No Sarsen in the Roadstone: What the A344 Excavations Reveal About Stonehenge’s Missing Stones

It is sometimes suggested that many of Stonehenge’s “missing” stones — sarsens or bluestones that once stood but are no longer visible — were broken up and reused, including as road metal or hardcore during the turnpike era or later repairs. The A344, which ran immediately past the monument and crossed the Avenue, is occasionally invoked in such discussions.

Two detailed archaeological reports from the Stonehenge Environmental Improvements Project (SEIP) — the works that closed and removed the relevant section of the A344 — allow us to examine this claim directly against primary evidence. They reveal what materials were actually used to build and maintain the road right beside Stonehenge.

The Reports

Powell et al. 2019 Andrew B. Powell (with contributions by Phil Harding, Rob Ixer, Matt Leivers and others). “Along the road to Stonehenge: investigations of the Stonehenge Avenue and within the World Heritage Site.” Wiltshire Archaeological and Natural History Magazine vol. 112 (2019), pp. 197–216. This is the full post-excavation report on mitigation works in 2014–15, including hand-excavated slots through the Avenue ditches and the edge of the Heel Stone ditch within the former road line, plus watching briefs and recording.

Wessex Archaeology 2011 Stonehenge Environmental Improvements Project: A344 Works, Stonehenge, Wiltshire. Archaeological Watching Brief Report. Report ref: 76860.03 (March 2011). This covers the earlier watching brief during site investigation trial pits (February 2011) along the A344, in the old visitor car park, and at Airman’s Corner.

Both reports are available via standard archaeological channels (the 2019 paper in WANHM; the 2011 report via Wessex Archaeology / OASIS).

What the Road Was Actually Made Of

2011 trial pits (12 small hand-excavated pits): The consistent sequence was modern tarmac over layers of hogging and road make-up/levelling. Materials included stone aggregate, limestone fragments, flint, gravelly sand, and in places clinker or industrial waste. These overlay truncated natural chalk. No sarsen or bluestone was recorded in the road construction layers. The pits were shallow and small, but they sampled the road fabric directly.

2014–15 excavations (Powell et al. 2019): Where the road crossed the Avenue, the make-up consisted of layers of compacted flint and crushed chalk rubble. In a more complex 3 m × 10 m slot near the old car park entrance there were additional layers: reddish-brown sand, spreads of large stones, greenish-yellow sand and sandstone (containing one fragment of ceramic building material), flint gravel, a band of large limestone blocks (c. 0.3 m wide × 0.2 m thick), limestone hardcore, and compacted flints.

The Avenue’s internal banks had been levelled during road construction, but the ditches themselves survived in form and fill sequence broadly comparable to earlier excavations outside the road line (e.g., by the Vatchers and Pitts). Periglacial striations and natural features were still visible in plan and section where road material was fully removed, indicating that truncation was not total everywhere.

Crucially, no sarsen or bluestone fragments are reported from the road bedding, sub-base, or make-up layers in either report.

This pattern is not unique to the A344. Local flint, chalk, and limestone appear to have been the standard materials for the area's turnpike roads. Whether the A303 itself conceals anything different has never been specifically tested.

Where Stonehenge Stone Was Found

Bluestone (mostly rhyolite, Group C, with one possible Group E piece that could relate to orthostat SH48) and sarsen fragments were recovered — but from the tertiary fills of the Avenue ditches (contexts 10070 southern ditch; 10081 northern ditch), not from the road construction. These were accompanied by worked flint. The assemblage is interpreted in the context of known stone-working activity near the Avenue terminals (previously recorded by Pitts) or as material that entered the ditches over time. Average artefact weights and the absence of microdebitage in sieved residues suggest these are not in-situ working floor deposits within the ditches themselves. Some material may be later than the main prehistoric use of the monument.

Animal bone, post-medieval pottery, and clay-pipe fragments appeared only in the uppermost fills — consistent with later intrusion or deposition into already silted features.

Discussion and the Question

The A344 section past Stonehenge was metalled in the early 1760s as part of the Amesbury–Shrewton turnpike and subsequently repaired and resurfaced. The materials used were the standard local ones available in the chalk downland or brought in: flint, crushed chalk, sand, and limestone.

Despite the road running immediately adjacent to Stonehenge, crossing the Avenue, and close to the Heel Stone, the detailed examinations reveal no incorporation of broken sarsen or bluestone from the monument into the road fabric.

William Stukeley's survey of 1721 already shows the outer circle incomplete, so any robbing of sarsen predates the construction of the A344 turnpike in the early 1760s, but pre-turnpike roads and tracks were a parish and landowner affair, maintained from whatever came to hand locally, so the absence of sarsen from those earlier surfaces and trackways is even more telling than its absence from the more organised turnpike era.

If, as sometimes suggested, missing or fallen stones from Stonehenge were broken up for road metal or hardcore, especially for this road so proximate to the site, we would expect to see traces in the sections meticulously excavated and recorded. Yet none are reported. The stone fragments present are contextualised within the prehistoric (or later) ditch silts and are discussed in relation to monument-associated activity, not road building.

The absence of sarsen or bluestone in the documented road construction layers is a negative observation that may have some bearing on discussions of whether the monument was ever fully completed in stone.

This observation doesn’t rule out reuse of stone elsewhere in the landscape or for other purposes, nor does it address every possible missing stone. Sarsen is exceptionally hard and durable; breaking it into usable roadstone would have been far more laborious than using abundant flint or quarried limestone. The reports simply show that, for this specific and well-examined road section, that does not appear to have happened.

Why This Matters

The SEIP excavations gave us an unusually clear window into both the prehistoric Avenue and the post-medieval road that overlay it. The contrast is instructive: the road was built with what was locally abundant and practical. The megalithic stones that left traces did so in the expected archaeological contexts, ditch fills associated with the monument, not used or scattered through road layers. The two reports discussed here provide a detailed record from the section of the A344 immediately adjacent to the monument, in which no sarsen or bluestone was identified within the make-up layers.

The lack of sarsen or bluestone in the road make-up provides no support for the suggestion that stones from Stonehenge were broken up for local road construction. Additional work on road fabrics elsewhere in the area would help place this negative observation in context.

Other explanations for the absence of stones have occasionally been proposed. These include the possibility that stones were broken up and reused in other local roads or tracks (though no supporting evidence has yet been identified in published reports from the A303 corridor or elsewhere in the immediate area), incorporation into buildings or field walls (for which there is similarly little documented evidence in the locality), or conversion into artefacts such as querns. The hypothesis that certain stone positions were never filled has also been discussed, although observations of parchmarks in 2013 provided evidence consistent with the former presence of posts in some previously uncertain locations. Earlier timber phases are well attested at Stonehenge. Was it a wooden monument being replaced in stone piecemeal, a process that was never completed?

---

References

Fitzpatrick, R. 2011. Stonehenge Environmental Improvements Project, A344 Works, Stonehenge, Wiltshire: Archaeological Watching Brief Report. Unpublished grey literature report, Wessex Archaeology, Salisbury. WA Report ref. 76860.03. OASIS ID: wessexar1-98456.

Pitts, M.W. 1982. On the road to Stonehenge: report on the investigations beside the A344 in 1968, 1979 and 1980. Proceedings of the Prehistoric Society 48, 75–132.

Powell, A.B., with contributions by Harding, P., Higbee, L., Ixer, R., Leivers, M., López-Dóriga, I., Mepham, L. and Norcott, D. 2019. Along the road to Stonehenge: investigations of the Stonehenge Avenue and within the World Heritage Site. Wiltshire Archaeological and Natural History Magazine 112, 197–216.

Wessex Archaeology. Stonehenge: Archaeology on the A303 Improvement.

Visualising the Secondary Solstice Axis with a Clever Lunar Trick

Independent researcher Simon Banton has published a clear and accessible blog post explaining how he used the Moon as a proxy for the ancient Sun to test and visualise Stonehenge’s secondary solstice axis.

Photo of the lunar proxy of the Midwinter Solstitial Sunrise over the Altar Stone - photo - Simon Banton

The Problem

The Earth’s axial tilt (obliquity) was slightly greater (~24°) when Stonehenge’s sarsens were erected around 2500 BC. This means the Sun no longer rises and sets at exactly the same horizon positions the Neolithic builders saw. Direct observation of the ancient winter solstice sunrise is impossible today.

The Clever Lunar Solution

During a major lunar standstill (which we’re in now), the Moon reaches more extreme positions on the horizon than the modern Sun. Banton timed his observations for July 2024/2025 so the Moon rose very close to where the winter solstice Sun would have appeared 4,500 years ago.

He captured the event from a position aligned with the proposed sightline — looking through the notch in Stone 58 and the edge of Stone 53 toward Coneybury Hill. The results are striking: the Moon acted as an excellent stand-in, confirming the alignment works.

Why This Matters

• The sightline is tightly framed and runs parallel to and directly above the long axis of the Altar Stone.
• It forms an ~80° angle with the primary solstice axis (summer sunrise to winter sunset) — exactly as expected for the solstice extremes at Stonehenge’s latitude in 2500 BC.
• This supports the earlier proposal by Gordon & Phyllis Freeman and reinforces the idea that the monument was deliberately designed around two intersecting solstice axes from the start.

You can read Simon’s blog post here: Using the Moon as a Proxy for the Ancient Sun

(Note: Simon has also written a more detailed peer-reviewed version which is currently behind a paywall, https://doi.org/10.1558/jsa.33686 )

Stonehenge Access All Areas - Film 1

Fremington Pottery

 I was very kindly given this Fishley Jug from the Fremington Pottery by Julian Richards. A treated gift. From ice age clay to a beautiful and useful object.

New Borehole Analysis Sharply Constrains the Fremington Clay: True Glacial Lake Limited to Low Elevations (~15–28 m OD), with No Lacustrine Deposit South of the Hele Ridge

Borehole Map from https://api.bgs.ac.uk/sobi-scans/v1/borehole/scans/items/703466

A detailed OD-referenced re-examination of the 1972 C.H. Brannam Ltd borehole campaign (BGS records SS53SW62–79) fundamentally refines our understanding of the Fremington Clay Series in North Devon. This 18-hole dataset, drilled for commercial pottery clay prospecting, provides the first systematic elevation control on the deposit. The results demonstrate that the genuine glaciolacustrine "potters clay" — the stone-free, homogeneous lacustrine unit formed in a proglacial lake — is far more restricted in both lateral extent and elevation than shown on BGS Sheet 293. Much of the mapped "Boulder Clay" is actually periglacial head and solifluction deposits. There is no Fremington lacustrine clay confirmed south of the Hele–Bickington ridge, and no lacustrine clay above ~28–29 m OD.

George Maw (1864) originally described a clear two-unit sequence: a lower stone-free potters clay resting on glaciofluvial gravel, overlain by an independent stony gravel/head produced during later erosion and periglacial reworking. Subsequent work by Stephens (1966) and the BGS (Edmonds et al. 1985) collapsed this distinction through a traceable sequence of conflations — Taylor (1956) had already grouped coastal and inland erratics together, inflating the apparent glacial signature; Stephens then treated the entire surface-to-bedrock sequence as a single boulder clay; and Edmonds mapped the valley floor accordingly, drawing an admittedly arbitrary boundary between boulder clay and pebbly drift. Crucially, Edmonds himself noted in the 1985 memoir that gravel was found only in holes that penetrated smooth clay — the diagnostic key to the whole problem — but never followed this observation to its mapping consequences. The 1972 Brannam logs allow us to restore Maw's distinction with quantitative precision.

Of the 18 boreholes, only 8 reached the diagnostic basal glaciofluvial gravel, confirming the full lacustrine stratigraphy (potters clay on gravel on bedrock). Their gravel-top elevations cluster tightly between 15.3 m and 28.0 m OD, with the basin floor around 20–23 m OD and margins rising to ~27–33 m OD. This matches independent data from the A39 Lake cutting (Cattell 2003: clay base 19.9–26 m OD), Maw's Roundswell well (~26.8 m OD clean clay top), and a 1982 A39 bypass borehole (703458) that positively identifies Fremington Clay at ~20.8 m OD — making 9 confirmed lacustrine holes across the full 31-borehole dataset. The ice-dammed lake surface therefore stood no higher than approximately 27–29 m OD.

Three boreholes (9, 10, 11) at surface elevations 41–43 m OD encountered "clay" tops at 38–40 m OD — 12–20 m above the confirmed lacustrine basin. These are unequivocally periglacial head. The remaining non-gravel holes show poor-quality "dirty" clay consistent with head masking thinner lacustrine deposits or basin-margin facies. The pattern is reinforced by a further sample of 13 archival boreholes: 12 are dominated by head throughout, with no trace of stone-free lacustrine clay. None of these additional holes, nor the 1972 set, records Fremington lacustrine clay south of the Hele–Bickington ridge. Notably, C.H. Brannam Ltd — guided by T.R. Wood's prior seismic survey and by generations of direct working knowledge — drilled entirely on the low ground around the existing pits and showed no commercial interest in the ground south of the ridge. A pottery firm whose livelihood depended on finding workable clay would not have overlooked accessible reserves: their silence on the southern ground is itself evidence (Sutton's Law in reverse) that the mapped Fremington Clay extent there reflects head and reworked material rather than genuine lacustrine deposit.

Implications for Irish Sea Ice Incursion and Erratics

This OD-constrained geometry severely limits models of onshore Irish Sea ice advance (likely Anglian/MIS 12). The Fremington potters clay records only a modest proglacial lake ponded in the Taw–Torridge estuary, not a widespread till sheet from grounded ice overriding the Devon landscape to 80–90 m OD. Ice must have been pinned near the Bristol Channel coast. The restriction of lacustrine clay to ~15–28 m OD directly challenges interpretations that rely on high-level glacial material and extensive ice incursion inland.

For producing a rigorous geological map of the glacial clays — loosely mapped and extensively mixed with head deposits — these boreholes supply critical control points. Future work will include a corrected map separating confirmed lacustrine clay from periglacial head, cross-sections (e.g., Lake–Roundswell transects showing the 11° northward basin-margin dip), and targeted investigation south of the Hele–Bickington ridge.

Relevance to Bluestone Transport

A thick, grounded Irish Sea ice sheet capable of glacial rafting or long-distance transport of erratics across Devon is incompatible with this restricted, low-elevation lacustrine basin. The data strengthen the case for predominantly fluvial/periglacial origins for most superficial "glacial" features and constrain maximum ice-surface elevations in the Bristol Channel. This directly informs the viability (or otherwise) of glacial hypotheses for Stonehenge bluestones.

The full April 2026 report (CC-BY-4.0), with tables, maps, and the unified borehole register (31 holes total), is available: 
Constraining the Fremington Glaciolacustrine Clays: and its implications for the Irish Sea Ice extent

and a related paper: A Case for Predominantly Fluvial and Periglacial Origins of the Fremington Clays, Devon

Stonehenge – Access All Areas. Coming soon

A major new 11-part documentary series from archaeologist Julian Richards

One whiteboard. Hundreds of questions. Decades of digging. Still no neat answers.

How did they move 30-tonne sarsen stones, and bluestones all the way from Wales? What was the route? Why choose those exact stones — and why from so far away? How did they shape, transport, and erect them with nothing but stone, wood, and muscle?

Why align the whole monument so precisely to the midwinter sunset and midsummer sunrise? What does the sun mean to the people who built it? Was this a place of the living, the dead, or something far more profound — a changing society marking time, ancestry, and the cosmos?

Was it a temple? A calendar? An ancestral monument? A healing centre? Or something we still haven’t imagined? Why are some stones missing? Why were others deliberately broken or buried? And why, after 5,000 years, does Stonehenge refuse to surrender its deepest secrets?

Julian Richards has spent over 35 years exploring Stonehenge and its landscape. In this no-holds-barred series he goes Access All Areas — tearing apart old theories that no longer work, testing the latest ideas, following the bluestone and sarsen trails, examining the evidence from digs, scans, and the stones themselves, and confronting everything that still doesn’t add up.

You’ll see the real process of archaeology: the “maybe it was Merlin” guesses, the failed hypotheses, the new data from modern science, the solar alignments, the landscape context, the ancestral connections — and the honest admission that we still don’t know it all.

No fluff. No tidy Hollywood ending. Just rigorous, passionate, on-the-ground investigation into Britain’s greatest prehistoric enigma.

Eleven episodes. From Sunday 3 May 2026.

First episode drops on YouTube: → Subscribe now at youtube.com/@julianrichards1483

If you think you know Stonehenge… think again. This is the series that goes deeper than ever before.

The Boring Hydrology of Stonehenge and Avebury

 

Why Water Tells Us Almost Nothing About the Monuments

In recent years, one researcher has spun elaborate tales of a post-glacial “high-water world” where Stonehenge sat on a watery peninsula, the River Avon was a tidal superhighway, and massive aquifers turned Salisbury Plain into a navigable inland sea until the megalithic age collapsed around 3000 BCE.

The reality is far more mundane.

Ground investigations for the A303 Stonehenge tunnel scheme produced dozens of boreholes that revealed the true nature of the underlying chalk. As detailed in the 2017 paper Stonehenge — a unique Late Cretaceous phosphatic Chalk geology (Mortimore et al., Proceedings of the Geologists’ Association), these cores exposed complex fault-controlled phosphatic chalk deposits and periglacial features formed millions of years ago — but nothing resembling the prolonged post-glacial flooding claimed by some researchers. Instead, they confirm a classic, well-drained chalk downland landscape.

The hydrology of Stonehenge and Avebury is classic chalk downland — well-drained, spring-fed, and locally variable, but fundamentally unremarkable. It explains the placement of a few features (like the Avenue’s route or proximity to the Avon) but offers virtually zero insight into the purpose, design, or cultural significance of the great monuments themselves.

Stonehenge: Dry Uplands Beside a Normal River

Stonehenge stands on a chalk spur at ~100–105 m OD, overlooking the Avon valley. Stonehenge Bottom (the dry valley to the south) is a periglacial feature incised during the Devensian cold stage. Borehole data from the A303 investigations show typical Chalk Group characteristics: solution features, periglacial head deposits, occasional Holocene alluvium in the valley floor, and localised organics — exactly what you’d expect in a well-drained upland with episodic floodplain activity.

• Water table: Modern and Holocene water tables fluctuate but do not indicate persistent deep flooding around the monument. Control boreholes on higher ground show intact chalk. Springs exist (e.g., at Blick Mead, ~2.5 km away), but these are normal chalk springs with stable ~11°C temperatures — attractive for Mesolithic campers, not evidence of a submerged landscape.
• River Avon: A typical chalk river with seasonal flow. The Avenue connects Stonehenge to the river, likely for processional or symbolic reasons (domain of the living vs. domain of the dead, per the Stonehenge Riverside Project). There is no evidence it was a “tidal superhighway” or that stones were routinely floated in vast post-glacial rivers. Bluestones and sarsens were moved overland or by river in manageable stages.
• Ditch and bank: Excavations show the henge ditch filled rapidly with chalk rubble from the weathering sides and bank, with only thin organic-rich “dark layers” that represent soil development or gradual accumulation on a damp base — not evidence of prolonged standing water or a deliberate moat. Chalk is highly permeable, so any water would drain quickly. The ditch was a symbolic boundary, not a water-holding feature fed by aquifers..

Claims of “long-duration flooded systems” at ~92.6 m OD rely on reinterpreting standard karstic, periglacial, and minor alluvial features as proof of exotic hydrology. Standard Quaternary science finds no such need.

Avebury: Even Drier and More Ordinary

Avebury sits in the upper Kennet valley on chalk with thin soils. The River Kennet is a small, intermittent chalk stream (winterbourne). Research shows:

• Low erosion and minimal alluviation in the Neolithic/Bronze Age.
• Springs and wells were important but unreliable in summer — people dug wells into the chalk aquifer.
• The henge ditch could hold seasonal water but was not part of a vast flooded network.

Silbury Hill is a massive, carefully engineered Neolithic mound beside the Kennet. Its surrounding ditch does fill with water seasonally in winter due to the naturally high water table in the valley floor — a phenomenon still seen today after heavy rain, creating a temporary “moat” effect. However, this is ordinary chalk hydrology, not evidence of a permanent flooded harbour or dramatic post-glacial inland sea. Silbury Hill is not a “lighthouse” for boats in a high-water world — its form, construction, and location are consistent with monumental/ritual functions on dry or seasonally damp ground.

Wansdyke, meanwhile, is a much later post-Roman linear earthwork, not Neolithic or prehistoric. It runs up and down the porous chalk ridges of the Marlborough Downs. On such highly permeable geology, it would be as likely to function as a canal as a Martian fault line — any water would simply drain away through the chalk aquifer, and its undulating profile makes no hydrological sense for navigation. Its alignment and profile are consistent with boundary or defensive purposes in a normal, well-drained chalk landscape, not a prehistoric waterway in a “high-water world.”

Silbury Hill and other features interact with local water (e.g., springs near the ditch), but again, this is normal chalk hydrology — not evidence of dramatically elevated post-glacial rivers.

What Hydrology Actually Explains (and What It Doesn’t)

It explains:

• Why people camped at springs like Blick Mead in the Mesolithic.
• Route choices (e.g., the Avenue following a natural corridor to the Avon).
• Practical considerations for construction (access to water for workers, but nothing extraordinary).

It explains almost nothing about:

• Why the monuments were built in these specific locations (solstitial alignment, landscape theatre, ancestral significance, social aggregation).
• The purpose of the stones, circles, or burials.
• The cultural or astronomical motivations.
• The transport and erection of megaliths (human organisation and effort, not super-rivers).

The builders chose well-drained, prominent chalk locations with good visibility and symbolic power. They were not fleeing floods or mooring boats at ritual harbours. The early Holocene was wetter than today with higher water tables in lowlands, but Salisbury Plain and the Marlborough Downs remained open, stable grassland ideal for large-scale monument building — not a saturated swamp.

Conclusion: Boring Is Good

Real hydrology at Stonehenge and Avebury is boring in the best scientific sense: predictable, consistent with regional geology, and fully compatible with mainstream archaeology. It requires no radical reinterpretation of sea-level curves, invented “90% terrace rules,” or global aquifer collapse to explain the monuments.

The monuments’ enduring mystery lies in the human story — ritual, cosmology, ancestry, and communal effort — not in dramatic (but evidence-free) flooded landscapes. Over-hyping the water distracts from what actually matters.

The chalk downs were dry enough, the rivers ordinary enough, and the springs local enough. That’s the truth the boreholes reveal — and it’s far more interesting than the alternative.

The Wreck of the Bluestone Raft - Not historically accurate