Physicist review: Frank W. Very, The Luminiferous Ether (1919)

Reviewer: PhD physicist (Step 2). Source tags: [read] from the document or project files; [web] with URL; [memory] with confidence. Page numbers are PDF pages (printed page + 4).

Source note: the only document is the 1919 pamphlet. Equations on PDF pp.37-38 were read by me from the page image (I rendered p.38 myself) [read: page image p.38]. PDF pp.1-35 and 43-60 were read via text/digests, and I re-read pp.28, 43-44, 50-51 directly. Fessenden's and Very's earlier papers are known to me only through the Investigator's notes (not downloaded to research/files/; none are there).

1. The mainstream view in brief (for the reader)

• Light. Maxwell's equations give electromagnetic waves in vacuum at c with no medium. Special relativity (1905) removes the need for an ether: c is the same in every inertial frame. Quantum theory adds that the field's energy comes in quanta, E = h nu (h = Planck's constant, nu = frequency). A photon is a quantum of the field, not a little sphere. It shows interference (waves) and single-point detection (particles) [memory, high].
• Electron. A point-like particle with intrinsic spin 1/2 (angular momentum hbar/2, hbar = h/2pi). No size has been found. Penning-trap and g-2 work limit any radius to about 10^-22 m [web: https://www.physicsforums.com/threads/how-does-a-penning-trap-determine-the-upper-limit-of-the-electron-radius.301175/ (a forum summary of Dehmelt's analysis; I did not reach the primary paper)]. A positive twin, the positron, was found by Anderson in 1932 [memory, high].
• Atoms and nuclei. Hydrogen is one proton plus one electron. The proton is about 1,836 electron masses. Nuclei are made of protons and neutrons (neutron found 1932). Mass is not "all electrical" [memory, high].
• Gravity. Newtonian gravity as the weak-field limit of general relativity (1915). It is attractive at all scales tested. Galaxies in groups and clusters attract. A large-scale acceleration (dark energy, found 1998) exists but has no link to Very's mechanism [memory, high].
• Entropy. The second law is statistical. Starlight absorbed by a medium heats it; it does not vanish [memory, high].

2. Corrections to Step 1

1. The key number is confirmed on the page image. p.38 prints "2.999 x 10^10 / (2 pi x 1.9 x 10^-13) = 1.660 x 10^22 revolutions per second". The deep read was right: the true value is 2.512 x 10^22 (I recomputed it). The printed figure implies r = 2.875 x 10^-13 cm. Coincidence worth noting, not a finding: that is within 2% of the classical electron radius e^2/(mc^2) = 2.82 x 10^-13 cm, and c/(2 pi r_classical) = 1.69 x 10^22. Very may have used a different radius or miscalculated; the book does not say.
2. The surface-pressure check is confirmed, with one nuance. E^2/8pi with e = 4 x 10^-10 esu and a = 7.5 x 10^-14 cm gives 2.0 x 10^32, matching Fessenden's printed figure. So his "diameter" was used as a radius. The Investigator found that Fessenden's own 1900 text says "diameter" (from IA OCR; I did not see the page image). The slip therefore originates with Fessenden and Very inherited it. Very's remark that the revised (Millikan) radius makes the pressure "more nearly 10^33" goes the wrong way: the pressure scales as a^-4, and the larger radius gives 7 x 10^30 dyn/cm^2.
3. The 69 MeV figure. h x 1.66 x 10^22 s^-1 = 1.1 x 10^-4 erg = 69 MeV (I recomputed). The deep read's "69 keV" belongs to the 1.66 x 10^19 Hz value (the "1000 times atomic frequency" check). Both are right; do not mix them.
4. Claims.md, claim 4 and the facts.json summary disagree. The summary says aberration is explained "with no appeal to Rayleigh/Brace", the key-claims list says Rayleigh and Brace show the ether is bound to the Earth. Both appear in the book (pp.36-37). Also, claims.md says "no space-filling light-bearing ether". On p.36 Very writes "We need not abandon the idea of a universal interstellar atmosphere, but it is not the atmosphere which conveys light". The aura is space-filling; what is denied is that it carries light as waves. The light is carried by ether-particles, and a "bound" ether shell is attached to matter. I think of this as three layers: aura (universal), bound ether (on atoms), ether-particles (light). Claim 1 is stated with that in mind.
5. Hydrogen = 1,844 electrons (p.28) [read]. Modern ratio of hydrogen-atom mass to electron mass is 1,837 [memory, high]; 1,844 is a period value. Very's hydrogen nucleus is 922 electrons spinning one way and 921 the other, plus one satellite (p.28). Uranium "about 439,000" is 238 x 1,844 = 438,872 [computed]; with the modern ratio it is 437,000.
6. Claim 7 and the 1909 paper. The Investigator's Very 1909 lead (matter destroyed into radiation and reproduced elsewhere) shows the entropy claim predates the book by nine years. This matters for claim 7 only as history.

3. Claim-by-claim

Claim 1: light is discrete electron-sized ether-particles at c, each with one Planck quantum; a separate universal "aura"

a. Mathematics. There is none for light as particles. Very gives no equation for a particle's energy, momentum, or how it yields wavelength. Internally the picture is vague rather than inconsistent. One tension: Very says the particles are "elastic spheres" electron-sized (p.37) and also carry exactly one quantum, but the quantum h nu depends on frequency while the size is fixed. The book never says how one size gives radio waves (kilometres) and X-rays. b. Experiments. Against: (i) Young-type interference and diffraction, including single-photon-at-a-time interference (G. I. Taylor 1909, low-intensity light) [memory, medium; not checked on the web]; (ii) polarisation and radio waves, whose wavelength greatly exceeds the supposed particle; (iii) emission-type "corpuscular" ideas in which light speed depends on source speed fail: de Sitter double-star argument (1913) and Alväger et al. 1964, where gamma rays from pions moving at 0.99975 c travelled at c to within 10^-5 [web: https://en.wikipedia.org/wiki/Tests_of_relativistic_energy_and_momentum; the double-star argument: https://en.wikipedia.org/wiki/De_Sitter_double_star_experiment]. Very's particles move at c "with the electron's surface speed", i.e. speed is fixed by the shed sheath's rotation (p.40), which is a source-linked speed; the book does not say that it is source-independent in the observer's frame. For: the photoelectric effect, Compton scattering, and the 1916 Millikan test show light delivers energy in lumps. Very is right that light has a particle side. He was writing in 1918, before Compton (1923) [memory, high]. c. Tested? Very proposes no test. His specific mechanism (a photon is the shed sheath of an overturned electron) has no measurable consequence stated. Photon properties have been tested massively, but not Very's structure. d. If true: relativity's c-invariance, Maxwell's wave theory in free space, and quantum electrodynamics' photon would need replacing, all of which predict accurately (QED to about 10 digits in g-2) [memory, high]. Verdict: Tested, not supported (strong form: light as sized elastic particles of an ether); Not testable as stated (the aura as a universal medium, with no measurable property given). The weak form, that light has particle character and that quantum energy exchange is lumpy, is Established, but this was Planck/Einstein's physics, not Very's.

Claim 2: the Planck quantum is the energy to overturn an electron gyroscopically and reverse its sign

a. Mathematics. None given. Very says only that the overturn costs "one of Planck's least quantities of energy" (p.39-41) [read]. The Investigator found that the note Very cites (Fessenden 1914, Science 39:533) supplies a link: work to upset the gyroscope is proportional to angular velocity, so to frequency [web: https://archive.org/stream/jstor-1640236/1640236_djvu.txt, via the Investigator; I did not open it]. That is a fair credit to Fessenden: he at least writes E proportional to nu. But Very's own text states a fixed energy for every overturn, which sits badly with E = h nu varying over 10^8 in frequency (radio to gamma). A test: if the overturn energy were tied to the electron's own rotation (the only frequency Very supplies, 1.66 x 10^22 s^-1) then h nu = 69 MeV, about 135 times the electron's rest energy (0.511 MeV) [computed]. Atomic photons are eV to keV. So the one number the book offers makes the claim inconsistent with atomic spectra by a factor of 10^4 to 10^7. b. Experiments. A classical gyroscope cannot reverse "electric sign": charge is conserved and an electron does not turn into a positron by spin reversal (that is pair-related physics at >1 MeV). A spin flip of an electron in a magnetic field costs ~10^-4 eV and emits a microwave photon, not a sheath. Nothing in Very ties photon emission to charge-sign change, and atomic transitions leave the electron's charge unchanged [memory, high]. c. Tested? No test proposed or performed. d. If true: QED's account of emission (atomic transition between energy levels, photon created by the field) would be replaced. Verdict: Not testable as stated (qualitative), and Refuted by calculation in its only numerical reading (rotation frequency x h = 69 MeV vs eV-keV observed).

Claim 3: the electron is a rotating vortical sphere of aura; surface at c; 1.660 x 10^22 rev/s; surface pressure ~10^32-33 "dynes"; volume elasticity 10^73

a. Mathematics. Checked with the page image (p.37-38) [read: page image]. - The rotation rate is wrong by arithmetic: 2.512 x 10^22, not 1.660 x 10^22. - "Same speed at all surface points" (p.38): a sphere's latitude circle has radius r cos(lambda). A fixed surface speed c means omega = c/(r cos lambda), which diverges at the poles. A rigid rotating sphere cannot have this: a rigid sphere has v = omega r cos(lambda), maximal at the equator. So it is not a rigid rotation, and "angular velocity increases from the equator to the poles" describes a flow the book never specifies. [computed; the text states the premise, the divergence is my inference] - The pressure is not in dynes. Stress is in dyn/cm^2. Fessenden's 2 x 10^32 is E^2/8pi with the diameter used as radius (reproduced). - Volume elasticity = (pressure)^(3/2)/area. Units: (dyn cm^-2)^(3/2) / cm^2 = dyn^(3/2) cm^-5, not dyn/cm^2. The exponent arithmetic (10^48 / 10^-25 = 10^73) is right as written. The justification ("the ratio of volume to surface being that of the cube to the square") is no derivation. Worked check: with proper stress ~7 x 10^30 dyn/cm^2 at Millikan's radius (or 3 x 10^32 at the smaller one), a bulk modulus of the same order as the pressure would already stop collapse in a simple model; there is no apparent need for 10^73, and the number has no stated meaning. - "About 1,000 times the frequency of atomic vibrations" (p.38): atomic optical lines are 10^14 to 10^15 Hz, soft X-ray lines 10^16 to 10^17 Hz. The ratio is 10^5 to 10^8, not 10^3 [memory, high]. b. Experiments. - Spin check I computed: a solid sphere of electron mass and radius 1.9 x 10^-15 m with equatorial speed c has L = (2/5) m r c = 2.1 x 10^-37 J s. The measured spin is hbar/2 = 5.3 x 10^-35 J s: 250 times larger. A classical spinning electron of that size cannot supply the observed spin without exceeding c. This is the old classical-electron-spin problem (Uhlenbeck-Goudsmit/Pauli, 1925-26) [memory, high], so a rotating-sphere model faces it; Very predates it. - Size: the electron has no measured size; bound ~10^-22 m, versus 1.9 x 10^-15 m (the proton is about 0.84 fm, so Millikan's value is bigger than the proton) [web: Penning-trap limit above]. - The Thomson/Millikan "radius" 2e^2/(3mc^2) = 1.88 x 10^-13 cm is the radius at which classical electromagnetic energy would equal mc^2 [computed]; it is not a size seen in scattering. c. Tested? Not directly. High-energy electron scattering (LEP Bhabha scattering and others) looks for compositeness and finds none down to far below 10^-16 m [memory, medium; the search result https://arxiv.org/pdf/hep-ph/0002172 is a LEP2 Bhabha paper, not checked in detail]. d. If true: the point-like Dirac electron, QED's precision g-2 agreement, and the Standard Model's lepton sector would have to change. Verdict: Internally inconsistent (arithmetic error in 1.660e22; units of pressure and elasticity; fixed surface speed incompatible with rigid rotation) and Refuted by calculation (spin: a c-limited sphere carries about 1/250 of the observed spin). The instructive part: Very took Fessenden's classical field-energy electron seriously and tried to give it a mechanical structure; the ideas of a rotating charge and an electromagnetic origin of mass are the pre-1925 mainstream.

Claim 4: bound-ether shell + corpuscular light explain Michelson-Morley, Rayleigh/Brace, aberration

a. Mathematics. None; qualitative. Internal tension: a shell of ether dragged completely by Earth with free light-bearing ether outside, and corpuscles moving at c in "free" ether. Aberration under full drag fails, as explained below, but Very reconciles it by saying corpuscles are unaffected by the dragged shell (Stokes 1845 corpuscular reading). b. Experiments. - Michelson-Morley null: Special relativity explains it without any medium. The ether-entrainment reading fails other tests: Sagnac 1913 (rotating interferometer sees a fringe shift), Michelson-Gale 1925 (Earth rotation measured), Lodge's whirling discs 1890s (no drag), and aberration [web: summary https://en.wikipedia.org/wiki/Hammar_experiment lists these as refuting full ether entrainment near masses and, for large masses, aberration and Michelson-Gale-Pearson; the primary papers I did not read]. - Fizeau 1851 found partial drag in moving water (coefficient 1 - 1/n^2), which neither full drag nor none reproduces; relativistic velocity addition does [memory, high]. - Rayleigh 1902 and Brace 1904 found no double refraction from motion through the ether; the Lorentz contraction/relativity predicts null here too [memory, high]. - A corpuscular theory with source-dependent speed is refuted (Alväger, de Sitter; see Claim 1). Very's corpuscles travel in a bound-ether shell, so the claim is partly a hybrid. c. Tested? The experiments he cites are real nulls and are accepted; his interpretation has not been tested separately. Kennedy-Thorndike (1932), Ives-Stilwell (1938) and later tests of Lorentz invariance support relativity [memory, high; not rechecked]. d. If true: special relativity would have to be replaced by a preferred-frame theory, and Sagnac and Michelson-Gale would need a separate explanation. Verdict: Historically superseded but instructive for the observation that these nulls fit a bound-ether picture (a respectable 1900-1920 reading, close to Stokes/Lodge); Tested, not supported for the full-drag form (Sagnac, Lodge, Michelson-Gale, aberration).

Claim 5: one kind of electron, no positive electron; all mass electrical (H = 1,844 electrons, U nucleus ~439,000)

a. Mathematics. The count is arithmetic from mass ratio. The hydrogen nucleus is 922 + 921 electrons of opposite rotation and a net charge of +1 left after the satellite (p.28) [read]. Ordinary charge counting does not work as stated: a "reversed" electron must have charge +e, so 922 of them plus 921 negative gives net +e only if the satellite is counted; this is consistent if +922e - 921e = +e. The count of rest masses is then 1,843 + 1; an exact bookkeeping could be made. I did not find an inconsistency in that sum. But binding energy and fusion/fission mass defects (few tenths of a percent) then have no accounting. b. Experiments. - Anderson's positron (1932) is a positive particle with the electron's mass; pair creation shows it is a distinct state, and annihilation gives two 0.511 MeV photons [memory, high]. - Rutherford scattering (1911) shows a compact nucleus of charge Ze, radius ~10^-15 m. Confining electrons there needs momenta p ~ hbar/r. For r = 7.4 fm (uranium), hbar c / r = 27 MeV per electron [computed], 50 times the electron rest energy. Electrons cannot stay in such a nucleus; the nuclear-electron problem was settled by the neutron (1932) and nuclear spin/statistics (nitrogen-14) [memory, high]. - The proton is about 1,836 electron masses but is a composite of quarks (deep inelastic scattering) with mass mostly from QCD binding energy, not 1,836 electrons [memory, high]. - Fessenden also held a nucleus of negative electrons [Investigator, via 1914 note]. Very's view is not original with him; the documented source is Fessenden. c. Tested? Yes, by the discoveries above. d. If true: the positron, the neutron, quark structure and nuclear physics would have to be reinterpreted. Verdict: Tested, not supported (positron, neutron, quarks). Reasonable in 1918 (before 1932): Historically superseded but instructive. The conclusion for the reader: "all mass is electromagnetic" was a serious program (Abraham, Lorentz, Kaufmann's mass-velocity data, 1901-1909) that failed because the binding forces (Poincare stresses) and nuclear forces are needed.

Claim 6: gravity is synchronous pulsation of electrons in the aura; confined to a "galactic cell"; galaxies repel at intergalactic distances

a. Mathematics. Very cites Hicks, Burton, Leahy on pulsating spheres in an incompressible fluid (pp.50-51) [read]. In that theory in-phase pulsating spheres attract with an inverse-square law (Bjerknes) [memory, medium], but, as Leahy notes (quoted by Very), beyond a quarter wavelength they repel. Very's answer is to set the quarter-wave to the radius of a "galactic cell" and calls the predicted repulsion a "powerful argument" in favour (p.51). That is an ad hoc rescaling: an adjustable length is tuned to fit the claim. No calculation of the force coefficient is given, so the Newtonian constant G is not derived. Also the electrons' pulsation needs to be phase-locked across the galaxy, the energy supply to be unexplained (he says "infinite source ... which is spiritual", p.44). b. Experiments. - Gravity tracks mass, not electron number or phase, to high precision (Eotvos-type tests; the equivalence principle is tested to ~10^-13) [memory, high]. - Galaxies attract: the Milky Way and Andromeda approach each other; clusters are bound [memory, high]. Very's evidence for repulsion is that "the relative crowding of the galaxies" argues mutual repulsion (p.43; p.47) [read], i.e. he reads clustering as repulsion; the observation of clustering, if anything, fits attraction. In 1918 it was not settled that "white nebulae" were galaxies; the Curtis-Shapley debate was 1920 [memory, high]. - Today's accelerating expansion is a different effect (a uniform repulsion from dark energy), not a galaxy-to-galaxy force. - Gravitational-wave detection (LIGO, 2015) shows gravity propagates at c as a field wave [memory, high; no URL checked]. c. Tested? Very's galactic-cell pulsation is not tested. He later wrote on Nipher's 1920 experiment; the Investigator could not read it (paywalled). Nipher's claim of an electrical effect on gravity was not replicated as far as I know [memory, low]. d. If true: general relativity and Newtonian gravity's universality, the equivalence principle and cluster dynamics would have to be rewritten. Verdict: Never properly tested (the pulsation mechanism, no test was proposed or performed); Tested, not supported for the specific claim of galaxy-to-galaxy repulsion on the "crowding" evidence; Historically superseded but instructive for the Bjerknes pulsation analogy.

Claim 7: matter is gradually converted to radiation and re-originates from the aura; entropy is not one-sided

a. Mathematics. No equations in the book; the argument is qualitative: "how would it be possible for the universe to endure unless both operations were actual?" (p.43) [read]. The earlier 1909 paper has dm/m0 = v/(V - v) for mass added by radiation [Investigator, via IA]; I did not check it. Worked check on "light absorbed by the aura gives matter". The Milky Way radiates about 10^10 times the Sun's 3.8 x 10^26 W = 3.8 x 10^36 W. By E = mc^2 that is 4.2 x 10^19 kg/s, or 6.7 x 10^-4 solar masses a year [computed]. Compared with the Galaxy's star-formation rate of order 1 solar mass per year [memory, medium], this is small, so the mass accounting does not forbid the idea, but nothing then says matter reappears as hydrogen atoms. b. Experiments. - The standard objection to light absorbed in space (Olbers' paradox) is that an absorber heats up and re-radiates, so it does not remove energy (Herschel 1848) [memory, high]. Very's answer is an absorption that goes into building matter; no such process or flux is observed. - The Steady-State theory (Hoyle, Bondi and Gold, 1948) proposed continuous creation of matter; the CMB (1965) and counts of radio sources refuted it [memory, high]. - The second law is verified in the sense that closed-system total entropy has not been seen to decrease. Entropy increase is a statistical law, so a "building-up region" that is itself entropy-lowering needs a heat sink; the cosmos does contain local ordering (stars, planets) paid for by radiated entropy. - Shapley's 1915 Messier 13 result (no selective absorption) is quoted by Very, who says the aura absorbs non-selectively (p.43 footnote) [read]. Non-selective absorption of starlight is limited by supernova brightness-distance relations (inverse-square dimming) [memory, medium]. c. Tested? Not by Very. Not tested as a specific mechanism. d. If true: energy-momentum conservation would need a matter-creation term (as in Steady State), and cosmology's CMB would be unexplained. Verdict: Never properly tested as stated (no measurable rate or mechanism), with the cosmological form (steady-state creation) Tested, not supported (CMB). The classical point (matter converting to radiation in stars) is Established.

4. The Investigator's findings, weighed

Kept here whatever the verdict: - Very 1909 (Science 30:491): first statement of claims 3, 4 and 7, with matter as "vacuities in an exceedingly dense ether, kept from collapsing by the centrifugal pressure of a rapidly rotating shell" and refraction via "bound ether attached to the molecules". It shows that the 1919 model is the end of a ten-year line, not a late improvisation. Evidence cited: Kaufmann's mass-velocity data (mass rising with speed), Lebedew and Nichols-Hull light pressure. These are real experiments; light pressure is Established [memory, high]. Kaufmann's electron data were later explained by relativity, not electromagnetic mass alone. - Fessenden 1900 (gravity, inertia), 1913, 1914: the true source of the numbers (2 x 10^32, the "breaking strain", the quantum-as-gyroscope idea). The diameter/radius slip and the "dynes" wording are Fessenden's. Fessenden's glass-prism test (stress 30 x 10^6 vs Kerr-equivalent 25 x 10^6 dyn, which he himself calls "probably accidental") is an experimental claim, interesting as a stress-optic (photoelastic) effect; the Kerr comparison is a coincidence of units I would not rely on [memory, medium]. His 1914 line that work scales with frequency fixes the weak point I flagged in Claim 2. - Fessenden's report of rejection (1890-91, editorial "probably wrong"): relevant to how heterodox ideas were treated, not to Very's physics. Weigh as context only. - Very's Norwood meteorite episode (1910): he first reported a fall, then exposed it as a fraud. It shows he corrected his own early report publicly. Relevant to character, not to the physics. - Historical record: reception of the 1919 pamphlet, funding after Lowell's death (1916), and the observatory's fate are unknown. Absence of a review is an anomaly worth analysis, not a neutral gap. Competing explanations: (a) a Swedenborgian-press publication that never reached physicists (New-Church Review reprint); (b) physicists considered it outside their field (relativity and Bohr/Rutherford were absorbing attention); (c) personal funding ended with Lowell. The AAS meeting of 20 Aug 1918, Popular Astronomy and the Pitt/Lowell archives would decide. None found yet. - Not yet read: NCR 1913 "Scientific Estimate of Swedenborg's Principia", the 1918 NCR printing, Very's 1920 Nipher paper, the second IA scan of the book. The IA scan could confirm the 1.660 figure from a second print, but my p.38 image already reads 1.660 clearly.

5. What a fair reader should take away

• Very is a competent observational astronomer (bolometry with Langley; Wikipedia: https://en.wikipedia.org/wiki/Frank_Washington_Very [Investigator]). The book is a theoretical speculation; its arithmetic contains checkable slips (1.660 vs 2.512; units).
• The strongest idea, that light has particle properties, is correct but was Planck/Einstein's. The Michelson-Morley null is the one place the author is close to a viable reading of 1918 evidence (a bound-ether picture), later displaced by relativity and the Sagnac family.
• The most instructive failure is the rotating electron: it fails the spin test by a factor of about 250 and has no unit-consistent pressure.

Verdict summary

Web checks done: electron size bound, Alväger/de Sitter, ether-entrainment experiments (3 searches; secondary sources, primary papers not opened). Unchecked [memory]: Taylor 1909, LEP compositeness details, Nipher non-replication, SN Ia dimming.