Thompson-JJ (J. J. Thomson): Science Contrarian review (STEP 2)

Reviewer: Science Contrarian (evidence over consensus; strongest honest case; same rigour on the inventor and on the critics). Date 2026-10-09. Source tags: [read] project documents (via facts.json, deep/ reads, digests, and the Investigator's notes), [web] URL given, [memory] with confidence. Not-read items are named.

Reading basis: claims.md, facts.json, all deep/ reads (Corpuscular Theory ch. II-III and VII; Conduction ch. XXI; Electricity and Matter ch. V; Positive Rays pp.106-131 printed; Roentgen Rays memoirs; 1896 note; Vortex Rings II-III), sites.md "Leads for review" and research/ notes. research/files/ is empty (no downloads were made), so the Investigator's leads are web summaries, not documents I could read. Thomson's 1913 Nature letter, the 1922 Proc. Roy. Soc. paper, and the Strutt, Merton and Egerton papers were NOT read; statements about them rest on secondary summaries and are marked so.

Web budget used: 4 searches (Kaufmann/Bucherer; Paneth-Peters and Wendt-Irion; Strutt 1914 title; vortex polygon stability). Medium/low [memory] facts that no verdict depends on are left tagged [memory].

Special features of this project. J. J. Thomson was a mainstream Nobel laureate (1906), not an outsider. The contrarian question here is therefore not "was he suppressed?" but (i) which of his claims are experimentally solid, (ii) where his own reasoning was weaker than the textbook retelling suggests, and (iii) whether the one episode where he stood against a cluster of transmutation claims (1913 helium/neon), and the one where he stood for an anomalous result (X3 / mass-3 line), were tested properly.

---

Corrections to Step 1

1. Kaufmann comparison: the columns of the 1907 table are swapped in the bottom two rows (new). The Corpuscular Theory deep read (section 3) reads the 1907 table as "Kaufmann ratio 3.09, 2.43, 2.04, 1.83, 1.65 | Thomson alpha 3.1, 2.42, 2.0, 1.66, 1.5" and concludes that at the low speeds Thomson's alpha is "about 10% low" against Kaufmann. The Conduction deep read (Art. 354, 1906) gives the same numbers with the opposite labels: observed p' = 3.1, 2.42, 2.0, 1.66, 1.5; Thomson's calculated p = 3.09, 2.43, 2.04, 1.83, 1.65. Kaufmann's own raw e/m values in Conduction Table 1 settle which is right [read: Conduction p.667-668]: with a slow-speed e/m of about 1.95 x 10^7, e/m = 1.17 at v = 2.48 and 1.31 at 2.36 give ratios 1.67 and 1.49, i.e. the observed values are 1.66 and 1.5 [my arithmetic]. So the 1907 book prints the two bottom rows with the columns interchanged, and the correct statement is: Thomson's calculated mass ratio is about 10% HIGH (1.83 vs 1.66; 1.65 vs 1.5), not low, at the two lowest speeds. The conclusion "second and third columns almost identical" holds to 1-2% at the top three speeds and fails by about 10% at the lowest two. Claims.md already lists the Corpuscular Theory misprints but not this one; the Corpuscular Theory deep read's "Findings" paragraph and item 3 of its error summary should be corrected.
2. Thomson's alpha/p disagrees with his own model at every speed. Both deep reads independently find that an exact integral of his stated model (point charge outside a sphere of radius a) gives 2.87, 2.15, 1.82, 1.64, 1.51 (at 2.85...2.36 x 10^10 cm/s), while his tabulated values are 3.09-3.1, 2.43, 2.04, 1.83, 1.65. The exact-model numbers match the OBSERVED values at the two low speeds (1.64, 1.51 vs 1.66, 1.5), not his own calculated column. The printed Conduction equation (1) for the field momentum R does not reduce to its stated slow limit (the second bracket term is wrong as printed). Combined with the swap in item 1, the fair summary is: the 1906-07 "agreement" of calculated and observed mass ratios cannot be reproduced from the printed formulas; the observed rise in mass is real, the quantitative agreement is not demonstrated by the text. I confirm the deep readers' reading and add item 1.
3. Claim 4: "supported by three counting methods" overstates. Deep read of Chapter VII shows: Method I (Barkla scattering) gives 21.5 corpuscles per air molecule with Thomson's own constants (he rounds to 25 to approach nitrogen's 28), and about 14 with modern e [calc in deep read], that is about A/2 and not A; Method II gives P/w of 2 to 4 per unit atomic weight from Becker's data, 0.27 from Crowther's, 0.01-0.02 from beta-ray values, a spread of two to three orders of magnitude which Thomson explains away by the definition of lambda; Method III gives about 1 for hydrogen but Drude's refraction counts give only a few per atom for heavy atoms. Thomson himself states only "of the same order as the atomic weight". Claims.md should say "three methods said to give a number of the order of the atomic weight; none gives it precisely, and the best (Barkla) modern reanalysis gives about half".
4. Claim 2 wording. "rising with speed as Kaufmann's radium data are said to show" is right, but the claim should note that the same data fit Lorentz-Einstein gamma about as well as Thomson's alpha (within 2-4%) [read: deep, section 3 table]. Claims.md omits that Thomson's text never names Abraham, Lorentz or Einstein.
5. Claim 5 wording is confirmed. "does not claim is proved" is accurate: [read: Positive Rays printed p.130] "few will be prepared to accept it until every other explanation has been shown to be untenable". I add that his book lists three explanations, not one, and ranks none as established.
6. Claim 6 / vortex rings. Step 1's error list (eq. 71 off by 3; eq. 74* k^4; seven vortices) stands. Web check on the seven-vortex point: the n-vortex ring on a circle is linearly stable for n up to 6 and unstable above 7; n = 7 is the marginal case [web: https://arxiv.org/pdf/2008.00365 and the Kurakin papers found at https://rcd.ics.org.ru/S1560354710010028/ ; the search summary credits Havelock 1931 for n >= 7 unstable and Kurakin and Yudovich 2002 for settling n = 7; I did not open the papers, so the exact n = 7 classification is [web, summary only]]. This is consistent with the deep reader's "seven is neutral at linear order, eight unstable" and with a mathematical result of Thomson's (up to 6 stable) that stood.
7. Claim 1 should be split. Claims.md bundles (a) same particle from all sources, e/m about 10^7, (b) a constituent of all atoms, (c) Zeeman-line emitters. Each is established; but Thomson's accompanying "number per atom about equal to atomic weight" belongs only to claim 4 and was wrong. Keep the split.
8. Step 1's "Source gap" note: none is needed for the claims themselves (texts are readable); but Thomson's 1913 Nature letter, the 1914 Proc. A papers by others and his 1922 paper are not in the folder and are not read by me.

---

Claim 1: corpuscles (electrons): same everywhere, e/m about 10^7 emu/g, constituent of all atoms, source of Zeeman light

a. Mathematics. Consistent. The Lorentz Zeeman algebra checks exactly (quadratic m w^2 - H e w - mu = 0; splitting He/m exact, half shift to first order) [read: Corpuscular Theory deep, section 8]. The order-of-magnitude argument that an ion's shift would be about 1/1700 of an electron's is sound. Thomson calls angular frequency "frequency" (cosmetic). b. Experiments. Consistent: Thomson's own 1897 deflection e/m, Zeeman's e/m of order 10^7 with negative sign, Lenard rays, photoelectric and thermionic e/m (Thomson cited these) [read: Conduction, Corpuscular Theory ch. I via digest only]; modern e/m 1.759 x 10^7 emu/g [memory, high]. Thomson's charge e = 1.0-1.2 x 10^-20 emu is 25-35% below modern 1.602 x 10^-20 [read: deep]; the inference chain (his ionic-charge cloud counts) was imperfect, but e/m, which is what claim 1 states, is right. c. Tested. Yes, repeatedly and independently: Kaufmann (1897, same time), Wiechert, Zeeman and Lorentz (1896-97), Millikan oil drop 1909 [memory, high]. Replicated. d. If true: already absorbed into mainstream. What had to change was the indivisible atom.

Strongest honest case. The experimental core is the best-established claim in the folder. Credit is shared: Thomson was not alone, and the contrarian should not overstate priority (Kaufmann and Wiechert published e/m for cathode rays in 1897 [memory, high]). Thomson's specific contribution, the inference that the particle is a universal constituent of matter with a mass near 1/1700 of hydrogen, is the step that others hesitated to take, and it was right. Caveat on rigour. The route to "constituent of all atoms" in Thomson's text is the same particle from many sources (cathodes of different metals, gases, radium beta rays, photoelectrons). That is evidence of universality of the emitted particle, not of its count inside atoms.

Verdict: Established. (Part: "number inside an atom about equal to its atomic weight": see claim 4.)

---

Claim 2: electromagnetic mass: all of a corpuscle's mass is field momentum, m = (2/3) e^2/a, a about 10^-13 cm, rising with speed as Kaufmann's data show; mass is bound ether on Faraday tubes

a. Mathematics. Mostly consistent, with local errors. The slow-motion result m_em = (2/3) e^2/a (emu) checks exactly (magnetic energy (1/3) e^2 V^2/a; field momentum (2/3) mu e^2 w/a) [read: deep, both]. The Heaviside field is correct. The printed closed form of field momentum in Conduction eq. (1) does not reduce to its stated limit (misprint or inherited error). "e = 10^20" is a misprint for 10^-20; the pulse magnetic force lacks a factor e. Two structural issues not resolved by Thomson: (i) the electrostatic energy U = (1/2) e^2 c^2/a = (3/4) m c^2 while m c^2 is defined from momentum: the "4/3 problem" [read: deep, section 5; memory, high that Poincare stresses resolve it]; (ii) Conduction Art. 363 balances work against kinetic plus radiated energy and concludes the acceleration starts at zero instead of Xe/m; with the Schott (acceleration) energy term omitted this is not the accepted radiation-reaction equation [read: Conduction deep, section 9; memory, medium-high on the Abraham-Lorentz equation]. Radius a is a free parameter fixed by the assumption it is meant to test: if m is entirely electromagnetic, a = (2/3) e^2/m = 1.9 x 10^-13 cm with modern values [read: deep]; there is no independent measurement of a. Hence the 10^-13 cm is not a prediction. b. Experiments. The velocity dependence of mass rises steeply as v approaches c: observed and real [read: Conduction Table 1; web: https://en.wikipedia.org/wiki/Kaufmann%E2%80%93Bucherer%E2%80%93Neumann_experiments]. But Kaufmann himself read his data as supporting Abraham's rigid sphere; Planck, Bestelmeyer and Ritz criticised his analysis; Bucherer 1908 favoured Lorentz-Einstein; the search summary reports that a later analysis (1938) found these early experiments only showed a qualitative increase and could not discriminate models, and that Lorentz-Einstein was confirmed by better equipment about 1940 [web: Wikipedia KBN page above, and https://arxiv.org/pdf/1205.5951; summary only, not opened]. Thomson's 1906-07 texts do not mention any of the alternatives (deep reads). Under relativity the gamma-dependence of mass is universal, so "mass varies with speed, therefore it is all electrical" does not follow [read: deep, error summary item 4]. Thomson himself says in the 1906 preface that corpuscles cannot carry the bulk of an atom's mass [read: claims.md; Corpuscular Theory digest p.13], so the claim was already limited to the corpuscle. c. Tested? Partly. The speed dependence was tested (Kaufmann 1901-06, Bucherer 1908, later others) and is explained by relativity. The claim that mass is purely electromagnetic was never tested on its own terms; the later progress (Lorentz, Poincare, Einstein) removed the need for it. The size a was not measured. The bound-ether picture (Faraday tubes carrying entrained ether, with the vortex-ring analogy) was never turned into a measurable prediction by Thomson; the Investigator's finds show he held to tubes of force as "physical realities" through 1925 and gave the electron a "dual structure" in 1928 (Beyond the Electron) [web, via Investigator: https://hdl.handle.net/2027/mdp.39015064384293 and https://archive.org/details/beyondelectron0000unse; not read by me]. d. If true: relativity would still hold, but the electron could not be point-like; yet scattering experiments today put the electron structureless to below about 10^-18 m [memory, medium]. Thus Thomson's a about 10^-15 m is not supported. A reader should also note the classical self-energy divergence persists in modern theory in disguised form.

Strongest honest case. Thomson's electromagnetic-mass work was not a crank idea: it is a correct classical calculation, it predicted the right qualitative speed dependence, it fixed an order-of-magnitude length (the classical electron radius) still used today, and the programme (Abraham, Lorentz, Poincare) led directly to relativity. His admission that mass is "the mass of the ether carried along by the tubes" is the same move as hydrodynamic added mass (checked exactly, Stokes) [read: deep, section 1]. Rigour against Thomson. The Kaufmann "agreement" cannot be reproduced from his printed formulas (Corrections 1-2) and the model-independence he claims is not available; once alternatives were on the table the data did not choose his model.

Verdict, split: Slow-speed formula m_em = (2/3) e^2/a as calculation: Established. Velocity dependence of electron mass: Established (but explained by relativity). "All the mass of the corpuscle is electromagnetic" with a about 10^-13 cm: Historically superseded but instructive. "Bound ether on Faraday tubes" as the general origin of mass: Not testable as stated.

---

Claim 3: Roentgen rays are thin electromagnetic pulses from a suddenly stopped charge; the rays make matter conduct by splitting molecules

a. Mathematics. Consistent where checked. Kink-field construction, pulse energy (2/3) e^2 v^2/delta and the acceleration fields and Larmor formula (2/3) e^2 f^2/c all check [read: Corpuscular Theory deep, section 7; Conduction deep, sections 8-9]. Minor: the Rontgen Rays reprint omits the maths; the absorption formula in the memoir has m missing from the denominator (misprint) [read: Rontgen deep, section 2.4]. Limit: the formula gives more energy than the particle had if delta < 2a; Thomson never discusses it [read: Conduction deep]. b. Experiments. (i) Ionisation of gases and insulators by X-rays at low fields, both polarities, no threshold found down to 3-4 volts: observed [read: 1896 note, pp.1-3]. (ii) The continuous X-ray spectrum is, in classical terms, bremsstrahlung from decelerated electrons, and this classical pulse picture is correct as far as it goes; it cannot give the sharp short-wavelength cutoff (Duane-Hunt 1915) or characteristic lines [read: deep; memory, high]. (iii) The claim that X-rays are NOT periodic waves, only impulses, was overturned by Laue-Friedrich-Knipping diffraction in crystals (1912) and Bragg's spectra [memory, high]. This was a direct test, and the pulse-only picture lost. (iv) Barkla's polarisation of X-rays (1905-06) already pointed to transverse waves; Thomson used Barkla's secondary-radiation data as support for his scattering formula [read: deep]. c. Tested. The 1896 conductivity measurements are qualitative (no numbers in the note [read]). The pulse theory had one proposed test: the angular distribution should be greatest at right angles to the cathode beam for a single-encounter stop [read: Rontgen deep, section 2.5]. Stokes noted target absorption confounds this. The decisive tests (crystal diffraction) came later and were not Thomson's. d. If true as an exclusive picture: no wave diffraction of X-rays; mainstream would lose crystallography.

Strongest honest case and the Investigator's finds. (1) Priority and independence: a telegram from St Petersburg and Lodge's letter recorded in the same minutes confirm the discharge effect independently [read: 1896 note, p.3]. (2) Thomson's localisation argument, recorded in Rayleigh's 1942 life: a Faraday-tube wavefront with "specks of great intensity" explains why X-rays ionise fewer than one molecule in 10^9 [web, via Investigator: research/2026-10-08-archive-org.md; not read by me]. This is a classical argument for the same fact the photon picture explains (X-rays deliver their energy in small concentrated units), and Thomson's 1925 "anchor ring" quantum and 1928 dual-structure electron follow from it [web, via Investigator]. Bragg, who also favoured corpuscular X-rays, held a neutral-pair hypothesis from 1907 [read: deep; memory high]. So the pulse theory was not a silly position: the unsolved wave-or-particle question was real and Laue's result (1912) did not close it for ionisation, whose mechanism (localised absorption, photoelectrons) is closer to Thomson's pulse picture than to a diffuse wave. (3) Wall (1924) extended the Stokes-Thomson theory (Phil. Mag. 48:378, not read) [web, via Investigator]. Rigour against Thomson. In the memoir the thickness delta is assumed, not derived (Stokes: "may probably be something comparable with the millionth of an inch"); a worked number in the deep read gives a 30 keV electron stopped in atomic distance radiating about 10^-5 of its energy, consistent with real tube efficiencies, but a pulse thin enough to radiate most energy would be below atomic size [read: deep; calc]. The pulse picture predicts no wavelength cutoff; it is the first thing quantum theory had to add.

Verdict, split: Ionisation by X-rays making matter conduct, as an experimental finding: Established. Classical pulse (bremsstrahlung) origin of the continuous spectrum, and Larmor power: Established (as classical theory; incomplete). "Rontgen rays are impulses, not periodic waves" in the strong form: Tested, not supported (crystal diffraction 1912; this was a real test, not a dismissal). Thomson's linked "specks" localisation idea: Historically superseded but instructive.

---

Claim 4: the atom is a uniform positive sphere with as many corpuscles as its atomic weight, ringed in stable patterns that explain the periodic law, valency, spectral series

a. Mathematics. Internally mostly consistent as classical mechanics: ring stability, the three-corpuscle equilibrium (0.577 R; the printed surd is illegible or a misprint), Mayer's magnets analogy [read: E&M deep]. Errors: radiation table n = 5 row off by a power of ten; table of rings N = 13, 8, 3 printed twice, no N = 23, valency of the 62 group swapped; like poles said to repel "inversely as the distance" [read: claims.md; deep]. The claim of n of about 1000 corpuscles in hydrogen gives a stored energy of 10^19 erg per gram [read: E&M deep], which is wrong by n^2 / n = 10^3 (modern n = 1 gives order-of-magnitude correct ionisation energy only by accident of the charge^2/radius method). The deep reader notes a coincidence: 10^12 J/g is close to hydrogen-fusion energy, accidental. Hidden assumption: classical stability of orbiting charges (no radiation), and equipartition in 3n modes: with 1000 corpuscles, gamma = 1.0007 versus measured 1.41 [read: deep], which Thomson addresses by arguing internal energy is lost by radiation, an unfalsifiable rescue. b. Experiments. Counting: see Correction 3. Periodic law: Thomson's inner-triangle recurrence at 3, 10, 20, 35 is a supposition (number of magnets proportional to combining weight) and does not reproduce 2, 8, 8, 18, 18, 32 [read: E&M deep, section 8]. Positive sphere: Rutherford's nuclear atom (1911) after Geiger-Marsden alpha scattering (1909) [memory, high]; Thomson's model predicts only small multiple-scattering angles, and the observed large-angle back-scattering was a direct test. Moseley 1913 and the Barkla-corrected count of A/2 gave atomic number [memory, high]. Spectral series: classical corpuscle normal modes cannot give the Rydberg-Ritz combination principle [read: deep, section 9; memory, high]. Thomson's Darwin periodic-orbit argument for sharp lines is a qualitative analogy with no orbit computed [read: Corpuscular Theory deep, section 4]. c. Tested. Yes. Alpha scattering (Geiger and Marsden, Rutherford) tested the positive-sphere model and refuted it; Bohr's model (1913) displaced the ring dynamics. Replicated many times. d. If true: no nucleus; the entire atomic physics of 1911-1925 would be unneeded.

Strongest honest case and the Investigator's finds. Thomson's 1904-1906 papers were the first serious shell model: chemical periodicity as recurring arrangements of electrons, and the demand that the number of electrons be small, not thousands, was derived by him from scattering, absorption and dispersion before the nucleus was found [read: deep]. The qualitative inference that spectra and chemistry reflect symmetries of the electron arrangement survived. The mathematical descendant, the Thomson problem (N charges on a sphere; Smale's 7th problem), is alive, and LaFave 2013 reports that its energy irregularities correspond to shell-filling and ionisation-energy patterns, conjecturing a role in the periodic law [web: https://arxiv.org/abs/1403.2591, abstract read by the Investigator; PDF not read]. This is a modern, partial, classical analogy; the mainstream explanation of the periodic law remains quantum. Hon and Goldstein (2013) note Thomson never used the "plum pudding" label [web via Investigator: Wikipedia, Thomson problem/Plum pudding]. Rigour against Thomson. The phrase in claims.md "supported by three counting methods" is exactly the weakness: each method has an escape clause (strongly bound corpuscles invisible to Method I, repulsion law at short range assumed in Method II, valency corpuscles in Method III; Darwin analogy for spectra). The numbers (21.5 vs "25") were rounded toward the answer sought, and the arithmetic with his own e inflates N/n; this is not fraud, but it shows the count was steered by the prior.

Verdict, split: Uniform positive sphere with electron-count about A: Tested, not supported (alpha scattering 1909-11; count about A/2, Moseley 1913). Ring arrangements explaining the periodic law and spectral series: Tested, not supported (cannot reproduce 2, 8, 18...; no combination principle); the classical-electrostatics partial analogy to shell-filling (LaFave) is a Plausible, untested side-line as a mathematical curiosity, not as physics. The structural idea (electron arrangements underlie periodicity): Historically superseded but instructive.

---

Claim 5: positive rays identify atoms by mass; the 1913 line at 22; mass 3 (X3, called H3); helium and neon from bombarded solids

This is the claim where the Investigator found the richest record, and where the contrarian's standard applies most. I treat five sub-claims.

5a. The positive-ray (parabola) method

a. Maths: consistent; (e/m)1/(e/m)2 = (y1^2/x1)/(y2^2/x2) at fixed x gives m2/m1 = y1^2/y2^2 [read: Positive Rays deep, section 1]. Worked check: neon-22 sits only 4.9% from neon-20 in deflection [read: deep]. Multiple-charge signature (prolongation at 1/2, 1/3 of normal distance) is sound. b/c. Experiments: Xe, Kr, Ar, Ne lines identified; detection of 1% helium in a mixture and 3 x 10^-6 cc of helium claimed [read: p.130]. This became the mass spectrograph (Aston 1919) [memory, high]. Replicated by Aston and beyond. Verdict: Established.

5b. The mass-22 line beside neon

Thomson (1913): "It must, I think, be a new element" [read: printed p.112], but also, after Aston's failures to separate it by charcoal fractionation (no change before/after, quartz-balance density check) and partial separation by clay-pipe diffusion, the "suspicion that the two gases ... may be indistinguishable in their chemical and spectroscopic properties", with a radioactive analogy (radio-lead, thorium) [read: p.137]. He did not yet use "isotope" (Soddy, 1913) [read: deep; memory, high]. Aston's 1919 mass spectrograph confirmed Ne-22 as a stable isotope of neon, now about 9% abundance [read: deep; memory, high]. Thomson's two-part hypothesis (new element, or inseparable twin) therefore resolves in favour of the second. Aston also worked in Thomson's laboratory (interested party: an assistant defending the lab's method, but his measurements were physical, with a quartz balance). The 22 line is also a clean example of an anomalous result handled well: alternatives (CO2++, NeH2) were tested and excluded [read: pp.135-136]; the controls are detailed. Verdict: Established (Ne-22 isotope), with Thomson's wording ("new element") historically superseded.

5c. Mass 3: X3, called H3

Observation: a strong line at m/e = 3 from bombarded solids and from hydrogen-containing salts. Thomson considers C4+ and H3; excludes C4+ with real arguments (no C2+/C3+ lines, no hydrocarbon source produces it, gas stable for days) [read: pp.141-143]. He tests properties: weeks of storage, destroyed by light plus oxygen, by sparking with oxygen, by hot copper oxide; survives heat in the absence of oxygen; palladium diffusion enriches it; it combines with mercury; yields persist for months from potash [read: pp.141-147]. His valency comment: H3 cannot be fitted to ordinary valency [read: p.146]. Mainstream: the mass-3 ion is H3+, formed in the discharge by H2+ + H2 -> H3+ + H; it is well established and astronomically important; the neutral stable H3 he argued for was abandoned by the mid-1930s [web: Kragh 2011 Centaurus abstract https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1600-0498.2011.00237.x ; Kragh 2012 Phil. Trans. A 370:5225, abstract, via the Investigator; HD+ also contributes at mass 3 after 1931 [read: deep; memory, high]]. Thomson's chemical tests were made on a stored neutral gas; the deep reader flags that how these map onto modern chemistry "is not settled by this text". Competing readings: (i) the stored-gas behaviour was another species or contaminant (the destruction by light plus oxygen, and the electrodeless-tube null result, "No lines which could be ascribed to X3 were detected" [read: p.147], are against a neutral H3 hypothesis; his own text records it); (ii) the observations of storing for weeks refer to a different species than the ion detected in the analyser; (iii) unknown. I have no document that tested his stored mixture independently; Kragh's abstract says results with neutral H3 ("active hydrogen") were "contradictory". Thomson kept the name X3 in his papers until 1937 [web: Charles University thesis snippet, via the Investigator; not opened]. Strongest honest case: his identification of a new mass-3 species was correct as an ion and he found the real species ahead of anyone; he flagged the weak point (spectrum null). His reasoning about the neutral gas is the questionable part. Verdict, split: Existence of a mass-3 line due to a triatomic hydrogen species (as H3+ ion): Established. Stable neutral H3 stored for weeks as an allotrope: Tested, not supported (Kragh: "contradictory", abandoned by the mid-1930s; I have not read the experiments themselves, so the strength of "tested" rests on Kragh's summary; if a reader requires primary sources this stays Never properly tested in my own reading).

5d. Helium and neon liberated by cathode-ray bombardment of metals and salts

Observation [read: pp.147-154]: traces of helium (and sometimes neon) in the analyser after bombardment of metals, minerals, salts; yield rises with bombardment time (so not just the 1 cc of air whose helium would give a stronger line than most of his signals); solution-and-evaporation of salts leaves the supply undiminished, up to ten or twelve cycles; control with aluminium electrodes saturated in helium gave no more than a blank half; cathode ruled out as source by independence of salt choice; yields vary between specimens of one salt; yield declines with long bombardment in nearly every case (unlike X3 from potash). Thomson's own hypotheses: absorbed helium (a), disintegration (b), helium compounds as impurities (c); he judges (a) weakened by his controls, (c) improbable, and offers a "loosened alpha-particle" bridge. He explicitly refuses to claim transmutation: "few will be prepared to accept it until every other explanation has been shown to be untenable" and names the decisive test (find the residue atoms; Be or C best) [read: printed p.130]. Contemporary evaluation, with each source's stake: Ramsay, Collie and Patterson, Masson, Baly reported helium and neon from discharges and X-ray bulbs and interpreted them as transmutation (they were the proponents; their claims are theirs, Baly called the case "overwhelmingly strong") [web, via Investigator: Morrisson, Modern Alchemy]. Thomson's Nature letter of 13 Feb 1913 offers "a different explanation" [web: https://www.nature.com/articles/090645d0, preview only]. The negative tests: Strutt (Proc. A 89:499, 1914), Merton (Proc. A 90:549, 1914), Egerton (Proc. A 91:180, 1915), Harkins and Wilson (1915), Allison and Harkins (1924) [web, via Investigator bibliography (Nelson); I confirmed Merton's paper title and that Robert John Strutt reported finding no helium in similar work via https://makingscience.royalsociety.org/items/rr_21_122 and a historical summary in the search result; the papers themselves were not read]. Redgrove (1922) says the claims "failed to be satisfactorily confirmed" and urges repetition after Rutherford's 1919 nitrogen result [web via Investigator: https://www.gutenberg.org/files/43240/43240-h/43240-h.htm]. Nelson (alternative-science author, a sympathetic party) writes "The matter never has been resolved" and lists "dozens of papers 1905-1927" [web via Investigator: https://www.alchemywebsite.com/nelson2_6.html]. The LENR community's own list of historical papers (New Energy Times) contains no Thomson, Collie or Strutt entry [web via Investigator; negative]. Rejection of the transmutation claims: handled by open experiment and open publication; Thomson, Strutt and Rutherford were the principal opponents [web via Investigator]. The challengers continued to publish replies about contamination objections (Collie and Patterson answered "occluded gases, permeable glass, and neon-contaminated hydrogen and oxygen") and a later revival by Riding and Baly is mentioned [web via Investigator]. Neither side's data were suppressed. The pertinent distinction: the negative papers tested the Collie-Patterson-Ramsay discharge-tube claims (gas in a tube, helium appears), not Thomson's own bombardment of solid salts and metals. I found no independent replication or refutation of Thomson's salt-bombardment helium, nor his own reading of the later result in 1922 (Proc. A 101:290; contents on deflagrated metals not read). The 1921 edition keeps both his explanations and still names the residue test [web via Investigator: https://archive.org/details/in.ernet.dli.2015.160862; not read by me]. So the bombardment helium is, on the record I have, neither confirmed nor refuted by a neutral controlled test. Mundane mechanisms to weigh (mainstream, [memory]): helium diffuses through glass and is occluded in solids and electrodes; metals and glass accumulate atmospheric helium and radiogenic helium (monazite, which Thomson himself says gave abundant helium [read: p.150]); ion bombardment releases it. Thomson's controls (time-dependence, solution and evaporation, aluminium blank) address several of these but not helium trapped in the crystal or glass walls as a persistent source with slow release; whether atmospheric helium in the crystal lattice (not surface) survives dissolution is not tested by dissolution, since dissolution would remove it. Thomson's solution test is therefore a strong control for adsorbed gas, not for helium inside or generated within the salt. His observation that K salts, particularly KI, gave "an exceptionally large" supply is unexplained; Na-He would be 23 - 4 = 19 (F), K-He 35 (Cl): inconvenient detection, as he says [read: p.153-154]. A positive test has been available since 1919 (the nuclear transmutation by alpha particles, Rutherford) and since 1932 (Cockcroft-Walton), but those need MeV particles, not tens of keV electrons, and nuclear physics today puts the thresholds of such reactions far above cathode-ray energies [memory, high]. So transmutation of K or Na into helium by 50 keV electrons is refuted by calculation in the sense of known nuclear energy scales (binding energies of MeV), not by experiment of 1913. Wendt and Irion (exploding tungsten, 1922) and Paneth and Peters (hydrogen to helium over palladium, 1926): Paneth and Peters retracted, attributing the helium to background from air and glass [web: https://en.wikipedia.org/wiki/Kurt_Peters_(chemist); https://www.deutsche-biographie.de/138755426.html]. I found no source reporting a replication or a refutation of Wendt and Irion; the only accounts found are Nelson's [web: https://www.alchemywebsite.com/nelson2_4.html]. They are related work in the same tradition. a. Mathematics: not applicable beyond the parabola formula; the (arithmetic) estimates of sensitivity (1 cc air helium about 3-4 x 10^-6 cc) differ between p.130 and p.148 of the book (3 vs 4 x 10^-6; modern about 5.2 x 10^-6) [read: deep; memory, high]. b. Consistent with established experiments? Helium emerging from solids bombarded or heated is consistent with established gas occlusion and radiogenic helium; transmutation by 50 keV electrons is not consistent with nuclear physics. c. Tested? Discharge-tube claims: tested by Strutt, Merton, Egerton and others, not supported. Thomson's solid-salt bombardment: Thomson tested with controls (his own); no independent neutral test found. d. If true as transmutation: the stability of nuclei under keV electrons would fail; an enormous change. Interested parties. Thomson was both an experimenter and the senior judge of this question (as President-level voice he is described as "had invented an absolutely new system of chemical analysis" by Percy Frankland at the Chemical Society, March 1913 [web via Investigator: Morrisson]); his Cavendish colleagues (Aston) and the method's reputation were at stake in 1913; the transmutation proponents (Ramsay, Collie, Patterson, Masson, Baly) had their own priority stakes. Absence of evidence. The record goes quiet on Thomson's own bombardment helium after 1922. Competing explanations: (1) he concluded it was contamination/occlusion and moved on (consistent with the 1922 paper's focus on gas analysis of liquid-air residues, radium vessels and deflagrated metals, whose result I have not read); (2) the topic was displaced by the isotope work (Aston 1919) and nuclear physics; (3) data remain in unpublished notebooks (CUL MS Add. 7654, NB35a, NB36 are cited by Falconer [web via Investigator]). What would decide: the 1922 paper's result, the 1913-1922 notebooks, and a modern bombardment of salts under helium-free ultra-high-vacuum with isotope measurement (He-3/He-4 and K-40 branching); a modern helium-leak-tested replication would settle it in a week. Verdict, split: Helium and neon appear in the analyser after bombarding solids: Plausible, untested as a precise phenomenon (consistent with occlusion; no neutral replication of his salt-bombardment protocol found). Transmutation of the bombarded atoms: Refuted by calculation (energy scales) in its strong form; Thomson himself did not assert it, so for his actual position ("liberated or loosened helium, undecided") Never properly tested. The Collie-Patterson-Ramsay discharge-tube transmutation claims: Tested, not supported (Strutt, Merton, Egerton, Allison and Harkins; summaries only).

---

Claim 6: vortex rings in a perfect fluid: computed interactions, linked rings, stability of up to 6 vortices on a circle; vortex atoms and kinetic theory

a. Mathematics. Consistent in structure with corrections: eq. 71 printed 3 times too small, eq. 74* needs k^4 (the k^3 version has wrong dimensions), the sphere total angle lost a factor 1/u; the "seven unstable" verdict is a rounding artefact, seven being neutral at linear order [read: Vortex deep, sections 4-6 and III.6, III.10]. The inverse-cube deflection and the sin 3 phi rules follow from the displayed equations; the equivalent impulses' printed coefficients are tied to the 71 error [read: deep]. Hidden assumptions (all stated by the deep reader): thin core, linear stability only, equal roots handled by appeal to Thomson and Tait, a slow-ring constant that does not affect the stated results. b. Experiments. The mathematical results on n point vortices on a circle (stable for n up to 6) were confirmed by Havelock (1931) and later work; n = 7 is marginal at linear order [web: see Correction 6]. The Mayer floating-magnet experiments (1878) were an experimental analogue [read: deep, III.7]. Modern superfluid and BEC vortex lattices realise related physics [memory, medium]. The atomic application (vortex atoms) was the programme of Kelvin (1867), taken up by Thomson in the 1882 Adams Prize essay; it was abandoned after 1890 [read: claims.md; web via Investigator: Wikipedia, Vortex theory of the atom]. c. Tested? The fluid mechanics is standard and has been tested (smoke rings, ring collisions) [memory, medium]; the vortex atom as a theory of matter was never given a spectrum prediction and was displaced by electron theory, Thomson's own work being a main cause. The Investigator's thread: Thomson's 1925 "anchor ring" quantum of closed lines of electric force moving at c has a family likeness to his vortex ring (his own later echo of the vortex idea) [web via Investigator]. This is a lead, not evidence of a quantitative link. d. If vortex atoms were true: a perfect fluid with persistent vortices as the basis of all matter; this would require ideal-fluid conservation of circulation to hold at atomic scales, impossible with ordinary viscosity; no such fluid is known.

Strongest honest case. The 1883 essay is rigorous mathematical physics for its time (Adams Prize), mostly correct, and its 6-vortex polygon result anticipates modern vortex-crystal stability. The deep reader found the errors; the best mathematics survived.

Verdict, split: The mathematical results for interacting vortex rings and polygons (with the stated corrections): Established as mathematics, with local errors noted. Vortex atoms as a theory of matter and chemistry: Historically superseded but instructive.

---

Summary table

What would settle the open points

• Claim 2: re-derive Thomson's alpha and p tables from the printed model and compare with raw Kaufmann (1901-03) data; the swap in Correction 1 alone would be settled by looking at the 1907 page image (the deep reader read the image, but a recheck of the table headers is cheap).
• Claim 5d: read Thomson 1922 (Proc. A 101:290), the Strutt/Merton/Egerton papers, and run a modern helium-free salt-bombardment test with isotope measurement.
• Claim 5c: the stored-gas properties of X3 (weeks of stability, light plus O2 destruction): a modern repeat with mass spectrometry on a stored hydrogen-bombardment product.
• Claim 3: nothing needed; settled by diffraction. Claim 4: settled by scattering.