henge

The earthwork and the ground

The enclosure ditch and bank — Established

Source: Cleal, Walker & Montague, Stonehenge in its Landscape: Twentieth-century excavations (English Heritage Archaeological Report 10, 1995) — the definitive excavation synthesis.

• The earliest monument (c. 2900 BC) is the circular earthwork enclosure: a ditch with an internal bank and a low external counterscarp bank. The main bank inside the ditch is the reverse of the classic henge arrangement — by the strict typology, Stonehenge is not a henge.
• Circuit roughly 110 m across (Darvill et al. 2012 p 1028); the ditch’s centre line a circle of radius 53.5 m and the bank’s medial line 48.3 m (Cleal pp 24–25).
• The ditch was dug as a chain of segments (p 63), and Fig 36 draws the 31 that were excavated (Hawley’s 1–28, the 1954 cuttings’ 98–100), all in the eastern and southern half: their edges, their ends and the chalk ridges between them. Its floor lay on average 1.68 m below the ground, 2.29 m at the western terminal of the north-eastern entrance and about 1.2 m in the south-west (p 67); the widest stretch Hawley gives is 4.27 m. The 16/17 ridge stands 0.9 m (p 65), the one ridge height given. Established; the ridge height used for every ridge Debated.
• The bank was 5.65 m wide in 1920 (p 94); Atkinson estimated it once stood 1.8 m (p 25), an estimate, so Debated.
• The ditch’s floor is drawn 1.68 m below today’s ground, Hawley’s average measured from the ground of the 1920s (p 67). That is where the floor is; the cut into the chalk is 1.2–1.3 m (p 67) because the topsoil above the chalk is part of that depth.
• The counterscarp: “only 13 inches (0.33m) high when sectioned in 1954” (p 24), the one height a source gives, so it is drawn at that as built (Debated, a floor on its first height); its crest where the LiDAR finds it, 59.0 m out.
• Today, on the 2022 1 m LiDAR (sector medians, research earthwork-digitisation-2026-10.md §5): the bank’s crest stands 0.36 m above the interior beside it and 1.17 m above the ditch’s floor where Hawley dug (his trenches “only partially backfilled”, p 25), 0.90 m where nobody did; the counterscarp 0.18 m.
• The ditch’s gaps are where Fig 36 ends it: the north-east entrance between segments 100 and 1, the southern between 17 and 18, and the 21/22 break Cleal p 68 reads as a blocked entrance. The bank and counterscarp open over the same two entrances, each as wide as the ditch’s gap (near 43° and near 174°), read off the reconciled plan: an inference from the ditch, so Debated (owner, 2026-10-14). They replace the app’s old gaps on the axis and at an unsourced 165°.
• The axis is not the entrance’s middle. The monument’s 49.9° axis crosses the tip of segment 1’s terminal (Fig 36 ends it 0.6° west of the axis, inside its own 1-sigma) and runs up the bank’s ramp at the edge of the north-eastern gap, whose middle is near 43°. The Avenue’s own line on the LiDAR runs about 2.5 m north-west of the axis, through the gap.
• Antler picks from the ditch bottom furnished the radiocarbon dates; the diggers backfilled placed deposits (cattle bone already old when buried) — the enclosure was never merely a quarry for the bank.

The North and South Barrows and the Heel Stone ditch — Debated

• Two small ditched rings inside the bank hold Station Stones 92 and 94 (Darvill p 1033: 94 “placed within the existing north barrow”; p 1034: 92 within the south barrow “by the later part of Stage 2”). Cleal p 276: the North Barrow’s V ditch c 0.85 m deep and 1.07 m wide with a low bank outside (now c 0.3 m); the South Barrow’s c 0.4 m deep, 10.8– 11.7 m across. Their rings are read off Figs 164 and 165; the North Barrow’s is inferred from its hachures and one excavated crossing, and its bank’s outer limit (9.75 m from its centre) is where the two readings overlap, so both are drawn Debated, from Stage 2.
• Where a barrow’s ring meets the enclosure bank, the ring dies into the bank’s inner slope and the crest runs unbroken. The North Barrow’s ring reaches r 49.15 m from the monument’s centre, past the bank’s 48.3 m crest; the South Barrow’s band runs at r 46.2–47.4 m on the inner slope. Simply summed, the rings would cut 0.85 m and 0.4 m slots through the as-dug 1.8 m bank, which no source draws: Darvill p 1034 has the South Barrow’s ditch “tight against the inside edge of the enclosure earthwork”, and the North Barrow’s arc across the bank is the half Fig 164 draws dashed and projected. So each ring is scaled by what is left of the bank, to nothing at the crest and beyond (Debated; the wave 4 review, research/earthwork-digitisation-2026-10.md §6).
• The Heel Stone’s ditch: c 10 m across, 1.1 m wide, 1.2 m deep with nearly perpendicular sides (Cleal pp 274–275), read off Fig 156. Dug “late in Stage 2 or perhaps early in Stage 3” (Darvill p 1034), and the Avenue was built “later in Stage 3, after the ditch around stone 96 had substantially silted up” (p 1035; Cleal p 268’s sequence “Stonehole 97 – Heelstone Ditch – Avenue bank”; Atkinson, as Cleal p 275 quotes him, had it refilled very soon after it was dug). So it is drawn open in no phase: listed in Stage 2 as notYetDug (its anchor may precede the digging; Fig 156 places it, so no row says locationUnknown), and listed as silted from Stage 3 on, when the Avenue exists, with no hollow drawn since no source gives one (Debated; the wave 4 review). Filled today: the LiDAR shows no ring. Its centre is Fig 156’s; its stone is a placeholder in Daw’s plan.
• The figures sit on their printed National Grid, which the LiDAR confirms; Daw’s placeholder Station Stones and Heel Stone stand 1.2– 1.8 m from the ring centres the figures draw round them, still inside each ring. How the two digitisations were reconciled, and where they disagree: research/earthwork-digitisation-2026-10.md.

In the app: HengeGeometry/Earthwork.swift models the circuit as an analytic profile on the erosion axis (as dug against today’s relief), the ditch’s segments from EarthworkPlan (the reconciled plans in Resources/earthwork/) and EnclosureDitch, the barrows and the Heel Stone ditch as RingFeatures on their own fine patches. The base terrain is not drawn inside the earthwork’s footprint, so a ditch below the coarse base grid shows. Tested in EarthworkTests (every segment where Fig 36 draws it, Hawley’s figures bounding it, one regular ditch where nobody dug, the floor rising over each ridge, each barrow containing its Station Stone, the bank’s crest unbroken past the barrows) and EarthworkLidarTests (today’s crest and floor against the 2022 LiDAR transects, within the EA’s ±0.15 m).

The ground surface — Established observation, Speculative palette

Reference imagery: aerial photograph of the monument from the north-west (supplied by the owner, 2026-07-31, from Nature’s news coverage of the Altar Stone work). What it shows, and what the renderer takes from it:

• Chalk downland is not one green: a patchwork metres across — thin worn turf reading pale where the chalk comes through, ranker growth, wear scars. The renderer’s two-band mottle is calibrated to this.
• The earthwork ring reads paler than the field — thin sward over chalky bank material. The wear channel on the ring carries this.
• The enclosure interior reads differently from the surrounding pasture (mown versus grazed in the photograph — a modern management artefact, not reconstructed). Likewise today’s visitor path round the stones and the desire lines to it are not worn into the turf: every phase has one pre-modern turf, the Avenue’s corridor the only path in it (owner answer 4a), and that only from Stage 3, when the Avenue was built (Darvill p 1035; MonumentPhase.avenueBuilt).
• The earthwork casts shadows (answer 5a): at a low sun the as-dug 1.8 m bank shades the interior (at 1.8 m, 1.8/tan h: 22 m at 5°). The renderer’s bank shadow is held to the analytic solution (ShadowSolver.profileShadowEdge) within the same 0.28 m as the stones’. The earthwork’s meshes meet the base terrain level, the base carved under them, so their rims cast nothing of their own.
• Soil is banked low against each stone’s foot, along the stone and on the ground under it (#79 fixed banks that had been turned across their stones and floated on low ground). The bank’s shape is the app’s choice, not archaeology: Speculative.
• Every colour triple in the renderer is an artistic reading — Speculative by construction; no source publishes reflectance spectra for this turf.

The wider landscape — Established

• Terrain: Environment Agency LiDAR (1 m, bare earth) heightfield, baked to a 768×768 grid at 40 m spacing (±15.3 km) by scripts/bake_terrain.py — the skyline the sun actually rises over.
• The round barrows and field boundaries beyond the enclosure in the reference photograph are not modelled.