Primer: Eugene Mallove, "Tesla and the Aether" (Infinite Energy #48, 2003)

Step 4 study document. It adds no new verdicts; the verdicts are the arbiter's (arbiter.md). Source tags: [read] = from the article or a saved file in research/files; [web] = from an Internet source (URL given); [memory] = recalled from training, with confidence. The worked example is reproduced in example.py (run: .venv/bin/python review/Malove-Eugene/example.py).

Source note (repeated from the arbiter). The article itself (4 pages) was read in full. The folder is spelled "Malove"; the author is Eugene F. Mallove (1947-2004) [memory, high]. Almost every experimental statement in the article is second-hand: the Correas' in-house monographs (AS2-13, AS2-14, AS2-16, AS2-28; paid downloads, not obtained by any step), a lecture report (Joie P. Jones, Society for Scientific Exploration, June 2001), or Mallove's own observations given without data. The numbers behind the PAGD claim come from a Mallove letter of 14 June 2001 and the Correa patents, which the Investigator found. Where a conclusion below rests on [memory] alone, it says so.

Interested parties. Mallove edited Infinite Energy (IE below), ran Cold Fusion Technology, Inc., and was the Correas' editor, fund-raiser and friend. Paulo and Alexandra Correa hold the PAGD patents and run aetherometry.com, aetherenergy.com and eugenemallove.com, which supply almost all of the further material (letters, witness statements, the 2002 e-mails with SEAS (Space Energy Access Systems), the group that offered an independent test). CSICOP (the Committee for the Scientific Investigation of Claims of the Paranormal), publisher of Skeptical Inquirer, is an organisation devoted to criticising such claims. Each party's description is reported as theirs.

QE revision note (2026-10-04). This primer was updated after the quality check (qe.md) and the arbiter's "QE revision". Changes: a new claim 7 on cold fusion (sections 1, 2, 3.8, 4, 6, 7); the 50/500 W figure dated to the March 2001 visit; the Reich-Einstein source replaced; the Correas' published Reich-anomaly paper, Mallove's own Stirling-engine and Field Meter observations and his stated certainties restored; omitted witnesses and the Joie Jones leads restored; the patent's statement on loading sensitivity added; acronyms expanded. An earlier review step charged Mallove with misnumbering a Correa monograph; that charge was wrong (the article prints the numbers correctly) and nothing in this primer repeats it.

---

1. The idea in one paragraph

Mallove argues that twentieth-century physics took a wrong turn when it dropped the aether, the medium that nineteenth-century physicists thought filled space, and that Nikola Tesla's experiments with high-frequency coils were touching something real that physics then forgot. His evidence is the work of Paulo and Alexandra Correa of Toronto. Their patented vacuum tubes (Pulsed Abnormal Glow Discharge, PAGD) gave, by Mallove's own observation, roughly ten times more output than input. Their Tesla-coil experiments, they say, show a second, "massfree" kind of electricity that is not ordinary electromagnetic (EM) radiation, and Mallove proposes calling it "cold electricity". He adds that biology shows similar signs: an acupuncture study he cites reported that stimulating a point on the foot lit up the visual part of the brain faster than any nerve could carry a signal. He also states, as settled fact and without citing any experiment, that cold fusion is real: "hundreds of experiments" show "irrefutable evidence" of nuclear reactions at low energy (LENR, low-energy nuclear reactions), elements transmute "sans even direct electric energy input", ordinary (light) water electrolysis with nickel or platinum gives excess heat, and none of this will be understood until physics "return[s] to fundamental questions about electricity and the aether" [read: p.1-2]. For Mallove, cold fusion and the aether are one story. He closes by saying the establishment ridicules such work instead of testing it, and that the question must be settled in the laboratory. That last principle is the one this primer takes most seriously: almost everything in the article is a claim about a measurement, so the useful skill is learning what a decisive measurement would look like.

2. What you need to know first

Electric power and its average. Power is energy per unit time, measured in watts (W = J/s). For steady direct current (DC), P = V I (voltage times current). When voltage and current change quickly, the meaningful number is the time average of their product over one period T: P_avg = (1/T) ∫ v(t) i(t) dt. This is not, in general, the average voltage times the average current. Read: Griffiths, Introduction to Electrodynamics, the chapter on electrodynamics (Poynting's theorem); any circuits text, e.g. Horowitz and Hill, The Art of Electronics, chapter 1.

Mechanical power of a rotating shaft. A shaft turning at angular speed ω (radians per second) and carrying torque τ (newton metres, N m) delivers P = τ ω watts. Torque has to be measured, normally with a dynamometer (a calibrated brake). An electric motor turns electrical power into shaft power with efficiency η < 1, so its electrical input is P_shaft / η. Read: any first-year mechanics text, e.g. Halliday, Resnick and Walker, Fundamentals of Physics, rotation chapters.

Conservation of energy (first law of thermodynamics). Over any interval, the energy a system delivers equals the energy put in plus the decrease of the energy stored inside it. A device can deliver more power than it takes in for a while if it is draining a store (a battery, a charged capacitor, a spinning flywheel). Only a run long enough to exhaust any plausible store tells you whether there is a new source. Read: Schroeder, An Introduction to Thermal Physics, chapter 1.

Gas discharges and negative resistance. A gas at low pressure between two electrodes conducts only above a breakdown voltage. As current rises, the discharge passes through distinct regimes: a "dark" Townsend discharge, a "normal glow" (voltage roughly constant while the glowing area on the cathode grows), an "abnormal glow" (the whole cathode glows and voltage must rise to push more current), and finally an arc, where the voltage drops as current rises. A region where voltage falls as current rises is called negative differential resistance, and a circuit with such an element and a capacitor tends to oscillate on its own. Read: Raizer, Gas Discharge Physics; for the circuit idea, the neon-lamp relaxation oscillator in any electronics text. [memory, high]

Resonant circuits and near fields. An inductor L and capacitor C together resonate at f = 1/(2π√(LC)). A Tesla coil is two such circuits, loosely coupled, that build up very high voltage at a few hundred kilohertz. Close to any antenna (closer than about one wavelength divided by 2π) the fields are mostly stored, not radiated, and do not look like the transverse waves of the textbook far field. Read: Griffiths, chapter on electromagnetic radiation (dipole radiation); Purcell and Morin, Electricity and Magnetism.

The aether and special relativity. Nineteenth-century physics pictured light as a wave in a medium, the luminiferous aether. If Earth moves through it, light speed measured in the lab should depend on direction. Special relativity takes the measured speed of light as the same for all uniformly moving observers. Read: Taylor and Wheeler, Spacetime Physics, early chapters; any modern-physics text (e.g. Serway, Moses and Moyer, Modern Physics, chapter 1).

Nuclear fusion, the Coulomb barrier and electrolysis heat (for claim 7). Two nuclei, both positively charged, repel each other; to fuse they must get within about 10⁻¹⁵ m, which in a hot gas takes temperatures of millions of degrees. "Cold fusion" (1989, Fleischmann and Pons) is the claim that fusion of deuterium (heavy hydrogen) happens at room temperature inside a palladium metal electrode, shown chiefly by heat beyond what chemistry can supply. Measuring that heat in an electrolysis cell needs an energy balance: the power in is V I, and part of it leaves as chemical energy in the hydrogen and oxygen gas, unless the gases recombine inside the cell. Read: Krane, Introductory Nuclear Physics (fusion and barrier penetration); Atkins, Physical Chemistry (electrochemistry); and, for the controversy from both sides, Mallove's Fire from Ice (1991) and Huizenga's Cold Fusion: The Scientific Fiasco of the Century (1992) [memory, high for the books].

Speed, distance and time resolution. A speed is distance divided by delay, v = d/t. You can only claim a delay as short as your instrument can resolve, so a timing instrument sets an upper limit on the speed you can demonstrate.

3. The textbook view

3.1 Power accounting

Average power is P_avg = (1/T) ∫₀ᵀ v(t) i(t) dt, where v(t) is the instantaneous voltage (volts), i(t) the instantaneous current (amps) and T the averaging time (seconds). A useful identity, true for any two signals, is

mean(v i) = mean(v) · mean(i) + cov(v, i),

where cov(v, i) is the covariance: the average of (v − mean v)(i − mean i). If the voltage is steady (a stiff battery whose voltage barely moves when current pulses), cov = 0 and V × mean(i) is correct, provided the ammeter really reads the average. If voltage and current pulse together, the covariance can be most of the answer. Meters differ in what they report: an average-reading meter, a "true RMS" meter (root mean square, √mean(i²)), and an oscilloscope sampling at some finite rate give different numbers for the same pulse train. [memory, high]

3.2 Glow discharges and self-oscillation

The PAGD tube is a real type of device operating in a real regime. Mallove describes "high vacuum glass tubes, typically with aluminum cathodes and anodes, which self-oscillate in a particular negative resistance region of current and voltage" [read: p.3]. The patent abstract (US 5,502,354) describes "a cold cathode vacuum discharge tube ... generating pulsed autoelectronic emissions ... in the abnormal glow discharge region", at "much lower current densities than predicted by the Fowler-Nordheim vacuum arc discharge region law" [web: https://patents.google.com/patent/US5502354A/en; patent holders' characterization]. The textbook analogue is the relaxation oscillator: a capacitor charges through a resistor R from supply voltage V_s until the tube strikes at V_strike, the tube dumps the charge and goes out at V_ext, and the cycle repeats with period

T ≈ R C ln[(V_s − V_ext)/(V_s − V_strike)].

The neon-lamp version (the Pearson-Anson effect, 1922) is a standard teaching circuit [memory, high]. Such oscillators produce short, intense current pulses separated by long quiet intervals, which is exactly the kind of waveform where the averaging in 3.1 matters.

3.3 Tesla coils and the near field

A Tesla coil's secondary resonates at f = 1/(2π√(LC)), typically 10⁵ to 10⁶ Hz [memory, high]. At 200 kHz the wavelength is λ = c/f = 1,500 m, so λ/2π ≈ 240 m. Every distance in an ordinary laboratory is deep inside the near field. There, an oscillating charge distribution has an electric field component pointing along the line from the source (a "longitudinal" or radial component). For a small oscillating dipole of moment p (coulomb metres), the radial field near the source is approximately the static dipole field,

E_r ≈ 2 p cos θ / (4π ε₀ r³),

falling as 1/r³ (θ is the angle from the dipole axis, ε₀ = 8.85 × 10⁻¹² F/m the permittivity of free space). Only far away (r ≫ λ/2π) does the familiar transverse, 1/r radiation field dominate. So "longitudinal electric effects near a Tesla coil" are textbook physics, not a contradiction of it [memory, high]. High-voltage coils also produce corona (ionised air), ozone, sparks that emit broadband radio noise and sometimes X-rays from spark gaps; electroscopes, neon bulbs and Geiger-counter electronics all respond to these [memory, medium-high].

Tesla's lit bulbs at a distance have a mainstream explanation: resonant near-field coupling. A 2007 group at the Massachusetts Institute of Technology (MIT) transferred 60 W at about 40% efficiency over more than 2 m with strongly coupled magnetic resonators (Kurs et al., Science 317:83) [web: https://ui.adsabs.harvard.edu/abs/2007Sci...317...83K].

3.4 The Earth-ionosphere cavity

Tesla's "resonating cavity surrounding the body of the Earth" [read: p.3] is real. The space between the conducting ground and the conducting ionosphere is a spherical-shell resonator. For an ideal, lossless thin shell of radius R the resonant frequencies are

f_n = (c / 2πR) √(n(n+1)), n = 1, 2, 3, ...

With R = 6,371 km, c/2πR = 7.49 Hz and f₁ = 10.6 Hz; the observed fundamental (Schumann resonance) is about 7.8 Hz, lower because the ionosphere is a lossy conductor, and the quality factor Q is only about 4-6 [memory, high for the formula and 7.8 Hz; medium for Q]. (The arbiter's simpler estimate, c divided by Earth's circumference, also gives 7.5 Hz.) The cavity carries ordinary electromagnetic modes; its low Q means energy put into it is dissipated within a few cycles.

3.5 The aether and the experiments against it

Michelson and Morley (1887) looked for the change in light's round-trip time along two perpendicular arms as Earth moves at v ≈ 30 km/s around the Sun. The expected shift of interference fringes when the apparatus is rotated is

ΔN ≈ (2L/λ)(v²/c²),

with L the effective arm length and λ the wavelength. With L = 11 m and λ = 500 nm, ΔN ≈ (2 × 11 / 5 × 10⁻⁷)(10⁻⁸) ≈ 0.44 fringe; they saw less than about a twentieth of that [memory, high]. Later tests (Kennedy-Thorndike, Ives-Stilwell, rotating optical cavities) put far tighter limits on any preferred frame [memory, medium on the exact figures]. Lorentz's aether theory, in which rulers and clocks are affected by motion through an undetectable aether, makes the same predictions as special relativity; physicists dropped it mainly because it adds an entity nothing can detect [memory, high]. Einstein's 1920 Leiden address allowed that space is "endowed with physical qualities" [memory, high]. Modern physics does give the vacuum properties (zero-point fluctuations, fields), but in a form that looks the same to every uniformly moving observer and offers no extractable energy in its lowest state [memory, high].

3.6 Nerve speed and fMRI timing

The fastest myelinated nerve fibres conduct at up to about 100 m/s; a foot-to-brain signal therefore takes on the order of 10-40 ms [memory, high for speed; medium for the evoked-potential latency]. Functional magnetic resonance imaging (fMRI) measures the blood-oxygen-level-dependent (BOLD) signal, a change in blood oxygenation that follows neural activity by 1-2 s and peaks at about 5 s, sampled every 1-3 s [memory, high]. fMRI cannot time a microsecond event.

3.7 Galvani

In the 1780s Luigi Galvani saw dissected frog legs twitch when a spark machine fired across the room. The standard account is that the spark launches an electromagnetic transient that induces a current in the nerve-and-wire arrangement, which acts as a crude antenna [memory, high]. Galvani's "animal electricity" and Volta's battery were both later shown to be real: nerves and muscles do carry electrical signals (Matteucci, du Bois-Reymond) [memory, high].

3.8 Cold fusion and the electrolysis energy balance (claim 7)

In an electrolysis cell driven at voltage V and current I, the electrical power in is V I. Splitting water stores energy in the gas: per unit current, that energy corresponds to the thermoneutral voltage, about 1.48 V for ordinary water (a little higher for heavy water) [memory, high]. If all the hydrogen and oxygen leave the cell, the heat the cell should release is

P_heat,expected = I (V − 1.48 V).

"Excess heat" means measured heat above this. Two textbook points matter. First, if a fraction f of the gas recombines inside the cell (on the electrode, for example), it releases f × I × 1.48 V of extra heat that is chemical, not nuclear. The fraction of current that ends up as escaping gas is the Faraday efficiency. Second, nuclear fusion of deuterium in a hot plasma gives neutrons, tritium or gamma rays in known proportions; the standard mainstream objection to cold fusion is that the heat claimed would require far more of these than is seen [memory, high]. Cold-fusion researchers answer that a different, lattice-assisted reaction makes helium-4 without the usual radiation, and some report helium roughly in proportion to the heat [web: https://www.lenr-canr.org/acrobat/DOEreportofth.pdf, via the arbiter; LENR-CANR is a pro-LENR site]. The 2004 US Department of Energy (DOE) review panel split roughly evenly on whether the excess-heat evidence was compelling, and about two thirds were not convinced that nuclear reactions occur [web: same; as reported by the arbiter].

A quick number (reproduced in example.py, Part F): a light-water cell at V = 3.0 V and I = 1.0 A should release 1.0 × (3.0 − 1.48) = 1.52 W. If 10% of the gas recombines unnoticed, the cell releases an extra 0.1 × 1.0 × 1.48 = 0.148 W, which appears as about 10% "excess heat". Light-water excess heat was reported at roughly 3-50% (Mills and Kneizys 1991; Notoya and others c. 1992), and a recombination explanation was later proposed by groups that measured Faraday efficiency during the run (Shkedi and colleagues, 1991-93) [web: https://en.wikipedia.org/wiki/Faraday-efficiency_effect, via the arbiter; primary papers not read by this review]. The calculation shows only that recombination of a plausible size can produce an excess in the reported range; whether it explains every report is a question for the primary papers.

4. Where the inventor departs

The common thread: in each case the starting observation is real or plausible (discharge tubes do self-oscillate in pulses; Tesla coils do produce longitudinal-looking near-field effects at a distance; the Earth does resonate; Galvani's frog legs did twitch), and the departure is in the inference from that observation to a new kind of energy or electricity, made without the numbers that would rule out the ordinary explanation.

5. Worked example: what "50 W in, about 500 W out" means, and what run would decide it

This takes the central claim (claim 1) and asks three questions a laboratory physicist would ask before believing or rejecting it. All inputs from sources are tagged; everything else is a labelled assumption for illustration, not data about the Correa device. example.py reproduces every number.

Step 1. The claim, stated with its units

Mallove, 14 June 2001: "I observed an input DC power to the PAGD reactor of 50 watts, with an output motor power (mechanical shaft power) of approximately 500 watts." The letter gives this figure for "my last visit", in March 2001, "when your PAGD inverter technology had improved considerably from my first visit" (August 2000; described as preliminary in the Correas' narrative, per the QE spot check). The meters he names, "conventional electric meters, a digital storage oscilloscope, and a computer data acquisition system", are those of "the on-site visits" in general; "no one else from Infinite Energy was with me" [web: https://www.eugenemallove.com/Letters/June_14_2001.html; read: research/files/eugenemallove.com-June_14_2001.html]. The article's "2000-2002" covers his visits, not this one figure.

Ratio: 500 W / 50 W = 10. The units match (both watts), but the quantities do not: one is electrical power into a tube, the other mechanical power on a shaft. Between them sit the tube, an inverter, and a motor.

Step 2. Measuring pulsed power: why the instrument matters

Take an illustrative rectangular pulse train: voltage V₀ = 500 V and current I₀ = 10 A for a fraction D = 1% of each cycle, zero otherwise (illustrative values only).

• True average power: P = D V₀ I₀ = 0.01 × 500 × 10 = 50 W.
• Average voltage × average current: (D V₀)(D I₀) = 5 V × 0.1 A = 0.5 W, too low by 1/D = 100.
• Pulse voltage × RMS current: V₀ × I₀√D = 500 × 1 = 500 W, too high by 1/√D = 10.

Now a stiff 50 V DC supply delivering 100 A pulses at 1% duty: v is constant, so cov(v, i) = 0, and V × mean(i) = 50 × 1 = 50 W is exactly right; but V × I_rms = 50 × 10 = 500 W is ten times too high.

The lesson is not that anyone made this error; no waveforms were published, so nobody can say. The lesson is that with pulsed signals a factor of 10 lives entirely in the choice of meter, which is why "power ratio" has to be defined by the integral ∫ v i dt measured at enough bandwidth on both sides of the device. Mallove named an oscilloscope, which is the right tool; the record does not show the integration.

Step 3. Shaft power and the motor behind it

P_shaft = τ ω. At an assumed 1,800 revolutions per minute (rpm; not stated anywhere), ω = 2π × 1800 / 60 = 188.5 rad/s, so 500 W needs τ = 500 / 188.5 = 2.65 N m. Shaft power is directly proportional to the torque figure, so if torque was estimated rather than measured with a dynamometer, an error in the torque is the same error in the watts. The SEAS Z-Prize rules, as the Correas quote them, excluded torque-based claims under 10x because "torque to wattage calculations ... may be inaccurate by an order of magnitude" [web: https://www.eugenemallove.com/emwords_SEAS.html; SEAS's wording as quoted by the Correas]. How Mallove's 500 W was obtained is not stated.

A motor with efficiency η delivers P_shaft = η P_elec. For small motors at 70-90% [memory, high]: P_elec = 500 / 0.9 to 500 / 0.7 = 556 to 714 W of electrical power must enter the motor. A complete account has to show where that power crosses the system boundary: from the 50 W tube circuit, or from energy stored in the inverter, a flywheel or a battery. In November 2002 Mallove described the configuration as an "existing inverter/flyweel configuration with the low power tube" and the goal as to "close the loop to eliminate the battery" [read: research/files/eugenemallove.com-emwords_SEAS.html], so a battery was part of the system then.

Step 4. The closed loop: a test that is cheap if the claim is true

Put a generator on the shaft and feed its output back to the tube input. If the overall return path (generator plus electronics) is only 50% efficient, it returns 0.5 × 500 = 250 W to a device that needs 50 W: a 5x margin, with 200 W net to spare. At 70%, 350 W returned, 300 W net. Mallove himself wrote that the motor "could easily be made self-sustaining with a DC generator on the output shaft" [web: June 2001 letter]. No record shows it was done; in 2002 he called closing the loop "this mysterious psychological milestone that is the holy grail of the new energy field" [read: emwords_SEAS.html].

One caution from the inventors' own patent: "increasing the charge pack size of 29 cells to 31, by 7% leads to a 10-fold reduction in frequency; further increases in the number of charge pack cells extinguishes the phenomenon" [read: research/files/gp-US5449989A.html]. The effect is very sensitive to what the tube drives. A generator on the shaft is a different load, so a loop that fails would still need interpreting: did loading stop the oscillation, or was there no gain? A fair test would agree beforehand on how to tell the two apart (for example, by checking that the tube still oscillates at its open-loop rate with the generator connected).

Step 5. How long must the closed loop run?

Power can be borrowed from a store; energy over a long run cannot. A store holding energy E can supply power P for a time t = E / P.

• One 12 V, 6 Ah gel cell of the type in the patent tests holds 12 × 6 = 72 Wh = 259 kJ. At 500 W it lasts 72/500 h = 8.6 minutes.
• The patent's drive pack was 46 such cells [web: US 5,449,989]: 3,312 Wh nominal, 1,656-1,987 Wh at the 50-60% effective capacity the patent itself reports. At 500 W that lasts at most about 4 hours.
• A 24-hour closed-loop run with 200 W net surplus delivers 200 × 24 = 4,800 Wh = 17.3 MJ, the contents of 67 of those cells, about 137 kg of lead-acid battery at 35 Wh/kg [memory, medium on specific energy].

So an unattended, battery-free run of a day, with the box open for inspection, would be decisive either way: a 137 kg hidden battery is not a plausible artefact, and a run that stops in minutes or hours would show where the energy came from.

Step 6. The patent's own numbers (US 5,449,989)

• 0.962 kWh gained / 0.246 kWh lost = 3.91, matching the stated "surplus factor of 3.9x".
• 221.7 Wh gain / 48.3 Wh loss = 4.59, but the patent states 485.2% ± 18% (n = 64); its comparison run, 182.2 / 37.7 = 4.83, matches its "observed 483%". The 459% vs 485% gap is unexplained (a scale factor or a typo); not proof of error [read: research/files/gp-US5449989A.html; arithmetic].
• In the same 10-second sample, drive-pack input read 36.6 to 57.82 W and charge-pack output 146.4 to 399.6 W [read: same file]. The ratio could lie anywhere from 146.4/57.82 = 2.5 to 399.6/36.6 = 10.9 depending on how the swinging readings are paired and averaged; the midpoint ratio is 5.8. This is the regime of Step 2.
• The patent itself warns that estimating battery charge from open-circuit voltage can distort the figures "by up to 10 fold" and that the cells had "grossly diminished (50-60% of manufacturer's theoretical estimate) effective charge capacity" [read: same file]. The inventors state this openly. These 1992-95 battery-to-battery figures are a different configuration from Mallove's 2001 shaft-power figure, so they neither confirm nor contradict it.

What the example shows

Nothing here refutes the claim: two numbers cannot be refuted by calculation, and the arbiter's verdict is Never properly tested. What the example shows is that (a) the stated measurement could be wrong by the full claimed factor through ordinary instrument choices, and (b) the claim, if right, is easy and cheap to make decisive with a closed loop and a long run. Both facts point the same way: the measurement, not the argument, decides it.

6. Evidence table

Where the record goes dark (absence of evidence). After the 16 November 2002 meeting, PAGD disappears from independent view: no SEAS/Loder report, no closed loop, no account of the "Berlin video". In April 2002 Axelrad wrote that the Correas were closing their laboratory for lack of funds; Mallove wrote that they had spent about $1.2M of their own money [web: aetherenergy.com Axelrad letter; new_energy_FAQ.html]. Mallove died in 2004; nothing in this record ties his death to any claim. The Correas kept publishing in their own journal, including "The PAGD Ping-Pong System" (J. Aetherom. Res. 4(9), 2026, not read) [web: aetherometry.com/correa_publications.php]. The arbiter sets out competing explanations without choosing: trade secrecy (against: the patents disclose the design, and Mallove said they are specified precisely enough for others to "verify it independently"); funding failure (against: a closed loop is cheap, as Step 4 shows); distrust of state agencies (the Correas cite an earlier episode with the US Department of Energy); the device not surviving independent conditions (Mallove raised "concerns about tube lifetimes", and the patent itself says that adding two cells to a 29-cell charge pack cut the oscillation frequency tenfold and more cells "extinguishes the phenomenon"; this fits fragility but equally fits a finely tuned real effect, and the Correas' wish to test "at our premises"; inference from absence only); institutional suppression (ridicule is documented, refusal to test is not); absorption or classification (no evidence either way). The records that would decide: SEAS/Loder notes, the terms of the proposed NDA (non-disclosure agreement), the "Berlin video", Carrell's 1999 report, the 2026 PAGD paper and the patent file histories.

Related work Mallove grouped with PAGD (kept for the reader; no verdict given here): the Graneau water-arc and air-arc experiments (excess "approaching 24%"), Wallman/Dammann underwater carbon arcs, BlackLight Power (Mills), Harold Aspden's 1969 electrodynamic law, Ernesto Manuel's 1969 pulsed abnormal glow patent, and Reich's orgone motor [web: https://www.eugenemallove.com/Writings/lenr_excess_heat.html; Axelrad letter]. A US Senate field hearing (S. Hrg. 107-564, 30 May 2001, Durham NH) carries a Mallove statement and a briefing by Thomas Valone that cites PAGD: advocacy entered into a government record, not a test [web: https://www.govinfo.gov/content/pkg/CHRG-107shrg80649/html/CHRG-107shrg80649.htm].

7. What a learner should take away

Solid. - Gas discharges have a negative-resistance regime, and tubes operated there self-oscillate in sharp pulses. The PAGD tube is a real device in a real regime. - Near a Tesla coil the electric field has strong components along the line from the source; at a few hundred kilohertz the whole laboratory lies in the near field. Tesla's and the Correas' "longitudinal" observations have a real physical basis. - The Earth-ionosphere cavity does resonate (Schumann resonances), and short-range resonant wireless power works (Kurs et al. 2007). - Mallove's method principle, "these matters can be adjudicated in laboratory experiment" [read: p.4], is right, and it applies to everyone, inventors included.

Wrong as stated, and why the mistake is instructive. - The 80-microsecond / 1,000x pair is internally inconsistent: an 80 µs resolution over a 1-1.5 m path can only show about 125-190x nerve speed, and BOLD fMRI cannot time microseconds at all. Instructive because it is a units-and-resolution check anyone can do in one line. - The reference to the acupuncture paper is wrong (wrong authors and title for those pages), and the paper was later retracted by most of its authors, with Jones and two others dissenting. Instructive: always check the citation before the claim. - Treating "relativity dropped the aether by fashion" as the whole story leaves out a century of tighter preferred-frame tests. But the half that is right is worth knowing: Lorentz's aether makes the same predictions as relativity and was set aside largely by economy of explanation. - Reasoning by exclusion ("not EM, not ions, therefore a new radiation") is only as good as the list of what was excluded and the numbers behind each exclusion. - Calling LENR "irrefutable" overstates the record, including Mallove's own: in his 2003 conference paper he wrote that the nuclear-ash hypothesis "has not been proved". Instructive: the strength of a word should match the strength of the evidence. That cuts both ways, since the 2004 DOE panel's even split on excess heat means "dismissed" also overstates it. - Two small slips in the article, for the habit of checking: the Time cover story on Tesla is dated 20 July 1931, not "July 10, 1931" (his birthday) [web: archive.org Time PDF, via the arbiter]; and Mallove's quotations of Johnson are verbatim but some are stretched in context (Section 6, row 6).

Genuinely open. - Whether the PAGD/inverter/motor system gave a real 10x shaft-to-input ratio. It was never properly tested; the decisive test (Section 5, Steps 4-5) is cheap and has never been reported. - What the Correas' Tesla-coil monographs actually measured (distances, count rates, shielding outcomes), and whether a shielded repeat free of radio-frequency (RF) interference would see the same thing. - Whether any version of the So/Jones/Kidney/Saito work or Jones's 2001 lecture contains a real latency measurement. Jones's later writing (Journal of Scientific Exploration (JSE) 15(2):183-197; later pieces on "torsion") is an unread lead; critical blog commentary (Respectful Insolence / ScienceBlogs, 2012) on his later pranic-healing experiments also exists [web: research/2026-10-04-google.md, search results only, none opened]. Both belong in the record, each as its author's view. - Whether the Reich thermal anomaly survives a neutral, gradient-controlled repeat. The Correas published their measurements (IE #37, 2001) and say the anomaly is reproducible; Einstein's 1941 explanation was convection; Mallove's own measurements were promised but not found. Reading the IE #37 paper is the next step. - What Mallove's own Stirling-engine and Field Meter observations would show with published controls (Correa, Correa and Mallove, IE 53, 2004, not yet read). - For claim 7: whether heat and helium-4 measured in the same cold-fusion cell, with calibrated calorimetry and a Faraday-efficiency check, agree across independent laboratories (the test the 2004 reviewers pointed toward); and whether the light-water reports all fall to recombination. - Why the 2002 independent test did not happen; the SEAS-side record would decide it.

8. Check yourself

1. A device draws current from a 24 V battery in 50 µs pulses of 40 A, repeated every 2 ms. What is the true average input power? What would you get if you multiplied 24 V by a true-RMS ammeter reading?
2. A claimant says a motor delivers 300 W of shaft power at 1,500 rpm. What torque must a dynamometer read? If the motor is 80% efficient, how much electrical power must it receive?
3. Mallove reports that fMRI's "minimum time localization" is 80 µs and that the signal travelled at least 1,000 times nerve speed. Taking a 1.2 m path and 100 m/s nerve speed, what is the largest speed factor 80 µs can demonstrate? What delay would 1,000x require?
4. A Tesla coil runs at 300 kHz. How far from the coil does the near field extend (roughly λ/2π)? Is a detector 20 m away in the near or far field?
5. A closed-loop device is claimed to give 150 W net surplus. A sceptic suspects a hidden 12 V, 7 Ah battery. How long a run would deliver ten times that battery's energy?

---

Answers

1. Duty cycle D = 50 µs / 2 ms = 0.025. Mean current = 0.025 × 40 = 1 A; the battery voltage is steady, so P = 24 × 1 = 24 W. RMS current = 40 × √0.025 = 6.32 A; 24 × 6.32 = 152 W, too high by 1/√D = 6.3.
2. ω = 2π × 1500/60 = 157.1 rad/s; τ = 300/157.1 = 1.91 N m. Electrical input = 300/0.8 = 375 W.
3. Largest demonstrable speed = 1.2 m / 80 × 10⁻⁶ s = 15,000 m/s = 150x nerve speed. 1,000x (100,000 m/s) needs a delay of 1.2 / 10⁵ = 12 µs or less. The printed numbers do not support each other (and BOLD fMRI resolves seconds, not microseconds).
4. λ = 3 × 10⁸ / 3 × 10⁵ = 1,000 m; λ/2π ≈ 160 m. A detector at 20 m is well inside the near field, where non-transverse field components are expected.
5. Battery energy = 12 × 7 = 84 Wh; ten times is 840 Wh; at 150 W that takes 840/150 = 5.6 hours. A 24-hour unattended run would deliver 3,600 Wh, about 43 such batteries.

9. Further reading

Mainstream. - D. J. Griffiths, Introduction to Electrodynamics (Poynting's theorem; dipole radiation and near/far fields). - E. M. Purcell and D. J. Morin, Electricity and Magnetism. - Yu. P. Raizer, Gas Discharge Physics (glow, abnormal glow and arc regimes). - P. Horowitz and W. Hill, The Art of Electronics (measurement, RMS vs average, relaxation oscillators). - E. F. Taylor and J. A. Wheeler, Spacetime Physics (the aether and special relativity). - A. Kurs et al., "Wireless power transfer via strongly coupled magnetic resonances", Science 317:83 (2007). - Z. H. Cho et al., Proceedings of the National Academy of Sciences (PNAS) 95:2670-2673 (1998), and the retraction, PNAS 103(27):10527 (2006) [web: https://www.pnas.org/doi/10.1073/pnas.0602520103]. - US Department of Energy, Report of the Review of Low Energy Nuclear Reactions (2004), for the cold-fusion context Mallove invokes (reviewers split roughly evenly on the excess-power evidence) [web: https://www.lenr-canr.org/acrobat/DOEreportofth.pdf, a copy on a pro-LENR site]. - J. R. Huizenga, Cold Fusion: The Scientific Fiasco of the Century (1992), the critical account by the co-chair of the 1989 DOE panel [memory, high]. - On light-water excess heat and recombination: the Wikipedia article "Faraday-efficiency effect" as an entry point to the Shkedi et al. papers (a secondary source; read the primaries). - W. Reich, The Einstein Affair (Orgone Institute Press, 1953), the Reich-Einstein correspondence of 1941 [web: via the arbiter's search results].

The inventor's and the Correas' own works. - E. F. Mallove, "Tesla and the Aether", Infinite Energy #48 (2003) (this project's document). - E. F. Mallove, letters of 14 June 2001 and 20 February 2002, and "LENR and 'Cold Fusion' Excess Heat" (ICCF-10, 2003), at eugenemallove.com (a Correa-run site) [web; saved in research/files]. - E. F. Mallove, Fire from Ice: Searching for the Truth Behind the Cold Fusion Furor (1991) [memory, high]. - P. N. Correa and A. N. Correa, "The Reproducible Thermal Anomaly of the Reich-Einstein Experiment Under Limit Conditions", IE #37 (May/June 2001); "A Modified Orgone Accumulator (HYBORAC) as a Drive for Low Delta-T Stirling Engine", IE #41 (2002); P. N. Correa, A. N. Correa and E. F. Mallove, "Power performance of Stirling motors driven from modified Orgone Accumulators - Part III", IE 53:9 (2004) [read: Mallove letters; web: aetherometry.com/correa_publications.php; none read by this review]. - Further PAGD reports listed by the Correas: Bahman, NET-Journal (1998, German); Minchrowski, PACE 11 (2000); Sapogin, IE 20:48 (1998); King, "The Supertube" (1996) [web: aetherometry.com reference list; not read]. - J. P. Jones, Journal of Scientific Exploration 15(2):183-197, and his later "torsion" pieces (supporting lead, not read); critical commentary on his later work, Respectful Insolence / ScienceBlogs (2012) (not read) [web: search results]. - P. N. Correa and A. N. Correa, US patents 5,416,391; 5,449,989; 5,502,354 [web: patents.google.com]. - P. and A. Correa, ABRI (Aurora Biophysics Research Institute [memory, medium]) monographs AS2-13, AS2-14, AS2-16 (Tesla coils, 1999) and AS2-28 (biological energies), collected in Experimental Aetherometry Vol. IIA-IIB (2003); the article's references print them as "ABRI Monograph S2-13", "S2-14" and "S2-16", correctly [read: p.4] [web: https://www.aetherometry.com/correa_publications.php; not read by this review]. - M. J. Seifer, Wizard: The Life and Times of Nikola Tesla (1996), Mallove's ref. 1. - J. Johnson, "The Strange Legacy of Nikola Tesla", Skeptical Inquirer, Summer 1994 (the CSICOP piece Mallove rebuts) [web: Center for Inquiry (CFI) PDF above; saved in research/files].