RankingsPrimerArbiterPhysicistMathematicianContrarianQESites logClaimsSourcesResearch notes

Thompson-JJ (J. J. Thomson): central claims

Project: 11 documents by or co-authored by J. J. Thomson (1883-1913), all primary; none is a secondary or AI-written analysis. All have a readable text layer (several are OCR scans in which equations are garbled; the deep readers read equations from page images).

Coverage note (not a source gap; updated after QE, 2026-10-09): deep reads now cover Vortex Rings Parts II, III and IV, Conduction Chapters V-VI (e/m and e values), XI (Arts. 160-163), XX (Arts. 346-350) and XXI, Electricity and Matter Chapter II (equation (1) and Table II), Chapter V (first half) and the end of Chapter VI, Corpuscular Theory Chapters I, II-III, IV-V, VI (part 1) and VII, Positive Rays pp.106-131, Roentgen Rays (Stokes and Thomson memoirs), and the 1896 note in full. Digest only: the other chapters of Conduction (about 400 pages), Electricity and Matter Chapters I, III, IV and the rest of VI, the rest of Corpuscular Theory Chapter VI, Positive Rays earlier chapters, Vortex Rings Part I, the textbook. The earlier loss of the Corpuscular Theory Chapter I deep read has been made good (Chapter I was deep read again).

Central claims

  1. Cathode-ray particles ("corpuscles", now electrons) are the same whatever their source, carry a negative charge with e/m of about 10^7 electromagnetic units per gram (about 1/1700 of the hydrogen ion's mass), are a constituent of all atoms, and are the emitters of the light in Zeeman-effect lines. Thomson's own first measured values were low: 7.7 x 10^6 (deflection, 1897) and 1.17 x 10^7 (heating method, 1897), against the modern 1.76 x 10^7; the 1/1700 ratio rests on later work (Kaufmann, Simon and others, about 1.86 x 10^7) plus the equality of the charge with the hydrogen ion's (his 1898 cloud-chamber charge 6.5 x 10^-10 esu was revised to 3.4 x 10^-10 in 1901-02). The same book of 1907 also claims that at very low pressure every gas gives the same two positive carriers (e/m about 10^4 and 5 x 10^3, masses one and two hydrogen atoms), a claim he later dropped.
  2. The inertia of a corpuscle is entirely electromagnetic: field momentum gives m = (2/3) e^2/a, rising with speed as Kaufmann's radium data are said to show, with radius a of about 10^-13 cm (5 x 10^-14 in Electricity and Matter); more generally, mass is the "bound ether" carried by Faraday tubes. (Thomson himself says in 1906 that corpuscles cannot carry all of an atom's mass.) The velocity formula is printed correctly in Electricity and Matter and misprinted in Conduction. His Table II compares two columns, 3.1, 2.42, 2.0, 1.66, 1.5 and 3.09, 2.43, 2.04, 1.83, 1.65; Electricity and Matter (all printings) and the Corpuscular Theory label the second as Kaufmann's observed values, Conduction (1906) labels the first as observed; the labelling is decided only by Kaufmann's original papers, and neither column is reproduced from Thomson's stated model.
  3. Roentgen rays are thin pulses of electric and magnetic force emitted when a charge is suddenly stopped, and the rays make matter conduct by splitting molecules (ionisation): the 1896 discharge experiments, Stokes's and Thomson's 1897-98 pulse theory, and the Larmor radiation law. In 1906 Thomson reads Haga and Wind's slit-broadening (diffraction) result as a pulse thickness of about 2 x 10^-8 cm, and reports Marx's finding that the rays travel at the speed of light.
  4. An atom is a uniform sphere of positive electricity holding as many corpuscles as its atomic weight, arranged in rings whose stable patterns change in steps, which may account for the periodic law, two-sided valency and spectral series. Thomson presents three counting methods as agreeing with N proportional to atomic weight; recomputation does not support the agreement, and Barkla's datum (printed both as .0002 and .00025 in Conduction) gives 21.5 per air molecule with Thomson's constants and about 14 (11 for .0002) with the modern charge, about half the atomic weight (the atomic number). He also finds, in his own metallic-conduction chapter, that free corpuscles would need about 6 cal per degree per cc of silver against a measured 0.6, and concludes free-corpuscle conduction is "at most only a part" of metallic conduction.
  5. The positive-ray (parabola) method identifies atoms and molecules by mass, and in 1913 it showed a line at mass 22 beside neon 20 (a new element or an inseparable form of neon; today neon-22, an isotope), a line at mass 3 called triatomic hydrogen H3 (today H3+, with HD+ also possible), and helium and neon liberated by cathode-ray bombardment of metals and salts, which Thomson links to disintegration of the bombarded atoms but does not claim is proved. Before 1913 Thomson had himself suggested (1906) that X-rays or radium might transmute atoms (keeping radium in contact with lead), reported lead heating about twice as much as zinc for equal absorbed X-rays (Bumstead), and reported negative searches for radioactive products after X-ray and radium exposure.
  6. (Early mathematical work, 1883) Vortex rings in a perfect fluid interact by laws that can be computed: path deflection falling as the inverse cube of distance, radius changes governed by sin 3(phi), attraction to a sphere, a stable linked-ring state, and n vortices on a circle stable for 3 to 6 (seven is neutral at linear order and, by later work, nonlinearly stable; eight is the first unstable case). A gas of such vortex atoms gives the kinetic-theory pressure with a small wall term (pv slightly below Boyle's value), makes the mean speed of a monatomic gas fall as temperature rises, predicts a thermal-effusion pressure law p/p' = (t/t')^m with m greater than 1 (kinetic theory: 1/2), and might explain chemical combination through the ratio of paired to free time and a valency rule (at most six links; carbon divalent; nitrogen, phosphorus and arsenic univalent). The effusion law and the speed prediction were later measured with the kinetic-theory result.

Corrections to earlier documents (found by deep readers; for the orchestrator)

  • Thomson's printed results with errors found by the deep readers, recorded so later roles do not repeat them as established: Vortex Rings eq. 71 is printed 3 times too small (reverses the p.75 crossing comparison), eq. 74* should read k^4, the sphere total-angle lost 1/u, and the "seven vortices unstable" verdict is a rounding artefact (seven is neutral at linear order, eight is unstable); Conduction eq. (1) in Chapter XXI fails its own slow limit and its calculated mass-ratio column p cannot be reproduced from the stated model; Corpuscular Theory "e = 10^20" is a misprint for 10^-20 emu, the pulse magnetic force formula lacks a factor e, the table of rings has N = 13,8,3 printed twice and no N = 23, and the text swaps the valencies of the 62 group; Electricity and Matter's radiation table has a power-of-ten slip in the n = 5 row and the heading says "each corpuscle" for group totals, and like poles are said to repel "inversely as the distance".
  • Digest errors fixed by the deep reads (apply when quoting the digests): Rontgen Rays digest p.85 (professorship 1884 not 1881) and p.77 (sphere around P of radius PS); Electricity and Matter 1909 digest p.126 and p.128 (n = 1000, 1.02 x 10^19 erg) and p.136; Corpuscular Theory digest p.48 (2.43 not 2.48) and p.49 (m = (2/3) e^2/a); Vortex Rings digest p.131-132 (volume XVIII not XVI, slow period 2 pi a / (n sqrt(n^2-1) V)).

QE revision (2026-10-09): corrections and additions applied (for the orchestrator)

Applied above: the arbiter's corrections 1, 3 and 7 (claims 4, 6 and 5) and QE C1 and C2 (claims 2 and 4); the Part IV and e/m deep reads added to claims 6, 1 and 5. For the orchestrator, from the new deep reads:

  • Kaufmann table (replaces arbiter correction 2 and the Corpuscular deep read's section 3 wording): Electricity and Matter Table II (PDF p.65 of the 1909 scan; the 1904 and 1907 printings agree) and the Corpuscular Theory (PDF p.48) label 3.09, 2.43, 2.04, 1.83, 1.65 as Kaufmann's observed values; only Conduction (PDF p.668) labels 3.1, 2.42, 2.0, 1.66, 1.5 as observed (the difference is a swap of the two Greek letters in the header). The e/m-product test favours the Conduction labelling to about 4% for one common slow-speed e/m near 1.93 x 10^7 (1.95 x 10^7 in the QE; 1.86 x 10^7 gives 3.00, 2.42, 1.91, 1.59, 1.42); the model test at the two lowest speeds and two of three books favour the Electricity and Matter labelling. "The 1907 book's headings are interchanged" is retracted. Kaufmann's 1901-03 papers decide it.
  • Equation (1) of the momentum of a moving charge is correct in Electricity and Matter (the 1/4 coefficient is exact); the misprint is in Conduction only. Earlier statements that "Thomson's equation fails its own slow limit" should say "as printed in Conduction". The calculated column still cannot be reproduced from the correct formula (3.1 ... is 8-12.5% above the model at the top three speeds, 3.09 ... is 7.5-13% above everywhere).
  • Barkla's datum: Conduction p.339 prints "about .0002" in words and ".00025" in the equation. Thomson prints Ne = 10 where 8.6 follows from .00025 (6.9 from .0002); "about 25 per molecule" comes from the rounded 10 with ne = .4.
  • Thomson's e/m history: the digest's "7.7 x 10^7" is a misreading; the page reads 7.7 x 10^6 and "770 times". Corpuscular Theory p.30 reads e = 3.1 (not 8.1) x 10^-10 esu.
  • Part IV of the vortex essay now has a deep read; "gas of vortex atoms reproducing kinetic pressure: not assessed" can be assessed. Its "at most six" rule rests on the Part III seven-vortex rounding error. Reynolds 1879 and Knudsen 1909-10 [web] measured m = 1/2; Stern's atomic-beam speeds agree with Maxwell's law [web].
  • facts.json C5: the statement that Electricity and Matter "rejects" an external-radiation source is replaced by Thomson's actual position (internal source because radioactivity is transient; magnitude "not a valid argument"; penetrating Roentgen rays could replace Le Sage's corpuscles for inverse-square gravity). Items from the deep reads absent from the reviews and arbiter, to be weighed there: the specific-heat contradiction (6 cal against 0.6 cal) and the "at most only a part" conclusion; the two universal positive carriers; Thomson's pre-1913 transmutation suggestions and null searches; Haga-Wind and Marx in his own text; the 50,000-year radium arithmetic; a reader's margin annotations in the project's copy of the 1904 scan. Also for the arbiter and sites.md: C. N. Wall (Phil. Mag. 48:378, 1924) and Nelson's "tritium" characterization of the mass-3 line (QE C8).
  • Small internal inconsistencies of Thomson's recorded for the reviewers: Conduction UV e/m 7.6 x 10^6 (table) against 7.3 x 10^6 (text); Simon's potential 4860-11840 V against "about 1100 volts"; corpuscle radius 5 x 10^-14 cm (Electricity and Matter; the stated numbers give 8 x 10^-14) against about 10^-13 cm (1906-07 books); Electricity and Matter's e = 1.2 x 10^-20 "electrostatic units" is an electromagnetic-unit value; Kaufmann's electrical-mass fraction "one-fourth" (Electricity and Matter) against "1/3" (Conduction); Corpuscular Theory p.100 prints a = 1.5 x 10^-16 where p.65 has 1.32 x 10^-16.
  • Digest errors for anyone quoting digests: Conduction digest part 04 p.135 (7.7 x 10^7 should be 10^6); Conduction digest part 21 p.668 (labels reversed); Corpuscular digest p.89 (contact-potential formula should have the factor 2/3; Peltier numbers 1/30 volt, log 1.33, not 1/80 and 1.88); Corpuscular text layer p.99 ("8.3 x 10^-15" should be 3.3; "1.8 x 10^22" should be 10^24) and p.70 ("n = 9 x 10^2" should be 9 x 10^23).