Rutherford-Ernest: Quality Engineering (Step 4b)

QE, 2026-10-09. Inputs read: prompt.txt; _text/manifest.json and _textprep.log; claims.md; facts.json (all 7 entries); sites.md (all of it, including the Leads); the research/ notes listing, plus the saved Bumstead 1906 text (research/pages/archive-amjsci-1906-01.txt, sections 7-8); physicist.md, mathematician.md, contrarian.md, arbiter.md and primer.md in full. I also checked the deep/ files by section, and I read the digests for every part of the books that the deep reads do not cover (1930 Ch. IX-XI, XIII §93 and XVIII; 1913 Ch. XI-XVII where they are digest-only; 1904 Ch. X). I searched all 77 digest files by keyword for anomalies, tests and claims. I checked load-bearing numbers against the extracted text in _text/. I ran example.py.

Tags: [read: file, PDF p.N] = checked by me in the extracted text or a digest; [calc] = my arithmetic (Python, project venv); [memory] = recall, with confidence.

Verdict: MINOR

The central verdicts hold: Claims 1-6 are Established in their core, and Claim 7 is split. Several defects need fixing. (1) Two parts of the review rest on digests alone, and the digests themselves were only partly read in Step 1b: Claim 6's departure-from-Coulomb verdict and the Claim 7 "1930 list". (2) Material in the project's own books was missed, and some of it changes sub-verdicts: the Dewar liquid-hydrogen anomaly, the "detonator by radiation" sub-claim, and where the gravity sentence first appears. (3) A few provenance errors need correcting. All of this can be fixed by updating facts.json, claims.md, arbiter.md (with a "QE revision" section) and primer.md after the deep reads below. Steps 2-4 do not need to be redone.

---

1. Coverage

• Manifest: 7 documents, 77 chunks. The digest folder has 77 chunk digests, one per chunk. No document was skipped. The source folder holds exactly these 7 PDFs. Coverage: complete.
• Deep reads exist for all 7 documents. Their page ranges are listed in claims.md and facts.json.
• facts.json has one entry per document, and each how_read is honest. Two of those how_read entries record a shortcut (see 2b).

2. Shortcuts

a. Central claim resting on a digest (defect). Claim 6, "alpha scattering from light nuclei departs from the Coulomb law at close approach", and the elements list for artificial disintegration rest on the 1930 book Ch. IX-X digest. Both are graded "Established (digest-level)". The same applies to the Claim 7 sub-claims that rest on 1930 Ch. XI (attempts to change decay rates; the list "3000 C, 20,000 atmospheres, gravitational fields"). Deep reads 1 and 2 below fix this.

b. The digests were only partly read in Step 1b (defect). facts.json records that the fact-finder read only "the first 100-330 characters of each page line" for most of the 1913 and 1930 books, and "contents and chapter headings for the rest" of the 1930 book. This sampling of the digests is the main cause of the missed material in Section 3. Every item there is in the digests in plain words.

c. Lost deep read not re-run (defect, known). The deep read of the 1913 book, PDF pp.131-192 (the alpha rays), was overwritten. Every step flags it, but nobody re-ran it. The digests show the project does hold primary text for alpha = helium: the 1913 book Sec. 62, PDF p.154 [read: digest f227f1.05], and the 1930 book PDF p.64, Rutherford and Royds, Phil. Mag. 17, 281 (1909) [read: digest d49e1e.02]. The arbiter's statement that "the project holds no deep-read primary text" is literally true, but it reads as "no primary text at all", which is not so. See deep read 3.

d. 1904 book Ch. X is digest-only, but it holds a Claim 7 passage (defect). The passage on the detonator and gravitation is in the project's own 1904 first edition (see Correction C1). Neither the Investigator nor the reviewers used it.

e. Acceptable shortcuts. Digest-only coverage is fine for the 1904 book Ch. I-IV, the 1913 book Ch. I-X (apart from the alpha chapter) and the 1930 book Ch. I-VII, XII and XIV-XVII. They are measurement methods and spectra that no verdict depends on.

3. Missed material (in the digests or extracted text; absent from facts.json, claims.md, all three reviews, arbiter and primer)

In order of importance to the reader:

1. The beta-decay energy crisis. 1930 book §93, PDF pp.436-441. Ellis and Wooster used calorimetry on radium E and found 350,000 ± 40,000 volts per disintegration. That matches the mean of the continuous beta spectrum (390,000 ± 60,000), not the 1,050,000 that the alternative theory predicted. Meitner and Orthmann got 337,000 ± 20,000. The book concludes that a beta disintegration "can release energy that varies within wide limits", and says no satisfactory interpretation exists [read: digest d49e1e.16, pp.440-441]. This is the one true energy-accounting anomaly in the project's record. It bears directly on Claim 2 ("the energy is the atom's own" and is accounted for). Its resolution is Pauli's 1930 neutrino hypothesis and Fermi's 1934 theory, with detection in 1956; Bohr had considered giving up strict energy conservation [memory, high]. A reader of this site needs this item: mainstream physics solved an apparent failure of conservation by postulating an unseen particle and confirmed it 26 years later.
2. Claims of transmutation by ordinary means, and how they were rejected. 1930 book §74, PDF pp.346-349 [read: digest d49e1e.12]. The claims: Miethe (1924-25), mercury to gold in a mercury-vapour lamp; Nagaoka (1925), gold and silver from a discharge under oil; Smits (1926), lead to mercury and thallium, later withdrawn; Wendt and Irion (1922), helium from exploded tungsten wire; Riding and Baly (1925), helium and neon from nitrogen; Paneth and Peters (1926), helium from hydrogen in palladium, later withdrawn. The refutations: Honigschmid and Zintl (Miethe's gold has atomic weight 197.2), Aston (no mercury isotope below 198), Allison and Harkins (a null repeat of Wendt-Irion), and helium released from glass. The book concludes that all claims of transmutation by agents other than swift alpha particles are "untrustworthy". There is also an earlier episode: Ramsay and Cameron (1907-08) reported neon and argon from water and lithium from copper. Rutherford and Royds traced the neon to air leaks, and M. Curie and Gleditsch could not confirm the lithium [read: digest f227f1.11 p.343-344; d49e1e.11 p.313]. This is the rejection record the Contrarian's role exists to examine. Paneth-Peters is the best-known antecedent of the 1989 palladium cold-fusion claims [memory, high]. It belongs in facts.json, in the arbiter under Claims 6/7, and in the primer as related work, with each claimant's own description beside the refutation.
3. Tests of a radiation "detonator". 1930 §76, PDF pp.352-355 [read: _text p.352; digest d49e1e.13]. These tests bear on the arbiter's verdict on Claim 7a ("Plausible, untested"). J. J. Thomson suggested that radium's own radiation might drive its decay. Rutherford tested this in 1903 by dissolving radium bromide into a volume 1000 times larger; the gamma activity was unchanged over a month [read: 1904 book sec. 206, PDF p.357-358; digest e1fda6.14]. In 1911 he found radon's decay rate constant to within 0.05 percent while its concentration fell more than a thousandfold, and Mme Curie confirmed it in 1924. Other tests, all null: - Danysz and Wertenstein (1925), alphas on uranium: at most 1 alpha in 5 million effective. - Herszfinkiel and Wertenstein (1928), alphas on thorium: at most 1 in 8 million. - Hevesy (1922), gamma and beta rays on uranium (at most 15 effective per 10^10 absorbed quanta) and on radium D (change under 0.2 percent). - Jorissen and Vollgraf (1915), cathode rays on uranium. - Walter (1926), 200 kV X-rays on uranium oxide.
4. The 1930 book says gravity was tested. PDF p.352: "powerful electric, magnetic, and gravitational fields have all been looked for, and all in the final analysis without a positive result" [read: _text p.352]. No reference is given. Rutherford co-wrote this book, and he was the one who planned the Rutherford-Compton test. The sentence does not by itself establish a controlled test, but it is pertinent evidence for the arbiter's Absence-of-evidence item 1 and must be quoted there.
5. Tests by heat and pressure. 1913 book Sec. 200, PDF pp.528-530 [read: digest f227f1.18]. Observers disagreed: some found a temporary rise in beta activity on heating, others a fall. Russell showed that both come from radium B and C redistributing inside quartz tubes. Rutherford and Petavel exploded cordite in a bomb holding radium emanation (about 2500 degrees and 1000 atmospheres) with no change at the moment of explosion. Schuster (2000 atmospheres) and Eve and Adams found no change. These belong with Claims 1 and 7b.
6. Pleochroic haloes. 1930 book PDF pp.204-211 [read: digest d49e1e.07]. The halo rings match today's alpha ranges, so the alpha energies (and so the decay rates, through Geiger-Nuttall) were unchanged over geological time. That bears on Claim 1. Joly and Rutherford (1913) estimated that some haloes took several hundred million years to form, which bears on Claim 5. Joly's "X haloes" and Iimori and Yoshimura's "Z haloes" (1926) have radii matching no known alpha range, and the book says their "origin is unknown". That is an open anomaly. Later halo controversies continue the theme (Gentry's polonium haloes [memory, medium]).
7. A residue of ordinary-matter radioactivity that was never resolved. 1930 book §124-125, PDF pp.574-581 [read: digest d49e1e.20-21]. Hoffmann (1920-27) found weak alpha activity in platinum and brass "not due to known radioelements". Ziegert's observations are "left for further investigation", and "alpha particles from a small residue" are "hard to explain". The book also put potassium's activity on mass 41 (the true isotope is K-40) with a half-life of about 1.5e12 years. This bears on the arbiter's Claim 7c ("Tested, not supported"). The verdict can stand, but the unresolved residue has to be recorded.
8. The uranium-radius paradox and the anomalies in Ch. IX. 1930 PDF pp.269-311 and 350-351 [read: digest d49e1e.10-12]: - Scattering bounds the uranium nucleus below about 4e-12 cm, yet alpha emission energy needs a radius of at least 6.3e-12 cm. Wave mechanics resolved the conflict (Gamow; Gurney and Condon). This is the cleanest instructive puzzle in the book, and the reviewers cite it only from memory. - Chadwick and Bieler found hydrogen recoils 100 times the Coulomb value for 8.2 cm alphas, and modelled the alpha as an oblate spheroid of 8 x 4e-13 cm. - Auger and Perrin, and Akiyama, reported forks that broke energy or momentum conservation. Wilson's beta forks broke momentum ("no satisfactory explanation"). Slater reported gamma rays from gold bombarded by alphas ("cannot be regarded as definite"). - Blackett's six nitrogen forks show energy changes from -0.33 to +0.02 of the alpha energy. The aluminium proton energy "is not constant from nucleus to nucleus".
9. Earlier versions of Claim 5's numbers, in the project's own 1904 book. Sec. 205-207, PDF pp.355-363 [read: digest e1fda6.14]: the total heat over radium's life is 2e9 to 5e10 cal/g; the sun could shine "50 to 500 times longer than Kelvin's estimate"; and the "2.5 parts per million" of radium in the sun is W. E. Wilson's 1903 figure (Nature, 9 July 1903), computed with Langley's solar constant (see Correction C4).

4. Accuracy spot checks (against _text/)

example.py runs (exit 0). It reproduces every number the primer prints: 1 in 52,837 and 1 in 61,562; 10^-10,199; the ratio 2,999 against 3,988 and 4,748; the closest-approach table; 124 cal/h; the factor 0.481; 7.7 MeV.

5. Retention

I found no case where pertinent material was dropped from the arbiter or primer because a reviewer judged it negatively. Le Bon (with his stake noted), Dewar, Bumstead, Curie-Danne, the Vienna claims, the New Energy Times dispute and the Rutherford-Compton gravity test are all kept with their sources. Smaller omissions, which the remediation should restore:

• R1. The primer drops Rutherford's 1933 "moonshine" remark, which the arbiter keeps. Both drop Richardson's 1921 "illimitable supply of power" (Papers Past; Investigator lead). Both statements are the record of what Rutherford, and others, expected of atomic energy (Claims 2 and 6).
• R2. Arbiter Claim 7c omits Gates' tests of Le Bon's claims: the quinine-sulphate effect (1904), and heat that transfers rather than destroys the active deposit (1903). It also omits Le Bon's own dispute of Gates' ultraviolet explanation (Investigator lead; physicist and contrarian kept them). The primer lists Gates only under "Related work".
• R3. The contrarian's point that the Chadwick-Mercier gold data in the 1930 book imply an effective charge about 3 below 79 (deep/Radiations 6.5) is dropped by the arbiter and primer.
• R4. Mantle convection as the second reason Kelvin's age failed is in the arbiter (Claim 5) but not in the primer.
• R5. Two Investigator leads appear in neither the arbiter nor the primer: the 1924 Franklin Institute address (helium binding; the satellite picture, whose text is saved at research/pages/archive-naturalartificia0000ruth.txt) and "Some Cosmical Aspects of Radioactivity" (1907, bearing on Claim 5). List both under further reading.

6. Process

• Source tags. Present throughout and honest, with search-summary-only web checks labelled as such. Exceptions:
• The physicist attributes "Compton, Radioactivity and the gravitational field, Phil. Mag. 1920" to the Investigator. sites.md has no such citation, so the citation is unsourced. The arbiter caught this (its item 16).
• The physicist's "1913 book p.26206 region" is a garbled page reference.
• The mathematician (Claims 1 and 7) says Bumstead's heating was "explained by secondary rays (sites.md)". sites.md says only that Bumstead examined whether secondary rays explain it, and the contrarian, who read the paper, reports that Bumstead argued they could not. The mathematician's sentence is a misreading. The arbiter's weighing is unaffected.
• Web checks. Done within budget by the reviewers and the arbiter. Medium- and low-confidence facts that a verdict depends on were checked (Blackett 1925, Bosch 1996, Royds, Harper's dates, Gando 2011) or flagged.
• Verdict scale. Applied consistently, with two exceptions. The arbiter's "Sign of the charge: Not established in 1911; established by 1913-14" is not a scale verdict; use "Historically superseded but instructive" (the 1911 sign was left open) or state it as a note. Its "Kept in the record (no verdict)" rows are acceptable as records.
• Primer. Follows STEP 4's nine sections in order, with answers at the end. It suits the reader profile: Gaussian units, e^2 = 1.44 MeV fm, cross-section and random walk are all introduced; the numerical example is full and reproducible. One small gap: "cosec" is used without the one-line definition cosec x = 1/sin x.
• example.py. Runs and matches the primer.

7. Corrections to earlier documents

• C1. Where gravity and the detonator first appear. The project's own Radio-Activity (1904, first edition), sec. 206, book p.339 (PDF p.357), already has both the detonator analogy and the gravity sentence: "It has not been found possible to alter the rate of change of any product by the application of any known physical or chemical forces, unless possibly it is assumed that the force of gravitation which is not under our control may influence in some way the stability of the radio-atoms" [read: _text lines 17221-17235]. sites.md says this is a 2nd-edition (1905) addition and that "the project has only the 1904 first edition"; that is wrong. Fix: sites.md lead 1; arbiter items 11 and 7e ("Rutherford named it in 1905"); primer Section 4 (detonator row) and Section 7, open item 1 ("named gravity in 1905" becomes 1904); claims.md claim 7.
• C2. Dewar's liquid-hydrogen heat was measured again. P. Curie and Dewar (Royal Institution, 1904) found the heat "about the same in liquid oxygen and hydrogen as in air at ordinary temperature" [read: 1930 book PDF p.184, printed p.159]. The 1906 book says the heat "is unaltered by plunging the radium into liquid hydrogen" [read: 1906 _text line 12039], and the 1913 book reports the same demonstration (Sec. 236, around PDF p.600) [read: _text lines 26159-26165]. The arbiter's "Never properly tested ... no confirmation or withdrawal found" (Claim 2; Absence item 3), the primer's evidence-table row and the primer's Section 7 open item 3 should become: Tested, not supported. The 1904 Curie-Dewar measurement found no excess. It was the same laboratory, not an independent one; say so.
• C3. Where 3.7e9 cal/g comes from. The 1913 book derives it in Sec. 238 [read: _text PDF pp.606-607]. Emanation with its products: 1.83e6 cal per hour per gram, so 2.44e9 cal/g over its life. An average alpha product gives about a quarter of that per gram, about 6e8 cal/g. Hence radium with its products through radium F gives about 3.7e9 calories. This replaces the arbiter's item 5 ("not derivable ...") and the mathematician's "could not trace its exact derivation".
• C4. The sun: the "factor 1.5" is Langley's solar constant. The 2.5 ppm comes from W. E. Wilson (1903), using Langley's surface emission of 8.28e6 cal/cm^2/h [read: 1904 book sec. 207, digest e1fda6.14]. The modern value is 5.42e6, so Langley's is 1.53 times higher. With Langley's figure the radium fraction is 2.53e-6, exactly the printed 2.5 ppm; with the modern figure it is 1.66e-6 [calc]. The 1913 "one-fourth" for a sun of pure uranium also becomes 0.30 with Langley's constant (0.46 with the modern one) [calc]. That accounts for most of the arbiter's "unexplained factor 1.8" (the 1913 book quotes Langley on the same page). Fix arbiter R3, the primer Section 4 sun row, and mathematician items 42 and 44.
• C5. How many elements Cambridge disintegrated. It was not "six elements". The forward method (1921-22) gave B, N, F, Na, Al and P. The right-angle method added Ne, Mg, Si, S, Cl, Ar and K. In total, every light element up to potassium except H, He, Li, Be, C and O [read: 1930 book PDF pp.316, 321; digest d49e1e.11]. Fix arbiter Claim 6 and the primer's evidence-table row.
• C6. Alpha = helium in the project's own texts. Rutherford and Royds is in the project at digest level (1913 book PDF p.154, Sec. 62; 1930 book PDF p.64). Change "the project holds no primary text" to "digest-level [read: digest]; the deep read was lost" in the arbiter source gap and the contrarian Claim 3.
• C7. Bumstead tested the arbiter's calibration explanation himself. In sec. 8 of his paper (saved at research/pages/archive-amjsci-1906-01.txt, printed pp.17-18) he measured the absorbed fractions with the radiometer, which measures energy, instead of the electroscope: zinc 88 percent and lead 91 percent, against the electroscope's 78 and 79 percent. So the factor of two in heating was not an ionisation-calibration artifact as far as he could test. Arbiter 7d and Absence item 2(ii) should record this. The verdict "Never properly tested" (one laboratory, no replication) stands.
• C8. The detonator sub-verdict (arbiter 7a). Triggering by radiation was tested and gave null upper limits: Rutherford 1903 and 1911; Danysz and Wertenstein; Herszfinkiel and Wertenstein; Hevesy; Jorissen and Vollgraf; Walter (Section 3, item 3). Proposed split:
• Trigger by alpha, beta, gamma, X-ray or cathode-ray bombardment of radioactive nuclei: Tested, not supported (at the stated limits).
• Trigger by "some unknown external force": Not testable as stated.
• Gravity: Never properly tested, with the 1930 sentence (Section 3, item 4) quoted beside it.
• C9. Cosmetic. facts.json, 1930 book entry: add §74, §76, §93 and §124 to the summary and key_claims, since they are currently missing. facts.json, 1904 book entry: add that sec. 206 has the gravity sentence.

8. Remediation

Deep reads (at most 8, most important first)

1. Radiations-From-Radioactive-Substances-d49e1e, PDF pp.269-349 (Ch. IX-X), Opus. Claim 6 needs a primary basis. Questions: - What exactly were the departures from Coulomb scattering for H, He, Mg and Al (distances, ratios, Chadwick-Bieler, Bieler, Rutherford-Chadwick 1925), and how did the book model them? - Which elements disintegrated, by which method, with what probabilities per million alphas? - What do Blackett's forks show about energy and momentum? - What is the Vienna-Cambridge disagreement as stated? - What does §74 report about each transmutation claim (Miethe, Nagaoka, Smits, Wendt-Irion, Riding-Baly, Paneth-Peters) and its refutation, quoted with page?
2. Radiations-From-Radioactive-Substances-d49e1e, PDF pp.350-369 (Ch. XI §75-77), Opus. Claim 7 sub-verdicts depend on it. Questions: - What does §76 say about attempts to change decay rates, including the "gravitational fields" sentence and any reference behind it? - What are the upper limits from Rutherford 1911 (radon), Danysz-Wertenstein, Herszfinkiel-Wertenstein, Hevesy, Jorissen-Vollgraf and Walter? - What do the uranium-scattering results and the uranium-radius paradox show? - What were Rutherford's 1927 satellite theory and its stated objections, and how does the book treat Gamow, Gurney and Condon?
3. Radioactive-Substances-and-Their-Radiati-f227f1, PDF pp.131-192 (the lost alpha-ray chapter), Sonnet. Re-run it, with this question: what does the book give for the alpha's charge (Rutherford-Geiger; Regener), the alpha count per gram of radium, and the Rutherford-Royds helium identity, with numbers and pages?
4. Radiations-From-Radioactive-Substances-d49e1e, PDF pp.422-447 (Ch. XIII §88-94), Sonnet. Question: what do the Ellis-Wooster and Meitner-Orthmann calorimetry of radium E and the continuous beta spectra show about the energy per beta disintegration, and how does the book state the unresolved problem (quote "varies within wide limits")?
5. Radioactive-Substances-and-Their-Radiati-f227f1, PDF pp.340-346 and 525-535, Sonnet. Questions: - What tests of temperature and pressure on decay rates are reported (Bronson, Russell's beta-activity explanation, Rutherford-Petavel cordite bomb, Schuster, Eve-Adams), with conditions and results? - What did Ramsay and Cameron claim about transmutation, and how was it rebutted (Rutherford-Royds air leak, Curie-Gleditsch)?
6. Radiations-From-Radioactive-Substances-d49e1e, PDF pp.204-211 and 574-581, Sonnet. Questions: - What do pleochroic haloes show about the constancy of alpha ranges over geological time and about ages (Joly, Joly-Rutherford), and what are the unexplained X and Z haloes? - What residual activity of ordinary matter is left unexplained (Hoffmann's platinum and brass alphas, Ziegert), and how is potassium and rubidium activity treated?
7. radio-activity-e1fda6, PDF pp.340-366 (Ch. X sec. 200-207), Sonnet. Questions: - What does the 1904 first edition say on external-energy theories, the detonator and gravitation (quote sec. 206)? - What was the 1903 self-irradiation test? - What are the total-energy estimate (sec. 205) and the sun estimate (Wilson's 2.5 ppm with Langley's constant; "50 to 500 times")? - What is the earth-heat calculation X = 3KT/R?
8. Radioactive-Substances-and-Their-Radiati-f227f1, PDF pp.595-611 (Sec. 236-238, heat emission), Sonnet. Questions: - What exactly do the Curie-Dewar liquid-hydrogen and liquid-oxygen measurements show? - How is the heat split among alpha, beta and gamma rays? - What is the full derivation of 2.44e9 cal/g (emanation) and 3.7e9 cal/g (radium through F)?

Changes to later files

• facts.json:
• Apply C1-C6 and C9.
• Add to the 1930 entry: the transmutation claims and refutations (§74); the tests of a radiation trigger and the gravity sentence (§76); the beta-energy problem (§93); haloes; the ordinary-matter residue.
• Add to the 1913 entry: Sec. 238 (derivation of 3.7e9), Sec. 200 (temperature and pressure tests) and Ramsay-Cameron.
• Add to the 1904 entry: sec. 205-207.
• claims.md:
• Claim 7: gravity and the detonator are in the 1904 first edition.
• Claim 6: all light elements to K except C and O.
• Consider adding a claim, or a "related claims" list, for the transmutation claims by ordinary means (rejected), so later steps carry them.
• Update the Reading limits after the deep reads.
• arbiter.md, adding a "QE revision" section:
• Claim 2: Dewar becomes Tested, not supported (C2). 3.7e9 provenance (C3). The beta continuous-spectrum anomaly and its later resolution, as an instructive item.
• Claim 5: Langley's constant explains 2.5 ppm and most of the "one-fourth" gap (C4). The 1904 "50 to 500 times".
• Claim 6: the element list (C5); deep-read support for the Coulomb departures, or a downgrade if the deep read does not support them; the transmutation claims (§74) and Ramsay-Cameron, each with its proponents' description and its refutation.
• Claim 7: re-split 7a (C8); quote the 1930 gravity sentence in 7e and Absence item 1; 7c adds the Hoffmann and Ziegert residue and Gates' tests of Le Bon (R2); 7d adds Bumstead's radiometer check (C7); 7e is dated 1904 (C1).
• Claim 1: Rutherford 1911 radon 0.05 percent; Rutherford-Petavel; Russell; haloes as geological-time constancy.
• Retention items R1 and R3. Fix the "Sign" verdict wording.
• primer.md:
• Section 4: detonator and gravity row from 1904 (C1); sun row with Langley (C4).
• Section 6, evidence table: Dewar row (C2); element list (C5); a 7a split row (C8); a row for transmutation claims by ordinary means (Tested, not supported, with the claimants named); a row for the beta energy spread (anomaly of 1927-30, explained by the neutrino [memory, high]).
• Section 7: open items revised (C2, C8); a "what is open" line on the X and Z haloes and the Hoffmann residue; one sentence on the uranium-radius paradox as the puzzle wave mechanics solved; "moonshine" and Richardson (R1); mantle convection (R4).
• Section 9: the 1924 Franklin address and "Some Cosmical Aspects" (R5).
• Define cosec on first use.
• sites.md: correct lead 1 ("the 2nd edition adds gravity") per C1.
• physicist.md and mathematician.md: no rewrite. The arbiter's QE revision supersedes the Compton citation, the Bumstead misreading and the "could not trace 3.7e9".

Corrections to earlier documents (for the orchestrator)

C1-C9 above. The stale statements to search for across review/, the entries and the site are: "named gravity in 1905", "Radio-activity 2nd ed. ... names gravitation", "Dewar ... never repeated", "Never properly tested (printed numbers show no anomaly)", "six elements", "3.7e9 ... not derivable", "unexplained factor 1.8" and "2.5 ppm ... 1.5 times larger".

Recheck

QE recheck, 2026-10-09. I read the current arbiter.md (including the new "QE revision" section, line 179 onward), primer.md, claims.md, facts.json and sites.md. I grepped for each stale phrase listed in Section 8 and for each missed item. Deep reads 1-8 exist: the files in deep/ were updated at 03:28-03:33 and the arbiter cites them.

Fixed: - Section 2a, 2b, 2c, 2d (digest-only reliance, lost alpha read, 1904 Ch. X): fixed. Deep reads were done, and the arbiter's QE revision cites them. The 1930 gravity/list sentence is now confirmed from the page image. - Section 3, items 1-9 (beta-energy anomaly with the neutrino resolution, transmutation claims and refutations, radiation-trigger null tests, 1930 gravity sentence, heat and pressure tests, haloes, Hoffmann/Ziegert residue, uranium-radius paradox, 1904 sec. 205-207 numbers): fixed. All are present in facts.json, claims.md, arbiter.md and primer.md. - C1 (gravity and detonator are in the 1904 first edition): fixed in arbiter.md (QE revision), claims.md and primer.md ("named gravity in 1904"). - C2 (Dewar): fixed. Tested, not supported, with "same laboratory" stated, in the arbiter, primer table and primer open items. - C3 (3.7e9): fixed, with a correction to this QE report. The deep read of 1913 sec. 238 shows that only 2.44e9 cal/g is derived. The 3.7e9 is stated by a rule of thumb, not derived. My C3 wording overstated this. The arbiter (item 5 note, line 311) and claims.md line 41 now say so, and I accept that. - C4 (Langley's constant): fixed. It explains the 2.5 ppm exactly. The "factor 1.8" shrinks to about 1.2, not to zero. - C5 (element list), C6 (alpha = helium at digest level), C7 (Bumstead radiometer check), C8 (7a split), C9 (cosmetic facts.json additions): fixed. - R1-R5 (moonshine, Richardson, Gates, Chadwick-Mercier, mantle convection, Franklin address, "Some Cosmical Aspects"): present in the arbiter and primer. - "Sign of the charge" verdict wording: fixed (arbiter line 195). Cosec is defined in the primer. - example.py: unchanged, still consistent with the primer. - Stale phrases: the pre-revision text in arbiter.md items 5, 10, 11, 12 and (e) (line 110) still contains "named gravity in 1905", "six elements" and "not derivable". Lines 19 and 311 say the QE revision supersedes them, items 5, 10, 11 and 12 carry inline pointers, and line 317 lists the phrases for searching. I accept this as a documented supersession, not a defect. primer.md and facts.json contain none of the stale phrases.

Remaining: - R6 / sites.md (1 issue). sites.md lead 1 (line 67) still says "The project has only the 1904 first edition" and presents the gravity sentence as a 2nd-edition (1905) addition. The 1904 first edition has the sentence (sec. 206). Section 8 required this correction. Fix: say that the 1904 first edition sec. 206 already contains the sentence, and that the 2nd edition (sec. 270) repeats it.

Final: OPEN (1 issues)

Orchestrator fix (2026-10-09)

Orchestrator fix (2026-10-09): sites.md lead 1 (line 67) now says the 1904 first edition sec. 206 already has the gravity sentence and the 2nd edition sec. 270 repeats it.

Final: PASS