RankingsPrimerArbiterPhysicistMathematicianContrarianQESites logClaimsSourcesResearch notes

Physicist review: Lodge-O (Oliver Lodge, "The Ether of Space", North American Review, May 1908)

Reviewer: PhD physicist (Step 2). Sources read myself: deep/ read of PDF pp.10-14; Lodge, The Ether of Space (1909) Appendices 1-2 and ch. on the whirling machine (research/files/lodge-1909-ether-of-space-gutenberg40911.txt) [read]; Thomson, Electricity and Matter (1904) pp.25-27 [read, OCR]; Investigator's sites.md leads [read]. The Lodge 1907 Phil. Mag. paper I did not re-read; I rely on the Investigator's quotation and on its restatement in the 1909 Appendix 2, which I did read. Tags: [read], [web] URL, [memory] with confidence, [computation] = my own arithmetic (run in the project venv).

Plain-language terms used throughout: ether = the hypothetical all-pervading medium that 19th-century physicists thought carried light waves. c.g.s. = centimetre-gram-second units (1 erg = 1e-7 J). esu/emu = electrostatic / electromagnetic c.g.s. units for charge and magnetic pole.

Source note

One primary document (a JSTOR scan of a lecture-style article). No source gap for the article itself. Its headline number (density 1e12 g/cm^3) is derived elsewhere; the Investigator found that derivation, so the review below judges it directly rather than on the article's say-so.

The core physics in one place (worked example, [computation])

This is the single most useful thing for the reader, because it shows what Lodge's central number actually is.

Lodge (1909 App. 2, [read]) models the electron (charge e, radius a, speed u) as a sphere in a perfect fluid of density rho, and reads its magnetic field H = e u sin(theta)/r^2 as a circulation of that fluid with speed w, by setting the fluid's kinetic energy density equal to the magnetic energy density: (1/2) rho w^2 = H^2/(8 pi) (emu, mu = 1). Integrating the energy over space gives an effective mass m = (8/3) pi rho a^3 (w0/u)^2, where w0 is the circulation speed at the electron's equator.

Now substitute w0/u = e/(a^2 sqrt(4 pi rho)) from his own line above. The density cancels:

m = (8/3) pi rho a^3 * e^2/(a^4 * 4 pi rho) = (2/3) e^2 / a.

That is exactly the textbook electromagnetic mass of a charged shell (the field energy divided by c^2, in emu). So the "fluid" step adds nothing to the mass formula; rho only re-enters through Lodge's extra hypothesis w0 = u ("the whirl at the equator is as fast as the electron"). Setting w0 = u gives rho = e^2/(4 pi a^4). Numbers (e = 1.6e-20 emu, a = 0.5e-13 cm): rho = 3e12 g/cm^3. Check by the other route Lodge states, rho = m/V with m = 9.1e-28 g and V = (4/3) pi a^3 = 5.2e-40 cm^3: rho = 1.7e12 g/cm^3 [computation]. Both give "of the order 1e12".

What this means: rho c^2 (his "elasticity", 9e32 erg/cm^3) is just of the order of the electric field energy density at the electron's surface, E^2/(8 pi) with E = e/a^2: 1.5e33 erg/cm^3 for a = 0.5e-13 cm [computation]. Lodge's "enormous ether density" is the electron's own field-energy density at an assumed radius, re-labelled as a property of empty space. Because rho scales as a^-3 (mass/volume) or a^-4 (his w0 = u route), the answer depends entirely on the assumed radius: for the classical electron radius 2.82e-13 cm the m/V density is 1e10 g/cm^3; for a = 1e-16 cm it is 2e20 g/cm^3 [computation]. Modern limits put the electron's size far below any of these (below about 1e-22 m = 1e-20 cm in Penning-trap work; [web] https://www.nobelprize.org/uploads/2018/06/dehmelt-lecture.pdf, search summary, medium reliability; electron treated as point-like in the Standard Model [memory, high]). So the "order of magnitude is independent of the model" argument (p.14) is true for the mass model, but the radius it needs is not a measured quantity.

Per-claim review

Claim 1: No action at a distance; a continuous, absolutely stationary ether carries strain, cohesion and potential energy

a. Mathematics: no equations in the article for this claim. Nothing internally inconsistent; but "stationary" and "carries all energy" are asserted, not derived. b. Experiments: Locality (effects propagate through fields at finite speed) is one of the best-confirmed features of physics: finite signal speed, radio waves (Lodge himself was an early radio pioneer, [memory, high]), gravitational waves (LIGO 2015, [memory, high]). A material stationary medium is a different claim. Michelson-Morley (1887) found no ether wind; aberration (Bradley 1725) and Fizeau's partial drag (1851) fit a nearly stationary ether; von Laue (1907) showed special relativity reproduces Fizeau's drag coefficient with no medium ([web] https://arxiv.org/pdf/2407.12960 and search summary on von Laue 1907). Modern rotating-cavity Michelson-Morley tests bound any direction-dependent change of light speed to about 1e-18 (Nagel et al., Nature Communications 6, 8174, 2015; [web] https://www.nature.com/articles/ncomms9174). Fairness point: Lorentz's ether theory with an undetectable rest frame predicts the same results as relativity, so these tests do not logically disprove a stationary ether; they show it has no measurable properties and does no work. Lodge's own position (the ether is the most stationary body there is yet cannot be moved mechanically, 1909 App. 2 [read]) is of exactly that kind. c. Tested? Lodge's whirling-machine null results (claim 5), Michelson-Morley and successors; replicated by many groups. Entanglement-type quantum correlations are non-local in a sense but cannot carry signals [memory, high]; Lodge could not have known this, and it does not rescue a mechanical ether. d. If true: relativity would need a preferred frame with detectable consequences at some level; so far none seen to 1e-18. Verdict (split): Locality / field-mediated interaction: Established. Material, absolutely stationary ether as the carrier: Historically superseded but instructive, with its observable consequences Tested, not supported (no ether wind or preferred frame found to about 1e-18); the unobservable-rest-frame version is Not testable as stated.

Claim 2: Matter is made of ether; the electron is a singularity; all mass is ether carried by Faraday tubes (Thomson)

a. Mathematics: the electromagnetic-mass formula m = (2/3)e^2/a is correct for its model (see worked example), but it is the textbook result; the fluid language is a re-description. Known internal problem: the field energy and field momentum of a classical charged sphere give masses differing by 4/3 (Abraham-Lorentz "4/3 problem"), cured only by adding non-electromagnetic stresses (Poincare, 1906) [memory, high]. Lodge says "the order of magnitude" is independent of the model; that is true, but it shows the model fixes nothing. b. Experiments: Kaufmann and Bucherer measured the electron's mass growing with speed in 1901-1908, the very years of Lodge's argument; results were first read as favouring the electromagnetic-mass electron, and later (about 1908-1915) as favouring relativity [memory, medium]. Today: the electron is point-like to below 1e-18 m ([web] as above); a classical-radius electron cannot hold its own mass as field energy (the field energy outside radius a_0 = 2.8e-15 m already equals mc^2). Most of the mass of everyday matter is not electron mass: about 99% of a proton's mass comes from the energy of the strong-force (QCD) field and quark motion [memory, high]. An honest echo of Lodge/Thomson: "mass is field energy" (E = mc^2) is correct in spirit for most visible mass, but the field is the QCD field, not an electromagnetic whirl in a fluid. c. Tested? The electromagnetic-mass programme was tested and abandoned by about 1915 (Historically superseded). The "ether singularity" is not a model that makes predictions. d. If true: relativity's treatment of mass as an independent parameter and the Higgs/QCD origin of mass would be wrong. Verdict (split): All mass is electromagnetic/ether mass: Tested, not supported (and Historically superseded but instructive). Electron as a singularity in a uniform plenum: Not testable as stated.

Claim 3: Ether density about 1e12 g/cm^3; elasticity about 1e33; whirl speed about c; 1 mm^3 holds 1000 tons of mass-equivalent and a station's output for 40 million years

a. Mathematics, checked step by step [computation]: - Mass per mm^3: 1e12 g/cm^3 x 1e-3 cm^3 = 1e9 g = 1000 tonnes. Correct. - Elasticity: rho c^2 = 1e12 x (3e10)^2 = 9e32 erg/cm^3 (= dyn/cm^2). Correct, but circular: a wave speed c = sqrt(elasticity/density) was used to define the elasticity from the density, so "whirl speed of order c" restates c. Nothing independent has been predicted. (The centrifugal-cord analogy, p.11, is a valid hydrodynamic fact for a cord, but doesn't establish that the ether is such a cord.) - Energy per mm^3: rho c^2 x 1e-3 cm^3 = 9e29 erg = 9e22 J. Lodge's "million-horsepower station for forty million years" is 9.4e23 J; his 1907 version (million kW, 30 million years) is 9.5e23 J. So his figure is about 10 times his own rho c^2 (the Investigator found the same factor already in 1907). Unexplained; I do not know the reason. This is a minor internal inconsistency, not fatal. - Kragh's "10,000 tons per cm^3" paraphrase (Investigator lead) is about 100 times too low versus Lodge: 1e33 erg / c^2 = 1.1e12 g = 1.1e6 tonnes per cm^3. Lodge's own 1000 tons per mm^3 (= 1e6 tonnes per cm^3) is consistent. - The derivation (above) assumes w0 = u and a radius, so rho depends on the assumed radius as a^-3 to a^-4. b. Consistency with experiment: the decisive check is gravity. In general relativity (and by E = mc^2 and the equivalence principle) every form of energy gravitates [memory, high]. An energy density of 1e33 erg/cm^3 corresponds to a mass density of 1.1e12 g/cm^3. For a uniform density rho the Hubble-type expansion rate is H = sqrt(8 pi G rho / 3) = 750 per second [computation], against the observed 2.2e-18 per second [memory, high]; a sphere of this density one astronomical unit in radius would outweigh the Sun by about 1e19 times [computation, 1.4e52 g vs 2e33 g]. The observed vacuum/dark-energy density is about 6e-30 g/cm^3, i.e. 5e-9 erg/cm^3 [memory, high; the web check did not return the number, Planck 2018 gives Omega_Lambda about 0.69 of the critical density 8.6e-30 g/cm^3, https://arxiv.org/pdf/1807.06209]. Lodge's figure is therefore about 2e41 times larger. Lodge's reply would be that the ether's mass is not "ordinary" mass and does not gravitate (his gravity is an ether tension, 1909 App. 1 [read]); but then it also would not be the inertia he says it is, and general relativity (1915) has no room for energy that has inertia but no weight. - Lodge's number does have a family resemblance to zero-point energy. A vacuum energy density with a cutoff at the classical electron radius is about hbar c / r_e^4 = 5e33 erg/cm^3 [computation], the same order as Lodge's. This is dimensional analysis (any energy built from one length a scales as 1/a^4), not agreement. Quantum field theory's zero-point energy depends on the cutoff, is not measured in absolute terms, and its real effects (Casimir force, Lamb shift) are differences [memory, high]. Kragh's account of Lodge as a "remote ancestor" is historical ([web] https://arxiv.org/abs/1111.4623, via Investigator); the spacefed.com article is advocacy by an interested party for zero-point-energy technology and is not an independent test. c. Tested? Lodge's 1892-93 solenoid-in-the-optical-square experiment (Investigator lead: four 7,000-turn coils, no fringe shift at 1/50 band) searched for a magnetic ether drift; he drew from the null that either there is no drift or the ether is at least 8 times denser than platinum (1907 paper Part II, via Investigator). That is a null result, not a measurement of density. Nobody has measured an ether density. Larmor, FitzGerald and Heaviside allowed a dense ether on paper [web via Investigator]; Larmor's remark that density might need to be even greater (1907 App. B) shows the number was not constrained. d. If true: gravity would have to ignore this energy; the vacuum would have to be a 1e12 g/cm^3 medium in which planets move without drag (a dense inviscid fluid is not a contradiction for d'Alembert-type potential flow, but the rigid elasticity 1e33 would then have to be invisible to matter). Verdict (split): Arithmetic given the inputs: internally consistent except the unexplained factor of about 10 in energy; 1e33 "elasticity", "whirl speed about c": Internally inconsistent only in the sense of circularity (defined from c) so Not testable as stated; a free-space ether density/energy of 1e12 g/cm^3 as a gravitating or inertial reality: Refuted by calculation (the Hubble rate and vacuum-energy numbers above, applying accepted general relativity); the inference "inertia is electromagnetic given this density": Never properly tested in the article (he says "if I had time, I would show"); the 1907/1909 derivation exists and reduces to the electromagnetic-mass formula.

Claim 4: Ether spins near a charge and a magnetic pole with moment of momentum e m; Faraday rotation is gyrostatic; E, H, V mutually perpendicular

a. Mathematics: Thomson's result is real (read myself, 1904 pp.25-27: momentum at each point is perpendicular to the plane through the charge, the pole and the point, so the momentum circulates "like a top spinning about AB", with moment of momentum "e m" [read, OCR]). In modern terms the electromagnetic field of a charge q and a magnetic monopole g carries angular momentum of magnitude q g / c (Gaussian units), directed along the line joining them and independent of their separation [memory, high]; this is the basis of Dirac's 1931 quantization condition q g / c = n hbar / 2. Units caution: in Gaussian units e (esu) times m (emu) divided by c has the dimensions of angular momentum. Thomson's and Lodge's "e m" omit c; the deep read suggested this reflects mixed units. I could not resolve from the OCR whether Thomson defined e in emu there; treat the factor of c as unchecked [memory, high on modern formula; Thomson's convention unverified]. Faraday's rotation (1821) is explained completely by the Lorentz force F = q v x B on the current, so the "gyrostatic" picture adds no equation. The E, H, V statement is the right-hand-rule geometry of F = I L x B and motional EMF (v x B): each is a cross product. "Any two imply the third" holds only when the two are perpendicular in the right way; a force and a field do not fix a velocity along the field. The link to "the three dimensions of space" is speculation (Lodge puts it as a question). b. Experiments: the field angular momentum is a standard, confirmed concept in electrodynamics (it appears in Feynman's disk paradox; [memory, high]). But magnetic monopoles have never been observed [memory, high], so e m refers to an idealized pole. For a real magnet the two poles' contributions have opposite sign, leaving a geometry-dependent total [memory, medium]. Faraday rotation and motors: Established. c. Tested? Faraday's 1821 demonstration, replicated daily. No experiment distinguishes "the ether spins" from "the field carries angular momentum". d. If true (as a material spin): no change needed; it is a relabelling of field momentum. Verdict (split): Field angular momentum of charge plus pole, Thomson's em: Established (as a field result, with the unchecked unit factor). Faraday rotation as evidence of a material ether spin: Not testable as stated (no measurable difference from the Lorentz-force account). E, H, V as the foundation of all physics and of 3D space: Not testable as stated.

Investigator's related leads kept: Thomson (1904) is the source of "e m" and Lodge's Electrons (1906) cites Thomson Phil. Mag. April 1899 p.416 [web via Investigator]. Lodge 1920 (Phil. Mag. 39, 170-174): ether graininess 1e-30 to 1e-33 cm and light-bending index n = 1 + 2GM/(r c^2) "Einstein's deflexion". My check [memory, medium]: the first-order light-bending from an index n = 1 + 2GM/(r c^2) reproduces the Newtonian-plus-spatial-curvature value 1.75 arcsec only if both potential terms are included; index models of this type give half the observed deflection if the spatial term is missing, so "matches Einstein" needs a derivation I have not seen. Puthoff's polarizable-vacuum model (Found. Phys. 2002) is a later related proposal, not a test, and I have not checked it.

Claim 5: A moving fluid can imitate rigidity (kinetic elasticity), so a fluid ether can make solid matter; Lodge's spinning discs showed the ether cannot be spun mechanically

a. Mathematics: the physics of kinetic rigidity is ordinary and correct: a spinning gyroscope resists tilting, a fast jet or rotating column acts stiffly, a vortex ring is long-lived in an inviscid fluid (Helmholtz 1858), Kelvin's vortex-atom (1867) and gyrostatic ether (1890) were serious attempts [memory, high]. Nothing here is inconsistent. The gap is the leap from "fluids can show elasticity" to "this one gives the specific elasticities of light and of electrons". The elastic response of a rotating fluid supports waves with the speed set by the circulation, but no Kelvin-type model ever reproduced Maxwell's equations together with Lorentz invariance [memory, medium]. b. Experiments, Lodge's discs: I read his 1909 statement: in 1893 he found light between steel discs 1 inch apart was not altered by more than 1/200 of their speed, later tightened to 1/1000 after runs of up to three hours (1909 ch. on the Ether machine [read]; Phil. Trans. 184, 1893; Proc. Roy. Soc. 1897 abstract via Investigator, also 3,000 rpm and a one-ton iron mass tests). The 1908 article gives the disc size and 4,000 rpm but no data [read]. The disc edge speed is about 190 m/s [computation, 0.457 m radius]. Plausibility check [computation]: if air co-moved with the discs, Fresnel's drag coefficient for air, 1 - 1/n^2 = 5.9e-4, is just below his bound of 1e-3, so his sensitivity was close to the expected ordinary effect, and a null at 1e-3 is what both a stationary ether and relativity predict. This is an honest, careful null result by a first-rate experimenter. Modern Lorentz-symmetry tests are about 1e15 times more sensitive (1e-18 versus 1e-3, Nagel 2015 above). c. Tested? By Lodge (1893, 1897, with assistants Davies and Hall per Investigator); related ether-drag nulls by Michelson-Morley (1887), Michelson and Gale (1925, Sagnac-type, [memory, high]). I have not read Hunt (1986) on the whirling machine ([web] Investigator citation HSNS 16:111-134). d. If true: relativity's invariance of c would need exceptions. Verdict (split): Kinetic elasticity as physics of fluids: Established. A fluid ether constituting matter: Historically superseded but instructive. Lodge's disc experiment: Tested (null result, real and careful, consistent with modern physics; sensitivity far below modern limits); it was not an anomaly, so nothing was dismissed without testing.

Investigator's lead, kept: Lodge's 1893 magnetic ether-drift experiment went unpublished for 13 years ("not published... its details ought to be recorded", 1907). Competing explanations from the Investigator: null result judged not worth publishing; delay until the density estimate gave it meaning; loss of details. The November 1893 Liverpool notebook would decide. I agree this is worth noting but there is no physics anomaly in it: a null result.

Claim 6: Visible cosmos about 1e-23 of water density (Kelvin); ether far denser than matter in aggregate; ether may have psychical significance

a. Mathematics: checked [computation]: 1/1000 arcsec = 1000 pc = 3.09e16 km (agrees with his 3e16); sphere volume 1.13e59 m^3 (his 1.12e59); 1.5e36 tons / 1.12e59 m^3 = 1.3e-23 times water. His vacuum analogy (cosmos is as rarer than a 1e-8 atm vacuum as that vacuum is rarer than lead) checks to about 1e12 each. Kelvin's solar mass is about 25% low [memory, high]. Appendix 1 (Investigator lead; I read it): the "critical disruptive tension" 1e33 is compared with a quantity T = gR = GM/R, which has units cm^2/s^2 (a potential per unit mass), against 1e33 dyn/cm^2 (a pressure). Those are different dimensions unless T is multiplied by a density the text does not supply; a mathematician should confirm. Also Lodge's "a globe of earth's density would need a radius of eight light years" gives about 27 light years with his own T = (4/3) pi G rho R^2 and T = 1e33 [computation]; the other three numbers in that appendix (1e13 and 1e7 light years; free fall at c for 26,000 earth radii) reproduce. So the "8 ly" looks like a slip of about 10 in T, curiously the same size as the energy discrepancy in claim 3 (I do not know the cause). b. Experiments: today's local mass density (stars plus gas plus dark matter, roughly 0.1 solar mass per cubic parsec near the Sun [memory, medium]) gives about 4e8 solar masses in a 1 kpc sphere and about 7e-24 g/cm^3, the same order as Kelvin's 1e9 suns and 1e-23. The cosmic mean matter density is far lower, about 3e-30 g/cm^3 [memory, high]. So the figure is a decent local galactic density, not the universe's. "Ether denser than matter" cannot be tested because the ether density has no measurement (claim 3). c. Tested? Astronomical density: measured, with large modern refinement. Psychical significance: Lodge was president of the Society for Psychical Research (1901-03) [memory, high]; Investigator lead: Raia (JHBS 2007), Noakes (Hist. Sci. 2005) on ether and psychical research [web via Investigator, not read]. Psychical research produced no replicated positive results in controlled tests [memory, medium]; no test addressed an ether. It was not dismissed without testing in the sense that no test of this specific idea was ever designed. d. If true: a non-physical mind-ether coupling would be a new interaction. Verdict (split): Cosmos density about 1e-23 of water: Established (as a local value; arithmetic consistent). Ether denser than matter: Not testable as stated. Psychical significance: Not testable as stated (Lodge presents it tentatively).

Corrections to Step 1

  1. claims.md note that the density "rests on this article's statement alone" is superseded: the derivation exists (Phil. Mag. April 1907, restated in the 1909 book App. 2) and I have read the 1909 version. It reduces to the electromagnetic mass formula plus an assumed radius and w0 = u.
  2. claims.md claim 3 says the density "gives" an elasticity of 1e33 and a whirl speed of order c. That is circular: 1e33 is rho c^2 by construction. Rephrase accordingly.
  3. facts.json says the 1e-23 density is internally consistent: confirmed. Add that it is a local, not cosmic mean, value.
  4. Deep read says Lodge's "e m" "leaves out the factor c, consistent with mixing units". I could not confirm the units convention from the OCR of Thomson (I read the passage; the value is printed as "e m" with no c); the modern formula has q g / c. Treat as unresolved.
  5. The p.5 garbled "108S dynes per square centimetre" is, per the 1909 Appendix 1 which I read, 1e33; the PDF page image p.5 should still be rendered to confirm.
  6. The deep read's energy discrepancy (about 10 x rho c^2) is confirmed and is already in the 1907 paper; the explanation is still unknown. I add a second, similar ~10x slip in the 1909 Appendix 1 (8 vs 27 light years).
  7. Step 1 treated "stationary ether" as mainly a conflict with relativity; the sharper conflict is energy density against gravity (Claim 3b) and the point that observed results leave a Lorentz-type stationary ether logically open but empty.

Investigator's findings: what I kept and how I weigh them

  • Missing derivation (1907, 1909 App. 2): kept; reduces to EM mass; the density depends on assumed radius.
  • Unpublished 1892-93 magnetic-drift null experiment: a real experiment of genuine interest as history; supports no density claim by itself; null consistent with modern physics.
  • Tighter disc bound and the 1897 abstract: kept; careful experiment; near the size of expected air drag.
  • Thomson 1904: confirmed by me; the strongest part of the article.
  • Larmor, FitzGerald, Heaviside precedents: show Lodge was within the mainstream of 1890s-1900s ether physics; the physics community of the time accepted dense ethers as possible.
  • 1920 restatement and Puthoff-type link; Kragh's zero-point lineage; spacefed (interested party, advocacy): kept as related work, none is a test.
  • Absence of evidence (Liverpool notebooks, Birmingham letters, Science 26 and Sci. Am. Suppl. 64 reception, Proc. RI 1908): not yet searched; I draw no conclusion from their absence.

Verdict summary

  1. Action at a distance / stationary ether: locality Established; material stationary ether Historically superseded but instructive (observable consequences Tested, not supported).
  2. Matter made of ether, all mass EM: Tested, not supported / Historically superseded but instructive; singularity picture Not testable as stated.
  3. Density 1e12 etc.: arithmetic mostly consistent but circular; as a real gravitating/inertial free-space density Refuted by calculation; inertia derivation Never properly tested.
  4. Thomson's e m spin: field angular momentum Established; material-spin reading Not testable as stated.
  5. Kinetic elasticity / discs: physics Established; fluid-ether matter Historically superseded; disc null Tested, consistent.
  6. Cosmos density Established (local); ether denser than matter and psychical significance Not testable as stated.