Aspden-Hearld: Arbiter (Step 3)

Inputs: facts.json, claims.md, physicist.md, mathematician.md, contrarian.md. Source documents re-read by the arbiter: the 2-page plasma essay (full text layer) and book pp. 19-20 (printed p.12, Maxwell's "Q" quotation). Tags: [read], [calc] (own arithmetic, project venv), [web] with URL, [memory] with confidence.

Source gap (repeated from claims.md and all three reviews). Physics Unified (212 pp.) is a scan with no text layer. Every reviewer worked from Tesseract OCR as a navigation aid, with a few equations (eqs 159, 160, 163, 167, 168) checked against page images. No reviewer read every page. The chapters on the Planck law, the Schroedinger equation, the electron g-factor, geomagnetism and field reversals, solar-system formation, and the detailed treatment of the aether experiments were NOT read in depth by any reviewer, so no verdict was given on those parts in the original text; the QE revision below (Q3, Q4, Round 2) supplies scale verdicts after deep reads. All experimental support in the plasma essay is by citation only; the cited papers (Kobel 1930; Graneau 1983/85; Roth 1978; Sethian et al. 1978; Aspden 1969/77/85/86), the three Correa patents and Aspden's 1996 report "Power from Space: The Correa Invention" were not obtained by anyone. Where a conclusion below rests on [memory] alone, it says so. To close the gap, obtain: Aspden, J. Franklin Inst. 287, 179 (1969); the Graneau papers; Roth 1978; Sethian et al. 1978; the Correa patents; and a clean text copy of Physics Unified (key equations 14, 157, 163, 168, 231).

0. Reconciling the three reviews: points of disagreement and how they were settled

Numbers recomputed by the arbiter [calc]

• alpha^-1 = 108 pi (8/1843)^(1/6) = 137.035915; CODATA 2022 = 137.035999177; difference 8.43e-5 = 0.615 ppm. Reviewers agree. Neighbouring odd integers: 1841 gives 137.061, 1845 gives 137.011 (reviewers' values; consistent).
• G from eq. 163, G = (e/m_e)^2 (4 pi)^2 (1/(108 pi))^6 (1/g)^8 with e/m_e = 5.273e17 esu/g: g = 5063 gives 6.6655e-8; g = 5062.75 gives 6.6681e-8; the g that gives exactly CODATA 6.6743e-8 is 5062.16. Aspden's extra "0.0004" factor brings 5063 to about 6.668e-8. So the book is about 0.1 percent low. G depends on g^-8, so G to 0.1 percent only needs g to 0.013 percent. All reviewers agree. Because g was fitted partly to G (book p.151 per two reviewers), this is circular.
• Dimensionless check: G m_e^2/e^2 = 2.40e-43. If gravity were (v/c)^2 times Coulomb, v/c ~ 5e-22 would be needed. So the electrodynamic law alone cannot supply gravity; the lattice assumptions carry the number. (Agrees with mathematician and contrarian.)
• Claim 6: m/k = 0.0408 m_e / k_B = 2.69e-13 K per (erg/g). phi (Sun + Earth, relative to infinity) = 9.50e12 cgs, giving T = 2.56 K. The book rounds to 2.6 K; the "10^-18" in the prose is an OCR/typo (all three reviewers agree).
• Galactic potential. The Sun term is a potential relative to infinity, so the consistent comparison is the Galaxy's potential relative to infinity. Using phi = v_esc^2/2: v_esc 445 km/s gives 270 K, 550 km/s gives 407 K, 580 km/s gives 453 K [calc]. The mathematician's 400 K corresponds to 550 km/s. The contrarian's 130 K used v^2 with v = 220 km/s (circular speed), which is not an escape-potential estimate; with v_esc^2/2 at 220 km/s one gets only 65 K. Local escape speed estimates from Gaia range from 445 to 580 km/s [web: https://arxiv.org/pdf/1807.04565 (580 +/- 63 km/s) and https://iopscience.iop.org/article/10.3847/1538-4357/ac4244 (445 km/s)]. So the Galactic term is between roughly 270 and 450 K, 100 times or more the Sun+Earth total, however one counts.
• Annual modulation. Sun term alone 2.39 K; Earth-Sun distance varies by +/-1.67 percent, so +/-0.04 K, 0.08 K peak to peak [calc]. Physicist and contrarian agree. I did not find a source saying whether existing CMB monopole data constrain this; flagged as a testable calculation.

1. Claim 1: Modified electrodynamics, F = (qQ/r^3){(V.r)v - (M/m)(v.r)V - (v.V)r}

Agreement. All three agree that: closed-circuit data really do leave the isolated-charge law undetermined (Maxwell's own words, [read: book p.12, printed, "any value of Q which appears to us to simplify the formulae"]); eq. 14 equals the standard magnetic (Lorentz/Biot-Savart) force plus one extra term, -(qQ/r^3)(M/m)(v.r)V; the extra term integrates to zero around a closed loop; the law violates Newton's third law for an isolated pair and needs a preferred frame and an aether to absorb momentum; no experiment was designed to test the mass-ratio term itself; the cited experiments (Kobel, Graneau, Roth, Sethian) were not obtained.

Disagreement. Whether to count the precision record of the standard law as a test. The physicist does ("Tested, not supported" indirectly); the mathematician and contrarian do not. Which is stronger: the strict reading of the scale. "Tested, not supported" requires a controlled experiment aimed at the claim. The standard law was never compared against this term; the statement that such data would have revealed it is a scale estimate by the physicist (extra term about (m'/m)vv'/c^2, 2e-4 for thermal electrons and slow ions, about 0.6 for relativistic electrons) that rests on [memory, medium] about beam-ion dynamics, and the mathematician's equally unverified point about electron-electron (M/m = 1) scattering. I keep it as a tension.

The mathematician's logical critique stands and was partly re-checked: the book's Chapter 1 argues a = b by assuming equal and opposite pair forces, then drops balance (book pp.13-14 per two reviewers; I confirmed only the p.12 Maxwell quotation myself). The claim that the law is "unique" is not shown: Aspden's own essay offers a third justification (avoid a "twisting problem"), and the book a second (Maxwell's Q = -1/r), so the law is justified three ways, not derived once.

Graneau: [web: https://www.sciencedirect.com/science/article/abs/pii/0375960185905912; search summary at the Graneau/Nasilowski result list] The wire-fragmentation experiments are real (kA pulses breaking 1 mm copper wire into about 100 tensile-fracture pieces); Graneau attributes this to Ampere longitudinal forces, others to self-induced emf imbalance and conventional forces. Unsettled. Importantly, Ampere's law is a different law from Aspden's eq. 14 [read/web], so Graneau does not support Aspden's term specifically.

Final verdicts (split). - Closed-circuit experiments cannot fix the isolated-charge term: Established (historical). - Aspden's derivation as a proof (uniqueness, conservation-law basis): Internally inconsistent (balance assumed then dropped; uniqueness asserted). - Aspden's law as the true law for isolated charges: Never properly tested. Not shown to be dismissed without testing; Aspden says it was ignored, and his 1969 and 1985-86 papers appeared in refereed journals [read: essay], so "ignored" is accurate and "suppressed" is not supported. Tension with standard precision data is flagged but unsourced [memory, medium].

2. Claim 2: Plasma energy partition; corona; "sun not fusion powered"

Arbiter's own check of the key argument. For an ion with velocity V and an electron moving with velocity v, the force on the ion from the extra term is proportional to (v.r)/r^3. With the ion nearly static, (v.r)/r^3 = d(1/r)/dt (r pointing from electron to ion), so the impulse of a complete flyby is the change in 1/r, which is zero [calc by mathematician, re-derived]. Aspden's cases 1 and 2 (electron approaching, receding) cancel exactly, and so do cases 3 and 4. When the ion moves, d(1/r)/dt = (v-V).r/r^3, so (v.r)/r^3 = d(1/r)/dt + V.r/r^3 and the leftover is second order in V; Aspden shows no calculation of it. The essay itself says the statistics are "an interesting, though quite complicated mathematical task ... which I trust some enthusiastic mathematician might undertake" [read].

Corona: see section 0. Neutrinos: [web: https://arxiv.org/pdf/2006.15115 (Borexino CNO neutrinos); https://www.nature.com/articles/s41586-020-2934-0] solar pp-chain and CNO-cycle neutrinos are directly detected, which measures the fusion reaction rates; the essay's reply that neutrinos "are not a measure of temperature" [read] is correct but does not address that they identify the reaction. The essay also states "Indeed I am sure it is not" [read].

Ion-hotter-than-electron plasmas: real and replicated across devices [memory, medium for coronal-hole UVCS ion temperatures; the reviewers cite search summaries]. Standard explanations (ion-cyclotron waves, reconnection, collisionless heating) are mainstream, though coronal heating is not completely settled [memory, medium].

Final verdicts (split). - Ion temperature can exceed electron temperature in plasmas: Established (as an observation). - Aspden's term as the cause: Never properly tested (no quantitative calculation by him or anyone found; no test aimed at it). His own four-case argument: Refuted by calculation at first order (shown above); the full statistical case remains open, because second-order and correlated-flow effects were not ruled out. - The corona "factor of 160": Internally inconsistent (essay compares photosphere temperature with corona temperature and calls it ion versus electron). - "The sun is not fusion powered": Tested, not supported (neutrino fluxes [web]). Not dismissed without testing; it was tested by SNO, Super-Kamiokande and Borexino among others [memory/web]. - Sethian et al. and Roth as support: weak or none (Sethian: equilibration, not net gain [web]; Roth not obtained).

3. Claim 3: Aether, preferred frame, Correa "power from space"

Agreement. A preferred frame compatible with all null results exists in principle (Lorentz aether theory is observationally equivalent to special relativity, [memory, high]). Aspden's book arranges the "E-frame" so that it follows the laboratory near matter [read: book pp.48-49, contrarian], which makes lab nulls automatic. The CMB dipole is real [memory, high] but is read by mainstream physics as motion relative to the matter/CMB frame, not as an aether drift. Aspden's aether is an unquantified energy and momentum bookkeeper; no energy budget for any device is given [read: essay].

Disagreement. Physicist: "Tested, not supported" (modern isotropy tests, about 1e-17). Mathematician and contrarian: "Not testable as stated". Ruling: the physicist's claim is [memory, medium] and not web-checked, and it applies to preferred-frame theories that predict drift, whereas Aspden's lattice, as the contrarian read it, is built not to. Aspden's specific force law (claim 1) does imply frame-dependent forces, which are testable in principle, but that belongs to claim 1.

Correa device. [web: https://www.rexresearch.com/correa/correa.htm, https://aetherscience.org/energy-science-report-8/, https://aetherometry.com/correa_publications.php, https://www.eugenemallove.com/emwords_SEAS.html, patents US 5,449,989 and 5,502,354 at https://patents.google.com/patent/US5449989A/] The claims of excess energy come from the Correas' own publications and advocacy sites; my search found no independent peer-reviewed replication and found disputed informal verification and non-verification attempts discussed on enthusiast forums. A patent grant and a citation of Aspden's 1969 paper in a patent are not tests. The contrarian's caution is apt: over-unity from a pulsed discharge with large reactive power is prone to measurement artefacts (power factor, spiky waveforms) [memory, medium]. Not obtained: the patents' data, Aspden's 1996 report.

Final verdicts (split). - Lattice aether defining a preferred frame, as formulated: Not testable as stated (fits null results by construction); the CMB dipole is not evidence for it. - Correa device draws excess energy from the aether: Never properly tested (no neutral controlled test on record that I could find). Whether it was "dismissed without testing" cannot be established from these sources. The decisive test is cheap: calorimetric input and output accounting by a neutral lab.

4. Claim 4: Gravity from electrodynamics, 2.587 GeV graviton, G

Agreement. Eq. 163 is dimensionally consistent and numerically reproduces G to about 0.1 percent; g (5063) is not independent of G: the book itself says the mass "gives the value of G exactly" and is "supported by the derivation of G" [read: p.151 per physicist and mathematician]. The electrodynamic law alone cannot supply gravitational strength (2.4e-43 gap, above); the lattice ansatz supplies it. The 2.60 GeV resonance was reported in 1977 and not confirmed by independent searches, according to the book itself [read: p.119 per all three]; the 2.260 GeV match to a 1976 Fermilab announcement rests on a popular book; its fate is unverified [memory, unverified; capped at medium confidence; see QE revision Q6].

Disagreement. The physicist calls "gravity is an electrodynamic effect" "Tested, not supported" via the equivalence principle and GW170817; the contrarian and mathematician say "Not testable as stated". The equivalence principle is tested to about 1e-15 [web: https://arxiv.org/abs/2209.15487, MICROSCOPE final result eta(Ti,Pt) = [-1.5 +/- 2.3 (stat) +/- 1.5 (syst)] x 10^-15]. But Aspden's gravitons are vacuum constituents and the book gives no explicit composition-dependence prediction that I can test, so the equivalence principle cannot be applied without writing that prediction. I side with "Not testable as stated" for the mechanism.

The contrarian's observation that the psi-particle thresholds (3.0948 vs J/psi 3.0969 GeV; 3.6830 vs 3.6861) are 0.07-0.08 percent from the formula with several adjustable choices stands; J/psi was discovered in 1974, so these are retrodictions [memory, high on J/psi mass; contrarian [calc]]. The graviton mass itself (2.587 GeV, stated in 1966) was a fit to hadron data of that era.

Final verdicts (split). - Gravity is an electrodynamic effect of the vacuum lattice: Not testable as stated. - G "calculated": Not a prediction. Arithmetic correct, g tuned, extra 0.0004 factor; circular. Against CODATA it is 0.1 percent low [web: https://physics.nist.gov/cgi-bin/cuu/Value?bg, via mathematician]. Verdict Internally inconsistent only in its claim to be a parameter-free derivation; as a numerical comparison it is Tested, not supported at today's precision. - A 2.587 GeV graviton exists: Tested, not supported (the book's own account of failed independent searches [read]; no such state in modern listings [memory, medium; not web-checked]).

5. Claim 5: Derived constants (alpha, m_p/m_e, g-factor, Planck, Schroedinger)

Agreement. All recomputed alpha^-1 = 137.0359148 and m_p/m_e = 1836.1523155 (412.665816/(sqrt(3/2) - 1)) and agree the arithmetic is right. All agree the integer 1843 selects the answer and the inputs (108 pi, the 1/6 power, 8, k = 3/2, muon-pair identification) are choices. The mathematician added that the book's own lattice sum gives a volume ratio 1844.6, not an odd integer, with 1845 being nearer; 1845 gives 137.011. The choice of 1843 is by a directional argument (verified in QE round 2: under the book's rule 1845 gives E < 0, so 1843 is uniquely selected given the premises; see Q4). Physicist and mathematician both note the book's own footnote gave Williams-Olsen 137.035963(15), which differs from 137.035915 by 3.2 sigma [calc], so the formula was already under strain c.1980.

Current values: alpha^-1 = 137.035999177, relative uncertainty 1.6e-10 [web: https://physics.nist.gov/cgi-bin/cuu/Value?alph]; m_p/m_e = 1836.152673426(32) [web: https://physics.nist.gov/cgi-bin/cuu/Value?mpsme]. The formula alpha is 0.615 ppm low, about 3,800 standard uncertainties [calc]. The physicist's "about 30,000 standard uncertainties" used a different (smaller) uncertainty; with 2.2e-8 absolute (1.6e-10 relative of 137.036) the figure is 8.43e-5/2.2e-8 = 3,800, and the arbiter adopts 3,800 (matches the mathematician and contrarian). The mass ratio is off by 1.9e-7 relative, about 1e4 times the present uncertainty [calc].

Final verdicts (split). - alpha and m_p/m_e formulas as exact results: Tested, not supported (excluded by precision measurement, repeatedly). The comparison was a normal measurement, not a controlled test aimed at Aspden; no replication question arises. - c.1980 numerical agreement: Historically superseded but instructive (a legitimate fit to then-current data, within error). - "No free parameters other than integers": Internally inconsistent (a fitted integer plus several chosen structural constants). - Derivations of Planck's law, the Schroedinger equation and the g-factor from lattice rotation: original text gave no verdict (not read in depth); now Planck and Schroedinger: Not testable as stated; g-factor: Tested, not supported (QE revision Q4). Reproducing a known formula from a model is not itself a test of the model.

6. Claim 6: Cosmic background as local lattice temperature; geomagnetism, domains

Agreement. The arithmetic reproduces 2.56 K, rounded to 2.6 K; the extra 0.1 K to reach 2.7 K is attributed after the fact to "one electron per 600 or so lattice particles" [read: p.178 per contrarian]; T_0 is set to zero near Earth by assumption; the CMB is a near-perfect blackbody 2.7255 +/- 0.0009 K [web: https://arxiv.org/pdf/astro-ph/9705101 via mathematician; https://lambda.gsfc.nasa.gov/product/cobe/about_firas.html via contrarian]. The potential used is only that of the Sun and Earth.

Which argument is stronger. The mathematician's point that the formula assigns a temperature to the value of a potential, which is defined only up to a constant, and that the Galaxy's potential, using the same relative-to-infinity convention, would add roughly 270-450 K, is the decisive one. The physicist's T(z) argument is independent: the CMB temperature follows T0(1+z)^(1-beta) with beta = 0.010 +/- 0.013 (all measurements combined) or (7.6 +/- 8.0)e-3 [web: https://arxiv.org/abs/2106.00119 and the search results listed under the query "CMB temperature redshift T(z)"], measured in gas clouds far from our Sun, where a local gravitational-potential origin has no reason to produce a (1+z) trend. The physicist's quoted "0.4 +/- 1.6 percent" from arXiv:1112.1862 was not re-verified by the arbiter; the newer sub-percent figure above replaces it.

The contrarian's favourable point is real: Aspden's formula is explicit and gives a checkable gradient inside the solar system (T falls with distance from the Sun; about +/-0.04 K annual swing), which is rare in aether writing. The test has not been reported as done in any source I found. It is cheap against existing COBE/WMAP/Planck data, though I do not know whether those data already rule it out.

Final verdicts (split). - The numerical match to 2.7 K: Tested, not supported (0.1-0.17 K off at present precision; fix ad hoc). - 2.7 K as local lattice temperature from gravitational potential, in the stated form: Refuted by calculation (the same eq. 230, applied with the Galaxy's potential, gives hundreds of K; the potential's zero is arbitrary). Also Tested, not supported by T(z) and the blackbody/anisotropy structure. - The solar-system gradient (T vs distance from the Sun; annual modulation): Never properly tested by anyone I could find, and no source says it was dismissed. - Geomagnetism, field reversals, solar-system formation from rotating vacuum domains: original text gave no verdict; now moment: Never properly tested; mass ratios: Internally inconsistent / Never properly tested; reversals: Tested, not supported (QE revision Q4).

The contrarian's distinctions, kept visible

Tested and failed: alpha and m_p/m_e formulas (precision measurement, [web]); the 2.60 GeV resonance as a graviton (the book's own account of failed searches); "sun not fusion powered" (neutrinos, [web]); local-potential CMB temperature versus T(z) and the Galactic potential.

Never properly tested: the mass-ratio term of the force law itself; Aspden's mechanism for ion-hot plasmas (never quantified, and the stated argument fails at first order); the Correa device (no neutral calorimetry on record); the solar-system CMB gradient. The Graneau wire-fragmentation question is real and not closed, but concerns Ampere's law, not Aspden's.

Dismissed without testing: no source on file shows any claim was dismissed without a test after being formally proposed. Aspden's own account is that his papers were "ignored", although they appeared in refereed journals (J. Franklin Inst. 1969, IEEE Trans. Plasma Sci. 1977 and 1986, Phys. Lett. A 1985, [read: essay citations, not independently verified]). Neglect, not documented suppression, is what the sources support. No funding block, retraction or data omission is documented in these materials. The polemic against academic publishing in the essay is a grievance, not evidence.

Rigour applied to the researcher (the contrarian's points, accepted): the essay asks someone else to do the statistics its own case needs, then cites experiments as support; it asserts certainty ("I am sure it is not") against direct neutrino evidence; it compares photosphere and coronal temperatures as if ion versus electron; numerical agreements carry ad hoc corrections (0.0004; one electron in 600) and a fitted integer.

Corrected claim list

1. Modified electrodynamics. The force between moving point charges contains Maxwell's undetermined (v.r)v' term with factor -(M/m) instead of +1 (as stated in the book and essay in c = 1 units; Gaussian units need 1/c^2). Closed-circuit data cannot exclude it (Established). The derivation does not prove uniqueness and uses balance of action and reaction then drops it (Internally inconsistent). The law itself was Never properly tested; tension with accelerator data is unsourced.
2. Plasma energy partition. Ion-hot plasmas exist (Established); Aspden's term as the cause: Never properly tested, with his four-case argument failing at first order (Refuted by calculation for that argument); the corona "160" comparison: Internally inconsistent; "sun not fusion powered": Tested, not supported. Sethian et al. found faster equilibration, not net energy gain.
3. Aether as energy source. Preferred-frame lattice aether: Not testable as stated. Correa "power from space": Never properly tested (no neutral controlled test found; claims come from the proponents). Chapter 3 experiments (QE Q4): only the Rydberg factor 2 - gamma differs from special relativity and is Tested, not supported [memory, medium]; the rest match SR or are null by construction; Townes and Silvertooth-ruby repeat: Never properly tested.
4. Gravity from electrodynamics. Mechanism: Not testable as stated. G: explicit split, as a parameter-free derivation Internally inconsistent (g tuned to G, extra 0.0004 factor), as a number Tested, not supported (0.1 percent low). 2.587 GeV graviton: Tested, not supported. Retardation/perihelion derivation (Ch.2, eq. 36): Internally inconsistent / not derived as claimed (QE revision, section Q2).
5. Derived constants. alpha^-1 = 137.035915 (book's revised 137.035953) and m_p/m_e = 1836.15232: Tested, not supported (about 3,800 and 2,100 standard uncertainties off for the two alpha values; about 1e4 for m_p/m_e). c.1980 agreement: Historically superseded but instructive (the revised alpha was within 0.7 sigma of Williams-Olsen). "No free parameters": Internally inconsistent. Particle-mass and moment matches: final per match (Q3): deuteron, neutron decay, pion, pi0, psi states: Historically superseded but instructive (now Tested, not supported at modern precision); muon mass, muon g, proton moment, muon lifetime, tau: Tested, not supported; quark masses: Historically superseded but instructive. Lattice sum r/d: Internally inconsistent (convention-dependent); 1843 is uniquely selected given the book's premises. Planck and Schroedinger derivations: Not testable as stated; g-factor series: Tested, not supported (Q4). 137.035953 is model-derived but post hoc (Q3, Appendix III).
6. CMB and cosmology. Local-potential temperature: Refuted by calculation in its stated form and Tested, not supported (T(z), blackbody); solar-system gradient: Never properly tested. Geomagnetic moment: Never properly tested (one-parameter fit); planetary mass ratios: Internally inconsistent (Earth) / Never properly tested; field reversals and 700,000-year domains: Tested, not supported (polarity timescale) (Q4).

What would settle it

• Claim 1: a measurement of the force between two isolated, controlled charged-particle beams or traps of very different mass (e.g. electron beam passing an ion in a trap) looking for an along-track term of size (M/m)vV/c^2 times Coulomb. A literature check of existing beam-ion measurements would be the cheap version.
• Claim 2: a full kinetic (particle-in-cell or Monte-Carlo) calculation using Aspden's term in a plasma with realistic distributions, computing the net ion energy gain to second order in V. If the net is zero, the mechanism is dead; if not, compare with Roth's bumpy-torus and coronal-hole ion temperatures.
• Claim 3: neutral-lab calorimetry of a Correa PAGD device with full input and output power (true RMS and integrated product of voltage and current, including reactive power), blind to the builder.
• Claim 4: a derivation of the composition dependence of Aspden's gravitational force; then compare to the equivalence-principle bound (about 1e-15, [web: https://arxiv.org/abs/2209.15487]). Independently: obtain the page images of eq. 163 and book p.151 and settle whether g or G was fitted first.
• Claim 5: already settled by NIST/CODATA; the only live question is whether the 1843 selection argument can be stated without reference to the answer.
• Claim 6: fit the annual (+/-0.04 K) and solar-system dependence predicted by eq. 230 against FIRAS/WMAP/Planck monopole data and the T(z) measurements.

Open questions

• Does the mass-ratio term give any net energy transfer in a realistic plasma at second order, and is the answer large enough to matter?
• Why do coronal heavy ions exceed electron temperatures? (Mainstream models exist; the problem is not fully settled [memory, medium].)
• Does the Graneau fragmentation require longitudinal Ampere forces, or is it conventional? (Contested, [web].)
• What does the Correa device really output? (No neutral test found.)
• The mid-book chapters on Planck's law, the Schroedinger equation, the g-factor, geomagnetism and cosmic domains were deep-read in QE round 2 and now carry scale verdicts (Q4); what stays open is whether the geomagnetic moment formula (a striking order-of-magnitude result with R tuned) would survive a second-body test (Sun, Jupiter).
• Whether the book's p.151 admission that the 1843-quantum route "puts in doubt" the early empirical graviton derivation changes the status of the G claim has not been verified from page images.

Verification log (arbiter web budget, 8; used 7 across the checks listed)

1. Sethian PRL 40, 451 abstract: success. 2. CMB T(z): success (search summaries). 3. Galactic escape speed: success (Gaia ranges 445-580 km/s). 4. Graneau debate: partial (search summaries; papers not read). 5. Correa independent replication: partial (only proponent and enthusiast sources; no peer-reviewed independent test found). 6. MICROSCOPE equivalence-principle result: success. 7. (Reviewers' NIST/CODATA and Borexino results accepted with their URLs.) Not web-checked, still [memory]: modern Lorentz-invariance bounds; accelerator/beam-ion consistency with the Lorentz force; coronal-hole ion temperatures; the 2.26 GeV particle's later fate (re-tried in QE round 2 and failed again, Q6: [memory, unverified]).

QE revision (after qe.md, verdict MINOR)

What changed in the source-gap statement. The opening "Source gap" paragraph is overstated. Physics Unified has full Tesseract OCR in _text/ (212 pp.; checked against page images, printed page = PDF page minus 7). The real gap is the cited external papers, the Correa patents and the 1996 report. In round 1 only the Chapters 1-2 deep read existed; that revision rested on direct OCR passages and the QE's digest notes and was marked provisional. Round 2 (this file as now edited) adds the deep reads of Chapters 3, 5-9 and Appendix III, replaces every 'provisional' and 'no verdict' below with a scale verdict (Q3, Q4, Q7), and adds a Round 2 subsection. Tags as before. Web budget: the session's search allowance was exhausted, so no new [web] checks could be made (see Q6).

Q1. New evidence read in this revision

• Appendix III revises eq. 157 for lattice motion at 390 km/s: the volume per lattice particle shrinks by 2.8e-7, "The resulting evaluation is 137.035953" [read: OCR, PDF p.211, lines 9685-9693].
• Deuteron binding energy "2.22464 Mev. This is exactly the measured binding energy" [read: OCR, PDF p.135-136]. Neutron decay "1.531173 ... energy measured from the beta spectrum" [read: OCR p.140]. Proton moment 2.792846 against 2.7928456, with the remark that "the last digit of the muon mass is critical to this agreement" [read: OCR p.149]. Muon lifetime 2.1973 us against 2.197134 +/- 0.000077 [read: OCR p.153]. Note the proton-moment sentence: the agreement depends on the last digit of an input (muon mass), which is a sign of fine tuning.
• Chapter 2 deep read (deep/Unified-Physics-8f6841.md): eq. 14 = Lorentz force plus one term; eq. 36 equals the GR orbit equation; its coefficient is fixed by unforced choices (n = 1, RMS, round trip, sign) and the choice of Fig. 12 over Fig. 13 is justified by "agreement" with GR [read: p.35; calc 42.98"/century if eq. 36 is accepted].

Q2. Retardation factor 1 + 3(v/c)^2 and Mercury (claim 4, new sub-verdict)

• Mathematically, eq. 36 does reproduce the GR perihelion advance, 6 pi GM/[c^2 a (1-e^2)], 42.98"/century for Mercury [calc, deep read]. The book itself never computes 43".
• It is not an independent derivation. Changing any one choice changes the coefficient (n = 0 gives 2/3 and 28.7"/century; one-way trip 1/4; Coulomb distribution 2; sign flip -1) [calc, deep read]. The book selects the distribution because it matches GR [read: p.35]. The n = 1 choice contradicts the book's own convergence rule [deep read].
• It is the same pattern as Gerber (1898): a retardation law tuned to the perihelion, with Seeliger's objection ("not the consequence of his premises") applying in kind. Gerber's light deflection was wrong [web via deep read: https://en.wikipedia.org/wiki/Paul_Gerber, secondary]. Aspden gives no light-bending, Shapiro-delay or binary-pulsar test in these pages.
• Verdict: Aspden's "derivation of GR from retardation": Internally inconsistent / not derived as claimed (coefficient set by selection against the known answer). As a heuristic reconstruction: Historically superseded but instructive. The back-cover claim that Einstein's result is recovered "without general relativity" is therefore not supported. It matches a known answer only after tuning.
• Related Chapter 1 findings that bear on claim 1: eq. 14 equals Lorentz plus -(M/m)(v.r)V, and the closed-circuit sentence on PDF p.21 is stated backwards (only a closed source circuit kills the middle term) [calc, deep read]. P does non-zero work, so the energy argument behind the derivation is broken (soft points a-d in the deep read). This reinforces, and does not change, the claim 1 verdict "Internally inconsistent" for the derivation. The Trouton-Noble argument (1903 null result, which Aspden says supports Maxwell's Q = -1/r over Lorentz) is not conclusive: special relativity also predicts the null result, and eq. 14 itself predicts a larger non-central couple for a capacitor with ions of different mass unless the frame follows the Earth [calc, deep read; mainstream point is [memory, high]]. Sethian et al.: the factor of 10^3 against the mass ratio of about 1836 is an order-of-magnitude coincidence that the authors themselves called unjustified [read: deep read p.23]. This agrees with the existing ruling that Sethian is weak support.

Q3. Particle-mass and moment matches (claim 5): final verdicts after the Chapter 7 (pp.130-162) and Chapters 5-6/Ch.6 (pp.119-129) deep reads

Source for every row: deep/Unified-Physics-8f6841.md (Chapter 7 section and graviton/psi section), checked there against page images [read: deep read]. Modern values are [memory, high] unless stated; none was web-checked (the web search allowance was exhausted, see Q6), so the modern column is standard CODATA/PDG recall. "Choices" counts the discrete or fitted steps made with the target value known.

No number in these pages was published before its measurement; every comparison value predates the book (1971-1977), and in every case where the measurement later moved, Aspden's number stayed with the old value [read: deep read section 3]. That pattern is what fitting produces. The pooled inputs M = 1836.152, E0 = 412.666, g = 5063 and alpha are themselves outputs of earlier fits (claim 5 and claim 4), so each match inherits them.

Counting what is fitted. Beneath all rows: the integer 1843 (one fit, see section 5 and Q4), E0 = 412.67 (from 1843 and the lattice), g = 5063 (fitted to G and to hadron masses), alpha. On top of those, the individual rows add the 3-5 choices listed. A match to 6-7 digits from a formula with that many tunable pieces, with the text's own admission that "the last digit of the muon mass is critical", is evidence of fitting, not prediction. One genuinely interesting caveat: some of the arithmetic reproduces exactly (deuteron, neutron, pion, g_mu, tau, pi0 steps [calc, deep read]), so the work is not careless; it is the freedom, not the algebra, that fails.

Graviton decay table and G (the support for g = 5063). The five-row table lands on 5062-5064 only with 1960s hadron masses and an open choice of hadron count; with modern masses the rows spread to 5058-5068 (the pion row works only with the obsolete 276). Page 158 (printed 151) concedes that taking the 1843 quantum literally would allow a larger particle (5526), so "this puts in doubt the early empirical derivation of the graviton mass", and falls back on G and the psi matches. G goes as g^-8, so the 0.0004 factor equals a shift of 0.25 in g [calc, deep read]. Verdict unchanged: the 2.587 GeV graviton is Tested, not supported by the book's own report that the 1977 2.60 GeV candidate was not confirmed independently.

Correction to the alpha remark in section 5 (kept from round 1). "Under strain c.1980" holds for 137.035915 only (3.2 sigma from Williams-Olsen 137.035963(15)); the book's revised 137.035953 is 0.7 sigma from it [calc]. Gaps to CODATA 2022 137.035999177 [web, NIST, cited in section 5]: 8.42e-5 (0.614 ppm, about 3,800 sigma) for 137.035915; 4.62e-5 (0.337 ppm, about 2,100-2,200 sigma) for 137.035953 [calc]. Both are Tested, not supported.

Appendix III confirmed (pp.211-212 deep read; page image p.211). PDF p.211 is the whole of Appendix III; p.212 is the back cover. The page says: lattice motion through space at 390 km/s (the U-2 CMB figure) frees 1 lattice particle in 385; each swells by (1.154)^3 = 1.54, an increase of 0.54; with the unit cell 5060 particle volumes, the continuum volume per particle falls by 0.54/(5060 x 385) = 2.8e-7, "effectively decreasing d^3 and thereby effectively increasing the 32 pi factor" and alpha^-1 by 2.8 parts in 10^7: "The resulting evaluation is 137.035953, an important change, especially as alpha^-1 is now being measured to accuracy of this order". Each step checks arithmetically (137.0359148 x (1 + 2.772e-7) = 137.0359529 [calc, deep read]). So the text is right to call 137.035953 derived within the model, but added post hoc: it was introduced after the Williams-Olsen value, the main text (p.193) had judged the free-particle effect negligible and a footnote on that page reverses this for alpha alone, and no other lattice quantity (m_p/m_e, G, g) is revised for the same effect. In fairness, the magnitude is prescribed by the model and the measured speed, not tuned (it stops 0.67 sigma short of 137.035963); but the sign and the "d^3" step are asserted, and the +/-60 km/s on the 390 gives a +/-0.04 ppm band, so the quoted last digit is over-precise. This agrees with what round 1 said and the earlier wording ("a 2.8e-7 correction made after the measurement") stands; the amendment is that "post-hoc adjustment" should read "post-hoc, model-prescribed correction". Implied but unstated by the book: an annual modulation of alpha of about +/-0.02 ppm, untested.

Q4. Chapters formerly marked "no verdict": scale verdicts after the pp.94-118, pp.163-200 and pp.54-80 deep reads

All from deep/Unified-Physics-8f6841.md [read: deep read; equations checked against page images there]. Modern values [memory] unless stated; none re-checked on the web in this round.

Planck's law (Ch.5, eqs 92-104). The algebra is consistent. What is derived is E = h nu, not the black-body spectrum. The h is put in through Omega = m_e c^2/hbar (eq. 92) and again by tying the standard unit to m_e c^2, so h comes out because it went in; the rotor's own kinetic energy is dropped. Verdict: Not testable as stated (no prediction that differs from E = h nu).

Schroedinger equation (eqs 105-110). The algebra is consistent, but eq. (107) is the de Broglie relation lambda = h/p built in through "alpha H = I omega"; the Helmholtz-to-Schroedinger step is then the standard textbook heuristic. Only the time-independent, single-particle equation appears. The identification of alpha in (116) is circular (Bohr's formula containing alpha is inserted to get alpha out). Verdict: Not testable as stated; the alpha identification step is Internally inconsistent (circular).

Electron g-factor series. Series g = 1 + x/2 - 0.32735 x^2 + 0.2143 x^3, with x = alpha/pi (book's g is 1 + a_e). Correction to the earlier wording: the -0.32735 multiplies (alpha/pi)^2, the same convention as QED's coefficient, so the comparison is like for like. - Comparison with QED: first order, +1/2 = the Schwinger term (1948), a match that comes from the chosen cavity radius h/2m_e c. Second order, -0.32735 against the QED coefficient -0.328478965: differs by 0.34 percent (6e-9 in g). Third order, 0.2143 against QED +1.1812: off by a factor of 5.5; the book fills the gap with a gravity term phi/c^2 = 1.06e-8 = 0.846 x^3 using the Earth and Sun potentials only [calc, deep read]. - The x^2 agreement depends on two hand-picked factors (sqrt 3 and the Thomson radius 2/3): without sqrt 3, C2 = +0.25 (wrong sign); with the classical radius, -0.616. - Against data: the book's 1.001159657 was already about 23 sigma from the 1977 Penning-trap value a_e = 1159652410(200)e-12 [memory, medium]; today it is about 4e4 sigma off [calc, deep read, a_e measured to 1.3e-13]. The gravity term omits the Galaxy (about 50 times larger) and predicts an annual swing of about 1.6e-10 in g, about 1,300 times the present uncertainty; no such variation is known [memory, low]. - Verdict: Tested, not supported (strong form: a classical cavity beats QED). Weak form (a cavity picture gives alpha/2pi and a C2 near -0.33): Historically superseded but instructive. The muon g (Q3) is the same pattern.

Lattice sum r/d = 0.302874 and the integer 1843 (new in this round; bears on claim 5). The lattice arithmetic reproduces (0.302874, sum 2.3137 [calc]). But the single-site potential is conditionally convergent: the book sums in cubic shells; the Ewald convention differs by exactly pi/6 and gives r/d = 0.3357 and 1/alpha of about 151.9 instead of 137 [calc, deep read]. This conflicts with the book's own statement that boundary conditions do not matter. Verdict: Internally inconsistent (convention-dependent, and the sigma step (136)). The earlier remark in section 5 that the choice of 1843 "is by a directional argument (not verified)" is now resolved: under the book's stated rule (r/d must rise to the first odd integer with E >= 0), 1845 gives E < 0 and is excluded, so 1843 is uniquely selected given the premises (this corrects the mathematician's "1845 is nearer" point); but the premises (odd-integer rule, cubic-shell convention, factor 1/2 in m, q = e) carry the answer from 137.017 (lattice alone) to 137.0359. Claim 5 verdict unchanged: Tested, not supported; "no free parameters" Internally inconsistent.

Geomagnetism and domains (Ch.9, pp.163-181; pp.182-200). - Earth's magnetic moment mu = 16 pi e r R^5 w^2 / (15 d^3 c^2), eq. 229: algebra and numbers reproduce (7.86e25 at R = 6.45e8 cm; 8.17e25 at 6.50e8). With the true radius it gives 7.44e25, 8 percent below the 1970s 8.06e25 (modern about 7.7e25, [memory, medium]); the match is made by tuning R to 6.483e8 cm ("about 100 km up"), then refitted to about 200 km when a cos 23.5 degree factor is added (216 km [calc]). Verdict: Never properly tested (one number, one free parameter, no second body such as Jupiter or the Sun; precession, dipole tilt and secular variation unexplained; the mainstream dynamo account stands [memory, high]). The unfitted 8 percent near-miss is the interesting part. - Planetary mass ratios m/M = 3 w^2/(25 pi rho_m G k): the Sun with k = 2 gives 1/764 [calc], reproducing the book (observed 1/745 is the sum of all planets, not Jupiter, which is about 1/1047 [memory, high]); the Earth with the printed inputs gives 1/94, not the book's 1/83 (a 16 percent miss against 1/81, not 2.5 percent). k is switched between 2 and 1 per body to fit (values that match exactly: 2.05 and 0.86). Verdict: Internally inconsistent for the Earth arithmetic; Never properly tested as a law (two bodies, one two-valued parameter). The rho_0 consistency (257 and 248 g/cc against the theory's 288 [calc]) is the one non-trivial cross-check and stays Plausible, untested, since it hangs on a rounded Earth spin rate (a different estimate, 3.6e-4 rad/s, would give 196). - Domain size D = 4.6e20 cm (480 light years) is not reproduced from eq. 226 (1.44e21 cm, about 1,520 ly [calc]). - Field reversals and the 700,000-year period: Tested, not supported by the geomagnetic polarity timescale. The record has about 183 reversals in 83 million years, irregular spacing (intervals from about 1e4 yr to tens of Myr), statistical rejection of periodicity, and the Cretaceous Normal Superchron (about 121-83 Ma, no reversals) [web, via the deep read: https://en.wikipedia.org/wiki/Geomagnetic_reversal for the 183-reversal rate, the randomness, and the last reversal at about 780,000 years; superchron dates [memory, high]]. A fixed domain lattice crossed at 390 km/s predicts no superchron (38 Myr would need about 49,000 ly without a boundary). The book's "last reversal 700,000 years ago" is the 1970s date (now about 780 ka), and its Fig. 40 fit is made with a free angle, spacing and time scale, "chosen deliberately", with the 12,000-year "Stockholm" event added after a conflict (the Gothenburg excursion, not accepted as a global reversal [memory, medium]). It also requires a "few seconds" gravitational upset at each crossing, a signature nobody has found [memory, medium]. The 700,000-yr period and the 400,000-yr cube fit are quoted as one number. - Extinctions, Steiner, ball lightning, Tunguska, tornado, earthquake lights (pp.179-201): Not testable as stated (associations or v-dependent estimates; ball lightning's 8e9 J/m^3 uses a freely chosen v of 200-500 m/s). The Steiner periodicity (274 Myr / 4 = 68.5 Myr) matches one number with one free integer; the periodicity is itself contested [memory, medium]. - Cosmic background as local lattice temperature (eq. 230-231, pp.183-185): the deep read confirms the claim 6 treatment (2.56 K [calc]; "one electron in 600" closes the gap; exponent garbles in the OCR are not the book's error as far as can be told; with the same formula the Earth-Sun potential changes by 3.4 percent over the year, about 0.09 K peak to peak [calc], and the Galaxy would add about 130 K at 220 km/s). Verdicts unchanged. New minor point: the claimed "exact conformity" with Einstein's light deflection (eq. 238) is asserted without calculation, and a reduction of c by 2 phi/c^2 alone would give half the GR deflection [memory, high], so it is unverified, probably off by 2. The "38 keV" lattice quantum does not equal twice 0.0408 m_e (41.7 keV) [calc].

Chapter 3 aether experiments (pp.54-80). Only one item gives a quantitative prediction different from special relativity (SR); the rest agree with SR or are null by construction through the E-frame. - U-2 dipole 390 +/- 60 km/s (Smoot, Gorenstein, Muller, PRL 39, 898, 1977): consistent with the modern CMB dipole speed of about 370 km/s [memory, medium; 0.3 sigma [calc]] and predicted by SR plus the Doppler effect; the label "aether drift detected" is a reading. Verdict: Established as a measurement (the motion relative to the CMB frame); the aether reading is Not testable as stated. Claim 3 ruling unchanged. - Lodge, Michelson-Morley, Kennedy-Thorndike, Champeney-Isaak-Khan, Fresnel drag (Jones 1972): null or standard-SR results reproduced by building the E-frame to follow the laboratory (nested E-frames for the rotor); the Fresnel step (63) is an analogy, not a derivation. Verdict: Not testable as stated (no difference from SR predicted). - Michelson-Gale-Pearson: it is the standard Sagnac effect, which SR computes; the book's "SR is silent on rotation" is wrong [memory, high]. Established (no difference from SR). - Townes maser (1958): the book expects 0.3 km/s; by its own scaling the reported 0.05 cps daily variation limits any signal to about 0.075 km/s, below the 0.2 cps that 0.3 km/s would give. Never properly tested at the book's sensitivity; no source on file shows a repeat. - Silvertooth ruby-laser repeat: the 79-cycle figure reproduces (L/lambda x v/c at 0.5 m, 632.8 nm [calc]); SR predicts a null for any gain medium, the book predicts a nonzero but unquantified shift. The proposed test is Never properly tested (Silvertooth's own null was private communication; his later claims unreplicated as far as recalled [memory, low]). The book's cancellation argument is second order in v/c against a first-order phase shift (an internal gap). - Speed-dependent Rydberg factor 2 - gamma (eq. 56): the one real discriminator; it differs from SR by 0.25 (v/c)^4. At beta = 0.064 that is 4.2e-6 [calc], about 20 times larger than the 2e-7 precision of stored-ion Ives-Stilwell experiments (Saathoff 2003; Reinhardt 2007) [memory, medium; web check failed]. The book's own data (Snyder-Hall 0.502 +/- 0.003) test only the beta^2 coefficient, where both give 0.5, so they do not discriminate. The formula also goes negative above beta = 0.866 and its text says "increase" where it decreases. Verdict: Tested, not supported [memory, medium; not web-checked], with the caveat that the book itself only proposed the test. - Sucksmith-Bates, k = 2: the factor 2 is spin g = 2; no new prediction (Not testable as stated).

Q5. Claim 4: explicit split of the G verdict

(a) "G is derived without free parameters": Internally inconsistent (g tuned to G; extra 0.0004 factor). (b) "G numerically 6.6655e-8 at g = 5063": Tested, not supported at CODATA precision (0.1 percent low). The arithmetic of eq. 163 was rechecked by the QE [calc: 6.66546e-8]. The value of g that gives exactly CODATA G is 5062.16 by the arbiter's e/m_e = 5.273e17; the QE found example.py gives 5062.12. This is a rounding or input difference in e/m_e and does not change any verdict.

Q6. The 2.26 GeV state (web check re-run in round 2; still unverified)

The web check was re-run in this round and failed again: the WebSearch tool refused the query ("2.26 GeV narrow state 1976 1977 Fermilab reported not confirmed") because the session's search allowance was spent, and a single fallback fetch of a Wikipedia search page returned no article on a 2.26 GeV state (it surfaced only a 9.5 GeV Upsilon-type resonance at Fermilab in 1976 and unrelated 2.46 and 9.46 GeV states), which is absence of evidence from one search page, not a finding. The fate of the 2.260 GeV report, and whether independent searches confirmed it, is therefore unverified: [memory, unverified]. What the deep read recalls, all unchecked: the August 1976 Fermilab report cited by the book (via Calder, The Key to the Universe, BBC 1977, p.122 [read: deep read of p.126]) may be Knapp et al., PRL 37, 882 (1976), a photoproduction baryon-like candidate [memory, medium-low]; the charmed baryon Lambda_c+ is now about 2.286 GeV [memory, high], outside the book's quoted 2.250-2.270 GeV window by about 25 MeV. No later confirmation is known to me; I cannot cite one. The book's formula gives 2260.1 MeV (g (2/3)^(1/3), [calc, deep read]) and the book's own text does not date any publication of this value before the August 1976 report, so it counts as a retrodiction either way.

Verdict, capped at medium confidence: the 2.260 match is Historically superseded but instructive (an agreement with a 1976 report that I could not verify, [memory, unverified], medium confidence at most). It supports nothing about the graviton and nothing in the arbiter's headline verdicts depends on it. The 2.60 GeV claim is separate and stronger: the book itself says independent searches (van Apeldoorn et al., Phys. Lett. 72B, 487, 1978) failed to confirm the 1977 Apostolakis et al. resonance [read: deep read of p.126]. The verdict "2.587 GeV graviton: Tested, not supported" rests on that admission plus the absence of such a state in modern listings [memory, medium; not web-checked], and is unchanged.

Q7. Summary of verdict changes (round 1 and round 2)

Still needed: a successful web check of 2.26 GeV, of the modern particle values in Q3, of the stored-ion Ives-Stilwell precision and of the 1977 Penning-trap a_e; page-image check of eqs 166, 175, 181, 190, 198-199, 202 (the Chapter 7 deep read already checked most of these); a page-image check of p.184 (exponents in 231), pp.186-187 and p.188. The external papers and patents listed in the Source gap paragraph remain unobtained.

Round 2 (new deep reads appended in deep/)

Eight deep-read sections now exist in deep/Unified-Physics-8f6841.md: Chapters 1-2 (round 1), Appendix III (pp.211-212), graviton/decay table/psi thresholds (pp.119-129), Chapter 3 (pp.54-80), Chapter 9 remainder (pp.182-200), Chapter 9 (pp.163-181), Chapter 7 (pp.130-162) and Chapters 5-6 (pp.94-119). Web search was unavailable to every deep reader, so all modern values stay [memory].

What the new reads found. - Nothing in them is directly evidence for any claim. Every particle comparison value predates the book; every one that later moved left Aspden's number with the old value (Q3). No match was published before measurement. - Several "reproduced" numbers do not reproduce: muon mass at the quoted last digit (eq. 184 misprinted; consistent arithmetic gives 206.7690-206.7692), the 336.9 MeV quark, the Earth mass ratio (1/94 not 1/83), the domain size (1,520 ly not 480 ly), the 38 keV quantum (41.7 keV). The proton-moment match uses a muon mass of 206.7689, not the theory's 206.7688, plus an ad hoc 1/(1-2 alpha) factor. - Corrections to earlier files: the integer 1843 is uniquely selected by the book's own rule (1845 gives E < 0), so the "1845 is nearer" point and "directional argument not verified" in section 5 are superseded (Q4); the book's lattice sum is convention-dependent by pi/6 (Internally inconsistent). The g-factor coefficient -0.32735 multiplies (alpha/pi)^2, not (alpha/2pi)^2 (Q4). The observed 1/745 is the sum of all planets, not Jupiter. Appendix III confirmed as read (Q3). - The physicist's "tested" claim for Chapter 3 is partly right: the one SR-discriminating prediction (Rydberg factor 2 - gamma) is excluded by stored-ion precision [memory, medium; not web-checked]; everything else in Chapter 3 either matches SR or is null by construction, which supports the arbiter's "Not testable as stated" for the aether claim. The contrarian's request that "never properly tested" be separated is honoured for Townes and the ruby-laser Silvertooth repeat.

Verdicts that change. The headline verdicts for claims 1-4 do not change. Changes are in claim 5 and claim 6 sub-items, where "no verdict" and "provisional" are replaced by the scale verdicts in the Q7 table: derivation items (Planck, Schroedinger) Not testable as stated; g-factor Tested, not supported; r/d Internally inconsistent; the particle matches final per row; geomagnetic moment Never properly tested; mass ratios Internally inconsistent / Never properly tested; reversals Tested, not supported; Rydberg factor Tested, not supported [memory, medium]. No claim's headline verdict is upgraded in Aspden's favour. The Corrected claim list above is updated to match.

Remaining limits. Everything tagged [memory] in this section, including all modern values and the 1977 CERN and Penning-trap values, still needs a working web check. The 2.26 GeV fate stays unverified (Q6).