Arbiter (Step 3): Lodge-O (Oliver Lodge, "The Ether of Space", North American Review, May 1908)

Inputs weighed: facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, deep/, sites.md and research/. Tags: [read] document; [web] URL; [memory] with confidence; [computation] my own arithmetic (project venv, constants c = 2.998e10 cm/s, G = 6.674e-8 c.g.s.).

Plain terms. Ether: the all-pervading medium 19th-century physicists thought carried light. c.g.s.: centimetre-gram-second units (1 erg = 1e-7 J; 1 dyne/cm^2 = 1 erg/cm^3). Verdict scale: the Step 2 scale.

Source gap note

No source gap for the article itself: one primary document (a 14-page JSTOR scan), fully read as text, with pp.10-14 deep-read and page images checked [read]. The derivation of the headline density, which claims.md said was missing, was found by the Investigator (Phil. Mag. 13, April 1907, pp.488-506, and the 1909 book, Appendix 2) and all three reviewers read it. So no conclusion rests on [memory] alone for the density. Lodge's Phil. Trans. 1893/1897 disc papers are downloaded in research/files/ (the 1897 abstract and the 1893 conclusion as quoted in the 1909 book were read in the QE revision, see below; the full 98-page 1893 paper has not been read by me). Still unread by anyone: Hunt 1986, Raia 2007, Noakes 2005, the Liverpool notebooks, Science 26:482 and Sci. Am. Suppl. 64:212, Proc. Roy. Inst. 1908 (the lecture's original), and Kelvin 1901. Claims about how the work was received rest on the Investigator's citations only.

Where the reviewers agree (all three)

1. The 1907/1909 derivation exists and its algebra is right. Substituting Lodge's own line into his mass formula makes the ether density cancel: m = (2/3) e^2 / a, the classical electromagnetic mass of a charged shell. Density re-enters only through his added assumption w0 = u (circulation speed at the electron's equator equals the electron's speed) plus a chosen radius [read: 1909 App. 2; computation by all three, re-run by me below].
2. "Elasticity 1e33" is rho c^2 by construction, so it is not a second line of evidence. Mass per mm^3 (1000 tonnes) is correct.
3. The power-station energy figure is about 10 times too large for Lodge's own 1e33 erg/cm^3, and the slip is already in the 1907 paper.
4. Thomson's "e m" field angular momentum is a real result (modern form L = e g / c, Gaussian units, independent of separation); "ether spin" adds no measurable difference.
5. The disc experiment is a careful null result with bound 1/1000 (1/200 in 1893), not a measurement of anything about the ether; it is a few times above the expected Fresnel air drag (5.9e-4 per physicist [computation]) so it could not have seen ordinary effects well.
6. The 1893 solenoid experiment gives only rho > about 173 g/cm^3 ("8 times platinum"), about ten orders below 1e12 (corrected in the QE revision; five orders is the gap in drift speed, not density).
7. Kelvin's 1.5e36 tons in a 1000-parsec sphere gives 1.3e-23 times water; arithmetic right, but it is the Sun's galactic neighbourhood, not the cosmic mean.
8. No suppression or funding block is documented. Larmor's critical letters were printed in Lodge's own 1907 paper (App. B), and the paper was reported in Nature, Science and Sci. Am. Supplement.

Recomputation where reviewers differed [computation]

Web checks (budget 8; 3 calls used, remaining items stay [memory] where noted)

• Electron size. LEP Bhabha scattering: radius below 2.8e-19 m (95% CL); g-2 bounds of order 3e-19 m under a chiral-scaling assumption; lower compositeness scales 0.8 to 6 TeV [web: search summary, results include https://link.springer.com/article/10.1007/BF01475783, https://arxiv.org/pdf/2506.02450]. Retagged from [memory, medium]; exact figures depend on model assumptions but all sources agree the bound is far below 1e-15 m. Lodge's radius (5e-14 cm = 5e-16 m) is 3 to 4 orders larger than the bound. Contrarian's figure (2.8e-19 m) is supported.
• Dark energy. Planck 2018 gives Omega_Lambda = 0.6847 +/- 0.0073 and H0 about 67.4 km/s/Mpc [web: https://arxiv.org/pdf/1807.06209]. With critical density 3 H0^2/(8 pi G) = 8.6e-30 g/cm^3 [computation], dark energy is about 5.9e-30 g/cm^3 = 5.3e-9 erg/cm^3. Physicist's number confirmed.
• Kaufmann-Bucherer. Kaufmann (1901-05) believed his data supported Abraham's purely electromagnetic electron; Bucherer (1908) reported support for Lorentz-Einstein; a 1938 re-assessment found all these experiments only showed a qualitative mass increase and were too imprecise to discriminate [web: https://en.wikipedia.org/wiki/Kaufmann%E2%80%93Bucherer%E2%80%93Neumann_experiments]. This refines both physicist ("later read as favouring relativity") and mathematician ("ambiguous"): the mass-velocity data never cleanly decided between models; the electromagnetic-mass programme was displaced by relativity and by electron-size bounds, not by a decisive Kaufmann-type experiment.
• Not rechecked and not verdict-bearing: Michelson-Gale, Puthoff's polarizable-vacuum light-bending factor, proton-mass-from-QCD (high confidence [memory]), the Nagel 2015 bound (cited by the physicist with a URL).

Corrected claim list (what the claims should now say)

1. Lodge: action at a distance does not occur; matter acts on matter via a continuous, absolutely stationary ether that carries all strain and potential energy. (Unchanged. Note: "absolutely stationary" conflicts, within Lodge's own account, with ether that carries momentum as circulation; Lodge reconciles it only verbally by calling the electron a pattern, [read: 1909 App. 2], mathematician's point.)
2. Matter is made of ether: the electron is a singularity in a uniform-density ether and all mass is the ether's, carried by Faraday tubes (Thomson). The only quantitative support is the electromagnetic-mass formula, which does not involve an ether at all. Lodge adds (p.14) that if the main mass is instead the positive charge, the density would be of much the same sort; a proton-sized unit gives about 1e14 to 1e15 g/cm^3, a matter-unit density, not that of empty space (QE revision).
3. REWRITTEN. Lodge assumes electron mass m = 9.1e-28 g, effective diameter 1e-13 cm and w0 = u, and obtains an ether density of order 1e12 g/cm^3. This is the electron's surface field energy density (about 1.5e33 erg/cm^3) written in mass units, and so depends on the assumed radius as a^-3 to a^-4 (about 1e10 at the classical radius). The "elasticity of order 1e33 c.g.s." and "whirl speed of order c" restate c^2 = elasticity/density; they are not independent results. One mm^3 then holds 1000 tonnes of mass-equivalent and about 1e23 J of energy; Lodge's "million-hp station for forty million years" is about 10 times too large for that.
4. Thomson's e m: the field of a charge e and a pole m carries angular momentum e m (L = e g / c in Gaussian units, separation-independent). Lodge reads this as ether spin and Faraday's magnetic rotation as gyrostatic evidence; E, H, V are three perpendicular vectors (cross products) possibly connected with three-dimensional space (posed as a question).
5. A moving fluid can imitate rigidity (Kelvin); Lodge's discs (1 yard; 4,000 rpm in the 1908 lecture, edge speed about 190 m/s; the 1893 and 1897 papers document runs chiefly at 3,000 rpm, about 144 m/s) showed no effect on light: "quite unaffected ... or, if affected at all, by something less than the thousandth part" with 1/200 as the figure he would assert as rigorous proof, both already in Phil. Trans. 184 (1893) p.777 [read: 1909 ch. V]. The 1908 article states the null without data. Lodge's own conclusion in 1897 was "no viscous connexion": whatever motion matter gives the ether "must be of an irrotational kind", which he says no optical experiment could detect [read: Proc. R. Soc. 1897 abstract, research/files/philtrans08143045.pdf]. The 1908 sentence "ether cannot be spun mechanically" is therefore stronger than the bound and than Lodge's own 1897 wording.
6. "Visible cosmos" density about 1e-23 of water (Kelvin: 1.5e36 tons in a 1000-pc sphere, a Galactic-neighbourhood scale; the present cosmic mean matter density is about 3e-30 to 4e-31 g/cm^3 [memory, high]). The ether is argued to be immensely denser than matter; tentative remark that it has psychical significance.

Final verdicts per claim

Claim 1: no action at a distance; stationary ether

• Agreement: locality is established; a material stationary ether is superseded. Disagreement: contrarian splits the stationary-ether bullet on whether time dilation (Kennedy-Thorndike, Ives-Stilwell) is added; physicist notes Lorentz's ether is empirically equivalent to relativity; mathematician calls the experimental question not separable.
• Which argument is stronger: the contrarian's. Lodge's 1909 route (stationary ether plus FitzGerald-Lorentz contraction) is not refuted by Michelson-Morley alone [read per contrarian: 1909 ch. IV]; it is tested by Kennedy-Thorndike/Ives-Stilwell, which need Lorentz's full transformation including time dilation, and then the ether has no measurable property. Modern rotating-cavity bound about 1e-18 [web per physicist: https://www.nature.com/articles/ncomms9174, not rechecked by me].
• Final verdict (split): Local, field-mediated interaction: Established. Material, mechanical stationary ether: Historically superseded but instructive. "Absolutely stationary" ether with Lorentz's transformations added: Not testable as stated (empirically equivalent to relativity); with only the contraction hypothesis Lodge cites: Tested, not supported.
• What would settle it: nothing remains to settle experimentally for the Lorentz-equivalent version; for any version that predicts a velocity-dependent effect, a more sensitive Kennedy-Thorndike-type cavity comparison (already at about 1e-18) is the test.

Claim 2: matter is ether; all mass is the ether's

• Agreement: the algebra yields classical electromagnetic mass and no more; the 4/3 problem is unmentioned by Lodge; the electron is point-like to below 1e-18 m. Disagreement: physicist and contrarian say "Tested, not supported"; mathematician says "Historically superseded" for the EM-mass part and "Not testable" for the fluid part.
• Stronger argument: the physicist/contrarian split is more precise, with one sharpening from the web check. The programme was not killed by Kaufmann-Bucherer (which were too imprecise [web]); it is not supported because (a) the electron has no size of 1e-13 cm to hold its field energy [web], and (b) for the electron the mass is not set by field energy in the Standard Model [memory, high]. A fair echo: most of a proton's mass is field (QCD) energy [memory, high], so "mass is field energy" survives in a different theory and is not Lodge's claim.
• Final verdict (split): All electron mass is electromagnetic/ether-field energy of a ball of about 1e-13 cm: Tested, not supported, and Historically superseded but instructive. Electron as an ether singularity in a uniform plenum: Not testable as stated (no pattern defined; Lodge 1920: how knots are tied "remains to be discovered" [web via Investigator: https://archive.org/details/londonedinburg6391920lond, not opened by me]).
• What would settle it: already largely settled for the strong form by scattering bounds. For the "singularity" form, a model that yields a dimensionless number (e.g. the fine-structure constant) from the ether pattern would be a test; none exists.

Claim 3: ether density 1e12 g/cm^3, elasticity 1e33, whirl about c, energy per mm^3

• Disagreement is real here, with three different verdict labels. Physicist: Refuted by calculation for a gravitating/inertial vacuum medium of that density; contrarian: Tested, not supported and Internally inconsistent; mathematician: Not testable as stated.
• Weighing. Each reviewer is right about a different component, and the claim splits cleanly:
• (i) The arithmetic from rho = 1e12: correct, except the station conversion (about 10 times).
• (ii) The inference "rho is about 1e12" is not a determination: it moves with the assumed radius (1e10 to 1e12 over 1900s models; unbounded as a goes to 0) and with w0 = u; the electron-size bound makes it larger still, not smaller [web]. This is not "internally inconsistent" in the strict sense (each step is valid); the contrarian's label is too strong. It is "Not testable as stated" as a property of space, because a free parameter (the radius) that is unmeasurable in his model sets it.
• (iii) "Elasticity 1e33" and "whirl about c": restatements of c, no independent content; Not testable as stated.
• (iv) As physics of the real vacuum: if this energy density gravitates (every form of energy does in general relativity [memory, high]) it is about 1.9e41 times the measured dark-energy density and gives an expansion rate H of about 780 per second [computation, with Planck values from the web check above]. That follows from accepted laws with no new experiment, so I accept the physicist's label for this component. Lodge's escape (ether mass "not felt mechanically" and gravity as ether tension) is a stated assumption, not a derivation, and I record it fairly: it does not contradict his own theory, but it removes the very inertia he attributes to the density, and general relativity has no energy with inertia and no weight. The "dark energy" mismatch is also the usual size of the zero-point-energy problem, so Kragh's "remote ancestor" is a statement of resemblance (cutoff-dependent dimensional analysis: hbar c / r_e^4 about 5e33 erg/cm^3 [computation by physicist]), not a test.
• Final verdict (split):
• Arithmetic from the assumed density (1000 tonnes/mm^3; rho c^2 = 9e32): consistent; Established as arithmetic.
• Power-station comparison: Internally inconsistent (about 10 times too large for his own 1e33 erg/cm^3; and 1909 says 1e29 erg in the same sentence as 3e11 kW centuries).
• Density 1e12 as a property of space: Not testable as stated (radius-dependent; unbounded).
• Elasticity and whirl speed about c: Not testable as stated.
• Such a density as a gravitating, inertia-carrying energy of the real vacuum: Refuted by calculation.
• The claim that inertia follows from this density: Never properly tested in the article (Lodge: "if I had time I would show"); the 1907/1909 version reduces to electromagnetic mass with the density cancelled.
• Potential-vs-stress comparison (App. 1): Internally inconsistent (units).
• Note for fairness: Larmor and FitzGerald thought dense ethers possible (Larmor even suggested a larger density), and Nature (1907) wrote "nothing paradoxical, nor ... improbable" [web via Investigator: https://archive.org/details/paper-doi-10_1038_075519a0]. These are statements of possibility, not measurements. Lodge was within the mainstream of ether physics of his day.
• What would settle it: for the radius-dependence, one calculation, which I have done above (rho versus a). For the gravity question, the calculation is the check; a direct test is impossible because the ether density has no measurement (the only experiment, 1893, bounds it from below by 173 g/cm^3; the model's predicted drift at 1400 gauss is 4e-4 cm/s versus a sensitivity of 30 cm/s, about 75,000 times too coarse [computation per mathematician; I confirmed the 173 g/cm^3 bound arithmetic only]).

Claim 4: Thomson's e m spin; Faraday rotation; E, H, V

• Agreement: all three accept e m as a correct field result; "ether spin" is not distinguishable from field angular momentum; the cross-product rule is established.
• Disagreement: only on the c factor (resolved above) and on whether "Established" applies given no monopole has been observed. Stronger: the contrarian's and physicist's, which call it Established as a theorem of Maxwell theory without experimental realisation. Dirac's later condition (e g / c = n hbar / 2, [memory, high]) is the closest experimental hook, but the monopole has not been found [memory, high].
• Final verdict (split): Field angular momentum of a charge plus a pole equals e g / c (Lodge's "e m", with c needed unless charge is in electromagnetic units): Established as a theorem of Maxwell's electrodynamics, not experimentally realised (no isolated magnetic pole). Faraday's rotation and the right-hand E, H, V rule as electromagnetism: Established. Ether spin as the physical content: Not testable as stated. E, H, V underlie the three dimensions of space: Not testable as stated (mathematically suggestive only in that the vector cross product exists in 3 and 7 dimensions [memory, high per mathematician]).
• What would settle it: a measured charge-pole (or real magnet-charge) angular momentum would confirm the field result; nothing would test "ether spin" separately from it.

Claim 5: kinetic elasticity; discs

• Agreement: kinetic elasticity is real fluid mechanics; the discs gave a careful null. Disagreement: contrarian "Never properly tested" for a fluid ether making matter; physicist "Historically superseded"; mathematician "Never properly tested" for the model and "Tested, not supported" for drag. Stronger: the contrarian and mathematician, since no model reproducing Maxwell's equations plus electrons was ever built (a plain rotating fluid's inertial waves are dispersive [memory, high per mathematician]), so there is nothing to test or to refute; but the programme was superseded, so both labels apply to different bullets.
• Final verdict (split): Kinetic rigidity in fluids (gyrostat, jet, vortex ring): Established. A perfect-fluid ether that makes matter and light: Never properly tested (a programme left unfinished, not dismissed without testing; Lodge in 1920 conceded the mechanism "remains to be discovered"); also Historically superseded but instructive. Lodge's disc experiment as a test of ether drag (bound 1/1000, 1893 bound 1/200): Tested, not supported (null, careful, consistent with all later results; not independently repeated in that geometry as far as the reviewers found; Hunt 1986 unread). "The ether cannot be spun mechanically": stronger than the bound supports.
• What would settle it: nothing needed for the null (modern Sagnac and cavity tests are far more sensitive); for the model, a derivation of Maxwell's equations and Lorentz invariance from the fluid.

Claim 6: cosmos density, ether denser than matter, psychical significance

• Agreement: the arithmetic is right; Kelvin's sphere is local, not the cosmic mean (about 7e-24 g/cm^3 local stellar plus gas versus about 3e-30 cosmic mean, [memory, high/medium]). The 25 % low Kelvin solar mass is high-confidence [memory].
• Final verdict (split): Cosmos density about 1e-23 of water from Kelvin's inputs: Established as arithmetic (and a fair local galactic value; not the cosmic mean). Ether immensely denser than matter: Not testable as stated (inherits claim 3; the contrarian's "Plausible, untested" is not adopted because its premise is not a measured quantity). Psychical significance of the ether: Not testable as stated; offered tentatively by Lodge and not tested by anyone for the ether specifically; Lodge was president of the Society for Psychical Research 1901-03 [memory, high].
• What would settle it: nothing for the ratio until an ether density can be measured; the psychical remark would need a defined, measurable mind-matter coupling.

What would settle it (summary)

Absence-of-evidence analysis (where the record goes dark)

Three places where the record goes silent.

A. The 1892-93 magnetic ether-drift experiment, unpublished for 13 years. Lodge, 1907: "This experiment has not been published, and in view of the above estimate its details ought to be recorded"; details "only partially recorded" [read via Investigator and contrarian: Phil. Mag. 13 (1907) p.493 and Part II]. Competing explanations: 1. A null judged not worth reporting in 1893. For: Lodge's own habit of reporting disc nulls with bounds; the result bounded rho only above 173 g/cm^3, which was unremarkable then. Against: Larmor's letters of Oct and Nov 1893 asked for exactly this test, so he knew others were interested. 2. Held until the density estimate gave it meaning (1907). For: Lodge's own sentence says so. Against: the estimate does not make the null more meaningful (a result bounding rho only above 173 g/cm^3 says nothing about 1e12), as the mathematician shows; but Lodge may have believed otherwise. 3. Details lost or partial. For: "only partially recorded" (his words). Against: he still gives the numbers (four 7,000-turn solenoids, 1/50 band, 1400 gauss). Nothing in the record suggests suppression by others (he was the author and the owner of the data). What would decide: the Liverpool notebook for November 1893 (30 notebooks, 1874-1913, in the Special Collections and Archives; not yet opened), the Larmor letters of 28 Oct and 1 Nov 1893, and Davies's or Hall's recollections. This is a modest anomaly, mostly a history-of-science question, not a physics anomaly.

B. Reception after 1908, and after relativity. The Investigator found contemporary reports (Nature 75:519, Science 26:482, Sci. Am. Suppl. 64:212) but I found no independent reply or test of the density number, and none of the density-derived consequences (drift 0.003 cm/s at 12,000 gauss, wave amplitude 1e-17 wavelength) was tested beyond his own runs. Lodge restated the same numbers in 1920 [web via Investigator]. Explanations: (1) ether physics was displaced by relativity before anyone designed a test (for: the timing, 1905 to 1919; against: the tests of the predicted drift were feasible and cheap); (2) the density is not testable as formulated (for: reviewers' conclusions above, the ether "cannot be moved by mechanical means" [read: 1909 App. 2]); (3) tests were made and not preserved (no evidence found; unexamined). Records that would decide: the Proc. Roy. Inst. 1908 discussion, Birmingham XOJL correspondence (over 2,000 letters, not searched), The Times 1919 letter, Hunt 1986. Not searched because of the Investigator's time choice, not because of a block.

C. The ether machine and its fate. The Liverpool guide lists two photographs of the Ether Machine with Lodge, George Hall and Benjamin Davies [web via Investigator: https://libguides.liverpool.ac.uk/library/sca/oliverlodge]. No reviewer found what became of the apparatus or whether anyone repeated it; the Chronicling America search was blocked. Plausible explanations are routine (equipment reused or scrapped after the null results; a test no longer needed after 1905), and no evidence of concealment exists. Deciding records: the Liverpool finding aid, the Wiley Digital Archives of the notebooks, the Hunt article.

No funding block or publication block is documented for Lodge, who was prominent (Principal of Birmingham University, former Liverpool professor); the notable absence is of independent replications and of any experiment aimed at his density, not of suppression.

Open questions

1. What caused the factor-of-10 error in the energy conversion (1907, 1908, 1909)? Is it a slip, or a different intended quantity (the 1909 text gives 1e29 erg in the same sentence)? The Liverpool notebooks might show.
2. Did Lodge or others ever calculate the drift his density predicts (4e-4 cm/s at 1400 gauss) and design a sensitive test?
3. What does the printed "eight light years" mean (27 light years = 8.3 parsecs)?
4. Is there a defensible reading of the ether's energy as not gravitating? (Light-bending part CLOSED in the QE revision: Lodge's 1920 index n = 1 + 2GM/rc^2 gives the full first-order Einstein deflection, 1.75 arc-seconds at the Sun's limb [computation]; the earlier "gives half" was wrong. The not-gravitating question remains open.)
5. What do Hunt 1986, Raia 2007 and Noakes 2005 add about the reception and the psychical link?
6. Which units did Thomson use for e m in 1904? (The text does not say.)

Corrections to earlier steps (to carry forward)

• claims.md header: density does not "rest on this article's statement alone"; claim 3 should say elasticity and whirl speed are restatements of c; claim 6 "visible cosmos" is a 1000-pc sphere.
• Deep read: "e m leaves out c consistent with mixed units" is wrong (mixing units requires c); "source of the 10x energy factor unknown": located as a conversion slip in the station comparison; p.5 exponent resolved to 1e33 by page image (journal p.727) and by 1909 Appendix 1.
• Contrarian: the 35-orders vacuum comparison should be 41 orders (J/m^3 conversion error); physicist: unchanged.
• Kragh and spacefed "10,000 tons/cm^3" is about 100 times too low versus Lodge's own 1000 tonnes/mm^3 (1e6 tonnes/cm^3). spacefed.com is an interested party advocating zero-point-energy technology; its framing of Lodge is its own.

Headline verdicts

1. Locality: Established; material stationary ether: Historically superseded but instructive (Lorentz-equivalent form Not testable as stated).
2. All mass is ether/EM: Tested, not supported; singularity picture Not testable as stated.
3. Density 1e12: arithmetic consistent; value Not testable as stated (radius-dependent); station comparison Internally inconsistent; as a gravitating vacuum density Refuted by calculation; inertia derivation Never properly tested.
4. e m: Established (field result; units need c); ether spin Not testable as stated.
5. Kinetic elasticity: Established; fluid-ether matter Never properly tested and Historically superseded; disc null Tested, not supported.
6. Cosmos density arithmetic: Established; ether denser than matter and psychical significance: Not testable as stated.

QE revision (after qe.md, verdict MINOR)

Inputs added: qe.md; the new second deep read in deep/The-Ether-of-Space-4dab18.md (article pp.2-9, section "Deep read (second section)"); and my own checks of the 1909 book ch. V, the 1897 Proc. R. Soc. abstract (philtrans08143045.pdf) and the 1920 Phil. Mag. text (philmag-1920-ser6-v39-djvu.txt, line ~11374), all in research/files/. No new web calls were needed. No headline verdict changes. Six items change wording, one open question closes, and several pieces of material that cut in Lodge's favour, which the earlier arbiter text lacked, are now recorded.

Corrections

New material recorded (this cuts both ways; stated fairly)

1. Lodge's escape clause: the positive charge (article p.14). He says the only way out of his density estimate would be the discovery that the negative electron is not the main matter-unit and that the main mass is the "more bulky positive charge", which would "probably land us in much the same sort of ethereal density" [read: deep read, p.14]. Check [computation, constants memory high]: a proton mass 1.67e-24 g in a sphere of radius 0.84e-13 cm is about 6.7e14 g/cm^3; nuclear matter is about 2.3e14 g/cm^3 [memory, high]. So his order-of-magnitude guess for a positively charged mass carrier was borne out in the sense that a "matter unit" density of 1e14 to 1e15 is real, and the proton is about 1,836 electron masses [memory, high]. Limits, equally pertinent: (a) it is the density of matter inside a nucleon, not of empty space; (b) the radius-dependence argument (rho goes as a^-3) applies to it unchanged; (c) it does not support an ether of uniform density with the electron as a singularity, and the electron-size bound still removes the electron version of the argument. Verdicts 2 and 3 do not change.
2. Lodge's own limit on the disc null. See A3/A4 and the corrected claim 5 above. This is Lodge's own statement that optical experiments cannot detect an irrotational motion of the ether. Verdict 5c stays Tested, not supported for viscous (rotational) drag; it should not be read as excluding every motion of the ether. The 1908 sentence "We cannot spin ether mechanically" is a stronger statement than Lodge's own 1897 conclusion.
3. Article pp.2-9 now read verbatim (second deep read), with the following results: - The p.5 footnote to the gravitational-disruption idea: "On doing the arithmetic, however, I find the necessary concentration absurdly great, showing that such a mass is quite insufficient." Lodge himself rejects his own disruption idea with an arithmetic he does not show. He says "I do not suppose that this can be the reason". I record this as his own hedge, not a reviewer's. - His further hedge at p.5: gravitation is too little understood "to adduce it as the strongest argument in support of the existence of an ether"; the "oldest valid and conclusive requisition" is the wave theory of light. - The 400-mile steel rod holds: Earth-Moon force 1.98e20 N against 1.5e20 N for a 400-mile-diameter rod at 30 long tons/in^2; a diameter of about 458 miles matches exactly. Consistent within about 30% in force [computation]. - Maxwell's "3,000 times" and "2,500 times" are not reproduced. My check: the Maxwell stress g^2/(8 pi G) is 5.7e10 Pa at the Earth's surface, about 124 times a 30 ton/in^2 steel (4.6e8 Pa), and at the Sun's surface 4.5e13 Pa, about 97,000 times steel (780 times the Earth value; Lodge's 2,500 does not match under either reading of "as great as that", which could mean 2,500 times steel or 2,500 times the Earth figure) [computation]. Maxwell's own lecture was not read; his figure may use a different stress or strength, so this is an open item, not an error found in Lodge. It is not verdict-bearing. - Newton's "diminution of pressure inversely as the distance" gives gravitation: a pressure deficit proportional to 1/r has a gradient proportional to 1/r^2 [computation]. This is the root of Lodge's tension picture of gravity (App. 1) and makes his later energy-density comparison with gravity (claim 3, component iv) a deliberate step, not an aside. - p.9: "Not speed, but sudden change of speed, is the necessary condition for generating-waves in the ether by electricity" is correct physics (radiation comes from accelerated charge) [memory, high]. - p.2 telepathy disclaimer: denying action at a distance across empty space is "not denying telepathy or other activities of a non-physical kind"; no experiment or witness named. pp.7-8: the ether as possible "material organism of beings exercising functions of life and mind", quoted from Maxwell via The Unseen Universe, which Lodge calls "too speculative". These give the context for claim 6's psychical remark, still Not testable as stated. - Digest errors found (10^88; page labels one late from p.11; omitted footnote) do not reach any verdict.
4. Related work. Silberstein 1920 (Phil. Mag. 39, a Stokes-Planck condensed-ether model reproducing the Einstein deflection), published beside Lodge's 1920 note, is a second contemporary ether model of light bending. Puthoff's polarizable-vacuum model is the later parallel (first-order match; [memory, medium]). I characterise both as effective-index descriptions equivalent to weak-field general relativity, not as independent tests of an ether. The exact index in Silberstein's equation (9) is garbled in the OCR and I did not resolve it; Lodge's own formula is legible.

Verdict table after the QE revision

Retention and process notes

• Nothing from the Investigator's leads was removed. Items previously lost from the arbiter are now present: the positive-charge clause, the irrotational caveat, the 1920 full-deflection result, Silberstein, the p.5 footnote and the p.2 and pp.7-8 context.
• Still not read by me: the full 1893 Phil. Trans. paper (98 pp.), Maxwell's lecture and the Hunt, Raia and Noakes papers. The A3 rpm discrepancy rests on QE's quotation of the 1893 paper and my reading of the 1897 abstract.
• Tags: exponents taken from the garbled 1907 OCR (1e12, 1e33, 3e17 kWh, 1e-9) are reconstructions from context and cross-checks, not direct reads [read: OCR garbled, reconstructed]; Lodge's Birmingham principalship is [memory, high].