Brush-Charles: Arbiter's report (Step 3)

Tags: [read] = from a Brush document (as relayed by Step 1 or a reviewer who read it); [web] URL; [memory] with confidence; [calc] = my own arithmetic, redone for this report (Python/SymPy); "reviewer X says" means I did not re-check it.

Source gap (repeated from Step 1, unchanged)

Most of the folder has no text layer. Only brush-rexresearch.pdf (Nature 1911, Lodge's letter, Brush's 1929 Proc. APS summary) is machine-readable. Everything else was read by OCR of page images (1910 reprint, 1914 Discussion pp.1-16 plus, after QE, the manuscript abstract pp.17-29, 1922 pendulum paper (complete: journal pp.167-183 of a 349-page volume; the rest of the volume is other authors' work, apart from the Gregorio exchange in the Minutes), 1917 steel paper (complete), 1928 Franklin Institute address, newspapers, Brush-Miller letters with the last page unreadable). NOT obtained: manuscript pages ~17-29 of the 1914 file; Brush's 1921, 1923, 1924, 1926, 1927, 1928 Proc. APS papers (so the drop-test plates and every calorimeter number are missing; no wattage or ice-melt figure appears anywhere readable); any Bureau of Standards report; any independent replication of the pendulum or drop experiments; the 1915 steel paper. Conclusions that rest on [memory] alone: modern limits on rock radioactivity heating, quench-ageing metallurgy of steel, Miller's later aether-drift results, and every statement that no replication exists (an absence in the folder, not a documented absence in the literature).

How I worked

I read claims.md and the three reviews in full, recomputed every number where reviewers differed, re-read the OCR of Brush's pendulum paper (pp.170-174 of the volume) to settle the pendulum arithmetic, and used 5 web fetches (Added mass, Eotvos experiment, Martensite, Speed of gravity on Wikipedia; one search that was refused because the session's web-search budget is exhausted). I therefore could not check the Bureau of Standards statement on the web, and say so below.

Where the reviewers disagreed, and how I settled each

Corrections to Step 1 (consolidated)

1. claims.md claim 4: "1 part in 32,000" is a period difference (zinc vs bismuth, equal-mass long pendulums, after subtracting 22%); "1 in 4450" is a weight difference scaled by the loaded fraction from a raw 1 in 11,100. They are not the same kind of quantity.
2. claim 5: only two compounds (barium sulphate, un-ignited barium aluminate) had been in the ice calorimeter by 1929, so "no heat, no impairment" rests on very few cases [read: 1929 paper, per physicist and contrarian].
3. claim 3: the source says Brush's x term explains "part of" Mercury (contrarian); no formula or value for x is in any readable text.
4. The Bureau of Standards result: secondhand, financed by Brush (1927 Cleveland Press, via Step 1), no numbers, no report, and it concerns heat only, never the drop effect.
5. Brush compared bismuth with bismuth (different grain) and zinc with zinc (different purity) and found an effect about 2.5% as large as zinc-bismuth for the bismuth pair [read: OCR, per contrarian]. Step 1 did not record this; it matters for claim 4 (see below).

Claim 1: A falling body's energy comes from the aether, not "potential energy of position"

Agreement: all three say Brush's paradox rests on confusing zero force with zero potential, that the field view is in the textbooks, and that Lodge made the same reply in 1911 [read]. Disagreement: only the mathematician calls part of it refuted; the contrarian splits it into a weak form (Established) and a strong form. [calc] neutral point 3.46e8 m from the Earth; potential there -1.28e6 J/kg vs -6.26e7 J/kg at the surface; energy to lift one pound (0.4536 kg) = 2.78e7 J, about 11 km/s of speed. The energy has not "disappeared"; the neutral point is a saddle of the potential, so the body stays put in unstable equilibrium. The paradox is Refuted by calculation. The weak form (energy belongs to the system, not the body) is textbook, hence Established, and not a Brush departure. The strong form (a literal aether reservoir exchanging energy) has no measurable consequence and, if it implies a rest frame, runs into Michelson-Morley [memory high]. Tested status (contrarian's scheme): never tested because not testable; not dismissed either. Lodge answered it on the merits. Settling experiment/calculation: none exists for the strong form until it is tied to claim 2 or 3. Final verdict: split. Weak form Established (textbook). Brush's neutral-point paradox Refuted by calculation. Strong aether-reservoir form Not testable as stated (historically superseded as a literal aether).

Claim 2: Gravity is a push from isotropic penetrating waves; each atom casts an inverse-square shadow

Agreement: Le Sage class; no derivation in Brush; screening, drag and heating unanswered; Brush flipped from very low frequency (1911, to avoid heating) to extremely high frequency (1929, above cosmic rays) [read]; waves must be both unabsorbed (to pass through the earth) and absorbed (to push). Disagreement: verdict label. Physicist: "never tested as mechanism" + superseded; mathematician: refuted; contrarian: superseded, but notes Brush has no equations to refute. Stronger argument: the mathematician's drag calculation is generic and does not need Brush's missing equations: any push from an isotropic flux of speed c on a body moving at v gives a drag of order (v/c) x gravity [memory high on the generic argument; Feynman, Poincare, Laplace are cited at https://en.wikipedia.org/wiki/Le_Sage%27s_theory_of_gravitation, which I did not re-fetch]. [calc] 18.6 m/s per year for the Earth, against 4.5e9 years of stable orbit. Brush's only reply is "waves too weak to heat", which concerns heating, not drag. 1929 has the frequency problem: [calc] h f at 6.54e27 Hz = 4.3e-6 J, about 27 TeV per quantum. Honest limit: a Brush-specific refutation would need his coupling constants, which he never gave. What is refuted is the class of model he described, in its definite parts. Tested status: never directly tested. His 1910-11 upper-atmosphere resonance-heating attempt was "unsatisfactory" by his own account [read], an unreported null. Not dismissed without testing; it was criticised on theory (Lodge 1911 [read]). Settling calculation: derive absorption cross-section per unit mass needed for G, then compute drag and heating; if both are below observed limits the mechanism lives, if not it dies. Standard result says they are far above. Final verdict: Refuted by calculation (the push mechanism as described, via orbital drag and heating). The geometric 1/r^2 shadow picture alone is internally consistent but has no parameters. Never directly tested; historically superseded but instructive as the wave version of Le Sage.

Claim 3: Non-inverse-square force for accelerating bodies (Mercury, lunar irregularities); gravity at c does not wreck the moon's orbit

Agreement: Brush's 52 ft and 1/24,000,000 are right [calc, resolved above]; no x given; GR explains Mercury (42.98 arcsec per century of 574.10 +/- 0.65 observed [web https://en.wikipedia.org/wiki/Tests_of_general_relativity], quoted by the physicist and mathematician; I did not re-fetch); the "few seconds of arc per century" retardation is wrong. Disagreement: the mathematician's assertion that 1/24,000,000 is unreproducible (resolved: wrong), and whether the x cancellation is merely "unquantified" or "impossible": mathematician and contrarian say the secular Laplace term has the same sign every orbit while x alternates, so cancellation fails; physicist says x vanishing for circular orbits makes it a poor canceller. Ruling: all three agree on the fatal numbers. [calc] drift about 9e5 arcsec per century against observed unexplained ~1e1 or less: the uncancelled effect is not "a few seconds". Cancellation would require x to track the retardation to about 1 part in 1e5 and be of opposite sign each way; with no formula this is unfalsifiable special pleading. In GR the static field of a uniformly moving source has no first-order aberration [web https://en.wikipedia.org/wiki/Speed_of_gravity], so the Laplace objection is answered in a way that needs no Le Sage mechanism. In 1914 the Mercury anomaly was open (Le Verrier 1859, unexplained until Einstein 1915), so the motive was reasonable. Tested status: never tested (no definite prediction); the finite-speed part is refuted by calculation under Brush's own stated assumptions. Not dismissed without testing: it was never specific enough to test. Settling calculation: numerically integrate the orbit with an explicit x(r, rdot); only then could Mercury and the moon be compared. Final verdict: split. (a) "Gravity at c via a static push does no harm to the moon": Refuted by calculation. (b) x-term as Mercury/lunar fix: Not testable as stated (no value); if x is odd in radial velocity (as described), it gives no perihelion advance (conditional calculation). (c) The Mercury hope itself: Historically superseded but instructive (GR, 1915).

Claim 4: Weight is not proportional to mass (pendulum 1 in 32,000; drops 1 in 4450)

Agreement: the arithmetic is right; effect sizes are 4e-5 to 2e-4, i.e. far above every equivalence-principle bound [web: Eotvos 1e-8; modern 2e-13 to 3e-15, https://en.wikipedia.org/wiki/E%C3%B6tv%C3%B6s_experiment]; Brush's own systematics (unequal bobs 40%, air vanes about 50%, scatter, one photograph negative of six) are comparable to the effect; no independent replication in the folder. Disagreement: the physicist and mathematician give the strong form "Tested, not supported" and call Brush's own data "within systematics"; the contrarian splits it and offers a specific conventional explanation (buoyancy and added mass). I tested the contrarian's calculation. Equal-mass bobs were set on a balance in air, so apparent weights are equal, and the buoyancy term in the restoring force cancels between bobs; only the extra inertia of entrained air (added mass) remains. The period difference is then (1+k)/2 x rho_air x (1/rho_zinc - 1/rho_bismuth), with k the added-mass coefficient. [calc] rho_air = 1.2 kg/m^3, rho_zinc = 7140, rho_bismuth = 9780: for k = 0.5, 0.7, 1.0 the result is 3.40e-5, 3.86e-5, 4.54e-5. Brush's raw equal-mass effect is 4.04e-5 and his corrected one 3.15e-5 [calc from [read: OCR]]. Sign (bismuth gains), ordering (zinc, iron, bismuth: the iron density is between) and size all match, and the bismuth-bismuth control giving about 2.5% of the zinc-bismuth effect is what a density effect predicts (bismuth against bismuth differs in grain, not density). Added mass is a real, old, documented effect: Bessel introduced it in 1828 specifically for pendulums, and a sphere's coefficient is 0.5 [web https://en.wikipedia.org/wiki/Added_mass; the cylinder coefficient is not on that page, my k range 0.5-1 for a squat cylinder is [memory medium]]. Caveats: I treated the rod as massless-inertia-wise (the rod's share of the total inertia lowers the effect by a factor close to 1 but I did not compute it; Brush's own 7 g sensitivity suggests a light rod), and the geometric subtraction of 22% (centre of oscillation of the fatter zinc bob) is a separate correction I did not re-derive. So this is a match to roughly 20-40%, which makes buoyancy/added mass a viable full explanation, not a proof. Brush says nothing about buoyancy or added mass in what I could read; his only air control is drag vanes, and he declined a vacuum test ("To this I can confidently answer no" [read: OCR p.178], via the contrarian) without running it. Drop test: [calc] a 30.6 g container, final speed 4.8 m/s; drag about 1.5-2% of weight at the end by physicist and mathematician (assuming 2-2.5 cm diameter, not from the source); a drag mismatch of 0.5-1% suffices to give 9e-5. Brush himself reports lateral air currents, mutual shielding of the two containers and guide-tube friction in the first millimetre [read]. Loose powder and hygroscopic water (his barium aluminate loses several % water) are untested alternatives. Loads and containers were swapped, which controls fixed left-right bias but not a load-linked drag difference. Tested status (contrarian's three categories): the claim IS testable and was tested by others in the general form (composition-dependence of free fall at 1e-8 to 1e-15 [web], but on materials that are not Brush's samples and not on silicates specifically as far as the folder shows). Brush's own measurements: tested by him only, never replicated, never done in vacuum. Not "dismissed without testing": no refusal is documented. The NYT 1921 Webster remark "if accurate, disproves Einstein" is a conditional, not a test. Settling experiment: repeat the zinc/bismuth pendulum in vacuum (or with equal-volume bobs of different mass). Buoyancy/added mass predicts the effect vanishes; Brush's hypothesis predicts it persists. A vacuum drop tube settles the container result. Final verdict: split. (a) Strong form (weight per unit mass differs by 1e-4 to 1e-5 between ordinary materials): Tested, not supported (Eotvos 1e-8; modern torsion balances 2e-13; MICROSCOPE 3e-15). (b) Brush's own pendulum and drop results as evidence: Never properly tested (single lab, no vacuum control, no replication) with a quantitatively matching conventional explanation (air added mass) not excluded; the claimed 1/4450 is a scaled quantity (raw 1/11,100).

Claim 5: Some silicates generate heat continuously; same materials fall slow; the correlation proves a common aether-wave cause

Agreement: no power figures in anything readable; the correlation rests on two compounds run in both experiments; the Bureau report is secondhand and unquantified; the mechanism contradicts itself (1911 low frequency, 1929 high frequency; collision frequencies about 1e13 s^-1 versus 1e27 s^-1 and above) [read; calc: 6.54e27 Hz is about 27 TeV per quantum]; ordinary hydration, adsorption and slow chemistry in freshly prepared wet-way silicates can give months of small heat output [memory medium]. Disagreement: the physicist's order-of-magnitude comparison of heat power with the fall energy is garbled (it mixes units; his own text admits it); I do not use it. The contrarian's version is sound: a 1e-4 impairment on a 30.6 g, 1.1 m drop is about 3e-5 J per drop, while the heat is a continuous output of a resting body; nothing in the papers links the two energy accounts. Contrarian adds a logical point the others missed: gravity acts on all matter in proportion to mass, but only a few silicates make heat, so the heat cannot be the same channel as gravity for all matter. Radioactivity: Brush says none is "more than minutely radioactive" [read]; the contrarian's estimate (microwatts per kg for ordinary rock) is [memory medium] and was not verified; I could not check it (web search budget exhausted). Tested status: heat by Brush in his own ice calorimeters; one claimed Bureau of Standards confirmation on air-dried Sandusky clay [read, secondhand]; I could not check the Bureau claim on the web (search refused; the physicist's attempt also failed). Correlation with fall: no independent test; knowledge of which container held which load (no blinding). "Never properly tested" fits both; not "dismissed without testing". Settling experiment: blind, multi-lab calorimetry (with blanks, humidity and gas control) of his barium aluminate and nickel-cobalt silicate, with the drop test run blind on the same samples and a vacuum drop tube. Final verdict: split. (a) Heat from some silicates: Never properly tested (one unverified secondhand Bureau statement; chemical sources not excluded). (b) Correlation with fall impairment and common aether cause: Never properly tested, and as an inference unsupported; also internally inconsistent in two respects (substance-specific heat vs universal gravity; frequency reversal 1911 to 1929).

Claim 6: Recently quenched steel releases heat for days to weeks

Agreement: all three: ordinary solid-state physics, careful controls (furnace-cooled bars show no heat, oil film, heat-capacity-matched water can, Dewar flasks), about 2 K peak, calibration 55 galvanometer divisions per K [read: OCR]; not tied to gravity in the 1917 paper. Disagreement: bar mass (contrarian about 6 kg, physicist implicitly smaller, mathematician 1.5 kg); not needed, since the energy is "of order 1 J/g" in all three and unverified. The mechanism (retained-austenite transformation, carbon redistribution, early tempering of martensite at room temperature) is [memory medium]; my web check of the Martensite page [https://en.wikipedia.org/wiki/Martensite] did not confirm room-temperature ageing (it mentions only tempering with heat), so this is still a memory claim and the replication status is unknown. Brush's own later claim that it is "recognised in metallurgy" is not backed in his own text. Tested status: tested by Brush with Hadfield's steels, published in Phys. Rev. 1917 (and, per his bibliography, Proc. Roy. Soc. 1917 with Hadfield and Main, not read). Replication not documented here. Final verdict: Plausible, untested (no replication source; mechanism mainstream-compatible). Its use as evidence for the aether programme: Not testable as stated.

Claim 7: The Eotvos experiments are flawed or too insensitive

Agreement: Brush gives no error analysis; "a hundred times too small" (1926 letter, [read: OCR]) is incompatible with his own claimed 1e-4 to 1e-5 in free fall; Miller did not perform the proposed test (he knew the apparatus only from manufacturers' circulars [read: OCR]); it was neither tried and failed nor refused. Disagreement: verdict label only. Mathematician: "Refuted by calculation" for "too insensitive"; physicist and contrarian: "Tested, not supported". Both are right of different parts: [web] Eotvos reached 1e-8 and later balances 2e-13 and 3e-15, five to ten orders below the 1e-4 to 1e-5 claim, so "too insensitive" fails arithmetically. The one escape (the effect depends on grain, purity, hydration, which Eotvos's samples did not vary) is the special pleading the contrarian flagged: it makes the claim unfalsifiable unless the dependence is specified beforehand. Which materials Eotvos used is not on the page I fetched [web]; unknown. Tested status: tested by others, repeatedly and with better apparatus (not as a response to Brush); Brush's own proposed test (balance on the Mt Wilson interferometer, lead vs sulphur or selenium) was Never run. Not "dismissed without testing" on the evidence available: the sources show inaction, not refusal. Settling experiment: none beyond what exists; a torsion balance with Brush's silicate samples would close the special-pleading loophole. Final verdict: split. (a) "Too insensitive": Refuted by calculation. (b) "Flawed" (no flaw specified): Tested, not supported (successive independent torsion-balance and satellite tests confirm free-fall universality). (c) Brush's proposed test: Never properly tested (never performed).

Corrected claim list

1. Falling-body energy: the textbook field view is established; Brush's neutral-point paradox is wrong (2.78e7 J/lb remains stored); a literal aether reservoir is not testable as stated.
2. Push gravity by isotropic penetrating waves: the class of model is refuted by drag (18.6 m/s per year for Earth) and heating; Brush's version supplies no equations; never directly tested. Priority context (Round 2): Gregorio's 1892-95 ether-absorption pressure theory is a different push mechanism from Brush's isotropic wave flux; nothing in the 1922 exchange tests either.
3. Non-inverse-square correction and finite-speed gravity (force law 1/D^(2+x), x only verbally described, QE revision): the static push at c would retard the moon by about 9e5 arcsec per century, 1e4 to 1e5 times what is observed; the x term has no value and, if odd in radial velocity, gives no perihelion advance; Mercury was later explained by GR.
4. Material-dependent weight: strong form not supported by Eotvos-type tests (1e-8 to 3e-15); Brush's pendulum effect (period difference 1 in 32,000, weight difference about 1 in 16,000) matches air added mass (3.4-4.5e-5 vs raw 4.0e-5); the drop effect (raw 1 in 11,100, scaled 1 in 4450) is within drag/air-current systematics he admits; neither was ever repeated in vacuum. QE revision: general pendulum tests (Southerns 1910, Potter 1923, 3-5e-6) were 10-20 times more sensitive than Brush's effect; and Brush's own vane result shows air inertia in his apparatus. Air added mass does NOT carry over to the drop test (identical outer containers), which stays unexplained and under-documented.
5. Silicate heat and its link to fall: heat unverified (single lab; the Bureau of Standards, per the 1928 address, was only 'working ... for the purpose of checking', no result reported there; chemical sources open); link rests on two compounds and is inconsistent in mechanism.
6. Quenched steel heat: plausible ordinary metallurgy; not evidence for the aether.
7. Eotvos flawed: not supported; "too insensitive" refuted by arithmetic; Brush's own test never done.

What would settle it

• Claim 4 pendulum: swing zinc/bismuth (equal mass) in vacuum, or equal-volume bobs of different mass. Prediction for buoyancy: disappears in vacuum. Prediction for Brush: persists at about 3e-5.
• Claim 4 drop: same twin-container drop in a vacuum tube, with blind loads and many more than six photographs and a stated standard error.
• Claim 5: blind, multi-laboratory calorimetry on Brush's own barium aluminate and nickel-cobalt silicate with blanks and humidity control, plus the blind vacuum drop on the same samples.
• Claim 3: numerical orbit integration with an explicit x(r, rdot); compare with lunar laser ranging and Mercury ephemerides.
• Claim 2: derive the cross-section needed for G and compute the drag and heating; no experiment is needed.
• Claim 6: modern differential calorimetry of fresh quenched steel against furnace-cooled controls.

Open questions

• Did anyone independently repeat Brush's pendulum or drop experiments, in air or vacuum? Nothing in the folder says; I could not check.
• What did the Bureau of Standards actually measure in Sandusky clay (power, uncertainty, humidity, date)? Is the report retrievable? Web check failed (search budget used up).
• Brush's 1921, 1923, 1924, 1926, 1927 and 1928 papers and his plates: do the raw drop tables show the six-photograph scatter and any vacuum trials?
• How large is the rod's contribution to the pendulum inertia, and does the bobs' 22% geometric subtraction fully account for the diameter difference? A full reduction-to-vacuum (Bessel) calculation would turn my 20-40% agreement into a number.
• Is there a modern calorimetric measurement of heat from freshly prepared wet silicates (non-radioactive, hydration/adsorption)? Not verified.
• Were bismuth, zinc or complex silicates among the materials in Eotvos-type tests? Not confirmed [web: page lists no materials].
• Is Brush's quench-heat effect described in modern metallurgy under room-temperature ageing of martensite? [memory medium], web check failed.

Summary of final verdicts (scale terms)

Reliance on [memory] alone: modern rock-radioactivity heat levels, steel quench-ageing, Miller's later work, and the added-mass coefficient for a short cylinder (k about 0.5 to 1).

QE revision (after qe.md, MINOR)

Tags as above. New evidence comes from the deep reads in deep/ (all [read]) and two web checks. Web: the Equivalence principle page lists "Southerns 1910, pendulum, 5e-6" and "Potter 1923, pendulum, 3e-6", with references "A Determination of the Ratio of Mass to Weight for a Radioactive Substance" (Southerns, Proc. Roy. Soc. 1910) and "Some Experiments on the Proportionality of Mass and Weight" (Potter, Proc. Roy. Soc. 1923) [web https://en.wikipedia.org/wiki/Equivalence_principle]. The web-search budget was exhausted and the Royal Society pages were not available, so which materials Potter used (bismuth?), and whether he cites Brush, remain UNVERIFIED. Southerns' title shows his test was a radioactive substance against ordinary matter, not Brush's pairing.

Claim 4 (changed in part)

• (b) pendulum results. Old: "Never properly tested ... no replication". Revised: pendulum equivalence tests at 5e-6 (1910) and 3e-6 (1923) were 6 to 13 times more sensitive than Brush's period effect (about 3e-5; weight effect about 6e-5). The general claim that bobs of different material differ in weight per unit mass at 1e-5 level was therefore already contradicted by pendulum experiments. Verdict: Tested, not supported (general pendulum form); Never properly tested / not replicated (the specific zinc-bismuth pairing with Brush's apparatus), since I could not confirm bismuth in Potter. If a later check shows Potter used bismuth against a light metal, the second part also becomes Tested, not supported. Contrarian category becomes: not dismissed; contradicted by contemporaneous pendulum tests, not by a targeted replication.
• Premise confirmed [read: deep Proceedings, p.189-194]: bobs were matched by weight on a Becker balance (no buoyancy correction; "in air" is not stated but is the only reading consistent with the text). Bobs were zinc 72 mm and bismuth 61 mm diameter, both 48 mm high.
• Strengthened [read: p.199-201]: Brush writes "The effect of the air vanes on the bismuth is to slightly lengthen its period", and a face-on card vane shifts the effect by about 2e-5 to 4e-5, the size of the whole effect. Pure damping would not shift a period at first order, so air inertia is demonstrably present. His claim that in vacuum both periods would be "equally shortened" contradicts this, and the bobs differ in volume by 37%. He rejected a vacuum run by argument only ("To this I can confidently answer no").
• Refinements [read]: (i) every run used compensating vanes on the bismuth; Brush puts the uncompensated increase at about 10%, so compare my 3.4-4.5e-5 added-mass range with about 1.1 times the reported effect; this lies inside the 20-40% agreement and the verdict is unchanged. (ii) The corrected effect is not independent of length: 31-32e-6 for the long pendulums but 19.4e-6 (single) and 25.9e-6 (double) for the short ones, a genuine gravitational difference would not depend on length. (iii) The North/South exchange values differ by up to 14 min (4-7%), and the "reversal" Brush flags for the short pendulums also occurs in a long run; same-metal bobs (bismuth-bismuth, zinc-zinc) differ by about 2-2.5% of the effect, a systematic floor. (iv) The 7 g mismatch was about 1/2 per cent, not 14 per cent (OCR error); "1/35,000" (1921) was uncorrected and was restated as 1/51,000; the three-metal "closure" (1/17 + 1/13 = 1/7.36) checks timing arithmetic, not gravity. (v) The 1922 paper dates: written April 1922, printed 16 Nov 1922.
• Drop test. The air added-mass argument does NOT apply: the metals were inside identical aluminium containers of equal outer shape [read: p.203-204]. In 1922 Brush gave only an order of magnitude (about 1e-5 s, equivalent to about 4e-5 in g), no table and no trial count. The "persistent collateral phenomenon" is not identified in the 1922 text; QE's guess that it leads to the heat work is a hypothesis with no support in the readable documents. My earlier phrase "within drag/air-current systematics he admits" rests on the later 1928-29 papers and is a possible, not demonstrated, explanation. Verdict for the drop result: Never properly tested (unreplicated, no vacuum, no error analysis, cause open).
• Settling experiment unchanged: zinc/bismuth pendulum in vacuum, plus a vacuum drop tube.

Claim 5 (strengthened caution)

The 1928 address says only: "The Bureau of Standards, with a calorimeter of its own design, is working with some of these minerals for the purpose of checking my findings" [read: deep Evidence, p.6]. That is work in progress, not a confirmation. The 1927 newspaper and 1929 summary reports of a Bureau result remain second-hand. The 1928 address gives no calorimeter design, no heat figures, and names only one silicate (nickel-cobalt, aged nearly two years); radioactivity is excluded by assertion, and the ether explanation is offered "in the absence of other explanation" [read]. Also [read]: 0.0125 mm lead in a 122 cm fall equals about 1/400,000 s, arithmetically fine. Verdicts unchanged: Never properly tested; the link is unsupported and internally inconsistent.

Claim 3 (no change)

The 13-leaf manuscript (PDF pp.17-29) is an abstract of the 1914 Discussion [read: deep Discussion]. It states the force law as 1/D^(2+x) (sign convention (2+x), not (2-x)), x "very small", positive for accelerating-in-line bodies and negative when energy goes to the ether, "alternately plus and minus" for Mercury, with the Moon hoped to fit too. There is NO formula or value for x, no orbit calculation. Moreover the abstract repeats the "very few seconds of arc in a century" retardation. So my conditional perihelion calculation stands, and the sub-verdict "(b) Not testable as stated" is confirmed. Added point: because x depends on acceleration or fall velocity, not position alone, the law is non-conservative as stated.

Claim 2 (context added)

Superseded by Round 2 below. The Gregorio-Brush exchange is now read in full (Proc. APS 61 Minutes, PDF pp.328-335) and the verdict is unchanged.

Claim 7 (detail added)

Miller's reply of 2 March 1926 [read: deep Miller, p.2-3]: he knew the Eotvos apparatus only from manufacturers' circulars, said an interferometer mount "would be easy", said mercury-flotation conditions could be arranged, said he had "already begun to look up the literature", and asked for more information. Letters of 1928-29 (pp.4-7) are on other subjects and never mention the test. Verdict unchanged: Brush's proposed test Never properly tested (no record that it was run; no refusal); the absence is an absence in this folder.

Claim 6 (no change)

The APS abstract of paper II adds nothing: the Cleveland meeting list (27-28 Oct 1916) carries only the title [read: deep Spontaneous]. The web check for room-temperature heat release from fresh martensite was not possible (search budget exhausted); still [memory, medium]. Verdict stays Plausible, untested.

Claim 1 (no change)

Round 2 (new deep read: Gregorio priority letter and Brush's reply)

Source: deep read of Proc. APS vol. 61 Minutes, PDF pp.328-335, in deep/Proceedings-of-the-American-Philosophica-771a0a.md (second section). All items below are [read] from that deep read unless tagged. The same file's first section (Brush's 1922 paper) was already used in the first QE round above; its findings on claim 4 stand unchanged.

What the new read found - Dates. Gregorio's letter is dated Palermo, 20 July 1922 and was read at the Stated Meeting of 3 November 1922. Brush's reply is dated Cleveland, November 1922 and was read at the meeting of 1 December 1922 (two meetings, not one). My earlier paragraph gave only the 3 November date. - Citations. Gregorio dates his proposal to 1892 (Palermo Academy), printed 1893 ("Nuovi Strumenti fisici ..."), and re-discussed in 1914. Brush worked not from the letter but from a translation of the 1895 Atti papers by Prof. Borgerhoff of Western Reserve, especially pp.177-181. My earlier "1892 (Palermo Atti, 1895)" merged these. - Gregorio's mechanism: ether passes through matter, matter absorbs part of its energy, so ether pressure between two bodies is lower than outside and they are pushed together. Alternatives he offers: molecular rotary vibration; in 1914, a "superatom" dynamic-fluid ether. He claims priority but calls Brush's work "valuable and commendable" and does not accuse him of copying. He also raises a third-party complaint against Olinto De Pretto (1921). - Brush's mechanism: free ether energy in isotropic wave form with inverse-square intensity. He calls this second postulate "original with me" and "essential", and says it is absent from Gregorio. He gives two objections to Gregorio: no explanation of how the pressure deficit persists after absorption stops, and no source for the energy gained by falling bodies. He does not dispute that Gregorio wrote first; he disputes that the theories are the same. - The Minutes never use "Le Sage", "shadow" or "screening". My earlier statement that this "places Brush's push theory in the Le Sage family" was my interpretation, not something the Minutes say [memory, high: Le Sage's theory used directional corpuscle flux screened by matter]. Both theories are push theories, and Brush's isotropic flux is the closer to the Le Sage pattern. Gregorio's own statement that matter absorbs part of the ether's energy is the same ingredient behind the heating objection in claim 2 (see Round 1 claim 2 logic), but the Minutes do not discuss heating. - Not checked: the 1893 and 1895 originals are not in the folder, so neither side's priority statements are verified against them. No experiment by Gregorio is reported.

Verdict changes. None. Claim 2 stays "Refuted by calculation (class: drag, heating); never directly tested". The only edits are to the wording of the Gregorio context (replaced the earlier Claim 2 paragraph above, and added a short note to claim 2 in the Corrected claim list and the verdict table). Claims 1, 3, 4, 5, 6, 7 are unaffected by the new read; the first section's claim 4 findings were already folded in above.

Changed verdicts, summary

Only 4(b) changed in Round 1; Round 2 changed no verdict. 4(b) changed (split into general pendulum form: Tested, not supported; Brush's specific pairing: Never properly tested; drop result: Never properly tested). Headline verdicts for claims 1-3 and 5-7 are unchanged. Still unverified: Potter's materials and whether he cited Brush; any 1920s published response to the falling-body or heat claims.