User:
redacted sentence: unpublished prose redacted/chapters/ for chapters 013 through 029 (files 013-.md to 029-.md; not all have notes). Each note is the text after the last "---" line, starting "Author's note.". Read ONLY those note sections (use sed/grep, not the whole chapters).
Policy: every fact in a note must be a real-world fact (real science, real history, real practice). Check every checkable claim: crop marks and aerial archaeology (taller/greener over ditches, short/yellow over walls), blackthorn redacted timings, sallow, fallow history (clover and turnips), cave breathing / chimney effect and seasonal reversal, predator sign (fox carries carcass, stoat neck bite, hawk plucking ring), snipe seasonality and scrapes, radiation fog and wind, tidal outfall sluices, hatches on geared screws, meander dynamics, gutter cuts in flood response, staddle stones vs rats, faggots' uses, chestnut paling rot resistance, fen drainage hierarchy, Bedford Level, Forty Foot and Twenty Foot drains, Old Bedford River 1637 and 21 miles straight, Roman roads and drainage/ditches, peat wastage (consolidation then oxidation; Holme Fen post 1851, ~4 m exposed), Russian/peat corer, spoil behaviour, bedding planes, sample handling, plot registers, transects, controls, volunteer wheat/fat hen/couch, chalk hydrology, clay-with-flints, dew ponds. Use WebSearch/WebFetch where you're not certain.
Return a concise list: for each claim that is WRONG, DUBIOUS, or OVERSTATED, give chapter number, the quoted phrase, what's wrong, and a corrected wording of the same length. Also list claims you confirmed only briefly (one line each, "confirmed"). Do not edit files. Your final message is raw data for the orchestrator, not for the user.
14 tool calls
🔧 Bash: cd redacted/chapters && for f in 0{13..29}-*.md; do echo "=== $f"; awk 'BEGIN{n=0} /^--… (Print text after last --- in chapters 13-29)
✅ 142 lines, 17256 chars
🔧 ToolSearch ({"query": "select:WebSearch", "max_results": 1})
✅ 1 lines, 52 chars
🔧 WebSearch: Holme Fen post 1851 iron post metres exposed peat shrinkage
🔧 WebSearch: redacted sentence: unpublished prose
🔧 WebSearch: Chiltern beechwoods clay-with-flints plateau beech oak soils
🔧 WebSearch: cave breathing chimney effect summer cold air flows out lower entrance winter reverses
🔧 WebSearch: peat subsidence after drainage consolidation initial then oxidation dominant long term
✅ 20 lines, 2468 chars
✅ 14 lines, 2663 chars
✅ 24 lines, 3041 chars
✅ 24 lines, 3265 chars
✅ 22 lines, 3550 chars
Findings (chapters 013–029 author's notes)
WRONG / DUBIOUS / OVERSTATED
- Ch 029 — "oak grows where there was beech" — DUBIOUS/reversed. Beech is the characteristic tree of the clay-with-flints plateaus (the Chiltern beechwoods sit on exactly that soil), with oak and hornbeam joining it; beech is not displaced by oak there. Same-length fix: "oak and hornbeam join the beech" (or "the woods thicken with oak").
- Ch 029 — "with dry valleys and no streams" — slightly OVERSTATED: chalk has winterbournes (seasonal streams) and spring-fed streams. Fix: "with dry valleys and few streams".
- Ch 020 — "Chestnut is used because it does not rot" — OVERSTATED. Cleft chestnut is rot-resistant (high tannin, ~15–20 years in the ground), not rot-proof. redacted sentence: unpublished prose
- redacted sentence: unpublished prose Fix: "gets more crooked with time, not less" (drops the absolute), or "and never straight".
- redacted sentence: unpublished prose "1851 iron post" is defensible as written; ~4 m exposed is correct. No change strictly needed; if wanted: "In 1851 an iron post was set into Holme Fen".
- redacted sentence: unpublished prose to rot" — confirmed in the long run (oxidation ~75% of cumulative subsidence by 5 years, >90% by ~18 years); consolidation dominates only the first year or so. Fine as written.
CONFIRMED (brief)
- Ch 013 crop marks: positive (taller/greener over ditches), negative (short/yellow over walls); cereals esp. barley best; Roman towns visible in drought summers — confirmed. Pasture masking crop marks — confirmed.
- Ch 013 blackthorn suckers, thorny, redacted in a few years; sallow (goat/grey willow, broad-leaved) on wet ground — confirmed (timings approximate but reasonable).
- Ch 013 fallow; clover and turnips ending fallow (Norfolk four-course) — confirmed.
- Ch 014 cave breathing: chimney effect, summer cold air out low entrance, winter reversed, reverses at equinoctial transitions, strong enough to affect flame — confirmed.
- Ch 014 fox carries carcass off; stoat bites back of neck; hawk plucks in a ring — confirmed (standard gamekeeper sign).
- Ch 014 scrapes; snipe seasonal movement (residents plus autumn/winter migrants) — confirmed.
- Ch 015 radiation fog: clear, calm, humid nights; forms first over water; drains into hollows; wind disperses — confirmed.
- Ch 015 tidal outfall sluice: only discharges when tide below inland level; spring tides/onshore wind lock it — confirmed. Pightle — real East Anglian word — confirmed.
- Ch 016 hatches/sluice doors on geared screws; straight cut = artificial; drainage law — confirmed.
- Ch 017 gutter cuts (emergency grips) — confirmed as practice. Staddle stones: damp gap and rat-defeating overhang — confirmed (traditional account). Flash — real term. Faggots: bread ovens, ditch/bank revetment, fascines under tracks — confirmed.
- Ch 018 tide mark — fine.
- Ch 020 cleft chestnut paling with twisted wire — confirmed (usually 2 or 3 wires).
redacted sentence: unpublished prose
- Ch 022 agger, camber, side ditches; Roman road surviving as bank — confirmed.
- Ch 023 subsidence fastest early, slows, never stops; Holme post ~4 m exposed; nine inches/year initially — confirmed. Gouge/Russian corer description (half-cylinder, T-handle, twist) — confirmed. Peat spoil drying/shrinking/cracking and faces slumping — confirmed.
- Ch 024 bedding plane definition and continuity — confirmed.
- Ch 025 sample boxing/labelling; long-term plot registers (Rothamsted practice) — confirmed.
- Ch 026 transect, controls, volunteer wheat/fat hen/couch as neglected-ground flora — confirmed.
- Ch 029 chalk drains freely, spring-line villages, clay-with-flints holding water on plateau tops, ponds in corners, dew ponds of puddled clay, flat-topped chalk hills = clay caps — confirmed.
Chapters 019, 027, 028 have no author's note (text after last "---" is story prose).