---
slug: the-machine-made-of-months
passed: 2026-10-14
by: the assay line (research/assay-line/cited-1014/), Codex fixer directed by claude-assaying-boyle-1014; source record by Codex, checked by Codex
checked: 125
confirmed: 95
wrong: 30
unverifiable: 0
observed: 0
stale: 0
---

This was a targeted correction pass against the 31 items in research/assay-line/cited-1014/the-machine-made-of-months/triage.json, with the layer files compared to 4df1d97700. Every triaged item's evidence was reread before editing, including the primary supplements, the full 2012 article, the 2021 proposal, NASA's Saros explanation and the authored code. The record also retains all 94 CONFIRMED claims from the checker's source rereading in N2-codex.md, with its evidence and qualifications; those carried claims are not presented as a new audit of all 162 original entries. The 13 MINOR items count in wrong and receive record-only findings. X10 is CONFIRMED because its rejection is supported; the remaining 17 WRONG items and 13 MINOR items are fixed within this layer. The repair worker confirmed the fix-checker's missed copies of X1 in other layers on 2026-10-14; X1 remains open for those copies, with a board note pending. Eight dated correction lines cover the WRONG groups and are mirrored in the wrapper; dateModified changes without an Updated line.

## Claims

- MINOR | X1: "from its own teeth" | research/antikythera/verify.mjs (baseline: git show 4df1d97700:research/antikythera/verify.mjs) | The verifier supplies MECH_YEAR and MODERN independently of the teeth. The hero now promises reconstructed mean ratios and names separate day conversions and comparisons. Other layers still repeat the universal derivation claim; their authored notes and preserved quotations remain open under the X1 finding.
- MINOR | X2: "nothing close would be built" | https://www.nature.com/articles/nature05357 | Nature 2006 compares the mechanism with known devices. The introduction now qualifies the comparison by what is known, without asserting what could never have been built.
- MINOR | X3: "four-year games" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | Freeth 2008 supplement, section 3.5, printed p. 20, places both two-year and four-year festivals in an Olympiad cycle. The page now names games of that cycle.
- MINOR | X4: "built entirely from those reconstructed tooth counts" | research/antikythera/verify.mjs (baseline: git show 4df1d97700:research/antikythera/verify.mjs) | The verifier takes a supplied 365.2425-day year and modern comparisons. The introduction now names those inputs alongside reconstructed tooth counts.
- MINOR | X5: "Nothing here is asserted; it is computed" | research/antikythera/verify.mjs (baseline: git show 4df1d97700:research/antikythera/verify.mjs) | The page now limits its computation claim to mean gear ratios and residuals, and identifies the browser as a separate illustration.
- WRONG | X6: "Kepler's law of the ellipse" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | Freeth 2006 supplement, PDF pp. 26-27, proves the pin-and-slot equivalent to Hipparchos' epicyclic lunar theory. Replaced the Kepler attribution in the caption and dek and added a mirrored Corrected line.
- MINOR | X7: "from the tooth counts alone" | research/antikythera/verify.mjs (baseline: git show 4df1d97700:research/antikythera/verify.mjs) | The ledger contains modern comparison values as well as ratios and day conversions. Its introduction now distinguishes those inputs from tooth-derived ratios.
- WRONG | X8: "254 orbits of the Moon take exactly 19 years" | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | Freeth and Jones 2012, paragraphs 100 and 104, distinguish an approximate astronomical relation from the exact 254/19 mean gear rate. The supplied periods differ by 0.094394 day over 254 months versus 19 model years. Qualified the body, dek, descriptions and wrapper; narrowed the unused OG template to the mean gear ratio and added a mirrored Corrected line.
- WRONG | X9: "they are spaced 6 months apart" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | Freeth 2008 supplement, section 4.4, printed p. 32, includes consecutive glyph months 25 and 26. The five/six-month spacing is now limited to the lunar scheme; the illustration explicitly uses synthetic positions. Included in the glyph-count Corrected line.
- CONFIRMED | X10: "it is forced arithmetic" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | REJECTED: Freeth 2008 supplement, section 4.3, printed p. 31, uses 6a + 5b = 223 with astronomical information to motivate the lunar scheme. Finding constants in the verifier does not refute that rationale. Retained the sentence "That is not decoration; it is forced arithmetic." The distinct scope and uniqueness errors are fixed under X9, X11 and X12.
- WRONG | X11: "spreading them as evenly as possible around the ring fixes the answer" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | Independent enumeration gives (3,41), (8,35), (13,29), (18,23), (23,17), (28,11), (33,5). The equation and even spacing alone do not select 33/5. The text now calls it the adopted lunar scheme, and the verifier enumerates and checks all seven solutions. Added a mirrored Corrected line.
- WRONG | X12: "38 eclipse-possibility glyphs, exactly as the bronze is engraved" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | Freeth 2008 supplement, Figure 20, printed p. 36, reconstructs 51 glyphs combining 38 lunar and 27 solar predictions. The replacement total is explicitly attributed to that reconstruction; the 38-marker animation is labelled a simplified lunar scheme. Added a mirrored Corrected line.
- MINOR | X13: "below the Saros" | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | Freeth and Jones 2012, paragraph 9, locates the subsidiary Exeligmos inside the Saros dial. Replaced "below the Saros" with "inside the Saros spiral".
- MINOR | X14: "the same place and hour" | https://eclipse.gsfc.nasa.gov/SEsaros/SEsaros.html | NASA describes an approximate eight-hour longitude shift and a return to about the same region after three Saros cycles. Qualified the day length, shift, longitude and time-of-day recurrence as approximate.
- WRONG | X15: "30 of the 31 gears named on this page have tooth counts read directly off the surviving fragments" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | Visually reread Freeth 2008 supplementary Figure 14, printed p. 22. m3, n1 and n3 are red hypothetical elements. The listed inventory is now 28 surviving gears plus those three hypothetical gears; adopted tooth counts can include estimates. Corrected the flags, public verifier copy and prose with a mirrored Corrected line. The separate total of 30 surviving gears is preserved.
- MINOR | X16: "no planetary gearing survives at all" | https://www.nature.com/articles/s41598-021-84310-w | Freeth 2021 proposes a Venus role for surviving Fragment D components. Replaced total absence with largely lost planetary trains, retaining the proposed identification as uncertain.
- MINOR | X17: "Every number on this page is computed" | research/antikythera/verify.mjs (baseline: git show 4df1d97700:research/antikythera/verify.mjs) | The footer now limits the verifier to mean gear ratios and residuals with supplied inputs, and calls the browser a separate illustration with a static check summary.
- WRONG | X18: "Saros · 223 months · 38 eclipses" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | The Saros canvas label now identifies a simplified lunar scheme. Nearby prose explains the 38 synthetic positions rather than calling them a count of eclipses. Included in the glyph-count Corrected line.
- WRONG | X19: "One turn of the crank is one year" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | Freeth 2006 supplement, PDF pp. 15 and 21, gives adopted a1=48, b1=223, and one b1 revolution per year. The heading now names the year wheel; prose and wrapper distinguish the interface from the side-crank ratio 48/223 year per turn. Added a mirrored Corrected line.
- MINOR | X20: "from its own tooth counts" | research/antikythera/verify.mjs (baseline: git show 4df1d97700:research/antikythera/verify.mjs) | The dek now promises reconstructed mean ratios and distinguishes separately supplied day scales and astronomical comparisons.
- WRONG | X21: "exactly 38 eclipse glyphs" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | The dek now attributes 51 glyphs, including 38 lunar predictions, to the 2008 reconstruction. Included in the glyph-count Corrected line.
- WRONG | X22: "30 attested ones" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | The dek now lists 28 surviving gears, whose adopted tooth counts include estimates, and three hypothetical gears. Included in the inventory Corrected line.
- MINOR | X23: "every figure printed by verify.mjs" | public/strata/the-machine-made-of-months/index.html (baseline: git show 4df1d97700:public/strata/the-machine-made-of-months/index.html) | The dek now limits verifier coverage to mean ratios and residuals. The separately implemented schematic readout is qualified under X31.
- WRONG | X24: "the same code that checks it offline" | public/strata/the-machine-made-of-months/index.html (baseline: git show 4df1d97700:public/strata/the-machine-made-of-months/index.html) | The inline script neither imports nor executes verify.mjs. The wrapper, footer and check summary now identify a separate browser illustration with mirrored values and a static summary. Added a mirrored Corrected line.
- WRONG | X25: "faithfully modelled from the real pin-and-slot geometry" | public/strata/the-machine-made-of-months/index.html (baseline: git show 4df1d97700:public/strata/the-machine-made-of-months/index.html) | The old inset used centre separation 1.088 wheel radii and pin radius 0.7, so its output could not complete a rotation; its old timing had a 27.79058-day period. The repaired inset shares e/rho=0.34 with the front illustration, has centre separation less than pin radius, and uses mean lunar rate minus apsidal rate. The caption and wrapper visibly identify the retained schematic exaggeration. Included in the model Corrected line.
- WRONG | X26: "exactly **38** eclipse glyphs" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | The wrapper now attributes 51 combined glyphs and 38 lunar predictions to the 2008 reconstruction, and describes the animation as a simplified lunar scheme. Included in the glyph-count Corrected line.
- WRONG | X27: "30 attested tooth counts" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | The wrapper now identifies 28 surviving gears and three hypothetical gears, with estimated tooth counts distinguished from direct counting. Included in the inventory Corrected line.
- MINOR | X28: "entirely-missing planetary gearing" | https://www.nature.com/articles/s41598-021-84310-w | The wrapper now says the planetary trains are largely lost and records the proposed, uncertain Venus role of Fragment D.
- WRONG | X29: "reconstructed gears are marked apart from attested ones" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | m3 and n1 now have a:false alongside n3 in the verifier and its public copy. The wrapper and honesty box accurately describe all three hypothetical gears. Included in the inventory Corrected line.
- WRONG | X30: "Saros dial's 38 eclipse glyphs" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | The meta description and JSON-LD now attribute 51 glyphs and 38 lunar predictions to the 2008 reconstruction. The OG description also narrows derivation coverage. Included in the glyph-count Corrected line; dateModified is 2026-10-14.
- WRONG | X31: (added) "anomaly: +19.8°" | public/strata/the-machine-made-of-months/index.html (baseline: git show 4df1d97700:public/strata/the-machine-made-of-months/index.html) | The old formula gives 19.783381 degrees at day 5 from the exaggerated illustration. The degree readout now explicitly says "schematic anomaly (exaggerated)" and refers to the illustrated pointer. Both displays share the corrected timing and geometry. Included in the model Corrected line.
- CONFIRMED | "A bronze computer, hand-cranked, lost in a shipwreck around 60 BC and pulled from the seabed in 1901." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/; https://antikythera.org.gr/ | hero-date-crank: The record's excerpt dates the sinking "within a few years after 70 BC". The excavation project independently says "circa 60 B.C." (https://antikythera.org.gr/). ISAW confirms salvage in 1900-1901; the inferred hand input is described in paragraph 9. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Every pointer it carries is run by a train of gears, and a train of gears is just a product of fractions." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | hero-gears-fractions: The supplement gives the fixed-axis rule "at a rate of –p/q times the first"; signed ratios multiply through a train. Read as mean period ratios, with the lunar anomaly separately described on the page. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "In 1901 sponge divers working a wreck off the islet of Antikythera brought up bronze statues, glassware " [...] " and a corroded lump the size of a thick book." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | recovery-cargo: ISAW paragraph 0 confirms "bronze and marble sculpture and glassware" and salvage "in 1900-1901". The thick-book comparison is approximate descriptive scale, not an asserted measurement. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "It split, drying, into fragments shot through with the green ghosts of gearwheels:" | https://en.wikipedia.org/wiki/Antikythera_mechanism | drying-fragmentation: The retrieved Discovery section repeats "cracked and shrank when it was brought up". ISAW paragraphs 4-5 independently describe an initially much less fragmented object and the corrosion products. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "thirty of them survive," | https://www.nature.com/articles/s41598-021-84310-w | surviving-thirty: 2021 abstract: "including 30 corroded bronze gearwheels". This is the total surviving inventory, distinct from the faulty 30+1 list on this page. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the largest with 223 teeth," | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | largest-223: 2006 table: b1 has best-fit count "223", with limits "223/224". It also has the largest listed outer radius. Confirmed as the adopted 2006 reconstruction, not an exact count of wholly preserved teeth. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "cut by hand into triangular points about a millimetre and a half apart." | https://en.wikipedia.org/wiki/Antikythera_mechanism | tooth-shape-pitch: The cited Gearing section says "equilateral triangles" and "average circular pitch of 1.6 mm" and describes hand tools as probable; about 1.5 mm is fair rounding. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the next geared astronomical clocks appear in medieval Europe and the Islamic world." | https://en.wikipedia.org/wiki/Antikythera_mechanism | medieval-clocks: The full Similar devices section describes medieval Byzantine and Islamic transmission, the Byzantine geared calendar, and later astronomical clocks. The millennium sentence is separately qualified under unique-millennium. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "It is a calculator for the sky: turn a crank on the side, and pointers on the front and back faces show the Sun's place in the zodiac, the Moon's place and phase, the cycle that keeps a lunar calendar in step with the year, and a spiral that predicts eclipses." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | sky-displays: ISAW paragraph 9 lists the input, zodiac/calendar dials, "235 lunar months", and "223-month eclipse prediction dial". Paragraph 59 describes the phase ball; paragraph 104 gives the Moon display. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "We can say how it works to the tooth because of X-ray CT scans made in 2005, which read the buried gears Freeth and colleagues published in Nature." | https://www.nature.com/articles/s41598-021-84310-w | ct-2005: The 2021 abstract explicitly dates the CT campaign to "2005" and says it "decoded the structure of the rear". The page separately acknowledges the uncertain front. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The great four-spoked wheel, b1 " [...] " 223 teeth " [...] " turns once a year and carries the Sun pointer around a ring of the twelve zodiac signs." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | b1-sun: 2006 supplement describes "large four-spoked gear wheel b1" and says "One revolution of b1 represents a year". Its Sun indication is the mean-Sun/date reconstruction, not a demonstrated exact true-Sun mechanism. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Off that single shaft hangs every other train." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | b1-shaft: 2006 supplement says "b2 is fixed to b1" and describes the branches driven from b2. This is the reconstructed surviving lunisolar train. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "it laps the zodiac about thirteen times a year, and the little ball " [...] " half bright, half dark " [...] " turns to show the phase, the angle between Moon and Sun." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | moon-laps-phase: ISAW describes a "revolving half-black, half-white ball"; paragraph 104 supplies "254/19 rotations per year", or 13.3684. Differential phase is a distinct mean period; the browser's exaggerated readout is addressed separately. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Two equal 50-tooth wheels, k1 and k2, mounted face-to-face on different centres; a pin on one rides a slot in the other." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | two-50-gears: ISAW paragraph 104: "Gears e5, e6, k1, k2 all have 50 teeth" and "a pin on its face that engages with a slot on k2" on eccentric axes. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The Moon pointer's mean rate is exactly 254 / 19 turns per year " [...] " and its speed is not steady: the pin-and-slot makes it surge and lag, the apse line itself creeping round once every 8.88 years." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | moon-rate-anomaly-apse: ISAW gives "254/19 rotations per year" and the apsidal rate (9*53)/(19*223). Its reciprocal is 4237/477 = 8.88259958 years. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Mesh a wheel of D teeth into one of E teeth and the second turns D/E times for each turn of the first." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | gear-ratio: 2006 supplement gives "at a rate of –p/q times the first". The page gives the magnitude; reversal of direction does not change this number. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Chain them and you multiply the fractions." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | gear-chain: Applying the quoted p/q rule at consecutive meshes multiplies the rate ratios. Coaxial gears share angular velocity. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "That is the whole secret: each pointer's rate is a single product of small whole numbers, chosen so the product lands on an astronomical ratio." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | each-pointer-product: Confirmed for the mean pointer periods shown in the adjoining ledger. The variable lunar correction also needs pin-and-slot geometry; see the universal-derivation findings. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The Moon train is (38/64)·(24/48)·(32/127) " [...] " and that equals 19/254 on the nose." | git show 80213b8a47:research/antikythera/verify.mjs | moon-train-exact: Independent exact arithmetic: (38/64)*(24/48)*(32/127)=19/254. Recorded in evidence/independent-arithmetic.json and reproduced by both verifier copies. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The bronze isn't close to a lunar month. It is one, to fourteen parts in a million." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | moon-fourteen-ppm: (19/254)*365.2425=27.32128937 days. Relative to 27.321661 this is -13.602 ppm, rounding to -14 ppm. The selected ISAW excerpt gives the cycle but not this residual; the code and independent arithmetic supply it. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The headline gear is d2, with 127 teeth " [...] " a prime, and an awkward thing to cut by hand." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | d2-prime: The CT table marks d2 "127" and "Definite". Trial division by 2, 3, 5, 7, 11 confirms primality; awkward is descriptive rather than a measured claim. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The maker needed the number 254 in the train, and put half of it into a single wheel." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | maker-254: The source supplies the 254-sidereal-month Metonic relation and the tooth table supplies d2=127. 127*2=254 is independently exact. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The back face carries two spiral dials." | https://www.nature.com/articles/nature07130 | back-dials: 2008 abstract explicitly describes the two major spiral back dials, upper Metonic and lower Saros. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The upper one is the Metonic calendar: a five-turn spiral of 235 cells, one per lunar month, that the pointer crawls around once every 19 years " [...] " the cycle that lets a Moon-calendar keep step with the Sun." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | metonic-235: ISAW gives "235 lunar months" and "19-year Metonic calendar". The 2008 abstract specifies "five-turn spiral". Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The lower one is the Saros, 223 months of eclipse prediction." | https://www.nature.com/articles/nature07130 | saros-223: 2008 abstract calls it a "four-turn spiral of 223 lunar months" and an eclipse-prediction dial. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The gearing for it is a clean ×12 " [...] " exactly three turns of the four-turn Saros spiral." | git show 80213b8a47:research/antikythera/verify.mjs | exeligmos-times12: Independent gear stage: (60/20)*(60/15)=12; a complete Saros takes four pointer rotations, so twelve rotations equal three Saros periods. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Gears can only have whole teeth." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | integer-teeth: Counts in the source table are integers; a toothed wheel used for a repeating gear ratio must have an integer count of teeth. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "You use the best whole-number approximations there are: the continued-fraction convergents of that ratio." | https://crypto.stanford.edu/pbc/notes/contfrac/converge.html | cf-best: Stanford theorem: if a closer rational approximation exists, "q' > q". This supports best approximation at bounded denominator, not an exclusive history of how the ancient makers designed it. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Expand 12.368… as a continued fraction and read off its convergents " [...] " each one the best fraction of its size:" | https://crypto.stanford.edu/pbc/notes/contfrac/converge.html | cf-each-best: The same continued-fraction theorem supports the stated denominator-bounded sense of size. Checked against the full theorem and its qualifications. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Sixth in the list sits 235 / 19" | git show 80213b8a47:public/strata/the-machine-made-of-months/index.html | cf-sixth: Recomputed the continued fraction of 365.242190/29.530589: 12/1,25/2,37/3,99/8,136/11,235/19,4131/334. The sixth entry is 235/19. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the Metonic cycle, named for Meton of Athens in 432 BC but known to Babylon before him." | https://en.wikipedia.org/wiki/Metonic_cycle | meton-432: The full Metonic article says "Meton introduced the 19 year cycle to the Attic calendar in 432 BC" and "The Babylonians applied the 19-year cycle from the late sixth century BC." The chosen excerpt alone omits the earlier Babylonian evidence. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "It is not an accident the machine uses 19 years: 19 is where the approximation gets suddenly, almost suspiciously good, and the makers cut that convergent straight into the tooth counts." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | makers-convergent: Independent numerical result: 235/19 is the sixth convergent. ISAW paragraph 234 calls the Metonic accuracy exceptional. This describes the ratio cut into the model, not evidence that ancient makers used the modern continued-fraction algorithm. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | n3=57 is hypothetical and gives the Olympiad dial a quarter-turn per year with axle spacing consistent with the estimated pitch radii. | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature07130/MediaObjects/41586_2008_BFnature07130_MOESM268_ESM.pdf | n3-reconstruction: 2008 supplement section 3.6 derives "57 teeth" and gives a 24.4 mm interaxial distance equal to the sum of estimated pitch radii. n3 is explicitly hypothetical. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The inscriptions describe the motions of all five planets known then, so there were probably planetary pointers on the front " [...] "" | https://www.nature.com/articles/s41598-021-84310-w | planet-inscriptions: 2021 abstract explicitly states "Sun, Moon and all five planets known in antiquity" and the front display. The page fairly uses probably for the reconstruction. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Reconstructions of it (including a 2021 model from the same group) are proposals, not readings," | https://www.nature.com/articles/s41598-021-84310-w | planet-proposals: 2021 conclusion says "we cannot claim that our model is a replica of the original". The page's proposal qualification is appropriate. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "One real surviving gear, r1 with 63 teeth, is still unaccounted for " [...] " it fits no train anyone has agreed on." | https://www.nature.com/articles/s41598-021-84310-w | r1-unaccounted: ISAW 2012 calls the function of the 63-tooth gear unknown. The 2021 paper offers a Venus interpretation but explicitly calls its model a reconstruction. The page's qualified lack of an agreed train is not refuted by a proposal; do not confuse this with no surviving candidate component. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Its cycles are Babylonian period-relations, so its errors are theirs:" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | babylonian-errors: 2006 supplement says the trains are "based almost completely" on the Babylonian Metonic and Saros relations. Confirmed for ideal period residuals, not every possible manufacturing error. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "modern 8.85" | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | modern-apse: 2006 supplement gives the modern value "8.8504 years"; 8.85 is appropriate rounding. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "And in 2025 one team argued that the hand-cut triangular teeth carried so much error that the mechanism may never have turned smoothly enough to be usable " [...] " while themselves cautioning that the scans may overstate the imperfections." | https://arxiv.org/pdf/2504.00327 | 2025-usability: The 2025 PDF conclusion supplies both possibilities and explicitly warns that corrosion, resolution, and incomplete teeth may exaggerate original errors. The page preserves that caveat. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "That argument is live;" | https://arxiv.org/pdf/2504.00327 | debate-live: The named 2025 argument is presented in the paper as conditional, "our results must be interpreted with caution". Confirmed as an unresolved claim as the page presents it, not proof that the machine failed. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "What is not in doubt is the design: the ratios are exact, whoever cut them and however well they ran." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | design-exact: Integer gear counts determine exact rational ideal ratios, independently of manufacturing error. This does not validate every proposed tooth count; the attestation overclaim is recorded separately. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Someone in the Greek world looked at the wandering Moon, the drifting calendar, the returning eclipse, and decided every one of them was a fraction " [...] " then cut those fractions into metal so a stranger turning a crank could read the heavens off a dial." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | greek-maker-design: The 2006 supplement explicitly bases the trains on Babylonian astronomical relations, and the 2012 paper places the instrument in Greek mathematical astronomy. The maker's particular thought process is rhetorical, not a recovered quotation. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Two thousand years later the metal is green and broken, the maker nameless," | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | age-corrosion: ISAW describes heavily calcified and corroded fragments after "nearly two thousand years under water". The named research does not identify a known individual maker; 2000 is rounded. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "19 over 254. 235 over 19. Six thirty-threes and five fives make 223." | git show 80213b8a47:research/antikythera/verify.mjs | closing-arithmetic: Independent arithmetic confirms 6*33+5*5=223 and the two listed fractions. This sentence itself does not assert the erroneous total glyph count. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Tooth counts: Freeth et al., Nature 444:587 (2006) & 454:614 (2008)." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | footer-tooth-citations: Both cited articles exist with matching first author, journal, volume, and first page; the 2008 supplement adds the n3 reconstruction. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Freeth et al. 2006 " [...] " "Decoding the ancient Greek astronomical calculator known as the Antikythera Mechanism," Nature 444:587" | https://www.nature.com/articles/nature05357 | citation-nature2006: Nature and Crossref agree on Freeth et al., title, Nature 444, pages 587-591, 2006, DOI 10.1038/nature05357. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Freeth, Jones, Steele & Bitsakis 2008 " [...] " "Calendars with Olympiad display and eclipse prediction," Nature 454:614 (Olympiad dial, n3 = 57)" | https://www.nature.com/articles/nature07130 | citation-nature2008: Nature and Crossref agree on Tony Freeth, Alexander Jones, John M. Steele, Yanis Bitsakis; Nature 454, 614-617, 2008. The tool's initial-M warning incorrectly treats the middle initial in "John M. Steele" as the first initial. The title is recognizably abbreviated. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Freeth & Jones 2012 " [...] " "The Cosmos in the Antikythera Mechanism," ISAW Papers 4" | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | citation-isaw2012: The retrieved header reads Tony Freeth and Alexander Jones, "The Cosmos in the Antikythera Mechanism", ISAW Papers 4 (February, 2012). Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Freeth et al. 2021 " [...] " "A Model of the Cosmos…," Scientific Reports 11:5821 (the proposed planetary gearing, used here only as a named uncertainty)" | https://www.nature.com/articles/s41598-021-84310-w | citation-nature2021: Nature and Crossref agree on Freeth et al., the expanded Cosmos title, Scientific Reports 11, article 5821, 2021. The page signals an abbreviated title with an ellipsis. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Szigety & Arenas 2025 " [...] " "The Impact of Triangular-Toothed Gears on the Functionality of the Antikythera Mechanism," arXiv:2504.00327" | https://arxiv.org/abs/2504.00327 | citation-arxiv-abs: arXiv abstract and PDF agree on title, Esteban Guillermo Szigety and Gustavo Francisco Arenas, submitted 1 Apr 2025, identifier 2504.00327. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Wikipedia " [...] " Antikythera mechanism (gear scheme and pointer formulae)" | https://en.wikipedia.org/wiki/Antikythera_mechanism | citation-wiki: The fetched Wikipedia page includes both the gear-train table and pointer formulas. Its selected excerpt on tooth shapes alone is insufficient for those formulas, but the full source supplies them. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The founding layer's subject is incommensurability " [...] " that no whole number of one length ever equals a whole number of another, the fact that breaks the diagonal of the square." | git show 80213b8a47:src/content/strata/incommensurable.md | nearby-founding: The pinned founding layer explicitly says "First stratum" and provides the parity proof of irrationality of sqrt(2). Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Two real mechanisms taken apart until every motion is re-derived from the parts that cause it " [...] " there the Antikythera's calendar read straight out of its tooth-counts,…" | git show 80213b8a47:research/antikythera/verify.mjs | nearby-rederived: As a description of the linked mechanisms and the calendar ratios, supported by the linked pinned layer and reproduced fractions. It does not supply a proof that physical cycle lengths are transcendental. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Sun · date pointer 1 (the input) 1 year b1 turns once / yr" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-1: b1 is the annual mean-Sun/date axis in the 2006 supplement. The table's annual input is b1, not the external a1 crank (see crank-year-heading). Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Moon · mean sidereal (38/64)·(24/48)·(32/127) = 19/254 yr 27.32129 d modern 27.321661 d · −14 ppm" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-2: (38/64)*(24/48)*(32/127)=19/254; times 365.2425 is 27.32128937 days, or -13.602 ppm relative to 27.321661. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Moon phase · synodic Moon − Sun differential = 235/19 / yr 29.5302 d modern 29.530589 d" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-3: 254/19-1=235/19 per year; 365.2425*19/235=29.53024468 days. NASA modern value: 29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Metonic pointer (38/64)·(96/53)·(53/15) = 19/5 yr/turn 5 turns = 19 yr 235 months = 6939.69 d" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-4: (38/64)*(96/53)*(53/15)=19/5 years per turn. Five turns equal 19 years; modern 235 months give 6939.688415 days. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Olympiad / Games dial (5/19)·(57/60) = 1/4 /yr 4 years exactly n3 reconstructed" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-5: (5/19)*(57/60)=1/4 per year. The proposed n3=57 is correctly identified as reconstructed. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Lunar apse precession (38/64)·(96/53)·(223/27) = 4237/477 yr 8.8826 yr modern 8.85 yr · +0.37%" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-6: (38/64)*(96/53)*(223/27)=4237/477=8.88259958 years; relative to 8.85 this is +0.36836%, rounded +0.37%. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Saros eclipse dial (apse)·(53/188)·(54/30) 4 turns 6585.24 d true 223 mo = 6585.32 d" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-7: 4237/477*(53/188)*(54/30)*4*365.2425=6585.244565 days, rounded 6585.24; modern 223 synodic months round 6585.32. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Exeligmos dial (Saros)·(60/20)·(60/15) = ×12 3 Saros 669 months = 19756 d" | git show 80213b8a47:research/antikythera/verify.mjs | ledger-8: (60/20)*(60/15)=12, equivalent to three four-turn Saros cycles; 669 modern months round to 19756 days. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "12 / 1 months : years 12.000 convergent 1" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-1: Recreated continued-fraction convergent 1: 12/1. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "25 / 2 months : years 12.500 convergent 2" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-2: Recreated continued-fraction convergent 2: 25/2. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "37 / 3 months : years 12.333 convergent 3" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-3: Recreated continued-fraction convergent 3: 37/3. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "99 / 8 months : years 12.375 convergent 4" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-4: Recreated continued-fraction convergent 4: 99/8. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "136 / 11 months : years 12.364 convergent 5" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-5: Recreated continued-fraction convergent 5: 136/11. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "235 / 19 months : years 12.368 ← the Metonic cycle" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-6: Recreated continued-fraction convergent 6: 235/19. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "4131 / 334 months : years 12.36826 convergent 7" | git show 80213b8a47:research/antikythera/verify.mjs | convergent-7: Recreated continued-fraction convergent 7: 4131/334. The displayed decimal is correctly rounded from this fraction; input is 365.242190/29.530589. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Metonic · 235 months / 19 yr" | git show 80213b8a47:public/strata/the-machine-made-of-months/index.html | canvas-label-1: ISAW paragraph 9 supplies the 19-year, 235-month Metonic calendar; the label matches the source. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The Antikythera mechanism is a bronze, hand-cranked astronomical computer from the second century BC, pulled from a shipwreck in 1901 and read to the tooth by X-ray CT." | https://www.nature.com/articles/nature05357 | dek-origin: 2006 abstract dates construction "around the end of the second century bc" and discovery to 1901; ISAW dates CT data to 2005. Crank is the standard inferred hand input. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | The animation includes Sun, Moon and phase, Metonic and Saros displays. | git show 80213b8a47:public/strata/the-machine-made-of-months/index.html | dek-animation: Confirmed only the presence of these display elements. The Kepler attribution, exaggerated amplitude and timing are corrected under X6, X25 and X31. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | 127 is half of 254; the adopted mean lunar gear rate is 254/19 turns per year. | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | dek-127: The source says "almost exactly" for physical 254 months/19 years. This dek uses the conventional period relation without the body's explicit exactly; confirmed as the design relation, with the exact-sky statement separately corrected. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the Metonic ratio 235/19 is, precisely, a continued-fraction convergent of the year: incommensurability solved in bronze;" | git show 80213b8a47:research/antikythera/verify.mjs | dek-metonic-cf: Independent continued fractions confirm 235/19; the incommensurability wording here is an analogy. The literal physical irrationality claim is separately unverifiable. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the live 2025 dispute over whether the hand-cut teeth could even turn is named." | https://arxiv.org/pdf/2504.00327 | dek-dispute: The 2025 paper discusses possible failure of gears under the studied error assumptions. Its caution is also preserved on the page. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The founding layer's subject is incommensurability " [...] " that no whole number of one length ever equals a whole number of another, the fact that breaks the diagonal of the square." | git show 80213b8a47:src/content/strata/incommensurable.md | relation-incommensurable: The pinned incommensurable layer is labeled "First stratum" and proves the square diagonal/side ratio irrational. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "a year is not a whole number of months, yet gears can only have whole teeth." | https://en.wikipedia.org/wiki/Metonic_cycle | relation-year-integers: The Metonic source gives a tropical year longer than twelve lunar months and shorter than thirteen; the gear count is integral. No transcendence assertion is required for this statement. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the best whole-number approximations are the continued-fraction convergents " [...] "" | https://crypto.stanford.edu/pbc/notes/contfrac/converge.html | relation-best: The Stanford theorem states the denominator-bounded best-approximation property for convergents. It does not establish how the historical designers discovered the ratio. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "235 months / 19 years, the Metonic cycle the machine runs on, is literally a convergent of the year-to-month ratio." | git show 80213b8a47:research/antikythera/verify.mjs | relation-metonic: Recomputed as the sixth convergent of the declared tropical-year/synodic-month ratio; the mechanism's Metonic period is 19 years. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "There, Linear A's language is lost but its arithmetic survives, so a stranger can audit a Minoan scribe's sums and even catch his off-by-one." | git show 80213b8a47:src/content/strata/the-accountant-we-cant-read.md | relation-linear-a: The cited pinned linked layer explicitly records HT 122's "97 for a 96 sub-total", and HT 13's readable arithmetic. Confirmed as an accurate description of that linked layer's contents; not a fresh full audit of the separate Linear A layer. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Here the maker is nameless, the mechanism corroded green, but the gear ratios survive intact," | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | relation-corrosion: ISAW paragraphs 4-5 document corrosion and surviving gearing; paragraph 100 derives period ratios. Read intact as recoverable model ratios, not a claim every original gear is present. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "so a stranger turning a crank can audit the heavens " [...] " and catch the machine's own off-by-thirds in the Saros." | git show 80213b8a47:research/antikythera/verify.mjs | relation-audit: The supplied NASA source explains the approximately eight-hour Saros shift and the Exeligmos correction. The claim is about inspecting that arithmetic, not an exact local repeat of an eclipse. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "There, grouped data is forced to confess what its grouping discards." | git show 80213b8a47:src/content/strata/every-number-honest.md | relation-grouping: The pinned Every Number Honest wrapper explicitly describes reclassification, aggregation, and selection, including opposite answers from the same table depending on grouping. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The mechanism is exactly as accurate as the Babylonian period-relations it encodes, no more, and the honesty is in saying so." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | relation-babylon: Supported for the exact idealized period relations used here, not real mechanical errors. The cited supplement says "based almost completely" and excludes lunar-anomaly amplitude. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the bronze geared astronomical computer recovered from a first-century-BC shipwreck in 1901 and read gear-by-gear by the 2005 X-ray CT campaign (Freeth et al., *Nature* 2006/2008)." | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | wrapper-recovery: ISAW paragraph 0 confirms first-century-BC sinking and 1900-1901 recovery. Paragraph 6 and the 2021 abstract independently date the CT campaign to 2005. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | The ledger expresses reconstructed mean pointer rates as products of tooth-count ratios. | git show 80213b8a47:research/antikythera/verify.mjs | wrapper-ledger: The displayed ledger products and the design relation 254 sidereal months per 19 years match the exact rational calculations. Physical exactness is separately corrected at claim 30. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "the live 2025 claim (Szigety & Arenas, arXiv:2504.00327) that the hand-cut triangular teeth may have been too imprecise for the machine to run, named and left unresolved." | https://arxiv.org/pdf/2504.00327 | wrapper-uncertainty: The arXiv PDF states the conditional failure argument and acknowledges that inferred original errors may be overestimated; the wrapper accurately characterizes it. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Tooth counts: Freeth et al., *Nature* 444:587 (2006) and 454:614 (2008); the planetary proposals (Freeth et al., *Sci. Rep.* 2021) are cited only as a named uncertainty." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | wrapper-citations: Bibliographic records confirm both Nature articles and the 2021 Scientific Reports proposal. The paper is indeed treated as uncertain on the page. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | The illustrated mean lunar, Metonic and Saros ratios follow from reconstructed tooth counts. | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | meta-2: Confirmed as a compact description of the mean period gear model and interactive controls. It does not make the separate full-glyph-count or direct-CT-attestation assertion. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "The Antikythera mechanism, made runnable. Every gear is a fraction; every fraction is the sky. The 127-tooth gear is half of 254 lunar months; turn the crank and check it." | https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fnature05357/MediaObjects/41586_2006_BFnature05357_MOESM1_ESM.pdf | meta-3: The 127-tooth d2 and 254-month Metonic design relationship are confirmed. The prime arithmetic is exact within that design relation. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "modern 27.321661 d" | https://eclipse.gsfc.nasa.gov/SEhelp/moonorbit.html | modern-sidereal: NASA states "27.32166 days", consistent with the displayed higher-precision 27.321661; the full gear summary and standard table give 27.321661. The excerpt alone only supports five decimals. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "modern 29.530589 d" | https://eclipse.gsfc.nasa.gov/SEsaros/SEsaros.html | modern-synodic: NASA explicitly gives "29.530589 days" for the mean synodic month. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "true 223 mo = 6585.32 d" | https://eclipse.gsfc.nasa.gov/SEsaros/SEsaros.html | modern-saros: NASA gives "6585.3223 days"; multiplying its displayed month length gives 6585.321347. Both round to 6585.32, so the page's two-decimal value is supported. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "235 months = 6939.69 d" | https://en.wikipedia.org/wiki/Metonic_cycle | modern-metonic: Independent calculation 235*29.530589=6939.688415 days rounds to 6939.69. The selected Metonic excerpt alone does not state this precise number. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.
- CONFIRMED | "Moon phase · synodic" | https://dlib.nyu.edu/awdl/isaw/isaw-papers/4/ | moon-phase-differential: The full ISAW article identifies the lunar phase display; the full Wikipedia gearing description explains Moon-Sun differential operation. The selected phase-ball excerpt alone does not derive the differential. Carried from the source rereading recorded in N2-codex.md at 80213b8a47.

## Findings

- [open note:1e5750] X1 (MINOR): Fixed within this layer: the hero now promises reconstructed mean ratios and names separate day conversions and comparisons. The repair worker confirmed missed copies in src/content/strata/the-one-that-broke-off.md:14, src/content/strata/the-tooth-that-keeps-its-word.md:13 and src/content/strata/the-wheel-that-gets-the-same.md:12, plus preserved copies of the latter two in research/lineage-edge-grammar/page.json:1 and public/strata/there-here/index.html:250. research/antikythera/verify.mjs:75-85 supplies MODERN and MECH_YEAR separately. These other layers are outside this worker's edit scope: narrow their authored notes and annotate the preserved quotations through research/lineage-edge-grammar/copy-corrections.mjs without changing the quoted words. Board note requested in repair-report.md; the directing instance must replace PENDING with the issued note id. No page correction line for this MINOR item.
- [fixed] X2 (MINOR): Nature 2006 compares the mechanism with known devices. The introduction now qualifies the comparison by what is known, without asserting what could never have been built. Recorded only; no page correction line for this MINOR item.
- [fixed] X3 (MINOR): Freeth 2008 supplement, section 3.5, printed p. 20, places both two-year and four-year festivals in an Olympiad cycle. The page now names games of that cycle. Recorded only; no page correction line for this MINOR item.
- [fixed] X4 (MINOR): The verifier takes a supplied 365.2425-day year and modern comparisons. The introduction now names those inputs alongside reconstructed tooth counts. Recorded only; no page correction line for this MINOR item.
- [fixed] X5 (MINOR): The page now limits its computation claim to mean gear ratios and residuals, and identifies the browser as a separate illustration. Recorded only; no page correction line for this MINOR item.
- [fixed] X6 (WRONG): Freeth 2006 supplement, PDF pp. 26-27, proves the pin-and-slot equivalent to Hipparchos' epicyclic lunar theory. Replaced the Kepler attribution in the caption and dek and added a mirrored Corrected line.
- [fixed] X7 (MINOR): The ledger contains modern comparison values as well as ratios and day conversions. Its introduction now distinguishes those inputs from tooth-derived ratios. Recorded only; no page correction line for this MINOR item.
- [fixed] X8 (WRONG): Freeth and Jones 2012, paragraphs 100 and 104, distinguish an approximate astronomical relation from the exact 254/19 mean gear rate. The supplied periods differ by 0.094394 day over 254 months versus 19 model years. Qualified the body, dek, descriptions and wrapper; narrowed the unused OG template to the mean gear ratio and added a mirrored Corrected line.
- [fixed] X9 (WRONG): Freeth 2008 supplement, section 4.4, printed p. 32, includes consecutive glyph months 25 and 26. The five/six-month spacing is now limited to the lunar scheme; the illustration explicitly uses synthetic positions. Included in the glyph-count Corrected line.
- [fixed] X11 (WRONG): Independent enumeration gives (3,41), (8,35), (13,29), (18,23), (23,17), (28,11), (33,5). The equation and even spacing alone do not select 33/5. The text now calls it the adopted lunar scheme, and the verifier enumerates and checks all seven solutions. Added a mirrored Corrected line.
- [fixed] X12 (WRONG): Freeth 2008 supplement, Figure 20, printed p. 36, reconstructs 51 glyphs combining 38 lunar and 27 solar predictions. The replacement total is explicitly attributed to that reconstruction; the 38-marker animation is labelled a simplified lunar scheme. Added a mirrored Corrected line.
- [fixed] X13 (MINOR): Freeth and Jones 2012, paragraph 9, locates the subsidiary Exeligmos inside the Saros dial. Replaced "below the Saros" with "inside the Saros spiral". Recorded only; no page correction line for this MINOR item.
- [fixed] X14 (MINOR): NASA describes an approximate eight-hour longitude shift and a return to about the same region after three Saros cycles. Qualified the day length, shift, longitude and time-of-day recurrence as approximate. Recorded only; no page correction line for this MINOR item.
- [fixed] X15 (WRONG): Visually reread Freeth 2008 supplementary Figure 14, printed p. 22. m3, n1 and n3 are red hypothetical elements. The listed inventory is now 28 surviving gears plus those three hypothetical gears; adopted tooth counts can include estimates. Corrected the flags, public verifier copy and prose with a mirrored Corrected line. The separate total of 30 surviving gears is preserved.
- [fixed] X16 (MINOR): Freeth 2021 proposes a Venus role for surviving Fragment D components. Replaced total absence with largely lost planetary trains, retaining the proposed identification as uncertain. Recorded only; no page correction line for this MINOR item.
- [fixed] X17 (MINOR): The footer now limits the verifier to mean gear ratios and residuals with supplied inputs, and calls the browser a separate illustration with a static check summary. Recorded only; no page correction line for this MINOR item.
- [fixed] X18 (WRONG): The Saros canvas label now identifies a simplified lunar scheme. Nearby prose explains the 38 synthetic positions rather than calling them a count of eclipses. Included in the glyph-count Corrected line.
- [fixed] X19 (WRONG): Freeth 2006 supplement, PDF pp. 15 and 21, gives adopted a1=48, b1=223, and one b1 revolution per year. The heading now names the year wheel; prose and wrapper distinguish the interface from the side-crank ratio 48/223 year per turn. Added a mirrored Corrected line.
- [fixed] X20 (MINOR): The dek now promises reconstructed mean ratios and distinguishes separately supplied day scales and astronomical comparisons. Recorded only; no page correction line for this MINOR item.
- [fixed] X21 (WRONG): The dek now attributes 51 glyphs, including 38 lunar predictions, to the 2008 reconstruction. Included in the glyph-count Corrected line.
- [fixed] X22 (WRONG): The dek now lists 28 surviving gears, whose adopted tooth counts include estimates, and three hypothetical gears. Included in the inventory Corrected line.
- [fixed] X23 (MINOR): The dek now limits verifier coverage to mean ratios and residuals. The separately implemented schematic readout is qualified under X31. Recorded only; no page correction line for this MINOR item.
- [fixed] X24 (WRONG): The inline script neither imports nor executes verify.mjs. The wrapper, footer and check summary now identify a separate browser illustration with mirrored values and a static summary. Added a mirrored Corrected line.
- [fixed] X25 (WRONG): The old inset used centre separation 1.088 wheel radii and pin radius 0.7, so its output could not complete a rotation; its old timing had a 27.79058-day period. The repaired inset shares e/rho=0.34 with the front illustration, has centre separation less than pin radius, and uses mean lunar rate minus apsidal rate. The caption and wrapper visibly identify the retained schematic exaggeration. Included in the model Corrected line.
- [fixed] X26 (WRONG): The wrapper now attributes 51 combined glyphs and 38 lunar predictions to the 2008 reconstruction, and describes the animation as a simplified lunar scheme. Included in the glyph-count Corrected line.
- [fixed] X27 (WRONG): The wrapper now identifies 28 surviving gears and three hypothetical gears, with estimated tooth counts distinguished from direct counting. Included in the inventory Corrected line.
- [fixed] X28 (MINOR): The wrapper now says the planetary trains are largely lost and records the proposed, uncertain Venus role of Fragment D. Recorded only; no page correction line for this MINOR item.
- [fixed] X29 (WRONG): m3 and n1 now have a:false alongside n3 in the verifier and its public copy. The wrapper and honesty box accurately describe all three hypothetical gears. Included in the inventory Corrected line.
- [fixed] X30 (WRONG): The meta description and JSON-LD now attribute 51 glyphs and 38 lunar predictions to the 2008 reconstruction. The OG description also narrows derivation coverage. Included in the glyph-count Corrected line; dateModified is 2026-10-14.
- [fixed] X31 (WRONG): The old formula gives 19.783381 degrees at day 5 from the exaggerated illustration. The degree readout now explicitly says "schematic anomaly (exaggerated)" and refers to the illustrated pointer. Both displays share the corrected timing and geometry. Included in the model Corrected line.