Larken (Edgar L. Larkin, Radiant Energy and Its Analysis, 1903): Arbiter (Step 3)

Arbiter: weighs physicist.md, mathematician.md and contrarian.md, and the Investigator's findings (sites.md, research/). Tags: [read] the book or a folder document, [web] with URL, [memory] with confidence, [calc] my own arithmetic (Python, project venv). The author spells his name Larkin; the folder says "Larken".

Source gap note

There is no source gap for the book itself: one scanned book (text layer good; 23 OCR pages; nine passages deep-read, tables and figures checked on page images). Two limits apply to every verdict below. 1. Everything in the folder is Larkin's own and single-source. The only independent records are the Investigator's: trade-press snippets (Google Books, snippet view only) for the storm of 16 July 1892, one editorial note in the Astronomical Society of the Pacific's journal, a 1906 critique, and later spiritualist-press reception. Of Larkin's own observations, the November 1882 sunspot jets, the "stellar floor" and the 1901 meteors have no second witness on record. The April 1882 aurora and sunspot do: the book (p.149) quotes S. P. Langley's letter of 17 April 1882 (corrected by the QE revision below; the first draft of this note said none had one). They are reported here as one observer's account, not dismissed. 2. Pages 1-123 and 182-204 were digest-only (mainstream spectroscopy, solar photography, instruments); the reviewers did not re-read them. Pages 143-150 and 17-19, 78-79, 158-160, 190-191 were digest-only when the reviews were written and have since been deep-read (QE revision below). Where a verdict leans on [memory] alone it is marked.

"Radiant energy" in the title means electromagnetic radiation in the 1890s sense (Melloni, Langley). The book describes no free-energy device and no non-Hertzian wave [read: pp.17-19, 39-40, 177-178], but it does state an aim: "taking energy out of space and chain it to his servitude. Men pay for power now; it should be free" [read: p.78], meaning solar radiation (QE revision, section D). The verdicts are about 1903 physics and astronomy.

Corrected claim list (final verdicts)

Scale: Established; Plausible, untested; Never properly tested; Tested, not supported; Refuted by calculation; Internally inconsistent; Not testable as stated; Historically superseded but instructive.

---

Claim 1: corpuscle sea, one wave motion, radioactivity "directs" energy

Agreement. All three reviewers: (a) is Established; the corpuscle medium is superseded; the passage "corpuscles bend in a magnet, therefore the entire wave motion of the universe is electromagnetic" [read: p.19] is a non sequitur (bending in a field shows the particle is charged, not that a light wave is a motion of such particles). Maxwell's wave is oscillating electric field E and magnetic field B and needs no medium. All three also note his amplitude/wavelength slip [read: pp.19-20] (wavelength is c/f, independent of amplitude).

Disagreement 1: verdict on radioactivity. Physicist: Refuted by calculation. Mathematician and Contrarian: Tested, not supported. My weighing: the physicist's energy budget is right (1 g of radium-226: 3.7 x 10^10 decays/s, 4.78 MeV per alpha, 0.028 W from alphas alone; about 0.14 W with decay products [calc; memory: medium]; Curie and Laborde measured about 100 calories per gram per hour in 1903, which is 0.116 W/g [web: Pierre Curie's Nobel lecture via https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf, and the March 1903 note, https://www.nature.com/articles/068622a0]). But a budget alone cannot exclude "the energy comes from space and the atom is a conduit"; a flux from space is not forbidden by arithmetic. What rules it out is experiment: exponential decay with an isotope-specific half-life (polonium-210: 138.4 days, so 16% of the activity remains after a year [calc; web: https://en.wikipedia.org/wiki/Polonium-210]), half-lives unchanged between laboratories and deep underground [memory: high], helium produced from radium emanation (Ramsay and Soddy, 1903 [memory: high]). So the stronger argument is the Mathematician's and Contrarian's: it is an empirical test, and it failed. The Contrarian adds fairly that in 1903 the source of radium's energy was a live puzzle, so an external-source idea was not absurd then; Larkin's p.18 wording ("simply more transparent to corpuscles from space ... or change their directions of motion") is also loose. Final: Tested, not supported.

Disagreement 2: Mathematician's energy-density objection. For a mechanical wave in a medium of mass density rho whose bits swing at speed v, energy density is about rho v^2. Sunlight at the earth carries u = 4.5 x 10^-6 J/m^3 [calc: 1361 W/m^2 divided by c]. For v much less than c one needs rho much greater than u/c^2 = 5 x 10^-23 kg/m^3, about 3 x 10^7 corpuscles per m^3 at Thomson's mass [calc]. With Larkin's "inches or feet apart" the swing speed would exceed c; with his mean cosmic density (about 10^7 to 5 x 10^7 corpuscles per m^3, spacing 3 to 5 mm [calc]) it is about 0.8 c. The Mathematician calls this internal inconsistency. The physics is sound as a constraint on any elastic-medium picture, but the book never gives the present density of the medium (the numbers are for his primordial cloud), so I rate it a strong objection, not a proven contradiction inside the book. The Physicist's separate point is decisive without it: Michelson-Morley (1887) found no drift for a medium at rest through which the earth moves "as a wire screen through water" (orbital speed 29.8 km/s, v/c = 10^-4 [calc]), and a free-electron gas would be a plasma that reflects waves below its plasma frequency.

What would settle it: nothing new is needed for (a)-(c). For (d) nothing is measurable as stated. If anyone revived (c), the deciding test is a decay-rate or calorimetric comparison under changed environment; this has been done for a century.

Kept in the record. Larkin held the corpuscle principle consistently: letter to Hallock (c.1906) "nothing exists but electricity ... I allude to 'refined' matter in my book 'Radiant Energy'"; Within the Mind Maze (1911) "nothing exists but electrons" [web: http://iapsop.com/ssoc/1906__hallock___luminous_bodies.pdf; https://archive.org/details/withinmindmazeor00lark]. The "electrical theory of matter" was a serious programme around 1900 to 1905 (Lorentz, Abraham, Kaufmann) and part survives (electromagnetic field energy contributes to the electron's mass) [memory: high]. The spiritualist press (Progressive Thinker, 1904 and 1906) took the book up for an electrical soul; that is the reception of readers with an interest in that conclusion, and it does not change the physics.

Claim 2: solar jets, light-speed waves, storms, earth currents, aurora

Agreement. The sun drives storms, aurora and telegraph earth currents (Established; Larkin saw this early and correctly). The light-speed mechanism fails the timing test: the 1859 storm began about 17.6 hours after the flare [web: https://geomag2.bgs.ac.uk/education/carrington.html], carried by a plasma cloud (coronal mass ejection, CME) that crosses 1 AU in 17 to 100 hours (104 h at 400 km/s, 17.5 h at 2,380 km/s [calc]). Larkin's own November 1882 jets were drawn on 19 November, two days after the storm he ties them to (17 November), and he writes that spots occur without storms and storms without spots [read: pp.126-128]. A real light-speed effect exists but is small: the "magnetic crochet" of about 100 nT at Kew in 1859, from flare X-rays raising ionospheric conductivity [web: BGS page above]. "Tuning" has no frequency, quality factor or quantity [read: p.137]. The Birkeland terrella (1896) and Arrhenius radiation-pressure aurora (c.1900) were contemporary rivals that were closer to the later answer [web: https://en.wikipedia.org/wiki/Terrella; https://archive.org/details/jstor-40671408].

Disagreements. (i) Date of the table. All three accept the Investigator: 16 July 1892 (a Saturday), not 1882 (a Sunday) [calc, verified by me]. Larkin's own Science letter, Electrical Engineer 14 (1892) p.79 and Electricity (1893) p.157 agree [web: https://archive.org/details/jstor-1766041; Google Books snippets in research/2026-10-08-google-books.md]. The book's p.134 heading is wrong. (ii) Was the table "copied" from a trade source? Contrarian says the evidence shows only agreement of figures, not direction of copying; the Physicist and the Investigator say "likely". I side with the Contrarian: the snippets match only the peak values (133 mA, 492 V), the book names no observer ("Made at Terminals, New York and Boston" [read: p.134]), and a snippet's "12:20 p.m." does not match the book's windows (12:05-12:29 Providence, 12:37-12:46 Boston). So: unsourced in the book, matched by 1892-93 trade-press peaks, direction unknown. (iii) Are the volts independent measurements? Nine of ten Boston rows give V/I = 3,700 ohms to 1%, against a stated line resistance of 2,700 ohms [read: page image p.134, checked by Mathematician, Contrarian, and consistent with 492 V / 0.133 A = 3,700 in the 1893 snippet]. That fits "volts computed as current times a resistance" but does not prove it; the Mathematician's caution that this is an inference stands. The printed 144 mA is a misprint of 133 mA (492/144 = 3,417 ohms, out of line with every other row; the text prints 133 [read: p.135]). (iv) Same-hour timing in 1892. Larkin's letter gives a telegraph disturbance at 10:30 a.m. (Galesburg, central time) and a sunspot "tongue" brightest 9:30-10:30 [read: research/files/larkin-jstor-1766041.pdf]. The Contrarian values it as the right kind of evidence for the prompt, light-speed part; the Physicist says a sudden commencement or an ordinary coincidence would fit equally and one same-hour match needs statistics over many events. Both are fair. The book's New York table (12:05-12:46 p.m., probably eastern time, about 11:05-11:46 central [calc, unverified time standards]) comes one to two hours after the 10:30 reading, so even Larkin's own two records do not line up as one prompt effect. I rate the coincidence as an honest first-hand observation of weak causal weight, and he himself left causation open in 1892 ("a question that future physicists must decide") while the 1903 book asserts it. (v) Magnitude plausibility. Physicist and Mathematician both find 1.6 to 2.2 V/km over roughly 290 to 306 km plausible for a great storm [calc: plane-wave estimate]. I accept the magnitudes as believable. (vi) Energy budget of the mechanism. The Mathematician shows that a 1 microtesla low-frequency plane wave at the earth needs a Poynting flux of 119 W/m^2, which over a sphere of radius 1 AU is 3.4 x 10^25 W, about 9% of the sun's luminosity [calc, rechecked: c B^2/(2 mu0) = 119 W/m^2; area 2.8 x 10^23 m^2]. This is an illustration of the mechanism's cost, not a number Larkin gave, and I keep it as such. (vii) The tube claim. Contrarian: "no new mode is seen in the auroral light" [read: p.139] was false in 1903 because the green oxygen line (557.7 nm) was not reproduced in a laboratory until 1925 (McLennan and Shrum) [web, per Contrarian: https://www.nature.com/articles/115047a0]; Physicist had this at [memory: medium]. I take the Contrarian's web-tagged version. The tubes show the light and not the driver, and Birkeland read the same result as particles.

Important absence flagged: the storm of 16 July 1892 does not appear in the modern storm lists the Physicist and I searched. My two further searches found nothing for July 1892 and did find the catalogued 13-14 February 1892 storm and aurora with telegraph disturbances [web search 2026-10-08, results citing a chronology at https://spaceweather.gc.ca/tech/se-chr-en.php and https://arxiv.org/pdf/1305.6661]. That is silence in the web sources I could reach, not contradiction. See the absence analysis below.

Final: as the table above. What would settle it: the Kew, Greenwich and Potsdam magnetograms for 16 July 1892 compared with the Greenwich sunspot and solar records of that morning (onset time, lag), and the original Electrical Engineer 14 p.79 article (observer, circuit, time standard). For the 1882 events: magnetograms of 15-19 November 1882 against Larkin's drawings.

Claim 3: gravitational contraction, meteors at 1%, catalysis

Agreement. The arithmetic is sound. For a uniform sphere the gravitational energy is U = (3/5) G M^2 / R (G the gravitational constant, M the solar mass, R the solar radius): 2.28 x 10^41 J [calc]. As water-equivalent heat (specific heat 4186 J/kg/K, mass M of water) this is 4.9 x 10^7 F, matching See's 49,048,200 F to 0.4% [calc, rechecked by three reviewers]. Dividing by the solar luminosity L = 3.83 x 10^26 W gives 18.8 million years; Larkin's 21.8 million is inside the range 19 to 24 million that reasonable structure assumptions give (Mathematician). Meteors: 1/100 of the earth's mass falling at 383 mi/s supplies about one year of radiation (0.94 year at modern luminosity) [calc]. All three call contraction superseded as the sun's actual energy source: 4.57 billion years needs L x t = 5.5 x 10^43 J, 240 times the contraction budget [calc; age: memory: high], and hydrogen fusion supplies it.

Slips (all agree, none touches the argument): Stefan's law done in ordinary degrees (9.7 not 16), Lane's "sixteen times" ignoring shrinking area, 62.5 x 1.41 x 28 printed 1,475 lb (should be 2,467), the "reduce its temperature 2 1/4 degrees per year" treating a heat budget as a cooling rate, and no use of the virial theorem (only about half of the released energy is radiated, so the true uniform-sphere clock is about 9 million years [web: https://www.astro.sunysb.edu/fwalter/AST101/k-h.html; https://en.wikipedia.org/wiki/Kelvin%E2%80%93Helmholtz_mechanism]).

Disagreements. (i) Catalysis: Physicist Refuted by calculation, Mathematician Internally inconsistent / Not testable, Contrarian Not testable. A catalyst changes a reaction's rate and adds no energy; that is a definition, not a measurement. Larkin offers catalysis "for what it is worth, if anything" [read: p.175], so a refutation would overstate what he claimed. Final: Not testable as stated. (ii) The meteoric bound: Mathematician "Established", Contrarian "Plausible, untested". The orbit-perturbation arithmetic gives 4 to 17 arcseconds per century for 1% infall against positions known to a few arcseconds [calc, Mathematician], an order-of-magnitude bound that Kelvin also used [memory: medium-high]; I take Established as a bound. (iii) 1903 framing: contraction was the mainstream view in 1903, not a contrarian one; the contrarians were the geologists. The Curie-Laborde radium-heat result (March 1903) was the new complication Larkin does not mention [web: https://www.nature.com/articles/068622a0]. Historians now dispute how much radioactive heating by itself undercut Kelvin, and radiometric dating (1905-1907 on) is what decided the age question [web: search summary 2026-10-08, e.g. https://3quarksdaily.com/?p=16026]; this does not change the verdict on the sun.

Settle: already settled by the age of meteorites and the solar neutrino flux [memory: high]. For the reader: compute U/L.

Claim 4: cosmogony

Agreement. Earth-Moon tidal evolution is Established: tidal recession of about 3.8 cm per year [web: https://ilrs.dgfi.tum.de/lw20/docs/2016/presentations/05-Gross_presentation.pdf]. Larkin's "3 to 5 hours" and "55 days" are Darwin's end states. Conserving angular momentum with modern constants gives two synchronous states: a = 2.3 earth radii, period 4.85 hours (inner, unstable) and a = 87 earth radii, period 47.2 days (outer, stable) [calc, Physicist and Mathematician agree]; the 55 days belongs to Darwin's constants. Everything beyond the earth-moon case, and the lunar fission, is superseded. The fall-speed arithmetic is right on his premises: v^2 = 3GM/R gives 37.5 km/s = 23.3 mi/s (his 25), 32 billion suns, 320, and 333,000 x 3.2 x 10^10 = 1.07 x 10^16 ("eleven quadrillion") [calc]. The step "a fast star must be bound" is invalid, and his reason (no parallax in Newcomb) is bad, though the speed (about 200 mi/s = 322 km/s) is right. With 322 km/s the same formula gives 3.7 x 10^10 solar masses [calc]; a 220 km/s rotation speed at 15,000 light years gives 5 x 10^10 [calc]; the Milky Way's mass is of order 10^11 to 10^12 [memory: medium]. So "far more mass than visible stars" is vindicated in order of magnitude, and "dead suns" is not: microlensing finds compact objects of 10^-7 to about 0.1 solar mass make up under 20% of the dark halo, and solar-mass objects are limited to about 12 to 40% [web: https://arxiv.org/pdf/astro-ph/9612102, via Physicist's search; MACHO claimed up to 20%]. Note the honest reading of that result: it limits dead stars as the dominant component; it does not exclude some contribution.

Disagreement: nebular spin by electrical induction. Physicist and Contrarian: Refuted by calculation. Mathematician: Never properly tested, because Larkin gives no charge amount and an induced dipole in a uniform field feels no torque. All three computed the same scale: for two solar-mass bodies electric repulsion equals gravity at a charge Q = M sqrt(4 pi epsilon_0 G) = 1.71 x 10^20 C [calc, I reproduced it], about one electron per 10^18 hydrogen atoms (9 x 10^-19 of an elementary charge per proton); in an ionised gas free charges cancel net charge within a Debye length (a few centimetres to a few metres depending on density and temperature [calc: Mathematician 4 cm, Contrarian 0.2-7 m]).

I add one calculation that none of the three made, and it favours "refuted". Larkin's primordial cloud is made of Thomson's negative corpuscles [read: pp.17-19, 279-284]. For two electrons the electric repulsion exceeds gravity by e^2 / (4 pi epsilon_0 G m_e^2) = 4 x 10^42 [calc]; a cloud of corpuscles with no positive background could not condense under gravity into nebulae at all, and the same ratio holds for his corpuscle of one-thousandth of a hydrogen mass to within a factor of order unity because it is charged like an electron. The book supplies no positive background. So the induction-spin step is refuted either way: if the cloud is charged, it repels itself far more strongly than gravity pulls; if it is neutral or ionised, there is no net charge to induce a spin. The Mathematician is right that Larkin gave no number and so the idea was never tested as written; but the numbers needed come from his own premises, and they fail. Final: Refuted by calculation for induction as the origin of nebular spin. Fair credit, as the Physicist and Contrarian both note: "areas" conservation (halving the radius quadruples the spin) is correct, and a neighbouring body twisting a lopsided cloud is the idea behind today's gravitational tidal-torque theory of galactic spin (Hoyle, Peebles, White [memory: medium-high]); Larkin writes "gravity and inductance" [read: p.290].

Also kept: Larkin's 1881 Science series said no known force could start a primeval nebula rotating and called the ring stage "a physical error" [web: https://archive.org/details/jstor-2900594]; the 1903 induction chapter is his answer to his own twenty-two-year-old problem. The croquet-ball experiments (induction charging) are correct textbook electrostatics; the step to nebulae is analogy.

Density check (Mathematician vs Contrarian, small difference): 10^8 to 5 x 10^8 solar masses of hydrogen in a 30,000-light-year sphere is 1.0 x 10^4 to 5 x 10^4 atoms per m^3, a spacing of 4.6 to 2.7 cm [calc, recomputed]; for 1000 times more numerous corpuscles it is 4.6 to 2.7 mm [calc]. Larkin's "miles apart" and "inches or even feet" [read: p.280] are off by a factor of order 10^5 and 10 to 100 respectively. This is an internal arithmetic slip, not central.

Settle: a charge-per-mass requirement stated by the idea's supporter; measurement of net charge of nebulae (none found). "Cycle back to corpuscles" has no energy or entropy accounting: Not testable as stated.

Claim 5: the stellar floor, rents and starless fields

Agreement. The only evidence is Larkin's visual report of a year of observation [read: pp.299-303]; no photometry, nights, second observer or photographs. The same-night contrast test is a valid control against a uniform atmospheric dust haze, and does not separate the real candidates (airglow, zodiacal light, diffuse galactic light, unresolved faint stars, dark nebulae) [Physicist, Mathematician, Contrarian agree]. He states his own weaknesses: "the dark places seem so to the eye through contrast probably" [read: p.302]; the floor is "either countless unresolved stars or a solid nebulous envelope" [read: p.299]. Telescope magnification cannot raise the surface brightness of a diffuse glow, so a 16-inch refractor does not make a faint floor more visible than the unaided eye (Mathematician, standard optics). The ASP printed his 1901 summary ("Orion is always bound in white") and noted the longer report was "too long to reproduce" [web: https://archive.org/details/jstor-40668215].

Disagreement 1: epoch of the table positions. Physicist treated the printed coordinates as J2000; Mathematician as 1900 epoch precessed to J2000. I recomputed both ways [calc]. Starless field 6 (18h10m, -18 25') lies 1.3 degrees from Barnard 92 (18h15.5m, -18 13' [web: https://en.wikipedia.org/wiki/Barnard_92]) if read as J2000, but 0.19 degrees if read as 1900 epoch and precessed. Field 2 (17h25m, -26 35') lies 1.8 degrees from the Pipe Nebula Bowl (17h33m, -26 30') as printed and 0.43 degrees after precession. Field 7 (18h13m, -18 01') is 0.6 to 0.8 degrees from B92 either way. Astronomers of 1903 used the 1900 epoch, and the fit is much closer on that reading, so the Mathematician's reading is the better one (the epoch is still not stated in the book; I do not claim certainty). Either way the starless fields fall on catalogued dark nebulae in Sagittarius and Ophiuchus; 7 of 8 are within 14 degrees of the galactic plane [calc, Mathematician]. A chance empty field is not plausible for a populated Milky Way field (probability exp(-N), 2 x 10^-9 for N = 20 stars [calc]); it signals obscuring matter, which is the Barnard reading.

Disagreement 2: "Refuted by calculation" for the floor (Contrarian only). The extragalactic optical background is about 15 nW m^-2 sr^-1 at least from galaxy counts and plausibly up to about 60 [web: https://www.arxiv.org/pdf/astro-ph/9907315; https://arxiv.org/pdf/0906.0358], against a dark-site sky of 22 mag per square arcsecond. I converted this: 22 mag/arcsec^2 in V is about 1,340 nW m^-2 sr^-1 [calc; web search 2026-10-08 got about 1,300, accurate to 10 to 20%]. So extragalactic light is about 1% to 4% of the sky glow and could not appear as a distinct pavement to the eye. That settles one reading: a floor made of the light of distant galaxies or stars outside the Galaxy cannot be what Larkin saw. It does not settle whether he saw a real visual impression of the Galaxy's own diffuse glow, airglow patches or contrast effects. So I split: the extragalactic-light floor is Refuted by calculation; the floor as an eye observation is Never properly tested.

Disagreement 3: "Orion bound in white" and Barnard's Loop. Physicist downgrades it (the Loop is too faint for the unaided eye; the bright Orion nebula complex is likelier) [memory: medium]; the Contrarian leaves it as a lead. I agree with the downgrade: weak lead only.

Final as in the table. What would settle it: a calibrated photometric map (modern all-sky surveys, or a photoelectric survey) of the fifteen table positions compared with Barnard's catalogue and airglow models; Larkin's own full 1901 report; any Lowe Observatory plates.

---

Weighing the Investigator's findings (kept whatever the verdict)

• Larkin's 1892 Science letter (own work): earliest form of the storm claim with causation left open; telegrams sent "by nature's electricity, without batteries" (compare the 1859 operators who ran without batteries [web: https://en.wikipedia.org/wiki/Solar_storm_of_1859]).
• Larkin's 1881 series (primeval rotation), Hallock letter (c.1906), Within the Mind Maze (1911), The Matchless Altar of the Soul (1916; Larkin says he saw Tesla make a thin wire glow in a dark room; a single published account, date and place not given) [web: https://archive.org/details/matchlessaltarof00larkrich]. Recorded as Larkin's testimony.
• Supporting by others: the trade-press peak values for the 1892 storm; William H. Knight "critically examined the work before publication" (LA Herald, 5 April 1903 [web: https://cdnc.ucr.edu/?a=d&d=LAH19030405.2.366.13]); the ASP editor's printing of the stellar-floor note without comment. The LA Herald review was by a self-declared non-scientist.
• Critic's view: "The Earthquake and Professor Larkin", Science 24 (1906) 178-180, signed D. S. J. (probably David Starr Jordan [memory: medium]), ridicules his 1906 earthquake article. This is an institutional critic's characterization of Larkin as a popular writer; it tests none of the book's claims, and earthquake-light reports are studied today [memory: medium].
• Interested parties. The book's closing advertisement is the Pacific Electric Railway's, whose Mount Lowe line ran the observatory's tourist trade and whose president H. E. Huntington is named [read: p.335]; I report it as the railway's commercial notice and as the setting of Larkin's employment, not as evidence on the science. The spiritualist press and Hallock/Stetson have a stake in an electrical-soul reading; their use of the book is reception, not support. USC calls him "among the most-read scientists of his time" (about 9,000 bylines) [web: https://libraries.usc.edu/article/wizard-echo-mountain]; that is a statement about readership.
• Dates: the Lowe directorship began 1 January 1901 (Larkin's own words, PASP 1902); the 1890 given by Wikipedia and the Altadena society conflicts with Lewis Swift's tenure to December 1900 [read: p.319-322]. The book implies he succeeded Swift after 31 December 1900 (p.319); it states no appointment year (QE correction; the earlier wording "the book itself says he was appointed in 1900" was not in the text). AAAIM: accepted 11 Aug 1900.

Absence-of-evidence analysis

Where the record goes dark, as an anomaly to analyse.

A. The storm of 16 July 1892 appears in contemporaneous sources (Larkin's Science letter; Electrical Engineer 14, "unexampled severity ... almost completely paralyzing the entire telegraphic service"; Electricity 1893) but not in the modern storm lists I and the Physicist could reach (the catalogued 1892 events are 13-14 February). Competing explanations: (1) modern catalogues are built from magnetometer thresholds and a few famous events, and an event notable chiefly for telegraph effects in the northeast United States could be missed; (2) the storm was real but moderate in magnetic amplitude, with large earth currents on long, poorly earthed lines (the trade press quotes only telegraph figures); (3) the trade-press and Larkin figures share one source and an error entered early (the weekday "Saturday, July 16" matches 1892 and supports the date, but not the amplitude); (4) my web access was limited to snippet views and open-web search, so the real record may exist in the Greenwich, Kew and Potsdam yearbooks and in newspapers of 17 July 1892. For: the weekday, three independent-seeming publications, and the plausible V/km size. Against: absence from the lists I saw. What decides it: the magnetograms and observatory yearbooks for that day, and Chronicling America/CDNC newspapers of 16-18 July 1892 (not yet searched; listed in sites.md).

B. Larkin's own observing logs (New Windsor 1880-1888, Knox College 1888-1900, the Lowe Observatory 1901-1904), his November 1882 drawings, his longer 1901 stellar-floor report, the San Francisco Examiner series of February to August 1902, and any Lowe plates, have not been found. Explanations: (1) private papers lost or discarded after his death in 1924 (ordinary for a popular writer; no sign in the record of a research programme feeding a successor); (2) institutional dispersal (Knox College, the Lowe Observatory after the 1905 fire reported by the Altadena society, and the later transfer of its instruments [memory: low, not relied on]); (3) the stellar-floor work was never meant for specialists, so no one tested or catalogued it; (4) suppression has no evidence in the record: no data reanalysed or omitted, no funding blocked, no rebuttal by name was found by the Contrarian. The nearest thing to a response is the 1906 ridicule of an unrelated article. What decides it: archive finding aids at Knox College, the ASP, the Huntington Library and the Southern California Academy of Sciences, plus newspaper archives.

C. No scientific rebuttal of the book's astrophysics, and no second observer for the stellar floor. Competing explanations: popular rather than specialist audience; Larkin's later reputation as a sensational popularizer; the floor claim was visual and unmeasurable by 1900s methods; and the dark-nebula work of Barnard and Wolf (1904 onward) absorbed the observation into a different explanation without naming it. This last point is a reading of the record, not a document I have seen.

What would settle it

1. Claim 1(c): already settled by a century of decay-rate and calorimetry work; a reader can compute the radium heat budget (0.028 W/g from alphas).
2. Claim 2: Kew, Greenwich and Potsdam magnetograms and Greenwich solar records for 16 July 1892 (onset time and lag against the 9:30-10:30 a.m. sunspot tongue); the original Electrical Engineer 14 p.79 table; for 1882, the same for 15-19 November.
3. Claim 3: nothing more; compute U/L and compare with 4.57 billion years.
4. Claim 4: the charge per unit mass Larkin's induction needs, set against the e^2/(4 pi epsilon_0 G m^2) ratio for his corpuscles (done above; it fails).
5. Claim 5: photometric map of Larkin's fifteen table positions; Barnard's atlas; the full 1901 ASP report.

Open questions

• Was the Boston/Providence table independently observed, or computed from currents? (observer, circuit, time standard unknown; 3,700 versus 2,700 ohm).
• Did the 16 July 1892 storm leave a magnetic record?
• What did Larkin actually see in the fields at his table positions: dark nebulae against Milky Way glow, airglow patches, or contrast illusions?
• Where are his logs and drawings?
• How much did the new radium-heat result of 1903 shape his choice not to mention radioactivity as a solar energy source (the book is silent)?
• Whether the Tesla observation in 1916 can be dated (single account).

Corrections to earlier documents

1. claims.md claim 2 and facts.json (main entry summary and key_claims "On 16 July 1882 ..."): the earth-current storm is 16 July 1892; the book's p.134 heading prints 1882 in error. Evidence: 16 July 1882 was a Sunday and 16 July 1892 a Saturday [calc, rerun]; Larkin's Science letter ("Saturday, July 16, 1892"); trade-press snippets. The Nov 1882 jets and the 17 Nov 1882 storm stay 1882. The deep read pp.124-142 sections 4, 7 and 8 need the same fix. Also: the table prints 144 mA but the text, and the 1893 snippet, give 133 mA.
2. claims.md claim 2: "(16 July 1882: up to 280 mA at 644 V New York-Providence)" should read "(16 July 1892, printed 1882 in the book ...)".
3. claims.md claim 5: "a year of testing is said to have ruled out atmospheric dust" overstates it: the control is a same-night field-to-field contrast against uniform dust haze only [read: p.300].
4. research/2026-10-08-google-books.md and sites.md say the table "was copied from an electrical-trade source": change to "matches figures published in the 1892-93 trade press; direction of copying unknown".
5. Physicist.md treats the table coordinates as J2000 and claims fields 6-7 are within 0.5 degrees of Barnard 92; recomputed here: 1.3 degrees as J2000, 0.19 degrees if precessed from the 1900 epoch (the more likely reading).
6. Deep read, "pressures look computed from currents": an inference, not established (Mathematician); keep as "consistent with".
7. facts.json mainstream_conflict item 8 and the Investigator's lead on Barnard's Loop: downgrade to weak lead (the Loop is too faint for the naked eye).
8. Physicist.md verdict for claim 1(c) "Refuted by calculation" and claim 3 catalysis "Refuted by calculation": this arbiter uses "Tested, not supported" and "Not testable as stated" respectively, with reasons above.
9. Dating of the Lowe directorship: 1 January 1901 (not 1890).

Web verification log (budget 8; used 4 new searches, plus reviewers' web checks relied on)

1. Curie-Laborde radium heat 1903, about 100 cal/g/h: [web: https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf; https://www.nature.com/articles/068622a0]. Retags the Physicist's [memory: medium]. Confirms; historians dispute how much it undercut Kelvin.
2. July 1892 storm in modern sources: two searches, nothing found; February 1892 storm found [web: https://spaceweather.gc.ca/tech/se-chr-en.php]. The check "failed" in the sense of finding no confirmation. Not a refutation.
3. Sky brightness 22 mag/arcsec^2 versus extragalactic background light: [web: https://arxiv.org/pdf/1709.09558; https://arxiv.org/pdf/0906.0358; https://www.arxiv.org/pdf/astro-ph/9907315]. Retags the Contrarian's [memory: medium] value and confirms my own conversion (about 1,300 nW m^-2 sr^-1).
4. Barnard 92 position and epoch recomputation: the Physicist's web source retained; the epoch arithmetic is my [calc]. Not checked on the web (memory, medium, no verdict depends on them alone): Darwin and Joly on radium and the sun (low); Roche limit for a fluid moon; modern escape speed near the sun (about 500 km/s); Weiss domains 1907; Barnard's 349 dark markings; the Kapteyn-type 30,000-light-year size.

---

QE revision (2026-10-08)

Inputs: qe.md (verdict MINOR) and the two new deep reads in deep/radiantenergyits00larkrich-bfeabb.md: Ch. XX "Auroral Displays", pp.143-150 (page image p.150), and "harnessing radiant energy", pp.17-19, 78-79, 145, 158-160, 190-191 (page image p.191). Tags as above; [calc] is my own arithmetic. No new web searches were needed; two memory items are marked.

Summary: no final verdict on claims 1-5 reverses. One sub-verdict is added (2g), one claim row is added (6, from material the first pass omitted), and several statements in this file were wrong or incomplete and are corrected.

A. Claim 2: April 1882 has a second witness, and a third timing datum

1. Correction. The Source gap note said none of Larkin's own observations has a second witness. For 16-17 April 1882 that is wrong. Larkin prints "Professor S. P. Langley, of the Allegheny, Pa., Observatory, wrote on the 17th": streamers "converged to a point 10 degrees south of the zenith", and Langley calls attention to "its coincidence with a remarkable sun spot, which is now visible to the naked eye ... two large and five small nuclei, united by penumbrae, traversed by 'bridges' of intense brightness", area "a thousand million square miles" at a minimum [read: p.149]. - Weight: an independent observer at another observatory, who independently uses the word "bridges". It is a secondhand quotation (Larkin prints it; the original letter was not consulted) and it confirms the coincidence of one great spot and one great aurora, not the light-speed mechanism. Larkin's group (four portions, largest spot 67,000 x 48,000 miles, seen on the 16th) and Langley's (two large and five small nuclei, on the 17th) are different descriptions at different times; the area figures are of the same order (67,000 x 48,000 is 3.2 billion square miles as a rectangle, about 2.5 billion as an ellipse [calc]) and are not reconciled on the page. The November 1882 jets, the stellar floor and the 1901 meteors remain single-observer.
2. The 16 April 1882 timing, weighed for 2(b) [read: pp.147-148]. The spot was seen at 15 degrees from the east limb on the 15th, quiet through the day, and "during the night activity set in". On the 16th: from 8 a.m. violent agitation; 9 a.m. two tongues pointing at each other; 12:30 p.m. their ends met and passed; 2 p.m. a circle, "clearly a cyclone"; 9 p.m. a yellowish-green auroral arc rose; 10 p.m. the great outburst; midnight to 1 a.m. the maximum. - The gap between visible solar change and aurora is therefore about 7 hours (from 2 p.m.) to 12 hours (from 9 a.m.), longer if the overnight change counts. Larkin gives no observing site on these pages and no time standard, and he does not discuss the lag [read: p.147-149]; so the gap is approximate. - Light-speed travel is 8.3 minutes [calc: 1.496 x 10^11 m / 2.998 x 10^8 m/s = 499 s]. A coronal mass ejection (CME) at 400 to 2,400 km/s takes 17 to 104 hours (above). Seven to twelve hours fits neither; it is faster than the usual CME, as the 1859 storm (17.6 hours) was. - Does it change the verdict? No. It is one more dated case where the sun's visible activity leads the aurora by hours, not minutes. It does not support a prompt light-speed effect, and a sub-day lag is not evidence against a fast CME either (fast CMEs can arrive in about 15 to 20 hours). The sequence is also consistent with ordinary coincidence of a great spot and a great storm. Larkin's own chart note (below) reminds the reader that the correlation is by year, not by hour. 2(b) stays Tested, not supported, now with three Larkin timings on record (1882 November, 1882 April, 1892 July), none prompt.
3. Loomis's chart, Cut 64 (pp.149-150), is Larkin's evidence for 2(a). It plots 1775 to 1875: auroras per year (117 in 1786, 4 in 1810), mean daily range of magnetic declination (Larkin: 12 degrees in 1786), and sunspots (the cut is captioned Wolf's relative number; Larkin's prose says number of groups, "about 110" in 1786). Larkin writes "The coincidence is complete" and then states his own exception: "in 1805 there were quite a number of spots, but few auroras, while in 1870 they agreed again"; he forecasts a spot maximum in 1903-4 [read: pp.149-150; page image p.150]. - Weighing. An 11-year-cycle correlation of aurora counts with sunspot number over a century is a real, well-known result and 2(a) stands as Established. Two cautions the deep reader raised: "12 degrees" a day is implausibly large for a daily declination range (minutes of arc is likelier [memory: medium]; unresolved, the chart's unit is not legible); and aurora counts before 1800 depend on observing effort. This is a correlation of yearly numbers, which supports "the sun drives it" but says nothing of mechanism or lag.
4. Mechanism as stated on pp.143-145, now read. Two tiers: (i) weak sun-warmed thermo-electric currents in ore beds flowing east to west cause the small daily magnetic variation, and Larkin says outright "Thermo-electric currents are weak, and are only able to set up minute daily variations" [read: p.144]; (ii) storms and aurora come from "a wave front of magnetic radiation from the sun" (Bigelow's sphere, Cut 62) striking an earth that, like a coherer, "cut[s] waves out of space" [read: pp.144-145]. Tier (i) is rated 2(g): the daily variation is an ionospheric dynamo effect first proposed by Balfour Stewart in 1882 [memory: high; not web-checked]. His claim that a magnet's deflection varies "in exact ratio" with the current is also loose (the standard result is a tangent law [memory: medium]). Tier (ii) is the claim already rated in 2(b)-2(d). That the sun is magnetic with a general and a sunspot field was established by Hale from 1908 [memory: high], so Bigelow's "spherical magnet" sun was a good guess whose "magnetic radiation" mechanism does not follow.

B. Claim 3: the "4,000,000 yet to run" cut-off (Mathematician's correction 5)

Weighed, and I side with the Mathematician over the deep read's stronger wording. The 49.05 million F budget is the heat already released in shrinking to today's size. For a uniform sphere the released heat goes as 1/R, so a further 4 million years at today's rate needs another 4/21.8 = 0.18 of the budget, i.e. shrinkage to R/1.18 = 0.85 R [calc]. Larkin also writes on p.179 that halving the diameter would release as much again, so he knew the 1/R law [read: p.179]. So the figure is an unexplained stopping point, not an impossibility ("cannot then be split" overstated it). It does not change the verdict on claim 3 (the sun's real age of 4.57 billion years needs 240 times the budget [calc, above]). The primer should soften its wording accordingly.

C. Claim 5: Larkin's own caveat beside the settling test

The settling test (calibrated photometric map of the fifteen table positions) must be read with Larkin's prediction: "It is doubtful if long time graphs will show this pavement as well as the eye, for the stars in the foreground will store their light in excess in front of the distant wall" [read: p.300; deep read of pp.297-308]. Modern surface photometry with foreground stars masked answers his objection in principle, but any photographic null result is one he anticipated, and a visual-only effect remains a possibility until the masked photometry is done. This does not change Never properly tested for the floor as an eye observation.

A related point from the deep read of pp.205-231: Larkin offered interstellar absorption ("quenched in space, by an absorbing medium, or by cosmic dust", p.219) as one of three explanations for the shortfall of faint-star counts (plates 100 million; his extrapolation 733 million). Interstellar extinction was established by Trumpler in 1930 [memory: high]. Add it beside the dark patches under "what is solid". No verdict changes.

D. New claim 6: energy "out of space", and "it should be free"

The first pass said only that the book describes no free-energy device. That is true, but incomplete. Larkin states a programme, in his own words: "The combined energies of man must be centered on the herculean work of taking energy out of space and chain it to his servitude. Men pay for power now; it should be free" [read: p.78]; "All problems so far presented to men for solution fall into insignificance in presence of the mighty problem of the ages - how change the lengths of waves from the sun and make them fit for human use?" [read: p.78]; "The problem of all ages is now to either lengthen them for use of man or find some way to cut them out of space and make use of them as they are" [read: p.190].

What he means [read: deep read]: - "Out of space" is solar radiation (and, in his corpuscle picture, the store that radioactive bodies draw on, p.18). Two routes: mirrors, boilers and accumulators (pp.159-160), and a coherer-like detector (p.145). The second is stated as a problem, with no design. - Solar constant: Langley's 3 calories per square centimetre per minute (Mount Whitney, p.157). This is 2.09 kW/m^2 [calc], about 50% above the modern 1.361 kW/m^2 [memory: high], which the reviewers' energy-density numbers already use. His "slightly less than three horse-power" per square yard is 2.35 hp from his own constant [calc: 2,093 W/m^2 x 0.836 m^2 / 746] and about 1.5 hp from the modern one, so it is about 25% high even by his own figure. - Machines: Ericsson, 176 square feet of mirror, 3 hp from about 35 hp; the Cawston Ostrich Farm motor, South Pasadena (1,788 mirrors, 70 square yards, boiler of 100 gallons, 210 lb steam), 11 hp indicated from about 130 hp, pumping 1,400 gallons per minute at maximum [read: p.159-160]. Both are 8.5% [calc: 3/35, 11/130]; both match his chain: air removes one third, the engine uses one eighth of the rest, 2/3 x 1/8 = 1/12 [calc]. The mirror area checks (33-foot top and 16-foot base circles: 72.7 square yards [calc] against 70). The pumping implies a lift of about 31 feet [calc], which he does not state. "Indicated" means steam-engine indicator horsepower, not delivered work, and no test date, duration or witness is given. He reads the 1/12 as "eleven-twelfths going to waste", a problem to solve; he does not mention a heat-engine temperature limit. - Coherer (p.145): the wave "switches in a local battery"; the battery does the work. "Cutting waves out of space" there is plain detection, not an energy source. Larkin does not draw that distinction on the page (a point from the deep read and mine). - Radium (p.18): "Radium is so active that its potential has been measured and with a delicate watt meter its power has been determined"; radioactive bodies are "simply more transparent to corpuscles from space". This is the claim 1(c) idea and keeps that verdict.

Verdicts: see table row 6. Interest to a reader looking for the Tesla/Moray sense of "radiant energy": Larkin's phrase is the same words with a different meaning; his space-energy is sunlight, and he proposes no device that draws on a source other than the sun. The reception of the title by later readers (the spiritualist press, 1904 and 1906) is a matter of what they wanted from it, as above.

E. Other corrections and additions

1. Lowe directorship: the book implies Larkin succeeded Swift after 31 December 1900 (p.319) and gives no appointment year (p.300 says "when taking charge" without a date). Corrected in the Investigator section above.
2. Digest chunk 04, p.191: Langley's table reads 19/100 for the visible share (page image checked by the deep reader), not 10/100; 1 + 19 + 80 = 100 [read: page image p.191]. Digest line to be corrected by the orchestrator; no verdict depends on it.
3. Records to obtain, added to "What would settle it": the sites.md "Other Larkin items": PASP 1901 Leonids report; Science 1892 "Meteoric Shower"; Scientific American 1909-1910 (the Orion "abyss", which bears on claim 5, "Signaling to Mars", "Motions of the Earth"); Los Angeles Herald 17 February 1903 lecture; Open Court, July 1906. Also, for claim 2: the original Langley letter of 17 April 1882 (source journal not named in the book) and Loomis's original compilation, to resolve the declination unit and the 1805 exception.
4. Open question added: what was the lag between the April 1882 spot changes and the magnetic record of 16-17 April 1882 at an observatory with a magnetograph (time standards for Larkin's site and Langley's)?

F. Corrections to earlier documents (additional to the list above)

1. arbiter.md, Source gap note 1 (above): corrected in place; the April 1882 aurora has Langley as a second witness [read: p.149]. Same error in facts.json (main entry summary, "no second witness is cited for any of them"): the orchestrator should fix it.
2. arbiter.md, Source gap note 2 and "Weighing the Investigator's findings", dates bullet: corrected in place.
3. arbiter.md framing sentence: "describes no free-energy device" now carries the p.78 aim. facts.json, claims.md (framing note) and primer.md (framing paragraph, section 1) need the same addition, with the pp.78, 190 quotations.
4. primer.md section 5, step 5, point 2: "cannot then be split" to be softened (section B above).
5. Not changed by this revision, for the orchestrator to apply to the source of truth: arbiter corrections 1-7 and 9 (1892 date, 133 mA, "consistent with computed", "ruled out" in claims.md claim 5, "copied", Barnard's Loop downgrade, Lowe 1 January 1901) remain unapplied in claims.md, facts.json, deep/ sections 4, 7, 8 of the pp.124-142 read, research/2026-10-08-google-books.md and sites.md, as the QE found. The deep read of pp.143-150 section 10 asks that the digest or facts entry record that the thermo-electric currents are for the daily variation only, and that 1805 is Larkin's own exception.
6. Primer reader-profile slips (QE): define AU (astronomical unit, 1.496 x 10^11 m) at first use in section 3.3; spell out Astronomical Society of the Pacific (ASP) once.