Vassilatos-Gerry: Mathematician review (Step 2)

Reviewer: PhD mathematician (checks derivations, notation, units, hidden assumptions). Date: 2026-10-08. Work reviewed: Gerry Vassilatos, Secrets of Cold War Technology: Project HAARP and Beyond (Borderland Sciences 1996; 162 PDF pages, two book pages per PDF page).

Tags: [read] = the book or a project file; [web] = Internet source with URL; [memory, high/medium/low] = recalled, with confidence; [calc] = my own arithmetic on figures tagged where they come from (scratch scripts: [scratch file] and c3.py, run with the project venv).

Source note. The book is a popular, secondary, interested account. It has almost no equations: the mathematics to check is a small set of numbers and scaling arguments scattered through the prose. Where the book gives no equation, I say so rather than invent one. Every verdict is therefore about "what the book claims", checked against its own numbers and against standard physics.

What I read myself, beyond the project files [read]: all of deep/ (three deep reads: pp.36-46 Tesla/Colorado Springs/Savo; Ch.7 first half pp.121-129; Ch.7 second half pp.130-138); facts.json; claims.md; sites.md "Leads for review" and "Corrections"; digest 12; and, in the extracted text, the passages on the 10,000 V per inch rule (book pp.35-37), Argus and Jim Creek/Cutler (book pp.177-178 and Chapter 6), Gamma Ray Bombs (pp.185-187), the Seafarer standing-wave passage (PDF pp.119-120, book pp.239-240), "Clear Atomic Light" (pp.182-183) and the gamma-ray communications conclusions (PDF pp.150-151). research/files/ is empty; the Investigator's evidence is in its notes and in the Leads, which I used.

---

0. Reading the verdicts

Verdict scale is the one in prompt.txt. Where a claim has a strong and a weak form, each gets its own verdict. The seven claims are those in claims.md.

---

1. Claim 1: "Radiant Energy" (book pp.9-35)

(a) Is the mathematics internally consistent? There is almost none. The book offers three quantitative statements. 1. "Voltages could often be increased at an amazing 10,000 volts per inch of axial coil surface. This meant that a 24 inch coil could absorb radiant shockwaves which initially measured 10,000 volts, with a subsequent maximum rise to 240,000 volts" [read: book pp.35-37, extracted text]. The arithmetic (24 in x 10 kV/in = 240 kV) is right. It is a rule of thumb, not a law, and the book calls it "a new induction law". 2. "Impulses of 100 microseconds or less" cannot be felt and do no harm [read: book pp.~30-31, extracted text]; trains above 0.1 ms cause "pain and mechanical pressures"; 1 microsecond gives "strong physiological heat". 3. Tesla measured "a zero current condition in these long copper secondary coils... Pure voltage was rising with each inch" [read: same pages].

Checks. - No "new induction law" is needed for (1). Two tuned circuits with coupling coefficient k (primary L1, C1; secondary L2, C2, with L1C1 = L2C2) are a pair of coupled ordinary differential equations. Energy conservation at full transfer gives (1/2)C1V1^2 = (1/2)C2V2^2, hence V2/V1 = sqrt(C1/C2) = sqrt(L2/L1). I integrated the four-variable system numerically (Runge-Kutta, L1 = 50 uH, C1 = 0.2 uF, k = 0.2, L2/L1 = 400): the secondary reaches 19.80x the primary voltage after 2.2 cycles and 19.998x later, against the predicted sqrt(400) = 20, with total energy conserved to 4 parts in 10^9 [calc: c3.py]. The step-up is a consequence of impedance ratio and resonance, not a new law. "Zero current" with rising voltage is the open top of a quarter-wave standing-wave resonator: current has a node (zero) at the free end and voltage an antinode. - The book's own rule is a strong internal check on its 100 MV claim (claim 2): at 10 kV per inch, the 10 ft (120 in) "extra coil" gives 1.2 MV, and even the whole 200 ft mast (2,400 in) gives 24 MV. Reaching 100 MV needs 10,000 inches = 833 ft of coil surface [calc]. - The 100 microsecond pain threshold matches the real strength-duration law of nerve stimulation (Lapicque): threshold current I(t) = I_rh (1 + tau_c/t), where I_rh is the rheobase (long-pulse threshold) and tau_c is the chronaxie, roughly 0.1-0.7 ms for motor nerves [memory, medium-high]. With tau_c = 300 us the threshold is 2x rheobase at 300 us, 4x at 100 us, 49x at 6.2 us, 301x at 1 us [calc]. So the book's qualitative statement is right, and Tesla's reading (very short pulses are not felt) is established physiology. "Do no physiological harm" is not: high-energy short pulses burn (heating at about 1 us is itself the book's own observation). - Shield penetration. The central claim is that the stream passes "glass and copper shields". The skin depth of copper (sigma = 5.8e7 S/m) at f is delta = 1/sqrt(pi f mu0 sigma). At 150 kHz (the book's own pulse rate, and about the bandwidth of a 6.2 us pulse), delta = 0.17 mm, so a continuous 1 mm copper wall gives e^-5.9 = 0.0027 (about -51 dB) from absorption alone, before reflection [calc]. A pulse of 1 us has more bandwidth and is attenuated more. So a sealed copper enclosure does stop such pulses. Observed "through the shield" effects in 19th-century experiments (Henry 1842 magnetising needles) belong to slow, low-frequency magnetic fields that copper screens poorly, and to unshielded leads and capacitive and ground coupling. The book cites no controlled shield test.

(b) Consistent with experiment? The conventional account (Maxwell's equations, skin effect, coupled resonators) reproduces everything quantitative in the chapter. The neutral ether gas has no operational definition other than "whatever passes the shield", so no experiment could distinguish it from a leakage path. The Michelson-Morley null result concerns ether drift and does not bear directly on a dynamical ether; I do not count it against the claim. [memory, high].

(c) Tested? Tesla's own published statements do not match the book's. The Investigator found that in US 685,957 and 685,958 (1901) "radiant energy" is a stream of "minute particles of matter positively electrified" thrown off by the Sun, charging an elevated plate against ground [web: https://patents.google.com/patent/US685957A/en]. That is a charged-particle hypothesis, the opposite of "neutral". Eric Dollard's reported reproductions and a "faster than light" line-interferometry result are cited by the book without any measurement given; I found nothing in the files that tests them.

(d) What would change? Electromagnetic shielding theory (Maxwell plus Ohm) would need an extra non-electromagnetic channel for energy and information; the null result for any such channel in shielded-room physics (screened rooms, MRI cages) would need explaining.

Investigator material kept. Tesla's patents describe particles, not a neutral gas (the book's reading and Tesla's own words diverge). Tesla's own 1900 definition of a "stationary wave" ("a wave with fixed nodal and ventral regions") [web: https://en.wikisource.org/wiki/The_Problem_of_Increasing_Human_Energy, via the deep read] is a standing wave, so the book's "stationary" versus "standing" distinction (pp.41-43 of the PDF, book pp.83-86) is not Tesla's.

Verdict. Strong form: Not testable as stated. Shield-penetration at the pulse widths named: Refuted by calculation. Weak form (pulse duration matters): Established.

---

2. Claim 2: Colorado Springs "ether engine" (book pp.72-93)

This is where the book gives most numbers, so most of my arithmetic is here. Symbols: V voltage, I current, f frequency, C capacitance, r radius, T period, Q charge.

2.1 The figures, checked against each other

2.2 Source of the figures (the Investigator corrected Step 1)

• 100 MV and 10,000 hp are Tesla's 1904 proposal for distribution ("I propose to distribute ten thousand horse-power under a tension of one hundred million volts") [web: https://teslauniverse.com/nikola-tesla/articles/transmission-electric-energy-without-wires, via the deep read]. In 1917 Tesla said "The highest potential I reached was something like 20,000,000 volts" and "I obtained in my antennae currents of from 1,000 to 1,100 amperes. That was in 1899" [web: http://www.tfcbooks.com/tesla/1917-05-08.htm, via the Investigator]. So the 1100 A is Tesla's own, and the book's 100 MV is not what he measured. My calculation shows the two Tesla numbers (20 MV and about 1100 A) are mutually consistent for a top-load of about 58 pF, which is what a ball of 0.5 m radius provides. That is a clean independent check that supports Tesla's own figures and removes the need for any "ether flow" to explain the current.
• 150,000/s and 0.0062 ms: no source anywhere in the files. The Colorado Springs Notes were not obtained.

2.3 "Disconnectable from the power supply"

The book says the transmitter could be "disengaged from the electrical power source after only a few seconds", the proof "encrypted and preserved in his cunningly captioned photographs" [read: PDF p.42]. The only quantitative principle on the table is Tesla's own 1917 statement that with no receivers "the plant consumes only a few horsepower necessary to maintain the electric vibration" [web: tfcbooks, via the Investigator]. That is the physics of a high-Q resonator: stored energy circulates, and the supply replaces only losses. If the quality factor is Q, the stored energy decays with time constant Q/w; for Q = 100 and f = 150 kHz that is 106 us, and for Q = 1000 about 1 ms [calc]. Persistence for seconds, or for the "twenty minute or more aether avalanche" the book infers from long photographic exposures (PDF p.44), cannot come from a resonator. The simpler reading of multi-minute exposures is that the station kept running. Stored terminal energy (1/2)CV^2 with C = 58 pF and V = 20 MV is about 12 kJ [calc], enough for roughly 10 W for 20 minutes, below any visible streamer display's demand. The book gives no supply power, so no full energy balance can be written; that is a missing number, not a verified result.

2.4 200 lamps (50 W each) at 26 miles

10 kW received at 42 km. Origin: O'Neill, Prodigal Genius (1944), assembled "by piecing together the fragmentary material published in a number of publications"; a Tesla list post calls it "generally accepted as being apocryphal" [web: https://pupman.com/listarchives/2007/Oct/msg00342.html, via the Investigator]. A radiation test of the claim: a small matched receiving antenna cannot have effective area above about 0.12 lambda^2 regardless of its physical size (lambda is wavelength). At 150 kHz lambda = 2 km, so the maximum effective area is about 4.8e5 m^2. Receiving 10 kW from a hemispherical wave at distance r = 42 km, with area 2 pi r^2 = 1.1e10 m^2, needs a radiated power of 10 kW x 1.1e10 / 4.8e5 = 2.3e8 W (230 MW) [calc]. That is far beyond any supply a station of that time had [memory, medium; the book gives none]. A conduction (non-radiating) mechanism is the only route open, and none is quantified in the book or in the files.

2.5 The Savo car (PDF pp.45-46)

Arithmetic [calc]: 80 hp = 59.7 kW; from the 12 V battery that would be about 5,000 A, so the book requires the power to come from the receiver. The Investigator found the source: Ahlers's summary of a 16 September 1967 interview of Peter Savo, held at tfcbooks (an "unidentified document circulated in the early 1980s") [web: http://www.tfcbooks.com/special/savo_account.htm]. It contains both receiver descriptions ("set into the dash"; "built in his hotel room, and carried it to Buffalo"), so the deep read's "editing seam" reading is withdrawn (see Corrections). If the energy came by broadcast the book rejects, a 6 ft rod matched at 1 MHz (lambda = 300 m, effective area at most 1.1e4 m^2) would need a power density of about 5.4 W/m^2 (about 45 V/m rms) to deliver 60 kW, and over a 50-mile route that is about 7 GW from a transmitter 10 km away [calc: E_rms = sqrt(30 P)/r]. For a self-contained receiver there is no equation to test. Two kinds of tube evidence the Investigator found: the "70-L-7" tubes Savo named date to about 1939 (Smithsonian Ken-Rad sample) or "first introduced in 1947" (Valve Museum) [web: https://americanhistory.si.edu/collections/object/nmah_702597 ; https://www.r-type.org/exhib/aaa1317.htm], later than 1931. Savo's identity is not established in the standard genealogies [web: https://treemily.com/blog/nikola-tesla-family-tree]; the Open Tesla Research site says no such nephew existed (its characterisation) [web: teslaresearch.jimdofree.com]. I keep this as one witness's account, interviewed 36 years after the event by someone who noted Savo "lacks a formal education".

(b) Consistent with experiment? Tesla's resonant transformers behave as the circuit theory predicts (claim 1). Long sparks (tens of metres) at multi-megavolt potentials were produced and photographed [memory, high]. No measurement of net energy gain, and no external power-gain accounting, exists in the files.

(c) Tested? The Colorado Springs station was a private experiment, demolished in 1900-1905 [memory, high]. A modern partial recreation, Toby Grotz's "Project Tesla" (1980s; the book's "Grotz"), exists [web: Google snippets via the Investigator; pages not opened]. The book itself records Grotz's opposing view that Tesla "merely stored energy in the earth" and calls it "a basic error". Neither the 26-mile lamp test nor the car has any replication in the files.

(d) What would change? Energy conservation would need an unmetered ambient source reachable at kilovolt-scale cost. Standard electrodynamics would need a non-electromagnetic energy carrier that enters through a metre-scale terminal and leaves as electrons.

Verdict (split). 100 MV at the terminal: Refuted by calculation and Internally inconsistent (book's own rule). 1100 A: Established as Tesla's figure, and explained by w C V. Self-sustaining after disconnection: Refuted by calculation for a resonator; otherwise Not testable as stated (no supply power). 200 lamps at 26 mi: Never properly tested (apocryphal reconstruction); by radiation it fails the energy budget by about 10^3. Savo car: Never properly tested (no object, one witness, tube anachronism, no energy accounting possible).

---

3. Claim 3: Marconi, the "Tesla Rule", and the "Houses" (book pp.47-153)

(a) Internal consistency. This claim is a historical and political argument, not a calculation. The logical structure can still be checked. The "Tesla Rule" says every post-war military project is "a very obvious and frail attempt at duplicating each item of Tesla Radiant Energy Technology" [read: PDF p.121, book p.242]; "Tesla is the rule. Tesla is the pattern" (p.122). Applied to HAARP the same book then says: "HAARP has nothing at all to do with Tesla Impulse Technology. HAARP employs technology which Tesla continually berated! HAARP is the very antithesis of systems developed by Tesla" [read: PDF p.127, book p.255]; yet "HAARP seeks the evocation of forces which Tesla successfully secured one hundred years ago... they will never reach their goal" (p.122), and "All of these schemes seek duplication of the Teslian aetheric methods" (p.135). So HAARP is at once a failed copy and the antithesis. A rule that classifies a similar project as a copy, a dissimilar one as a failed copy, and a contrary one as an "antithesis that only uses the name" cannot be falsified by any project. That is a logical defect (unfalsifiability), independent of the physics.

(b) Experiment/record. Marconi's priority was litigated. In Marconi Wireless Telegraph Co. v. United States (21 June 1943) the Supreme Court held broad claims of Marconi's U.S. Patent 763,772 invalid as anticipated by John Stone Stone; the ruling did not invalidate Marconi's earlier transmission patent [web: https://en.wikipedia.org/wiki/Guglielmo_Marconi, citing 320 U.S. 1]. That is real evidence that Marconi's patent position was not secure, but it names Stone, not Tesla, in the source I checked; whether the opinion also relied on Tesla and Lodge I did not check [memory, medium]. The book's claim that the 1901 signal "could not have crossed by Hertzian waves" (Ch.2) says "calculations have proven" a "Teslian component" without showing one (facts.json), so I can test nothing.

(c) Tested? Not applicable to the oligarchy thesis: no named entities, dates or records are given for "Houses" that could be checked.

(d) What would change? Nothing in physics; this is historiography.

Verdict. Tesla Rule and oligarchy: Not testable as stated. Application to HAARP: Internally inconsistent. Marconi priority: contested in the court record (Stone, 1943); I leave the historical weighing to the physicist and reader.

---

4. Claim 4: Cold War radiation technology "documented in patents" (book pp.70-153)

A patent records a description and intended operation, not performance. The claim bundles several separate assertions; I take them in turn.

4.1 Nuclear EMP. Real and well recorded: Starfish Prime (9 July 1962), a 400 km burst, damaged equipment in Hawaii about 1450 km away (about 300 streetlights, alarms, a microwave link) [web: https://en.wikipedia.org/wiki/Starfish_Prime]. Three errors in the book's chapter: (i) Argus I was "detonated at a height of 12,900 miles" [read: book pp.177-178, PDF p.89]. Argus was detonated about 300 miles up (the New York Times headline of 19 March 1959 reads "3 Atomic Devices Detonated 300 Miles Up") [web: https://en.wikipedia.org/wiki/Operation_Argus]. 12,900 miles is 43 times that and more than the Earth's circumference (about 24,900 mi) [calc]. (ii) "completely destroyed powergrid integrity in several cities" is not recorded for Argus (a 1.7 kt weapon over the South Atlantic [web: same page]); the documented grid and streetlight damage is Starfish. (iii) A 12,900-mile altitude would put the burst outside the inner belt altogether, contradicting the sentence's "well within the newly discovered Van Allen Belts". Verdict: Established (for EMP as a weapon effect); the book's Argus numbers are errors.

4.2 "Clear Atomic Light" from buried explosions (book pp.182-183). The book says yields "as small as 1 or 2 Tons" (about 4-8 GJ), reflected by "simple plastic sheets, or polished metals" into "cruciform beams". The companion volume names a "De Ment" patent [web: archive.org text, via the Investigator]. A few comments, none a refutation. X-rays reflect efficiently only at grazing incidence (critical angle of order a degree or less for keV photons), so "plastic sheets" as mirrors for the X-ray part is doubtful [memory, medium]. A nuclear-explosion-pumped X-ray laser (Excalibur) is a real Livermore concept, so the weak form (a nuclear explosion as an intense one-shot radiation source) is plausible [memory, high]. No performance data, test log or open patent drawing in the files shows the book's beams. Verdict: strong form Never properly tested (no open record); weak form Plausible, untested.

4.3 Mossbauer "gamma-ray bombs" (book pp.185-187). The text says that sound energy applied to Mossbauer isotopes "forces" gamma emission, that one could "pump such isotopes until all their radioactive emissions had been wasted... the mutability of radioactive halflife was demonstrated, where the entire sum of gamma ray emissions could be theoretically released in a single massive pulse", and that a "small Gamma Ray Bomb would effectively carbonize every living thing within a specified 'kill' radius", with "not one long lingering radioactive residue". - The Mossbauer effect is recoil-free gamma emission and absorption in a crystal; acoustic motion of the source shifts or sidebands the line [memory, high]. It changes the line shape, not the decay rate. The excited state that emits the 14.4 keV line in iron-57 lives about 98 ns [memory, high], so there is no stored reservoir of excited nuclei to dump. - The only stored gamma energy is the nuclear energy of the sample's atoms that have not yet decayed. For 1 kg of cobalt-57 (half-life 272 days, which feeds the iron-57 emission): N = 1.06e25 atoms, about 0.136 MeV per decay, so the total available is 2.3e11 J (about 55 tons of TNT equivalent); its activity is 3.1e17 Bq (8.4 megacuries) and it radiates 6.8 kW continuously [calc]. To "carbonize" a city of 100,000 people (heating and vaporising about 4e6 J per kg of 70 kg bodies, 2.8e13 J, about 6.7 kt TNT equivalent [calc; the 4e6 J/kg is my round figure for heating to carbonisation plus vaporising water]) you would need about 120 kg of cobalt-57, an activity near 1,000 megacuries that would decay in months, and that is before attenuation. Nothing of the kind has ever existed. - Half-lives are essentially constant. The known exception is bound-state beta decay in fully ionised atoms: bare rhenium-187 has a half-life of 32.9 years against 41.6 billion years for the neutral atom [web: https://en.wikipedia.org/wiki/Isotopes_of_rhenium; Bosch et al., Phys. Rev. Lett. 77, 5190 (1996)]. That requires stripping all electrons, not sound. Small (percent-level) chemical and pressure shifts occur for electron-capture isotopes such as beryllium-7 [memory, high]. - The book reads "Mossbauer isotopes" and "phase modulation" correctly as laboratory physics, then builds a bomb on the unsupported step that the whole emission can be forced out at once. Verdict: Refuted by calculation (energy available and decay-rate constancy).

4.4 Radioactivity "neutralisation" (Reich, McCullough). Decay-rate constancy under chemical, thermal and pressure changes is tested to well below the percent level [memory, medium; the literature is large and I did not re-survey it]. The orgone/ORANUR reports are the claim's only source in the book; no controlled replication is in the files. Verdict: generic decay-rate mutability: Tested, not supported; orgone-specific: Never properly tested.

4.5 Gamma-ray communications "cannot be jammed", reach submarines "at any depth", 30 GW coherent emitters (book PDF pp.150-151, quoted above). The book says "Passing through several thousand feet of seawater, Gamma Rays suffer neither dispersions nor diminutions of strength." - Attenuation is exponential: transmitted fraction = exp(-(mu/rho) rho x), where mu/rho is the mass attenuation coefficient. For 1 MeV photons in water, mu/rho = 0.0707 cm^2/g [memory, high, standard NIST value]. Through 1 m of seawater (rho = 1.025 g/cm^3) the fraction is 7.1e-4; through 10 m, 3e-32; through 305 m (1,000 ft), exp(-2,209) = 10^-959 [calc]. Even at the photon-attenuation minimum (about 20-50 MeV, mu/rho about 0.017 cm^2/g [memory, medium]), 305 m gives exp(-531) = 10^-231 [calc]. Not "neither diminution". - Air: the atmospheric column is 1,033 g/cm^2, so a 1 MeV photon from orbit is attenuated by exp(-0.0636 x 1,033) = exp(-66) and at the minimum exp(-18) [calc, mu/rho for air from memory, medium]. "Gamma-ray beams from orbit" do not reach the surface. - Collimation without divergence: diffraction does not forbid narrow gamma beams (wavelength about 10^-12 m); the missing element is a source. The "gamma-ray laser" has not been built [memory, high]; the 30 GW figure is a patent assertion. The patents the book lists (Eerkens 3,430,046; Pieckenbrock 3,557,370; and others) are real bibliographic entries [read: book bibliography]; I did not independently confirm each number. - For contrast, what actually works for submarines is ELF/VLF: the skin depth in seawater (sigma about 4 S/m) is about 503 sqrt(rho/f) m with rho = 1/sigma, which is 29 m at 76 Hz [calc: 503 x sqrt(0.25/76)] so 100 m depth is an attenuation of e^-3.5. Verdict: Refuted by calculation.

(b-d) Consistent with experiment; tested; what would change. Tested and recorded: nuclear EMP (Starfish, Argus; Soviet and U.S. tests) [web/memory as above]. Not on any public record: any gamma bomb, CAL beam or gamma communications. To make the claims true, nuclear physics would need adjustable decay rates and coherent gamma amplification at kilowatt-to-gigawatt scale, and photon attenuation tables would be wrong by hundreds of orders of magnitude.

Investigator material kept. The companion volume (Vassilatos 1997, 666 pp of patent texts) sources CAL to a "De Ment" patent and says Farnsworth produced "contained nuclear fusion reactions" through "cathode optics" [web: https://archive.org/details/gerry-vassilatos-declassified-patents-of-the-cold-war-and-sdi-1997-borderland-sc]; it was not read here beyond the Investigator's note.

---

5. Claim 5: HAARP as an industry-owned auroral power test (book pp.242-276)

5.1 Numbers that check out

• 300,000 A x 200,000 V = 6.0e10 W = 60 GW [calc, agrees with deep read].
• Power chain of the prototype: 3 shelters x 6 pairs x 10 kW = 360 kW; 3 x 300 kW generators = 900 kW (40% efficiency) [read: PDF p.129]. The deep read could not decide whether "double crossed dipoles" meant 36 transmitters served 36 or 72 dipoles. The arithmetic resolves it: 18 energised elements x 2 orthogonal dipoles (one 10 kW transmitter each) = 36 transmitters = 360 kW, i.e. 20 kW per element. The same 20 kW per element gives 48 x 20 = 960 kW and 180 x 20 = 3.6 MW, which are exactly the official figures for 1999 and 2007 [web: https://haarp.gi.alaska.edu/faq, via the Investigator]. So the book's prototype numbers are correct and the earlier ambiguity is resolved: one element = two 10 kW transmitters (see Corrections).
• Arecibo's 18 acres: pi x (500 ft)^2 = 18.0 acres [calc].

5.2 Numbers that fail

1. The scaling argument (book p.259 area). The book says the 30 m HAARP wave versus Arecibo's 0.03 m radar wave means "The IRI has to be 1000 times the size of Arecibo... this is precisely why the IRI requires the intended 1 Gigawatt... performs the same ionospheric work as 1.6 Megawatts". The ratio 1000 is the wavelength ratio, used linearly. For an aperture of area A radiating power P at wavelength lambda, the on-axis power density at distance R is S = P A/(lambda^2 R^2) (gain G = 4 pi A/lambda^2). So at equal area and range the density scales as 1/lambda^2, a factor 10^6, not 1000. To match the flux of a 1.6 MW radar of 18 acres with a 33-acre array at 30 m, the power ratio is (18/33) x (30/0.03)^2 = 5.5e5, i.e. 870 GW rather than 1 GW [calc]. So the book's own premise gives the opposite of its conclusion by a factor of nearly 10^3. In addition, "ionospheric work" depends on power absorbed, not flux; 0.03 m (10 GHz) waves pass through the ionosphere almost unabsorbed, while 3-10 MHz waves near the plasma frequency are absorbed [memory, high]. The comparison is not meaningful as stated.
2. The "1 Gigawatt" is an effective radiated power (transmitter power times antenna gain). My estimate for the prototype: area 18 x (80 ft)^2 = 1.07e4 m^2, G = 4 pi A/lambda^2 = 149 (21.7 dBi) at 30 m, giving an effective radiated power of about 54 MW; for the full array, 33 acres = 1.34e5 m^2, G = 1,865, and 3.6 MW x 1,865 = 6.7 GW [calc]. This matches the order of "about 4 GW" cited for the final array [memory, medium]. The book writes "Gigawatt of HF energy" (power, not energy) throughout.
3. The HIPAS "250 MW" figure: deep read gave two readings; with the corrected ratios the sentence cannot be both "one-quarter the rated output of HAARP" and consistent with the 360 kW the book gives for the prototype, unless the figures are effective radiated powers. As stated, power types are mixed (transmitter power versus effective radiated power).
4. Long-delay echoes: "A signal path of 12 light-seconds is a path exceeding 84 earth circumferences" [read: PDF p.132]: 12 s x 299,792 km/s = 3.60e6 km = 89.8 circumferences (84 corresponds to 11.2 s), so "exceeding 84" is true but loose. "Requiring a reflective layer averaging some 6 times further away than the moon": a 12 s round trip puts a reflector at 4.7 lunar distances; the whole path is 9.4 lunar distances. The deep read flagged this as a mismatch. A mean delay of 7.7 s gives a path of exactly 6.0 lunar distances, so the author appears to have quoted the path length of an average echo as if it were the reflector distance (twice too large). The error is of the "path versus distance" kind, not a number pulled from nowhere (a small refinement of the deep read; see Corrections).
5. Circular polarisation: "Cross dipoles conduct HF radio energy with a phase difference of 180 degrees. The resultant radiant field rotates like a screw" [read: PDF p.135]. Two orthogonal dipoles fed 180 degrees apart radiate a linearly polarised wave along a diagonal; circular polarisation needs 90 degrees [memory, high; standard antenna theory]. Agrees with the deep read.
6. Site area: 500 ft x 800 ft = 9.2 acres, not "very roughly 12 acres"; the 12-acre screen then exceeds the site that contains it [calc]. Gakona is in south-central Alaska, not on the "Alaskan North Slope" [memory, high].
7. The Investigator reported that the prototype's element count conflicts (48 in the book versus 18 in the UAF history). It does not: the book says 48 towers stand and "only 18 are ever actually energized", with 360 kW (PDF p.129), exactly the UAF "18 antenna elements with a net radiated power of 360 kW" [read: PDF p.128-129; web: https://haarp.gi.alaska.edu/faq]. The "48 x 6 by 8 grid" matches the later 48-element, 960 kW stage. The author's quoted 1996 sentence that testing began in December 1994 with 48 elements is the one place he or his source runs the stages together; the discrepancy is in the stage description, not in the book's hardware figures.

5.3 The 60 GW claim, as a worked example

The claim is that pumping the electrojet draws "usable power" to a ground screen: "Sixty Gigawatts of natural, non-polluting energy at no cost for those who own the system" [read: PDF p.135]. P = IV gives the power delivered to a load only when the full current I flows through a voltage drop V across that load. Three checks follow [calc, all using Investigator-supplied or memory-tagged geophysical numbers]. 1. Aperture bound. The auroral oval is about a 5-degree-wide ring at about 20 degrees colatitude, area about 7.6e12 m^2 [calc; memory, medium for width and location]. 60 GW spread over it is about 7.9 mW/m^2, matching bright-aurora energy fluxes of order 1-100 mW/m^2 [memory, medium]. The prototype's 12-acre screen (4.9e4 m^2) intercepting the flux at that altitude would receive 0.4 kW at 7.9 mW/m^2, or about 5 kW at 100 mW/m^2. Getting 60 GW needs a collector gain of about 10^8 in area, which is the entire content of the unspecified "tornadic stream". 2. Induction bound. A conductor lying on the ground feels the geoelectric field. The Geophysical Institute reports about 10 V/km and pipeline currents of order 50 A in mild activity [web: https://www.gi.alaska.edu/alaska-science-forum/aurora-makes-waves-ground-too, via the Investigator]; with a polar-cap potential of 200 kV over about 4,400 km the ionospheric field is about 0.045 V/m [calc; memory, medium on the geometry], so the screen's 240 m width sees about 11 V. Even if it carried the whole 3e5 A, P = 3e5 x 11 = 3.3 MW, 18,000 times below 60 GW [calc]. At 10 V/km and 50 A, a 60 GW tap would need a conductor 1.2e8 km long [calc], about 3,000 times the Earth's circumference. 3. Economics. 60 GW continuous is 1.9 EJ per year [calc], about 0.3% of world primary energy (about 600 EJ) and about 1% of the energy in the oil burned in a year (about 190 EJ) [memory, medium]. As an "alternative to oil" it would be small even if fully collected.

The geophysical magnitudes are real (electrojet currents of order 10^5-10^6 A; Davis's "half a million amperes"; storm-time dissipation of order 10^11 W) [web: https://www.gi.alaska.edu/alaska-science-forum/auroral-electrojets, via the Investigator]. The claim fails at the step the book itself cannot specify: how the energy gets from there to a ground terminal. The book states the method honestly: "In the absence of all other evidence... one is compelled to consider the thought that the DP is an experimental device designed to draw power from the aurora" [read: PDF p.134]. That is an inference by elimination, and the elimination chain has errors (the "not a research tool" statement is contradicted on p.128; the "not an ELF system" claim is contradicted below).

5.4 Other items

• "Not an ELF system... HF wave energies are completely incapable of effecting the coherent ELF pulsations" [read: PDF p.125]. Mainstream experiments generated ELF/VLF waves by amplitude-modulating HF heating of the electrojet at Tromso, HIPAS and HAARP [memory, high]. The book is wrong on this point.
• "Major announcements on the successive breakthroughs of the DP soon" [read: PDF p.138]: HAARP's operator history (construction 1993, 180 elements and 3.6 MW by 2007, transfer to the University of Alaska Fairbanks in August 2015) records no energy announcement, and ten megawatts of prime power come from diesel generators [web: https://haarp.gi.alaska.edu/faq]. A 29-year silence in a facility that is publicly documented and open to visitors is evidence against this reading of HAARP, not against the aurora; the competing explanations (thesis wrong; work classified or moved; HAARP never an energy project) are set out by the Investigator and the third seems most economical to me, since the facility's logged power draw is in diesel megawatts.
• Ownership: Google Patents legal events for Eastlund's US 4,686,605 show assignments to Atlantic Richfield (1985), then APTI (1987), then BAE Systems (2006), with none to Raytheon [web: https://patents.google.com/patent/US4686605A/en]; APTI was bought by E-Systems in 1994, E-Systems by Raytheon in 1995, and APTI spun out in 1998 [web: https://en.wikipedia.org/wiki/Advanced_Power_Technologies]. So the book's "owned by Raytheon" and "swept the patent floor" describe a temporary 1995-1998 ownership of the patent holder, not of HAARP; HAARP itself was a U.S. Air Force, Navy and DARPA programme [memory, high].
• The Eastlund patent does cite only Tesla items among its non-patent references (Liberty 1935; NYT 1915 and 1940) [web: same patent page], although the book says it is unrelated to Tesla.

(b-d). Consistency: the electrojet exists and carries real current. Mainstream HAARP experiments (artificial airglow, ELF generation, plasma turbulence) are documented in the research literature [memory, high]. The claim of net power extraction has no experiment or neutral test. If it were true, the magnetosphere-ionosphere coupling would have to allow a ground tap to modify magnetospheric current closure, and an external ground circuit would need a closing path through the sheath at the ground-ionosphere boundary; no framework for this exists in magnetospheric physics.

Verdict (split). 60 GW at one terminal at no cost: Refuted by calculation (aperture and induction bounds above; the funnelling step is unspecified). Mechanism ("bunching", "tornadic stream", "valve"): Not testable as stated. Prediction of announcements: Tested, not supported. Claim of "not ELF": contradicted by experiment.

---

6. Claim 6: the auroral record (book pp.130-135)

(a) Mathematics. The numbers the book gives about the record are mostly right or loose in an explainable way (see 5.2 item 4). The chapter offers no equation that links the record to energy extraction.

(b) Consistency with experiment. - Lemstrom: documented as a real historical programme (1872-1884, Finnish Lapland, with experiments at Sodankylae in 1882-1884) [web: https://pubmed.ncbi.nlm.nih.gov/38494601/ ; https://hgss.copernicus.org/articles/15/17/2024/, via the Investigator]. Corliss's anomaly catalogue: "Lemstrom's experiments were widely reported but apparently never replicated" [web: archive.org text of Geophysical Anomalies, via the Investigator]. The book itself admits his passive apparatus "did not produce consistent available outputs of the energy" (PDF p.132). No energy measurement is given anywhere, so "millions of electrical watts" (PDF p.130) is a figure of speech [read]. - Ground-touching aurora: Corliss catalogues the reports; the Norwood 1929 case is quoted in Corliss from Nature 127:108 (1931), while the book dates it "winter of 1929" [web: archive.org Corliss text, via the Investigator]. The mainstream view is that aurora forms at roughly 90-300 km [memory, high]; the low-altitude reports are witness accounts of luminous phenomena. They remain in the record. They give no energy measurement. - Long-delay echoes: first heard in 1927 by Jorgen Hals near Oslo and reported to Stormer, echoes about 3 s after the signal, investigated with van der Pol, "still considered anomalous, with several proposed explanations" [web: https://en.wikipedia.org/wiki/Long_delayed_echo, via the Investigator]. The book says "1928"; the correct statement is first heard 1927, tested 1928. This is a real unexplained effect, and I record it as pertinent evidence beside the book's storage speculation. - Siple-Roberval: injected VLF signals were amplified. Helliwell's experiments report "growth rates of the order of 100 dB/s and total gains up to 30 dB" (1974) and 30-50 dB (1988) [web: Google snippets via the Investigator; pages not opened]. The energy comes from resonant radiation-belt electrons [memory, high]; a power gain of 10^3 from stored particle energy is consistent with energy conservation because the source is the electrons. This is the strongest item in the author's favour and I keep it as such. It shows signals can trigger magnetospheric energy release at the 30 dB level; it does not show a ground terminal receiving that energy.

(c) Tested? The observations are reports and one documented experiment series (Siple). The inference "a transmitter can pull auroral energy down" has not been tested as far as the files show, and the one operating test bed, HAARP, has no such result (Claim 5).

(d) What would change? Mainstream magnetosphere-ionosphere coupling would need a mechanism by which a weak ground transmitter triggers a macroscopic downward current closing to ground. Gyroresonant triggering by whistler-mode waves is accepted (Siple); a ground-closing one is not.

Verdict (split). Reported observations (Lemstrom luminous effects, Siple amplification, long-delay echoes): Established as observations; the echo mechanism is open. Ground-touching aurora: Never properly tested. "A transmitter can pull energy down to ground": Never properly tested, and not supported by the aperture and induction bounds in 5.3.

---

7. Claim 7: Farnsworth's 1966 fusion (book p.137)

(a) Mathematics. None. The text says "In 1966, this goal was indeed fulfilled in the hands of Dr. Philo Farnsworth... successfully produced a self-sustained thermonuclear fusion reaction, and a consequent electrical output of enormous value" [read: PDF p.137]. Two category errors: an electrostatic fusor is not "thermonuclear" (its ions are accelerated by grid voltage, not in thermal equilibrium), and "self-sustained" requires the reaction power to match or exceed the input.

(b) Consistency with experiment. Order of magnitude: a deuterium-deuterium reaction releases about 3.65 MeV on average (two branches, 3.27 and 4.03 MeV) [memory, high]. A fusor yield of order 10^10 neutrons per second (Hirsch's published range, order 10^9-10^10; I did not re-check [memory, medium]) equals roughly 1e10 x 2 x 3.65 MeV = 12 mW of fusion power against an input of, say, 50 kV x 10 mA = 500 W [calc]: about 10^-5 of break-even. A theoretical limit on non-thermal plasmas, also applying to the fusor (Rider, Phys. Plasmas 4, 1039, 1997 [memory, medium]), says net gain from such devices is not expected.

(c) Tested? Yes, at ITT's Fort Wayne laboratory. The University of Utah finding aid for the Farnsworth papers (MS 0648) says ITT funded "fusors" and that "For Scientific reasons unknown to Farnsworth and his staff, the necessary reaction lasted no longer than thirty seconds"; ITT's board pressed to end the work in December 1965 and it was funded one more year [web: https://archiveswest.orbiscascade.org/ark:80444/xv98312, via the Investigator]. Wikipedia, citing the same papers, says further funding was refused in 1967 and that "no fusor has ever come close to break-even" [web: https://en.wikipedia.org/wiki/Fusor]. The family archive and ITT each have a stake in how the ending is told. This record documents fusion reactions at ITT; it does not document self-sustained fusion. It also documents a funding decision, which is the factual core the book's "suppression" story starts from; motive is unrecorded.

(d) What would change? Plasma physics would need a way around the non-equilibrium and grid-loss limits; nothing in the record shows that.

Verdict. Self-sustained fusion with electrical output: Tested, not supported. Suppression by ITT: Never properly tested (a documented end of funding, no record of intent).

---

8. Cross-checks and items nobody else is likely to check

Seafarer standing-wave arithmetic (digest 12 "not yet deep read"; now checked)

Book: "At the SEAFARER operating frequency of 76 cycles, the wavelength is 2447.37 miles. Taking half the earth circumference as 12444.5 miles, there would be approximately 5.035 standing wave zones" [read: PDF p.120, book p.240]. - The wavelength is correct for the speed the book uses: 186,000 mi/s / 76 Hz = 2,447.37 mi exactly [calc]. (With c = 299,792 km/s it is 2,451 mi, 3,945 km.) - 12,444.5 / 2,447.37 = 5.085, not 5.035 [calc]. Almost certainly a digit slip (8 as 3), because 5.085 is what the numbers give; the digest's "5.08" is right. - The conceptual error is larger. Between adjacent nodes of a standing wave lie half a wavelength, so half the earth's circumference holds 12,444.5 / (2,447.37 / 2) = 10.17 half-wavelength zones, not about 5.0 [calc]. The author counted wavelengths, not zones. - ELF in the Earth-ionosphere waveguide has a phase velocity of roughly 0.7-0.8 c [memory, medium], so the effective wavelength is shorter (1,800-2,000 miles), which would give 6.3-6.8 wavelengths on half the circumference [calc]. These are refinements; the book's use of free space is not what an engineer would use. - The book's conclusion (an ELF signal cannot be focused on a target; wavelength about 3,900 km) is the mainstream view. Its "genocide" scenario is not credible by the system's numbers: the two ELF transmitters used 2.6 MW of input power and radiated about 8 W of ELF [web: https://en.wikipedia.org/wiki/Project_ELF] (radiated fraction 3e-6 [calc]). The book itself notes "analysts" imagining "nearly one megawatt" peaks (Chapter 6, book p.~241) and rejects them.

Other VLF figures in Chapter 6

The book gives "the 10 Megawatt Jim Creek Station, call letters NLK" and "NAA is a 2 Megawatt station which operates at 14.8 Kilocycles" [read: Chapter 6, extracted text]. Jim Creek (NLK) radiates on 24.8 kHz with a power of 1.2 MW [web: https://en.wikipedia.org/wiki/Jim_Creek_Naval_Radio_Station]; NAA at Cutler is generally listed near 24.0 kHz [memory, medium; the Cutler page I fetched had no figures]. The book's 10 MW and 14.8 kHz are wrong or unsupported.

Other numeric/units items

• "Gigawatts of energy" (PDF pp.129, 135) and "for centuries" mix power with energy.
• 1.4: Argus 1.7 kt yield, 12,900 miles altitude (error), above.
• Edison "in the latter part of the Twentieth Century" (p.136; Edison died in 1931) and Saudi "65 percent" (p.136) are the deep read's flags; I agree they are errors.

---

9. Corrections to Step 1 (and to earlier documents)

1. claims.md claim 2 and the Colorado Springs deep read ("1100 A has no source"). The Investigator's correction is right: 1100 A is Tesla's own number (1917 Edison Medal speech; 1919 My Inventions caption) [web: Leads in sites.md]. I add the arithmetic: at the book's 150 kHz, I = w C V gives C = 58 pF for 20 MV and 1100 A, a 0.5 m-radius ball. That reproduces Tesla's two numbers without any "measure of aether flow". The book's 100 MV with 1100 A needs a 10 cm ball [calc].
2. Deep read, section 3.7, "editing seam". Withdrawn: both receiver descriptions are in Ahlers's 1967 summary [web: tfcbooks savo_account, per the Investigator]. The deep read's statement that the two descriptions "do not truly conflict" stands.
3. Deep read Ch.7, E8.3 (power chain). The ambiguity is resolved: 18 energised elements x 2 orthogonal dipoles, each with one 10 kW transmitter, gives 36 transmitters and 360 kW (20 kW per element). The same figure gives 960 kW for 48 and 3.6 MW for 180 elements, matching the official stages. The deep read's alternative of 4 dipoles per tower is not needed.
4. Investigator, sites.md ("The prototype element count also conflicts: 48 ... 18"). Not a conflict: the book says 48 towers stand and 18 are energised (PDF p.129), identical to the UAF history's 18 elements at 360 kW; the 48-element count is the later 960 kW stage. The one genuine slip is the author's 1996 sentence mixing the stages.
5. Deep read Ch.7, long-delay-echo arithmetic ("6 times the moon: mismatch"). The 12 s check is right, but the text says "averaging", and a mean echo delay of 7.7 s gives a path of exactly 6.0 lunar distances. The author treated the path length as the reflector's distance (a factor-two error), not an arbitrary number.
6. Deep read Ch.7, E8.5 (the 1000x / one-six-hundredth / 1.6 MW chain). The deep read said "his arithmetic holds up; the question is whether his scaling is right." I add the quantitative consequence: with the correct 1/lambda^2 law the equivalent of 1.6 MW is about 870 GW, not 1 GW, so his premise leads to the opposite conclusion.
7. digest 12 and claims.md ("Digest 12 notes that the book's 76 Hz Seafarer count ... Not yet deep read"). Checked: wavelength 2,447.37 mi is correct for c = 186,000 mi/s; 12,444.5/2,447.37 = 5.085, so "5.035" is a slip; and the quantity counted is wavelengths, not nodal zones (10.17 zones).
8. facts.json, Chapter 4 and Chapters 5-6 entries. They report "Argus" and "Jim Creek NLK 10 MW; Cutler NAA 2 MW at 14.8 kHz" as the book's statements, which is correct, but should add that the book's Argus altitude (12,900 miles) is about 43x too high (about 300 mi) [web: Operation Argus], the Argus grid-damage statement is unsupported, and NLK is 1.2 MW at 24.8 kHz [web: Jim Creek page].
9. claims.md claim 4. The wording "documented in patents" conflates the existence of a patent with the existence of working technology. It should be split: (i) the patents exist (the bibliography entries; the Investigator confirmed the 19th-century collector patents and the Eastlund family); (ii) the performance the book describes is not documented by them. The mathematician's verdicts are per item (section 4).
10. claims.md claim 6. "long-delay echoes of 3 to over 12 seconds were heard in 1928" should read: first heard 1927 (Hals), tested 1928; the 12 s figure is the author's statement and no table of delays is given.
11. Deep read, Colorado Springs (3.3) and Investigator on the 1100 A. My arithmetic agrees with the deep read's Q = 6.8 mC, C = 111 pF per metre, 5.6e5 J, and 8e10 W numbers; they are internal arithmetic, not delivered power.

No other error in Step 1 found in the parts I checked. Not checked by me: Chapter 1 history (pp.9-35), Chapter 2-3 (Marconi, Houses), the Chapter 8 rocketry history, and the numbering of the patents in the book's bibliography.

---

10. What would settle the open points

• Colorado Springs Notes (1978 edition): frequencies, supply power, capacitances, and any record of 150,000/s and 0.0062 ms. A supply-power figure would let the energy balance of claim 2 be written.
• A shielded-room test of "passes shields" (claim 1): sealed 1 mm copper cage, 6 us pulses, calibrated receiver inside.
• For claim 5: a Raytheon or E-Systems technical report on electrojet energy extraction (DTIC), and the HAARP power log; a measurement of ground-screen current during an auroral event.
• For claim 4: any open test record of CAL or Mossbauer devices; an independent decay-rate measurement of Co-57 under strong ultrasound would be a cheap, clean test of the half-life claim.
• For claim 7: Farnsworth papers MS 0648, Boxes 16-17, for neutron counts and input power.