceedings of the Geologists’ Association 136, 101087.
Leivers M (2021) The Army Basing Programme, Stonehenge and the emergence of the Sacred Landscape of Wessex. Internet Archaeology 56. https://doi.org/10.11141/ia.56.2
Mason AJ (2015) The Palaeoproterozoic anatomy of the Lewisian Complex, NW Scotland: Evidence for Two ‘Laxfordian’ Tectonothermal Cycles. Journal of the Geological Society 173, 153-169.
Mason AJ, Parrish RR, Brewer TS (2004) U-Pb geochronology of Lewisian orthogneisses in the Outer Hebrides, Scotland: implications for the tectonic setting and correlation of the South Harris Complex. Journal of the Geological Society of London 161, 45–54.
McAteer CA, Daly JS, Flowerdew MJ, Connelly JN, Housh TB, Whitehouse M (2010) Detrital zircon, detrital titanite and igneous clast U-Pb geochronology and basement-cover relationships of the Colonsay Group, SW Scotland: Laurentian provenance and correlation with the Neoproterozoic Dalradian Supergroup. Precambrian Research 181, 21-42.
McCarroll D, Stone JO, Ballantyne CK, Scourse JD, & 3 others (2010) Exposure-age constraints on the extent, timing and rate of retreat of the last Irish Sea ice stream. Quaternary Science Reviews 29, 1844-1852.
Neilson JC (2008) From slab breakoff to triggered eruptions: Tectonic controls of Caledonian post-orogenic magmatism. PhD thesis, University of Liverpool, 518 pp.
Neilson JC, Kokelaar BP, Crowley QG (2009) Timing, relations and cause of plutonic and volcanic activity of the Siluro-Devonian post-collision magmatic episode in the Grampian Terrane, Scotland. Journal of the Geological Society, London 166, 545-561.
Newall RS (1959) Stonehenge, Wiltshire. HMSO London, 24pp.
Oliver GJH, Wilde S, Wan Y (2008) Geochronology and geodynamics of Scottish granitoids from the late Neoproterozoic break-up of Rhodinia to Palaeozoic collision. Journal of the Geological Society, London 165, 661-674.
Parker Pearson M (2023) Stonehenge. Journal of Urban Archaeology 7, 147-168.
Parker Pearson M, Bevins R, Ixer R & 3 others (2020) Long-distance landscapes: from quarries to monument at Stonehenge. Megaliths and geology. Boaventura, Mataloto & Pereira (eds), pp 151-169.
Parker Pearson M, Pollard J, Richards C, Welham K (2017) The origins of Stonehenge: on the track of the bluestones. Archaeology International 20, 52-57.
Parker Pearson M, Pollard J, Richards C, Welham K & 6 others (2021) The original Stonehenge? A dismantled stone circle in the Preseli Hills of West Wales. Antiquity 95, 85-103.
Scourse JD (1997) Transport of the Stonehenge Bluestones: Testing the Glacial Hypothesis. Proceedings of the British Academy 92, 271-314.
Snoeck C, Pouncett J, Claeys P, Golderis S & 6 others (2018) Strontium isotope analysis on cremated human remains from Stonehenge supports links with west Wales. Nature Scientific Reports (2018) 8:10790.
Stevens T, Baykal Y (2021) Detrital zircon U-Pb ages and source of the late Palaeocene Thanet Formation, Kent, SE England. Proceedings of the Geologists’ Association 132, 240-248.
Strachan RA, Prave AR, Kirkland CL, Storey CD (2013) Detrital zircon geochronology of the Dalradian Supergroup, Shetland Islands, Scotland: Implications for regional correlations and Neoproterozoic-Palaeozoic basin development. Journal of the Geological Society 170, 905-916.
Strahan A, Cantrill TC (1904) Geology of the South Wales Coalfield. Part VI. The country around Bridgend. Mem. Geol. Surv. Great Britain.
Strahan A, Cantrill TC, Dixon EEL, Thomas HH, Jones OT (1914) The geology of the South Wales Coalfield, Part XI. The country around Haverfordwest. London: HMSO.
Sutcliffe AJ (1995) Insularity of the British Isles 250 000-30 000 years ago: the mammalian, including human, evidence. Geological Society, London, Special Publications, 96, 127-140.
Thomas HH (1923) The source of the stones of Stonehenge. Antiquaries Journal 3, 239-260.
Thorpe RS, Williams-Thorpe O (1991) The myth of long-distance megalith transport. Antiquity 65, 64-73.
Thorpe RS, Williams-Thorpe O, Jenkins DG, Watson JS (1991) The geological sources and transport of the bluestones of Stonehenge, Wiltshire, UK. Proceedings of the Prehistoric Society 57, 103-157.
Waldron JWF, DuFrane SA, Schofield DI, Barr SM (2025) Is Britain divided by an Acadian suture? Geology 53, 847-852.
Waldron JWF, Floyd JD, Simonetti A, Heaman LM (2008) Ancient detrital zircon in the closing Iapetus ocean, Southern Uplands terrane, Scotland. Geology 36, 527-530.
Waldron JWF, Schofield DI, Pearson G, Floyd JD, Crowley QG, Simonetti A, Dokken RJ, Pothier HD (2014) Ganderia-Laurentia collision in the Caledonides of Great Britain and Ireland. Journal of the Geological Society 171, 555-569.
Williams-Thorpe O, Thorpe RS (1992) Geochemistry, sources and transport of the Stonehenge Bluestones. Proceedings of the British Academy 77, 133-161.
