henge

The stones and their journeys

The sarsens — Established

• About 30 uprights of the outer circle with a continuous lintel ring; five trilithons in the inner horseshoe, graded toward the Great Trilithon. Silcrete (“sarsen”), dressed hard on the inward and processional faces.
• Each standing sarsen is drawn at its own width and thickness, read off the base outline of the Rees 1989/90 plan as digitised by Tim Daw (CC BY-SA 4.0): 1.1–2.5 m across and 0.7–1.4 m thick, not one shared section. The laser-scan volumes (issue #3) will refine the sections.
• Stone 11 is the short, narrow one: 2.74 m high (Cleal, Walker & Montague 1995, App. 5) and 1.12 m wide. As it stands the app draws it at that height. As it was, the app draws the circle complete, with Debated support from the 2013 parchmarks for the missing south-west uprights (Banton et al. 2014, Antiquity 88). Stone 11’s full height in that view is this app’s reconstruction, cited to no source, chosen by the owner on 28 Sep 2026 so that the lintels it carries are held up: the stone survives short and narrow, and whether it was broken, replaced or always small is not settled (the in-app badge is issue #71). The rule that raises it catches only a circle upright short of the lintel seat by more than the real ring’s own unevenness — 0.19 m, the largest departure of any top from the mean in Petrie’s 1877 levels. Stones 22 and 23 stand a little under the model’s level seat, but Petrie levels their tops within the ring’s spread, so they keep their recorded heights in both views.
• Provenance: West Woods, Marlborough Downs, ~25 km north — Nash et al. 2020, Science Advances 6:eabc0133, geochemical match of 50 of the 52 surviving sarsens against the core drilled from Stone 58.
• Deeper petrology, weathering behaviour and lichen cover are worked through in research/sarsen-geology.md — audited and tiered, and since 2026-09-27 applied in the renderer: as it stands, the stones carry lichen by species (Ramalina, Candelariella, a trace of Xanthoria; cover Debated, colours Speculative), rust in concave pockets, rain streaks and pick-dressed faces; as it was carries no lichen, because the zonation is a modern artefact (the note’s §7). The one recommendation not taken is the note’s absolute albedo for the fresh stones, left to the owner.

Each stone’s surface: where the values come from — Debated

The colour, finish, tooling and features the app gives each stone (StoneSurface) cite the 2011 laser survey’s own report, page by page: Abbott & Anderson-Whymark 2012, Stonehenge Laser Scan: Archaeological Analysis Report, English Heritage Research Report 32/2012. Its findings are cited; its tables, scan data and images stay out of the repository under the licence rule. The purple-grey (53, 154, 56) and orange (54, 55, 156) sarsens are pp 20–21; the circle’s pick-dressed faces and the unworked outsides of 14, 15 and 16, and 14’s unworked sides, are p 22; the trilithons’ fine inner faces pp 20–21; the ribs on the horseshoe’s outer faces p 14; the oval tenon on 56 p 23; the unworked Heel Stone pp 13 and 24. Press reports of the survey and web pages, which the values were first taken from, are credited as what prompted them and never cited (owner decisions D2, D12; #68). Every value keeps the Debated tier it had.

Five values the app draws are not borne out by the report. They are kept until the owner decides, and each says so in place of a citation:

• lintel 101 drawn orange; the report’s orange sarsens are 54, 55 and 156;
• a cavity in 54; the report records none, and the large natural hollow it does describe is at the foot of 60 (p 41);
• raw outer faces on the south-west circle uprights other than 14, 15 and 16; the report finds those faces pick dressed (p 22);
• the sides of 14 drawn pick dressed, like the rest of the circle; the report finds “the sides and external face of Stone 14” not worked (p 22; “sides and back of Stone 14”, p 13);
• the natural surface as the default for stones with no transcribed account; the report describes the Bluestone Horseshoe, the bluestone lintels 36 and 150, stone 45 and the Slaughter Stone as dressed (pp 13, 24–25).

Stones 36 and 68 — Debated

The sources disagree about whether these two stand. The owner decided on 28 Sep 2026 (D14) that the newest survey decides, and that each stone’s lore records the disagreement. D14 took the 1989–90 survey plan, as Tim Daw digitised it, for the newest; the 2011 laser survey is newer. On 29 Sep 2026 (answer 3A) the owner kept 68 standing, as the 2011 survey records it, and closed #72 as not planned. On 30 Sep 2026 (answer 2A, #73) the owner applied the same rule to 36: the ruin draws it fallen and largely buried, as the 2011 survey records it:

Source	Stone 36	Stone 68
Daw’s digitisation of the Rees 1989/90 plan	standing	fallen
Petrie 1880, pp 10–11 (surveyed 1874–77)	“a square stone lying with edge up”, north end buried	top levelled; “worked groove down north side”
Cleal, Walker & Montague 1995, App. 5	1 ft 2 in above ground	8 ft 2 in above ground
Engleheart & Newall 1930	fallen, three-parts buried; a lintel with two mortises	—
Abbott & Anderson-Whymark 2012 (the 2011 laser survey, the newest)	fallen and largely buried, lifted in 1954 and reburied (p 25)	one of the six Bluestone Horseshoe stones “standing and intact” (p 24)
What the app draws	as it stands: lying and largely buried, following the 2011 survey (2A); as it was: unchanged, Daw’s row at Cleal’s 0.36 m	standing, following the 2011 survey

In the monument as it stands the app draws 36 lying, its long axis level, most of it under the turf (MonumentScene.buriedDespiteRow, held by testStone36LiesLargelyBuriedInTheRuinOnly). What is surveyed: where it lies and the outline it shows at the turf, 1.26 × 0.82 m, from Daw’s row, and how far it rises, Cleal’s 1 ft 2 in (0.36 m). That figure is the appendix’s part A for 36, from Atkinson’s records (1950s–60s); a lettered part is not a standing height, and since Atkinson lifted and reburied 36 in 1954, its epoch against the 2011 state is uncertain. It is read as the rise because it is the only surveyed one. What is reconstructed: how deep it goes. Petrie records a square stone, so the block is drawn as deep as the outline is across, 0.82 m, and 0.46 m of it lies below the ground. The report’s own figures for the lintel (about 1.83 m long, p 50) and the laser volumes of its exposed part are not used: the volumes stay out of the repo until Historic England grants permission, and Daw’s outline is already the part of its length that shows; the rest of the real stone, north end first as Petrie saw it, is under the turf, and the app does not draw it. The complete monument is unchanged: 36 is drawn from Daw’s row at Cleal’s 0.36 m, as before. The note says the app “draws it lying, largely buried, following the newest survey”, and testStone36sNoteSaysWhatTheRuinDraws holds the note to the scene.

Stone 68 is placed from Daw’s row, which records it fallen, and the app draws the row’s block at the row’s height, which for 68 is Cleal’s 2.49 m: it stands, as the 2011 survey, Petrie and Cleal record it. Its note says so: the app “draws it standing, following the newest survey”, and testStone68sNoteSaysWhatTheRuinDraws holds the note to the scene.

The bluestone circle as it stands

The seven upright stones — Established. Daw’s digitised plan disagrees with the published surveys about which bluestones of the circle stand. It calls 33, 35 and 40 empty sockets, though Petrie numbered all three in 1877, and it calls a dozen stones standing that the surveys record lying. Under the owner’s rule the newest survey decides: the 2011 laser survey names the circle’s upright stones as “31, 33, 34, 37, 46, 47 and 49” (Abbott & Anderson-Whymark 2012, RR 32/2012, p 41). Its own count on p 25 is eight standing, eleven fallen, two displaced and nine buried stumps, without saying which eighth stone it counted, so the app follows the named seven.

Which of the rest are fallen, and which displaced — Debated. The report contradicts itself on the count (eight standing on p 25, seven named on p 41) and does not say which of its two displaced stones are which, so every circle stone missing from p 41’s list is read as fallen. The check on that reading is p 25’s own count: it gives eleven fallen, and the stones read as fallen number eleven (a test holds it). Stone 35 is one the 2011 report does not name: p 40 counts nine buried stumps without numbers, so 35’s status comes from the newest survey that names it, Cleal’s appendix (35A and 35B, p 547), which agrees with Petrie’s “Only a stump” (p 10).

The corrections are a file of their own, stones/bluestone-status-corrections.csv, one row per stone with the page that decides; Daw’s CSV is left byte for byte as imported (#48a). The ruin applies them: the seven stand at Cleal’s heights, and every other stone lies at Daw’s row and outline, rising the highest figure Cleal’s appendix prints for it. The complete monument is unchanged.

Stone	Daw’s plan	As it stands	Deciding source	Petrie 1880	Cleal 1995, App. 5
32	standing	fallen	2011, p 41 (not upright)	p 10, “Foot buried”	p 547, 0.63 m
33	empty socket	standing	2011, p 41	p 10, listed (Hoare 4)	p 547, 1.68 m
35	empty socket	stump	Cleal 1995, p 547 (2011 p 40 gives a count only)	p 10, “Only a stump widely cracked like two stones”	p 547, 35A and 35B, 0.18 m at most
36	standing	fallen, largely buried	2011, p 25	p 10, “a square stone lying with edge up”	p 547, 0.36 m
38	standing	fallen	2011, pp 45–46	no row	p 547, 0.51 m
39	standing	fallen	2011, pp 45–46	p 10	p 547, 0.61 m
40	empty socket	fallen	2011, pp 45–46	p 10, “Overthrown before 1747”	p 547, 0.18 m
41	standing	fallen	2011, p 41 (not upright)	p 11	p 547, two pieces, 0.51 m at most
42	standing	fallen	2011, p 41 (not upright)	p 11, “Fallen in 1747 before trilithon fell”	p 547, two pieces, 0.48 m at most
43	standing	fallen	2011, p 41 (not upright)	p 11, “On edge. Rises 50 deg to east”	p 547, 0.18 m
44	standing	fallen	2011, p 41 (not upright)	p 11	p 547, 0.15 m
45	standing	fallen	2011, p 25 (“now upper” face)	p 11	p 547, 0.25 m
48	standing	fallen	2011, p 41 (not upright)	p 11	p 548, 0.36 m

“Fallen” is Daw’s word for a stone that lies. The 2011 report does not say which of these are displaced rather than fallen where they stood, and the app does not guess. The Bluestone Horseshoe is not in the file yet: there too Daw’s column differs from the 2011 survey (p 41: 67 and 71 fallen, 64 and 65 stumps, 61a a displaced fragment, 68 upright).

The bluestones — Established

• The smaller stones of the circle and horseshoe: dolerites, rhyolites and tuffs from the Preseli Hills, west Wales, some 225 km away (Parker Pearson et al. 2025, Archaeology International 27, p 113; the app said 250 km until wave 2); specific outcrops identified at Carn Goedog and Craig Rhos-y-felin (Bevins, Ixer and colleagues, 2010s), with excavated Neolithic quarrying traces (Parker Pearson et al.).

The Altar Stone came from Scotland — Established, 2024

Primary: Clarke et al., “A Scottish provenance for the Altar Stone of Stonehenge”, Nature 632, 570–575 (14 August 2024), doi:10.1038/s41586-024-07652-1. News summary: Gaind & Smith, Nature 632, 484–485 (2024), doi:10.1038/d41586-024-02584-2 — the article behind the owner’s reference photograph.

• The Altar Stone — the six-tonne greenish sandstone slab lying at the monument’s heart — is not from Wales as long assumed. U–Pb and Lu–Hf dating of its detrital zircon, apatite and rutile grains fingerprints the Orcadian Basin of north-east Scotland, at least 750 km away (the news piece says “hauled 800 kilometres”).
• Implication carried by the paper’s authors: long-range Neolithic transport — likely by sea — and a degree of connectivity across Britain not previously credited for c. 2500 BC.
• In the app the Altar Stone lies recumbent at the heart of the setting, drawn as its own grey-green micaceous sandstone rather than as a sarsen or a bluestone; its lore note carries this finding and the Clarke et al. 2024 citation. Since wave 2 that note calls it “the six-tonne sandstone slab” rather than the slab “lying flat”: whether it ever stood is Debated (Darvill et al. 2012 p 1033 call it a former standing stone; Parker Pearson et al. 2025 p 118 judge it most likely laid flat on purpose), and that reading has its own note, monument.altarStone.setting, which says the app’s recumbent slab is a choice between the readings. monument.altarStone.ruin (Established) records 55b and lintel 156 lying on it (Clarke et al. 2024 p 570); Daw’s plan agrees, and LoreStonesTests checks the footprints overlap.

Construction sequence — Established (dates); Debated (readings)

The app’s phase list is Darvill, Marshall, Parker Pearson & Wainwright 2012, “Stonehenge remodelled”, Antiquity 86(334): 1021–1040 (owner decision D4), plus the two surveys. The rows are Sources/HengeGeometry/Resources/chronology/, each citing its page, and ChronologyTests holds them to the paper (#69, timeline C1, 2026-10-05). Years in cal BC; the span is Table 3’s “suggested dates” (printed p 1026), the anchor the year the sky lands on (the owner’s answer 2, 2026-10-05).

Phase	Span (Table 3, p 1026)	Anchor	In the app
Stage 1: enclosure, Aubrey holes, timber (pp 1027–1029)	3000–2620	2900 BC	draft
Stage 2: trilithons, sarsen circle, Station Stones, D and E (pp 1030–1034, printed)	2620–2480	2500 BC	ships (C5): the first new phase, the one of Stages 1 and 2 whose defining features are mostly drawn rather than listed
Stage 3: central bluestone circle, ditch recut, Avenue (pp 1033–1035)	2480–2280	2300 BC	draft
Stage 4: the bluestones rebuilt as oval and outer circle (pp 1035–1036)	2280–2020	2200 BC	ships, as “as it was”
Stage 5: carvings, Y and Z holes (pp 1036–1037)	2020–1520	1600 BC	draft
Petrie’s survey	1874–77	1877	draft (today’s survey walked back through the cited stone events; the poses before the resets of 1919, 1959 and 1963 are not printed, so those stones are listed)
Today	—	the present	ships, as “as it stands”

• The era strip (timeline C5, 2026-10-08) sits on the phase card: a snapping scrubber over the shipping phases, Stage 2, Stage 4 and today, with a step either side, adjusted a phase at a time by VoiceOver and by ⌥⌘[ and ⌥⌘] on the Mac. Nothing is rebuilt until the finger lifts, so no position between two phases is ever drawn. Stage 1 stays a draft: of what it defines, the enclosure and the Aubrey holes are drawn and the timber settings, stones B and C and 97 only listed (the design plan’s rule, Chronology.clearsListedRule). Each phase on the strip, and the gap after the building, has its own note in the lore.
• “As it was” is labelled Stage 4, c. 2200 BC (Darvill et al. 2012) in all nine languages (the owner’s answer 4), and draws exactly what it drew before the phase list existed. Three differences from Darvill’s Stage 4 are named rather than hidden: stone E stands although D and E were removed in Stage 3 (p 1034), by the owner’s ruling D9; the inner bluestones are drawn as today’s 19-stone horseshoe where Stage 4 had an oval of c. 25 (Table 3, p 1026) or about 24 (p 1035), and the authors think it “highly likely” the horseshoe was cut from it in the Roman period (p 1036) — Debated; and stone 11 may always have been “narrower and shorter” (p 1033).
• Every stone the app draws without a source for it as drawn carries a reconstruction mark beside it (timeline C2, the owner’s ruling D9 and answer 5 of 2026-10-05): position where the app placed it (the lost sarsens 13, 17, 18, 20 and 24, the bluestone fills), height for stone 11, raised from its printed 2.74 m (Cleal p 547) and 96 in (Petrie p 10) to carry the lintel ring, presence for stone E, and in the ruin outline for stones 35 and 40, which lie on the plan’s placeholder outlines where it marks only empty sockets. The phase card’s “Reconstructions” chip names stone E and stone 11 first; VoiceOver reads the whole list; a tap rings the marked stones for four seconds. The renderer never reads a mark, so no picture changes.
• Hypothetical restorations (#37, #38; the owner’s answers 14a–16a of 2026-10-06): one toggle, off by default and offered only where a phase has a reading to draw, which today is Stage 2 alone (on the strip since C5). With it on, a stone stands in hole 97 beside the Heel Stone, the pair either side of the line to the midsummer sunrise (Darvill’s preferred reading, p 1034), and stones D and E stand in their excavated holes across the entrance with the Slaughter Stone, the three-stone façade (Cleal p 285, Darvill p 1034). Where the holes are is Cleal Fig 156, read off its grid and offset to Daw’s frame (ADS terms; stones/cleal-1995-fig156-entrance.csv). Each stone carries a hypothesis mark, and a badge says “Debated” for as long as one stands. It never draws in “as it stands”. Stone E stands in its excavated hole in the default view too, 2.2 m from the mirror of 95 across the axis where the app drew it while the two were read as a portal pair (owner answer 15a, re-pinned 2026-10-07). The computed first gleam of the 2500 BC midsummer sunrise (49.27°) falls between the bearings from the centre to hole 97 (48.04°) and to the Heel Stone (50.88°), by 1.2° and 1.6°; reported, not tuned.
• Changing phase moves the sky to the phase’s year (answer 6): with time travel owned, the same month, day and time; in the free version, that year’s midsummer sunrise (answer 3), so an arbitrary date stays paid while the phases stay free. Switching back to Today lands on now in either version: the present is not travel. Reduced motion jumps instead of walking.
• Erosion of the earthwork and the freshness of the stones are presentation choices in every phase, not citations: no source in the register gives a value between “as dug” and today.
• Features known but not drawn (the Q and R holes, the Y and Z holes, the Avenue’s banks, the timber settings) are listed per phase, waiting for a licensed plan or a mesh; no ring formula stands in for them.
• Petrie’s 1877 is today’s survey walked back (timeline C3, 2026-10-06): chronology/stone-events.csv records what happened to which stone since, each row citing a primary page — 22 and its lintel 122 fell on 31 December 1900 (Gowland 1902 p 37; Cleal p 345); the leaning stone 56, 60.5° above the horizontal and leaning north-east, was set upright in 1901 (Gowland pp 40, 44); 6, 7 and lintel 107 were straightened in 1919 (Hawley 1921 pp 20, 29); lintel 158 was re-erected in 1958 (Fig 190), and p 346 plans six stones for that season without naming them, read here as 22, 122, 57 and 58 (Debated: the sixth is unidentified, and the season’s temporary removals are unnamed); 4, 5 and 60 were straightened in 1959; 23 collapsed in 1963 and was re-erected with 28, 53 and 54 (Cleal pp 345–346; that 23 went back in its own socket and that the other three stood before are Debated readings). Walking back starts from the bluestone status overlay, not Daw’s column, and undoes each event: a stone put back in its own socket stands as surveyed, 56 is tipped by Gowland’s 1901 angle (Debated for 1877, when it lay between 66° and 60.5°), and a stone whose earlier pose no page prints is listed, never guessed (StoneLedgerTests). Petrie’s own table is the independent check: every stone he measured stands or leans, none he records fallen stands, and his lintels in situ are seated or listed. Not in the ledger, because no primary page read records them: the 1920 work on the north-eastern arc (29 and 30 are named only by a popular history; Hawley’s 1922 report is still to read) and the fall of Station Stone 91, which Petrie saw leaning outward (p 12) and which lies today, so 91 is listed in 1877 rather than drawn lying. No year before Petrie’s survey has a scene. Neither shipping phase changes.

What the lore says about the stones — wave 2 (#31, #33, #48, #52–#54)

Drafted and refuted in research/lore-stones-draft-2026-10.md, landed under the owner’s answers of 2026-10-05 (“all recomended”, items 7–10). One kind of claim per note; the figures in each are held to their source by LoreStonesTests (Petrie’s and Cleal’s tables, Darvill et al. 2012 Table 3, Gowland’s feet), and every note ships in all nine languages.

Note	Tier	Says	Rests on
monument.purpose	Debated	What it was for is not settled; the cemetery, ancestors, healing, unification and calendar readings, none agreed	Parker Pearson et al. 2009, 2020, 2025; Darvill & Wainwright 2009; Darvill 2022a; Magli & Belmonte 2023
monument.axis	Established	The horseshoe and circles share the solstitial axis; the app measures the year against it	Darvill et al. 2012 pp 1030, 1033; Ruggles 1997 p 218
monument.axis.calendar	Debated	Built as a calendar? Darvill’s 365.25 days against Magli & Belmonte’s “numerology”	Darvill 2022a p 319; Magli & Belmonte 2023 p 745; Ruggles 1997 p 203
monument.design	Established	The four settings, outside in; the complete sarsen ring is “most likely”	Darvill et al. 2012 pp 1022, 1030, 1033
monument.design.sequence	Established	Sarsens in Stage 2 (2620–2480 cal BC), bluestone circle and oval in Stage 4 (2280–2020 cal BC)	Darvill et al. 2012 p 1026 (Table 3), p 1036
monument.design.horseshoe	Debated	The oval made a horseshoe, “highly likely” in the Roman period	Darvill et al. 2012 p 1037
monument.trilithons	Established	Graded towards the Great Trilithon (Petrie’s estimates); 56 is 6.55 m, the tallest	Petrie 1880 pp 11–12; Cleal et al. 1995 App. 5; Nash et al. 2020
monument.trilithons.joints	Established	Tenons, mortises, tongue and groove; 67, 69, 70, 72’s tenons dressed off	Gowland 1902 pp 77–78; Parker Pearson et al. 2020 pp 175, 333
monument.stoneHoles	Established	The holes, Gowland’s 2.1–2.5 m for 56; two sarsen blocks under 56 and the mauls under 55; antler picks dating the raising	Gowland 1902 pp 53–54, 71–72, 80–82; Hawley 1921 p 36; Darvill et al. 2012
monument.stoneHoles.raising	Debated	Raising is reconstruction; the “ramp” by 56 is a later pit	Gowland 1902 pp 81–82; Darvill et al. 2012 p 1025
monument.sarsen.core	Established	The Stone 58 core: 58 more badly cracked than expected in 1958, the core kept by Robert Phillips, back in 2018 (not 2019)	Nash et al. 2021; Nash et al. 2020
monument.sarsen.hauling	Debated	Sledges, 1,500 haulers or half that, the route open	Parker Pearson et al. 2020 pp 354, 357; Nash et al. 2020
monument.bluestone.name	Established	“Bluestone” is a label for 15 geological groups	Darvill et al. 2012 p 1022; Clarke et al. 2024; Parker Pearson et al. 2020
monument.bluestone.survivors	Established	Some 43 survive: 29 in the circle, 14 in the horseshoe	Parker Pearson et al. 2020 pp 163, 174–175; Darvill et al. 2012 p 1037
monument.bluestone.original	Debated	About 80 once; more than two-thirds gone	Parker Pearson et al. 2020; Darvill et al. 2012; Darvill & Wainwright 2009
monument.waunMawn	Debated	The 2021 proposal, its critics, and the proposers’ own 2022 revision: c. 3400 BC, “few, if any” stones reached Stonehenge	Parker Pearson et al. 2021, 2022b; Darvill 2022b; John 2024a
monument.altarStone.ruin	Established	Prone and broken under 55b and lintel 156	Clarke et al. 2024 p 570; Darvill et al. 2012 p 1033; Petrie 1880 p 12
monument.altarStone.setting	Debated	Upright or flat? Not known; the app’s flat slab is a choice	Darvill et al. 2012; Ruggles 1997; Clarke et al. 2024; Parker Pearson et al. 2025; Bevins et al. 2023a

Held back, by the owner’s defaults: the carvings (left to #55 and #61), the curved lintels (RR 32/2012, unread), Bevins et al. 2022b’s Cerrig Lladron result (unread), and Magli & Belmonte’s “accurate in space, not in time” (read only in the arXiv copy, print page unset). Lore bodies are capped at 600 characters; stone 36’s note is waived by name.

What is recorded as cut into each stone — wave 3, data only (#3, #49, #51, #55)

StoneRecords (HengeGeometry) holds, per stone, the joints and carvings the sources record, each with a tier and a page. Nothing in it is drawn yet: no file that builds, bakes or renders a stone reads it (StoneRecordTests scans them), the stone meshes hash identically before and after, and ApprovedLookTests passes unedited. Stone 56’s tenon, which is drawn, stays in the stone surface table with its survey height.

Stone	Feature	Tier	Rests on
9b	one tenon, on the end of the lying piece	Debated (Petrie prints it in 9’s row and names 9b; reading it as 9b’s is ours)	Petrie 1880 p 10
12	two tenons, on its end (lying)	Established	Petrie 1880 p 10
15	two tenons, on its end (lying)	Established	Petrie 1880 p 10
67, 69, 70, 72	tenons dressed off by later pick dressing	Established	Parker Pearson et al. 2020 p 175
68	worked groove down its north side	Established	Petrie 1880 p 11 (the side); Parker Pearson et al. 2020 p 175
36	a pair of mortises (a reused bluestone lintel)	Established	Engleheart & Newall 1930 pp 152–153; Parker Pearson et al. 2020 p 175; Darvill et al. 2012 p 1030
150	a pair of mortises (the other bluestone lintel)	Established	Parker Pearson et al. 2020 p 175; Darvill et al. 2012 p 1030 (Petrie 1880 p 12 calls it an impost, a lintel, and does not mention mortises)
53	12 axe-heads (found 1953)	Established	Crawford 1954, pp 25–31 (the article; page unread)
4	25 axe-heads (found 1953)	Established	Crawford 1954, pp 25–31 (the article; page unread)
3	3 axe-heads (found 1953)	Established	Crawford 1954, pp 25–31 (the article; page unread)
53	4 new carvings, identified by Leong et al. 2025a; motif and date of finding not given in the register	Established	Leong et al. 2025a, pp 112–121 (the article; page unread)

Gowland’s rule (1902, p 78), held as data beside the list: a circle upright carries two tenons, a trilithon upright one. Petrie’s 12 and 15 and the drawn 56 keep it. Monument-wide, 115 axe-heads are known after the 2011–12 recording, 44 before it and 71 found (Leong et al. 2025a); the counts are as of each source and never summed across sources. Stone 66, the tongue to 68’s groove, has no row in the surveyed plan (Petrie: “Only a stump”), so it carries no record. English Heritage RR 32/2012’s per-stone tallies are not used (the licence rule); Crawford’s page within pp 25–31 and Leong’s within pp 112–121 are still to be read, so both are cited by the article’s range.

A stone’s mean section (#3, the unblocked half): StoneSection.crossSection(volume:height:) gives V/h, the mean horizontal section, with the circle and square of equal area. It is held to a cylinder, a triangular and a rectangular prism and a cone frustum, worked on paper. No stone is fed through it: the 2011 scan’s per-stone volumes stay out of the repository until Historic England grants permission.

Medieval reception — Established as historiography

From the Nature news piece: Henry of Huntingdon (twelfth century) counted it among the wonders of England; legend gave the building to Merlin. The lore panel may quote this as historiography, never as history.

A popular infographic — cited, not reproduced

Source: “Infographics of the construction process of the megalithic monument of cromlech at Stonehenge (England)”, Alamy stock image 2NEC3YA (contributor-owned, commercially licensed; fetched as a PDF listing 2026-07-31, at the owner’s direction). The artwork itself is copyrighted stock material and is not vendored or reproduced anywhere in this repo — only its factual claims are recorded here, cited to their source, the way any secondary reference is treated.

What it shows, general-audience quality, English-as-a-second-language phrasing throughout — treat as Speculative/Debated unless a claim is independently corroborated:

• Etymology panel: “megalithic” from Greek megas (big) + lithos (stone); “cromlech” as stone + circular monument — Established, ordinary etymology.
• First-phase panel: circular ditch, two original accesses, 110 m diameter, livestock bone deposits in the ditch fill, only one causeway still open today — all corroborated by our own established figure (Cleal et al., see the earthwork page) and by the two causeways Earthwork.swift draws on Cleal Fig 36’s ditch gaps (near 43° and 174°, Debated: the bank’s gaps are drawn on the ditch’s; see the earthwork page). The app draws both open in every phase, since no source gives a phase in which the southern gap was closed.
• Third-phase panel: the central structure “surrounded by two rings made of holes, consisting of two horseshoes and two circles of megalithic [stones]” — matches our sarsen/bluestone circle-and- horseshoe model. Labels the Altar Stone and a stone it calls the “Stone of the sacrifice.”
• Labels “Hole Y” and “Hole Z” on a ring outside the sarsen circle — these correspond to the real Y and Z holes, a genuine excavated ring of pits from a much later phase (c. 1600 BC), and are not yet modelled in this app. A legitimate gap to add, separate from the two claims below.

Two claims from this source are flagged, not carried into the app:

• “166” as the stone count of the original construction. This number does not match any figure in our established sources (the sarsen count, bluestone count, or any published total we’ve found) and is not corroborated elsewhere. Recorded here as what the source says; not repeated as fact.
• “Stone of the sacrifice.” Almost certainly this source’s gloss on the Slaughter Stone — the fallen sarsen at the north-east entrance, a name coined by antiquarian observers (its surface iron staining was once read as blood) with no excavated evidence of sacrifice ever performed there. Scene.swift’s own comment already says as much: “nothing was slaughtered on it.” If this label is ever shown in-app, it needs the correction attached, not the euphemism alone.