Becquerel-Henri: Mathematician review (PhD mathematician, Step 2)

Source gap carried forward: only three short documents are in the folder; Becquerel's 1896 Comptes Rendus notes are absent. The Nature 1901 article has no text layer; I re-read all three pages from rendered images [read: page image p.396-398]. Harper's 1902 (Duncan) and Thomson 1903 were read from the text layer [read]. The folder is pure history with a few quantitative statements, so my checks are mostly arithmetic and unit checks; I redid each one.

Corrections to Step 1

1. Claim 3, "non-deviable part like X-rays", is not Becquerel's 1901 position. In Nature p.397 he says the non-deviable radiation is "of which the nature is still unknown" and notes Villard's very penetrating undeviable rays [read: page image p.397]. The X-ray identification comes from Duncan (Harper's) and Thomson, who even calls alpha "Roentgen-like". claims.md should attribute it to the 1902-03 popularisers, not to Becquerel.
2. Claim 2 overstates "intrinsic to uranium, independent of chemical form, no decay". Becquerel p.396 says activity "appears to be bound up with the presence of the element uranium" and that metal is 3-4x more active than its salts [read: p.396]. On p.398 he writes that, per Crookes, Debierne, Giesel and himself, uranium's activity is "due in great part, though not altogether" to a small quantity of actinium or another active body [read: p.398]. Polonium and active barium are described as losing activity with time [read: p.397-398]. So "no noticeable decay over years" applies to uranium only, and "independent of chemical form" is not stated in the source (the metal/salt difference is just uranium content per gram).
3. Thomson's speed is not "consistent with" Becquerel's number. Thomson gives "over 120,000 miles per second" [read], which is 0.64c (c = 186,000 mi/s). Becquerel's 1.6 x 10^10 cm/s is 0.53c, about 99,000 mi/s [read: p.398 + my arithmetic]. These are compatible only because Becquerel's figure is for one particular Hrho = 1600 component of a spread; faster components exist. Not a contradiction, but not the same number.
4. Harper's speed confirmed as a slip: the text reads "sixty millions to ninety millions of miles per second" [read], about 320-480 times the speed of light. Likely 60,000-90,000 (0.32-0.48c) misprinted, but that is a guess [memory: low confidence]. Step 1 was right to flag it.
5. Facts.json numerals (Hrho = 1600, 1.6 x 10^10 cm/s, e/m = 10^7, 4 x 10^-13 emu/cm^2, 1 mg per 1000 million years) all verified legible on the images [read: page image p.398]. Step 1's sourcing is otherwise sound.

Arithmetic checks (done by me, numpy/python)

Becquerel's p.398 numbers are mutually consistent.

• Velocity. A charge q of mass m moving at speed v perpendicular to a field H follows a circle of radius rho = m v/(q H), so H rho = (m/q) v, i.e. v = (q/m)(H rho). With q/m = 10^7 emu/g and H rho = 1600 gauss-cm: v = 1.6 x 10^10 cm/s. This matches his stated value exactly. That is 0.534 c, and gamma = 1/sqrt(1 - v^2/c^2) = 1.18, so the Newtonian formula is already off by about 18% (the measured q/m is really of the "transverse mass" gamma m_0; Kaufmann/Abraham/Lorentz sorted this out around 1901-1905 [memory: medium]).
• Unit of e/m. The modern value 1.7588 x 10^11 C/kg equals 1.76 x 10^7 emu/g (1 emu of charge = 10 C) [memory: high, from constants]. So his "10^7" is the right order and correct to about a factor 1.8, consistent with a 1901 measurement.
• Mass loss. Charge current 4 x 10^-13 emu/cm^2 (per second) divided by e/m = 10^7 emu/g gives mass flux 4 x 10^-20 g s^-1 cm^-2. Multiply by 1000 million years = 3.156 x 10^16 s: 1.26 x 10^-3 g = 1.26 mg. Becquerel's "about 1 mg" is correct. Thomson's 0.001 mg per million years (computed 0.00126 mg) is also correct.
• Thomson's ice figure. Classical kinetic energy flux = 1/2 (4 x 10^-20)(1.6 x 10^10)^2 = 5.1 erg s^-1 cm^-2. Over 10^6 yr = 3.156 x 10^13 s: 1.6 x 10^14 erg = 1.6 x 10^7 J per cm^2. Latent heat of ice is 334 J/g, so 4.8 x 10^4 g of ice, which at 0.917 g/cm^3 is 527 m, about 0.33 mile. Thomson's "more than a quarter of a mile" is correct. (My first script divided by the wrong power of ten; I corrected by hand: 4.8 x 10^4 g / 0.917 g cm^-3 = 5.3 x 10^4 cm = 527 m.) Using the relativistic (gamma - 1) m c^2 gives 5.5 erg s^-1 cm^-2, only 8% more.
• Where the arithmetic is valid but the framing is wrong. These numbers count only the beta-particle kinetic energy. Curie and Laborde (1903) measured radium's total heat as about 100 cal/hour per gram [web: https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf; https://history.aip.org/exhibits/curie/article.htm], about 0.12 W/g. A gram of radium has surface of order 1 cm^2, so the beta estimate (5 erg/s = 5 x 10^-7 W) is about 10^5 times too small. The dominant energy is in the alpha particles and daughter decays, which were not identified in 1901. So the "tiny mass loss therefore no energy problem" argument is internally sound as arithmetic but physically misleading: the energy is mass-energy released from nuclear binding, which is not visible in the mass of the emitted electrons.

Claim-by-claim review (claims.md numbering)

Claim 1. Uranium emits penetrating radiation, spontaneously, that blackens plates and discharges electrified bodies.

• a. Math consistent? There is no mathematics here beyond description. Nothing inconsistent.
• b. Consistent with experiment? Yes. It is the founding observation of radioactivity; every later measurement (ionisation chambers, Geiger counters, autoradiography) reproduces it [memory: high].
• c. Tested? Becquerel's cloudy-weather accident and dark dissolve-and-recrystallise tests are reported by Thomson [read]. Independently confirmed by Schmidt and Mme Curie (thorium), the Curies (pitchblende), Rutherford (ionisation, 1899) [read: Nature p.397]. Replicated many times; Becquerel shared the 1903 Nobel Prize [read: Harper's; the prize co-award is also [memory: high]].
• d. If true, what changes? Nothing; it is already the foundation of nuclear physics. In 1896 it required dropping the idea that atoms are permanent and that radiation requires an external excitation.
• Verdict: Established.

Claim 2. Emission is intrinsic to the element (uranium, thorium, radium, polonium, actinium), independent of phosphorescence/temperature, with no noticeable decay over years.

• a. Math consistent? No mathematics. The 3-4x metal-versus-salt ratio is a mass-fraction effect: uranium is about 85% by mass in U-metal versus roughly 40-70% in typical salts, so a factor 3-4 is larger than expected from mass fraction alone [memory: low]; I did not verify and do not lean on it. Becquerel does not claim a mechanism.
• b. Consistent with experiment? Mostly. Independence of temperature and chemistry holds to excellent accuracy for alpha/beta decay rates; Curie and Dewar's liquid-oxygen/hydrogen test found radium heat output independent of temperature [web: https://history.aip.org/exhibits/curie/article.htm]. "No decay over years" is right for uranium (half-life 4.5 x 10^9 yr for U-238 [memory: high]) but false for polonium (138 days [memory: high]) and was stated as decaying by Becquerel himself for polonium and actinium. Becquerel's own hedge that uranium's activity may owe partly to actinium or another body was a known 1900 confusion (uranium X / thorium-X type daughters, later explained by Rutherford and Soddy's separation of daughters) [read: p.398; [web] https://web.lemoyne.edu/giunta/ruthsod.html].
• c. Tested? Yes, by Becquerel, the Curies, Crookes, Debierne, Giesel [read]. Later replicated, with the correction that individual radioisotopes have characteristic half-lives.
• d. If true, what changes? Nothing now; at the time, conservation of energy appeared threatened (see claim 4).
• Verdict: Split. Strong form ("every active element emits without noticeable decay"): Tested, not supported. Weak form ("emission is a property of the element, independent of excitation, temperature and phosphorescence, and long-lived for uranium"): Established.

Claim 3. The radiation is a mixture of a deviable negative component (e/m about 10^7 emu/g, speed up to about 0.5c, like cathode rays) and a more penetrating non-deviable part.

• a. Math consistent? Yes. See arithmetic above: v = (e/m)(H rho) reproduces his number exactly. Hidden assumptions: a single velocity per Hrho (it is a magnetic-rigidity selector, not a speed selector, unless e/m is known; he cross-checks with the electric deflection, which gives a second equation for the two unknowns e/m and v, as in Thomson's 1897 method); non-relativistic mechanics at 0.53c (error about 18%, flagged above); the screen experiment of the Curies, in which the radium radiation charges an isolated screen negatively and the source positively, supplies the charge flux. The parabolic path in an electric field is correct for a constant transverse force [read: p.398]. One loose end: the 4 x 10^-13 emu/cm^2 figure is a current per unit area from a particular geometry (the Curie screen) and Becquerel applies it to "the radiating surface" without stating that only the beta fraction contributes. This is fine at the order-of-magnitude level.
• b. Consistent with experiment? Yes for the beta component: beta particles are electrons (e/m measurement by Kaufmann 1901-02 and Becquerel [memory: high]). The non-deviable part was later found to be gamma rays (Villard 1900, identified as EM radiation by Rutherford and Andrade [memory: high]); alpha particles (helium nuclei, Rutherford 1908 [memory: high]) are deviable only with very strong fields and are positive. Duncan's "deviable = cathode, non-deviable = X-ray" is therefore half right, and Thomson's "alpha is Roentgen-like" is wrong.
• c. Tested? Yes, by Becquerel, Giesel, Meyer and von Schweidler, the Curies, Kaufmann; replicated in later magnetic-spectrometer work [memory: high].
• d. If true, what changes? Nothing; it extends J.J. Thomson's electron to a new source, which implied the electron is a constituent of atoms.
• Verdict: Established for the beta component (Becquerel's own claim). Historically superseded but instructive for the 1902-03 popular framing "undeviable = X-rays, no alpha" (alpha and gamma were not yet separated).

Claim 4. The apparently inexhaustible energy of radium is explained by continuous transformation into lower-energy forms (Rutherford-Soddy via Thomson), not by absorption of energy from space (Mme Curie's speculation).

• a. Math consistent? Thomson's argument is a clean exponential-kinetics argument. If parent activity recovers as 1 - e^(-lambda t) and the separated daughter decays as e^(-lambda t), then the half-times coincide (both ln 2 / lambda). That is exactly the "infants" analogy [read: Thomson]. The ice calculation is correct (above). Hidden assumption: that the energy lies in the nucleus, which Thomson does not know: he says "a state in which it has less energy" without a mechanism. The inference "mass loss only 0.001 mg per Myr, therefore no contradiction" is arithmetic but, as noted, undercounts total energy about 10^5-fold because only beta energy was counted.
• b. Consistent with experiment? Yes. Rutherford and Soddy 1902: thorium X lost half its activity in about four days while thorium regained the same amount [web: https://web.lemoyne.edu/giunta/ruthsod.html]. Curie-Laborde calorimetry 1903 gave the energy rate [web: https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf]. Later: E = mc^2 and nuclear binding energies. The "radiations from space" alternative was not supported by later work (activity unchanged underground or shielded) [memory: medium; I did not find a single source and do not rely on it].
• c. Tested? Transformation theory: Rutherford and Soddy, replicated (Curie, Debierne, Boltwood and many) [memory: high]. The Curie "space radiation" speculation, in Duncan's text, is explicitly presented as speculation ("Heaven only knows") [read]; no one ran a test I can cite specifically for it, though the shielding argument above is the natural test [memory: medium].
• d. If true, what changes? Atoms are not permanent; the energy of matter is stored in atoms at scales far beyond chemistry. All of this has already been incorporated into mainstream physics.
• Verdict: Split. Transformation explanation: Established. "Energy absorbed from space" (Curie's aside, as reported by Duncan, not a worked-out hypothesis): Never properly tested as a specific proposal (it was dropped, not formally refuted); it is Not testable as stated in the form given (no frequency, no mechanism).

Claim 5. Induced radioactivity of nearby bodies; Thomson's ion-layer proposal (positive ions near a negatively charged body pull cathode rays out of it, no radioactive substance needed).

• a. Math consistent? Thomson gives no numbers, so it is qualitatively consistent only. I tried the order-of-magnitude check. Treat the ion layer as a slab of thickness d and volume charge density rho = n e. Its field is E = rho d / epsilon_0. Air contains about 10^3 ions/cm^3 = 10^9 /m^3 [memory: medium, from standard atmospheric electricity texts; not verified on the web]. Then rho = 1.6 x 10^-10 C/m^3 and E = rho d / epsilon_0 = 18 V/m per metre of thickness. Pulling electrons out of a metal needs about 10^9 V/m (field emission; Fowler-Nordheim 1928 [memory: medium]), so d would have to be about 5 x 10^7 m, larger than the Earth's radius. Even allowing a factor 10^3-10^4 for tip enhancement on a lightning-rod or leaf point, ambient fields (about 100 V/m fair weather [memory: medium]) reach at best 10^6 V/m, still 10^3 below. Electrical breakdown of air occurs near 3 x 10^6 V/m [memory: high]. So the mechanism is refuted by calculation to the extent my memory-sourced inputs are right; the inputs are medium confidence, but the conclusion has about three orders of margin.
• b. Consistent with experiment? The observation (Elster and Geitel: negatively charged wires become radioactive in open air) is real. The explanation is not cathode rays but positively charged radon decay products attracted to the negative wire [web: https://www.ebsco.com/research-starters/physics/elster-and-geitel-study-radioactivity (search summary); radon progeny discussion in search results]. Thomson himself found radioactive emanation in tap and well water in 1902-03 [web: https://zenodo.org/records/1945007; https://www.nature.com/articles/068343d0], which undermines his own "no radioactive substance needed" line. Note the irony: Thomson's first step (a negative body attracts positive ions) is correct and is the real mechanism; only the "ions pull out cathode rays" step is wrong. Elster and Geitel's original hypothesis, a radioactive constituent of air, was the right one.
• c. Tested? The temporary "induced activity" from radium emanation was verified by Rutherford, the Curies and others [read: Nature p.397]. Thomson's hypothesis was not tested directly; it was superseded when emanation and decay-chain progeny were understood [memory: high; partly [web] above].
• d. If true, what changes? If Thomson's mechanism were real, activity would not require radioactive elements at all and "radioactivity" would split into nuclear and electrical kinds; cosmic and atmospheric ionisation would need rethinking. The calculation above shows it does not.
• Verdict: Split. (i) Induced radioactivity near radium/emanation: Established (deposition of decay products, Historically superseded in its 1901 description as a mysterious property of the bodies). (ii) Thomson's ion-layer cathode-ray mechanism: Refuted by calculation (about three orders of magnitude short in field strength; inputs medium confidence, see 5a), and later superseded by radon-progeny deposition. Instructive: a correct qualitative step followed by an unchecked quantitative leap.

Summary of mathematical themes

• Becquerel's 1901 numbers are internally consistent and reproducible by hand (v = (e/m) H rho; 1 mg per 10^9 yr). He was a careful experimenter; the 1901 weakness is conceptual (energy accounting), not arithmetic.
• The popular accounts carry the errors: Duncan's mph slip, Thomson's alpha identification and ion-layer hypothesis.
• No claim depends on [memory] alone for its verdict except the non-deviable-ray identification, the atmospheric ion numbers (medium confidence) and the Villard/Rutherford later history (high confidence).