The chalk has eaten my homework

That is the new line : "The chalk downs of Salisbury Plain and adjacent areas are geochemically lethal, and are perfect places for the efficient degradation and destruction of any scattered patches of till and glaciofluvial deposits that might once have existed." Not “we found till.” Not “here is an erratic train.” Not “here is a glacitectonite, a tunnel valley, or an outwash braidplain that cannot be explained any other way.” The gentle rolling downs of Wessex have become, overnight, a geochemical sinkhole: porous, aggressive, lethal to carbonate — and somehow fatal to resistant silicates as well — so thorough that an entire Irish Sea till sheet can vanish without a trace while a curated cluster of Welsh monoliths sits conveniently where a Neolithic monument later appears.

Chalk does dissolve. That is ordinary downland karst and periglacial geomorphology, not a revelation. White Chalk is often more than 98% CaCO₃. Meteoric water charged with carbonic acid removes that matrix. Solution pipes, decalcified gravels, coombe rock and head have been mapped across Wessex for well over a century. Clay-with-flints is a remanié of chalk dissolution mixed with Palaeogene cover and later cryoturbated material. Jukes-Browne showed long ago that pure chalk contains far too little insoluble residue to yield the plateau clays we see. None of that requires a Pleistocene ice sheet on Salisbury Plain.

What the “lethal chalk” story gets wrong is the next step. Chalk porewaters and rendzina soils are alkaline to circumneutral, buffered by excess carbonate. They are not an acid bath that eats silicates. Even if acidic soils formed on former Palaeogene caps, quartz, zircon, rutile, tourmaline, and dense igneous clasts do not dissolve away on Quaternary timescales. Frost may shatter a dolerite erratic. Periglacial attrition reduces it to gravel, sand and heavy-mineral grains. It does not vaporize silicate chemistry into thin air.
What a glacier leaves behind, even after every scrap of carbonate matrix has gone, is that durable fraction: quartz, zircon, rutile, tourmaline, micro-lithics of dolerite and rhyolite, and far-travelled pebbles concentrated in residual soils and river gravels. That test has been run.

Stream sands draining the Plain carry detrital zircon suites whose age spectra match recycled southern British Mesozoic and Palaeogene sources — not a first-cycle dump from the Lower Palaeozoic of Wales or a crystalline northern shield. Out of hundreds of grains, a lonely ~464 Ma zircon is the antithesis of the Welsh igneous signal required if ice had transported dozens of multi-tonne bluestones and ground the rest into rock flour. The apatite record is dominated by a Cenozoic reset signal, not Palaeozoic basement. The minerals any real till sheet must leave behind are present in the drainage. They show no evidence of a Welsh ice sheet.

Clast counts in the Wiltshire Avon terraces show local flint, sarsen and Greensand chert — not the exotic suite an Irish Sea or Welsh ice stream would impose. Field surveys still fail to locate a bluestone erratic in an undisturbed Quaternary context away from the monument and its debitage. The Newall boulder matches Craig Rhos-y-felin rhyolite of the Stone 32d group, moved by people. Coastal patches in North Devon or Somerset do not put a grounded ice lobe on the Wiltshire chalk plateau. The Anglian limit of preserved evidence remains well to the north.

Numerical ice-sheet models are not a substitute for deposits on the ground. A model can be made to grow ice almost anywhere if the climate forcing, sliding law and bed friction are generous enough. That is why older, lightly constrained experiments have been cited as showing it was “perfectly possible” for ice to reach east of the chalk scarp. Possibility in a sensitivity run is not a reconstruction. The reconstructions that are actually tied to mapped margins and dates — BRITICE-CHRONO and the later ensemble tests against flow geometry and deglaciation timing — put Devensian Irish Sea ice to the Celtic Sea shelf edge and onto parts of the north Devon and Somerset coasts. They do not put a grounded lobe over the Wiltshire chalk plateau. The Anglian ice sheet, the one usually invoked for a 450-ka till on Salisbury Plain, is even less a creature of those high-resolution Devensian models. Its limit of preserved evidence remains well north of Wessex. And even if a future ensemble were forced across the downs, the prediction would still be a first-cycle Welsh and Irish Sea mineral load in the Avon sands and terrace gravels. That load is not there. A model that contradicts the grains is a discarded run, not a hidden ice sheet.

The argument has retreated to a closed loop of auxiliary hypotheses:
1. Missing moraines and till → cold-based ice.  
2. Need large megaliths moved → warm-based basal plucking.  
3. Missing gravel-sized erratics across the Plain → corrosive chalk chemistry.  
4. Why the bluestones survive at one later monument → left unexplained.

Each failed prediction is reclassified as a feature of the model.

Absence of evidence is not evidence of a dissolved till sheet. After 150 years of stratigraphy, terrace clast surveys and grain-scale U–Pb geochronology, the record across Salisbury Plain is not silent. It is negative. The “lethal chalk” hypothesis is an ad hoc excuse. It does not survive contact with the mineral grains.

A stranger in the family grave

Thirty new genomes from southwestern Britain: Bronze Age newcomers burying their dead in thousand-year-old Neolithic tombs — and why we still can't say whose ancestry the last seven per cent was.

---

Around 3200 BC four people were buried together at Monkton-up-Wimborne in Dorset. Close family — two first-degree pairs, one mitochondrial lineage between them. Either four siblings, or a woman and three children aged about five, nine and ten. The grave was sealed.

Fifteen hundred years later somebody opened it, laid a man in the middle of the earlier cut and capped him with flint. Not related to them. Not descended from them. From a population that didn't exist in Britain when they were buried.

The same at Sisters Long Barrow, where a young woman was placed in a scoop cut into the top of the mound two millennia after the three men beneath her — and at Sale's Lot, where the secondary burial is also a woman. Two of the three reused monuments in the new paper by Vuković and colleagues received women. The authors are right that three burials prove nothing. It is a well-timed three, six weeks after aDNA showed the Upton Lovell "shaman" — goldworking toolkit, ceremonial cloak, battle axe — was female.

(a) Map of southwestern England showing the location of each of the investigated burial sites. Black square on the inset map indicates the location of the study area. (b) Radiocarbon-dated individuals in the study per site (y-axis). For each individual, the 95.4% calibrated probability interval is shown in grey. Individual labels on the x-axis are coloured blue (male) and red (female). The orange dotted vertical line represents the point estimate for the Amesbury Archer; its 95.4% calibrated probability interval is 2470–2239 BCE. - Fig 1 from Vuković et al.

The paper covers thirty individuals from twelve sites in Gloucestershire and Dorset, c. 3800–1400 BC. Everyone before about 3100 BC looks like a European Neolithic farmer; everyone after about 2550 BC looks Bell Beaker. The archaeological reading — monuments keeping their meaning for people with no ancestral claim on them — is careful and, I think, right.

I'm interested in a different part.

The question they tried to answer

British ancestry changed by something like ninety per cent between 2450 and 2000 BC (what that number measures). The interesting remainder is the seven to nine per cent of Neolithic-farmer ancestry in English Chalcolithic and Early Bronze Age people.

Everything turns on where it came from. Descended from Neolithic Britons, and some Neolithic families had descendants. Carried across the Channel by incomers who already had farmer ancestry, and it says nothing about British survival at all. Same number, opposite meanings — at length and formally.

This paper tests it better than anyone has. The usual weak point is the local Neolithic reference, pooled from wherever data happen to exist. Here it is nine individuals from the same regional transect, in some cases the same monuments as the targets.

What came back

The earliest Bronze Age individual — the Sale's Lot woman, c. 2620–2460 BC, buried within a generation or two of the Amesbury Archer — comes back entirely continental. Most of the rest are estimated at zero, or at a positive number whose error bars include zero. Likewise across a further 118 English Bronze Age individuals from the published record.

One exception, which the paper flags as its clearest positive: a woman from Fir Tree Field Ring Ditch, at 45.3 per cent local Neolithic ancestry. The confidence interval runs from 4.2 to 86.3 per cent.

The strongest evidence in the study for Neolithic British ancestry surviving cannot distinguish a trace from most of a genome. It rests on fewer than fifty thousand markers. And she lived around 1400 BC, in the window where farmer ancestry across Bronze Age Britain rises again — a rise generally put down to renewed contact with the continent.

The part that matters

These are whole genomes, shotgun-sequenced and publicly deposited. Every downstream analysis then reduces them to single random base calls at 1.2 million pre-chosen positions — the 1240K panel, the field's common currency for a decade.

Not laziness: it is the only way to talk to the comparative dataset. But the panel discards the rare variation carrying information about specific recent shared ancestors, which is the one thing that could separate a Dorset Neolithic grandmother from a Rhineland one. Sequenced, then thrown away at harmonisation, in a paper published three weeks ago.

The honest qualification is that coverage is thin — median 0.27×, range 0.014–0.78× — so the better analysis probably wouldn't have run anyway. Which makes the constraint structural rather than accidental. (These libraries also mean English Chalcolithic–EBA shotgun data now exist, where when I wrote there weren't any.)

Where it leaves things

The arithmetic is unchanged: seven to nine per cent, well measured, unaddressed. What has changed is that the failure has been reproduced independently, with a better local reference than published data allow, by people with a laboratory.

The man in that Dorset grave wasn't descended from the family whose bones he was laid among. That much is solid, and striking. Whether anyone in Bronze Age Britain descended from Neolithic Britain, we still can't say.

---

Vuković, N., Bernhardsson, C., Edlund, H. et al. Diachronic reuse of Neolithic burial monuments by Bronze Age newcomers in Southwestern Britain. Sci Rep 16, 26819 (2026). https://doi.org/10.1038/s41598-026-66094-z

Rock on a Rope, how to level the lintels

 

In an earlier post on setting out the Stonehenge sarsens I suggested a simple way of establishing a horizontal: https://www.sarsen.org/2014/08/how-to-construct-level-sarsen-circle.html , place two posts either side of a small pool of water (an “Aurochs skin full of water”) and sight across their tops. Adjust the posts so their tops both have the same length of post above the water surface, the line of sight is level and can be transferred around the circle. 

A complementary approach starts from a different but equally basic physical fact: a free-swinging pendulum released from rest reaches essentially the same height on the opposite side of its arc. A tall central timber mast, a long fibre rope, and a heavy stone or timber bob are all materials the Late Neolithic builders already used. Release the bob from one marked position and it will rise to the matching height on the far side; adjust a second post until the bob just reaches it. For more accuracy, you release the pendulum from post A and it marks on post B the end of the first swing. When it comes back to post A the second time, it's going to be just a little bit lower because of air resistance. Halfway between the first and the second mark on post A will be the same as a mark on post B. Subtle and easy to get an accurate reading. Repetition is quick, free and easy so a reliable and repeatable measure can be recorded.  Rotate the plane of swing or shift the mast a short distance and a small set of levelled reference posts can be established around the intended circle. Once three or four reliable points exist, ordinary sighting or taut-string transfer finishes the job.

The two methods form a small family of “gravity levels”:

• Water-pool / skin method – immediate visual reference, excellent for a local datum, but needs a still surface and a reliable container.
• Pendulum method – needs no liquid, generates widely spaced equal-height points from one central mast, and is cheap to repeat when wind deflects the swing.

Neither is attested archaeologically. Both are simply possible with the technology of the time and address the practical problem of creating a consistent horizontal on sloping chalk so that the sarsen uprights could finish with level tops for the continuous lintel ring. And avoid the need for a water filled pig intestine or portable troughs as are often suggested as alternatives.

Pendulum method being tested

Beyond Stonehenge, Treasures in the Landscape - Stonehenge Access All Areas, Ep 9

The Sarsen Transport Routes - A Least Cost Model

A terrain model you can argue with. Open this to understand this post: https://sarsenroute.netlify.app/ 

---

Fifty of the fifty-two sarsens at Stonehenge came from West Woods, near Marlborough. That was settled in 2020 by matching their chemistry to the source. It leaves the interesting part open: how thirty-tonne stones travelled 25 km south, across the Vale of Pewsey, to the monument.

Several routes have been proposed over the years, mine among them. None has field evidence behind it. So rather than argue for another line on a map, I set out to ask a narrower question:

If the only things that mattered were the shape of the ground, what it is made of, and the water on it — where would the cheapest road run?

The answer is not a line. It is a corridor: 37 km², about 5% of the landscape searched. Within that band the terrain barely cares which way you go.

Doing it blind

I am one of the people with a published route, which is a problem. So the method was written down and committed before any data arrived, and the published proposals — including my own — were sealed in a file that stayed closed until the result was frozen. No proposed route, mine included, is scored anywhere in the paper.

---

https://sarsenroute.netlify.app/ 

What the ground says

Four things came out that I did not expect.

The vale crossings are on sand, not clay. National geological mapping shows the Vale of Pewsey floor as one great smear of Gault clay and greensand. Zoom in to 1:50,000 and the five places where roads actually cross the vale turn out to sit on greensand benches — firm ground — with the real clay lying to the west, between Stert and Urchfont. Across all 360 model runs, not one metre of route touches Gault. The clay is real. Nothing goes near it.

There are only five crossings in eleven kilometres of vale — and adding every footpath and bridleway to the metalled roads adds no new ones. The same five places. When footpaths don't create new crossings, the crossings are being fixed by the ground.

Better data made a question harder, not easier. Two ways down off the Marlborough Downs, 2.8 km apart, are what the model has to choose between. At 50 m resolution they differ by 1.7% in cost. At 1 m, measuring the gentlest line the ground allows, they differ by 1.2 percentage points of gradient. More resolution made them more alike. Huish is not steep ground.

What separates them isn't the slope. It's use. One descent — Walkers Hill — carries a worn way running continuously from 219 m down to 140 m at a steady 6%, with the dished profile of a holloway. Andrews and Dury drew a road down it in 1773 and none down the other. The Huish face has a narrow farm track and nothing else.

Andrews and Dury's map of 1773

The bit I find most interesting

With no thumb on the scale, the model preferred the unused descent.

That isn't a bug, it's a bias — and I suspect it's general. A 50 m terrain model averages a scarp face. A smooth, unwalked hillside averages nicely. A hillside with a deep worn track in it averages badly. So the model kept choosing the easier- looking slope precisely because nobody had ever used it.

Feed in the evidence that one descent was used and the other wasn't, and it moves. But the model needed telling. It could not work that out from the ground alone.

What it can't do

A least-cost path is not a route. It knows nothing about ownership, tradition, season, weather or who you were allowed to walk past. It says what the ground permits, not what people did.

And it doesn't reproduce my own route. I think the stones came down Walkers Hill and crossed the vale beside Honey Street, on a low rise you can see on the ground. The rise is real — 1 m laser survey and the terrain model both find it, standing a couple of metres above the ground either side. What isn't shown is that it was made, or that anyone used it. The model doesn't go that way, and I've left that disagreement in the paper rather than tidy it out. A model that agreed with me would be worth less.

Have a go yourself

The model is online as a set of sliders: sarsenroute.netlify.app

Three hundred and sixty routes, pre-solved. Turn up the fear of hills, or of wet feet, or of river crossings, and watch the corridor move. There is no correct setting — that's the point. The route is a function of what you think a hauling party was most afraid of, and the tool makes that dependence visible instead of hiding it behind one confident line.

The full technical paper, with the method, the data and the things that went wrong, is on https://www.researchgate.net/publication/413532460_Modelling_the_sarsen_transport_route_West_Woods_to_Stonehenge_A_blind_least-cost_analysis_over_layered_open_data and https://www.academia.edu/172184320/Modelling_the_sarsen_transport_route_West_Woods_to_Stonehenge_A_blind_least_cost_analysis_over_layered_open_data.

Data: OS Terrain 50, BGS Geology, Environment Agency LIDAR (all Open Government Licence), OpenStreetMap. Analysis and drafting with Claude and Grok; the errors are mine.

Avebury 3D Fly Through

A 3D topographical fly through of the Avebury Landscape - https://avebury-3d-flythru.netlify.app/ . Play with it and get back to me with comments. It is schematic and only shows a sample of the monuments, but importantly it shows Silbury in relation to its neighbouring hills.

All built with Open Source Data so is freely shareable.

The various periods which you can switch between in the 3d model.

Click to expand

Why We Still Don’t Know Where Britain’s “Missing” Neolithic Ancestry Came From

I’m not a geneticist by trade. I’m an amateur who got stuck on a question that the big Beaker papers leave half-open, and then spent a long time trying to close it with public data only—the same call sets, tables and annotations everyone else can download.

The question sounds small. After the great ancestry turnover in Britain around 2450 BC, English Chalcolithic–Early Bronze Age genomes still carry roughly 7–9% ancestry that looks Middle/Late Neolithic-related. Where did that residue come from?

• Local British farmers absorbed into incoming communities?

• Or continental people who were already mixed before they arrived?

Those are different pasts. One implies survival and incorporation on this island. The other implies the “British-looking” fraction was already baked into migrant groups. Group averages hide the difference. You need to know whether almost everyone carries a thin layer, or whether a few people carry a lot and most carry none.

I thought public IBD, f-statistics, Y/mt labels, and the shape of published ancestry estimates would settle it. They don’t. What I think I have produced—and why I’m bothering to post—is a clearer map of why they don’t, and therefore where the problem has to be solved next. Not a funding pitch. A diagnosis.

What I actually found

1. The residue looks population-wide, not a mosaic of survivors.

On the Booth × Olalde overlap (n = 28), once one extreme individual (I2462) is set aside, the group is statistically homogeneous at about 7.15%. Between-person scatter is small; models where only a minority carry the component and everyone else carries none are excluded. Booth’s gradual rise through time reproduces—and it is a rise in everyone. That is a real result, and as far as I can tell it hasn’t been stated this way with a scripted dispersion bound behind it.

It still doesn’t tell you provenance. An already-mixed import is uniform from day one. A local pulse absorbed over ten to fifteen generations is uniform by the time we sample Early Bronze Age people. Same snapshot. Different histories.

2. The “trace it home” instruments fail on public products.

Long shared segments (IBD), allele-frequency contrasts aimed at British private drift, and uniparental “insular” screens all break under ordinary audits: leave-one-out, site pruning, shared-depth truncation, positive controls. The discriminating information—rare variants, fine Y structure, long-segment continuity at time depth—isn’t reliably in the 1240k / AADR-shaped data we all use. Sometimes the burial still holds it; the assay and the label conventions don’t.

3. Those are two different kinds of failure.

Three instruments fail because information was discarded (capture + harmonisation). The residue-shape instrument fails because information was erased (admixture before these people lived). Collapsing both into “we need more aDNA” is how you waste the next decade.

I also built a Sardinian re-imputation instrument to ask whether public IBD nulls are partly old pipelines rather than ascertainment. It validates within poles; the decisive cross-pole test is specified and not run due to the constraint of my available resources.

Where and how to solve it (not “who should fund it”)

If the goal is provenance of that 7–9%, the map is annoyingly specific:

A. Where the assay threw the information away

Hold the same individuals and change the data product: shotgun (or denser) sequencing → modern imputation → rare-variant / fine-lineage / IBD methods. Especially useful where the same person already exists in both capture call sets and shotgun releases (a clean ascertainment test). Published library-quality metrics for the people I’d prioritise suggest this is practical for most of that set, not all—some look too thin on public annotation alone.

B. Where time erased the information

Don’t keep sequencing well-sampled Early Bronze Age individuals expecting residue structure to reappear. Sample closer to the transition—Chalcolithic / earliest Bronze Age—where an absorbed local pulse might still show heterogeneity. That is a dating-and-excavation problem as much as a sequencing problem.

C. What not to confuse with a solution

More Beaker genomes won’t answer this. Ireland isn’t a free continuity control across the same interval. Collapsing Y strings to “I2” and calling it local doesn’t survive a shared-depth check. And a homogeneous ~7% background is evidence about structure, not a passport stamp for “British Neolithic survival.”

Why post this as an amateur?

Because the literature often states the residue, then slides past the provenance question—or treats every failed test as “low power” instead of “wrong information class.” I’m not claiming a final β. I’m claiming a usable negative architecture: four instruments, two kinds of limit, and a concrete split in where the next honest attempt has to go.

The draft paper: 

https://www.academia.edu/171955479/The_last_seven_per_cent_a_known_unknown_in_the_British_Beaker_transition_Four_instruments_two_kinds_of_limit_why_the_provenance_of_Neolithic_ancestry_in_Early_Bronze_Age_Britain_cannot_be_resolved_on_published_data

https://www.researchgate.net/publication/412356174_The_last_seven_per_cent_a_known_unknown_in_the_British_Beaker_transition 

Full deposit (pre-registrations, results, code, audit trail):

https://doi.org/10.5281/zenodo.21995822 (https://zenodo.org/records/21995822)

https://github.com/TimDaw37/british-neolithic-ibd-audit-supplement

If I’ve got something wrong in the genetics, I want to hear it. If the diagnosis is roughly right, the next useful work isn’t another average—it’s either denser data on the right people, or older people on the right dates.

───

• “Uniform at 7% is a result. It just isn’t a provenance result.”

• “Some missing answers are missing chips. Some are missing centuries.”

• “I’m an amateur. The deposit is public. Please break it if you can.”