Larken: Primer
Edgar Lucien Larkin, "Radiant Energy and Its Analysis: Its Relation to Modern Astrophysics" (Los Angeles, copyright 1903). The folder is named "Larken"; the author spells his name Larkin, and this primer uses his spelling.
Source note (no source gap for the book; two limits). The project folder is one scanned book of 342 pages with a good text layer (23 pages OCR'd). Nine passages were deep-read in the first pass and two more after the quality check (QE: pp.143-150; pp.17-19, 78-79, 145, 158-160, 190-191), and every table and figure that a verdict depends on was checked on page images [Arbiter, "Source gap note"]. Two limits apply to everything below. 1. Everything in the folder is Larkin's own, and most of his observations are single-source. The independent records come from the Investigator: trade-press snippets for the storm of 16 July 1892, an editorial note in the Publications of the Astronomical Society of the Pacific (PASP), a 1906 critique, and later reception in the spiritualist press. Three of Larkin's own observations (the November 1882 sunspot jets, the "stellar floor", the 1901 meteors) have no second witness on record; they are reported here as one observer's account, not dismissed. One does have a second witness: for the great sunspot and aurora of 16-17 April 1882 the book itself quotes a letter from S. P. Langley of the Allegheny Observatory (section 7) [read: p.149]. (An earlier version of this primer said none had one; the QE found Langley's letter.) 2. Pages 1-123 and 182-204 (textbook spectroscopy, solar photography, instruments) were covered by digests only, except pp.17-19, 78-79 and 190-191, which were deep-read after the QE together with pp.143-150 (aurora history and Larkin's storm mechanism) and pp.158-160 (solar engines) [Arbiter, "QE revision"]. Where a statement rests on [memory] alone it says so.
Framing. "Radiant energy" in the title means electromagnetic radiation in the 1890s sense of Melloni and Langley. The book describes no free-energy device and no non-Hertzian wave [read: pp.17-19, 39-40, 177-178]. It does, however, state an aim: "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]; and "The problem of all ages is now to either lengthen them [the sun's waves] 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]. By energy "out of space" Larkin means sunlight: his examples are solar engines with mirrors and boilers (John Ericsson's, and the Cawston motor in South Pasadena) [read: pp.158-160]. A reader drawn by the Tesla or Moray sense of "radiant energy" will find the same words with a different meaning (claim 6 in section 6). It is a popular astrophysics text of 1903 by the director of the Lowe Observatory on Echo Mountain above Pasadena, mixing correct textbook physics with Larkin's own large ideas: a universe made only of J. J. Thomson's "corpuscles" (electrons), the sun as an electrical driver of storms on earth, and a luminous "floor" behind the Milky Way. Most of the arithmetic is careful. The interest for a learner lies in seeing which of his starting points were sound physics, which were the best physics of 1903 and later replaced, and where his inferences went beyond what the evidence supported.
Tags used: [read: p.N] from the book; [web] with URL; [memory] recalled, with confidence; [calc] arithmetic checked in example.py or by the reviewers. Bracketed role names ([Arbiter], [Physicist], [Mathematician], [Contrarian], [Investigator]) say which review file a point comes from.
1. The idea in one paragraph
Larkin believed that the whole universe is made of one thing: the tiny charged particles that J. J. Thomson had just discovered in 1897, which Larkin called "corpuscles" and we call electrons. Light, radiant heat and the "actinic" (chemically active, ultraviolet) rays were, he said, one wave motion carried by a sea of these corpuscles filling space, so the old ether was not needed. From this he built a chain of ideas: storms on the sun send electromagnetic waves at the speed of light that shake the earth's magnetism, drive surges in telegraph wires and light the aurora; the sun's heat comes from its slow gravitational shrinking (the accepted view in 1903); the stars, planets and moons were built from a primordial cloud of corpuscles that gravity gathered, electrical induction set spinning, and tidal friction shaped; and, looking through the observatory's 16-inch telescope, he reported a faint luminous "floor" lying behind the stars of the Milky Way, broken by dark "rents" and "starless fields". He also set his readers a goal: to take energy "out of space" and make power "free", by which he meant catching sunlight with mirrors and engines, or finding some way to use the sun's waves directly [read: pp.78, 158-160, 190]. Some of this is exactly right (the sun does drive magnetic storms and aurora; heat and light are one kind of radiation; the dark patches he saw are real), some was the best science of its day and has since been replaced (contraction as the sun's power source), and some steps do not follow from his premises.
2. What you need to know first
Electromagnetic (EM) waves (Maxwell). A light wave is an oscillating electric field E and magnetic field B, at right angles to each other and to the direction of travel, moving at c = 3.00 x 10^8 m/s in vacuum. Frequency f and wavelength lambda satisfy c = f lambda; the amplitude (how strong the fields are) is a separate number that does not enter c or lambda. Radio, infrared ("radiant heat"), visible light and ultraviolet ("actinic rays") differ only in wavelength. The wave carries energy at the rate given by the Poynting flux S = E x B / mu_0 (mu_0 = 4 pi x 10^-7 H/m, the permeability of free space). Read: D. J. Griffiths, Introduction to Electrodynamics, the chapter on electromagnetic waves.
Gravitational potential energy of a sphere. Bringing mass together under gravity releases energy. A uniform sphere of mass M and radius R has gravitational potential energy U = -(3/5) G M^2 / R (G = 6.674 x 10^-11 m^3 kg^-1 s^-2 is the gravitational constant). The minus sign means energy was released when it formed; the magnitude (3/5) G M^2 / R is what is available as heat. You will derive this in section 5. Read: any calculus-based mechanics text (Halliday, Resnick and Walker, Fundamentals of Physics, gravitation chapter), and for stars B. W. Carroll and D. A. Ostlie, An Introduction to Modern Astrophysics, the chapter on the interiors of stars.
The virial theorem (plain version). For a ball of gas held together by its own gravity and in balance, the total thermal (heat) energy K and the gravitational energy U obey 2K + U = 0. So when the ball shrinks and releases an amount of gravitational energy, half of it goes into heating the gas and only half can be radiated away. A self-gravitating gas ball that loses energy therefore gets hotter, which is J. Homer Lane's 1870 result that Larkin quotes [read: p.180]. Read: Carroll and Ostlie, same chapter.
Heat, specific heat and old units. Heat Q needed to raise mass m by temperature change Delta T is Q = m c Delta T, with c the specific heat (water: 4186 J kg^-1 K^-1, or 1 British thermal unit (BTU) per pound per Fahrenheit degree). A Fahrenheit degree is 5/9 of a kelvin. Joule's mechanical equivalent of heat: 778 foot-pounds per BTU (Larkin uses Joule's older 772). Larkin often expresses an amount of energy as "degrees F" of an equal mass of water: that is a heat budget, not a temperature. Read: any intro thermodynamics text (e.g. Halliday et al., heat chapters).
Radioactive decay. A radioactive sample decays as N(t) = N_0 exp(-t / tau), with a fixed half-life t_half = tau ln 2 that is characteristic of the isotope and does not change with temperature, pressure or location. Each decay releases a definite energy (radium-226: 4.78 MeV per alpha particle; 1 MeV = 1.602 x 10^-13 J). Read: K. S. Krane, Introductory Nuclear Physics, chapter on radioactive decay; or any intro modern physics text (e.g. Serway, Moses and Moyer, Modern Physics).
Space weather. The sun's magnetic activity throws out clouds of magnetised plasma (coronal mass ejections, CMEs) that reach the earth in roughly 1 to 4 days and squeeze the earth's magnetic field, producing magnetic storms, aurora, and ground currents in long wires. Flares also emit X-rays that arrive in 8.3 minutes and briefly change the upper atmosphere's conductivity. Read: NASA or NOAA Space Weather Prediction Center introductions; the British Geological Survey page on the Carrington event, https://geomag2.bgs.ac.uk/education/carrington.html [web].
Surface brightness. A telescope collects more light from a point-like star, but it cannot make a spread-out glow (a nebula, the night-sky background) look brighter per unit area than it does to the naked eye; it only magnifies it. This is why faint diffuse light is a hard thing to see through a large telescope. Read: any observational astronomy text (e.g. C. R. Kitchin, Telescopes and Techniques).
3. The textbook view
3.1 Light is a field wave, not a motion of particles
Maxwell's equations in empty space (no charges, no currents) give the wave equation
nabla^2 E = (1/c^2) d^2E/dt^2, with c = 1 / sqrt(mu_0 epsilon_0),
where nabla^2 is the Laplacian (sum of second partial derivatives in x, y, z) and epsilon_0 = 8.854 x 10^-12 F/m is the permittivity of free space. The wave needs no medium. The energy density is u = epsilon_0 E^2 (averaged, half electric, half magnetic), and the intensity is I = u c. Sunlight at the earth carries I = 1361 W/m^2, so u = 1361 / c = 4.5 x 10^-6 J/m^3 [calc, Mathematician]. The Michelson-Morley experiment (1887) found no change in light speed as the earth moves at 29.8 km/s through any supposed medium [memory: high], and Einstein's relativity (1905) removed the need for one.
Thomson's discovery (1897) was that cathode rays are particles with charge-to-mass ratio e/m = 1.76 x 10^11 C/kg, bent by magnets and electric fields. That shows the particles are charged; it says nothing about whether a light wave is a motion of such particles.
3.2 Radioactivity carries its own energy
A gram of radium-226 has 3.7 x 10^10 decays per second (the definition of the old unit, the curie). At 4.78 MeV per alpha: 3.7 x 10^10 x 4.78 x 1.602 x 10^-13 J = 0.028 W per gram from radium's own alphas, about 0.14 W with its decay products [calc, Arbiter; memory: medium for the chain]. Curie and Laborde measured about 100 calories per gram per hour in March 1903, which is 0.116 W/g [web: https://www.nature.com/articles/068622a0]. The activity falls exponentially with a fixed half-life, and helium is produced as the atom changes (Ramsay and Soddy, 1903 [memory: high]). The energy is stored in the nucleus; the atom is used up.
3.3 The sun drives magnetic storms by a plasma cloud, not by a light-speed wave
A flare's light and X-rays reach the earth in d/c = 1.496 x 10^11 m / 3.0 x 10^8 m/s = 499 s = 8.3 minutes, where d is one astronomical unit (AU), the mean earth-sun distance, 1.496 x 10^11 m. The great storm usually begins hours to days later, when the CME arrives: 1 AU at 400 km/s takes 104 hours, at 2,380 km/s it takes 17.5 hours [calc, Arbiter]. In 1859 the storm began about 17.6 hours after Carrington's flare [web: https://geomag2.bgs.ac.uk/education/carrington.html]. There is one real prompt effect: the "magnetic crochet", a deflection of about 100 nT (nanotesla) seen at Kew in 1859 within minutes, caused by flare X-rays raising the ionosphere's conductivity [web: same page]. It is small next to the storm itself. The earth's surface field is about 30,000 to 60,000 nT.
A changing magnetic field induces an electric field in the ground (Faraday's law, curl E = -dB/dt), and a long telegraph line earthed at both ends picks up a voltage of order (field in V/km) x (length in km). Great storms produce a few volts per kilometre [calc, Physicist and Mathematician: 1.6 to 2.2 V/km over roughly 300 km is plausible].
Two further facts Larkin did not have. The sun is itself magnetic, with a general field and strong fields in sunspots, as George Ellery Hale showed from 1908 [memory: high]. And the small, regular daily swing of the compass needle is caused by electric currents in the sunlit upper atmosphere (an ionospheric "dynamo", first proposed by Balfour Stewart in 1882) [memory: high; Arbiter, QE revision A].
3.4 Where the sun's energy comes from
Gravitational contraction (Helmholtz 1854, Kelvin 1862) releases U = (3/5) G M^2 / R for a uniform sphere. Divided by the luminosity L (power radiated, 3.83 x 10^26 W) this gives the Kelvin-Helmholtz time, about 19 million years (about 9 million once the virial theorem's half-share is taken into account) [calc]. This was the best known source until radioactivity (1896-1903) and nuclear physics. The sun's age from meteorites is 4.57 billion years [memory: high], and the power comes from hydrogen fusion in the core (Eddington 1920, Bethe 1938-39), confirmed by solar neutrinos measured since 1968 [memory: high]. Contraction does still power young stars before fusion starts, and part of Jupiter's internal heat [memory: high].
3.5 How galaxies and moons get their spin
A cloud conserves angular momentum J = I omega (I the moment of inertia, omega the spin rate). For a sphere I = (2/5) M R^2, so halving R quadruples omega: Larkin's "conservation of areas" is correct [read: p.290]. Today's account of where the spin of a galaxy comes from in the first place is tidal torque: the gravity of neighbouring lumps twists a lopsided cloud (Hoyle, Peebles, White [memory: medium-high]). The earth-moon system shows tidal friction at work now: the moon recedes about 3.8 cm per year while the earth's day lengthens [web: https://ilrs.dgfi.tum.de/lw20/docs/2016/presentations/05-Gross_presentation.pdf].
3.6 Dark patches in the Milky Way
E. E. Barnard and Max Wolf showed from 1904 onward, with wide-field photographs, that many dark patches in the Milky Way are clouds of dust in front of the stars (dark nebulae), not empty holes [memory: high]. Barnard 92 and the Pipe Nebula in Sagittarius and Ophiuchus are examples [web: https://en.wikipedia.org/wiki/Barnard_92].
4. Where Larkin departs
| Topic | Textbook (then or now) | Larkin [read] | Where the step fails or holds |
|---|---|---|---|
| Nature of light | One EM field wave, no medium needed | Heat, light and actinic rays are one wave motion (agrees, after Langley and Melloni, pp.177-178); the wave is carried by corpuscles that "saturate space" and replace the ether (pp.17-19) | First part holds. Second: "corpuscles bend in a magnet, therefore the entire wave motion of the universe is electromagnetic" (p.19) is a non sequitur; bending shows charge, not that light is a motion of particles. A medium at rest is also contradicted by Michelson-Morley (1887). He also ties wavelength to amplitude (pp.19-20); they are independent |
| Radioactivity and enzymes | Energy stored in the nucleus; catalysts add no energy | Radium, polonium and enzymes "merely direct radiance already saturating space" without themselves changing (pp.18, 174) | Tested and not supported: they decay with fixed half-lives and transmute. In 1903 the source of radium's energy was a live puzzle (Marie Curie herself floated, in passing, the idea that radium might draw energy from space [memory: high; see the Becquerel-Henri primer]), so the idea was not absurd then [Contrarian] |
| Sun-earth link | CME plasma, arriving 1-4 days later; small prompt X-ray effect; the daily compass swing from ionospheric currents | EM waves from solar jets and sunspot bridges at light speed shake the earth's field, which is "tuned" to the sun's (pp.124-137). His evidence for the link is Loomis's chart of aurora counts, daily compass swing and sunspots, 1775-1875: "The coincidence is complete", with his own exception that 1805 had "quite a number of spots, but few auroras" (pp.149-150). Two tiers of mechanism: weak sun-warmed thermo-electric currents in "vast beds of unlike ores" for the small daily swing only, and "a wave front of magnetic radiation from the sun" (after Frank Bigelow's sun as a "spherical magnet") for storms, the earth acting like a coherer that "cut[s] waves out of space" (pp.144-145) | The link is right and he saw it early; Loomis's century of yearly counts is a real correlation, but it says nothing about mechanism or lag. The mechanism fails the timing test: his own November 1882 jets were drawn 19 November, two days after the 17 November storm, and he writes that spots occur without storms and storms without spots (pp.126-128). His own record of 16 April 1882 (spot changes from 9 a.m. to 2 p.m., aurora from 9 p.m.) puts 7 to 12 hours between the visible solar change and the aurora, again not 8.3 minutes (pp.147-148) [Arbiter, QE revision A]. A magnetic sun was a good guess (Hale, 1908); the daily swing is an ionospheric current, not ore-bed thermo-electricity. "Tuned" has no frequency or number attached |
| Sun's power | Contraction in 1903; fusion now | Contraction from a sphere 25 trillion miles across at absolute zero, Helmholtz and See: 49,048,200 F water-equivalent, 21.8 million years (pp.168-179) | In 1903 this was the textbook. His arithmetic checks (section 5). Missing: the virial half-share, and the past/future split (section 5, step 6) |
| Catalysis | A catalyst changes a rate, supplies no energy | Catalysis "may have been the first mode of energy" in the cold nebula, "offered for what it is worth, if anything" (p.175) | Hedged by Larkin himself; a catalyst cannot be an energy source by definition, and he also says chemistry is "completely arrested" by the cold (p.171) |
| Spin of nebulae | Angular momentum conservation; tidal torques | Electrical induction from a charged neighbour sets the cloud spinning ("gravity and inductance", p.290) | His own premise defeats it: a cloud of negative corpuscles repels itself 4 x 10^42 times more strongly than gravity pulls [calc, Arbiter]; a neutral cloud has no net charge to induce. The "areas" law and the idea of a neighbour twisting a cloud are sound |
| Hidden mass | Milky Way about 10^11-10^12 solar masses, mostly dark matter | Fast stars imply far more mass than visible: about 11 quadrillion dark earth-like bodies, one live sun in 320 (pp.291-293) | "More mass than we see" is right in order of magnitude; "dead suns" as the main component is limited by microlensing surveys [web: https://arxiv.org/pdf/astro-ph/9612102] |
| Stellar floor | Dark nebulae on a background of unresolved stars, airglow and diffuse galactic light | A luminous floor behind the Milky Way with over 100 rents and tubular "starless fields"; a year of comparisons offered as ruling out dust haze; the control is a same-night contrast between fields, which excludes only a uniform haze (pp.299-303) | The dark patches are real and catalogued; the "tubes" reading was replaced by dust. The floor as something he saw is an untested single visual report |
| Energy "out of space" | Sunlight is a real energy flow (1.361 kW/m^2 above the air); a heat engine turns only a fraction of it into work | Take energy "out of space" and make power "free": mirrors and boilers (Ericsson 3 hp from about 35 hp of sunlight; Cawston motor 11 hp indicated from about 130 hp), or "cut" the sun's waves "out of space" as a coherer does (pp.78, 145, 158-160, 190) | The solar engines are real machines and his figures are consistent (both about 8.5%) [calc, Arbiter]. His solar constant (Langley's 3 calories per square centimetre per minute, 2.09 kW/m^2) is about 50% above today's [calc, Arbiter]. The coherer is a detector: the wave only switches in a local battery, which does the work, so it is a trigger, not a source. "Free" is an aim with no design behind it |
The common pattern: Larkin's numbers are usually right (or right on his inputs), and his observations are honestly reported, often with his own caveats ("the dark places seem so to the eye through contrast probably" [read: p.302]). Where he goes wrong it is in the step from an observation or a fact to a mechanism: from "corpuscles bend in a magnet" to "light is a corpuscle wave", from "storms follow sunspots" to "a light-speed wave causes them", from "charged spheres attract light objects" to "induction spins nebulae", from "a coherer is set off by waves" to "waves can be cut out of space for power".
5. Worked example: the Helmholtz contraction budget of the sun
This is the one place where Larkin gives a full quantitative chain (pp.165-179), and it was mainstream physics in 1903. We redo it exactly, see that his numbers are right, find the two physics slips, and then see why the whole budget is about 240 times too small for the real sun. All numbers come from example.py in this folder; run it with
.venv/bin/python review/Larken/example.py
Symbols. G = 6.674 x 10^-11 m^3 kg^-1 s^-2 (gravitational constant); M = 1.989 x 10^30 kg (sun's mass); R = 6.957 x 10^8 m (sun's radius); L = 3.828 x 10^26 W (sun's luminosity); c_w = 4186 J kg^-1 K^-1 (specific heat of water); rho = density; r = a radius inside the sun. Modern constants are [memory: high].
What Larkin prints. The sun's mass is "two octillion tons", its diameter 866,000 miles; contracting from 25 trillion miles (or from infinity) releases heat that would raise "two octillion tons of water" by 49,048,200 F (p.172; rounded to 49,050,000 on p.178); "the best estimates of physicists are that the sun is radiating heat enough to reduce its temperature 2 1/4 degrees per year, and 49,050,000 divided by this gives 21,800,000 years as the life of the sun" (p.179) [read: text layer, page image p.179 checked by the deep read].
Step 1. Check his inputs. 2 x 10^27 short tons x 907.2 kg = 1.81 x 10^30 kg, 9% below the modern mass (he writes "nearly"); 866,000 miles = 1.394 x 10^9 m, within 0.2% of the modern diameter [calc].
Step 2. Derive the energy. Build the sun shell by shell, bringing each thin shell in from far away. When the sphere already built has radius r, its mass is m(r) = (4/3) pi r^3 rho. A shell of thickness dr has mass dm = 4 pi r^2 rho dr. Bringing it from infinity to radius r releases dU = G m(r) dm / r. Integrate from r = 0 to R with uniform rho = M / ((4/3) pi R^3):
U = integral from 0 to R of G (4/3 pi r^3 rho)(4 pi r^2 rho) / r dr = (16/15) pi^2 G rho^2 R^5 = (3/5) G M^2 / R
(SymPy does this integral in example.py, step 2.) Starting from a finite radius R_0 instead of infinity gives (3/5) G M^2 (1/R - 1/R_0). Because 1/R_0 is tiny for R_0 = 12.5 trillion miles (R/R_0 = 3.5 x 10^-8), Larkin is right that infinity and 25 trillion miles give the same answer [calc].
Step 3. Put in numbers. U = 0.6 x 6.674 x 10^-11 x (1.989 x 10^30)^2 / 6.957 x 10^8 = 2.28 x 10^41 J [calc].
Larkin's unit: heat this to an equal mass of water. Delta T = U / (M c_w) = 2.28 x 10^41 / (1.989 x 10^30 x 4186) = 2.73 x 10^7 K. A kelvin step is 9/5 Fahrenheit degrees, so Delta T = 49.2 million F. See's 49,048,200 F agrees to 0.4% [calc]. Note that this Delta T = (3/5) G M / (R c_w) is a bookkeeping number: the sun is not water and was never heated to 49 million degrees.
The "stage" figures follow the 1/R law: the share of the heat released before the sun's edge passed a radius R_x is R/R_x. With his diameters: before Neptune's orbit 1/6,443 of the total, 7,612 F (he prints 7,202, 7,250 and, by his own subtraction, 7,450); before Jupiter about 42,500 F (he gives 43,960, which matches Jupiter's true orbit rather than the rounded "one billion miles"); before the earth's orbit 1/215; before Mercury's 1/83. These last two match him exactly [calc; read: pp.172, 178-179].
Step 4. How long does it last? Larkin: 49,050,000 / 2.25 = 21.8 million years. Modern: L x 1 year / (M c_w) = 3.828 x 10^26 x 3.156 x 10^7 / (1.989 x 10^30 x 4186) K = 1.45 K = 2.6 F per year, so his 2 1/4 is close [calc]. (A side note the script turned up: Larkin's own solar constant, Langley's 3 calories per minute per square centimetre [read: p.157], implies L = 5.9 x 10^26 W and 4.0 F per year, so the 2 1/4 he quotes did not come from his own 3-calorie figure [calc]. He attributes it to "the best estimates of physicists".) In time units: t = U / L = 2.28 x 10^41 / 3.83 x 10^26 = 5.9 x 10^14 s = 18.8 million years [calc].
Step 5. The two physics slips.
1. Virial half-share. By the virial theorem only half of U can be radiated; the other half heats the gas (this is why Lane's law, which Larkin quotes on p.180, says a contracting gas sphere gets hotter). The uniform-sphere clock is then 9.4 million years. A centrally concentrated sun (a "polytrope" of index n, a standard model in which pressure goes as density to the power 1 + 1/n; n = 3 is the classic model of the sun) has U = 3/(5 - n) G M^2 / R, which for n = 3 is 2.5 times the uniform value: 47 million years, or 23.6 million with the half-share [calc]. Anne Sewell Young's result that a centrally denser sun gives "more heat" [read: p.179] is the right direction.
2. Past versus future. The 49,050,000 F is the heat already released in shrinking down to today's size (p.178). Dividing by today's rate gives about 22 million years of past sunshine. Splitting it into "18,000,000 years" past and "4,000,000 yet to run" (p.179) needs a step he does not supply: future heat must come from shrinking below today's size, which this budget does not contain. By the same 1/R law, a further 4 million years at today's rate needs another 4/21.8 = 0.18 of the budget, which means shrinking to R/1.18 = 0.845 (about 0.85) of today's radius [calc; example.py, step 4]. Larkin knew the 1/R law (on p.179 he says that halving the diameter would release as much heat again), so the 4 million years is an unexplained stopping point, not an impossibility [Mathematician; Arbiter, QE revision B]. (Also 18 + 4 = 22, not 21.8.) And "reduce its temperature 2 1/4 degrees per year" treats the sun as cooling, which contradicts Lane's law two pages later [Mathematician; deep read].
Step 6. How fast must the sun shrink? Set the rate of release equal to L: dU/dt = (3/5) G M^2 / R^2 x (-dR/dt) = L, so
-dR/dt = L R^2 / ((3/5) G M^2) = 4.7 x 10^-7 m/s.
The diameter shrinks by twice that: 8.0 inches per day (16 with the virial half-share) [calc]. Larkin's "nine inches per day" (p.167) is the period's textbook figure and the right size. At 450 miles per arcsecond of the sun's disc (866,000 / 1,924, correct), 9 inches a day takes 8,676 years per arcsecond, not his 9,400 [calc]. So the shrinking was far too slow to measure in 1903, which is why the theory could not be tested directly.
Step 7. The decisive comparison. The sun has shone for 4.57 billion years (meteorite ages [memory: high]) at roughly today's luminosity (in fact it was somewhat dimmer when young, which changes this by tens of percent, not by factors of a hundred [memory: high]). Energy needed: L x t = 3.83 x 10^26 W x 4.57 x 10^9 x 3.156 x 10^7 s = 5.5 x 10^43 J. That is 242 times U, or 485 times the half that can be radiated [calc]. Hydrogen fusion converts 0.7% of the mass of the burning core (about 10% of the sun) into energy: 0.007 x 0.1 x M c^2 = 1.25 x 10^44 J, about 10 billion years at today's L [calc]. Only nuclear energy closes the books.
Step 8. Larkin's meteor check (his own 1882 text, p.165). The speed of fall onto the sun from far away is the escape speed v = sqrt(2 G M / R) = 617.8 km/s = 383.9 mi/s (his 383 is excellent). A hundredth of the earth's mass (5.97 x 10^22 kg) arriving at that speed carries (1/2) m v^2 = 1.14 x 10^34 J, which pays for 0.94 year of sunshine (his "one year"). To power the sun entirely by infall would need 2.0 x 10^15 kg/s, about 1% of an earth mass per year, growing the sun's mass by 3.2 x 10^-8 per year [calc]. The Mathematician estimated that this would drift Mercury's longitude by hundreds of arcseconds per century, against positions known to a few arcseconds; at 1% of the heat it would be 4 to 17 arcseconds per century, near the limit. So "not more than one per cent" is the right order, and it is the argument Kelvin used against the meteoric hypothesis [Mathematician; memory: medium-high].
What the example teaches. Larkin's arithmetic is good, and so is the physics he took from Helmholtz, Kelvin and See: gravitational energy really is released when a gas ball shrinks, and that really is how young stars shine. The two slips (forgetting that half the energy heats the gas, and splitting a past budget into past and future without saying how far the sun must shrink next) make his clock a little generous, not wrong in kind. What made the whole picture wrong was an energy source nobody in 1903 had yet identified, and the test that settled it came from outside astronomy: radiometric ages of rocks and meteorites. The book appeared in the same year that Curie and Laborde measured radium's heat [web: https://www.nature.com/articles/068622a0], and Larkin does not mention radioactivity as a possible solar source.
6. Evidence table
Verdicts are the arbiter's final verdicts, by part [Arbiter, "Corrected claim list"].
| Claim | Tested? | Result | Replicated? | Verdict |
|---|---|---|---|---|
| 1a. Heat, light and actinic rays are one EM wave | Yes (Melloni, Langley; Hertz; whole spectrum since) | Confirmed | Yes, universally | Established |
| 1b. A sea of corpuscles is the wave medium, replacing the ether | Yes, indirectly (Michelson-Morley 1887, before the book) | No medium detected; relativity needs none | Yes (many ether-drift repeats) [memory: high] | Historically superseded but instructive (stationary-sea form also Tested, not supported) |
| 1c. Radioactive bodies and enzymes only "direct" energy from space | Yes (decay law, half-lives, helium production, calorimetry, from 1902-03 on) | Energy is internal; atoms change | Yes, for a century | Tested, not supported |
| 1d. "Nothing exists but electrons", extended to souls (later works) | No measurable form | - | - | Not testable as stated |
| 2a. The sun drives magnetic storms, aurora and telegraph earth currents | Yes (Larkin's own basis: Loomis's 1775-1875 chart, pp.149-150, with his exception of 1805; the April 1882 spot and aurora, with Langley as a second witness, p.149) | Confirmed | Yes | Established |
| 2b. The driver is a light-speed EM wave from solar jets | Yes (storm onset times, 1859 onward; his own November 1882 drawings came after the storm; his own April 1882 record shows a 7-12 hour lag, pp.147-148) | Storms lag flares by hours to days (CMEs); a small prompt X-ray effect exists | Yes | Tested, not supported |
| 2c. Earth-current table (book p.134) | Partly: peak values match the 1892-93 trade press [web: Google Books snippets, research/2026-10-08-google-books.md] | Real magnitudes; the book misdates it (1882 for 1892) and misprints 144 for 133 mA; volts fit current x 3,700 ohm, so may be computed | No magnetogram for 16 July 1892 found yet | (No scale verdict; record corroborated in part) |
| 2d. The earth's field is "tuned" to the sun's | No | - | - | Not testable as stated |
| 2e. Magnetism is "a vortex in corpuscles" | Superseded by electron spin and domains (1907 onward) [memory: medium] | - | - | Historically superseded but instructive |
| 2f. Crookes and Geissler tubes reproduce the aurora | Yes | Same light; the green oxygen line was not made in a laboratory until 1925 (McLennan and Shrum) [web: https://www.nature.com/articles/115047a0] | Yes | Established as to the light; "every colour, no new mode" Tested, not supported |
| 2g. Sun-warmed thermo-electric currents in ore beds cause the daily magnetic variation (p.144; too weak for storms, he says) | Superseded by the ionospheric dynamo (Balfour Stewart 1882 onward) [memory: high] | - | - | Historically superseded but instructive |
| 3a. The sun's heat comes from gravitational contraction | Yes, through the age of the earth and meteorites | Budget about 240 times too small | Yes | Historically superseded but instructive (1903); Refuted by calculation (as a statement about the real sun) |
| 3b. Meteoric infall at most about 1% | By planetary orbits | Bound checks | Same argument as Kelvin's | Established (as an upper bound) |
| 3c. Catalysis as the "first mode of energy" | No | - | - | Not testable as stated |
| 4a. Earth-moon tidal evolution | Yes (lunar laser ranging: 3.8 cm/yr) | Confirmed | Yes | Established |
| 4b. Tidal friction builds everything; moon born by fission; gemmation of suns | Yes (later astronomy) | Replaced | - | Historically superseded but instructive |
| 4c. Electrical induction spins nebulae | By calculation from his own premises | Charge needed fails either way | - | Refuted by calculation |
| 4d. Far more mass than visible stars; the extra mass is "dead suns" | Yes (rotation curves; microlensing) | More mass: yes. Dead suns as most of it: limited to a minority | Yes | Missing mass Established; "dead suns" Tested, not supported |
| 4e. Cycle back to corpuscles | No | - | - | Not testable as stated |
| 5a. Dark patches at his table positions | Yes (Barnard and Wolf photographs, 1904 onward) | Dark nebulae at or near his positions (e.g. field 6 lies 0.19 degrees from Barnard 92 if his coordinates are 1900 epoch) [calc, Arbiter] | Yes | Established |
| 5b. "Tubular" empty sight-lines | Yes | Dust clouds, not holes | Yes | Historically superseded but instructive |
| 5c. A luminous floor visible to the eye beyond the Milky Way | No (one visual observer, no photometry) | - | No second observer found | Never properly tested |
| 5d. The floor as light from beyond the Galaxy | By calculation | Extragalactic light is only 1-4% of the dark-sky glow | - | Refuted by calculation |
| 6a. Concentrated sunlight can drive a steam engine (Ericsson; Cawston, pp.158-160) | No test of these machines is in the record; the book's figures only | Figures consistent: both about 8.5% (3/35, 11/130) [calc, Arbiter] | - | Established |
| 6b. Power "should be free" (p.78) | No (an economic aim; no plan beyond mirrors and accumulators) | - | - | Not testable as stated |
| 6c. Waves can be "cut out of space" or lengthened for use (pp.145, 190) | No (posed as a problem, no design) | The coherer passage mistakes a trigger for a source | - | Not testable as stated |
| 6d. Radioactive output is energy drawn from space (p.18) | As 1c | As 1c | As 1c | Tested, not supported |
7. What a learner should take away
What is solid. - Larkin's calculations are mostly careful and right on his inputs: the contraction budget, the meteor bound, the escape speeds of the earth (6.94 mi/s) and sun (383 mi/s), the Kepler-law binary-star masses, the Joule-equivalent "223 feet per second" rule [read: pp.163-164; Mathematician's table of checks]. The slips are numerical (2,467 printed as 1,475; Stefan's law done in ordinary instead of absolute degrees, 9.7 not 16) and do not touch the arguments. - He was right, and early, that the sun drives magnetic storms, aurora and telegraph currents, and he recorded telegraph lines working "by means of nature's electricity, without batteries" in 1892 [web: https://archive.org/details/jstor-1766041], as operators had in 1859. His own account of the great sunspot and aurora of 16 April 1882 has an independent witness: S. P. Langley wrote from the Allegheny Observatory on 17 April that the auroral streamers "converged to a point 10 degrees south of the zenith" and called attention to "its coincidence with a remarkable sun spot, which is now visible to the naked eye ... traversed by 'bridges' of intense brightness" [read: p.149]. Langley confirms the coincidence of spot and aurora, not Larkin's light-speed mechanism. - The dark "rents" and "starless fields" he tabulated lie on real dark nebulae. His observation was sound; his "tubular hole" reading was replaced within a few years by Barnard's dust clouds. - He foresaw interstellar absorption as one possibility. Photographs showed about 100 million stars where his extrapolation gave 733 million, and among three explanations he offered that their light "is quenched in space, by an absorbing medium, or by cosmic dust" [read: p.219; deep read]. Interstellar absorption was established by R. J. Trumpler in 1930 [memory: high]. - Concentrated sunlight really does run engines, and his figures for Ericsson's and Cawston's machines agree with each other (about 8.5% of the sunlight on the mirrors) [calc, Arbiter]. - The "areas" (angular momentum) argument for spin-up is correct, and his question from 1881 (what starts a nebula rotating?) was a real question; today's answer, tidal torque from neighbours, is close in spirit to his "a neighbouring body twists the cloud", although by gravity, not induction.
What is wrong, and why the mistakes are instructive. - An observation does not fix its mechanism. Bending in a magnet shows a particle is charged; it does not make light a corpuscle wave. Storms following sunspots shows a link; only timing tells you what carries it, and the timing (hours to days) rules out a light-speed wave. Larkin's own 1892 letter left the cause "a question that future physicists must decide"; the 1903 book asserts it. - Check your own premises against each other. A cloud of only negative corpuscles cannot collapse under gravity at all (electric repulsion beats gravity by 4 x 10^42 for two electrons), so the induction-spin step fails on his own terms. - Energy budgets are powerful but incomplete. The contraction calculation is correct physics that ran out of energy; the lesson is that a missing source (nuclear energy) can hide behind correct arithmetic, and that an independent clock (rock and meteorite ages) is what exposes it. - A trigger is not a source. A coherer "cuts waves out of space" only in the sense of detecting them: the wave switches in a local battery, and the battery does the work [read: p.145; Arbiter, QE revision D]. The same distinction applies to any device said to draw energy from a weak signal. - A catalyst changes rates, not energy. Calling catalysis or radioactivity a "conduit" for energy from space was tested: half-lives do not change with surroundings, and the atoms themselves transmute.
What is genuinely open (as a matter of the record, not the physics). - Did the storm of 16 July 1892 leave a magnetic record? It appears in Larkin's Science letter and in the 1892-93 trade press but not in the modern storm lists the reviewers could reach. That is silence, not contradiction; the Kew, Greenwich and Potsdam magnetograms and newspapers of 17-18 July 1892 would decide it [Arbiter, absence analysis A]. - Were the volts in the earth-current table measured, or computed from currents (nine of ten Boston rows give exactly 3,700 ohms against a stated 2,700)? - What did Larkin actually see as the "stellar floor": the Galaxy's own diffuse glow, airglow patches, or a contrast effect? A photometric map of his fifteen table positions, and his longer 1901 report to the Astronomical Society of the Pacific (ASP; never printed), would help. Larkin himself doubted that photographs would show it: "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]. Modern surface photometry with the foreground stars masked out answers his objection in principle, but a photographic null result is one he foresaw [Arbiter, QE revision C]. - Where are his observing logs and drawings (New Windsor 1880-88, Knox College 1888-1900, Lowe Observatory 1901-04)? None has been found. The likeliest explanations are ordinary loss of a popular writer's papers and dispersal after the 1905 fire reported at the observatory [memory: low, not relied on]; the record shows no sign of suppression, and none of a research programme that a successor took up [Arbiter, absence analysis B]. - 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? Langley's original letter (the book does not name the journal) and Loomis's original compilation (the unit of the daily declination range, which Larkin gives as "12 degrees"; the 1805 exception) would also help [Arbiter, QE revision E]. - Other Larkin writings not yet read: the PASP 1901 Leonids report; Science 1892, "Meteoric Shower"; Scientific American 1909-1910 (the Orion "abyss", which bears on the stellar floor; "Signaling to Mars"; "Motions of the Earth"); the Los Angeles Herald report of his lecture of 17 February 1903; Open Court, July 1906 [sites.md; Arbiter, QE revision E]. - Larkin's 1916 statement that he saw Tesla make a thin wire glow in a dark room is a single published account with no date or place [web: https://archive.org/details/matchlessaltarof00larkrich].
Kept in the record (whatever the verdicts). S. P. Langley's letter of 17 April 1882 on the aurora and the bridged sunspot, quoted in full by Larkin, is an independent observer's account [read: p.149]. The book was critically read before publication by William H. Knight and published with the Southern California Academy of Sciences' support [web: https://cdnc.ucr.edu/?a=d&d=LAH19030405.2.366.13]. The ASP printed his 1901 stellar-floor summary ("Orion is always bound in white") and said the full report was "too long to reproduce" [web: https://archive.org/details/jstor-40668215]. A 1906 Science piece signed "D. S. J." (probably David Starr Jordan [memory: medium]) ridiculed his earthquake article; that is an institutional critic's characterization of Larkin as a popular writer and tests none of the book's claims. The spiritualist press (Progressive Thinker, 1904 and 1906) and writers such as Hallock used the corpuscle principle for an "electrical soul"; that is reception by readers with an interest in that conclusion. Larkin himself held the principle consistently ("nothing exists but electrons", 1911 [web: https://archive.org/details/withinmindmazeor00lark]). The book closes with an advertisement for the Pacific Electric Railway, whose Mount Lowe line ran the observatory's tourist trade [read: p.335]; it is the railway's commercial notice and the setting of his employment, not evidence on the science. USC's library calls him "among the most-read scientists of his time" (about 9,000 bylines) [web: https://libraries.usc.edu/article/wizard-echo-mountain]. He took charge of the Lowe Observatory on 1 January 1901 in his own words in PASP 1902 (not 1890, as some sources give); the book itself gives no appointment year, only that Lewis Swift directed the observatory until 31 December 1900 [read: p.319; Arbiter].
8. Check yourself
- Larkin says the contraction heat would raise "two octillion tons of water" by 49 million F. Show that this number does not depend on what the sun is made of, only on G, M, R and the specific heat of water. If the sun had twice the mass and the same radius, what would the "degrees F" be?
- A storm on the sun is seen at 10:00. Under Larkin's mechanism, when should the magnetic storm begin? Under the CME picture with a cloud speed of 800 km/s? What observation would tell the two apart?
- Why does the virial theorem halve the Kelvin-Helmholtz time, and which law quoted by Larkin himself (p.180) expresses the same physics?
- Larkin's primordial cloud is made of negative corpuscles. Using e = 1.602 x 10^-19 C, m_e = 9.11 x 10^-31 kg and 1/(4 pi epsilon_0) = 8.99 x 10^9 N m^2 C^-2, compute the ratio of electric repulsion to gravitational attraction between two electrons. What does it imply for his cosmogony?
- Why can a 16-inch telescope not make a faint, uniform "stellar floor" easier to see than the naked eye does, while it does reveal fainter stars?
9. Further reading
Mainstream sources - D. J. Griffiths, Introduction to Electrodynamics (electromagnetic waves; Faraday's law). - B. W. Carroll and D. A. Ostlie, An Introduction to Modern Astrophysics (stellar interiors, the Kelvin-Helmholtz timescale, the virial theorem, nuclear energy generation; the Milky Way and dark matter). - K. S. Krane, Introductory Nuclear Physics (radioactive decay). - J. D. Burchfield, Lord Kelvin and the Age of the Earth (1975), the history of the contraction clock and its overthrow [memory: high]. - British Geological Survey, "The Carrington event" (flare, crochet and storm timing): https://geomag2.bgs.ac.uk/education/carrington.html - P. Curie and A. Laborde, radium's heat (1903): https://www.nature.com/articles/068622a0 ; Pierre Curie's Nobel lecture: https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf - Kelvin-Helmholtz mechanism (overview with the virial factor): https://en.wikipedia.org/wiki/Kelvin%E2%80%93Helmholtz_mechanism ; SUNY Stony Brook course notes: https://www.astro.sunysb.edu/fwalter/AST101/k-h.html - Birkeland's terrella (the rival laboratory aurora of 1896): https://en.wikipedia.org/wiki/Terrella
Larkin's own key works - Radiant Energy and Its Analysis (1903): https://archive.org/details/radiantenergyits00larkrich (the book reviewed here; the project folder's copy). - "Magnetic Storm, Aurora, and Sun-Spots", Science 20 (1892) 65-66: https://archive.org/details/jstor-1766041 (also in research/files/). - The 1881 Science series on primeval rotation and cosmic rings: https://archive.org/details/jstor-2900594 - Report on the stellar floor, PASP 14 (1902): https://archive.org/details/jstor-40668215 - Within the Mind Maze (1911): https://archive.org/details/withinmindmazeor00lark - The Matchless Altar of the Soul (1916): https://archive.org/details/matchlessaltarof00larkrich - Letter to Charles Hallock in Luminous Bodies (1906): http://iapsop.com/ssoc/1906__hallock___luminous_bodies.pdf
In this folder: example.py (the worked example), arbiter.md (final verdicts and what would settle each), physicist.md, mathematician.md, contrarian.md, sites.md and research/ (the Investigator's finds).
Answers
- Delta T = U / (M c_w) = (3/5) G M^2 / R / (M c_w) = (3/5) G M / (R c_w). Composition enters nowhere; only the mass, the radius and the specific heat of water. It is proportional to M, so twice the mass at the same radius doubles it to about 98 million F. (This also means Larkin's "two octillion tons", 9% low, would give about 45 million F; See's figure matches the true mass [calc].)
- Light-speed mechanism: at 10:08 (8.3 minutes later). CME at 800 km/s: 1.496 x 10^11 m / 8 x 10^5 m/s = 1.87 x 10^5 s = 52 hours later, about 14:00 two days on. The test is the onset lag measured over many flare-storm pairs: the 1859 storm began 17.6 hours after the flare, and storms routinely follow flares by one to four days. A small effect within minutes (the crochet) exists, but it is not the storm.
- For a self-gravitating gas ball in balance, 2K + U = 0 (K thermal energy). If the ball shrinks so that U becomes more negative by Delta, K rises by Delta/2; only the other Delta/2 can leave as radiation. So the energy available for sunshine is half of (3/5) G M^2 / R, and the clock halves (18.8 to 9.4 million years). Lane's law, "a gaseous sphere, losing heat, under its own gravity, must become hotter" [read: p.180], is the same statement.
- F_e / F_g = e^2 / (4 pi epsilon_0 G m_e^2) = (8.99 x 10^9 x (1.602 x 10^-19)^2) / (6.674 x 10^-11 x (9.11 x 10^-31)^2) = 2.31 x 10^-28 / 5.54 x 10^-71 = 4.2 x 10^42. A cloud of only negative corpuscles would fly apart; gravity could never gather it into nebulae. To collapse it needs an almost exactly neutralising positive background, and then there is (almost) no net charge for electrical induction to act on. For two solar-mass bodies, electric repulsion would equal gravity at a charge of only Q = M sqrt(4 pi epsilon_0 G) = 1.7 x 10^20 C, about one spare electron per 10^18 atoms, and an ionised gas cancels any such excess within a short distance [calc, Arbiter].
- A telescope gathers light over a larger aperture but spreads the image of an extended source over a proportionally larger area of the retina (magnification), so the brightness per unit area of a diffuse glow is at best what the naked eye sees (less after losses in the optics). A star is a point: its light stays concentrated in one spot however much it is magnified, so collecting more light makes it brighter. This is why faint diffuse features are best seen with low magnification or wide-field photography, which is how Barnard settled the dark-nebula question.