Thompson-JJ · Source catalogue

Thomson's Cambridge Adams Prize essay (Macmillan, London, 1883, 153 PDF pages, Internet Archive/Google scan, OCR on 152 of 153 pages so equations are garbled in the text layer and were read from page images by the deep readers). The set subject was 'a general investigation of the action upon each other of two closed vortices in a perfect incompressible fluid' [read]. Part I treats the vibrations of one slightly disturbed ring. Part II treats two rings that pass each other far apart (compared with their diameters), treating each ring in the far (dipole) field of the other and integrating perturbations along the undisturbed straight path; the deflection of a ring's path falls off as the inverse cube of the shortest distance, the radius change follows a sin 3(phi) rule, and the summary splits collisions into Class I and Class II at a 60 degree boundary; a ring passing a fixed sphere is attracted to it and its radius swells and then relaxes. Part III treats linked rings (equal and unequal strength: a quick and a slow vibration, an exact moment-of-momentum condition for unequal rings, eq. 105) and n equal straight vortices on a circle, concluding that 3 to 6 are stable and 7 or more unstable, and cites Mayer's floating magnets (Nature 1878) through Kelvin's analogy. Part IV applies the results to a vortex-atom theory of gases and of chemical combination (pairing of rings; pressure of a gas of vortex atoms equals the kinetic-theory pressure). The deep reader reproduced most results independently: eq. 69 (in-plane bending) and the radius change are exact (the latter once k^3 is corrected to k^4); the out-of-plane bending eq. 71 is printed 3 times too small, which reverses one printed comparison (crossing vs non-crossing paths, p.75 of the PDF) but not the signs, the 30 and 60 degree rules, or the Class I/II summary; the total-angle formula for the sphere lost a factor 1/u; and the 'seven vortices are unstable' verdict rests on a rounding error (full precision gives a zero root, i.e. seven is neutral at linear order and eight is the first unstable case). [read: deep/A-treatise-on-the-motion-of-vortex-rings-a73ca5.md] QE REVISION (2026-10-09): Part IV, which this entry first listed as digest only, was deep read by an Opus reader (QE remediation) from the second scan; the results are in the treatiseonmotionofvotexrings.pdf entry. Because Part IV's 'at most six links per atom' rule rests on the Part III seven-vortex instability, it inherits the rounding error found here (seven is neutral at linear order, eight is the first unstable case). [read: deep/treatiseonmotionofvotexrings-0ee5cd.md Part IV]

Key claims

• Two vortex rings that pass at distances large compared with their diameters deflect each other's paths by an amount inversely proportional to the cube of the shortest distance between them.
• In such an encounter one ring's radius grows and the other's shrinks, by an amount proportional to sin 3(phi) with phi the angle between the line of centres at closest approach and the common perpendicular; the sign flips at phi = 60 degrees (Class I / Class II).
• A vortex ring passing a fixed sphere is attracted to it, its radius first swells then relaxes to its original value, and the greatest swelling is proportional to the sphere's volume and inversely proportional to the sixth power of the closest distance.
• Two linked vortex rings can move steadily and stably; each mode of displacement has a quick vibration (about 2m/(pi d^2)) and a slow vibration (about n sqrt(n^2-1) V/a), and unequal rings can be linked only at one definite moment of momentum (eq. 105).
• n equal straight vortices equally spaced on a circle are stable for n = 3 to 6 and unstable for n = 7 or more (the deep reader finds n = 7 is actually neutral at linear order, and n = 8 is the first unstable case).
• A gas of vortex rings has the pressure of the kinetic theory (Kelvin's argument as quoted); Thomson's own Part IV result is pv = T/3 minus a small wall term, so pv is 'a little less' than Boyle's value; pairing of rings that overtake one another is proposed to account for chemical combination (vortex-atom theory). See the treatiseonmotionofvotexrings.pdf entry for the checked Part IV predictions.

Relevance: Thomson's first major work and the mathematical root of his later atom models: the limited number of vortices per ring and Mayer's floating-magnet analogy return in the 1904 'plum-pudding' ring arrangements (Corpuscular Theory ch. VI; Electricity and Matter ch. V). The vortex-atom idea (Kelvin 1867) is a mechanical picture of matter as modes of a fluid, which Thomson later replaced with corpuscles bound by Faraday tubes (and which he echoes in a vortex analogy for bound mass).

Mainstream conflict: The vortex-atom theory of matter was abandoned (superseded by electron/nuclear atom and quantum theory). The hydrodynamic results themselves are standard far-field/first-order perturbation vortex dynamics, mostly correct as checked; the stated errors (eq. 71, the 7-vortex rounding, the 1/u factor) are mathematical, not conflicts with mainstream physics. Part IV makes definite predictions that differ from kinetic theory (thermal-effusion exponent m > 1; mean molecular speed of a monatomic gas falling as temperature rises); measurement gave the kinetic-theory result (see the treatiseonmotionofvotexrings.pdf entry).

A second, cleaner scan of the same 1883 Adams Prize essay (Macmillan; printed at the University Press, Cambridge by C. J. Clay and Son; preface dated Trinity College, 1 October 1883; Internet Archive scan digitised in 2007, 152 PDF pages, only 12 OCR pages so the text layer is better than the first copy). The digest gives the table of contents (Parts I-IV, paragraphs 4-60), the preface (notation altered after submission with help from G. H. Darwin, one of the examiners, who lent his notes; paragraph 60 and parts of 58-59 added after submission; thanks to Larmor for proof reading), the introduction, and a fuller view of Part IV (vortex atom theory of gases, paragraphs 56-60): a gas of vortex atoms has the pressure of ordinary kinetic theory, a monatomic gas gains energy only by increasing mean ring radius, a diatomic gas can gain energy by shrinking the core-line distance between its two atoms so mean velocity need not fall with temperature, and the ratio of paired time to free time governs combination and chemical affinity (nitrogen taken to have a very small ratio and combine poorly). Thomson calls it the only theory that tries to give a mechanism for intermolecular forces and says it is not fully developed. [read: digest chunks 01-07] QE REVISION: PART IV DEEP READ (pp.133-148). Thomson says a full dynamics of a vortex gas 'must form the subject of a future paper' and offers 'some results which can be obtained with very little additional calculation' [read: p.133]. (1) Boyle's law: from p = (2/3) sum(I V') he gets pv = T/3 - (1/6) rho (surface integral of (u^2+v^2+w^2) p' dS) over the wall; dropping the wall term is 'Boyle's law', keeping it makes pv 'a little less than the value given by Boyle's law, which agrees with the results of Regnault's experiments' [read: page image p.136]. The sign follows from his formula, but he derives no density dependence, discards a ring-ring term he himself calls positive (which would raise pv), and Regnault's hydrogen deviates the other way [memory, high]; his remark that elastic spheres 'cannot explain any deviation' was already wrong in 1883 (van der Waals 1873) [memory, high]. (2) Mean molecular velocity: a ring's energy rises with its radius while its speed falls, so for a monatomic gas 'the mean velocity of the molecules decreases as the temperature increases', unlike ordinary kinetic theory, though the mean momentum increases; he restricts this to monatomic gases and calls it a likely 'crucial experiment' [read: pp.136, 138]. (3) Thermal effusion: quoting Maxwell (Phil. Trans. 1879), kinetic theory gives p/p' = (t/t')^(1/2) for the pressures on two sides of a small hole; the vortex theory gives p/p' = (t/t')^m 'where m is a quantity greater than unity' [read: page image p.137]; the reader reproduced it (m = 1 + k, k about 0.5-0.85 for thin rings, so m about 1.5-1.9) [calc]. He himself warns the experiment would be difficult and that boundary effects are not fully treated [read: pp.137-138]. (4) Chemistry: molecules of a compound exchange partners; a compound exists when 'paired time' much exceeds 'free time'; disturbance lowers the ratio (dissociation); the ratio governs affinity, combination only between temperature limits, and a link to the dielectric strength of gases (Phil. Mag. June 1883); qualitative, no number computed. Two paired rings stay together only if their strengths are equal (checked, with a printing slip m for m'). (5) Valency: an atom is 1-6 linked rings of equal strength; a stable compound must divide into at most six 'primaries' of equal strength, 'No single atom would be capable of uniting with more than six atoms of another element', supported by Lothar Meyer and gaseous WCl6 [read: p.144]. His table makes nitrogen, phosphorus, arsenic, potassium, rubidium, thallium, bromine, chlorine, fluorine, iodine and hydrogen univalent; carbon (from CO), oxygen, sulphur and others divalent; aluminium, antimony, bismuth, boron, indium trivalent; silicon and tin quadrivalent (mercury undecided) [read: page image p.148]. NO and NO2 need a 'figure of eight' doubled nitrogen ring, which behaves 'like one with twice as many links' [read: page images pp.147-148]. By 1883 carbon was firmly tetravalent and nitrogen tri/pentavalent [memory, high]. The reader's bookkeeping shows the rule is applied inconsistently: HCN and CH3NH2 need primaries made of mixed atoms, and once those are allowed, his acetylene argument against tetravalent carbon and his cacodyl argument against trivalent arsenic both fail [calc]. 'PH5' (as printed) does not exist; PCl5 or PF5 is probably meant [memory, high]. (6) Spectra: low-valency elements should have simpler spectra (Na, K, Li, H, Cl cited) [read: p.148]; true of alkalis and hydrogen, not chlorine [memory, medium]. The 'at most six' rule depends on the Part III seven-vortex result, which is a rounding artefact (seven is neutral at linear order, eight first unstable) [read: deep/A-treatise-...-a73ca5.md III.6]. TESTS ON RECORD: Reynolds (Phil. Trans. 1879, porous plates) and Knudsen (1909-10, narrow tubes) found equilibrium at p1/p2 = sqrt(T1/T2), i.e. m = 1/2; Stern (1919-20) measured silver-atom beam speeds in agreement with Maxwell's law (speed rising as sqrt(T)) [web: sharpgiving.com/thebookofscience/items/p1879d.html; arxiv.org/pdf/2006.01574; frontiersin.org/journals/physics/articles/10.3389/fphy.2019.00208/full; nobelprize.org/uploads/2018/06/stern-lecture.pdf]. Thomson cites Maxwell but not Reynolds. The machine-checkable consequence: a temperature ratio of 2 gives a pressure ratio 1.41 (kinetic) against about 2.8-3.6 (vortex) [calc]. [read: deep/treatiseonmotionofvotexrings-0ee5cd.md]

Key claims

• The vortex-atom theory gives a mechanism for intermolecular forces that the kinetic theory with power-law attracting atoms does not.
• The pressure of a gas of vortex-ring atoms equals the pressure of the ordinary kinetic theory of hard elastic particles (Kelvin's argument as quoted); Thomson's own result adds a wall term that puts pv slightly below Boyle's value.
• The proportion of time two overtaking rings stay paired, relative to the time free, sets chemical affinity and the proportions of compounds formed.
• Thomson's own Boyle-law result: pv = T/3 minus a wall term, so pv is slightly below Boyle's value, which he says agrees with Regnault (true for air, N2 and CO2, not for hydrogen; no density dependence is derived).
• In a monatomic vortex gas the mean molecular velocity falls as the temperature rises (while the mean momentum rises), unlike ordinary kinetic theory; stated as a possible crucial experiment.
• Thermal effusion: the pressure ratio across a small hole between gases at temperatures t and t' is p/p' = (t/t')^m with m greater than 1 for a vortex gas, against m = 1/2 in kinetic theory (measured: m = 1/2, Reynolds 1879, Knudsen 1909-10).
• Valency equals the number of linked rings (at most six), a compound must divide into at most six 'primaries' of equal strength, which makes carbon divalent and nitrogen, phosphorus and arsenic univalent, with a 'figure-of-eight' doubled nitrogen ring for NO and NO2.
• Dielectric strength, affinity and dissociation of gases are governed by the ratio of paired time to free time of vortex rings (qualitative, no value computed).

Relevance: Duplicate of the other vortex-ring copy and the only source in the folder with a readable text layer for Part IV. Part IV is the clearest case in the folder of the 1883 theory making a definite, quantitative, falsifiable prediction that differs from kinetic theory (thermal-effusion exponent, speed falling with temperature), which was later measured; it also links the vortex 'links' of 1883 to Thomson's corpuscle-ring account of valency in 1904-07.

Mainstream conflict: The vortex-atom picture of matter and of chemical affinity is not part of accepted physics (superseded). Part IV's two sharpest predictions (m greater than 1 for thermal effusion; falling mean speed of a monatomic gas with temperature) were contradicted by measurement (Reynolds 1879, Knudsen 1909-10, Stern 1920); the valency table disagrees with the chemistry of 1883 and today and needs an ad hoc doubled nitrogen ring; the 'at most six' limit rests on a rounding artefact. Thomson himself says the theory is not fully developed.

The 701-page Cambridge University Press monograph that summarises the Cavendish programme on ionised gases (OCR on 700 pages; equations garbled, read from images by the deep reader in the chapter that was deep read). Chapters in the digest: I conductivity of gases in a normal state; II the conducting state; III mathematical theory of conduction with ions; IV effect of a magnetic field on ions; V determination of the charge-to-mass ratio e/m of an ion; VI the charge on the negative ion; VII physical properties of gaseous ions (cloud condensation); VIII ionisation by incandescent solids; IX flames; X photo-electric effects; XI ionisation by Roentgen rays; XII rays from radio-active substances; XIII emission of ionising radiation by the elements in general; XIV chemical action; XV spark discharge; XVI-XVII discharge at low pressure and its theory; XVIII the arc; XIX cathode rays; XX Roentgen rays; XXI properties of moving electrified bodies. The deep-read Chapter XXI gives: (i) the field of a uniformly moving charge (Heaviside's field), correct; (ii) field momentum of a slowly moving charge R = (2/3) mu e^2 w / a, so effective mass m = m' + (2/3) mu e^2 / a, correct in his electromagnetic units (the printed general-velocity closed form, eq. 1, fails its own slow limit); (iii) Kaufmann's 1901 radium beta-ray table (five points, speeds 2.36 to 2.83 x 10^10 cm/s, e/m falling from 1.31 to 0.63 x 10^7 emu/g) with Thomson's calculated mass ratios p agreeing with Kaufmann's p' within 10 per cent, taken to suggest that all the corpuscle's mass is electrical (radius about 10^-13 cm); the deep reader finds Thomson's p column cannot be reproduced from his stated model (8-14 per cent above the exact integral, within about 3 per cent of the later relativistic gamma factor) and that the same data fit gamma about as well; (iv) in Art. 355 Thomson himself states that corpuscles cannot supply all the mass of an atom (the number of corpuscles would be far larger than Arts. 161 and 192 allow); (v) the 'bound ether' picture of Faraday tubes; (vi) the kink derivation that a suddenly stopped charge emits a thin pulse, identified with Roentgen rays, with pulse energy (2/3) mu e^2 w^2 / delta and ratio 2a/delta to the field energy, and the Larmor power (2/3) e^2 f^2 / V; (vii) a radiation-reaction energy balance (Art. 363) with a time scale about 10^-23 s, not equivalent to the later Abraham-Lorentz equation. [read: deep/Conduction-of-electricity-through-gases-23223e.md; digest] QE REVISION (2026-10-09), three further deep reads. (A) e/m and e, Chapters V-VI [read: PDF pp.130-175]. Thomson prints two cathode-ray e/m values of his own, both dated 1897: 7.7 x 10^6 emu/g from magnetic plus electrostatic deflection ('mean of the values of m/e is 1.3 x 10^-7 or e/m = 7.7 x 10^6'; hence 'the hydrogen ion must be 770 times that of the ion in the cathode rays') and 1.17 x 10^7 from a heating (thermocouple) method which he calls 'not so reliable' [read: page images pp.134-139]. The master table (book p.145, PDF p.158) lists Lenard 1898 (6.39 x 10^6 and 6.8 x 10^6), Kaufmann 1897-8 (1.86 x 10^7), Simon 1899 (1.865 x 10^7), Wiechert 1899 (1.01-1.55 x 10^7), Seitz 1901 (6.45 x 10^6) and 1902 (1.87 x 10^7), Starke (1.84 x 10^7), Reiger and Becker 1905 (1.82 and 1.8 x 10^7), his own UV (7.6 x 10^6) and hot-carbon (8.7 x 10^6) values of 1899, Kaufmann's radium value at small velocities 1.77 x 10^7 (1901-2) and Ewers 1906 (1.7 x 10^7); the modern value is 1.76 x 10^7 [memory, high]. His own first value was therefore low by a factor of about 2.3, and the methods he criticised as giving too large a value (cathode-fall voltage) came closest. The charge e: Thomson 1898 cloud method 6.5 x 10^-10 esu in air and 6.7 x 10^-10 in hydrogen; 1901-2 with radium 3.4 x 10^-10; H. A. Wilson 1903 3.1 x 10^-10; modern 4.80 x 10^-10 [memory, high]. The '1/1700' (Art. 78, p.173) comes from e equal to the hydrogen-ion charge E plus e/m about 1.7 x 10^7 against E/M = 10^4; the earlier ratio '770' (Art. 59, p.135) is carried in the same book without a note that it was superseded. Internal oddities recorded by the reader: Thomson's UV value is 7.6 x 10^6 in the table and 7.3 x 10^6 in the text; Simon's potential range is 4860-11840 V on p.141 and 'about 1100 volts' on p.142. [read: deep section 'New section: e/m and e values']. (B) Chapter XI [read: PDF pp.333-344]. Art. 161 counts corpuscles per atom from Barkla's scattering datum, which this edition prints in words as 'about .0002 that of the primary' and in the displayed equation as (8 pi/3) N e^4/m^2 = .00025 (both printed on the same page; not reconciled) [read: page image p.339]. Chain: pulse model; Barkla's polarisation test; radiated energy of one accelerated corpuscle; opacity (8 pi/3) N e^4/m^2 independent of pulse thickness; Barkla's two laws; numbers: with e = 1.2 x 10^-20 emu and e/m = 1.7 x 10^7 'we find Ne = 10', ne = .4, so about 25 corpuscles per air molecule, 'approximately equal to its atomic weight' [read: page images pp.339-340]. Recomputed [calc]: Ne = 8.6 for .00025 (he rounds to 10) and 6.9 for .0002; 21.5 per air molecule with ne = .4 (about 20 with ne = .43), 17.2 for .0002; 14 with the modern charge (about 11 for .0002). Art. 160/163: Bumstead (Phil. Mag. [6] xi, 292, 1906) found that with equal absorbed Roentgen energy lead develops 'about twice' (Art. 163: 'nearly twice') the heat of zinc, measured with a radiometer; Thomson: 'a most interesting and important result, as it suggests that Roentgen rays may furnish a means of tapping the stores of inter-atomic energy'; the result 'suggests that the rays produce some change in the atoms'; he proposes keeping radium in contact with lead 'whether ... other elements would gradually be produced' and asks whether the many elements in radioactive ores are 'partly due to transmutation' [read: pp.333, 343-344]. Dorn's ejected corpuscles (10^9 cm/s) are argued to come from an 'explosion' of the atom, since direct action by the pulse would need nearly one calorie per cm^2 (an 'inadmissible amount') [read: page image p.334]. His own null searches: 'I have tried a large number of experiments to see if the incidence of the Rontgen rays on metals gives rise to any products which are radio-active ... I was, however, unable to detect any residual radio-activity in the metal after an exposure of several hours to strong Rontgen rays', and similar radium experiments by Bumstead (Proc. Camb. Phil. Soc. xiii, 125, 1905) and himself (ibid. p.124) 'with negative results'; no detector, sensitivity or number of trials is given [read: p.344]. (C) Chapter XX [read: PDF pp.657-662]. Marx (Ann. der Phys. xx, 677, 1906), by a Lecher-bridge phase method, found that the velocity of Roentgen rays through air equals the velocity of light; no number, distance or error is given. Haga and Wind (Wied. Ann. 68, 834, 1899; Wind, Phys. Zeit. ii, 292, 1901), with a V-shaped platinum slit a few thousandths of a mm wide, obtained shadow broadening 'which would be produced by diffraction, and which have not been explained in any other way'; light would need a wavelength of order 2 x 10^-8 cm to give it (exponent read from a small superscript in the page image, some residual doubt between 6 and 8), and 'If we regard the Rontgen rays as due to discontinuous pulses, this will be the order of the thickness of the pulse' [read: page image p.662]. Thomson thus reads the diffraction result as a pulse thickness (about 1/3000 of a sodium wavelength) rather than as a wavelength; he does not say whether Marx's result bears on the pulse picture. Fomm's 1896 bands are credited to contrast effects by Haga and Wind. Laue, Friedrich and Knipping (1912) later showed crystal diffraction of X-rays [memory, high]. CORRECTION from the Electricity and Matter deep read: equation (1) for the electromagnetic momentum is misprinted in this book (Conduction) only; in Electricity and Matter p.44 it is printed correctly (see that entry). Table labels: this book heads the 3.1 ... 1.5 column 'rho', observed by Kaufmann, and 3.09 ... 1.65 as calculated; Electricity and Matter (all printings) and the Corpuscular Theory label them the other way round. [read: deep/Conduction-of-electricity-through-gases-23223e.md]

Key claims

• A charged body moving at speed w carries electromagnetic field momentum (2/3) mu e^2 w / a, so a small charge behaves as if its mass were increased by (2/3) mu e^2 / a.
• The mass-to-charge ratio of the rays from radium (corpuscles) rises with speed in the way expected if all the corpuscle's mass is electromagnetic (Kaufmann's five points, calculated and observed ratios agreeing within 10 per cent).
• On the assumption that the whole corpuscle mass is electrical, the corpuscle radius is about 10^-13 cm.
• The corpuscles in an atom cannot account for all of the atom's mass (Art. 355): the number needed would exceed what the investigations of Arts. 161 and 192 allow.
• When a charged particle is suddenly stopped it emits a thin pulse of intense electric and magnetic force, which Thomson believes constitutes the Roentgen rays; thinner pulses are more penetrating.
• A charge accelerated by a force radiates power (2/3) e^2 f^2 / V, so a particle starting from rest under a steady force has initially zero acceleration and reaches half its mechanical value after a time of order 10^-23 s.
• Thomson's own first cathode-ray e/m was 7.7 x 10^6 emu/g (deflection, 1897) and 1.17 x 10^7 (heating method, 1897), against the modern 1.76 x 10^7; his 1898 cloud-chamber charge 6.5 x 10^-10 esu was revised to 3.4 x 10^-10 in 1901-2 (H. A. Wilson: 3.1 x 10^-10); the 1/1700 mass ratio comes from e/m about 1.7 x 10^7 divided by the hydrogen-ion 10^4, not from his 1897 numbers (which gave 1/770).
• Secondary Roentgen radiation counts about 25 corpuscles per air molecule (Barkla datum .00025, printed also as .0002 on the same page), 'approximately equal to its atomic weight' (21.5 with ne = .4 and Thomson's constants, 14 with the modern charge).
• Roentgen rays may release inter-atomic energy: lead absorbs about twice as much heat as zinc for equal absorbed X-ray energy (Bumstead 1906), which Thomson takes as a hint of change in the atom; he suggests testing for transmutation with radium in long contact with lead, and reports his own null search for radioactive products after several hours of X-ray exposure.
• Roentgen-ray diffraction (Haga and Wind) implies a length of about 2 x 10^-8 cm, which Thomson reads as the thickness of the pulse; Marx (1906) found Roentgen rays travel at the speed of light in air.

Relevance: Thomson's summary of the ionisation and corpuscle experiments and, in Chapter XXI, the theoretical core of his electromagnetic-mass and X-ray-pulse claims.

Mainstream conflict: The 'all mass is electrical' conclusion is not supported by later relativity (mass rise with speed is kinematic, gamma fits the same data) and the point-charge energy gives the 4/3 problem; Thomson's Art. 363 energy balance is not the Abraham-Lorentz law. The pulse picture is bremsstrahlung and is correct as far as it goes, but cannot give characteristic lines or the short-wavelength cutoff. The remaining chapters were not checked in this step. The e/m and e history is ordinary progress of measurement (Thomson's first value was low); the Barkla count with the modern charge gives about A/2, the atomic number (Moseley and Barkla later), not A; Roentgen-ray diffraction is wave diffraction of short-wavelength light (Laue 1912), with the pulse picture applying to bremsstrahlung only.

Six lectures given at Yale in May 1903 on 'the bearing of the recent advances made in Electrical Science on our views of the Constitution of Matter and the Nature of Electricity' [read]. Ch. I: the two-fluid and one-fluid theories of electricity called 'mathematical fictions'; Faraday's lines of force; the Faraday tube as unit of charge; moving tubes produce magnetic force; field momentum; Maxwell's vector potential from the momentum of a charge. Ch. II 'Electrical and bound mass': the mass of the 'bound ether' carried by Faraday tubes is 4 pi mu N^2 per unit volume; a vortex-column analogy (the mass trapped by a vortex column is proportional to the square of its strength); Heaviside's flattening of the tubes with speed; Kaufmann's radium table; Thomson's view that the whole mass of any body is the mass of the ether around it carried by the Faraday tubes of its atoms, which requires an ether density immensely greater than any known substance. Ch. III: the stopped-charge pulse as Roentgen rays; Faraday tubes as discrete threads in a continuous ether giving a 'fibrous' light wavefront (specks), used to explain why fewer than a billionth of the molecules in a Roentgen beam are split; radiation pressure and Arrhenius/Poynting; radiation loss from an orbiting charge, a rate that rises as the eighth power of speed. Ch. IV: atomic structure of electricity. Ch. V (deep read): primordial unit as an electrical doublet; atom as a uniform positive sphere (Kelvin's Aepinus atom) with corpuscles inside; equilibrium positions (two at half the radius, three at 0.577 of it, checked correct); stability of rings only up to 7 or 8 corpuscles; 'corpuscular temperature' of an atom; element evolution on cooling (hydrogen taken to contain about a thousand corpuscles); a table of radiation from n corpuscles on a ring (the n=5 row is a power of ten too large and the heading says 'each corpuscle' but the numbers are the group total); the estimate that one gram of hydrogen stores 1.02 x 10^19 erg (arithmetic correct, premise n = 1000 later disowned by Thomson himself in 1906); Mayer's floating magnets (the book says like poles repel 'inversely as the distance', a slip for inverse square) with a table of ring structures up to 42 magnets; a link to the periodic law and a qualitative link to spectra (nine vibration periods for three corpuscles; degeneracy splitting into doublets and triplets that widen with atomic weight). Ch. VI: radio-activity as atomic change, the energy balance of radium, and a discussion of an external-radiation (Le Sage-type) source. [CORRECTED by the QE deep read: Thomson places radium's energy inside the atom because radioactivity is a transient property, 'in all cases in which we have been able to localize it'; he states that 'the magnitude of the energy radiated from radium is not a valid argument against the energy being derived from radiation'; and he proposes that very penetrating Roentgen rays could replace Le Sage's corpuscles as a mechanism for inverse-square gravity. The earlier wording 'rejection of an external-radiation source' was too strong.] The pages are a scan of a Google Books copy; OCR is poor on equations. [read: deep/ELECTRICITY-AND-MATTER-1909-828b3e.md; digest] QE REVISION: CHAPTER II (pp.60-66) [read: page images 60-66]. Equation (1) is printed as I = (e^2/2a) V^2/(V^2-v^2)^(1/2) { theta(1 - (1/4)V^2/v^2) + (1/2) sin 2theta (1 + (1/4)(V^2/v^2) cos 2theta) }, sin theta = v/V, with mass due to the charge I/v; it is correct: it equals the exact integral of the field momentum of a uniformly moving point charge outside a sphere of radius a, and reduces to (2/3) e^2/a for slow motion [calc]. The misprint is in Conduction (a 1/2 coefficient that makes the formula blow up at low speed), not in Thomson's formula. Kaufmann's e/m table here: v = 2.83, 2.72, 2.59, 2.48, 2.36 x 10^10 cm/s, e/m = .62, .77, .975, 1.17, 1.31 x 10^7. Table II (p.65) heads the column 3.1, 2.42, 2.0, 1.66, 1.5 as rho, 'determined by equation (1)', and 3.09, 2.43, 2.04, 1.83, 1.65 as rho', 'found by Kaufmann in his experiments'. The 1904 and 1907 printings carry the same labels; the Corpuscular Theory (1907) also labels 3.09 ... as Kaufmann's; only Conduction (1906) labels them the other way. The reader found that neither column is reproduced by equation (1) at all five speeds (3.1 ... matches at the two lowest speeds and is 8-12.5% above the model at the top three; 3.09 ... is 7.5-13% above it everywhere), that the 3.1 ... column is within about 4% of (e/m)_0/(e/m)_v for one slow-speed value near 1.93 x 10^7 (the 3.09 ... column is not), and that the 3.09 ... column follows the later relativistic factor gamma within about 1% for V = 2.98 x 10^10 (Thomson could not have used gamma in 1903). The labelling is undecided from these pages; Kaufmann's 1901-03 papers are needed. Other points: Thomson's e = 1.2 x 10^-20 'in C.G.S. electrostatic units' is an emu value (slip); a = (2/3) e (e/m) gives 8 x 10^-14 cm, not the printed 5 x 10^-14; Kaufmann's electrical-mass fraction is 'about one-fourth' here and 'about 1/3' in Conduction. CHAPTER VI END (pp.167-179). Radium arithmetic: 100 cal/h = 4.2 x 10^9 erg/h; if the whole gram became alpha particles at 2 x 10^9 cm/s, the energy is (1/2)(1 g)(4 x 10^18) = 2 x 10^18 erg, giving 4.8 x 10^8 hours, 'about 50,000 years' (an upper limit), 'about 5 x 10^10 calories' in all; arithmetic correct [calc]. Real half-life of Ra-226 is about 1600 years [memory, high]. External source (p.174): 'We might suppose that the radium obtained its energy by absorbing some form of radiation which is passing through all bodies on the surface of the earth, but which is not absorbed to any extent by any but those which are radio-active'; plausible because Roentgen and radium rays penetrate 'several inches of lead'. Le Sage precedent (pp.175-177): 'ultra-mundane corpuscles' give gravity by shadowing; Maxwell's objection (the absorbed energy would keep bodies white hot) is answered by supposing the energy re-emitted as a more penetrating radiation. Roentgen rays replacing Le Sage's corpuscles (p.177): 'waves of electric and magnetic force possess momentum in their direction of propagation; we might therefore replace Le Sage's corpuscles by very penetrating Rontgen rays ... two bodies would attract each other inversely as the square of the distance ... If the absorption of these waves per unit volume depended only upon, and was proportional to, the density, the attraction ... would be directly proportional to the product of their masses'. Costs he states himself: changes in gravitation 'would be propagated with the velocity of light; whereas, astronomers believe they have established that it travels with a very much greater velocity', and the energy absorbed 'must be enormous - so great that the energy emitted by radium would be but an exceedingly small fraction'; his 'simple calculation' of the energy per gram is asserted, not shown. His reason for an internal source (pp.178-179): the transience of every localised radioactivity (Th X, emanations, induced activity last at most a few days), 'This is what we should expect on the view that the source of the radio-activity is a change in the atom; it is not what we should expect if the source were external radiation.' Gaps noted by the reader: the radiation postulated for radium (absorbed only by radioactive bodies) differs from that for gravity (absorbed by all matter); the parents' own energy source rests on the transformation hypothesis. MARGIN ANNOTATIONS: the project's copy of the 1904 printing (electricitymatter1904.pdf, PDF pp.69-75) carries a reader's margin annotations; see that entry. [read: deep/ELECTRICITY-AND-MATTER-1909-828b3e.md]

Key claims

• The whole mass of any body is the mass of the ether bound by the Faraday tubes of its atoms, so all mass, momentum and kinetic energy are those of the ether.
• Roentgen rays are thin pulses of intense electric and magnetic force emitted when a cathode-ray corpuscle is suddenly stopped.
• An atom is a sphere of uniform positive electricity containing many negative corpuscles whose stable arrangements (rings) change in steps as the number increases, and this recurrence may account for the periodic law and for series and doublet/triplet structure in spectra.
• A symmetric ring of corpuscles radiates very little (six corpuscles at one tenth the speed of light radiate less than one five-millionth of what one does), so a classical atom of many corpuscles can be stable against radiation until perturbed.
• Atoms store energy far beyond chemical energy: one gram of hydrogen, if it contains 1000 corpuscles within a radius of 10^-8 cm, holds about 10^19 erg, enough to lift a million tons more than a hundred yards.
• Radio-activity is a change inside the atom (Thomson's reason: localisable radioactivity is transient); an external penetrating radiation as the energy source is discussed and not excluded on magnitude grounds, and penetrating Roentgen rays are proposed as a replacement for Le Sage's corpuscles in a theory of gravity.
• Equation (1) of Chapter II, the electromagnetic momentum of a moving charge, is correct as printed in Electricity and Matter, and gives mass (2/3) e^2/a at low speed.
• A gram of radium emitting 100 calories per hour could last at most about 50,000 years if its whole mass turned into alpha particles at 2 x 10^9 cm/s, emitting about 5 x 10^10 calories in all.
• Very penetrating Roentgen rays could replace Le Sage's corpuscles: absorption would give momentum transfer, an inverse-square attraction proportional to the product of the masses if absorption is proportional to density, with gravity then travelling at the speed of light.
• Radium's energy output is 'not a valid argument' against its being drawn from external radiation; Thomson places the source in the atom because radioactivity is transient.

Relevance: Thomson's popular but central statement of the electromagnetic-mass view and of the first model atom; the 1903 lectures predate his 1906 revision of the number of corpuscles per atom.

Mainstream conflict: The uniform positive sphere was replaced by Rutherford's nucleus (1911); 1000 corpuscles for hydrogen was disowned by Thomson (Phil. Mag. 1906); classical ring/magnet shell structure does not give the quantum shell periods; the 'all mass is bound ether' claim depends on a dense ether that relativity discarded. The classical radiation-suppression of symmetric rings is correct electrodynamics.

The Charles Scribner's Sons 1907 issue of the Silliman lectures (193 PDF pages, OCR on 192). The deep read is an edition check: the 1907 title page is Scribner 1907 but the imprint page still reads 'Copyright, 1904 By Yale University / Published, March, 1904', and no second-edition, revision or new-preface statement appears. A bag-of-words comparison of the body text gives 1907 vs 1909 98.7 per cent and 1904 vs 1907 and 1909 about 92 per cent (the gap is OCR noise in the 1904 scan). Same chapters (I to VI, last printed page 162), same preface signed Cambridge, Aug. 1903. The deep reader concludes the 1907 and 1909 printings are reprints of the 1904 Yale text under new imprints; subtle unannounced corrections are not excluded by this check. The Silliman foundation notice records an 1883 legacy of eighty thousand dollars for lectures on the 'presence and providence, the wisdom and goodness of God' in nature, the fund reaching Yale in 1902 with this volume the first of the series. [read: deep/Electricity-and-matter-jjthompson-1907-d22ed9.md] QE REVISION: Table II of the Kaufmann comparison appears at PDF p.67 (printed p.48) of this printing, with the same labels as the 1904 and 1909 printings: 3.1 ... is rho 'determined by equation (1)' and 3.09 ... is rho' 'found by Kaufmann'. [read: page image, deep/ELECTRICITY-AND-MATTER-1909-828b3e.md]

Key claims

• The 1907 Scribner printing carries the same text as the 1904 Yale printing (no stated revision).

Relevance: Establishes that the three printings in the folder are one text, so the content of Electricity and Matter is covered by the 1909 deep read and need not be read three times.

Mainstream conflict: none (bibliographic)

The 1904 first printing of the Silliman lectures (189 PDF pages; 18 OCR pages so most of the text layer is original type and cleaner than the other two printings). Title page: Charles Scribner's Sons, New York, 1904; imprint 'Copyright 1904 by Yale University. Published March 1904'; the Silliman Foundation notice; preface signed Aug. 1903; contents Ch. I Representation of the electric field by lines of force (p.1), II Electrical and bound mass (p.30, probably 36), III Effects due to the acceleration of Faraday tubes (p.68, probably 58), IV The atomic structure of electricity (p.71), V The constitution of the atom (p.90), VI Radio-activity and radio-active substances (p.140). The text opens 'My object in these lectures is to put before you in as simple and untechnical a manner as I can...' and ends at printed p.162, identical in content to the 1907 and 1909 printings. The deep reader also used its 600 dpi page image to read the blurred surd for the three-corpuscle radius ('.57 times the radius', correct value 1/sqrt 3 = 0.577) and confirmed the same plates. [read: digest; deep/Electricity-and-matter-jjthompson-1907-d22ed9.md] QE REVISION (C7): this PDF is a composite and carries a reader's margin annotations in the project's copy (PDF pp.69-75). They record that p.72 is missing from the original scan and that the next page (p.73) was inserted from a different (1907) printing of the book (the inserted page carries printed number 56); one (p.75), beside the passage on the pulse, produced by stopping the particle, that constitutes the Roentgen rays, links that passage to an external cause of radioactivity (see the Electricity and Matter 1909 entry, Chapter VI end). The annotations are not Thomson's text and not primary evidence. Table II at PDF p.64 carries the same labels as the 1907 and 1909 printings. [read: text layer; page images; deep/ELECTRICITY-AND-MATTER-1909-828b3e.md section G]

Key claims

• Same claims as the 1909 printing (see that entry); the 1904 printing is the original publication of these claims.

Relevance: Earliest printing of the lectures and the cleanest scan; the page image for the n=3 surd was taken from it.

Mainstream conflict: same as the 1909 printing

A 3-page note in the Proceedings of the Royal Society. Apparatus: Ruhmkorff coil and exhausted bulb inside a tin-plate-covered packing case, a hole about 1 inch across covered with aluminium or tin foil, an electrified plate about 2 inches above the hole joined to a quadrant electrometer with a quartz-fibre needle (no magnetic control). Method: charge the plate to high potential, disconnect the quadrant pairs, expose to the rays. Results [read]: a violent leakage from the plate whether it is positively or negatively charged; an uncharged plate acquires no charge; the leak is a more delicate detector than any photographic plate known to him and persists through a zinc plate a quarter inch thick; leakage is detectable only when the bulb is at a pressure low enough that the positive column has almost disappeared; leakage occurs with the plate only 3 or 4 volts above the case and he found no lower limit; unlike ultra-violet light (Elster and Geitel) both signs discharge, any metal works, and it happens when the plate is surrounded by solid paraffin, sulphur, ebonite, mica or paraffin oil, where insulation is otherwise practically perfect at 10 to 15 volts; little difference in leakage along and across the rays through paraffin. Conclusion: dielectrics traversed by the rays become conductors, 'a splitting up of its molecules ... resembling that by which a current passes through an electrolyte'. All results are qualitative (no currents, pressures or rates). The page also records that a telegram from Borgman and Gerchun (St Petersburg) reporting that the rays discharge electricity, and a letter from Lodge, were read after his paper. [read: deep/On-the-Discharge-of-Electricity-produced-by-Rot-Rays-ad79e8.md]

Key claims

• Air and solid or liquid insulators become conductors of electricity while Roentgen rays pass through them.
• Roentgen rays discharge a charged plate of either sign and of any metal, unlike ultra-violet light, which discharges only negative charge and only clean electropositive metals.
• The Roentgen-ray leakage occurs at potential differences as low as 3 or 4 volts with no lower limit found, and is strong only when the bulb is highly exhausted.
• The passage of the rays through a substance is accompanied by a splitting up of its molecules enabling electricity to pass as through an electrolyte.

Relevance: First experimental step in Thomson's gas-ionisation programme that led to the 'ions' and then 'corpuscles' of the later books.

Mainstream conflict: none; the conclusion is the accepted discovery that X-rays ionise gases and other matter. The deep reader notes the 3 to 4 volt 'no threshold' result corresponds to ionisation-chamber behaviour at low field [memory medium].

Longmans, Green and Co. monograph (163 PDF pages, OCR on 162; page numbers differ from PDF by 23). The apparatus is a discharge tube with a perforated cathode through which positive particles pass into crossed electric and magnetic fields and onto a photographic plate; each kind of particle traces a parabola with e/m = y^2 A / (x B^2), so the mass of unknown particles is read relative to the hydrogen line (m2/m1 = y1^2/y2^2 at equal x, equal charges); multiply charged atoms show a prolongation to 1/2, 1/3, 1/4 of the normal distance. Claimed advantages over spectroscopy: gas of about 0.01 cc, one per cent of helium easily seen, 3 x 10^-6 cc of helium in 1 cc of air detected. Findings read by the deep reader: (1) atmospheric residues give xenon (128 as Thomson gives it), krypton 82, argon 40, neon 20, and a line of atomic weight 22 'just under the neon line'; he rules out doubly charged CO2 (removing carbonic acid made no change) and writes 'It must, I think, be a new element', while also noting a double-charge prolongation and that Mendeleev's table has no room unless a group of elements sits near neon. F. W. Aston at the Cavendish tried to separate it from neon by fractional absorption on cooled coconut charcoal (no difference in proportion, tested by line intensity and a quartz density balance) and by diffusion through a clay pipe stem (some enrichment, no spectral difference); Thomson then raises 'the suspicion that the two gases, although of different atomic weights, may be indistinguishable in their chemical and spectroscopic properties', compared with radio-lead and thorium, without using the word isotope. (2) A line at 3 times the hydrogen mass, 'X3', is produced by cathode-ray bombardment of metals, salts, minerals and meteorites; he excludes C4+ (no multiple-charge signature, carbon compounds do not give the line) and concludes it is triatomic hydrogen H3, argued from its persistence (potash bombarded for months without falling off), its association with hydrogen, its combining with oxygen under light or sparking, its reaction with mercury vapour under discharge and with hot copper oxide, and its stability for weeks; no spectral lines for X3 were found in an electrodeless discharge. The deep reader notes the modern assignment of the 3 line to H3+ (formed in the discharge) and possibly HD+, and leaves the mapping to reviewers. (3) Helium, and sometimes neon, appear when metals, minerals and salts are bombarded; controls reported: onset only after about an hour of bombardment and growing with time, repeated dissolving and evaporating of salts (ten or twelve times, with undiminished helium from KI), a half-aluminium blank that had or had not been exposed to helium, and independence of the cathode. He weighs three explanations (absorbed helium, disintegration of the atoms, helium-compound impurities), offers a 'loosened alpha-particle' hypothesis and states that 'the view that helium can be got from other chemical elements raises questions of such a fundamental character that few will be prepared to accept it until every other explanation has been shown to be untenable'. Priority note: first published account of helium in vacuum tubes was Ramsay (Nature, 18 July 1912). The deep reader records, with a [memory, medium] tag, that similar 1913 reports by Collie and Patterson were challenged by Strutt and others and the episode is generally ascribed to air contamination, to be verified by reviewers. [read: deep/Rays-of-Positive-Electricity-and-Their-A-a3a643.md; digest] QE REVISION: items from the earlier chapters (digest only, not deep read, not weighed by the reviewers): (a) Thomson argues from carbon-monoxide and phosgene counts, nickel carbonyl and H2 + Cl2 under light that the atoms in compound molecules are not oppositely charged [read: digest pp.84-86]; the Corpuscular Theory (1907) argues the opposite for like atoms in O2 (oppositely charged, used to explain the paramagnetism of oxygen) [read: digest p.155 of that book]. (b) Mercury ionises by losing 'one or eight' corpuscles and nothing in between; 'Eight corpuscles is a very large number for an atom to lose' [read: digest pp.66-70]. A reader should treat these as unreviewed. Nelson's identification of the mass-3 line as 'tritium' (sites.md) is another author's characterization, recorded there.

Key claims

• A parabola method with crossed electric and magnetic fields identifies the mass of positive-ray particles by comparison with the hydrogen line, from samples of about 0.01 cc of gas.
• A positive-ray line at atomic weight 22 beside neon 20 in atmospheric residues is due to a new element (or to a form of neon indistinguishable from it by charcoal absorption and spectrum).
• The line at three times the hydrogen mass (X3) is triatomic hydrogen H3, manufactured by cathode-ray bombardment of hydrogen-containing substances, and is a stable form more persistent than ozone.
• Cathode-ray bombardment of metals, minerals and salts liberates helium (and sometimes neon) at a rate not accounted for by absorbed air, possibly by disintegration of the bombarded atoms or by loosened alpha-particles.
• Multiple charges on atoms of one chemical group increase with atomic weight (Ne 2, Ar 3, Kr 4 in the plates shown). [From the earlier chapters, digest only; no reviewer judged this claim.]

Relevance: Thomson's own later experiments: the method that led to Aston's mass spectrograph and the discovery of stable isotopes (neon-22), plus the contested 1913 helium/neon-from-bombardment results, which bear on transmutation claims.

Mainstream conflict: The 22 line is accepted as neon-22, a stable isotope (Aston 1919) [memory high]; the 3 line is attributed to H3+ rather than neutral H3, which is not stable [memory high]; helium and neon from bombardment of ordinary materials are not accepted as transmutation, being attributed to contamination or occlusion [memory medium, to verify].

Seven chapters, expanding a spring 1906 Royal Institution course (193 PDF pages, OCR on 192; printed pages are PDF minus 15). Ch. I corpuscles in vacuum tubes: negative carriers have a mass about 1/1700 of hydrogen and are the same from any source; positive carriers are never lighter than a hydrogen atom. Ch. II electromagnetic mass: m = m' + (2/3) e^2 / a (algebra checked), velocity factor alpha, Kaufmann's table (Thomson's alpha column agrees with the observed ratios at the top three speeds within 1 to 2 per cent and 10 per cent low at the bottom two; the deep reader's recalculation of his stated model differs by 8-12 per cent at three speeds), 'all the mass is electrical', corpuscle radius 10^-13 cm (printed 'e = 10^20' is a misprint for 10^-20 emu), 'each corpuscle may be said to extend throughout the whole universe' (true of the field, but half the mass lies within 2a), and the implicit 3/4 m c^2 field energy (the 4/3 problem). Ch. III: Heaviside field; stopped-corpuscle pulse as Roentgen rays (magnetic pulse formula is missing a factor e); Larmor radiation; the Zeeman effect with Lorentz's circular-orbit derivation (omega = omega0 +/- (1/2) He/m, checked), the conclusion that the light of Zeeman lines comes from corpuscles not atoms because the shift is proportional to e/m about 10^7, and classical multi-corpuscle models of the anomalous patterns. Ch. IV-V: metallic conduction by free corpuscles (equilibrium theory) and the 'second theory' of doublet chains. Ch. VI part 1: positive sphere with corpuscles; derived equilibria and Q for 1 to 4 (Q = 0.9, 2.1, 3.6, 5.4 e^2/a, correct), octahedron stable for 6, cube unstable for 8; planar ring stability up to 5 in a hollow ring and a table i = number needed inside (n up to 40), the greedy graphical rule and a ring table N = 1 to 100 (printing error at N = 23/24; several entries are metastable not ground states); Mayer's floating magnets; the 'outer ring 20' series giving valencies that sum to 8 (Thomson says the agreement with chemists' 8 is 'quite accidental'; the text swaps the two valencies of the 62 group). Ch. VII number of corpuscles in an atom: Method I Barkla secondary X-rays (0.00025 of the primary; N = 25 n, equal to atomic weight; with Thomson's inputs 21.5 and with modern e about 14), Method II cathode-ray opacity (2 to 4 corpuscles per unit atomic weight from his formula), Method III refraction and dispersion (hydrogen gives n near 1; Drude's counts a few per atom) reconciled by splitting corpuscles into a few loose valency corpuscles and many tightly bound ones; answer to the spectral-line objection with Darwin periodic orbits and Boscovich belts; atomic mass attributed mainly to the positive electricity; a table of (mu-1) for gases (Cuthbertson and Metcalfe). [read: deep/The-corpuscular-theory-of-matter-ab57d0.md; digest] QE REVISION. KAUFMANN TABLE: this book (PDF p.48, printed p.33) heads 3.09, 2.43, 2.04, 1.83, 1.65 as 'the values found by Kaufmann' and 3.1, 2.42, 2.0, 1.66, 1.5 as alpha 'calculated on the preceding hypothesis', the same labelling as Electricity and Matter; only Conduction (1906) labels them the other way (see the Electricity and Matter 1909 entry). CHAPTER I (deep read, pp.16-42): Thomson gives e/m = 1.7 x 10^7 for the corpuscle, the hydrogen ion 10^4 (ratio 1,700), e = 3.1 x 10^-10 esu = 10^-20 emu (the text layer reads '8.1'; the page image reads 3.1) and m = 6 x 10^-28 g; Townsend 'found e = E'. For positive rays he reports, from his own experiments with a phosphorescent screen: at not-too-low pressure a band; at very low pressure and strong field two patches with maximum e/m about 10^4 and about 5 x 10^3, 'the same two types of carriers whatever kind of gas is in the tube' (air, CO, O2, H2, He, argon, neon); the 5 x 10^3 value 'corresponds to the positive particles which are given out by radio-active substances'; his 'most obvious interpretation' is that all gases give off positive particles with the mass of the hydrogen atom or twice it, 'independent of the nature of the gas'; and 'We know of no case where the mass of the positively charged particle is less than that of an atom of hydrogen' [read: page image p.38]. Evidence is qualitative ('the appearance of the phosphorescence was the same') with no spread, field or pressure figures; helium bands at 10^4, 5 x 10^3 and 2.5 x 10^3 at other pressures show the universality is limited to very low pressure and strong field. Alpha particles carry e/m 5 x 10^3, taken as helium with a double charge. Today the e/m 10^4 and 5 x 10^3 patches read as H+ and H2+ (or He2+) from hydrogen and helium present [memory, high]. CHAPTERS IV-V (deep read, pp.64-117): metallic conduction by free corpuscles in temperature equilibrium; Ohm's law; the Wiedemann-Franz ratio k/c = (4/3) a^2 theta/e^2 (6.3 x 10^10 absolute units at 0 C, in Thomson's terms 3 (k_B/e)^2, the classical value; Jaeger and Diesselhorst's tabulated values for pure metals are 6.65-7.35 x 10^10, within about 10 per cent); Lorentz's radiation constant a = 1.2 x 10^-16 observed against 1.32 x 10^-16 from theory (but 'the same expression ... whatever may be the mass or charge'); Peltier and Thomson effects; a Hall-effect argument that negative corpuscles give the sign seen in bismuth and silver, with iron, cobalt and tellurium of the opposite sign explained by a collision term, rejecting two-carrier explanations. THE SPECIFIC-HEAT CONTRADICTION (pp.98-100): Rubens and Hagen's long-wave reflection requires the interval between collisions to be under 3.3 x 10^-15 s, hence n greater than 1.8 x 10^24 free corpuscles per cc of silver; each carries energy a theta, so they would absorb about 6 cal per degree per cc, whereas 'to raise the temperature of a cubic centimetre of silver one degree, only requires about 0.6 calories' for the atoms and corpuscles together: 'We thus get to a contradiction.' [read: page image p.100]. Checked: correct [calc]. It is the classical electron specific-heat problem, resolved only by Fermi-Dirac statistics (Sommerfeld 1927) [memory, high]. CONCLUSION (p.90, written before the specific-heat page): the free-corpuscle mechanism is 'at most only a part and not the whole of the process of metallic conduction', on grounds including the large resistance change at fusion without a matching thermoelectric change; Chapter V develops 'the second theory' of corpuscles hopping between polarised doublets and ends: 'the results would hold if the carriers were bodies having a much greater mass than that of a corpuscle', which qualifies the p.70 argument that heavy carriers are excluded. Errors: printed a = 1.5 x 10^-16 on p.100 against 1.32 x 10^-16 on p.65; the digest's contact-potential formula should read V = (2/3)(a theta/e) log(N1/N2). Thomson's atom count for silver (1.6 x 10^23 per cc) is about 2.7 times the modern value. [read: deep/The-corpuscular-theory-of-matter-ab57d0.md]

Key claims

• The mass of a corpuscle (electron) arises entirely from the electromagnetic field around it, m = (2/3) e^2 / a, with a radius about 10^-13 cm.
• Corpuscles are the same from any source, about 1/1700 the mass of hydrogen with e/m about 10^7 emu/g, and are a constituent of all atoms.
• The light in Zeeman-effect spectral lines comes from vibrating corpuscles, not atoms, because the splitting is proportional to e/m (Lorentz's circular-orbit derivation).
• An atom of N corpuscles in a uniform positive sphere has stable ring arrangements that change at certain N, so recurring ring structure and two-sided valency (positive and negative valencies summing to a constant) imitate the periodic law.
• The number of corpuscles in an atom is of the order of the atomic weight (secondary X-ray scattering, cathode-ray opacity and dispersion), so most of an atom's mass belongs to the positive electricity.
• Roentgen rays are thin pulses emitted when a corpuscle is stopped.
• At very low pressure and strong field every gas gives the same two types of positive carrier, with e/m about 10^4 (mass of a hydrogen atom) and 5 x 10^3 (twice that, the same e/m as alpha particles); no positive carrier is lighter than a hydrogen atom, whereas the negative carrier is about 1/1700 (Thomson later dropped the universality claim).
• A classical free-corpuscle gas in a metal explains Ohm's law and the Wiedemann-Franz ratio, but the number of free corpuscles needed for the observed optical behaviour of silver would absorb about 6 cal per degree per cc against a measured 0.6 cal total (Thomson's own contradiction; the electron specific-heat problem solved by Fermi-Dirac statistics in 1927).
• Free-corpuscle conduction is 'at most only a part' of metallic conduction; a second theory (corpuscles hopping between polarised doublets) gives the same results and 'would hold if the carriers were bodies having a much greater mass than that of a corpuscle'.
• The sign of the Hall effect in bismuth and silver is that of negative corpuscles; the opposite sign in iron, cobalt and tellurium comes from a magnetic collision term, not from positive carriers (qualitative).

Relevance: The fullest statement of Thomson's corpuscle/plum-pudding atom in the folder, including all of the arithmetic of the 'number of electrons per atom' argument.

Mainstream conflict: Mass velocity-dependence is relativistic not purely electromagnetic; uniform positive sphere replaced by nucleus (Rutherford 1911); electrons per atom is about A/2 (atomic number) by later work; ring/valency numbers do not survive a full numerical treatment and the outer-20 coincidence is called accidental by Thomson; anomalous Zeeman patterns need spin, not multi-corpuscle modes; the Lorentz normal Zeeman triplet is correct.

Volume III of Harper's Scientific Memoirs (Harper and Brothers, New York, 1899, ed. J. S. Ames; translated and edited by George F. Barker; 94 PDF pages, OCR on 9). Contains Roentgen's memoirs (digest only), G. G. Stokes's Wilde Lecture 'On the Nature of the Roentgen Rays' (delivered 25 July 1897, Manchester Literary and Philosophical Society) and J. J. Thomson's 'A Theory of the Connection between Cathode and Roentgen Rays' (Phil. Mag., Feb 1898, signed 16 Dec 1897; the editor omitted the mathematics). Stokes (who did no experiments himself) argues that cathode rays are streams of electrified molecules, that each striking molecule launches a single ether pulse of opposite-signed halves (zero net area) so the Roentgen emanation is 'a vast succession of independent pulses' not a train of waves; he explains the near-absence of diffraction by cancellation of oblique contributions (shell thickness assumed 'comparable with the millionth of an inch', giving a beta of 0.002 inch, a shadow-edge zone of 1/250 inch), the absence of refraction by the lack of time for ether-molecule harmony, higher vacuum and faster molecules as thinner pulses and harder rays, and Becquerel rays as intermediate. Thomson's paper: the field of a moving charge, a suddenly stopped particle starts a thin electromagnetic pulse of thickness equal to the particle diameter (or 2a + VT for collision time T) in which the magnetic force varies inversely with thickness and goes as sin(theta), brightest at right angles to the cathode rays for a single stopping encounter; absorption by ions proportional to the pulse thickness, so thin pulses are penetrating; agreement with Stokes that the rays are impulses, not short waves. The deep reader found the Stokes numeric arithmetic correct and a misprint in Thomson's ion-absorption energy (mass missing from the denominator, not affecting the conclusion), corrected the digest on a date (professorship 1884, not 1881) and a sphere radius label. [read: deep/rontgen-rays-f49feb.md; digest]

Key claims

• The Roentgen emanation is a vast succession of independent thin pulses of ether (electromagnetic) disturbance, one per cathode particle stopped at the target, rather than a periodic train of short waves.
• A pulse thinner than a millionth of an inch (Stokes's assumed value) would diffract only within about 1/250 inch of a shadow edge for 4-inch geometry, accounting for sharp shadows.
• A suddenly stopped charged particle emits a thin pulse whose magnetic force is proportional to sin(theta), is greatest at right angles to the cathode beam, and varies inversely with pulse thickness (Thomson).
• The absorption of the pulse by ions in matter is proportional to its thickness, so thinner pulses (from faster particles, more abrupt stops) penetrate more.
• The rays are not refracted because the pulse is gone before the ether and the molecules of glass can vibrate in harmony.

Relevance: Gives the primary Stokes and Thomson statements of the pulse theory of X-rays that Thomson repeats in the 1906-07 books, with Stokes's independent numerical estimate.

Mainstream conflict: The pulse picture is classical bremsstrahlung and is correct as far as it goes, but X-rays also show a wave nature (Laue 1912, diffraction by crystals) and characteristic lines and the Duane-Hunt cutoff that need quantum theory [memory high]; Stokes's polarisation discussion and ether-molecule refraction theory belong to the ether era.

Charles Griffin and Company, London, 1909, Vol. I of the Poynting and Thomson Text-Book of Physics, fifth edition, preface signed J. H. P. and J. J. T., January 1909. An elementary experimental text, chapters I to XVIII: weight and mass, acceleration of gravity (Huygens, Kater, Bessel, geodesy), gravitation and the mean density of the earth (the gravitational-constant chapter drawn from Poynting's work), elasticity, strain, torsion, bending of rods, spiral springs, impact (Hertz), compressibility of liquids, gas laws, thermal effects of strain, capillarity and Laplace's theory, diffusion of liquids and gases, viscosity of liquids. Includes a section on experiments on the qualities of gravitation (gravitation is not selective, not affected by the medium, not directive: Mackenzie's quartz-crystal test) and Newton's pendulum experiment on mass. No electrical theory, corpuscles or ether in the digested lines I read; Thomson's role as co-author is not distinguished from Poynting's in the digest. [read: digest] QE REVISION: the digest (p.214) records the kinetic-theory thermal-effusion law in a book Thomson co-signed: for a porous diaphragm between gases at equal pressure and different temperatures, molecular speed is proportional to the square root of temperature, the number flux is inversely proportional to the square root of absolute temperature, 'so gas streams from cold to hot'. This is the law that Thomson's 1883 vortex-gas prediction (p/p' = (t/t')^m, m greater than 1) contradicts. The advertisement pages (digest p.272) list Theo Kassner, 'Gold Seeking in South Africa' (Griffin), probably unrelated to E. E. W. Kassner of the Ternion project; noted only as a possible name lead. [read: digest]

Key claims

• Gravitation is not selective, not screened or affected by an intervening medium, and not directive in crystals (as tested by Poynting and by Mackenzie's crystal-sphere experiment).
• Weight is proportional to mass at the same place and mass is conserved in chemical change (Landolt's sealed-tube experiments).
• Thermal effusion: in a gas at equal pressure on both sides of a porous diaphragm, flux is inversely proportional to the square root of absolute temperature and gas streams from cold to hot (kinetic-theory law, stated in a book co-signed by Thomson).

Relevance: Peripheral: a standard textbook co-authored by Thomson, relevant only as evidence of his conventional teaching; no part of the project's central claims rests on it.

Mainstream conflict: none (standard physics of the period)