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

Study document (Step 4). It adds no new verdicts; every verdict here is the arbiter's (arbiter.md). Source tags: [read] from the pamphlet (PDF page = printed page + 4); [read: page image p.N] from a rendered page (I re-rendered PDF pp.37-38 and confirmed the printed numbers); [web] with URL; [memory] from training, with confidence; [computed] arithmetic in example.py.

Source note: the project folder holds one document, the 1919 pamphlet (a Google Books scan; 63 of 64 pages were OCR). There is no source gap for the central text. The exponents on PDF pp.37-38 are garbled in the text layer, so those numbers come from page images. Fessenden's 1900-1914 Science notes and Very's 1909 Science paper are not in the folder; the Investigator found them on the Internet Archive and the reviewers read them there. Very's 1913 and Oct 1918 New-Church Review papers, his 1913 "What becomes of the Light of the Stars?" and his 1920 note on Nipher were read by nobody, so conclusions about them are open, not negative. After the QE round (qe.md), PDF pp.11-59 have all been deep-read (deep/The-Luminiferous-Ether-eb1408.md), so no part of the pamphlet now rests on a one-line digest alone; the arbiter's "QE revision" section weighs what those reads found, and this primer follows it.

Who he was: Frank Washington Very (1852-1927) was an American astronomer. He worked with Samuel Langley at the Allegheny Observatory on infrared and lunar-temperature measurements and later ran the Westwood Astrophysical Observatory in Massachusetts, with Percival Lowell as patron [web: https://en.wikipedia.org/wiki/Frank_Washington_Very; Lowell patronage per the Investigator]. Langley's 1890 lunar paper used Very's observations and "for unknown reasons omitted his name" (same Wikipedia page). He read this paper to the American Astronomical Society on 20 August 1918 [read, per facts.json]. He wrote inside a Swedenborgian framework: the paper was first printed in the New-Church Review (the journal of the church founded on Emanuel Swedenborg's teachings), it elaborates his 1913 "critique and appreciation of Swedenborg's Principia", and it advocates "a return to Swedenborg's scientific doctrine of a limited ether" [read: p.9]; the word "aura" is "borrowed from Swedenborg" [read: p.11]. The Principia is Swedenborg's 1734 natural-philosophy treatise [memory, high]. Whether Very was himself a member of the New Church is not stated in the pamphlet; the Investigator found a 1900 New-Church Messenger note on a paper by him (sites.md), so his Swedenborgian writing goes back at least to 1900.

Who Fessenden was (he appears throughout): Reginald Aubrey Fessenden (1866-1932), a Canadian-born engineer and radio pioneer, known for early amplitude-modulated voice radio transmission (1906) [memory, high]. Very borrows his electron numbers and his idea of an elastic ether from Fessenden's Science notes of 1900-1914 (section 4).

---

1. The idea in one paragraph

Very wanted to keep what was right about the old idea of an "ether" (a medium in space) while accepting the new evidence (1900-1918) that light comes in discrete lumps of energy, Planck's quanta. His answer has three layers. A thin universal medium, which he calls the "aura" (Swedenborg's word), fills all space but does not carry light. Every electron is a tiny sphere of aura spinning so fast that its surface moves at the speed of light, and every atom carries a "bound" shell of ether with it. Light is a stream of separate "ether-particles", each about the size of an electron, made when an electron flips over like an upset gyroscope and throws off its outer skin; the energy of that flip is Planck's quantum. From this picture he argues that the Michelson-Morley experiment found no "ether wind" because each atom carries its own ether; that all atomic mass is made of electrons (hydrogen = 1,844 of them); that gravity is a synchronised pulsing of electrons in the aura and stops at the edge of each galaxy, so galaxies repel; and that matter turned into starlight is eventually rebuilt from the aura, so the universe need not run down.

---

2. What you need to know first

Gaussian (CGS) units. Very works in centimetres, grams, seconds. Force is the dyne (1 dyn = 1 g cm/s², 10⁻⁵ N), energy the erg (1 erg = 1 dyn cm = 10⁻⁷ J), pressure dyn/cm² (0.1 Pa). Charge is in esu (electrostatic units, also called statcoulombs). In these units Coulomb's law has no 4πε₀: F = q₁q₂/r², the field of a point charge is E = q/r², and the energy stored per unit volume in an electric field is E²/8π (in SI it is ε₀E²/2). The electron charge Very quotes from Millikan is e = 4.774 × 10⁻¹⁰ esu (= 1.59 × 10⁻¹⁹ C). Read: Griffiths, Introduction to Electrodynamics (4th ed.), Appendix C on units; E. M. Purcell, Electricity and Magnetism, 2nd ed. (1985), which uses Gaussian units throughout (the 3rd ed. with D. Morin switched to SI) [memory, high].

Energy stored in an electric field. Bringing charge together costs work, and that work is stored in the field. For a total charge e spread on the surface of a sphere of radius a, the stored energy is U = e²/(2a). The smaller the sphere, the larger the energy. Griffiths ch. 2 (electrostatic energy); Feynman, Lectures on Physics Vol. II ch. 8.

Pressure, stress and bulk modulus. A stress is a force per unit area (dyn/cm²). The bulk modulus K, which Very calls "volume-elasticity", says how hard a material resists being squeezed: K = −ΔP/(ΔV/V). Because ΔV/V has no units, K has the units of pressure. Water has K ≈ 2 × 10¹⁰ dyn/cm²; steel about 1.6 × 10¹² [memory, high]. In an elastic medium of density ρ, compression waves travel at √(K/ρ) and shear waves at √(μ/ρ), where μ is the rigidity (shear modulus). Any physics text on elasticity or waves; Halliday, Resnick and Walker cover the bulk modulus and the speed of sound.

Rigid-body rotation and angular momentum. A body spinning at angular velocity ω (rad/s) has angular momentum L = Iω, where I is the moment of inertia: I = (2/5)ma² for a uniform solid sphere, (2/3)ma² for a thin spherical shell. A point at distance r from the axis moves at speed ωr. A gyroscope resists having its axis tipped; tipping it takes torque and, if done against resistance, work. Taylor, Classical Mechanics, ch. 10.

Waves, interference and the photon. When two waves overlap, they add; where crests meet troughs they cancel. Two slits a distance d apart, lit by light of wavelength λ, give bright fringes spaced λL/d on a screen at distance L. The fringe spacing depends on λ: this is the basic evidence that light behaves as a wave. Since Planck (1900) and Einstein (1905) we also know light's energy is delivered in lumps E = hν (h = Planck's constant, 6.626 × 10⁻²⁷ erg s; ν = frequency). Modern physics holds both at once: the photon is a quantum of the electromagnetic field, detected at points but distributed by a wave rule. Any intro modern physics text (Krane, Modern Physics; Serway, Moses and Moyer).

Special relativity, the minimum. Einstein (1905): the laws of physics are the same in every frame moving at constant velocity, and light's speed is c in every such frame regardless of the source's motion. There is then no need for, and no way to detect, a medium at rest. Griffiths ch. 12; Taylor and Wheeler, Spacetime Physics.

Topology in one sentence (the "hairy ball theorem"). You cannot comb the hair on a ball flat everywhere without leaving at least one tuft or bald spot. In mathematical terms: every continuous field of arrows lying along the surface of a sphere is zero somewhere. This matters for Very's claim that every point of the electron's surface moves at the same speed c.

---

3. The textbook view

3.1 Light and the ether, 1800-1919

Thomas Young and Augustin Fresnel showed in 1801-1820 that light interferes and diffracts, so it is a wave [memory, high]. Polarisation showed the wave is transverse (it vibrates across its direction of travel). In the 19th century a wave was assumed to need a medium, the "luminiferous ether". Transverse waves need a medium with rigidity, like a solid, yet planets move through it freely; Very quotes Stokes on this "heavy strain on our credulity" [read: page image p.37].

Maxwell (1865) showed light is an electromagnetic wave. In empty space his equations give a wave equation

∇²E = (1/c²) ∂²E/∂t²,

where E is the electric field, ∇² the Laplacian (sum of second space derivatives) and c = 1/√(ε₀μ₀) ≈ 3.00 × 10¹⁰ cm/s. The wave needs no material medium; the fields themselves oscillate.

Aberration. Because the Earth moves at v ≈ 30 km/s around the Sun, a telescope must be tilted forward by an angle α ≈ v/c ≈ 10⁻⁴ rad = 20.5 arcseconds to catch starlight (Bradley, 1729) [memory, high]. With corpuscles (particles) this is like tilting an umbrella in rain while walking: simple. With waves in an ether it depends on how the ether moves near the Earth. Stokes (1845) proposed that the Earth drags the ether along; Fresnel (1818) that the ether is almost stationary and matter drags it only partly.

Michelson-Morley (1887). If light moves at c relative to an ether at rest, an interferometer moving at v through that ether should see the light take slightly different times along two perpendicular arms. Turning the instrument by 90° should shift the interference pattern by

ΔN ≈ 2Lv²/(λc²),

where L is the arm length, λ the wavelength and ΔN the shift counted in fringes. With L = 11 m (the folded path), v = 30 km/s, λ = 5.9 × 10⁻⁵ cm: ΔN ≈ 0.37 fringe [computed in mathematician.md and arbiter.md; I re-derived: 2 × 1100 × (3 × 10⁶)² / (5.9 × 10⁻⁵ × (3 × 10¹⁰)²) = 0.37]. They saw at most a few hundredths [memory, high]. Rayleigh (1902) and Brace (1904) also found no double refraction caused by motion through the ether, to about v²/c² ≈ 10⁻⁸ [memory, high].

Three responses were on the table: (i) the Earth drags the ether completely (Stokes-type); (ii) moving bodies contract along their motion by √(1 − v²/c²) so the effect cancels (FitzGerald 1889, Lorentz 1892); (iii) special relativity (1905), which makes the null result a consequence of the principle of relativity. Tests that separate them: - Fizeau (1851): light in moving water is dragged only by the fraction (1 − 1/n²) of the water's speed (n = refractive index), not fully, and not at all in air [memory, high]. Against complete drag by matter. - Michelson-Gale-Pearson (1925): a 600 m × 340 m interferometer detected the Earth's rotation by the Sagnac effect, 230 against 237 parts per 1,000 predicted [web: https://en.wikipedia.org/wiki/Michelson%E2%80%93Gale%E2%80%93Pearson_experiment]. Against an ether dragged around with the rotating Earth. - de Sitter (1913) and successors: if light from a star in a binary orbit left at c relative to its source, the light from the approaching and receding phases would arrive at different times and the orbits would look distorted. No such distortion is seen; X-ray binaries bound any source-speed dependence below 2 × 10⁻⁹ (Brecher 1977) [web: https://en.wikipedia.org/wiki/De_Sitter_double_star_experiment]. Against "emission" theories where light's speed depends on the source.

Quanta. Planck (1900) needed E = nhν for oscillators to get the black-body spectrum. Einstein (1905) proposed light itself comes in quanta E = hν; Millikan's photoelectric measurements (1914-16) confirmed eV_stop = hν − W (V_stop = stopping voltage, W = work function) [memory, high]. Very cites this relation in Millikan's form, h/e = ΔV/Δν [read: p.19]. Compton (1923) showed the quanta carry momentum hν/c. Yet interference experiments with light so faint that only one quantum is in the apparatus at a time still build up ordinary fringes, one dot at a time; the clean modern version uses true single-photon sources (Grangier, Roger and Aspect, 1986) [memory, high]. (G. I. Taylor's 1909 faint-light experiment is often cited as the first such test [memory, medium; not checked on the web in this review].) So each quantum obeys a wave rule.

3.2 The classical extended electron, 1881-1915

J. J. Thomson (1881) found that a charged sphere moving slowly carries field momentum as if it had extra mass

m_em = 2e²/(3ac²),

where e is the charge, a the radius, c the speed of light. Note this is (4/3) × U/c², with U = e²/(2a) the field energy; the stray 4/3 was a known puzzle [memory, high; Feynman Vol. II ch. 28 discusses it]. If all of the electron's mass is electromagnetic, solving for a gives

a = 2e²/(3mc²) ≈ 1.9 × 10⁻¹³ cm

with m = 9.0 × 10⁻²⁸ g [computed, example.py step 1]. This is "Millikan's radius" that Very uses. It is a model-dependent number, not a measured size. (The modern "classical electron radius" e²/(mc²) = 2.82 × 10⁻¹³ cm drops the 2/3.)

A charged sphere pushes itself apart: the outward stress on its surface is

P = E²/8π, with E = e/a² just outside,

in dyn/cm². Something must hold it together. Poincaré (1906) added non-electrical inward stresses ("Poincaré stresses") for this; Fessenden (1900) instead had the ether's elasticity do it [memory, high for Poincaré; web for Fessenden, below]. This line of work (Abraham, Lorentz, Poincaré) was serious physics until about 1915-1925.

What happened next: in 1925 Uhlenbeck and Goudsmit found the electron has intrinsic angular momentum (spin) ħ/2 = 5.27 × 10⁻²⁸ erg s, where ħ = h/2π [memory, high]. Quantum mechanics treats the electron as point-like; experiments (Penning traps, g − 2) find no structure, bounding any radius near 10⁻²² m = 10⁻²⁰ cm [web, secondary: physicist.md cites a forum summary of Dehmelt's analysis]. The positron (Anderson 1932) and the neutron (Chadwick 1932) were found; nuclei are protons and neutrons, about 10⁻¹³-10⁻¹² cm across [memory, high].

3.3 Gravity, galaxies and entropy, briefly

Newtonian gravity, F = Gm₁m₂/r² (G = gravitational constant), is the weak-field limit of general relativity (1915); it is attractive at every scale tested, including between galaxies in groups and clusters [memory, high]. Andromeda approaches us at about 300 km/s (Slipher measured it in 1912) [web: https://en.wikipedia.org/wiki/Andromeda_Galaxy]. The accelerating expansion found in 1998 is a separate, cosmological effect.

C. A. Bjerknes (1870s-1880s) showed in water that two spheres pulsating in phase attract and out of phase repel, roughly as 1/r²; this "pulsation gravity" analogy was popular in ether theories [memory, high].

The second law of thermodynamics says the entropy of an isolated system does not decrease. Stars convert mass to radiation (E = mc²; the Sun loses about 4 × 10⁹ kg/s) [memory, high]. Radiation can make matter only as particle-antiparticle pairs above 1.02 MeV per pair.

---

4. Where Very departs

Very's starting points are real. Quanta were real; the null results were real; the extended electron with all-electromagnetic mass was a respectable research program; Fessenden really had proposed that the ether's breaking strain sets the electron's size [web: https://archive.org/stream/jstor-1628439/1628439_djvu.txt]. Where he goes beyond them:

The sharpest single point of departure, and the only place the book does arithmetic, is the electron in row 4. That is the worked example.

---

5. Worked example: Very's rotating electron, by the numbers

All numbers are reproduced by example.py (run: .venv/bin/python review/Very-Frank/example.py). The book's text, from the page images [read: page image pp.37-38]:

"the electron is 'about 3/4 × 10⁻¹³ cms. in diameter. The ionic equivalent being about 4 (±1) × 10⁻¹⁰ e. s. units [now more accurately fixed by Millikan at e = 4.774 (±.005) × 10⁻¹⁰] we find the electrostatic tension and pressure at the surface of the corpuscle, about 2 × 10³² dynes.' The revised measurement of the electron's radius makes this more nearly 10³³ dynes."

"volume-elasticity of electron = (surface pressure)^(3/2) / (surface of electron) = (10³²)^(3/2) / 10⁻²⁵ = 10⁷³."

"2.999 × 10¹⁰ / (2π × 1.9 × 10⁻¹³) = 1.660 × 10²² revolutions per second at the equator, if we adopt Millikan's value of the radius of the electron. This is about 1,000 times as great as the frequency of the characteristic vibrations of the lighter atoms."

Step 1. Where the radius comes from. Set Thomson's electromagnetic mass equal to the electron mass and solve: a = 2e²/(3mc²) = 2 × (4.774 × 10⁻¹⁰)² / (3 × 9.0 × 10⁻²⁸ × (2.999 × 10¹⁰)²) = 1.88 × 10⁻¹³ cm. That matches Millikan's 1.9 × 10⁻¹³ cm. Very's input is consistent.

Step 2. The rotation rate. A point on the equator travels a circumference 2πa per turn. At speed c it makes

f = c/(2πa) = 2.999 × 10¹⁰ / (2π × 1.9 × 10⁻¹³) = 2.999 × 10¹⁰ / 1.194 × 10⁻¹² = 2.512 × 10²² rev/s,

not the printed 1.660 × 10²². The printed figure is low by a factor 1.513. It would follow from a radius c/(2π × 1.660 × 10²²) = 2.875 × 10⁻¹³ cm, close to e²/(mc²) = 2.78 × 10⁻¹³ cm; that may be what happened, but the book does not say so. The angular velocity is ω = c/a = 1.58 × 10²³ rad/s.

Very adds that "the velocity of electronic rotation is the same at all points of the surface ... so that the angular velocity increases from the equator to the poles". At latitude λ the circle of latitude has radius a cos λ, so ω(λ) = c/(a cos λ): double at 60°, 57 times larger at 89°, and infinite at the pole. By the hairy ball theorem (section 2) no smooth flow on a sphere can have the same nonzero speed everywhere. The stated flow has singular points at the poles.

Step 3. The surface stress. The field just outside is E = e/a². The outward stress is P = E²/8π. Units: E² in Gaussian units is dyn/cm², so P is a pressure, not a force in "dynes". Numbers:

Fessenden's printed 2 × 10³² comes out only if his diameter is used as a radius. The arbiter checked Fessenden's 1900 page image: the word "diameter" and the "dynes" are Fessenden's own, so Very inherited the slip [web: https://archive.org/download/jstor-1628439/1628439.pdf, PDF p.5, per arbiter.md]. Because P scales as 1/a⁴, Millikan's larger radius gives a smaller stress, 7 × 10³⁰, not "more nearly 10³³". The 10³³ matches Fessenden's small radius done correctly.

Step 4. The units of Very's elasticity formula. A bulk modulus has the units of pressure, g cm⁻¹ s⁻². Very's expression has units

(dyn/cm²)^(3/2) / cm² = g^(3/2) cm^(−7/2) s⁻³,

which differs from a pressure by a factor g^(1/2) cm^(−5/2) s⁻¹ (example.py step 4 checks this with SymPy). So the formula cannot be a statement about a bulk modulus whatever numbers are put in. The exponent arithmetic is right as printed: (10³²)^(3/2)/10⁻²⁵ = 10⁴⁸ × 10²⁵ = 10⁷³. But the 10³² belongs to the small Fessenden radius and the 10⁻²⁵ cm² area belongs to Millikan's radius (4πa² = 4.5 × 10⁻²⁵ cm²). Using one radius throughout gives 4 × 10⁷⁰ (Millikan) or 10⁷⁶ (Fessenden's radius, e = 4 × 10⁻¹⁰; 1.8 × 10⁷⁶ with e = 4.774 × 10⁻¹⁰), a spread of six orders of magnitude.

Fairness to the sources: the number 10⁷³ is Fessenden's, and inside Fessenden's model it is coherent. His density 0.66 g/cm³ and rigidity 6 × 10²⁰ dyn/cm² give a shear-wave speed √(6 × 10²⁰/0.66) = 3.0 × 10¹⁰ cm/s, which is c, and his compressional-wave speed 5 × 10³⁶ cm/s gives K = ρv² = 0.66 × (5 × 10³⁶)² = 1.65 × 10⁷³ dyn/cm² [computed; Fessenden's figures per arbiter.md page-image check]. So what fails is Very's route to the number, not necessarily Fessenden's model (which is untested and would need compression waves at about 10²⁶ c).

Step 5. Angular momentum: the test that came later. Treat the electron as a uniform solid sphere of mass m = 9.109 × 10⁻²⁸ g and radius a = 1.9 × 10⁻¹³ cm with equatorial speed c. Then

L = Iω = (2/5)ma² × (c/a) = (2/5)mac = 0.4 × 9.109 × 10⁻²⁸ × 1.9 × 10⁻¹³ × 2.999 × 10¹⁰ = 2.08 × 10⁻³⁰ erg s.

The measured spin is ħ/2 = 5.27 × 10⁻²⁸ erg s, 254 times larger. A thin shell, (2/3)mac = 3.46 × 10⁻³⁰, is still short by a factor 152. To carry ħ/2 a solid sphere this size would need a surface speed of about 254c. The rotational kinetic energy, (1/2)Iω² = 0.2mc², is also extra energy that Thomson's mass formula does not count. Spin was discovered in 1925, six years after the book, so this is a later test of the model, not a mistake Very could have caught.

Step 6. What the rotation rate would mean as a quantum (a reviewers' construction). Very does not equate the quantum with h × (rotation rate); he says the radiant period "appears to depend on some function of the radius of the electronic orbit" (pp.39-40). Still, for scale: h × 1.660 × 10²² = 1.1 × 10⁻⁴ erg = 69 MeV (104 MeV with the corrected rate), about 130-200 times the electron's rest energy (0.511 MeV), against 2 eV for a visible photon. And his comparison "about 1,000 times ... the lighter atoms" gives ratios of 3 × 10⁷ (visible light), 5 × 10⁶ (hydrogen's Lyman limit, 3.29 × 10¹⁵ Hz), 2 × 10⁵ (carbon's K-alpha X-ray line). Only the K lines of the heaviest atoms (uranium, about 2 × 10¹⁹ Hz) come near 1,000, and Very says "the lighter atoms".

What the example shows. The inputs are honest period values. The chain from them has four separate problems: an arithmetic slip (factor 1.51), an inherited diameter/radius slip, a "revision" that goes the wrong way, and a formula that fails the units check. The picture of a surface everywhere at speed c is geometrically singular. And the model, judged by a measurement made after 1919, carries 1/150 to 1/250 of the electron's spin. The arbiter's verdict follows these lines: the numbers are internally inconsistent; as the real electron the model is refuted by calculation (spin, size); the idea of an extended electron held together by a medium is historically superseded but instructive.

---

6. Evidence table

Verdicts are the arbiter's, including its "QE revision" section (which changed no headline verdict but revised the claim 4 and claim 6 sub-verdicts shown here).

Evidence kept in the record beside the verdicts

• Very, Science 30:491 (1909), "The Conservation of Mass and the Passing of Matter": first statement of claims 3, 4 and 7 ("vacuities in an exceedingly dense ether, kept from collapsing by the centrifugal pressure of a rapidly rotating shell"); a mass equation dm/m₀ = v/(V − v); cites Kaufmann's mass-velocity data and the Lebedew and Nichols-Hull light-pressure measurements, which are real [web: https://archive.org/stream/jstor-1635576/1635576_djvu.txt].
• Fessenden, Science 12:740 (1900): the numbers Very borrows; a stretched-glass prism compared with the Kerr effect (30 × 10⁶ against 25 × 10⁶ dynes), which Fessenden himself calls "probably accidental"; a proposed test that gases should weigh slightly more than solids [web: https://archive.org/stream/jstor-1628439/1628439_djvu.txt]. Fessenden's account, in his own words, that his 1890-91 theory was held back about a year and then printed with an editorial calling it "probably wrong" (Science 12:325, 1900) [web: https://archive.org/stream/jstor-1629562/1629562_djvu.txt]: one source, his own, about his own work, not Very's.
• Fessenden 1913 and 1914: "every continuous medium theory must involve quanta"; gyroscopic upsetting work proportional to frequency; a nucleus formed of negative electrons. A vortex route to quanta in print before quantum mechanics [web, archive.org links in sites.md].
• Dayton Miller's Mount Wilson ether-drift result (1920s) is the one respectable positive result on record; Shankland et al. (1955) attributed it to temperature effects and statistics, which is that group's reading of the data [web via contrarian: https://en.wikipedia.org/wiki/Dayton_Miller]. Nobody in this review rechecked Miller's data.
• Character evidence on Very's handling of data: he exposed the 1910 Norwood "meteorite" as a fraud after first reporting the fall (Investigator; jstor-1636137), and he states plainly he "could not verify" a Franklin quotation given by Millikan [read: p.42].
• The religious and speculative frame is in the text: the aura as an "infinite source of energy which is spiritual" (p.44); synchronous pulsation as "the finger of God" (p.49, Exodus viii, 19); and telepathy (p.55). In a passage he says is "left undecided", an earthly atmosphere of free ether-particles at rest relative to the earth "is perhaps the physical basis of exterior thought ... and especially of that telepathic thought which is received by harmoniously attuned instruments" [read: p.55]. It is offered as a hypothesis, with no experiment or mechanism.
• Very's own measurement: firefly spectro-bolometry. Langley and Very, "The Cheapest Form of Light", Am. J. Sci. XL (1890), cited at PDF p.21: the firefly's light has short-wavelength emission that "if it were thermally produced would correspond to a temperature of several thousand degrees", given off at body temperature "without the corresponding infra-red rays" [read: p.21]. Cold light from living things is mainstream [memory, high], and the observation is real. Very uses it by analogy (light from satellite electrons, not from heat), not as a test of his electron model. It is the same 1890 period of work as the Langley credit omission above.
• Eichenwald 1903 (p.23): Very's one experimental reason for an aura that matter does not drag (quoted in section 4). He gives no data or journal reference, and nobody in this review read Eichenwald's paper.
• Hirn on the ether's mass (pp.34-35): planetary motion shows no massive resisting medium, which Very uses to confine ether atmospheres to the neighbourhood of massive bodies (section 4). He also corrects d'Auria (1897) for reading Maxwell's ether density as relative to air instead of water, and notes a small numerical error in Maxwell's own value [read: pp.34-35]. Hirn's original work was not read here.
• Shapley and M13, and grey absorption (p.43, footnote): Very accepts Shapley's 1915 finding that light from the globular cluster M13 shows no colour-dependent (selective) absorption, says he reached it independently, and denies that this shows colourless (grey) absorption is negligible, citing his own study of white nebulae (Astron. Nachr. 4536) [read: p.43]. The logic is sound: a colour test cannot detect grey absorption. Trumpler (1930) later found interstellar extinction, mostly reddening by dust in the galactic plane, and M13 lies off the plane [memory, high]. So his logical point stands; his attribution to the aura does not.
• A numerical point in the model's favour (the mathematician's steelman). Take Fessenden's compression-wave speed v = 5 × 10³⁶ cm/s and a galactic cell radius R = 3.8 × 10²² cm (40,000 ly) as a quarter wavelength: the pulsation frequency is v/(4R) = 3.3 × 10¹³ Hz, in the infrared, the same order as the "periodicity comparable to that of light" Very asserts (pp.37-38) [computed, mathematician.md and arbiter.md]. The book does not do this sum. It shows only that Fessenden's constants and Very's galactic scale are order-of-magnitude compatible.
• Thornton's cube rule (p.32): H, B, Al, Cu, Ru and Os, "the densest atoms of their respective thermal periods", have atomic weights proportional to 1³ to 6³; Very uses it to support six electron "stories" in the heaviest atoms. Checked with modern atomic weights [memory, high; example.py step 8]: it holds for Al and Cu (within 1.5%), fails for B (about 35% high) and for Ru and Os (about 12-20% low). Exercise 6 below lets you check it yourself.
• Other material from the deep reads (context, no verdict). Lenard 1902 and Bragg's corpuscular X-rays, which Very reinterprets (pp.13-20); Swan's 1897 resin stress figures, offered as evidence of polar inflow and outflow at the electron (pp.26-27); Millikan's Kepler-law derivation of the Rydberg constant "to one part in one thousand", which Very uses to reject Marion Erwin's 1,732-electron hydrogen atom (p.28 fn), showing him discarding an atom-count model on measured data; Lodge's steel-rod comparison (p.44); Zöllner's Transcendental Physics as "a semblance of experimental evidence" for hyperspace (pp.49-50); Newton's "aetherial spirits" letter, which Very ties to his aura (pp.56-57); Swedenborg's flying-machine design and R. W. Brown's 1916 critique of the vortex theory (pp.54-58). Very's own estimate that Nova Persei 1901 had 1,149 solar masses (Astron. Nachr. 3771, cited p.45) is far above any modern stellar or nova mass [memory, high]. One new arithmetic slip: his star-spacing expression on p.46, (92.9 × 206,265)/10⁶, evaluates to 19.2, not the printed 192 km, and mixes miles with kilometres; his conclusion that star collisions are negligible survives either way [read: page image p.46, per arbiter.md].
• Nipher. Very wrote a 1920 note on F. E. Nipher's gravitation experiments [Nipher's initials: memory, high] (Investigator, sites.md). Nobody in this review has read it or Nipher's papers; whether Nipher's result was ever repeated is unknown and was not checked.
• The record goes dark. No contemporary review, reply or citation of the pamphlet has been found, and the fate of the Westwood observatory after Lowell's death (1916) and of Very's papers after 1927 is unknown. The Investigator's search was brief. The arbiter weighs these explanations: the pamphlet went to a Swedenborgian journal and a private series (but Very did read it to the AAS); it made no prediction to answer; relativity and the Bohr atom absorbed attention; the religious frame put physicists off; loss of patron. For suppression of this pamphlet the evidence found is nil; the Langley credit omission concerns other work. Records that would decide: the AAS 1918 abstract in Popular Astronomy, Ogden's 1928 memoriam (Popular Astronomy 36:391), Lowell Observatory archives, the Allegheny Observatory file "Very, Frank (Westwood Astrophysical Observatory)" at Pittsburgh, Boston obituaries of Nov 1927.

---

7. What a learner should take away

Solid. Very saw early (1918) that discrete light quanta remove the reason for a continuous light-carrying medium. That direction was right. His inputs (Millikan's charge and radius, Thomson's mass formula, the null experiments, the photoelectric relation) are correctly quoted. He asked a real question that classical physics never answered: what holds an extended charged electron together? He argued that Andromeda is of "truly galactic dimensions" (p.45) when that was still open (settled 1923-25), though his distances are far too small. An interstellar medium that absorbs starlight exists (dust and gas), though it is not his aura; and his footnote point that a colour test (Shapley's M13) cannot rule out colourless absorption is logically correct (p.43). His one first-hand measurement, the firefly's cold light (p.21), is real. And the mathematician's steelman shows Fessenden's constants and Very's galactic scale give a pulsation frequency of the right order (3 × 10¹³ Hz) for his claim, though the book never does the sum.

Wrong, and why it is instructive. - Units are a free error detector. Very's elasticity formula fails a two-line units check. Fessenden's "dynes" for a stress is another. Always check units before believing a number. - Keep one value of each input. Mixing a pressure from one radius with an area from another moved his answer by six orders of magnitude. A 1/a⁴ dependence is very sensitive to which a you pick. - Check the direction of a correction. "The revised radius makes this more nearly 10³³" goes the wrong way for a quantity that falls as 1/a⁴. - A picture needs a rule. Particles of fixed size cannot produce fringes that scale with λ unless a wave rule tells them where to land. Very gives none. Modern physics keeps the quanta and keeps the wave rule. - An explanation must derive the result. "In harmony with" the Michelson-Morley null is not the same as predicting zero fringe shift. Very's own statement that the bound ether does not carry light leaves the null unexplained under either reading of what light's speed is relative to. - Sound arithmetic on bad inputs. Very's galaxy-collision argument is internally reasonable: on his 1918 distances (galaxies a few hundred light-years across) collisions come within about 10⁹ years (example.py step 7). It fails because his distance scale, built on taking the 1885 supernova in Andromeda for an ordinary nova, is about 300 times too small [memory, high]. An earlier draft of this review (mathematician.md, and the arbiter's first table) misread his 80,000 ly, a distance to Curtis's nebulae, as a galaxy size and said the argument failed on his own numbers; the arbiter withdrew that. Check what a number in a source refers to before using it against the author. - Every flexible medium explains everything. An aura with no measurable property can absorb any anomaly, which is why it cannot be tested.

Genuinely open (in this record). Whether the printed 1.660 × 10²² and "10³³" are misprints or Very's own: compare the second scan (https://archive.org/details/luminiferousethe00veryrich) and the Oct 1918 New-Church Review printing. Whether the "collapse" Very feared means an inward stress (Fessenden's "condensation" of the aura) or is a sign error, since a charged surface is pushed outward. What Very's unread 1913 papers and 1920 Nipher note say. Why the pamphlet drew no recorded reply. Whether Very would accept the literal zero-drag reading of p.53, and in what frame he meant light to move at c: the pamphlet does not say. What the original reports of Eichenwald, Hirn, Thornton, Swan, Hicks and Leahy actually show, since all are known here only as Very reports them. And, in physics generally, the self-energy of a point charge remains a real issue that quantum field theory handles by renormalisation rather than by a mechanical model [memory, high].

---

8. Check yourself

1. Compute the electrostatic stress E²/8π at the surface of a sphere of radius 1.0 × 10⁻¹³ cm carrying charge 4.774 × 10⁻¹⁰ esu. By what factor does it change if the radius doubles?
2. Show that (dyn/cm²)^(3/2)/cm² is not a unit of pressure. What extra factor would you need to multiply it by to get dyn/cm²?
3. Very says each point of the electron's surface moves at c. What is the angular velocity at latitude 75° for a = 1.9 × 10⁻¹³ cm? Why can't this hold right at the pole?
4. Very writes (p.53) that the bound ether slows light by its "density and electro-magnetic state, and not by the motion of the bound ether". (a) What does that predict for the speed of light, measured in the lab, travelling along a tube of water (n = 1.33) that flows at v = 5 m/s in the same direction? (b) Fizeau's experiment found c/n + v(1 − 1/n²). How large is the difference from (a) in m/s? (c) What would complete drag predict, and why does Fizeau's result count against both readings?
5. A stream of identical, fixed-size particles passes through two slits 0.1 mm apart. Using the wave-rule formula, how far apart are the fringes on a screen 1 m away for λ = 500 nm and for λ = 650 nm? What does Very's model need to supply to reproduce this?
6. (Optional, numerical.) Very cites Thornton's rule (p.32) that H, B, Al, Cu, Ru and Os have atomic weights proportional to 1³, 2³, ..., 6³. Using modern atomic weights (1.008, 10.81, 26.98, 63.55, 101.07, 190.23), compute weight/n³ for each. Where does the rule hold, and by how much does it miss elsewhere?

---

9. Further reading

Mainstream: - D. J. Griffiths, Introduction to Electrodynamics, 4th ed.: ch. 2 (electrostatic energy), ch. 9 (electromagnetic waves), ch. 12 (relativity), Appendix C (units). - R. P. Feynman, R. B. Leighton, M. Sands, The Feynman Lectures on Physics, Vol. II, ch. 28 "Electromagnetic Mass" (the extended-electron problem, the 4/3 puzzle, Poincaré stresses). Free online at feynmanlectures.caltech.edu. - E. T. Whittaker, A History of the Theories of Aether and Electricity (1910; revised 1951): the ether theories Very was working within, including Stokes, Fresnel and the vortex ethers. - J. R. Taylor, Classical Mechanics, ch. 10 (rotating rigid bodies, moment of inertia). - E. F. Taylor and J. A. Wheeler, Spacetime Physics (special relativity at the right level). - K. S. Krane, Modern Physics (photoelectric effect, Compton effect, spin, nuclei). - Wikipedia articles used as checks: Michelson-Gale-Pearson experiment; De Sitter double star experiment; Sagnac effect (links above).

Very and his sources: - F. W. Very, The Luminiferous Ether (Westwood Astrophysical Observatory Occasional Papers No. 2, Four Seas Co., Boston, 1919); the project PDF. Second scan: https://archive.org/details/luminiferousethe00veryrich. - F. W. Very, "The Conservation of Mass and the Passing of Matter", Science 30:491-496 (1909): https://archive.org/stream/jstor-1635576/1635576_djvu.txt. - F. W. Very, New-Church Review XXV (Oct 1918) pp.528-576 (first printing): https://archive.org/details/new-church-review-1894_1918-10_25_4; "A Scientific Estimate of Swedenborg's 'Principia'", New-Church Review XX (1913), attribution from a bookseller listing. - R. A. Fessenden, "A Determination of the Nature and Velocity of Gravitation", Science 12:740-745 (1900): https://archive.org/stream/jstor-1628439/1628439_djvu.txt; "Inertia and Gravitation", Science 12:325 (1900); "Transformation of Gravitational Waves into Ether Vortices", Science 38:553 (1913); "Gyroscopic Quanta", Science 39:533 (1914): https://archive.org/stream/jstor-1640236/1640236_djvu.txt. - E. Swedenborg, Principia (1734), the new English edition Very annotated. - S. P. Langley and F. W. Very, "The Cheapest Form of Light", Am. J. Sci. XL, p.97 (Aug. 1890): the firefly spectro-bolometry Very cites at PDF p.21 [read: p.21; not read in the original].

---

Answers

1. E = e/a² = 4.774 × 10⁻¹⁰ / 10⁻²⁶ = 4.774 × 10¹⁶ statvolt/cm; P = E²/8π = 2.28 × 10³³/25.13 = 9.1 × 10³¹ dyn/cm². Doubling a divides E by 4 and P by 16, giving 5.7 × 10³⁰.
2. dyn/cm² = g cm⁻¹ s⁻². Raised to 3/2: g^(3/2) cm^(−3/2) s⁻³; divided by cm²: g^(3/2) cm^(−7/2) s⁻³. Dividing by g cm⁻¹ s⁻² leaves g^(1/2) cm^(−5/2) s⁻¹, so you would need to multiply by g^(−1/2) cm^(5/2) s (for example by dividing by the square root of something with units of pressure and multiplying by an area). Nothing in Very's argument supplies such a factor, so the formula is not a bulk modulus.
3. ω = c/(a cos 75°) = 2.999 × 10¹⁰ / (1.9 × 10⁻¹³ × 0.2588) = 6.1 × 10²³ rad/s, about 3.9 times the equatorial value. At the pole the circle of latitude has zero radius, so a finite speed c would need infinite angular velocity, and the direction "around the axis" is undefined there; by the hairy ball theorem a smooth surface flow must stop somewhere.
4. (a) Zero drag: the moving water changes nothing, so the lab speed is c/n = 2.998 × 10⁸/1.33 = 2.254 × 10⁸ m/s, whatever v is. (b) Fizeau's fraction is 1 − 1/1.33² = 0.435, so the measured speed is larger by 0.435 × 5 = 2.2 m/s [memory, high for Fizeau's formula]. Tiny, but Fizeau's interferometer detected it; so moving matter does affect light at first order, contrary to the literal p.53 statement. (c) Complete drag adds the full 5 m/s; Fizeau found less than half of that. So the result sits between the two: it rules out zero drag (Very's words) and complete drag (a Stokes-type ether carried by every atom). The arbiter rates the literal p.53 reading Tested, not supported; Very never mentions Fizeau, so whether he would accept that reading is not known.
5. Spacing = λL/d: 500 × 10⁻⁹ × 1 / 10⁻⁴ = 5.0 mm; 650 nm gives 6.5 mm. The spacing scales with wavelength, yet Very's ether-particles all have the same size; his model would need a rule, tied to the frequency of each particle, that tells it where to arrive. The book does not give one.
6. weight/n³ = 1.008 (H), 1.351 (B), 0.999 (Al), 0.993 (Cu), 0.809 (Ru), 0.881 (Os); example.py step 8 does the same. The rule holds within 1.5% for Al and Cu; B is 34% above the hydrogen value, Ru 20% and Os 13% below. A rule that fits two of five non-trivial cases is not strong support for Very's six electron "stories".