Morray-TH (T. Henry Moray): Step 2 review by a PhD mathematician (re-run after QE "MAJOR")

Reviewer role: checks derivations, notation, units, hidden assumptions and the arithmetic of the evidence. Source tags: [read] = in the project documents or the deep reads (page numbers are the PDF page markers); [web] = fetched, URL given; [memory] = recalled, confidence stated; [calc] = computed here (Python in .venv). Source folder (read only): source folder: Morray-TH. Inputs used: qe.md, facts.json, claims.md, the six deep reads in deep/, the digests, and my first review.

What changed since the first review (read this first)

1. The first review had no text for the 4th (1960) and 5th (1978) edition scans of The Sea of Energy. They are now read. The record contains witnessed lamp loads up to about 5.7 kW nameplate (about 8 kW per one affidavit) between 1929 and 1938, a state Securities Commission inspection with meters, a sealed 158-hour endurance test, and a Fletcher-arranged test that Moray reportedly did not attend. My first review said "no instrumented check" and that observations stopped at 1.4 kW. Both statements were wrong as statements about the documents and are corrected below.
2. I kept what stands: the energy-budget refutations (cosmic radiation, Appendix IV, Appendix III) and the errors in the book's own mathematics (Chapter 10 RLC, per-person lamp count). They are not touched by the new record.
3. New mathematics done this round: the energy a battery would need for each recorded test; a lamp-brightness and iron-heating check on whether nameplate wattage is delivered power; a series-capacitor test reading; an atmospheric-electricity collector bound; the mass balance of the uranium "upgrading"; an arithmetic audit of the Cosray table and of the gold statistics.
4. Verdict changes are summarised at the end.

Source gap (rewritten)

• No document records a neutral, metered measurement of the device's input and output power [read: 5th ed. pp.32-133 deep read; 4th ed. pp.116-197 deep read]. Every power figure is a count of lamps and appliances by nameplate rating, or a later retelling. Moray's own 1960 summary says instruments for present-day energies "are of no value" for the valves and his own are "not giving ABSOLUTE readings" [read: 4th ed. pp.192-194]. Conclusions about what the device did therefore rest on friendly, non-neutral testimony; conclusions about physics rest on [calc], [web] and [memory].
• The witness letters, affidavits and the 1932 "Brief History" reach us only through books edited by Moray or his sons. Several were solicited decades later (A. B. Jensen: events 1938, statement 1956, sworn 1963; Gardner 1963; Todd and Bringhurst 1971; Jurgensen 1978; Fletcher 1979). Moray himself wrote on the 1932 booklet that it was "composed and edited too much like promotion" [read: 5th ed. p.62]. No original Securities Commission, REA, Bell Labs, utility or Frazer report is in the folder.
• Depth of reading: 5th ed. pp.32-133 and 4th ed. pp.116-197 are deep-read. 5th ed. pp.134-279 and 4th ed. pp.1-115 are digest only (so the Moray Products litigation, REA shooting episode and the 1942-43 rebuild details are less certain). I did not examine any physical object, photograph, or the audio/video.
• Handwritten witness notes (Betz, Wachs, Noffsinger), the AEC fragment and the Sen. Bennett fragment are read only through poor OCR. The FDA 1,039-page volume was searched, not read.

---

Mathematics found in the source (kept from the first review, with additions)

M1. Chapter 10, the RLC (resistor-inductor-capacitor) derivation [read: page image pp.113, 115 of the tesla.hu transcription]

Symbols: Q = capacitor charge (coulombs), C = capacitance (farads), L = inductance (henries), R = resistance (ohms), I = current (amperes), t = time. - Equation L Q'' + R Q' + Q/C = 0, underdamped for R^2 < 4L/C, Q = Q0 e^(-Rt/2L)[cos(wt) + (R/2Lw) sin(wt)], w = sqrt(1/(LC) - R^2/(4L^2)). Checked symbolically [calc]; correct. - Errors as printed: (i) period "T = 1/sqrt(...)" lacks 2 pi, and the next line (T = 2 pi sqrt(CL)) has it, so the page contradicts itself; (ii) with E = Q/C and I = dQ/dt the line "E - IR = L dI/dt" gives -Q/C, so the sign needs I = -dQ/dt; (iii) "mu = R/2L + alpha" should be "-R/2L + alpha" (p.114). - "Every oscillation, whether large or small, is completed during the same interval of time" (p.115) is true of a linear RLC (period independent of amplitude) but the same paragraph says the period depends on L and C, which is a different statement; the page conflates "independent of amplitude" with "independent of circuit". Minor. - Decisive point. The book's own result is geometric decay of the amplitudes ("lost in heating the circuit", p.115). It then asserts, without any equation, that for Radiant Energy "the oscillations can continue forever because of the actions of the universe". The correct model is the driven equation L Q'' + R Q' + Q/C = V(t). The book never writes V(t), its amplitude, or a power. Resonance multiplies the voltage across the capacitor by the quality factor; it does not raise the average power the source supplies. - New this round: Moray's own 4th-edition theory paragraph says "Oscillations are started in the first stage ... by exciting it with an external energy source ... the oscillations are sustained by harmonic coupling to the cosmic wave frequencies", and that the machine then needs no further excitation [read: 4th ed. p.116]. In the recorded demonstrations the excitation was a physical energy input every time: magnet stroking for 4-10 minutes [read: 5th ed. pp.42, 50], a shock at the terminals (Jensen got "a rather vigorous shock" before the switch was thrown [read: 5th ed. p.46-49]), rubbed rubber and iron discs (Yates 1930 [read: 4th ed. p.179]), a "hand electro-static plate" (Todd [read: 5th ed. p.112]). A started-by-stored-charge decaying oscillation looks the same on the lamps for a short time as a sustained one. The driven-versus-decaying distinction needs an energy audit of the run, which the 158-hour test came closest to (see Claim 1).

M2. Appendix IV (C. Warren Simmonds): required energy density [read: text layer p.130]

Radiation of energy density rho moving at speed c through an ideal absorbing collector of area A delivers P = rho c A. For 50 kW and 1 m^2: rho = 5e4/3e8 = 1.67e-4 J/m^3 (book: 166e-6; correct). This is a required density, not a source. Sunlight is 1361/3e8 = 4.5e-6 J/m^3, so 50 kW per m^2 needs 37 times full sunlight captured at 100 percent [calc; solar constant [web https://en.wikipedia.org/wiki/Solar_constant]]. Cosmic rays are about 1 eV/cm^3 = 1.6e-13 J/m^3 [web https://en.wikipedia.org/wiki/Cosmic_ray], about 1.0e9 times too weak [calc]; for isotropic arrival the usable flux through a flat collector is at most rho c/4, a shortfall of about 4e9 [calc]. "Regardless of its exact form" hides the assumption that radiation of this density exists.

M3. Appendix III (Richard E. Maxwell): Poynting-vector estimate [read: text layer p.128]

P = Ex Hy d^2 = 3979 W with Ex = 100 V/m and Hy = 39.79 A/m. 39.79 A/m = B/mu0 with B = 5e-5 T (Earth's static field) [calc], and 100 V/m is the order of the static fair-weather atmospheric field ([web https://en.wikipedia.org/wiki/Atmospheric_electricity]: about 100 V/m). Static fields do not carry power through a surface. In a travelling plane wave E/H = 377 ohms; here E/H = 2.5 ohms. A real plane wave with E = 100 V/m has H = 0.265 A/m and 26.5 W/m^2 [calc]. The appendix also concedes a 24-hour periodicity and near-zero power at night, which the author says contradicts the reported night operation.

M4. Chapter 7: lamps per person

• Corrected figure. Moray's own 4th edition (1960, pp.8-9) and the 1945/"1926" pamphlets give 1,193,600 100-W lamps per person (1.1936e8 W); the sons' 5th edition (1978) gives 1,693,600 (1.69e8 W) [read: QE spot check 5; claims.md]. facts.json's "119,360 W" is a factor-1000 slip.
• Against sunlight intercepted by Earth, 1361 W/m^2 x pi x (6.371e6 m)^2 = 1.74e17 W [calc]: with 2e9 people (1930s) the 1960 figure gives 1.37 times all sunlight and the 1978 figure 1.95 times; with 4e9 people 2.7 and 3.9 times; with 8e9 people 5.5 and 7.8 times [calc]. Either version claims more arriving power than the sun supplies. If cosmic rays arrive isotropically at 1 eV/cm^3, the power through the whole Earth is about rho c/4 x 4 pi R^2 = 6e9 W, about 1 W per person [calc].

M5. Internal consistency of the endurance-test documents [read: 5th ed. pp.46-54 and Hayes pp.50-52; 4th ed. pp.148-151]

• Duration: 74 h 21 min + 83 h 34 min = 157 h 55 min (Jensen). Hayes says "eighty-four hours" for the second run and "158 hours" total. The test was two runs with the trunk opened between, not one continuous week (Hayes's 1929 "continuously for a week").
• End load: 575 W iron + 60 W lamp = 635 W (Hayes, Judd). Hayes 1929 says 100 W + 575 W = 675 W. "Five 100-watt globes ... total power output of 560 watts" (Hayes) does not add (5 x 100 = 500); Judd says four globes, Jensen five. Pilot lamp 10 W (Jensen, Judd) versus 20 W (Hayes).
• Hayes's "635 watts ... 0.878 of a horsepower": 635/746 = 0.851 hp [calc]. A small slip, but it appears in a sworn statement used as the central proof against batteries.
• Size and weight of the device are inconsistent across sources: 7 x 7 x 27 in (1932 history) versus 18 x 24 x 24 in (Fletcher 1979) versus 12 x 18 in (Todd) versus two boxes about 10 x 20 x 4 in (Hayes) versus 10 x 10.5 x 26 in (Nuenergy). The detector is "one ounce" (Moray's note), "less than 1 ounce" (1930 telegram note), "the vital eight ounces" (1931 narrative), "two ounce bit" (John Moray) [read: 5th ed. pp.58, 59, 76, 94].
• Performance figures conflict: tubes meeting requirements 10-20 of 100 (1960) versus 2 of 100 (1962 speech); 1 lb per kW (1962) versus 1 lb per W (1968); 4,000 W with "50 kW potential" (Hayes 1931) versus 50,000 W "obtained from one unit" (1960, no witness) [read: 4th ed. pp.193, 195-196].

---

New mathematics on the witness record

N1. Energy audit: what a hidden battery would have to hold [calc]

Lead-acid specific energy 35-40 Wh/kg [web https://en.wikipedia.org/wiki/Lead%E2%80%93acid_battery] (upgraded from [memory] to [web]).

The headline: the energy argument against a stored-energy box is strong for the 1928 endurance test and for the 1938 Securities Commission test if the lamp counts are delivered power and if the leads carried no external power. It is not strong for the 1926 Strawberry test, which is the one field test with a measured distance from power lines (26 miles at last power line by odometer [read: 5th ed. pp.40-43]). A 15-minute single-lamp run needs about one kilogram of lead-acid. So the best-documented remote test does not exclude a concealed battery.

N2. Is nameplate wattage delivered power? [calc; assumptions flagged]

• Several observers say the lamps behaved oddly: "burned cold except each had a hot spot about the size of a dime" (Todd [read: 5th ed. p.112]); the book itself says the bulbs "become exceedingly hot in one spot about the size of a dime" because the gas becomes incandescent under high-frequency current [read: 4th ed. p.131]; "softer ... more like daylight" (Pyper); photographic negatives needed "about ten times the normal exposure" in the lamp areas (photographers [read: 4th ed. pp.127-131]). A tungsten filament at rated power glows over its whole length. A lamp lit at high frequency by capacitive coupling or a gas discharge can emit visibly while taking far less than its rating. So the nameplate sums (4-8 kW) are upper labels, not measurements. This also weakens the energy audit in N1 in the direction of less energy required: if the 30 lamps took a few watts each, the Securities Commission test needed far less than 5.5 kWh. Because lamp loads cannot be turned into watts, the one thing the witness record cannot do is establish a power in watts.
• Filament lamp brightness scales roughly as power to the power 2 (light output ~ V^3.4, power ~ V^1.6) [memory, medium: not web-checked]. A lamp "75% as bright" is then about 87% of rated power [calc], used in N1.
• The iron: a flat iron brought to "sizzling" in about 5 minutes (Fletcher [read: 5th ed. p.43-45]) or about 2.5 minutes (Yates [read: 4th ed. p.126]). Heating a 1-2 kg iron (assumed; specific heat 0.45 kJ/(kg K), temperature rise 100 K; assumptions [memory, low-medium]) by 100 K needs 45-90 kJ, so 150-600 W of real heating depending on mass and time [calc]. Fletcher's 5 minutes is consistent with 150-300 W rather than 575 W; Yates's 2.5 minutes is consistent with a near-rated iron. This is a crude calorimeter, not a measurement, but it is the only energy-sensing observation in the record besides the lamps, and it does not demand the full nameplate.

N3. The 1938 series-capacitor test [read: 5th ed. p.127, E. G. Jensen, 21 Mar 1938] [calc]

Output stayed lit through a 0.025 microfarad capacitor, a glass-plate capacitor, or both in the antenna lead; a 60 W lamp in the same lead "did not light under any condition"; removing the lamp put out the output; shorting antenna to ground killed it and "no electricity could be felt". - The capacitor's reactance Xc = 1/(2 pi f C) is 106 kilohms at 60 Hz, 64 ohms at 100 kHz and 6.4 ohms at 1 MHz [calc]. Output through it therefore says the antenna lead carries alternating current at a frequency well above mains, consistent with every "high frequency" description (Fletcher, Lovesy, Yates). Moray's "high frequency" is the one property of the record that is internally consistent. - A 60 W lamp that stays dark in the lead means the lead current was well under the 0.5 A that a 120 V, 60 W lamp draws (60/120) and probably under about 0.1 A (estimate, a cold filament is about 10 times lower resistance than hot). Then 100 W through that lead would need at least about 1 kV; at 0.05 A it is 2 kV RMS at unity power factor [calc]. Equivalent readings: a high-voltage, low-current antenna circuit, or most of the lamp power not coming through the antenna lead at all (stored energy or a path not in the diagram). The test cannot decide between these. It does show that the antenna lead was not a low-voltage, high-current feed. - Jensen's own meters on the antenna and ground "were electrically dead" [read: 5th ed. p.117]. If the device was running when he metered, a reading of zero is a real puzzle, since by every other account the power flows through those leads. A meter of the period (a moving-coil or iron-vane instrument) does not respond at radio frequency [memory, high]. The statement does not say whether the box was running at the time. I treat it as unresolved, not as evidence either way.

N4. How much could the antenna collect from known natural sources? [calc; [web] https://en.wikipedia.org/wiki/Atmospheric_electricity]

• Fair-weather field about 100 V/m. An 80-ft (24 m) high wire would sit at a few kilovolts relative to ground in a static field. The fair-weather conduction current density is about 3.5e-12 A/m^2 (global current about 1,800 A over 5.1e14 m^2; the 1,800 A is [memory, medium], not shown in the fetched page). A collector of 10,000 m^2 (a 200 m x 50 m footprint) therefore yields about 3.5e-8 A, so about 8e-5 W at 2.4 kV [calc]. That is eight orders of magnitude below 1 kW. Point-discharge (corona) collection from sharp points can do better in a storm field, but still milliwatts in fair weather [memory, medium]. This is the natural-source figure Moray's antenna should be checked against.
• Radio: a 1 V/m AM field gives S = 1/377 = 2.7 mW/m^2 [calc], so 1.4 kW needs about 5e5 m^2 of ideal collector. Nuenergy's alternative (spark-gap amateur stations until 1928) fails the same arithmetic and also contradicts its own p.19 statement that radio signals "are much too weak" [read: Nuenergy deep read, pp.9-21].

---

Claim 1. The RE device (box, antenna and ground only) lit lamps and heated irons without an internal source: 655 W (1925), loads up to about 5.7 kW (1938) in witnessed tests, and "50 kW" in later statements

a. Mathematically consistent? The claim is a set of reports, so the check is arithmetic and consistency. (i) The numbers are inconsistent with each other in the ways listed in M5. (ii) No wattage is a measurement; each is a sum of nameplates (N2). 655 or 665 W is the rating of the Hotpoint iron plus a 100 W lamp [read: 5th ed. p.36-37], not a measured output. The sums by witness: Fletcher 1928 about 1.4 kW (12 x 75 W + 500 W per Fletcher's 1979 memory; contemporaries say 3 x 100 W + 575 W); Welling 1930 1,475 W; Lovesy 2,150 W; Yates 1930 2,475 W (32 x 50 + 3 x 100 + 575); Todd 1937 about 4 kW; A. B. Jensen 1938 5.5 kW; Bringhurst 1938 about 5.7 kW (20 x 150 + 15 x 100 + 600 + 575); Gardner 1936 about 8 kW (40 x 200) [read: 5th ed. pp.32-133; 4th ed. pp.132, 179-183; sums [calc]]. The 4-50 kW figures in the book's own summary have no witness, date or measurement attached ("50,000 watts" [read: 4th ed. pp.195-196]). (iii) Strongest feature, mathematically: the endurance test. If the rating counts as delivered power the box delivered 17.4 kWh in 158 hours (N1), which is 435-500 kg of lead-acid, against two boxes of about 10 x 20 x 4 in. Hayes's own sentence "This alone is sufficient to dispose of any suggestion of a battery" is arithmetically right about the energy, even though he used the wrong horsepower figure and applied it to the end-of-test 635 W instead of the 17 kWh. The test used a water-pipe ground and a 200 ft antenna 80 ft up [read: 5th ed. p.46-52]. (iv) Weaknesses, also mathematical: nameplate wattage at high frequency is not delivered power (N2); the test ran in two segments, was re-tuned, and was started by a physical excitation; the main run was 110-120 W, so it is an energy argument at about 17 kWh, not at 635 W; the pilot lamp and end loads disagree by observer.

b. Consistent with established experiments? For the energy accounting: no known source supplies kilowatts through an antenna: radio fields give milliwatts per m^2 (N4), the fair-weather atmospheric field gives microwatts, cosmic rays give 5e-5 W/m^2 at most (see Claim 3). The new witness record shows an unexplained behaviour (lamps lit through a series capacitor while a 60 W lamp in the lead stays dark; HF brush discharges) that is physically ordinary for a high-frequency oscillator but not evidence of a source. The established fact that matters: every power-producing device with a known mechanism needs a source; the account describes a load, not a source.

c. Tested? By whom? Replicated? Observed, in the tabulated deep read, on about 25 dated occasions between 1925 and 1938 by at least 30 named witnesses. Instrumented items: Jensen's meters (1938), Yates's terminal-short and condenser-discharge checks (1929-30), Welling's check of the house meter (1930), the crystal-set test (Yates 1930), the series-capacitor test (1938). Seals on the trunk with recorded numbers (1928). Not present: any metering of the input to or output from the box, any measurement of voltage, current or frequency at the loads, any antenna-less run, any run with a dummy detector. Fletcher proposed exactly that last kind of test in 1928 ("by redesigning the system the same performance can be obtained without the use of the antennae" [read: 5th ed. p.76]). The detector was withheld in every examination except possibly the 1938 inspection [read: 5th ed. p.66-68, 117]. Replication: none that succeeded. Murray O. Hayes, Moray's patent agent who said he had seen everything, testified in Reno in 1932 that he and associates tried to build it from the patent papers and failed [read: 4th ed. p.169; editors' report, no transcript]. The Moray tubes in later hands were said to show no radioactivity, with no instrument given [read: Nuenergy deep read]. In the 1950s and 1957 demonstrations are mentioned; I read nothing on instruments there.

d. If true: Either the energy balance fails (a source about a billion times denser than any measured radiation, or a new channel) or a hidden or stored source existed. The cheapest settling measurement is a metered run: an RF power meter or a calorimeter on the box input and output, with an antenna-less control and a dummy detector.

Verdict (split): - (i) "Lamps and irons lit, up to a few kilowatts nameplate, for hours, in the presence of non-neutral witnesses, with no battery visible": Never properly tested. It was witnessed repeatedly, with some meter checks, but never metered or independently reproduced. The record also includes a test offered by Fletcher and, per Fletcher, not attended by Moray (see Claim 6), so the right label is "tested by friends, offered a neutral test, not completed". It was not dismissed without any look: several experienced engineers looked and said they could not explain it. - (ii) "A stored-energy box cannot explain the 158-hour endurance test": Plausible, untested as an arithmetic argument. It is strong for 17 kWh in the stated boxes, but it depends on unmetered nameplate numbers and on whether the ground and leads were isolated; it does not exclude a hidden feed from the building's bonded water pipe or a hidden wire (N5 below). - (iii) "The 1926 Strawberry run excludes hidden power": Not supported by the arithmetic (0.6 kg battery, N1). Never properly tested. - (iv) "50 kW": Not testable as stated (no witness, date or measurement) and Refuted by calculation if the source is any known radiation (Claim 3).

N5. The hidden-feed hypothesis, mathematically. Knudsen (1931) suspected a hidden wire in the 90-foot antenna pole [read: 5th ed. p.125]. A feed conductor hidden in the antenna lead would make every disconnect test (lights off when the antenna is removed) come out as reported, and a water-pipe ground bonded to the building neutral would survive pulling the main switch [read: deep read note, 5th ed. p.50]. The tests that bear on it: freshly erected antennas at remote sites (Strawberry, Dugway), the counterpoise substitution (Pyper: a hand-held elevated wire in place of earth), and the capacitor test (the lead carries HF at low current, which a DC or 60 Hz hidden feed would not). The remote tests have no power measurement and no duration for Dugway. This hypothesis has not been refuted by any record I read.

---

Claim 2. It was not induction, hidden wires or batteries; demonstrations were repeated where observers chose

a. Mathematically consistent? Logically an argument by exclusion with hidden assumptions: that "no wire and no battery seen" implies "no source inside", and that the detector (the one element that could contain a source) is irrelevant. The record shows the detector was shown in Moray's hand, not examined, in 1925 (pieces pocketed), 1927 (a part cut off and pocketed), September 1928 (Fletcher), the 1928 endurance test (examined "with the exception of the detector"), 1931 (Baldwin), 1933 (Jurgensen) and 1963 (H. Eyring) [read: 5th ed. p.66-68]. The one exception is the 1938 Jensen inspection (parts "weighed and tested"), which does not say the detector was opened [read: p.117]. Start-up needed excitation every time and the device died if the antenna was left off too long [read: 5th ed. p.32, 37].

b. Consistent with established experiments? Induction from power lines at 26 miles (Strawberry) is negligible; I accept that part. The better checks are Yates's (terminals shorted with no spark: not a battery; condenser charged from mains gives an arc: condensers present; crystal set on the lead-in gave hum-free reception: not mains pickup), Pyper's shorting of the aerial to ground ("no spark, it just turned off the power"), and the series-capacitor test. These are consistent with a high-frequency oscillator excited by a source or by stored charge. Fletcher stated both alternatives in 1928: an outside source, or "battery action" in the tubes started by atmospheric energy [read: 5th ed. p.44-46, 75-76]. His conclusion: "I am just as puzzled as ever" [read: p.75].

c. Tested? By friends and by two sets of semi-neutral observers: the Securities Commissioner (Gull) with battery expert A. B. Jensen in 1938 (holes bored in the box, parts weighed and metered), and the Welling/Gaxiola visit of 1930 (state Secretary of State and the Mexican Consul General; house meter checked; they tuned the device themselves) [read: 4th ed. pp.132-133, 181-183]. Both are reported years later by Moray's side. Not done: a blank run with the detector swapped for a dummy; any antenna-less run; any neutral custody of the device.

d. If true: the same as Claim 1.

Verdict: Never properly tested. More than the first review said (meters were used, in 1938 and 1930), less than the family says (no metered power; detector withheld). The sequence "witnessed, partially inspected, detector withheld, neutral test offered, not completed" is classification-wise "not dismissed without a look" but also "no neutral, controlled test on record".

---

Claim 3. The energy source is oscillating cosmic radiation, captured by resonance; the condensers keep oscillating because the surroundings keep exciting them; energy is conserved

a. Mathematically consistent? Split, as before. (i) The resonance mathematics as printed is a decaying RLC with typos (M1). (ii) "Oscillations continue forever because of the universe" adds an energy source to an equation with none, and the source power is never computed. (iii) The new fact that the first stage is "excited externally" [read: 4th ed. p.116] means the device in operation could be a primed, decaying oscillator with a trickle of input; the book never quantifies which. (iv) The book computes the 1.5-million-lamps figure and the 50 kW density itself (M2, M4) and never compares them with measured radiation. Calculation, unchanged: cosmic radiation at 1 eV/cm^3 = 1.6e-13 J/m^3 [web https://en.wikipedia.org/wiki/Cosmic_ray] times c gives at most 4.8e-5 W per m^2; a 50 kW load needs 1e9 times that (4e9 with isotropic arrival) [calc]. A larger antenna does not help: 5e4/1.2e-5 = 4e9 m^2 = 4,000 km^2 of ideal collector [calc]. Atmospheric electricity (N4) gives 8e-5 W for 10,000 m^2.

b. Consistent with established experiments? Measured cosmic-ray energy density and flux agree with the figure used. The fair-weather atmospheric field and current are measured (N4). Galactic radio background is many orders below cosmic-ray power [memory, high]. The "oscillations of the universe" frequency is never defined; Appendix III says solar-type sources would vanish at night, contradicting the reported night operation. Perreault's (Nuenergy) alternative of local spark-gap stations to 1928, then radioactive dopants, is unsupported: its "military records", tube inspection (no instrument, no result) and Pb-210 argument (half-life about 22 years, so not "burned out" in 25-40 years [memory, high]) are weak, and it contradicts its own sentence that radio signals "are much too weak" [read: Nuenergy deep read, pp.9-19].

c. Tested? No test of the cosmic hypothesis is on record. The 1968 proposal drops it for a radioisotope-excited semiconductor with an admitted nuclear source [read: 1968 Air Force proposal, as in first review].

d. If true: The radiation energy density at Earth's surface would have to be about 1e9 times what instruments measure, or there would have to be a form of energy invisible to every detector except Moray's. It would show in the Earth's heat balance and in every radio telescope.

Verdict (split): (i) Quantitative strong form (kW from cosmic radiation at known densities): Refuted by calculation (4.8e-5 W/m^2 vs 5e4 W; atmospheric and radio collection fare no better). (ii) "Resonance with the oscillations of the universe sustains energy": Internally inconsistent as written (own equations show decay; no source term) and Not testable as stated (no frequency, amplitude, detection). (iii) An unknown source or a primed/stored-energy mechanism: Never properly tested (the cleanest test would be a metered energy balance).

---

Claim 4. The secret is the "Moray valve"/detector, a cold-cathode semiconductor "stone" that preceded the transistor; the 1968 version is a radioisotope-excited semiconductor cascade at about 1 lb per watt

a. Mathematically consistent? No equations are given. Quantitative statements: the Patent Office's 1932 inoperability grounds (unheated cathode, radioactive ionization without plate potential [read: docx rejection, 23 Dec 1932]) are a correct statement that a passive diode cannot deliver net power. New numerical data from Moray's own 1960 status: only 10-20 tubes in 100 meet requirements, 29 tubes per unit at about $500 each (about $14,500 per unit), development cost $5,380,000, 90% complete, instruments "of no value" for the valves [read: 4th ed. pp.192-196]. A unit that needs 29 tubes selected 1-in-5 to 1-in-10 and "perfectly balanced" is a delicate oscillator, which is in line with its observed fragility (a felled tree, a jarred bench, or a hammer blow put the lights out [read: 5th ed. p.46-52, 4th ed. p.183]). The 1930s claim (50 kW in 60 lb, 833 W/lb) and the 1968 claim (1 lb per W or "well below one pound per watt") differ by a factor of about 1,000 [calc]; the 1962 speech says 1 lb per kW. The detector's mass is stated four ways (M5). Moray's valve date: a sworn certificate of 14 Nov 1927 describing "pure germanium" is blanked in the book [read: 5th ed. p.67-68, 90]. Transistor action was demonstrated at Bell Labs 17 Nov-23 Dec 1947 [web https://en.wikipedia.org/wiki/Transistor]. A germanium detector in a 1920s box is possible; "preceded the transistor" requires current gain, which nothing in the record shows.

b. Consistent with established experiments? Point-contact detectors are real [memory, high]; a detector rectifies power already present. A negative-resistance element converts DC bias into oscillation but needs a bias source. Betavoltaics: microwatts, tens of grams [web https://en.wikipedia.org/wiki/Betavoltaic_device]. Nuenergy's mechanism (radium-chloride/thorium/Po-210 doped germanium and zinc sulfide pellets) is Perreault's reading of Moray, with arithmetic and nuclear errors: bismuth-209 does not become Po-210 by exposure to tube emissions (it needs neutron capture) [memory, high]; "electron velocity about 1 cm/s" and "Ohm's law only applies to conventional tubes" are wrong [memory, high].

c. Tested? The detector was never examined by anyone outside the family in a recorded test. Nuenergy says surviving tubes were inspected and "there does not appear to be any radioactive material", without instrument, method or tube count [read: Nuenergy p.12]. Hayes's failed rebuild (1932) is the only recorded copy attempt.

d. If true: a passive solid-state element would be a net energy source; that contradicts thermodynamics, not just semiconductor theory.

Verdict (split): (i) "A semiconductor detector/valve of undisclosed composition exists": Plausible, untested. (ii) "It preceded the transistor": Never properly tested; probably Not testable as stated since no function is specified. (iii) "It produces net power from nothing": Refuted by calculation if no source exists (energy balance). (iv) "The 1968 radioisotope proposal": Plausible, untested as a modest-power concept.

---

Claim 5. Matter dissociates into energy; the technology transmutes elements (lead to gold, ore upgrading, uranium) and heals disease (Cosray, US 2,460,707, 1949)

5A. Gold and silver ("transmutation")

a. Mathematically consistent? The mechanism fails: 8 MeV photons (the Varian linac [read: Rex pp.6-7]) can remove nucleons but not build Z = 19 up to Z = 79; adding charge needs fusion past a Coulomb barrier of order 100 MeV [memory, medium]; binding energy per nucleon peaks near iron, so building gold from lighter elements absorbs energy [memory, high]. Hooper's own statement that nothing above Z = 19 was in the solution [read: Rex p.3] sits uneasily with John Moray's 1981 statement that the "environment" contains arsenic and iron pyrites and antimony (Z = 33, 26, 51), which are classic natural gold carriers [read: Rex pp.21; hosts [memory, high]]. A result of 929.76 oz/ton of gold in a residue is 3.19% gold by mass [calc], i.e. visible gold crystals in a few grams of salts, which is what contamination or assay-flux gold would look like. The Rex deep read shows the test design in more detail than the first review had. One irradiated no-solution control pair exists (Radiation Dynamics linac, Aug 1963): A2 0.12 and A3 0.26 oz Au/ton against 17.2-100 oz/ton with solution [read: Rex pp.13-14], a 66- to 833-fold contrast [calc]. n = 2, nonzero (so gold appeared with no solution), likely not covering the reagent-grade sample (if the labels are 2A/3A), no unirradiated blank, no environment-only blank, no assay blank, no replicate at any condition [read: Rex pp.12-14, controls table p.4 of the deep read]. Statistics: Rudolph's Varian mean is 15.16 oz/ton with a "97.5% certain above 10.51"; Olsen's mean is 11.79 with a 95% interval 9.22-14.36 (p.18 misprints 14.66). Rudolph's mean is outside Olsen's interval, so they cannot describe the same data by the same method. The back-calculated standard deviations are about 18 and about 10 [calc, from the deep read]. A one-sided 97.5% lower bound of 10.51 and a "99% lower limit about 10.5" for a hand-edited distribution cannot both be right. Olsen's reading of a 95% interval as "97.5% chance of lying above 9.22" is a misreading of a frequentist interval. The distribution was negative-exponential with three outliers and zeros "spread out" [read: Rex pp.7, 20-21]. Conditions were deliberately not held constant [p.20]. Pilot plant: nine runs, three of nine below 1-5 oz/ton; mean 19.2 with "standard deviation was quite high" [p.17]. Subsequent linac runs "very few have shown any worthwhile results" [p.7]. The 1970 table: oz/ton falls by exactly 2 when sample mass is doubled (30 g: 65-67; 60 g: 33-35), which is what a per-gram normalisation artefact predicts [calc/read: p.19]. John Moray contradicts himself on the "bombardment tube" (1964 vs 1982) [p.22]. b. Consistent with established experiments? Accelerator transmutation of Bi or Hg to gold has been done atom by atom [memory, medium: Seaborg 1980; not web-checked]; macroscopic gold from salts at 8 MeV is not seen. Fire assay has known flux and reagent gold, so blanks are standard [memory, high]. The reaction is "sensitive to interference" from plastics, stainless steel, rubber and brass [read: Rex p.5], the pattern of an uncontrolled artefact. c. Tested? By the Institute itself, at outside accelerators (High Voltage Engineering, Electronized Chemicals, Radiation Dynamics, Varian), with the Institute's technician and solutions, assays "by a wholly independent firm" (unnamed, no certificate, not stated to be blind) [read: Rex pp.2-3, 13]. The first review's statement that assays were "by the Institute's own people" is not supported by the text and is withdrawn. Their own repeat attempts failed ("efforts to duplicate ... failed" [read: Rex p.6]). No outside group has replicated. The 1940-41 Boaz Mine/Union Assay Office assays in the 1945 pamphlet are not examined (no certificates in the folder).

5B. Uranium ore "upgrading" [read: Isotopes in Ionic Recovery of U3O8, pp.33-40]

The booklet reports 0.04-2.6% ore raised to 30-67% U3O8, a US Bureau of Mines radiometric report (28 Sep 1956, Sample UP 6234-4: 21.06, 18.82, 27.94, 18.22% "U3O8 equivalent"), an AEC Grand Junction statement that a 30% sample "appears to be fused or reduced uranium dioxide ... will probably assay between 75 and 85%", and commercial assays (Crisman & Nichols; Black & Deason). Mass balance [calc]: with no uranium created, the surviving mass fraction is f0/f1: 0.047/21.947 = 0.21%, 0.09/53.44 = 0.17%, 2.6/67.22 = 3.9%. That means discarding 96-99.8% of the mass; no mass of "after" is given anywhere (only one "1 lb" input). This is exactly what ordinary physical or chemical concentration does, while the booklet says "no silicates or sands are removed" without a mass balance. Transmutation instead would need 26 to 590 times more uranium atoms in the same rock. The booklet's figure for the 1958 series (4.59% ore: 19.6-51.4% after treatment on nominally one lot) shows large scatter. The Bureau of Mines report has no "before" value (the booklet adds "less than 0.09%"), is radiometric (it measures gamma activity and assumes equilibrium, so radium and daughters would raise it [memory, high]), and notes an unidentified fibrous isotropic substance that "might be an artificial product". The AEC statement is a visual estimate and describes a reduced uranium dioxide, not a natural ore. "Aging" in the booklet refers to decay, which lowers uranium; reversing it is the opposite of the physics cited. Arithmetic: "several hundred percent" understates the author's own factor of about 470 (0.047 to 21.947) [calc]. No chain of custody, split sample, blank or isotopic analysis.

5C. Cosray therapy [read: Photo Actinia pp.22-54]

The clinical table (785 treated) is unsigned, undated, with no control group, no definition of "relieved", no follow-up and all patients also on conventional therapy [read: deep read, pp.49-53]. My arithmetic audit of its 48 rows [calc, from the deep read's transcription]: treated sum 783 (stated 785), relieved 706, "no effect" 75 (stated 42), 783 - 30 dropouts = 753 (stated 750), the sinus row has 309 treated but 304 + 3 = 307, sinus is 39% of all patients and 43% of all "relieved", cancer is 54/78 = 69% against a printed 75%, several printed percentages are wrong (blood pressure 10/11 printed 99%, chronic bronchitis 6/6 printed 87%). "Metabolistic general" 44/44 is not a disease. The USDA item is a June-July 1940 "written request and order" to build three devices, a build order, not a request for clinical testing; the "extended clinical tests" were by private physicians, with no results, names or dates [read: pp.48-49]. A 1941 Moray text says the medical equipment still needed a researcher. US 2,460,707 (filed 30 Apr 1943; granted 1 Feb 1949; continuation-in-part of an application filed 15 Nov 1940) [web https://patents.google.com/patent/US2460707A/en]. FDA seizure: default entered 3 Apr 1963; the Institute contested and the court denied its motion to vacate on 20 Aug 1963; article destroyed 19 Mar 1964 [read: FDA NJ 8137, QE spot check 4]. The charge was false and misleading labelling; the court did not rule on therapeutic efficacy. Hearing text records a lessee reporting no improvement (a single anecdote). d. If true: low-energy photons would have to build heavy nuclei, overturning binding-energy systematics and the Coulomb barrier; processes that raise assayed uranium 100-fold in unweighed rock would need a mass balance and an isotope check first; for Cosray a controlled clinical trial is the ordinary requirement.

Verdict (split): - (i) "Matter slowly dissociates into usable energy" as the device's power source: Not testable as stated (no rate, no measured mass loss); the quantitative form (50 kW from ordinary matter) is a mass loss of about 17.5 g per year [calc], not impossible but unmeasured. - (ii) "Gold and silver from non-gold material by photon bombardment": the mechanism is Refuted by calculation (nuclear binding energy and Coulomb barrier at 8 MeV). The reported assays are Never properly tested by any neutral party: the one internal control pair is thin, unreplicated and nonzero, with no unirradiated, environment-only or assay blanks, and the Institute's own later runs failed; their statistics are internally inconsistent. Not replicated. - (iii) "Uranium ore upgraded from below 0.1% to 20-85% U3O8 by ionic treatment": Never properly tested, and Not supported by the numbers presented: the data are consistent with ordinary concentration of 96-99.8% of the mass, contamination or radiometric artefacts; a transmutation reading is Refuted by calculation in the same sense as (ii). Third-party assays exist but only as radiometric or visual statements on treated product, with no "before" values. - (iv) Cosray therapy: Never properly tested clinically (an uncontrolled tally; its own totals are internally inconsistent, a local Internally inconsistent finding for the table, not for the device); it was stopped by a default judgment and a seizure, not by a controlled trial.

---

Claim 6. The work was blocked (Patent Office, scientists, politicians, REA, agents); the device was destroyed; Moray was attacked

a. Mathematically consistent? Not a mathematical claim. Corrected patent-ground statement: the Patent Office actions say "apparently inoperative device", "cathodes are not heated to a point where they will emit an appreciable number of electrons" and "No natural source of electric wave energy is known to the Examiner and proof of the existence of such a source is required" [read: gov-Commerce letter image per QE spot check 2; 4th ed. pp.163-164]. The first review wrongly said the natural-source ground appears only in the sons' gloss. Both grounds are the same two calculations as Claims 3 and 4 and were physically sound.

b-c. Evidence and what was tested? The new record lets one sort the episodes. The following are documented in the folder (all reported through Moray's side): - Tests Moray accepted and ran: Eyring's 3-4-mile criterion (done as Emigration Canyon, Dec 1925, without Eyring present; Judd countersigned in 1929) and Fletcher's endurance test (done, 158 hours, in Moray's own basement, with the detector withheld). - Test offered, Moray did not attend (per Fletcher): Fletcher (22 Oct 1956, 5 June 1957) and A. G. Foster (letter 31 Oct 1973, notarized, recounting Fletcher) say Fletcher arranged for a New York electrical concern or Bell Labs to build and test the device and that Moray agreed, then sent a telegram that he would not come. Moray annotated "100% untrue"; John Moray says no such correspondence exists in the files [read: 5th ed. pp.84-88]. Two tellings 16 years apart, consistent, no document either way. Unresolved. - Test Moray declined: a specific-wattage test requested by Carl Eyring ("the amperage would burn out his device"), reported by Moray's side only [read: 5th ed. p.74]. - Test cut short: Knudsen, 1931: the detector burned out during his check after he threw the switch repeatedly and added his own coil; Moray had agreed on condition of a written statement; Knudsen later suggested a hidden wire in the pole [read: 5th ed. pp.118, 121-125]. - Non-neutral rebuild, negative: Hayes, Reno 1932 (above). - Declines to examine because of concealment: Millikan (Nov 1929), Bell Labs (1937, "he was not allowed to examine all the parts"), Wilkinson of BYU (1959), Eyring (1963, 1970), Sen. Bennett (1961, "a pig in a poke"), AEC (1970) [read: 5th ed. pp.79, 82, 89, 91-93, 97-102]. Moray's reply, repeated: demonstrations stopped on legal advice to protect patent rights and the key part was shown only under contract. - Funding: NASA agreed a $5 million programme per John Moray, then judged the probability too low (21 July 1961); Bennett says Moray asked for $50,000 without disclosing the location or process [read: 5th ed. pp.99-102]. REA 1938-41: Crim's tests "entirely satisfactory as far as performance went"; the 1939 Frazer episode (a "glass test" through twelve glass sheets, then Frazer smashed the device) [read: 4th ed. pp.174-176; digest-only for 5th ed. p.181]. - Misattribution in the book: John Moray credits Fletcher with the sentence "it is established, I should say absolutely, that what we have is something entirely new"; in Judd's own letter it is Judd's conclusion [read: 5th ed. p.45 vs p.74]. Moray's own signal on method: by his own 1960 account 10-20% of tubes work and instruments do not give absolute readings; he says the machine is not yet practical.

d. If true: nothing in physics changes. It bears on Claims 1-2 in one way: the decisive tests (neutral custody, antenna-less run, metered power) were offered or implied by several people and not carried through, for reasons that are partly protective (patents, secrecy) and partly unexplained.

Verdict (split): (i) The patent rejections and their grounds: Established as a documented fact; the grounds are physically sound. (ii) "Dismissed without testing": not supported as stated. The device was seen and tested repeatedly by friends and several experienced outsiders; the neutral, whole-device test was offered and not completed (by the family's account, because disclosure was refused; by Fletcher's, because Moray did not come). (iii) "Suppressed and attacked": Not testable as stated from this folder; the shooting (2 Mar 1940), threats and the REA story are one-sided sources and digest-only for the 5th edition's later chapters.

---

Claim 7. Moray held the title "Dr."/D.Sc. (an Uppsala doctorate was claimed)

Unchanged. The Uppsala librarian found no thesis for 1910-1922 or in the pre-1962 title registry (7 Oct 2011) [read: Uppsala email]; the 2018 enquiry has no recorded answer. A negative search is weak evidence. Verdict: Never properly tested. Not relevant to the physics; the claims stand or fall on the arithmetic above.

---

Corrections to my first review

1. "No instrumented check exists": wrong. Jensen's meters (1938), Yates's checks, Welling's meter check, Pyper's shorting, and the capacitor test exist. None measured power.
2. "Observations stop at 1.4 kW": wrong; witnessed lamp counts reach 5.7-8 kW nameplate (1936-38).
3. "No test was offered": wrong; Eyring and Fletcher offered tests; Moray accepted two; the Fletcher-arranged New York test is disputed.
4. "The Patent Office's natural-source ground is only in the sons' gloss": wrong; it is in a Patent Office action.
5. "Transmutation assays by the Institute's own people; no blanks": too absolute. An unnamed independent firm is asserted; a two-sample no-solution control exists. Still no unirradiated, environment-only or assay blank.
6. "Cosray: no clinical data on either side": there is an uncontrolled tally with wrong totals and no controls.
7. "USDA request for clinical testing": it was a build order.
8. Hayes's "failed rebuild" and Knudsen's hidden-wire suspicion were missing from the first review.
9. 119,360 W (facts.json) should be 119,360,000 W; Moray's own per-person figure is 1,193,600 lamps, the sons' 1,693,600.

Source honesty: remaining [memory] items

• Lamp brightness exponent (about 2 against power) [memory, medium; not web-checked]; flat-iron mass and specific heat [memory, low-medium]. Both only feed N2, whose conclusion is qualitative.
• Global fair-weather current of about 1,800 A [memory, medium]; the fetched page confirms the 100 V/m field and 400 kV potential only [web https://en.wikipedia.org/wiki/Atmospheric_electricity]. The conclusion (eight orders below 1 kW) is not sensitive to a factor of 10.
• Coulomb barrier of order 100 MeV, binding energy per nucleon, Seaborg 1980 Bi-to-Au, radiometric "equivalent U3O8" behaviour, fire-assay flux gold, Pb-210 half-life 22 years [memory, medium to high]; verdicts on 5A and 5B do not rest on any one of these alone.
• Knudsen's affiliation: the book says Stanford; he was an acoustics physicist closely tied to UCLA [web https://en.wikipedia.org/wiki/Vern_Oliver_Knudsen]; the book's affiliation is probably an error, not verdict-bearing.
• Lead-acid 35-40 Wh/kg is now [web https://en.wikipedia.org/wiki/Lead%E2%80%93acid_battery].

Verdict changes (summary)

• Claim 1: from "Never properly tested; strong form Refuted" to a four-way split: lamp-count demonstrations Never properly tested (not "never observed above 1.4 kW"); a stored-energy explanation of the 158-hour test Plausible, untested as an argument (strong for 17 kWh, silent on a hidden feed); the Strawberry remote test not exclusive of a 0.6 kg battery; 50 kW unwitnessed and Refuted if from known radiation.
• Claim 2 and Claim 6: Never properly tested remains, but "dismissed without testing" is replaced by "tested by friends and some experts, neutral whole-device test offered and not completed" (Fletcher test unresolved; Hayes rebuild failed, non-neutral); patent-ground wording corrected.
• Claim 5: split into gold (mechanism Refuted; assays Never properly tested, thin internal control, inconsistent statistics), uranium (new; consistent with ordinary concentration, Never properly tested, transmutation Refuted), Cosray (Never properly tested; the tally is arithmetically inconsistent). Claims 3, 4 and 7 keep their verdicts.

---

Investigator findings (2026-10-04)

Inputs weighed: sites.md "Leads for review" 1-28 (runs 1-3), the 29 research/ notes dated 2026-10-03 and 2026-10-04, and research/files/ (printed 1944 Senate REA hearing text, Chicago Tribune and Electrical World excerpts, Bearden's Wayback page, the Internet Archive compilation with the 1945 booklet and 1946 therapy treatise, FBI log excerpts, NARA index-sheet images, Utah newspaper pages, finding aids). My earlier review mentioned none of them, so all are weighed here. Method: for each finding, the claim it bears on, what the arithmetic or logic says, and whether a verdict moves. Tags: [read] = in the research files I read this round; [web] = URL; [memory] = recalled, confidence stated; [calc] = computed here (venv Python). The Investigator reports leads, not verdicts; everything below is my weighing. Material is kept whatever my judgement of it.

A. Source-honesty frame for these findings

• The government record is by interested parties. The REA investigators' report (V. A. Young's staff) is part of a hostile inquiry into Administrator Slattery and frames the projects as "revolutionary and fantastic ... followed through to negative conclusions" [read: HT_1944 pp.1344-1345]. Slattery's account is second-hand from R. B. Craig [read: pp.1370-1376]. Electrical World (11 Nov 1944) is the private utilities' trade weekly [read]. Bearden has his own scalar-vacuum theory [read]. Moray's side (family books) is also an interested party. I report each as that party's characterization.
• Single-source items stay in the record: Tugman's letter is documented in one printed source (the 1944 report, quoting a May 1940 letter; the letter itself is not in the folder). Slattery's rumour of a German assassination attempt is documented in one printed source; no press story has been found, and silence in the Utah newspapers searched (poor OCR, later-1940 pages unopened) is not contradiction.

B. Findings that bear on the mathematics or the energy accounting

B1. Wood's August 1939 report and the REA payroll (Claims 1, 2, 6; leads 1, 10). [read: Chicago Tribune 11 Dec 1944; HT_1944 p.1344-1345]. REA engineer Franklin P. Wood (probably the Denver AIEE member of 1919, identity not proven) wrote after visiting Salt Lake City: "unmitigated ass, crackpot, nitwit, religious fanatic, essentially honest, unable to originate or perpetuate a hoax or fraud. In some inexplicable way he has, I believe, stumbled onto something that he himself cannot understand ...". The Salt Lake Tribune printed only the favourable half. Weighing: this is a trained engineer's impression of Moray's honesty and of the device's behaviour, with no number in the quoted text. It does not move any verdict on the arithmetic. It does bear on N5 (hidden-feed hypothesis): Wood saw a 1939 demonstration and, as far as the quote goes, did not say he found a feed; nor does the quote say he looked for one. It cuts against "hoax by Moray" and is silent on "self-deception or an unrecognised feed". The investigators' own sentence that Moray "unsuccessfully demonstrated on numerous occasions" is their characterization, set against the family's "Crim's tests entirely satisfactory" [read: 4th ed. pp.174-176]; both are one-sided and neither gives a measurement. Payroll arithmetic: 17 Feb 1939 to 28 Feb 1941 is 742 calendar days, so at most $18,550 at $25/day [calc]; REA also bought equipment. "For the next 2 years Mr. Frazier spent considerable time ... assisting, begging, cajoling, and threatening Mr. Moray to develop the radiant energy device, all to no avail": two years of paid access by the government's own engineers and no delivered, testable device is a documented fact in an interested party's words. It is exactly the situation "Never properly tested" describes (opportunity for a neutral test existed and no test result is on record). Slattery says Moray was still on the roll in Nov 1942, against the report's Feb 1941; unresolved and not mathematical.

B2. Tugman (University of Utah physics head), May 1940 (Claims 1, 2; lead 11). [read: HT_1944 p.1345]. Twelve years of requests for a demonstration met excuses; demonstrations were "made to men not fitted to be critical"; an agent was told "to stay away from physicists and engineers"; "a very clever hoax". Weighing against the folder: it agrees with the record that the detector was withheld in every examination but one and that several physicists (Millikan 1929, Bell Labs 1937, Eyring) were not shown the whole device [read: 5th ed. pp.79-102]. It is in tension with one part: Harvey Fletcher (a physicist, Bell Labs) did attend the 1928 test, and A. B. Jensen (battery expert) the 1938 one, so "physicists were kept away" is too strong as stated; Tugman may mean himself and the university. Tugman had his own stake (stock sold to Utah investors, university reputation). His letter reports refusals, not a failed test. Effect on my Claim 2 verdict: none in label ("Never properly tested"), but the sentence "not dismissed without a look" is now supported on both sides: looked at by friends and some experts (family record), refused to a named qualified sceptic for 12 years (Tugman). I also note the arithmetic of time: 12 years before May 1940 is 1928, the year of the Fletcher test.

B3. Slattery's second-hand "small aluminum disk ... run a small motor and fans by current, unseen current" (Claims 1, 2; lead 12). [read: HT_1944 p.1371]. Weighing: Craig's description via Slattery, with no date, power or instrument. New and mathematically interesting only in one respect: a motor load. All the lamp and iron accounts fit a high-frequency output (series-capacitor test, N3). An ordinary small induction or universal motor presents an inductive reactance X = 2 pi f L; with an assumed L of 0.01-0.1 H, X is 6-63 kilohms at 100 kHz [calc, L assumed], so it would draw almost nothing from a 100 kHz supply and would not run. If a motor and fans really ran (the 1938 Securities Commission test also lists "a motor" [read: 5th ed.]), the output feeding them was either at a low frequency or rectified, or the motor was unusual. That is not consistent with the simple "HF only" reading of N3, and none of the accounts gives output frequency or waveform. Treat as an open inconsistency in the record, not as evidence against: second-hand, undated, and "small" (watts, not kilowatts). It also strengthens the point that a motor with a measurable shaft load would be a far better power test than a lamp count (N2), and no such measurement exists.

B4. The 1931 "nationally known physicist" and the slowly filling condensers (Claims 1, 3; 1945 booklet, IA compilation). [read: IA compilation lines 1381-1395]. The unnamed physicist noted that "the condensers fill slowly ... much as in filling a bucket", unlike ordinary capacitors, and that lamps were hotter at one spot due to gas incandescence. Maths: a capacitor charged through a source resistance R_s rises as 1 - exp(-t/(R_s C)); slow filling means a large R_s C, i.e. a high-impedance source, which is ordinary, not anomalous. Example [calc, assumed values]: a 0.025 microfarad condenser (the size in the 1938 capacitor test) charged to 2.4 kV holds E = 0.5 C V^2 = 0.072 J and Q = 6e-5 C. The fair-weather collection current for a 10,000 m^2 collector is about 3.5e-8 A (N4), so charging takes 6e-5 / 3.5e-8 = 1,700 s, about 29 minutes. A time constant of 100 s would need R_s = 100 / 2.5e-8 = 4e9 ohms. A slow-filling condenser therefore indicates a high-impedance, low-current source, consistent with N3 (lead current well under 0.1 A) and with Tesla-type collection (B7), and says nothing about power. It also fits the stored-charge-priming reading of M1 (magnet stroking 4-10 min). It does not support kilowatt output; a high source impedance limits the maximum power transfer to V^2 / (4 R_s) (matched load), which for 2.4 kV and 4e9 ohms is 0.36 mW [calc]. The physicist's name was not given in the booklet and I found no second account.

B5. How the figures grew in later retellings (Claim 1; leads 20, 26, Google Books note items 3, 7). Same test, different statements [calc, lead-acid 37 Wh/kg, midpoint of 35-40 [web https://en.wikipedia.org/wiki/Lead%E2%80%93acid_battery]]:

(The Indian Science Congress phrase gives "lights" with no wattage; the 100 W lamp is my assumption, and the ratio is for that assumption.) The first row is the only figure with witnessed seals and durations. The later rows turn an argument that is arithmetically strong at 18 kWh (N1) into one that is absurd at 1.5-7.9 MWh, so inflation actually weakens the case against batteries not at all (nobody hides 200 t) but weakens the credibility of the retellers. This confirms my M5 point that 50 kW has no witness. Bearden's other statement, that independent investigators took the device "completely apart several times", conflicts with the folder's record (detector withheld in every examination but the 1938 inspection) [read].

B6. The Moray detector, the Army "sound detector", and Losev/Lilienfeld (Claim 4; leads 13, 21, 25). [web NARA catalog images; read: research notes]. The War Department Adjutant General index sheets of 14 Dec 1936 file Brig. Gen. Shinkle's report as "Invention of sound detector by F. [T.] Henry Moray", cross-filed under Moray's attorney K. K. Steffensen; in Feb 1939 a Magdiel enquiry (via Sen. Elbert D. Thomas) calls it a "demonstration of Radiant Energy device". The family text agrees the 1936 test was of the "Sound Detector Device" driving loudspeakers. Weighing: an independent record that an Army officer saw a demonstration; the report itself is not in hand. A receiver driving a loudspeaker needs milliwatts, so this test is consistent with an ordinary detector plus amplification (which needs a bias source) and bears on Claim 4's "valve" at the receiver level, not on the power claim; I do not read it as supporting or refuting Claim 1. Losev's 1923-24 biased zincite negative-resistance amplifier/oscillator [web https://en.wikipedia.org/wiki/Oleg_Losev] and Lilienfeld's US 1,745,175 (1930) [web https://en.wikipedia.org/wiki/Julius_Edgar_Lilienfeld] make my earlier line "A germanium detector in a 1920s box is possible; 'preceded the transistor' requires current gain" more specific: a 1920s crystal device with current or power gain did exist (Losev's, with a battery bias). So a Moray valve with gain is not mathematically impossible; the power gain relies on a DC bias supply, and the net-power claim still needs a source. Verdict on Claim 4 (i) stays "Plausible, untested"; (ii) "preceded the transistor": the priority claim is weakened (others were earlier) but "Never properly tested" stands for Moray's own device.

B7. Tesla's US 685,957 (1901) is the same antenna-ground-condenser principle (Claim 3; lead 24). [web https://patents.google.com/patent/US685957A/en, per the Investigator note]. Elevated plate and ground charge a condenser that is periodically discharged through a load. Tesla's mechanism: positively charged solar particles. Weighing against numbers: (a) atmospheric-electric collection gives about 8e-5 W for 10,000 m^2 (N4). (b) Solar wind at Earth's orbit: speed 250-750 km/s, 3-10 particles/cm^3 [web https://en.wikipedia.org/wiki/Solar_wind]; with 5 /cm^3 at 400 km/s (assumed mid-values) the proton flux is 2e12 per m^2 per s at about 0.8 keV each, an energy flux of 2.7e-4 W/m^2 [calc], so even 100% capture over 10,000 m^2 would give about 3 W; and these particles do not reach the ground (stopped far above it by Earth's magnetic field and atmosphere [memory, high]), so this is an upper bound for the particle idea in space, not at the antenna. That is the same order as the cosmic-radiation bound in Claim 3 (4.8e-5 W/m^2 maximum); it does not rescue kilowatts. Lineage, not support: Moray's book quotes Tesla's "wheelwork of nature" [read: IA compilation]. Claim 3 verdict unchanged.

B8. Correction to my N4 radio bound (Claims 1, 3). N4 divided the power by the physical area of a collector. The better bound for a resonant antenna in a radio wave is the effective aperture A_e = G lambda^2 / (4 pi), with G = 1.64 for a half-wave dipole [web https://en.wikipedia.org/wiki/Antenna_aperture]. At 1 MHz (lambda 300 m) A_e is about 1.2e4 m^2, larger than any antenna's physical area, so tuning does enlarge the "area". Capture is then P = S A_e with S = E^2 / 377 [calc]:

This is an upper limit assuming a lossless, matched, correctly sized antenna. An 80-ft wire is much shorter than a wavelength at these frequencies, so its real aperture and efficiency are far below the bound [memory, medium: electrically small antennas]. The correction changes my N4 sentence "1.4 kW needs about 5e5 m^2 of ideal collector". Correct statement: 1.4 kW from a plane wave needs about 5e5 m^2 of physical area, or an A_e of that size, which a half-wave dipole reaches only below about 150 kHz [calc: 1.2e4 x (1 MHz/f)^2 = 5e5 at f = 150 kHz], and then needs about 1 V/m, a field found only within kilometres of a transmitter. So resonance does let a small antenna capture more than its geometric size suggests, but the field strength, not the area, then limits the power. Typical broadcast fields far from a station are mV/m [memory, medium] (microwatts to milliwatts). For Strawberry (26 miles from the last power line) and Dugway (25-30 miles from Johnson's Pass) the relevant question is distance from a transmitter, which the accounts do not give. Consequence: Nuenergy's alternative (spark-gap stations to 1928) is not excluded by area alone; it fails because 1.4 kW needs a field near 1 V/m at long wavelength, and because the tests at remote sites show no distance to any transmitter. Verdicts unchanged; the argument is now correctly stated. Bearden's "subharmonic resonance across 29 stages" adds no energy source: stage-by-stage resonance multiplies voltage (Q), not available power, and is bounded by P = S A_e per stage input at best.

B9. Cosray applicators were radioactive, by the family's own 1946 text (Claim 5C, Claim 4; lead 19). [read: IA compilation lines 4690-4748; copyright 1946 [web Catalog of Copyright Entries per the Investigator note]]. The "Short Treatise on Alpha, Beta, Gamma Ray Therapy" says the applicators emit alpha, beta and gamma rays, "will penetrate through the human body (proven by Geiger Muller Counter tests)" and "1/2-inch thickness of lead", have a half-life "about 1600 years", yet "will not injure healthy tissues because of the internal phantom characteristics" and need "no special shielding". Weighing as logic and arithmetic: - Half-life 1,600 years is radium-226's [web https://en.wikipedia.org/wiki/Radium]; the specific activity of Ra-226 is 3.7e10 Bq per gram [calc: ln2 N_A / (T_half x 226 g/mol)], the definition of the curie. So the text points to radium, as the Investigator noted. - Alpha and beta particles cannot pass through the body or half an inch of lead: alpha particles of a few MeV are stopped by tens of micrometres of tissue [memory, high]; only the gamma rays of the decay products (which Wikipedia says accompany almost every emission in the chain) penetrate. The sentence "the rays penetrate through the body and half-inch of lead" can be true only of the gamma component. A source that penetrates 1.3 cm of lead and the body cannot at once be harmless to "healthy tissue" and need no shielding: penetration means energy deposited along the path, and lead shielding would be needed. "Internal phantom characteristics" is not a physical quantity with any definition given. - The text says "the second problem [measuring depth dose] does not exist", which is the reverse of radiotherapy practice: dose at depth is the problem to be measured [memory, high]. Verdict effect: for Cosray the clinical claim remains "Never properly tested" (an uncontrolled, internally inconsistent tally), and I add a local finding that the 1946 radiation description is Internally inconsistent (penetrating radiation that is also harmless and needs no shielding). It adds a safety-relevant fact the previous review lacked: the device was described by its makers as a radium-like source, which fits the FDA's later seizure but is a different question from efficacy. It also reopens the question of the valve: the family's radium-chloride valve recipe and Perreault's "no radioactivity" in the surviving tubes concern different objects; a gamma-spectrum measurement of either would test the statements, and none is on record.

B10. The 1945 booklet's transmutation letter and the early Hagmann trail (Claim 5; leads 7, 22). The 1945 booklet already quotes a September 1943 letter on transmutation of metals (catalytic "growing" of metals from colloids by synchronised current and rays) [read: Investigator note on IA compilation; line 1668 mentions the cyclotron mercury-to-gold report]. So the gold idea predates the 1963-70 tests by about 20 years. Mark J. Hagmann is very probably the BYU-trained physicist who lectured on historical transmutation cases in 1966 [web alkahest.com bulletin No. 29 via pdfcoffee, per the Investigator note] and was sued in 1970 by Research Institute and Moray for return of papers (case 195569) [web Daily Record 25 Sep 1970 p.1; the PDF text in research/files confirms "195569 - Research Institute, Inc. and T. Henry Moray vs MARK J. HAGMANN; Pltf - return of papers"]. Identity is not proven. Weighing: no number from these items bears on my gold arithmetic (3.19% gold; 1.25e8 J per gram of gold-bearing residue against 1.6-200 J/g delivered, claims.md). A technically trained collaborator is a possible witness to the blank and control questions I raised in 5A; his papers are the sort of record that would decide them. No verdict change.

C. Findings that bear on Claim 6 and the historical record (no mathematics, weighed for the record)

C1. REA and Technocracy (Claim 6; leads 3, 10, 12, 26). Printed: Carmody approved an inquiry in fall 1938 "long enough to get a clear view"; Frazer (WPA investigator, "reputedly an advocate of technocracy") and Wood ("also a believer") were involved; the Wood report is dated August 1939 [read]. Felix J. Frazer served subpoenas for the La Follette Committee in 1937 [web HathiTrust, per note] and was probably the Technocracy economist later of the Builders of the Adytum (identity not proven). Bearden's alternative, a Soviet-infiltration reading with an REA man "Z", is his characterization and rests on a Salt Lake Tribune of 2 Dec 1941 not yet read. The printed record fits the Technocrat reading better than a utility or intelligence one, but all three are inferences from names. Effect on Claim 6: it moves the REA part of "blocked or ignored" from "one-sided family account" to "documented in the government's printed record, framed by a hostile staff"; it shows the government employed Moray for two years rather than blocking him, and that the Senate record's tone is mockery of REA, not a technical rebuttal. Electrical World (private utilities' trade press) ridiculed the episode ("retailed perpetual motion to REA at $25 a day") and adds no test data [read]. Verdict on 6 (iii) below.

C2. The 1940 shooting (Claim 6; leads 12, Utah newspapers). Slattery (June 1943): "a mysterious story went around that the Germans had tried to assassinate him, and that story got into the Salt Lake papers" [read: HT_1944 p.1371]. One printed source, second-hand, a rumour as he reports it, and it is the first source outside the family's books for an attempt on Moray's life. Utah Digital Newspapers searches (about 25 queries) have not found the story; the Tooele paper of 26 Mar 1940 shows Moray speaking at a funeral about three weeks after the family's date of 2 Mar 1940 with no mention of an attack [web, via note]. Both facts stand: a corroborating rumour in 1943, and no press story found. Later-1940 pages were not read. Not testable as stated, unchanged; the Slattery mention is now in the record.

C3. Business and court records (Claim 6; leads 4, 5, 6, 15; silence E). Third District (Salt Lake) civil cases against Moray: Ashton-Jenkins Co. (case 46754, $2,500 on mortgage, Dec 1930), Brewster/Progressive Heating (65983, 1941), Home Owners' Loan Corporation (67422), Sanford (67845, 68210; $1,290 judgment 27 May 1942 for non-delivery of stock in "a corporation to use radiant energy in aircraft"), Green/Christensen (80824, 1947-48), Roy Ware (104172, 1954-56); Moray Manufacturing Company no. 022465 (Utah, struck off); Washoe County, Nevada "In re Moray Products Company" (notice Apr-May 1932; no online index) [web, per notes]. They give debt litigation across 25 years, consistent with an under-capitalised venture and with a stock-selling pattern that Tugman, Sen. Thomas and a Dun & Bradstreet report (per the Senate record) also describe. They say nothing about whether the device worked. The Hayes testimony that the patent-paper rebuild failed (my Claim 1 weighing) is expected in the Washoe file, still unobtained. A 1930 Salt Lake Telegram list of "ten greatest Utahns" including Moray and a 1949 Ogden report of visitors told the "cos ray machine" treated cancer and that transmutation work was "being proved at Harvard" are newspaper statements of belief in circulation, not tests [read: UDN pages; OCR poor]. A 1942 Ogden letter by J. G. Teuscher says "investigated Moray's invention" and public statements on its merits (OCR garbled): not relied on.

C4. Federal files and the FBI record (Claim 6; leads 2, 14; silence B). The FBI FOIA log 2000: request 917273, subject MORAY, T HENRY, 74 pages reviewed, 72 released, requester Ravnitzky; a second request 924753 closed for lack of a notarised consent; the Q1 2013 log answers a 2011 request (1178194-002) "No Records" [read: research/files excerpts]. Silence B is the contradiction between a 74-page release in 2000 and "No Records" in 2013. Competing explanations (destruction under a schedule; a main-file index versus cross-references; the "-002" scope; different spelling) are not decidable from the logs; no release is posted on the FBI Vault, MuckRock (0 requests), governmentattic or theblackvault. CIA request F-2000-01342 (22 Jun 2000): outcome not posted; the CIA reading room holds only the FTD translation of a 1974 UFO-Nachrichten piece by "Henry Moray", already in the folder. The FDR Library Carmody finding aid lists REA Chronological File Boxes 102-103, card indexes 91-94 and Box 99 "Senate Hearings [1944]" but names no Moray [read: finding aid note]. Univ. of Utah/Archives West: no finding aid naming Moray. Chronicling America, Swedish newspapers (queries for "Henry Moray", "Cain Larson", "Moray Upsala", 1900-1990): negative. Nothing here changes a verdict. These are the records that would decide Silences B-E, and I keep them listed as pending.

C5. Records the Investigator located but I have not read (kept in the record). (a) University of Ottawa, Arthur Bray fonds (30-003), Wilbert B. Smith files: S1-SS5-F4 "Correspondence: 'Radiant Energy' Invention (by Dr. T. H. Moray)", 1928-1956; F24 Marcel C. Schiess 1928-1957, "correspondence and reports", "witness accounts of the machine in use", technical drawings; S1-SS4-F4 two letters on a demonstration "reducing cosmic energy to electrical energy"; F6 Carl H. Betz 1945-54; F8 Buchli 1957-58; F17 Hooper 1958-59 [web arcs-atom.uottawa.ca per note; none digitised]. Because they reached a Canadian government engineer outside the family's editing, they could show whether the books quote the 1928-38 letters in full, and could hold the original of the Fletcher-to-Schiess letter on which the disputed "Fletcher-arranged test" rests (Claim 6). (b) A Brief Analysis of the Moray Radiant Energy Device: What it Is, Possibilities, Testimonies, 50 pp., undated, by T. H. Moray [web Google Books lE4FNQAACAAJ]: probably the first printing of the witness testimonies. (c) J. Cecil Alter, Utah, the Storied Domain (1932) pp.439-440, a biography "Utah's electrical wizard, inventor, scientist, editor, poultry expert", which may show what education or degree was claimed in 1932 (Claim 7); the snippet does not show it. (d) A 1943 date for the 20-page Radiant Energy in a 1950 UN list against the copyright date 23 Dec 1945 [web]. (e) Copyright entries: 1946 therapy sheet (c 11 Jul 1946), 1956 55-leaf "Radiant energy" with new matter (c 15 Jan 1956): the folder's printing citing post-1945 physics is probably the 1956 one, which matters for dating statements inside it. (f) Okuto and Crowell papers: located and citations confirmed (Phys. Rev. B 6, 3076, 1972; 10, 4284, 1974; Solid-State Electronics 18, 161, 1975); none free; they concern impact ionisation and its energy conservation and are peripheral [web crossref]. (g) The Army AGO file 413.68 and 095 Steffensen; NARA RG 221, RG 46 and RG 241; Washoe, Salt Lake County and Utah State Archives case files; the Hagmann 1970 file. None obtained.

C6. Replications (Claims 1, 4). Hobby videos (2016-2025), a Reddit "blueprint", an open "specification (V1)" on the Internet Archive and a "Manual of Free Energy Devices" were found; none reports metered input and output [web, per Google note]. I treat this as the state of the record: no replication with measurements exists in anything found. It does not close the claim, because the original's detector and the excitation are unknown; it does mean "replicated" cannot be answered "yes" anywhere. The only recorded copy attempt remains Hayes 1932.

D. Weighing by claim (a-d), with verdict changes

• Claim 1. (a) The inflation table (B5) and the motor point (B3) add inconsistencies to M5; no new internal contradiction in the 1928 record. (b) N4 corrected (B8): resonance enlarges the effective aperture of a small antenna, but power is then limited by the field strength; kilowatts need about 1 V/m at long wavelength, i.e. near a transmitter. B4 shows slow condenser charging is ordinary for a high-impedance source. (c) New tests: none beyond Wood's 1939 visit (no number), the Army's 1936 sound-detector demonstration (not the power claim) and Craig's motor description (second-hand); no replication. (d) Unchanged. Verdict: no change to the four-part split. Wood and Craig add observers, not measurements; sub-verdict (i) "Never properly tested" is further supported by two years of paid government access ending with no test report on record.
• Claim 2. Tugman (B2) and the REA record (B1) add one named physicist's refusals and one engineer's favourable-on-honesty, silent-on-feed impression. No change ("Never properly tested"). I add the sentence: a qualified sceptic (Tugman) says he was refused a demonstration for 12 years; Fletcher and A. B. Jensen are on the other side of the record; neither side gives a neutral custody test.
• Claim 3. Tesla 1901 (B7) and Bearden (B8) are restatements, not tests. Solar-wind bound about 3 W per 10,000 m^2 even in space; atmospheric bound 8e-5 W; radio bound corrected (B8). No change: strong form still Refuted by calculation; the resonance claim still Not testable as stated and, as printed, Internally inconsistent. The Investigator's finding that Tesla's principle is the same is correct and I keep it.
• Claim 4. Losev and Lilienfeld (B6) show 1920s solid-state gain existed with bias. The Army record concerns a "sound detector". No change in labels. (ii) "Preceded the transistor": the novelty is shared with others, wording refined.
• Claim 5. (iv) Cosray: add a local verdict "Internally inconsistent" for the 1946 statement that the radioactive applicators penetrate body and lead yet harm nothing and need no shielding (B9). Clinical efficacy remains "Never properly tested". (ii) and (iii) unchanged; the 1943 letter and Hagmann trail (B10) give no new numbers.
• Claim 6. Revised split: (i) patent rejections Established (unchanged). (ii) "Dismissed without testing": still not supported. (iii) REA: the employment (17 Feb 1939 to 28 Feb 1941, $25/day), the Wood report (Aug 1939), Tugman's letter and Frazer's role are now Established as documented facts in a printed Senate record, which is a hostile party's document; the family's "REA tests entirely satisfactory" and the investigators' "unsuccessfully demonstrated" are both unmeasured characterizations. "Suppressed because it worked": Not testable as stated; the printed record shows the government paid for and observed it for two years, then it ended amid a political attack on REA and Senator King's and Senator Thomas's complaints. (iv) Shooting: Slattery's rumour adds one second-hand mention; Not testable as stated. The disputed Fletcher-arranged test stays unresolved; the Ottawa Schiess file is the most promising record.
• Claim 7. No change (Never properly tested). Swedish newspapers searched under "Henry Moray" and the alias "Cain Larson": negative; a newspaper silence is weak evidence for a student. The Alter 1932 biography is the next record to read.

E. Source honesty for this section

• [memory, medium] Electrically small antennas have far less than the ideal aperture; typical distant broadcast fields in mV/m; solar-wind particles do not reach the ground (high). Verdicts do not rest on any of them.
• [web] checked this round: solar wind speed and density (https://en.wikipedia.org/wiki/Solar_wind; flux and energy flux are my calculation from mid-range values, not on the page); effective aperture formula and dipole gain 1.64 (https://en.wikipedia.org/wiki/Antenna_aperture); Ra-226 half-life 1,600 y and that nearly every emission in its decay chain is accompanied by gamma rays (https://en.wikipedia.org/wiki/Radium). The alpha-range statement is [memory, high]; the Wikipedia page fetched gave no range.
• Items taken from the Investigator's notes without my own re-fetch: the Google Books snippets (Indian Science Congress 1934, Alter 1932, Rosicrucian Digest 1956), the IEEE biography of Hagmann, the Utah court indexes, the NARA image identifications, the Ottawa catalogue. I read the saved copies of the Senate hearing text, the FBI log excerpts, the Chicago Tribune and Electrical World excerpts, the Bearden text, the IA compilation passages, the UDN page texts (OCR poor) and all research notes.
• Not weighed because not available: the Wood report in full, the Shinkle report, the 74-page FBI file, the Washoe and Utah case files, the Ottawa files, and the physicist's name in the 1931 demonstration.

F. Verdict changes from this section (summary)

• Claim 5 (Cosray): added "Internally inconsistent" for the 1946 radiation description (penetrates body and lead, harmless, no shielding); clinical claim stays Never properly tested.
• Claim 6: REA episode facts (employment, Wood report, Tugman letter) raised from one-sided family account to Established as documented in a hostile printed Senate record; "dismissed without testing" and "suppressed" unchanged (not supported / not testable as stated).
• Claims 1-4, 7: labels unchanged; N4 radio bound corrected to the effective-aperture form (kilowatts need about 1 V/m at long wavelength), and the Moray-figure inflation table (18 kWh primary versus 1.5-7.9 MWh retellings) added.

---

Investigator findings (2026-10-08)

Inputs weighed. sites.md "Leads for review" 29-43 (follow-up leads of 2026-10-06, 10-07 and 10-08), which my 2026-10-04 section did not cover (it weighed leads 1-28), and the eight research/ notes dated after it: Q01 (Rutherford sentence), Q02 (infrared paragraph of the 1968 proposal), Q04 (sulphide screens), Q08 (resonance and matter), Q09 (two notes: the orchestrator's note of 10-07 on the hypotheses put to the review, tagged by its author as unresearched, and the researched note of 10-08), Q11 (thorium "10,000 times") and Q12 (Le Bon's amalgamated tin). research/files/ holds the same items as on 10-04 (nothing added since); the saved pages in research/pages/ were used by the Investigator and are not republished. Method as before: for each finding, the claim it bears on, the arithmetic or logic, and whether a verdict moves. Tags: [read] = in the research notes or project text as quoted there; [web] = URL; [memory] = recalled, confidence stated; [calc] = computed here in the venv. The Investigator reports leads, not verdicts; the weighing is mine, and every finding stays in the record whatever I think of it.

A. Source-honesty frame

• Most of these findings are the Investigator's matches of Moray's pages to earlier published texts. They are [web] comparisons made against Rex Research's transcription of Le Bon (Legge translation) and Google Books snippets. A transcription can differ from the printing Moray used (Q11 itself notes Rex inserts "[radon]"), so "nearly verbatim" is the Investigator's reading of two texts, not a collation of editions. I did not re-fetch Le Bon; I rely on the quoted passages in the notes.
• Foster 1990 (lead 42) is one participant's account, relayed in a compilation by an author who advocates Moray's device (Perreault, Nuenergy 2001 p.30). It is evidence and is weighed; no second source and no original letter has been found, and silence elsewhere is not contradiction.
• Plauson's patent text (lead 39) is the patentee's own description. Campbell (1906) was a rival worker on weak natural radioactivity. Ramsay and Spencer first read the effect as atomic disintegration and Le Bon advertised that. The 10-07 Q09 note records hypotheses put to the review, with numbers that were [memory]; the 10-08 note checks them.
• Copying someone else's text is a fact about authorship, not about whether the physics is right. I weigh the physics of each passage on its merits and report the provenance beside it.

B. Findings and their arithmetic

G1. The "experiments of Rutherford ... taken up again" sentence and its numbers are Le Bon's of 1905 (Claim 5; leads 29, 30; Q01). [read: Q01; web https://www.rexresearch.com/lebonmat/lebonmat.htm]. Radiant Energy printed pp.53-56 (and the later editions) follow Le Bon, Book IV Ch.III, with "I think" dropped and the footnote (a detection-threshold idea after Strutt) dropped. Checks [calc; alpha mass 6.645e-27 kg]: | Speed | Kinetic energy | Comment | |---|---|---| | Le Bon/Moray "30,000 km/s" (one-tenth of c) | 18.7 MeV | said to be "very difficult to explain" | | Rutherford 1903, "swiftest" 2.5e9 cm/s | 13.0 MeV | Rutherford 1908 lecture [web https://www.nobelprize.org/prizes/chemistry/1908/rutherford/lecture/] | | Ra C' (Po-214) alpha, 7.69 MeV [memory, high] | 1.92e7 m/s, 19,200 km/s | modern value | | Th C' (Po-212) alpha, 8.78 MeV [memory, high] | 2.05e7 m/s, 20,500 km/s | fastest natural alpha | | Rutherford's "critical speed" 9,000 km/s | 1.7 MeV | where alphas vanish from detection | So the "puzzle" of an alpha at 18.7 MeV is a puzzle about an overestimated speed: no natural alpha exceeds about 8.8 MeV, and a 8.8 MeV alpha is explained by the energy released in the nuclear decay (a quantity not known in 1905). Likewise "70,000 alphas per second per gram for uranium and thorium": uranium-238 alone gives 12.6 kBq/g [web https://en.wikipedia.org/wiki/Uranium-238, from "1 mole emits 3e6 alphas per second"; derived here]; with all eight alpha emitters of its chain in equilibrium that is about 1e5 per gram [calc]; thorium-232 gives 4.1 kBq/g and about 2.4e4 with its six alphas [calc, half-life 1.405e10 yr, memory high]. So the figure fits equilibrium uranium within a factor of 1.4 and overstates thorium about threefold. These are 1905 numbers printed in a post-1945 text, with no update; Moray's "1926" printing, which cites pile physics, sits on top of them (internal dating inconsistency, already noted by the Investigator). The reading of the sentence is settled by Q01 (Rutherford's 1902-06 velocity measurements); the three readings of why Moray kept it (A copying, B adopting Le Bon's thesis of sub-threshold emission, C a bridge to the 1968 direct-conversion proposal) are inference. My arithmetic bears on B: Le Bon's detection-threshold footnote is arithmetically possible (alphas below 1.7 MeV do exist, in the Bragg tail), but ordinary matter emitting them at rates that matter is a separate, measurable claim; Moray's text gives none. Lead 30 (Moseley's 1913 radium cell as the "1913" of Safenev) is dealt with under G7.

G2. The infrared paragraph of the 1968 proposal is correct detector physics, and does not point at the microwave background (Claim 4, Claim 3; lead 31; Q02). [read: Q02]. Checks [calc]: the CMB at 2.725 K peaks (Wien, 2898 um K / T) at 1,063 um, photon energy about 1.17 meV. Ge:Hg's impurity level of about 0.09 eV gives a cutoff of 13.8 um [1.2398 eV um / 0.09 eV]; the passage's range ("beyond 6 microns") is about 100 times shorter in wavelength than the CMB peak. The proposal's logic, that a narrower gap responds to longer wavelengths but is also crossed by thermal excitation, is right, and a rule of thumb makes it quantitative. HgCdTe works at 77 K with a 10 um cutoff (gap 0.124 eV), where gap/kT = 18.7; keeping gap/kT about 18 for a 1.17 meV gap needs T of about 0.75 K [calc; rule of thumb from the 77 K HgCdTe case, my construction]. So even a hypothetical intrinsic CMB photon detector would need sub-kelvin cooling, not "reasonable cooling". The passage is internally consistent; whether it was borrowed from a 1960s review is open (no source found). Plasma frequency check: 8.98 kHz x sqrt(5) = 20.1 kHz for 5 electrons per cm^3 [calc], as the note says. Nothing here supports or weakens a verdict: it is competent background, followed (p.1) by the unsupported sentence that the Institute's "energy transponders" detect, count and convert.

G3. The sulphide-screen passages are Le Bon's, with one Moray sentence (Claim 5; Claim 4; leads 32, 33; Q04). [read: Q04]. "Herein be the door to greater discoveries" and "Herein may lie a prototype of the action of the universe" are Moray's additions; the rest, down to the raindrop image (Crookes) and the 18-month calcium-sulphide sentence, is Le Bon's, changed from first person to impersonal ("I have succeeded" becomes "One may ... succeed"). Mathematically: nothing is quantified, so there is nothing to check except the logic. A zinc-sulphide screen scintillates under alphas from uranium and thorium minerals and mantles, which Le Bon's footnote says; that is ordinary alpha radioactivity and is not evidence of "dissociation" of non-radioactive matter without a blank. The experiment Q04 suggests (screen on a thoriated mantle, a uranium mineral and a non-radioactive control) is the missing controlled comparison. The photostimulable storage (CaS:Eu,Sm read by infrared) is real; a storage phosphor releases at most what it absorbed (see the bound under G7, Foster). The change of "energy" to "matter or particles" in the 1958-59 booklet is a change of a quantity's category, not a mathematical result.

G4. Resonance "breaking matter": the booklet's mechanism fails the energy-per-quantum and the oscillator arithmetic (Claim 5, Claim 3; leads 34, 35; Q08). [read: Q08; the passages are quoted in the note and checked there against page images]. The borrowing: Isotopes pp.20-23 follow Glasser (1944) on X-ray ionization nearly word for word; p.25 turns Preston's description of an oscillating Leyden-jar discharge into "disintegration and transformation" of atoms; the silver-to-cadmium example and the "superultra sonic reaction" have no earlier source found. My checks [calc]: 1. Oscillator. A series RLC discharge oscillates with decay rate alpha = R/2L; successive peaks fall by the factor exp(-pi alpha / omega_d). "The first discharge will over-run itself and severe recoils set in" describes one reversal, and each reversal is smaller than the last. A free damped oscillation cannot grow, so "if these oscillations become violent enough, disintegration ... will take place" needs an external drive or negative resistance, which the passage does not supply; "This can be proven mathematically" is not followed by any mathematics. Driven at resonance the steady amplitude is Q times the static response, and Q is finite in every real medium. 2. Radiant Energy p.34 is in tension with itself: "Science claims complete resonance is not producible" (correct: infinite Q does not exist) followed by "in perfect resonance the oscillations will become more and more vigorous ... forever, or to the dissociation of matter" (true only for an undamped linear idealization) and "the impressed force which, in the case of the universe, is immense". The size of that force is exactly what Claim 3's arithmetic puts at 1.6e-13 J/m^3 (4.8e-5 W/m^2). The resonance gain Q cannot supply the missing nine orders of magnitude (it multiplies amplitude, not available power). This is the same point I made in B8 for Bearden's 29 stages. 3. Energy per atom. A Leyden jar of 1 nF charged to 30 kV stores 450 J; spread over 1 g of silver (5.6e21 atoms) that is 0.50 eV per atom [calc]. Nuclear changes need 10^5-10^7 eV in one nucleus. Quanta: 1 GHz photons carry 4.1e-6 eV, so supplying even the 215 keV that 109Ag to 109Cd costs would take 5.2e10 photons absorbed together by one nucleus; the rate of n-photon absorption falls as intensity^n, which is why multiphoton effects ionize atoms (tens of eV) and do not cross nuclear thresholds. Sound in a solid has phonon energies of hundredths of an eV; bubble collapse reaches 20,000 K, kT = 1.7 eV [web https://en.wikipedia.org/wiki/Sonoluminescence, per the note]. 4. Silver to cadmium, recomputed from the masses in the note: 109Ag to 109Cd is uphill by 215.1 keV, 107Ag to 107Cd by 1,416 keV; 109Ag + n releases 6.81 MeV then beta decay of 110Ag 2.89 MeV; 107Ag + n releases 7.27 MeV; the proton Coulomb barrier at r = 1.2(A^1/3 + 1) fm is 9.8 MeV; 9.8 MeV / 1.7 eV = 5.8 million [calc]. The note's numbers are right. One gram of silver at 1 MeV per atom is 8.9e8 J = 248 kWh [calc]. The real route (neutron capture) needs a neutron source, which the booklet does not describe and whose cost is not in the energy credit. 5. Vacancies. Glasser's text (inner-shell vacancy filled from outer shells, keV events in the electron cloud) leaves Z unchanged. Moray's added sentence is the step that carries the silver-to-cadmium claim, and no data accompany it. What remains true in Moray's p.34-35: structures fracture at mechanical resonance; there are real combined effects (sonophotocatalysis, sonoluminescence, laser isotope separation by tuned light). All change molecules or the electron cloud. Competing readings in Q08 (a theory assembled from reading; real chemistry with the wrong name; a real transmutation needing a neutron source) are all open on the record; (c) has no data behind it.

G5. Thorium "ten thousand times more active" is Le Bon quoting Rutherford, and the number is geometry, not energy (Claims 4, 5; leads 36, 37; Q11). [read: Q11]. The booklet pp.4-6 follow Le Bon, Book IV Ch.III sec.6, removing both Rutherford credits; the folded cylinder of Radiant Energy pp.58-59 is Le Bon's experimental part with "1 to 2 per minute" changed to "a definite discharge". Rutherford's apparatus (thoria vessel at +300 V, negative wire inside) gives a wire "over 10,000 times as active per unit surface". Independent check of the number [calc, assumptions]: ThO2 powder at 5 g/cm^3 holds 4.06 kBq/g x 0.88 x 5 = 1.8e4 Bq/cm^3 per chain member; a thick layer emits from the top alpha range (about 15 um) so about (1/4) n R = 6.7 Bq/cm^2 per member, about 40 Bq/cm^2 for the six alphas; a wire of 0.03 cm^2 carrying 36 kBq (the Q11 estimate, 10% escape of thoron from 100 g) shows 1.2e6 Bq/cm^2. The ratio is about 3e4, the same order as Rutherford's "over 10,000". So the figure is a surface-density gain from collecting daughter ions on a small area. The energy figures in Q11 check: Pb-212 to Pb-208 releases 11.78 MeV [calc from masses]; 36 kBq x 11.78 MeV = 0.068 microwatt; the whole chain for 100 g thoria is 360 kBq x 42.6 MeV = 2.5 microwatts. Concentration adds no energy: the deposit's decay power is at most the thoron escaping from the source. Whether Moray's tubes did this is inference (Q11 readings A, B, C): a radium-bearing coating and a negative cathode would collect radon daughters (the mechanism is real; the power is microwatts to milliwatts for practical amounts). The match of the 1968 "2.5 to 8 MeV" range to Tl-208 gamma (2.61 MeV) and 6-8.8 MeV alphas is true but weak, since most alpha emitters lie in 4-9 MeV.

G6. Le Bon's amalgamated-tin "radioactivity" is the photoelectric effect, anticipated, repeated and reinterpreted (Claim 5, Claim 3, Claim 4; leads 38, 39, 40; Q12). [read: Q12; web as cited there]. Checks [calc]: 295 nm is 4.20 eV and 360 nm 3.44 eV (E = 1239.8 eV nm / lambda), so amalgamated tin responding to 296-360 nm implies an effective work function of 3.4 eV or less, below clean tin or mercury (about 4.4-4.5 eV, [memory, medium]). That fits a photoelectric threshold and not a nuclear one (a nuclear process has no sunlight-sensitive threshold of a few eV and does not depend on charge sign or on emery-cleaning). One arithmetic point I add: Le Bon turns "40 degrees in 10 seconds" against "6 degrees in 60 seconds" for a uranium screen into "40 times" and then multiplies Rutherford's 70,000 alphas per second per gram by 40 to get "3,000,000 particles/second" (70,000 x 40 = 2.8e6). That step assigns the per-gram rate to a screen of unstated mass and treats equal currents as equal particle numbers. One 4.2 MeV alpha makes about 4.2 MeV / 35 eV = 1.2e5 ion pairs, so for equal current the photoemitting strip releases about 1.2e5 times as many charge carriers as the uranium screen does alphas [calc; 35 eV per ion pair, memory, high]. The "3,000,000" is not a measured count. Upper bound on the energy: a photocell cannot deliver more than the sunlight it absorbs, and its electrons carry less than the photon excess over the work function, about 0.8 eV at 295 nm if W is 3.4 eV; for 100 cm^2 total sunlight is about 10 W (1000 W/m^2 x 0.01 m^2) and the Q12 estimate of the current is about 1 nA, so about 1 nW at under 1 V [calc, assumptions in the note: 10 pF, 1 kV, 10 s]. Plauson's photoelectric amalgam patches on a collector balloon (US 1,540,998) are the same bound: nanoamperes per 100 cm^2, so a balloon of 100 m^2 would give microamperes at sub-volt emission, microwatts at most [calc], and the structure is a photocell, not a collector of cosmic energy. Reception record, weighed as history: Elster and Geitel (1889-92) had the amalgam result; Ramsay and Spencer (1906) repeated Le Bon and could not reproduce the positive leak or the passage through metal shields; Millikan and Winchester (1907) treated the data as photoelectric; Campbell (1906) called Le Bon's experiments "perfectly inadequate" and reported another worker's failed repetition. This is a rare case in the record where a rejected claim was repeated and the disagreement was settled by measurement, and it is the contrast to Claim 6's "never properly tested". Moray copies Le Bon's thesis ("dissociation of matter", Radiant Energy p.46), not the amalgam experiments, and elsewhere uses the mainstream term "photo-electric effect". Ives (lead 40): shared ancestry (Elster and Geitel's photosensitive metals), different mechanism (film thickness and standing waves); no formula of either bears on the other.

G7. The Q09 chain: vacancy current, "pump", detector crystal, hidden reservoir (Claims 3, 4, 5; leads 41, 42, 43; Q09 notes). [read: both Q09 notes]. Provenance first: "diamond" in Moray's text is Le Bon's radium gauge (Brazilian stones only), copied without Le Bon's footnote; "and sometimes other properties" is Moray's own insertion into Le Bon's paragraph (with copper dropped from the activator list); the sentence on an unweighable mass loss is Le Bon's and exact. No Moray document read names diamond as his detector. Arithmetic, each re-done here: 1. Mass converted. 50 kW x 86,400 s / c^2 = 4.8e-8 kg = 48 micrograms per day [calc]; so a year is 17.5 mg (see Corrections: my own earlier text says grams). 1 A of beta current at 2.89 MeV is 2.89 MW and 2.8 mg per day [calc]. 2. A correction to the note's decay rates. 2.89 MeV is the end-point energy of the 110Ag beta decay (it equals Q). The mean beta energy is about a third to two-fifths of Q [memory, medium] and an antineutrino carries the rest; at a mean of 1 MeV, 1 kW needs 1.7e5 curies, not the 5.8e4 curies the note gives (which is the minimum, if every decay delivered all 2.89 MeV to the load) [calc]. The conclusion (watts of output need at least that much radiation) only strengthens. 3. Charge and current. Atoms changed at 1 A: 6.24e18 per second, 96.6 g of silver per day [calc]. Spherical symmetry: for a radial current out of a point or sphere the Ampere-Maxwell total (J + epsilon0 dE/dt) is zero, because E = Q(t)/(4 pi epsilon0 r^2) and J = -(dQ/dt)/(4 pi r^2); so no magnetic field. That is correct and narrow: any asymmetry restores B. It says nothing about energy. "Near-ideal current source" is also right for MeV electrons pushing against kilovolts. 4. Moseley 1913 check. 20 mCi is 7.4e8 Bq; the radon bulb plus RaB and RaC betas give at most about 2 betas per decay, so if every beta escaped the current would be 2 x 7.4e8 x 1.6e-19 = 0.24 nA [calc]. The reported 0.01 nA (secondary sources, [web https://en.wikipedia.org/wiki/Atomic_battery], primary paper not read) is 4% of that, which is plausible for betas absorbed in the glass and the finite insulation. 150 kV x 0.01 nA = 1.5 microwatts; even the escape-everything ceiling at 150 kV is 36 microwatts [calc]. This is the right scale for a "vacancy current" device: very high voltage, microwatt power. The abstract's "more than a million volts" to stop the fastest betas agrees with the RaC end-point energy of about 3 MeV [memory, high], and the correction (the bulb, not the outer sphere, charged) is in the 10-08 note. 5. Foster 1990 (lead 42). If a radium-doped pellet stores radiation energy in traps, the storable energy is bounded by what the source deposits: a source of activity A whose chain gives about 30 MeV per decay (4.8e-12 J) deposits A x 4.8e-12 J x t. For 1 microcurie (3.7e4 Bq) over 30 years that is 3.7e4 x 4.8e-12 x 9.5e8 = 170 J [calc], and only the fraction trapped (a few per cent in the best known cases such as Wigner energy [memory, low]) could be released. The note's trap-density example gives 4.8 J/cm^3 (1e19 traps/cm^3 x 3 eV) [calc]. A 2,000 A heater supply at an assumed 10 V is 20 kW; one second of it holds 20 kJ, about 4,000 times the 5 J example [calc, voltage assumed, not in the account]. The account (breakers tripped, the pole transformer blown) therefore fits an electrical short through the pellet or die far better than a release of stored radiation energy, and the energies as stated cannot decide between that and "both". The account is one person's, 1990, unverified. It would decide nothing about the kilowatt claim even if exact: it concerns a pressing, not an operating device. 6. Neutrinos and the hidden reservoir. Solar neutrino flux 6.5e10 per cm^2 per s at a mean energy of about 0.3 MeV gives about 30 W/m^2 [calc], as in the note; a metre of lead stops about 1e-19, so 1e-17 W/m^2 needs an enhancement of about 1e21 to give watts [memory for the cross-section; the note]. Weber's claimed 1e23 coherent enhancement was tested at least four times with null results (Franklin 2010, [web https://www.colorado.edu/physics/sites/default/files/attached-files/gravitywavesandneutrinos.pdf], per the note); Borexino's neutrino luminosity within about 10% of the photon luminosity (+9/-11%) and N_eff = 2.99 +/- 0.17 bound a hidden channel to a few per cent of the Sun's output and to little extra radiation; helioseismology agrees to 0.2% rms for 1990s models and about 1% after the abundance revision [memory, medium for the last]. These bound the stellar-physics side. They say nothing about a mechanism in a box, and the 10-08 note is right to leave the hidden channel "bounded, not excluded". 7. Compton "pump". Background absorbed dose of 1e-7 Gy/h is 2.8e-11 W/kg, so 100 kg absorbs 2.8 nW [calc]; 50 kW is 1.8e13 times that. Compton scattering degrades photon energy and cannot concentrate it. The "taking up again" and Rutherford-school precedents (Moseley, 1899 electrometer, lead 30) show that direct conversion of decay particles is real physics; they do not change its scale.

C. Recomputed and agreed (kept so the reader can see what was checked)

Natural silver 51.8/48.2%; Q-values 215 keV, 1,416 keV, 2,891 keV; barrier 9.8 MeV; 1 A = 96.6 g Ag/day; 85,000 ion pairs per 2.89 MeV electron; 1 kW = 2.16e15 decays/s; ThO2 chain 2.5 microwatts; Wien and gap numbers of G2; Pb-212 chain 11.78 MeV; the 8-GHz-to-15-MeV photon-frequency conversion (3.6e21 Hz). No arithmetic error found in the Investigator's or follow-up numbers except the one in section E item 2; the one wrong sentence found was in my own review.

D. Weighing by claim (a-d), with verdict changes

• Claim 1 and 2. (a)-(d) unchanged. Nothing in these notes bears on the lamp-load record or the custody tests, except Foster's account (a pressing, G7.5), which is not a test of the device. No change.
• Claim 3. (a) The p.34-35 resonance text is internally in tension (G4.2): finite Q is admitted, then unbounded growth, with the "immense" impressed force unquantified. (b) Photoelectric amalgams, Compton and neutrino pumps, and the stellar checks all give powers many orders below kW (G6, G7.6-7): nW, microwatts, 3 nW against 50 kW. (c) Tested: no test of the cosmic hypothesis; Plauson's photoelectric patches are a design, not a test. (d) Unchanged. No change: (i) strong form Refuted by calculation; (ii) Internally inconsistent and Not testable as stated; (iii) Never properly tested.
• Claim 4. (a) The 1968 infrared paragraph is correct (G2) and not a CMB reference; Le Bon-derived sulphide and diamond passages quantify nothing (G3, G7). (b) Thorium-daughter plating on a cathode is a real effect worth microwatts-to-milliwatts (G5). (c) Foster 1990 is the only account of a Moray "detector" being made after Moray, and it is a pressing, single-source, unverified. No change: (i) Plausible, untested; (ii) Never properly tested / Not testable as stated; (iii) Refuted by calculation if no source exists; (iv) Plausible, untested as a modest-power concept. I add one sentence to (iv): the same arithmetic caps a radioisotope-excited device at the source's decay power (microwatts per millicurie-scale source, G5, G7.4), which is the "modest-power" qualifier.
• Claim 5. Split verdicts: 5A(i) "matter slowly dissociates into usable energy": Not testable as stated, unchanged; Le Bon's core (a loss of mass too small to weigh can free a great deal of energy) is quantitatively right by E = mc^2 (48 micrograms per day at 50 kW), which is why weighing cannot test it; what would test it is radiation and assay. New 5A(v), the resonance mechanism of Isotopes pp.20-26 (tuned current, rays and sound change silver to cadmium): Refuted by calculation as a mechanism (energy per quantum: 4e-6 eV for 1 GHz, 0.5 eV per atom even if a 450 J capacitor discharged into 1 g; 215 keV to 9.8 MeV per nucleus; a free damped oscillation cannot grow), and Never properly tested as an experiment (no silver or cadmium datum in the folder). This extends the existing "(ii) mechanism Refuted by calculation" from photon bombardment to the booklet's resonance mechanism; it is a new sub-verdict, not a reversal of an old one. New 5E, Le Bon-derived physics in the books: Historically superseded but instructive for (1) the 1905 alpha speeds and counts (G1), (2) universal spontaneous dissociation, (3) light-driven charge loss as "dissociation of matter" (reinterpreted as the photoelectric effect within a year or two of the repetition, G6), while Rutherford's 1900 excited-radioactivity result as quoted is Established (activity per area, G5), and Le Bon's own flagged numbers remain his. I am not removing Le Bon-derived passages from the record: they are Moray's printed text and the reader needs both the text and its source. (ii), (iii), (iv) of 5 unchanged: gold mechanism Refuted by calculation / assays Never properly tested; uranium Never properly tested and not supported by the numbers; Cosray Never properly tested, with the 1946 radiation description Internally inconsistent (B9, 10-04).
• Claim 6. Not mathematical. The Q12 history (a rejected claim repeated within months and settled by measurement) is a comparison case, not evidence about Moray. No change.
• Claim 7. The uncredited copying bears on authorship of the books, not on whether Moray held a degree. No change (Never properly tested).

E. Corrections to earlier documents

1. mathematician.md, Claim 5 verdict (i) (my own text): "a mass loss of about 17.5 g per year [calc]" is wrong by a factor of 1,000. 50,000 W x 3.156e7 s / (2.998e8 m/s)^2 = 1.76e-5 kg = 17.5 mg per year. The arbiter (arbiter.md, "Mass for 50 kW") and the physicist already have it right; the Q09 note's 48 micrograms per day agrees. Please replace "g" by "mg" in that line.
2. research/2026-10-07-followup-Q09.md section 3.3 and A09 (the "58,000 curies per kW" figure): it takes 2.89 MeV, the beta end-point, as the energy per decay delivered. With a mean beta energy of about a third to two-fifths of that, 1 kW needs about 1.7e5 curies; 58,000 is the lower bound (G7.2). Mean-energy fraction is [memory, medium], not web-checked.
3. sites.md lead 38 / Q12, "Per unit of current ... about 10^5": right as stated; I note that Le Bon's "3,000,000 particles/second" step multiplies a per-gram rate by a ratio of currents (G6) and is not a particle count.
4. Q01 (and lead 29) "70,000 per gram per second ... Le Bon's": not corrected by the notes; for the record, natural thorium gives about 4.1 kBq/g (2.4e4 with its chain), uranium about 12.6 kBq/g (about 1e5 with its chain) [calc]. This is a 1905 figure, not an error of Moray's own.
5. No correction needed to the 10-04 section of this file for the leads 1-28; its summary stands.

F. Source honesty for this section

• [web] fetched this round: https://en.wikipedia.org/wiki/Uranium-238 (gives "1 mole emits 3e6 alpha particles per second"; the 12.6 kBq/g is my division); the Isotopes of polonium page did not give alpha energies, so Po-212 8.78 MeV and Po-214 7.69 MeV remain [memory, high] and verdicts do not rest on them. The remaining web references are to pages the Investigator cited (Nobel lecture, Atomic battery, Sonoluminescence, Franklin 2010, Rex Research Le Bon); I did not re-fetch them.
• [memory] items used: mean beta energy fraction (medium), photoelectric work functions (medium), Wigner-energy storage fraction (low), neutrino cross-section in lead (the note's figure), helioseismology after the abundance revision (medium). None decides a verdict.
• Not weighed because not available: Moseley's full 1913 paper, Foster's original letter, Rutherford's 1900 paper, Nye 1974, the rest of the Athenaeum 1906 Le Bon correspondence, the Le Bon printing Moray used, any Moray notebook showing a silver, cadmium, amalgam or screen experiment.

G. Verdict changes from this section (summary)

• Claim 5: added 5A(v) resonance-transmutation mechanism as Refuted by calculation (mechanism) and Never properly tested (experiment); added 5E Le Bon-derived physics as Historically superseded but instructive, with Rutherford's excited-radioactivity result as Established. No earlier label changed.
• Claims 1-4, 6, 7: labels unchanged; Claim 4(iv) gains the decay-power cap; Claim 3 gains the p.34 internal-tension finding.
• Correction to my own earlier text: 17.5 g per year should be 17.5 mg per year.