This is the working evidence file behind the glacial position taken on The Gift of the Ice. It lists what has actually been found west of Stonehenge, who reported it, where the primary record sits, and how much weight each item will bear. It also carries the case against, because a file that only collects the friendly evidence is a scrapbook, not a reference.
Two separations run through everything here, and they matter more than any single find.
First: region versus Plain. There is secure evidence that ice reached north Somerset and the Bristol Channel coast. There is no secure evidence that ice ever reached Salisbury Plain. Those are different claims, roughly 50 miles apart, and the whole bluestone argument lives in the gap between them.
Second: report versus interpretation. For every found stone below, two questions apply. Was foreign material genuinely found and recorded at that spot? And how did it get there? The confidence scores on this page answer only the first question. The second is contested in every single case.
The frame: supply, not delivery
Everything in this file is old. The Anglian glaciation, the largest ice advance to touch southern England, ran from roughly 478,000 ya (BC 476,000) to 424,000 ya (BC 422,000), with ice at its fullest around 450,000 ya (BC 448,000). The Somerset lowland deposits are older still on the usual reading. The first stone did not stand at Stonehenge until about 5,000 ya (BC 3,000), and the enclosure ditch was dug about 5,100 ya (BC 3,100).
So nothing here shows ice delivering stones to a monument. What it can show is erratic supply: ice carrying far-travelled rock into the region, hundreds of thousands of years before anyone was around to care. A glacial case for the bluestones needs this evidence. This evidence alone does not make that case.
Map 1: Glacial deposits west of Stonehenge. What the ice left in the ground, between the Bristol Channel and Bath, plotted against Stonehenge. Two clusters and a scatter. Yellow triangles mark the Kenn lowlands and Failand ridge: the secure, GCR-designated glacial deposits. Red circles mark the Bath plateau drifts, whose origin is contested; Donovan (1995) keeps glacial as one option among several. White squares mark individual erratic-bearing gravels and till-like outliers, strung along the Avon valley and scattered across Sedgemoor: not a cluster, just strays. Everything shown is Anglian-age or older, roughly 450,000 ya (BC 448,000) and beyond: erratic supply to the region, not delivery to Salisbury Plain. Coastline and relief from open elevation data (AWS Terrain Tiles / SRTM, public domain). Sites after Donovan 1995, Campbell et al. 1998 and Brian John.
The clusters, and the scatter
Cluster 1: Kenn lowlands and Failand ridge (secure)
Centre roughly 51.42, -2.83. The sites, all in North Somerset behind the Clevedon and Portishead coast:
- Court Hill — a buried glacial gully cut through the Failand ridge, near Tickenham, just east of Clevedon. The original erratic finds here go back to Trimmer (1853); the modern glacial reading to Hawkins and Kellaway (1971).
- Nightingale Valley — till and glaciofluvial deposits on the Failand ridge, near Clevedon.
- Kennpier — till and erratic-rich outwash gravels of the Kenn Formation, in the low Kenn lowlands south-east of Clevedon.
- Holly Lane, Clevedon — the old quarry section in Clevedon where the modern glacial story began.
- Weston-in-Gordano — till-like material stratified within interglacial marine deposits, in the Gordano valley between Clevedon and Portishead.
- Yew Tree Farm — interglacial deposits, read as Marine Isotope Stage 15, overlying the Kenn Formation.
- Kenn Church — Stage 7 marine molluscs, at Kenn village itself.
This is the strongest glacial evidence in the region, and its glacial character is not seriously disputed. These are Geological Conservation Review sites, described at length in Campbell et al. (1998).
Dating is the hedged part. The interglacial deposits sitting above the Kenn Formation are usually read as Marine Isotope Stage 15, which would put the glaciation beneath them at around 650,000 ya (BC 648,000) or older. At minimum it is Anglian or earlier: no younger than about 450,000 ya (BC 448,000). Either way these deposits sit more than 25 km inland from the present coast, which is the point Brian John presses: ice demonstrably got well into Somerset.
Cluster 2: Bath plateau drifts (contested)
Centre roughly 51.38, -2.34. High-level spreads on the Great Oolite plateau east and south-east of Bath, sitting at 150 to 200 m OD:
- Bathampton Down — drift on the plateau immediately east of Bath, above the Avon.
- Claverton Down — plateau patches just south of Bathampton Down, above the University of Bath.
- Farleigh Down — drift east of Bathampton, above Monkton Farleigh and the Bathford to Box road.
- Kingsdown — the easternmost locality, near Box; the source of the single far-travelled sarsen (Prestwich 1890, on Winwood’s authority).
- Midford to Freshford plateau — the western extension of the drifts, on the high ground south of Bath.
- Limpley Stoke — at about 150 m, in the Avon valley south-east of Bath.
These are spreads of flint, quartz, chert, sandstone, shale and coal in clay and loam, lodged in karstic fissures and solution cavities on hilltops where no river could recently have put them. Donovan (1995) is the modern account. He treats a glacial origin as one possibility alongside reworked early Tertiary gravels and ancient fluvial deposits, favours a composite origin, and dates emplacement to the early or middle Pleistocene at the latest. The far-travelled clasts are the interesting part: the single Kingsdown sarsen, and Palaeozoic sandstone, coal and shale (Varney 1921). Kellaway read these drifts as glacial; Donovan does not commit, and even doubted whether the shale was truly far-travelled.
Honest weight: this cluster is suggestive, not secure, and must never be written up as established till.
The scatter: Avon valley and Sedgemoor (not a cluster)
These are individual sites, not a cluster: places where far-travelled material turns up in secondary contexts, mostly river gravels, plus a couple of till-like outliers. They fall into three loose areas.
Avon valley:
- Newton St Loe — erratic-bearing Pleistocene gravels on the Bristol Avon, south-west of Bath, roughly 51.38, -2.43. The only surviving fossiliferous exposure on the river, known for its mammoth and horse remains.
- Hampton Rocks Cutting — recycled glacial erratics in fluvial gravels, in the Avon valley at Bathampton, east of Bath.
- Stidham Farm — erratic-bearing gravels near Keynsham, on the Avon between Bath and Bristol.
- Ham Green — an erratic site near Pill, beside the Avon gorge west of Bristol.
Sedgemoor:
- Greylake No. 2 Quarry — a basal diamicton on King’s Sedgemoor, roughly 51.10, -2.88, that may mark the maximum eastward spread of glacial deposits into the Somerset lowlands.
- Langport railway cutting — erratic-free, about 6 km south of Greylake, read as sitting just beyond the glacial limit.
Mendip margin:
- Bleadon Hill — gravels queried as glaciofluvial even in the GCR account, on the western tip of the Mendips near the River Axe, inland of Weston-super-Mare, roughly 51.32, -2.94.
Sources: Campbell et al. (1998), Donovan (1995), John (2011).
Map 2: Reported erratic stones west of Stonehenge. Individual finds and reports of foreign stone, plotted where they were recorded, each badged with a confidence score out of 10. The score rates one thing only: whether foreign material was genuinely found and recorded at that spot. It says nothing about how the stone travelled. Report and interpretation are different questions, and both are contested. The trilithon glyph marks the foreign debitage assemblage at Stonehenge itself, scored 7. Unlocated reports (wall-built and field-cleared boulders, scored 1 to 2) are not plotted. Reports: Cunnington 1924 (Boles Barrow); Judd 1902, Thomas 1923 and Ixer and Bevins, various (debitage); Trimmer 1853 (Court Hill); Kellaway 1971; Donovan 1995. All contested.
Reported erratics, scored for confidence
Reported erratics are a different evidential class from the deposits above. These are individual found stones, often recorded once, by one observer, in the 19th or early 20th century. Some have since vanished into walls and field clearance. Confidence rests on who reported the find and whether the stone, or a sample of it, still exists.
The reporting is dominated by a few advocates, Kellaway above all and John following him, whose glacial readings others dispute. So, again: the score rates the report, not the transport.
| Report | Confidence | Primary record | Principal pushback |
|---|---|---|---|
| Boles Barrow bluestone | 8/10 | Cunnington footnote on an 1801 letter; published Cunnington (1924) | Museum stone may not be the Boles stone (Green 1997; Pitts 2000) |
| Stonehenge foreign debitage | 7/10 | Judd (1902, 1903); Thomas (1923); Ixer & Turner (2006); Ixer et al. (2017) | Presence agreed; transport is the fight |
| Court Hill, Kenn | 6/10 | Trimmer (1853), via Campbell et al. (1998) | Trimmer’s diluvial explanation long dead; the finds stand |
| Bath plateau-drift clasts | 5/10 | Donovan (1995), reporting Prestwich (1890) and Varney (1921) | Single finds, single observers; Kellaway doubted the shale |
| Saltford in the Avon valley | 4/10 | Kellaway (1971); gravels in Campbell et al. (1998) | Gravels real; specific erratic identifications thin |
| Bleadon in the Mendip margin | 4/10 | Campbell et al. (1998), queried entry | Recorded but enigmatic |
| Kellaway’s Salisbury Plain erratics | 3/10 | Kellaway (1971) | Poorly relocated; never independently confirmed |
| Flat Holm & Steep Holm | 3/10 | Burl (1999); John (2018) | No published provenanced sample traced |
| Unlocated wall-built / field-cleared boulders | 1-2/10 | Passing mentions in Kellaway and John | Cannot be checked; some noted as gone |
Where each plotted report was recorded:
- Boles Barrow bluestone — an Early Neolithic long barrow on Salisbury Plain, near Heytesbury, Wiltshire, about 12 miles (19 km) west of Stonehenge.
- Stonehenge foreign debitage — in and around the Stonehenge Layer, at the monument itself.
- Court Hill, Kenn — the buried glacial gully through the Failand ridge, near Tickenham, east of Clevedon, North Somerset.
- Bath plateau-drift clasts — the far-travelled clasts of the Bath drifts; the headline find, the lone sarsen, is from Kingsdown near Box, east of Bath.
- Saltford, in the Avon valley — erratics in the Avon gravels around Saltford, on the river between Bath and Bristol.
- Bleadon, on the Mendip margin — on Bleadon Hill, the south-western tip of the Mendips, near the River Axe inland of Weston-super-Mare.
- Kellaway’s Salisbury Plain erratics — scattered, poorly relocated finds on the chalk of Salisbury Plain north-west of Stonehenge; never independently confirmed.
- Flat Holm and Steep Holm — two islands in the Bristol Channel off Weston-super-Mare (Flat Holm the northern, in Welsh water; Steep Holm the southern, English), with erratics also reported from the channel floor.
Boles Barrow, in full
Excavated in 1801 by William Cunnington, the barrow itself is an Early Neolithic long barrow about 12 miles west of Stonehenge, built in the centuries around 5,500 ya (BC 3,500). Cunnington’s original letter to H.P. Wyndham describes the large stones in the mound as the same species as the big stones at Stonehenge, meaning sarsen. The bluestone claim comes from an undated footnote he added later to his own copy of that letter: that among the stones he discovered “the Blue hard Stone, ye same to some of the upright Stones in ye inner Circle at Stonehenge”. A note in Robert Newall’s hand, reported by Pitts (2000), adds that Cunnington wrote of bringing ten stones to his house at Heytesbury and later finding some of them blue and grey. The find was published by B.H. Cunnington (1924).
A spotted dolerite boulder later stood in the garden of Heytesbury House, known there as the Stonehenge Stone, and was given to Salisbury Museum by Siegfried Sassoon in 1934. It weighs about 611 kg (1,338 lb) and has been identified as spotted dolerite matching the Stonehenge orthostat groupings (Group 3; Bevins, Ixer and Pearce 2014). Thorpe and colleagues sampled it in the late 1980s (Thorpe et al. 1991).
Why it matters: if a bluestone truly lay sealed in a long barrow centuries before the monument’s bluestone settings, human haulage for that stone is very hard to sustain, and the erratic reading gains its best single exhibit. Burl (1999) adds the supporting argument that the stone cannot have been robbed from Stonehenge, because Petrie’s punctilious 1877 survey found no stone missing since Wood’s survey of 1740.
Why it is contested, and this page carries it in full: Cunnington’s recorded Boles stones weighed 28 to 200 lb and were first described as sarsens; the museum boulder is roughly six times heavier than the largest of them. Green (1997) and Scourse (1997) argue there is no unequivocal evidence the museum stone was ever in Cunnington’s possession, let alone in the barrow. Pitts (2000, pages 198-204) gives the forensic account of the whole chain and is the single best starting point. Critics also note that even a genuine bluestone found high in the mound need not pre-date the monument’s bluestone settings. Score 8/10 that a foreign stone was genuinely reported. The meaning of the report is wide open.
The Stonehenge foreign debitage
The foreign fragment assemblage at Stonehenge, in and around the so-called Stonehenge Layer, has been reported continuously since Judd’s petrological note on Gowland’s 1901 excavation (Judd 1902, 1903). Thomas (1923) provenanced the main bluestone types to Preseli. The modern classifications are Ixer and Turner (2006) and Ixer, Turner, Molyneux and Bevins (2017). Counts differ with how finely the lithologies are split: the published debitage classifications distinguish twenty-plus rock types, and John (2018) counts 46 or more. Either figure is awkward for a tidy quarry story and comfortable for an erratic assemblage; the quarrying camp answers that several outcrops were exploited. The presence of the foreign material is agreed by everyone. Its journey is the entire argument.
Flat Holm, Steep Holm and the erratic train
Erratics of Welsh origin are said to lie on Flat Holm and Steep Holm and on the channel floor, which would mark a carry-line running east from Pembrokeshire (Burl 1999; John 2018). This file has not traced a published, provenanced, sampled Preseli erratic from either island. Until one surfaces, the train is an argued pattern, not a chain of confirmed finds. Write it that way.
Other evidence strands
The shape of the stones themselves
Many of the bluestones are rounded, abraded, slabby and boulder-like rather than fresh angular pillars, and several carry smoothed or faceted surfaces. John (2018, 2024) reads these profiles as ice-worn: erratics, not quarry products. Thorpe and colleagues made a similar argument a generation earlier (Thorpe et al. 1991). The quarrying reading treats the same shapes as natural jointing exploited by extraction, and rests on provenance geochemistry tying the spotted dolerites to particular Preseli outcrops (Bevins, Ixer and Pearce 2014), plus excavation claims at two Preseli outcrops that John (2024) reviews and disputes.
The defensible claim for this site is narrow and visual: these profiles are consistent with erratic boulders, and not obviously consistent with freshly quarried monoliths. That is all. Rounding alone proves nothing; frost-shattering, in-situ weathering and rivers all round stone too. This is exactly what the naked renders are for. Show the surfaces, cite both readings, and let people look.
Landforms beyond the deposits
The Fremington till in north Devon shows ice on the Bristol Channel coast west of Exmoor (Campbell et al. 1998). The late Devensian ice margin crossed the northern Isles of Scilly around 24,000 ya (BC 22,000), proving Irish Sea ice ran far down the western approaches in a much later glaciation (Scourse 1991). And John (2011) reads some Mendip dry valleys and gorges as glacial meltwater features; the conventional account keeps them periglacial and fluvial, and that remains the majority view. Flag the meltwater reading as a minority one whenever it is used.
Periglacial, not glacial, at the monument itself
One ice-age signal exists at Stonehenge and is beyond dispute: the natural frost-sorted stripes in the chalk beneath the Avenue, running on the solstitial axis (Cleal, Walker and Montague 1995). They formed under deep cold in the last (Devensian) glaciation, which ended about 11,700 ya (BC 9,700). They prove intense periglacial conditions on the Plain. They do not prove ice cover, and they carried no stones. Their value to this site is as the natural thread people later noticed, the same thread picked up by the Mesolithic posts about 10,000 ya (BC 8,000). Keep them out of the transport argument.
Maybe, just maybe: glacial till
Two threads from the A303 tunnel paperwork, neither of them looked at with a glacier in mind. First, the Arup Atkins air-photo team noticed hummocky ground on the west side of Stonehenge Bottom under low winter sun; Wessex read it as very old phosphate workings, and it has never been archaeologically sectioned. Second, Wessex Archaeology’s evaluation trench 511, beside the A303 west of Countess Roundabout and across the road from Blick Mead, has the words glacial till written against context 51132 in the context register.
That is a field description in a road-consent trench, not a Quaternary section. Nobody in Trench 511 was looking for ice; they were looking for postholes. The forty-litre environmental sample came from the context next door (51135, light green sand and degraded chalk), went through flotation for seeds and charcoal, and the coarse fraction above 4 mm, the only part that could say what rock it was, was sorted, weighed and discarded. Whether anything from 51132 itself survives on a shelf, only Wessex can say. The full working, including the correction I had to make to my own first reading of the table, is on the blog.
Report confidence: the phrase “glacial till” is in the record, 9/10. What it describes: unscored, because nobody has examined it. Brian John’s test for a degraded till after 450,000 years on chalk (cementation, flow structures, far-travelled clasts, fabric) has not been applied, and until someone applies it this is a question, not a finding. It sits on the page for the same reason Kellaway’s erratics do: an unchecked entry in the record is worth chasing and worth nothing else.
The case against, carried honestly
Green counted more than 50,000 pebbles from the terraces of the rivers draining Salisbury Plain and found not one glacially derived rock (Green 1997; restated in correspondence in British Archaeology, 1999). If ice had reached the Plain, its debris should be in those terraces. This is the single most damaging fact for the glacial-transport case, and every page on this site that touches the argument must be able to survive it. The usual erratic-camp reply is that a narrow lobe, or an older and largely stripped deposit, could leave little trace. That is possible. It is also unfalsifiable as stated.
Scourse (1997) adds the glaciological argument: modelled ice from the Bristol Channel has no plausible way of flowing up onto the Plain and dropping monoliths where they were later found. John (2018) contests the framing. The debate is real and unresolved.
Provenance geochemistry ties the spotted dolerites tightly to particular Preseli outcrops (Bevins, Ixer and Pearce 2014; Williams-Thorpe et al. 2006 for the wider dolerite work). The quarrying camp reads that precision as incompatible with the scatter an ice sheet would sample. John (2024) disputes how precise the provenancing really is.
On the Altar Stone: recent provenance work argues a far northern origin (Clarke et al. 2024). The secure part is a Laurentian source north of the Iapetus Suture. The claim of the Orcadian Basin specifically is overclaimed on present sampling: the two museum specimens are tied to the stone by portable XRF chemistry rather than direct sampling, and the Old Red Sandstone comparison samples were bought, not field-collected. That critique is carried in full on the Altar Stone pages and is only flagged here.
Where each strand sits
Secure. Ice reached north Somerset and the Bristol Channel margins, around 450,000 ya (BC 448,000) and earlier: Cluster 1, the Fremington till, and, for a much later glaciation, the Scilly limit.
Suggestive and renderable. The morphology of the bluestones. The Boles Barrow report. The foreign debitage assemblage.
Contested. The origin of the Bath plateau drifts. The erratic train and the Holm reports. The Mendip meltwater reading. Kellaway’s Salisbury Plain erratics.
Against. Green’s pebble counts. Scourse’s glaciology. The precision claimed by provenance geochemistry.
The honest bottom line, and the line every page should hold: the region was glaciated; whether ice ever crossed the 50 miles from the proven deposits to Salisbury Plain is unproven in both directions. This site shows the stones, scores the reports, cites both camps, and does not pretend the gap away.
→ Back to: The Gift of the Ice · Wider argument: Banks, Ditches and the Broken Ring
References
- Bevins, R.E., Ixer, R.A. and Pearce, N.G. 2014. Carn Goedog is the likely major source of Stonehenge doleritic bluestones: evidence based on compatible element geochemistry and principal components analysis. Journal of Archaeological Science 42, 179-193.
- Burl, A. 1999. Glaciers and the bluestones of Wales. British Archaeology 45.
- Campbell, S., Hunt, C.O., Scourse, J.D., Keen, D.H. and Stephens, N. 1998. Quaternary of South-West England. Geological Conservation Review Series 14. London: Chapman and Hall. Chapter 10 (the Somerset lowland sites) is online via the JNCC.
- Clarke, A.J.I. and colleagues. 2024. A Scottish provenance for the Altar Stone of Stonehenge. Nature 632. Treated on this site as overclaimed beyond “Laurentian, north of the Iapetus Suture”; see the Altar Stone pages.
- Cleal, R.M.J., Walker, K.E. and Montague, R. 1995. Stonehenge in its Landscape: Twentieth-Century Excavations. London: English Heritage.
- Cunnington, B.H. 1924. The “Blue Stone” from Boles Barrow. Wiltshire Archaeological and Natural History Magazine 42, 431-47.
- Donovan, D.T. 1995. High Level Drift Deposits East of Bath. Proceedings of the University of Bristol Spelaeological Society 20(2), 109-126. PDF via ubss.org.uk.
- Green, C.P. 1997. The provenance of rocks used in the construction of Stonehenge. In Cunliffe, B. and Renfrew, C. (eds), Science and Stonehenge. Proceedings of the British Academy 92. Oxford. The 50,000-pebble count restated in correspondence, British Archaeology, 1999.
- Hawkins, A.B. and Kellaway, G.A. 1971. Field meeting at Bristol and Bath with special reference to new evidence of glaciation. Proceedings of the Geologists’ Association 82(2), 267.
- Ixer, R.A. and Turner, P. 2006. A detailed re-examination of the petrography of the Altar Stone and other non-sarsen sandstones from Stonehenge as a guide to their provenance. Wiltshire Archaeological and Natural History Magazine 99, 1-9.
- Ixer, R.A., Turner, P., Molyneux, S. and Bevins, R.E. 2017. The petrography, geological age and distribution of the Lower Palaeozoic Sandstone debitage from the Stonehenge landscape. Wiltshire Archaeological and Natural History Magazine 110, 1-16.
- John, B.S. 2011. The Glaciation of Somerset. Stonehenge and the Ice Age (blog), October 2011. brian-mountainman.blogspot.com/2011/10/glaciation-of-somerset.html
- John, B.S. 2018. The Stonehenge Bluestones. Newport: Greencroft Books.
- John, B.S. 2024. The Stonehenge bluestones did not come from Waun Mawn in West Wales. The Holocene. DOI: 10.1177/09596836241236318.
- Judd, J.W. 1902. Petrological note in Gowland, W., Recent excavations at Stonehenge. Archaeologia 58 (2nd series 8), note at p. 106.
- Judd, J.W. 1903. Note on the nature and origin of the rock fragments found in the excavations made at Stonehenge by Mr Gowland in 1901. Wiltshire Archaeological and Natural History Magazine, 47-61.
- Kellaway, G.A. 1971. Glaciation and the stones of Stonehenge. Nature 233(5314), 30-35. DOI: 10.1038/233030a0. (Volume 233; some bibliographies mis-cite it as 232.)
- Kellaway, G.A. 2002. Glacial and tectonic factors in the emplacement of the bluestones of Salisbury Plain. Survey of Bath and District 17, 57-71.
- Pitts, M. 2000. Hengeworld. London: Century. Pages 198-204 for the Boles Barrow stone: the forensic account.
- Prestwich, J. 1890. Report of the Kingsdown sarsen, on H.H. Winwood’s authority. Cited in Donovan 1995.
- Scourse, J.D. 1991. Late Pleistocene stratigraphy and palaeobotany of the Isles of Scilly. Philosophical Transactions of the Royal Society of London B 334.
- Scourse, J.D. 1997. Transport of the Stonehenge bluestones: testing the glacial hypothesis. In Cunliffe, B. and Renfrew, C. (eds), Science and Stonehenge. Proceedings of the British Academy 92. Oxford.
- Thomas, H.H. 1923. The source of the stones of Stonehenge. The Antiquaries Journal 3, 239-260.
- Thorpe, R.S., Williams-Thorpe, O., Jenkins, D.G. and Watson, J.S. 1991. The geological sources and transport of the bluestones of Stonehenge, Wiltshire, UK. Proceedings of the Prehistoric Society 57(2), 103-157.
- Trimmer, J. 1853. The original erratic finds at Court Hill, Kenn. Cited in Campbell et al. 1998.
- Varney. 1921. Far-travelled clasts in the Bath plateau drifts. Cited in Donovan 1995.
- Williams-Thorpe, O., Jones, M.C., Potts, P.J. and Webb, P.C. 2006. Preseli dolerite bluestones: axe-heads, Stonehenge monoliths, and outcrop sources. Oxford Journal of Archaeology 25(1), 29-46.
Fifty miles. That is the gap. Kenn at one end, till and channels and outwash, all bagged and labelled. Stonehenge at the other. In between, not a docket, not a chit, not one pebble with a foreign accent. Green sieved fifty thousand of them. Nothing.
So this is what I have. A delivery, signed for at Clevedon. No delivery further on. And a lump of grey-green sandstone forty miles up the road with no paperwork at all. I don’t know how it got there. Neither does anybody else, whatever the man in the fleece says.