QE (Step 4b): Lodge-O (Oliver Lodge, "The Ether of Space", North American Review, May 1908)

QE agent: Opus. Inputs read: _text/manifest.json, the single digest (digest/The-Ether-of-Space-4dab18.01.md), deep/The-Ether-of-Space-4dab18.md, facts.json, claims.md, sites.md, the research/ notes index, physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md and example.py. Spot checks were made against the extracted text (_text/The-Ether-of-Space-4dab18.txt), a page image I rendered of PDF p.5 (/tmp/qe-lodge/p-05.png, 200 dpi), and the Investigator's downloaded primary sources in research/files/. Tags: [read], [read: page image p.N], [web] URL, [memory] with confidence, [computation] = my own arithmetic in the project venv.

Verdict: MINOR

The central verdicts are well supported and will not change. The worked example is correct, and example.py runs and reproduces every number in the primer. The defects are a few arithmetic and wording slips, one wrong physics statement that leaves a point favourable to Lodge as "unresolved" (the 1920 light-bending index), and material in the article and in the Investigator's downloaded primary sources that no one read or carried into the primer. All of them can be fixed by updating facts.json, claims.md, arbiter.md (QE revision section) and primer.md, plus one deep read and three reads of research files.

---

1. Coverage

• There is one document in the manifest, it was digested in full, and nothing was skipped. No coverage defect.
• Investigator files. Ten files are in research/files/. All three reviewers read the 1909 book (Apps. 1-2), the 1907 Phil. Mag. paper (§12, App. B) and Thomson 1904 (pp.25-27). Nobody read philtrans04916537.pdf (Lodge, "Aberration Problems", Phil. Trans. A 184, 1893, 98 pp.: the primary report of the disc experiment). philtrans08143045.pdf (the 1897 abstract) was used only through the Investigator's summary. The arbiter and primer list the 1893/1897 papers as "unread by anyone" even though both are in the folder. Claim 5's disc experiment therefore rests on secondary summaries while the primary sits unread. Shortcut defect (moderate): see section 3, items M3-M4.
• The 1920 Phil. Mag. volume (philmag-1920-ser6-v39-djvu.txt) was downloaded but nobody opened it. Reading it would have settled arbiter open question 4. See A1.

2. Shortcuts

• Article pp.2-9 rest on the digest plus reviewers' direct quotes from the text layer; there was no deep read. These pages carry claims 1, 2 and 5 (stationary ether; mass is ether; kinetic elasticity; disc test). The reviewers and primer quote the text layer directly (I checked "absolutely stationary ... our standard of rest", p.8, and "4,000 times a minute", p.9 [read]). So the claims themselves are not built on digest alone. But the digest has two proven errors on these pages (below), and the numbers on pp.4-5 were never checked by anyone. Minor defect: one Sonnet deep read of pp.2-9 is requested.
• Digest errors found:
• p.5: the digest gives the disruption stress as "10^88 dynes/cm²". My page image of p.5 (journal p.727) reads 10^33 dynes per square centimetre [read: page image p.5]. That matches the mathematician's page image and the 1909 App. 1 "[10³³ c.g.s.]" [read: research/files/lodge-1909…txt line 4004].
• From p.11 on, the digest's page labels are one page late. The "Supplementary Remarks" start on PDF p.11 in the extracted text (the "singularity" passage is at text line 585, inside "=== page 11 ==="), but the digest puts them at p.12. The deep read and primer use the correct p.11. facts.json inherits some digest page numbers ("[read: p.14]" for the electron-mass argument, which the deep read also places at p.14, so that one is correct). Only the digest is affected.
• The digest gave the parallax as 1/2 arc-second. The deep read already corrected this to 1/1000".
• The density derivation was a source gap in Step 1. It is now closed: all three reviewers read Lodge's printed algebra in the 1909 App. 2 and the 1907 §12. This is good practice.
• The 1907 Phil. Mag. text is a djvu OCR with garbled exponents ("10!", "10°°", "10-%"). The reviewers reconstructed 10¹², 10³³, 3×10¹⁷ kWh and 10⁻⁹ from context and cross-checks, and each reconstruction is consistent with the surrounding arithmetic ("100 foot-lbs per atomic volume"; "a fiftieth of a wave-length" over 1,408 cm) [read; computation]. This is an acceptable shortcut, but it should be tagged as a reconstruction, not as [read]. See P3.

3. Missed material (from the digests, the article and the downloaded sources)

M1. Lodge's own escape clause, the positive charge, is dropped from the arbiter and primer. Lodge says (p.14) that the "only way out" of his density estimate would be "the discovery that the negative electron is not the real or the main matter-unit ... whereas the main mass is the more bulky positive charge", and that this would "probably land us in much the same sort of ethereal density" [read: deep read p.14]. Only the contrarian mentions it, in passing. This is now known to be the case: almost all the mass of matter is in nuclei, and a proton is about 1,836 electron masses [memory, high]. [computation] Proton mass ÷ volume with radius 0.84×10⁻¹³ cm: 1.67×10⁻²⁴ g ÷ 2.5×10⁻³⁹ cm³ ≈ 7×10¹⁴ g/cm³. Nuclear matter density is about 2.3×10¹⁴ g/cm³ [memory, high]. So Lodge's guess that the positive charge would give "much the same sort" of density is borne out in order of magnitude (10¹⁴ against 10¹²). This is pertinent to claims 2-3 and is the kind of thing the reader should weigh. The primer should state it fairly, with its limits: it is a density of matter units, not of empty space, and the radius-dependence point applies to it too.

M2. Unchecked numbers on pp.4-5 (gravitational stress). No reviewer checked these. - Moon-Earth rod: "would ... tear asunder a steel rod four hundred miles thick, with a tenacity of 30 tons per square inch" [read: p.4]. [computation] The Earth-Moon force is 1.98×10²⁰ N. A 400-mile-diameter rod at 30 long tons/in² (4.6×10⁸ Pa) holds 1.5×10²⁰ N. Consistent (within about 30%). - "Maxwell calculates that the gravitational stress near the earth ... is 3,000 times greater than what the strongest steel could stand; and near the sun it should be 2,500 times as great as that" [read: pp.4-5]. - [computation] The Maxwell stress g²/(8πG) at the Earth's surface is 5.7×10¹⁰ Pa, about 125 times a 30 tons/in² steel, not 3,000. - The Sun/Earth surface-gravity ratio squared is 780, not 2,500. - So neither figure reproduces from the obvious formula. Maxwell's own wording (1865 "Dynamical Theory", note on gravitation [memory, medium]) was not read. The figures may use a different steel strength or a different stress. This is open, not an error found. - Newton's "Optical Queries" argument (p.4): pressure lower near bodies, falling off "inversely as the distance", gives the law of gravitation. This is a correct potential-gradient statement (a pressure deficit ∝ 1/r gives a force ∝ 1/r²). It is the root of Lodge's App. 1 "tension" gravity. The primer does not mention it.

M3. The disc experiment's primary sources qualify claim 5 in three ways nobody recorded: - Speed. The 1908 article (p.9) and the 1909 book (line 3083) say "4,000 times a minute". The 1893 Phil. Trans. says "hitherto I have not spun them at more than 3000 revolutions to the minute" (the makers "recommend no speed over 4000") [read: research/files/philtrans04916537.pdf, Phil. Trans. 184 p.758]. The 1897 abstract says "whirled to higher speeds, chiefly at 3000 revolutions a minute" [read: philtrans08143045.pdf]. The primer's "up to 4,000 rpm, edge speed about 190 m/s" should say that the documented runs were chiefly at 3,000 rpm (edge speed about 144 m/s [computation]), and that 4,000 is the lecture's figure. - 1/1000 in 1893. The 1893 conclusion, quoted in 1909 ch. V, already says the ether is "unaffected ... or, if affected at all, by something less than the thousandth part". The 1/200 was the "rigorous proof" figure [read: 1909 lines 2093-2100]. The primer and arbiter present 1/1000 only as a later tightening. - Lodge's own limit on what the null means. 1897 abstract: the conclusion is that "whatever motion moving matter may confer upon the ether must be of an irrotational kind", and "It was demonstrated theoretically in the previous memoir that no optical experiments could be competent to detect motion of this latter character" [read: philtrans08143045.pdf]. So Lodge himself said his null excludes only viscous (rotational) drag, not all ether motion. This bears on verdict 5c and on his 1908 sentence "We cannot spin ether mechanically". It should appear beside the verdict.

M4. Silberstein 1920 (related work, not found by the Investigator). Lodge's 1920 note is a comment on "Dr. Silberstein's paper" on the eclipse results and the Stokes-Planck aether, in the same Phil. Mag. volume [read: philmag-1920 txt lines ~10830-11190, 11384]. Silberstein's paper derives the refractive index that reproduces "Einstein's formula ... quite apart from his theory" and notes it follows from Einstein's approximate line element. It is a second contemporary physicist's condensed-aether model of light bending. It belongs in the Investigator's related work and in the primer's "Lodge after relativity" paragraph.

M5. Smaller items from the digest that are absent from the primer. None is load-bearing. - p.2: Lodge denies action at a distance across empty space but explicitly does not deny telepathy. This is context for claim 6's psychical remark. - p.7-8: the "Unseen Universe" (Balfour Stewart and Tait) and Maxwell's tentative suggestion that the ether might be the organism of beings with life and mind. facts.json records this; the primer does not. - p.9: "sudden change of speed, not speed itself, generates ether waves" (X-rays from sudden stoppage). This is correct physics (radiation from accelerated charge; the Larmor formula [memory, high]) and could go under "What is solid". - p.9: Sir William White's report that naval steel plates are cut by a fast-spinning soft-iron disc, and Kelvin's spring balance made of spinning rigid bodies (two more kinetic-rigidity examples). - pp.13-14: the Hooke (1705) and Poynting argument that a continuum must be denser than a porous aggregate. Lodge himself says his estimate does not rest on it. The mathematician mentions it; the arbiter and primer omit it. - p.14: Lodge's argument that the ether's density must be finite because light has a finite speed ("the ether has given itself away").

4. Accuracy spot checks

Accuracy defects found:

• A1. Light-bending factor (mathematician §3.6; physicist claim 4 note; arbiter open question 4; primer §6 "unresolved"). The mathematician states that n = 1 + 2GM/(rc²) "gives half the Einstein deflection". That is wrong.
• [computation] For an index n = 1 + 2GM/(rc²), the ray deflection is ∫∇⊥n dz = 4GM/(c²b). For the Sun's limb this is 8.5×10⁻⁶ rad = 1.75″, the full Einstein value. The index 1 + GM/(rc²), using only the time part of the metric, gives the half value 0.87″ [memory, high: standard isotropic-coordinate result].
• The 1920 Phil. Mag. text agrees. Silberstein shows this index reproduces "Einstein's formula" and "would follow at once from Einstein's approximate line-element", and Lodge's note adopts it [read: philmag-1920 txt lines 11147-11175, 11374-11378].
• So Lodge's 1920 ether index does reproduce the observed light bending to first order. Puthoff's polarizable-vacuum model does the same; the Investigator's lead is correct on the first-order match [memory, medium for Puthoff's K].
• The point is favourable to Lodge. The primer's "unresolved" should be replaced with this result, with its limit: it is a fitted index, an effective description equivalent to weak-field general relativity, not an independent prediction.
• A2. "Five orders below 1e12" (arbiter "Where reviewers agree" item 6; mathematician §3 item 3). The 1893 bound ρ > 173 g/cm³ is 1e12/173 = 5.8×10⁹, about ten orders of magnitude below 10¹² [computation]. The "five orders" (factor 76,000) is the gap in drift speed, not in density. The primer (Step 8) correctly says "ten orders".
• A3. Primer Step 5, point 1: "the density comes out 1.6 × 10^10, about 200 times below 10^12". 10¹²/1.64×10¹⁰ = 61; 200 is the ratio to the 3.3×10¹² that Lodge's own radius gives [computation]. Reword as "about 60 times below 10¹² (200 times below the 3.3×10¹² his stated radius gives)".
• A4. Contrarian vacuum mismatch "35 orders" (10²⁶ J/m³). The arbiter already caught and corrected this to 41 orders. No further action.
• A5. facts.json mainstream_conflict says "the field angular momentum ... (reviewers to confirm)", and claims.md says the density "rests on this article's statement alone". Both are superseded by Step 2 and need updating (see remediation).

5. Retention check

Question: was pertinent material from the source or from the Investigator's findings left out of the arbiter or primer because a reviewer judged it negatively?

• No retention defect of the prohibited kind. The arbiter and primer keep the material that cuts against the verdicts. This includes:
• Lodge's careful nulls, Larmor's critical letters printed by Lodge, the Nature 1907 "nothing paradoxical" statement, and FitzGerald's and Larmor's dense-ether precedents.
• Kragh's "remote ancestor" and Rowlands's quote, tagged as historians' characterizations.
• spacefed.com, tagged as an interested party.
• The 1920 restatement.
• The 13-year delay in publishing the 1893 experiment, analysed under absence of evidence with three explanations and the deciding record (the Liverpool notebook).
• The ether-machine photographs and Lodge's staff (George Hall, Benjamin Davies).
• The psychical-research context (Raia, Noakes).
• Omissions that are not judgement-driven but should be fixed, because each one weakens a point in Lodge's favour:
• R1 = M1: Lodge's positive-charge escape clause (borne out in order of magnitude by nuclear density). It is in the contrarian only.
• R2 = A1/M4: the 1920 index reproduces the full Einstein deflection, and Silberstein's parallel model. Both are left as "unresolved" or "half" because a reviewer was wrong.
• R3 = M3: Lodge's own statement that optical experiments cannot detect irrotational ether motion. This is the inventor's own limit on his null, and it bears on how "we cannot spin ether mechanically" should be read.
• R4: Investigator related-work leads that reached the arbiter but not the primer: Puthoff's polarizable-vacuum model (first-order match to Lodge's index); Heaviside's 1883 perfect-liquid analogue; Barone 2004 and the Nernst/Wiechert vacuum lineage; the 1907 offprint circulated with La vie et la matière (1909). The primer's Further reading or "Lodge after relativity" should list Puthoff and Silberstein at least, tagged as related work, not as tests.

6. Process

• Source tags. Present and mostly honest throughout. Issues:
• P1. The primer sources Lodge's Birmingham principalship to "(arbiter)". That is not a source tag; use [memory, high].
• P2. The arbiter and primer list the 1893/1897 Phil. Trans. papers as "unread by anyone" while both are in research/files/. They should say "downloaded, not read" until remediation reads them.
• P3. Reconstructed exponents from the garbled 1907 OCR (10¹², 10³³, 3×10¹⁷ kWh, 10⁻⁹) should be tagged "[read: OCR garbled, reconstructed from context]" where first used.
• Web checks.
• The physicist made several checks (Nagel 2015, Planck 2018, von Laue, Dehmelt) and the contrarian made 3.
• The mathematician made none while relying on [memory, medium] items: the electron size bound, the dark-energy value and the light-bending factor. The first two were later web-checked by the arbiter (3 of 8 calls used). The third, which the mathematician got wrong (A1), was never checked even though the downloaded 1920 text settles it. This is a minor process defect.
• The arbiter's own web budget was under-used while open question 4 stayed open.
• Verdict scale. Applied consistently and with clear splits.
• The contrarian's "Refuted by calculation" for the power-station conversion was re-labelled "Internally inconsistent" by the arbiter, correctly (the slip is against Lodge's own 10³³).
• The physicist's "Internally inconsistent only in the sense of circularity" for the elasticity was resolved to "Not testable as stated", correctly.
• Verdict 5c ("Tested, not supported" for drag) is fine, but should carry Lodge's irrotational caveat (M3).
• Primer against STEP 4.
• All nine sections are present and in order. Every symbol is defined. Units (esu, emu, Gaussian) are introduced from scratch, which suits the reader profile.
• The worked example is real, step by step, and reproduced by example.py with comments.
• Five check-yourself questions, answered at the end; I verified answers 2-5 numerically.
• Textbooks are named without chapter numbers except where they are certain (Feynman II ch. 27-28 and Griffiths ch. 8 are correct [memory, high]).
• Tone is fair. No new verdicts.
• example.py runs (S7).
• Source gap. None for the article. The derivation gap was closed and reported as such in the arbiter and primer. Correct.

7. Remediation (one round)

Deep read (manifest document): 1. The-Ether-of-Space-4dab18, PDF pp.2-9, Sonnet. Question: record verbatim, with correct PDF page numbers, Lodge's statements behind claims 1, 2 and 5. Check the numbers on pp.4-5: - the 400-mile steel rod at 30 tons/in²; - Maxwell's "3,000 times" and "2,500 times as great"; - Newton's pressure ∝ 1/distance; - the p.5 footnote ("absurdly great"); - the p.9 disc passage (speed, "tested carefully", no data); - "sudden change of speed ... generates ether waves"; - the p.2 telepathy disclaimer; - the p.7-8 Unseen Universe passage.

Note every place where the digest differs (the known error is p.5 "10^88" → 10^33). Render page images for any garbled number.

Reads of Investigator files (for the remediation agent; not in the manifest):

2. research/files/philtrans04916537.pdf (Lodge, Phil. Trans. A 184, 1893, "Aberration Problems"), Sonnet. Use pp.~755-777 of the journal, around the disc-speed passage and the p.777 conclusion. Question: the speeds actually used, the bound and how it was derived, the apparatus, and the "theoretically demonstrated" claim that optical experiments cannot detect irrotational ether motion.
3. research/files/philmag-1920-ser6-v39-djvu.txt, around lines 10800-11400 (Silberstein's Stokes-Planck aether paper and Lodge's note, pp.161-174), Sonnet. Question: exactly what index each author proposes, what deflection it gives, and what Lodge adds (circulation-energy deficiency, the electron as a region of ceased circulation).
4. research/files/philtrans08143045.pdf (1897 abstract, 2 pp.), as part of item 2: quote the irrotational conclusion.

facts.json: - Correct the digest-inherited p.5 threshold (10^33). - Add that the density derivation is in Phil. Mag. 13 (1907) §12 and the 1909 App. 2 and reduces to electromagnetic mass plus w₀ = u and an assumed radius. - Add the positive-charge escape clause (p.14) to key_claims. - Record the 4,000 rpm (1908) versus 3,000 rpm (1893, 1897) discrepancy. - Replace "reviewers to confirm" (e m) with the arbiter's resolution. - Add the 10^33 p.5 threshold as a key claim.

claims.md: - Remove "rests on this article's statement alone". - Claim 3: elasticity and whirl speed are restatements of c (per arbiter). - Claim 6: "visible cosmos" is Kelvin's 1,000-pc sphere. - Claim 5: "documented runs chiefly at 3,000 rpm; Lodge's own conclusion was 'no viscous connexion', with irrotational motion undetectable optically".

arbiter.md, add a "QE revision" section: - A1: the 1920 index gives the full Einstein deflection to first order (strike "gives half" from open question 4 and close it); add Silberstein 1920 as related work. - A2: "five orders" → "about ten orders" (density); five orders is the drift gap. - M1: add Lodge's positive-charge escape clause and the nuclear-density comparison (about 10¹⁴ g/cm³), stated fairly. Verdicts 2-3 do not change: it is a density of matter units, not of empty space. - M3: add the irrotational caveat to verdict 5c and the 3,000 rpm documentation. The verdict does not change. - P2: the 1893/1897 papers are downloaded; update after item 2 is read. - No headline verdict is expected to change.

primer.md: - §5 Step 5, point 1: fix "about 200 times below 10^12" (A3). - §6 table row 5c: "chiefly 3,000 rpm (lecture says 4,000)", "1/1000 already Lodge's 1893 best estimate", and the irrotational caveat. - §6 "Lodge after relativity": replace "unresolved" with the full-deflection result and its limit (an effective index fitted to the eclipse data, equivalent to weak-field general relativity); add Silberstein 1920 and Puthoff (related work, tagged). - §7 "What is genuinely open" / "What is solid": - add the positive-charge escape clause and nuclear density (M1), with the caveat; - add "acceleration, not speed, radiates" (p.9) as a solid point; - remove the light-bending item from "open". - §1 or §6: one sentence each on the telepathy disclaimer, the Unseen Universe passage and the Hooke/Poynting argument (M5), and on the pp.4-5 gravitational-stress figures (moon rod consistent; Maxwell's 3,000×/2,500× not reproduced by g²/8πG, source unread). - Tags: P1 and P3. - Optionally add a sixth check-yourself question on the Moon-rod arithmetic (F = GMm/r² against the rod's tensile capacity), which suits the reader profile.

example.py: no change needed. Optionally add a step for the light-bending integral (4GM/(c²b) = 1.75″) and the Moon-rod check.

sites.md (next Investigator run): add Silberstein 1920 (Phil. Mag. 39:161-170) under related work; mark philtrans04916537.pdf as read once item 2 is done.

Recheck

Checked against the current arbiter.md (QE revision section), primer.md, facts.json, claims.md, example.py (re-run) and sites.md.

Residual note: mathematician.md still contains the original "gives half" wording; it is a reviewer record superseded by the arbiter correction, not a deliverable. No headline verdict changed.

Final: PASS