Becquerel-Henri: Arbiter (Step 3)

Source gap (repeated from Step 1, partial). The folder holds three short documents: Becquerel's 1901 Nature article (scanned, no text layer; read from page images by all reviewers), and two popular Harper's pieces (Duncan 1902, J. J. Thomson 1903; text layer). Becquerel's 1896 Comptes Rendus notes and his 1903 Nobel lecture are absent, so the 1896 discovery rests on his 1901 retelling and popular accounts. Verdicts on Claims 1-3 rest on [read] material plus well-known later history; where a conclusion rests on [memory] it is marked below. The page-image numerals (Hrho = 1600, 1.6 x 10^10 cm/s, e/m = 10^7 emu/g, 4 x 10^-13 emu/cm^2, 1 mg per 10^9 yr) were independently confirmed legible by all three reviewers and re-derived by me, so the earlier "re-check the scan" warning can be dropped.

Framing. This is not a suppressed-claim project. It is the founding record of radioactivity, plus three instructive wrong turns: Becquerel's "no contradiction with energy" remark, Mme Curie's "energy from space" aside, and Thomson's ion-layer hypothesis. All three reviewers agree on this; nothing in the folder documents suppression or blocked tests.

Recomputed numbers (where reviewers differed or relied on each other)

Web checks (budget 8)

1. Curie-Laborde 1903: 1 g radium emits about 100 g-cal/h [web: https://history.aip.org/exhibits/curie/article.htm ; https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf]. Upgrades the physicist's [memory: medium] item. Curie and Dewar showed the heat is independent of temperature (same sources).
2. Field emission from metals needs about 1-10 GV/m [web: https://www.classe.cornell.edu/Research/SRF/SrfCavitiesAPrimerThree.html and search summaries]. Upgrades the "10^9 V/m" input (previously [memory] medium) that Claim 5b depends on.
3. Not checked: atmospheric ion density 10^3/cm^3 and fair-weather field of 100 V/m (mathematician, [memory: medium]). I did not confirm these. The 5b verdict below does not need them to be exact, since even a 1000-fold error leaves the mechanism short.
4. Not checked: the claim that radioactivity is unchanged underground or shielded (physicist and mathematician [memory: medium]). I did not find a source in the time budget, so I do not use it for any verdict.
5. Reviewers' other web sources (Rutherford-Soddy, Elster-Geitel, Kaufmann) were accepted as given; I did not re-open them.

Corrected claim list

Reviewers agreed on these corrections to claims.md, and I adopt them:

1. Claim 1: unchanged.
2. Claim 2: the Nature article does not state "independent of chemical form" (metal is 3-4x more active than its salts [read p.396]), and it does not support "no decay" for all elements. It says polonium and radium-derived activity vary with time, and that uranium's activity is "due in great part, if not altogether" to a small quantity of actinium or another active body [read: page image p.398; QE revision: earlier "though not altogether" was a misquote]. Restate: "Radioactivity is a property of certain elements, independent of light excitation, temperature and phosphorescence; for uranium it is steady over years."
3. Claim 3: Becquerel says the non-deviable rays are "much less penetrating than the deviable rays" and adds Villard's separate "very penetrating rays which are not capable of deviation" [read p.397; QE revision: not "more penetrating"]. Becquerel himself says the nature of the non-deviable rays "is still unknown" [read p.397]. The "like X-rays" identification belongs to Duncan and Thomson (1902-03). The speed "up to half c" is for one radiation (Hrho = 1600), not a maximum; Thomson cites over 120,000 mi/s (0.64 c) for the fastest radium rays. The nonrelativistic formula is off at 0.5 c, and e/m = 10^7 is about 43% below the modern 1.76 x 10^7.
4. Claim 4: Becquerel's article does not give the transformation answer. It says only that the tiny mass loss means "no contradiction" with energy [read p.398], which hides an undercount of energy output by several orders of magnitude (up to about 10^5 if the sample were pure radium; QE revision). Thomson, not Becquerel, supplies the Rutherford-Soddy answer. Curie's "space energy" idea appears in Duncan only as speculation.
5. Claim 5: split into 5a (induced activity: observation) and 5b (Thomson's ion-layer mechanism).
6. Facts.json fix: "radium 100,000 times uranium" is from Harper's (Duncan; also Thomson, "more than 100,000 times that of uranium"), not Nature (Nature p.397 says "several thousand times").
7. Duncan's "60 to 90 million miles per second" exceeds c by a factor of over 300; the cause (probably 60,000-90,000) is a guess [memory: low].

Claim 1: Uranium emits penetrating radiation spontaneously, blackening plates and discharging electrified bodies

• Agreement: all three say Established. No disagreement.
• Tested status: tested and replicated (Schmidt, Mme Curie, Kelvin/Beattie/de Smolan, Rutherford [read p.397]).
• Final verdict: Established.

Claim 2: Emission is an intrinsic property of the element, independent of excitation, temperature, chemical form; no decay over years

• Agreement: all split it into a weak form (intrinsic, no light/heat dependence) which is Established, and a strong form ("no decay", "independent of chemical form") which is wrong.
• Disagreement: label for the strong form. Mathematician says "Tested, not supported"; physicist and contrarian say "Historically superseded but instructive". Stronger: superseded. Becquerel never claimed the strong form in the paper (he reports polonium decaying [read p.397]); it is claims.md's overstatement, and no experiment was designed to test "no decay for all elements". It was a reasonable early position for uranium, replaced by half-lives.
• Tested vs never tested: tested. Crookes, Becquerel and Rutherford-Soddy found separable, decaying components [read, Thomson; web via contrarian].
• Final verdict (split): weak form Established; strong form Historically superseded but instructive. The "no decay for uranium over years" part is consistent with a half-life of 4.5 x 10^9 y [memory: high].

Claim 3: Radiation is a mix of a deviable negative part (e/m about 10^7, speed up to about c/2) and a non-deviable part (Becquerel: less penetrating, plus Villard's penetrating non-deviable rays; "like X-rays" is the popularisers')

• Agreement: beta component is electrons: Established. The "non-deviable = X-rays" framing is incomplete (alpha = He nuclei, positive, and deviable only in strong fields; gamma = EM radiation).
• Disagreement: none on verdict. Detail: the physicist notes the relativistic error and e/m 43% low; the mathematician quotes 18% for the nonrelativistic error at 0.53 c. Both are right (different quantities). With Becquerel's own e/m the relativistic beta is 0.47; with the modern value 0.68.
• Tested vs never tested: tested (Becquerel, Kaufmann, Giesel, Curies). Mathematics internally consistent.
• Final verdict (split): Beta = electrons, deviable part: Established (1901 numbers rough). Non-deviable "like X-rays": Historically superseded but instructive; the attribution is to the popularisers, not Becquerel.

Claim 4: Radium's energy comes from transformation into lower-energy forms, not from absorbing energy from space

• Agreement: transformation is Established. The Becquerel "no contradiction" remark is an evasion; all agree the arithmetic is right and the physical conclusion wrong.
• Disagreement: verdict for the "space energy" alternative. Physicist: Tested, not supported / Refuted by calculation. Mathematician: Never properly tested / Not testable as stated. Contrarian: Historically superseded, never directly tested.
• Resolution: the contrarian's reading is the most careful. No source in this review cites a controlled test of a space-energy hypothesis (my own web check of the underground-shielding claim was not completed, and I do not rely on it). The physicist's "refuted by calculation" is overstated: an energy budget shows nuclear decay accounts for the heat (Q = 4.87 MeV with 1600 y half-life gives 0.029 W/g for Ra-226 alone, and about 0.12 W/g with daughters, matching Curie-Laborde), but that shows the alternative is unnecessary, not impossible. The mathematician's "Not testable" is also too strong: the idea predicts variation with shielding, so it is testable in principle. It was a one-sentence aside with no mechanism, so no one tested it as such.
• Tested vs never tested vs dismissed: transformation: tested. Space energy: never specifically tested; displaced by a theory that made quantitative predictions that succeeded (equal decay and regrowth times, thorium X half-life about 4 days [read, Thomson; web]). Not dismissed without a hearing, simply overtaken.
• Final verdict (split): Transformation: Established. Space-energy source: Historically superseded but instructive (never directly tested in the sources reviewed; unnecessary given the energy accounting). Becquerel's 1901 "no energy problem" inference: Refuted by calculation (the beta-only estimate undercounts output by several orders of magnitude, up to about 10^5 for pure radium; see QE revision).

Claim 5: Induced radioactivity; Thomson's ion-layer hypothesis

5a. Induced activity near radium or emanation. All agree: real (Rutherford, Curies, Elster-Geitel), explained by deposition of radon/thoron decay products, more on negatively charged bodies. Becquerel's own account that the activity is removed by acid and left behind on evaporation [read p.397] already points to a deposited substance. Final verdict: Established.

5b. Thomson's mechanism. - Disagreement: physicist and mathematician say Refuted by calculation; contrarian says Historically superseded (and, at math stage, not testable without numbers). - Resolution: the calculation is robust. Needed field is 1-10 GV/m [web: Cornell SRF primer]; ambient fields are about 100 V/m, and even a slab of ions of any plausible thickness gives about 18 V/m per metre. The margin is 3 to 7 orders of magnitude depending on tip enhancement, far larger than the uncertainty in the medium-confidence atmospheric inputs. The contrarian's point that Thomson gave no numbers is fair: the refutation is mine, not Thomson's refuting. It applies to the mechanism as stated ("ions pull cathode rays out of metal"); Thomson reported unspecified recent experiments with no data [read], and I could not check them. - Tested vs never tested: no dedicated test of the cathode-ray mechanism found; the acid-removal observation and radon-progeny decay times are indirect evidence against it. Not dismissed without testing, but overtaken. - Final verdict: Refuted by calculation (field strength short by at least 10^3), and Historically superseded but instructive (correct first step, that a negative body attracts positive ions, then an unchecked quantitative leap).

What the reviewers collectively got right or wrong

• Physicist: strongest on the energy undercount and relativistic correction; overreached on "refuted by calculation" for Curie's space-energy aside.
• Mathematician: most careful on attribution of claims; "not testable" for space energy overstates.
• Contrarian: the cleanest on tested/never tested/superseded and correct that nothing here was suppressed; understated how decisive the field estimate is for 5b.

What would settle it

• Claim 2 strong form: measure the activity of a freshly purified uranium sample versus time (daughter ingrowth) and compare with an aged one; the same measurement for polonium shows decay with half-life 138 days [memory: high].
• Claim 3: remeasure e/m for beta rays at several Hrho using modern e/m and the relativistic momentum relation; consistency across speeds settles the speed claim.
• Claim 4 space energy: count radium-source heat or a decay rate deep underground versus surface in shielded calorimeters; no change settles it directly. (I know of no such test cited here.)
• Claim 5b: charge a wire in a radon-free chamber and measure induced activity; absent radon, no activity would settle the matter against Thomson. This is the Elster-Geitel setup with the radon removed.

Open questions

• Which 1896 data exactly led Becquerel to drop the refraction/polarisation idea? The Comptes Rendus notes are needed.
• What were Thomson's "recent experiments" on induced activity? No data are given.
• Does anything remain of Mme Curie's space-radiation speculation in the original French source, beyond Duncan's retelling?
• Atmospheric ion densities and fields used in 5b are [memory: medium], unchecked.

Source list

• [read: page image p.396-398] Nature 63, 21 Feb 1901; [read] Duncan, Harper's 1902; Thomson, Harper's 1903.
• [web] https://history.aip.org/exhibits/curie/article.htm ; https://www.nobelprize.org/uploads/2018/06/pierre-curie-lecture.pdf (Curie-Laborde heat; Curie-Dewar temperature independence).
• [web] https://www.classe.cornell.edu/Research/SRF/SrfCavitiesAPrimerThree.html (field emission 1-10 GV/m).
• [web, from reviewers, not re-opened] https://www.chemistryworld.com/features/transmutations-and-isotopes/3004868.article ; https://history.aip.org/exhibits/rutherford/sections/exploring-radioactivity.html ; https://www.ebsco.com/research-starters/physics/elster-and-geitel-study-radioactivity ; https://arxiv.org/pdf/1208.6527 ; https://www.chemlin.org/isotope/radium-226.
• [memory: high] alpha = He nucleus; e/m = 1.76 x 10^7 emu/g; U-238 half-life 4.5 Gyr; Po-210 138 d. [memory: medium, unchecked] atmospheric ion density and fair-weather field; shielding independence of decay rates.

QE revision (after qe.md, verdict MINOR)

New evidence came from the two deep reads in deep/ (page-image reading of Nature pp.396-398; Thomson pp.291-293) and the QE spot checks. No verdict changes. Four defects are corrected above and restated here.

1. Misquote of Becquerel on actinium. The page image reads "due in great part, if not altogether, to a small quantity of actinium or of another radio-active body" [read: page image p.398, deep/Nature-Feb-1901-184ae1.md section 1]. My earlier "though not altogether" (also in physicist and mathematician) reversed the hedge. Becquerel was entertaining that uranium might itself be inactive. Thomson's twelve-month result settles it: the uranium solution, made inactive by removing activity onto barium, "had recovered all its former activity" while the barium became "absolutely inert" [read: Thomson p.292, deep/J-J-Thomson section 2]. Uranium regrows its own activity, so uranium is the source of the (UX) activity, not actinium; Becquerel's activity-carrier was a separable daughter [memory: high]. Thomson says Becquerel "fortunately kept his solutions", i.e. a chance observation, with no measurements or dates. Claim 2 verdict unchanged: weak form Established, strong form Historically superseded but instructive.
2. Penetration of non-deviable rays. Becquerel writes: "The non-deviable rays are much less penetrating than the deviable rays; the polonium radiation is limited in air to a kind of sheath of some centimetres" [read: p.397], and then reports Villard's "very penetrating rays which are not capable of deviation". So by 1901 he already had three components, in effect alpha < beta < gamma in penetration [memory: high for the later names]. He himself says the nature of the non-deviable kind "is still unknown". Claim 3 heading and corrected list are fixed. Verdict unchanged: beta = electrons Established; "non-deviable = X-rays" is the popularisers' (Duncan, Thomson) and Historically superseded but instructive. Becquerel was nearer the modern picture than they were.
3. Claim 4c, size of the energy gap. The "10^5" divides the Curie-Laborde heat of pure radium (0.12 W/g, 1903 [web]) by a screen current from a 1900 preparation. The same article says the Curies' products were only "several thousand times" as active as uranium; pure radium is about 3 x 10^6 times uranium per gram (about 3.7e10 versus 1.2e4 decays per second per gram [memory: high, not web-checked]). So the beta-screen estimate undercounts by several orders of magnitude, up to about 10^5 for pure radium. The verdict rests on a sounder argument: in equilibrium a radium decay with its daughters releases about 24 MeV, mostly as alpha energy absorbed inside the source where Becquerel's screen cannot see it, against well under 2 MeV of beta energy, part of it carried off by neutrinos [memory: high]. Becquerel's inference from tiny mass loss to "no energy problem" is wrong whatever the numbers. Verdict stays: Refuted by calculation.
4. Attribution of "100,000 times uranium". Harper's, in two places: Duncan, and Thomson ("more than 100,000 times that of uranium" for purified radium [read: Thomson p.3 of the printout]). Nature p.397 says "several thousand times". facts.json should say so.

Other new evidence, all [read] and used only as support, no verdict effect: - Nature p.397: radium salts "increase spontaneously in activity for some time after their preparation" while polonium's diminishes (Curies, Giesel). This is the ingrowth and decay later explained by Rutherford-Soddy, seen before it was understood; Thomson p.292 repeats it for radium ("when freshly prepared it is not nearly so active"). It supports Claim 2's strong-form verdict. - Thomson p.292-293: helium occurs only in radioactive minerals, and "helium may be very closely connected with the gaseous emanation". Only a conjecture; the alpha = helium link is [memory: high]. - Thomson p.293 on induced activity: the Earth's charge makes pointed conductors radioactive, and "freshly fallen rain" is radioactive for about an hour. Correction to the QE note: the rain finding is attributed to C. T. R. Wilson, not to Elster and Geitel (they did the charged-wire experiment). Both fit airborne radon progeny and so support 5a and the 5b rebuttal; Thomson gives no data for his own experiments. - Giesel's "polonium" was deviable and the Curies' was not, so the preparations were different substances [read: Nature p.397]; the Curies' and Giesel's chemical separations of 1899-1900 gave mixtures. - Rutherford-Soddy test as Thomson states it: time to lose half the activity (thorium X) should equal time to regain half (thorium), "satisfied with great exactitude", no numbers given. - Burns: Pierre Curie's deliberate arm exposure (about 1.5 h, a sore lasting more than three months) and Becquerel's pocket burn are in Harper's (Duncan); optional history.

Checks still open: atmospheric ion density and field, and decay-rate independence underground, remain [memory: medium] and unchecked; 5b has 10^3 or more of margin so no verdict depends on them. No new web checks in this revision (2 of 8 used earlier, plus reviewers' sources).

Verdict table (changes only in wording; no verdict changed)