Soddy: Arbiter (Step 3)

Project: Frederick Soddy, Matter and Energy (1911) and The Interpretation of Radium and the Structure of the Atom (4th ed., Aug 1920; 1922 reprint). Inputs weighed: facts.json, claims.md, deep/, physicist.md, mathematician.md, contrarian.md, sites.md and research/. Tags: [read] from the extracted text or a page image I looked at, [web] with URL, [memory] with confidence, [calc] my own arithmetic (Python, .venv, scratch in /tmp/soddy_arb).

Source gap note (repeated from claims.md): the folder holds only two popular books. Soddy's papers with Rutherford and Ramsay, his lead-atomic-weight and isotope papers, and his economics (Cartesian Economics, Wealth, Virtual Wealth and Debt) are NOT in the folder. The Investigator found some online and read extracts only. Everything below about those works rests on [web] extracts or [memory]. Deep reads cover Interpretation PDF pp.153-165 and 268-300 and Matter and Energy PDF pp.201-236; the rest of both books rests on digests, plus the passages the reviewers read themselves (claim 7). Where a verdict rests on [memory] alone it says so.

Reader's key. alpha particle = helium-4 nucleus (charge +2e, mass 4 u). beta particle = electron from the nucleus. Mean life T = 1/lambda (Soddy's "period of average life"); half-life = 0.693 T. Calorie = 4.184 J. Z = atomic number (protons), A = mass number. Soddy's "periods" are mean lives, so they look 1.44 times longer than modern half-lives.

---

1. How the three reviewers line up

The three reviews agree on every central verdict. Soddy's books are mainstream-successful physics, so there is no suppressed-anomaly story to adjudicate. The disagreements are in details, and the arbiter's work was to recompute them:

Numbers I recomputed ([calc], all reproduce the reviewers within rounding)

• Ra-226: 3.66e10 decays/s/g. Heat: alpha energy of Ra only (4.87 MeV) = 24.6 cal/h (matches "25 from radium" in the 1909 edition); four alphas of the chain (24.4 MeV) = 123 cal/h; 28 MeV = 141; 29.2 MeV including beta Q-values = 147 cal/h. Soddy's 133 lies inside this range; 100 (1909-11) was 25-30% low. The ceiling depends on how much beta/gamma energy the calorimeter absorbs; the three reviewers' ranges (121-141, 125-147, to 141) are all this same range.
• 133 x 8,760 x 2,500 = 2.91e9 cal; per-atom count 29.2 MeV x 2.66e21 atoms = 2.98e9 cal. "A million times coal": 1.33e6 against coal plus its oxygen (2,182 cal/g), 3.6e5 against coal alone. Quarter ton of coal: 0.27 t at 100 cal/h, 0.36 t at 133.
• Mean lives: Ra-226 2,308 y (Soddy 2,500, +8%); U-238 6.45e9 y (Soddy 7.5e9 in 1911, +16%; 8.0e9 in 1920, +24%); Th-230 108,800 y (Soddy 100,000, -8%). Product T(Th-230) x T(Ra): 2.51e8 modern against Soddy's 2.50e8. The mathematician's point stands: the t-squared growth experiment (Ra/Ra_eq = t^2 / (2 T_Io T_Ra) = 2.0e-7 after ten years, observed 2.1e-7) measures that product, which agrees with modern values to 0.4%, even though each factor is off by 8%.
• Uranium:radium ratio: modern atom ratio 2.79e6, mass ratio 2.94e6; Soddy 3.0e6 (1911), 3.2e6 (1920). If 3.2e6 is a mass ratio (Boltwood and Rutherford measured by weight), the atom ratio is 3.04e6 and the uranium life 7.6e9 y, not 8.0e9 [calc]. The book does not say which it used. A 5% hidden assumption.
• Radon: 481,250 s = 5.570 d mean; x 0.6931 = 3.861 d against printed 3.84 d; modern 3.82 d / 5.52 d. Emanation per gram of radium at equilibrium: 6.5e-6 atoms per Ra atom = 0.64 mm^3 at STP; Soddy's 0.6 is right.
• e^40 = 2.4e17 (Soddy: "a thousand billion", understated, conservative); e^-10 = 4.5e-5 (true that under 1/1000 remains after 25,000 y).
• Joly's uniform radium in the earth: 5.97e27 g x 2e-12 x 0.164 W/g = about 1,960 TW against measured total surface heat flow of about 44-47 TW [web: https://en.wikipedia.org/wiki/Earth%27s_internal_heat_budget]. Roughly 40 times too much. Sun: uranium plus chain gives 9.4e-8 W/g, so a uranium sun would need about 2 solar masses.
• Fission (200 MeV per U-235) gives 8.2e13 J/kg, about 6.8 times the 1.2e13 J/kg from all of radium's decay in Soddy's figure.
• Coulomb barrier for a 5.5 MeV alpha: nitrogen about 4.2 MeV, copper about 12.5 MeV (r = 1.2 (A^1/3 + 4^1/3) fm). Alpha-particle bombardment of nitrogen is feasible at MeV energies; copper to lithium by radon alphas is not.
• Gold from silver: Delta A = 197 - 109 = 88 (or 90 from Ag-107). Decay never raises A.

---

2. Web checks by the arbiter (budget 8; 5 used)

1. Blackett 1925 forks: [web, search summary only] about 23,000 photographs with about 400,000 tracks and eight forked tracks, two-pronged forks, so the alpha is captured (14N + alpha -> 17O + p). Page: https://www.nndc.bnl.gov/discovery/abstracts/8/17.pdf and https://arxiv.org/pdf/1202.0954. The original paper was not opened; sources differ slightly on the totals. The physicist's second-hand "8 in 23,000" is supported at summary level. Verdict does not depend on the exact numbers.
2. Word "isotope" and Margaret Todd: [web] https://en.wikipedia.org/wiki/Margaret_Todd_(doctor) , https://www.chemistryworld.com/feature/transmutations-and-isotopes/3004868.article and the Scottish National Dictionary entry (https://www.dsl.ac.uk/entry/snd/snd00090062): the word was suggested to Soddy by Margaret Todd in 1913 (a dinner at the Beilbys' per Chemistry World) and adopted by Soddy; the dictionary says it expresses the idea Fleck put forward. This settles the physicist's unchecked [memory, high] point; the 1951 NYT credit to Soddy alone is the reporter's shorthand, not a competing claim. Retag [web].
3. Rutherford's June 1920 mass-3, charge-2 particle: [web, search summary] reported in the 1920 Bakerian lecture, the "X3" notation, nitrogen minus X3 leaves boron-11; Rutherford's later work showed only protons and alpha particles; helium-3 was first identified by Alvarez and Cornog in 1939 (https://en.wikipedia.org/wiki/Helium-3 , https://physicstoday.aip.org/news/helium-3-isotope). Confirms "Tested, not supported" and corrects a loose phrase in the physicist: He-3 was proposed in 1934 by Oliphant and found in 1939, so the 1920 particle is an early mistaken hint, not a mis-measured He-3. The "later withdrawn" gloss now has a [web] source at summary level.
4. Rutherford "moonshine", British Association, 11 Sept 1933: [web, mostly one popular article reposted on mirror sites, https://nuclearfaq.ca/yonge.htm] The Times report of the next day; Szilard's chain-reaction idea followed. The wording is the newspaper's paraphrase; tag the primer's use "as reported by The Times". No verdict depends on it; it matters only to the contrarian's point that the leading authority was dismissive of practical atomic power five years before fission.
5. Hess 1912 as source of the 133 cal/h figure: [web] search found no source; it stays unattributed. A Glasgow thesis (https://theses.gla.ac.uk/80920/1/27555572.pdf) repeats "exact experiments have proved ... 133 calories per hour" without naming the experimenter. The 133 figure is therefore Soddy's quoted value; the modern per-atom count (123-147) supports it. Not reached: the lead-weight figure 206.05 (physicist's check failed; I did not repeat it; it stays "Soddy's stated value, experimenter unnamed"). Still [memory] and not verdict-bearing: Rutherford-Royds 1909, Gamow 1928, Cockcroft-Walton 1932, Bethe, Oklo, packing fraction 1927.

---

3. Corrected claim list

QE revision (section 10) adds to this list: C6e (Ramsay and Cameron copper to lithium), C8 (life and the second law; forces; Le Sage), C9 (actinium as a branch of uranium); and adds the helium and "thirteen per cent" inconsistencies to C3 and C2.

claims.md states seven claims. The reviewers split several into parts with different fates. The corrected list below is the version later steps should use. Changes from claims.md are marked "Changed".

C1. Radioactivity is spontaneous, each atom has a fixed chance per second independent of its age; so decay is exponential (mean life 1/lambda, half-life 0.693 of it). Changed: add that the strong form "independent of every other consideration" holds for alpha and most beta decay but not for electron capture and bound-state decay (fully ionised Re-187 about 33 years against about 41 billion years neutral).

C2. Radium's heat is internal atomic energy: about 133 cal/h per gram (1920; about 100 in 1909-1911), about 2.9e9 cal per gram over the change of the radium atom and its quick products (to Po-214), about 1.3e6 times the energy of coal plus oxygen (3.6e5 times coal alone); consistent with conservation of energy because the radium atom itself changes. Changed: (i) the 1911 value of 100 cal/h is a printed-evidence inference, not a guess: the sentence "in thirty-eight hours, its own weight of water" fits 100 and the 1909 first edition prints 100 outright [web: https://archive.org/details/interpretationof00soddrich]; (ii) "complete change" excludes the Pb-210, Bi-210, Po-210 tail (about +23%); (iii) the "million times" has a stated basis; (iv) 133 is inside the modern 123-147 range, 100 was about 25-30% low.

C3. Radium survives only because it is continuously made: in secular equilibrium the amount of a product relative to its parent equals the ratio of their mean lives; Soddy's mean life of radium 2,500 y (modern 2,308, +8%), uranium 7,500 (1911) to 8,000 (1920) million years (modern 6,450, +16 to +24%); chain U, UX, ionium, radium, emanation, ..., lead with eight alphas and six betas. Changed: say these are Soddy's numbers with the modern offsets; add six betas (92 - 16 + 6 = 82); add that the "58 grams ionium per gram of radium" sentence is internally inconsistent with the table (40); add the uranium-radium weight-versus-atom-ratio 5% assumption; note Soddy's 1911 "a few per cent" was too tight.

C4. The alpha particle is a helium atom (charge 2, mass 4); radium becomes its emanation by expelling one; helium accumulates in uranium and thorium minerals. No change. Note "uranium appears to be made up of eight atoms of helium with one of lead" (1911) is a mass-balance description, not a literal composition.

C5. Atoms of one element can differ in mass (isotopes); displacement rules (alpha: Z down 2, A down 4; beta: Z up 1); ordinary lead is not one thing (206.05 and 207.9 found for uranium-lead and thorium-lead); Aston's 1920 spectra show many elements are mixtures with near-whole-number masses; priority for the 1913 generalisation. Changed: (i) the integer rule "to one part in a thousand" is superseded by Aston's own 1927 packing-fraction result; some table entries (sulphur, xenon, mercury) were wrong or incomplete in 1920; (ii) ordinary lead is four isotopes (204, 206, 207, 208), not a mixture of two; (iii) credit: the 1920 book names Russell, Fajans and Soddy "in the same month"; the 1922 Nobel lecture adds Hevesy and gives the order (Hevesy and Russell partly correct; Fajans complete and first; Soddy a fortnight later, independently) [web: https://www.nobelprize.org/uploads/2018/06/soddy-lecture.pdf, printed p.387]. Both are Soddy's own accounts (he has a stake in priority); the word "isotope" was suggested to him by Margaret Todd [web, see check 2]; (iv) the figure 206.05 is not traced to a named experimenter; (v) thorite lead 207.90 is Hoenigschmid's; Soddy's own density-derived value is 207.74.

C6. Atomic structure and nuclear energy. Split into four: - C6a. The atom has a tiny dense nucleus whose charge is the atomic number (Rutherford, Moseley). - C6b. Alpha bombardment of nitrogen knocks out hydrogen (Rutherford 1919); Soddy's surmise of the product (carbon-12 plus two hydrogen) is wrong, the real reaction is 14N + alpha -> 17O + p. - C6c. Rutherford's June 1920 report of a mass-3 doubly charged particle from nitrogen and oxygen (Soddy reports it; he is the reporter, not the claimant). - C6d. Releasing nuclear energy by induced transmutation is real but unsolved; success would bring great energy and great danger; Soddy's estimate that two to four million volts would be needed. - Also (changed): Soddy's speculative proton radius of about 1e-16 cm. Changed: claims.md read as if the mass-3 particle and the hydrogen were both observed results; they have different fates.

C7. Radioactivity and the wider picture. Split into five: - C7a. Radioactive heat in the earth (Strutt, Joly); the earth's interior "steadily getting hotter" with alternating "geological" and "incandescent" ages (both books). - C7b. The sun: 1911 "mainspring of the universe" (hedged); 1920 radium cannot supply the sun's heat but the source may be internal. - C7c. "Energy and wealth are synonymous"; civilisation lives on a diminishing capital of coal. - C7d. Gold and silver may belong to one slowly changing series, gold changing faster. - C7e. The philosopher's stone and the Fall of Man may echo a forgotten race that mastered transmutation. Changed: claims.md lumped these as "speculative asides"; their fates differ and C7d is self-contradictory under its own premise.

---

4. Verdicts

C1. Exponential decay of single atoms

Agreement: all three, Established. Disagreement: none in substance. The contrarian adds a living literature on claimed solar or neutrino modulation of decay rates (Jenkins, Fischbach, Sturrock for; Norman 2008 and Pomme and Pelczar 2022 against) [web: https://arxiv.org/abs/1106.2374v1 , arXiv 0810.3265, via the contrarian; I did not re-open them]. The strongest honest reading: the disputed modulations are a few parts in 1,000 at most, are not reproduced by the better-controlled groups, and would modify "lambda is constant" by a tiny amount, not the exponential law. The critical 2022 review is by critics of the claim; I record its stance, not just its conclusion. - Mathematics internally consistent: yes (1.45 for 1/ln2 = 1.4427; 3.84 d and 5.57 d are 0.5% off from each other). - Consistent with experiment: yes. Tested and replicated: yes, universally. If false: all of nuclear and weak-decay physics and radiometric dating would change. Final verdict: Established. Qualifiers: strong form (independence from surroundings) is false for electron capture and bound-state decay (Re-187, [web: https://www.osti.gov/biblio/404092], via the contrarian); claimed modulation of decay rates: Tested, not supported (disputed, weakly). Soddy's phrase was correct for everything he could measure. Settles it: already settled for the law. A strict test of constancy: long-term multi-site decay counting in shielded and unshielded settings (done by PTB and NIST groups; continuing).

C2. Radium's heat is internal atomic energy

Agreement: all three, Established. They differ only in qualifiers, and my recomputation shows the differences are one range (123-147 cal/h) described three ways. Stronger argument on the 1911 value: the mathematician's (a printed sentence in the 1911 book itself fits 100 and not 133; I verified the sentence in the text). Kelvin, H. E. Armstrong, L. Poincare and others proposed external-energy sources (radiation, the ether) [web, via the Investigator: snippets of Badash 1966, Campos 2015]. That was a serious rival and it was decided by experiment: the heat tracks the amount and decay of radium's own emanation (Rutherford and Barnes 1903-04, [web: https://www.nature.com/articles/068622a0, abstract-level via the contrarian]); the per-atom energy is the mass difference, E = mc^2. Kelvin's own letters were not read by anyone; his side is known only through Soddy's reply and press snippets. - Mathematics: consistent (multiplying by the mean life is the right integral). - Experiment: yes; tested by Curie and Laborde 1903 [memory, high], Rutherford and Barnes. Replicated. - If false: conservation of energy would need a supply of energy from an unknown field. Final verdict: Established. Qualifiers: "complete change" means radium plus products to Po-214; "million times" basis stated; 1909-11 constant 100 was 25-30% low and Soddy himself revised it; the mean life is 8% high. Settles it: nothing open. A reader can reproduce the heat from decay energies (24.6 cal/h from radium alone, 123-147 with the short-lived products).

C3. Equilibrium, lives, uranium to lead

Agreement: Established, with numbers off. The mathematician's additions are accepted: the 58-versus-40 slip; the weight/atom-ratio 5%; the transient-equilibrium case not mentioned; the t-squared growth test measures T_Io x T_Ra and reproduces it to 0.4% (modern product 2.51e8, Soddy 2.50e8, [calc]). The reviewers differ on the weight to give the numerical errors: the contrarian emphasises that Soddy's 1911 "few per cent" was overconfident (8% for radium; true), the physicist that they are within 1911-20 analytical error (also true). Both go in. - Mathematics: consistent in the secular case; one sentence internally inconsistent (58 versus 40, verified on the page image). - Experiment: Boltwood, Rutherford, Soddy and Hitchins 1915-19, direct alpha counting. 1911 text honestly says the uranium-to-radium growth test "has not yet produced a definite result" [read: Matter and Energy p.224]. - If false: radiometric dating and the age of the earth. Final verdict: Established for the law, the chain and the logic of the growth test. Historically superseded but instructive for the numbers (radium +8%, uranium +16% to +24%, uranium II 3e6 y about eight times too long [web: Wikipedia isotope pages via the mathematician]). One sentence (58 g) is Internally inconsistent. Settles it: nothing open; the original raw data for the 1904 and later volume measurements are not in the folder.

C4. The alpha particle is helium

Agreement: Established. Note the nuance: the thin-glass experiment is Rutherford and Royds 1909 [memory, high; not web-checked here], the Ramsay-Soddy 1903 experiment is one in which Soddy took part, so his account is partly his own report; replication by others settles it. Final verdict: Established. Settles it: nothing open.

C5. Isotopes and the displacement rules

Agreement: Established for the concept, the rules and the lead result; "Historically superseded but instructive" for the strict integer rule, the two-isotope description of common lead, and several of Aston's 1920 table entries (sulphur, xenon 133 and 135, mercury; all [memory, high]; Aston flagged some with "(?)"). Priority: the physicist and mathematician report both accounts and so does the contrarian. Weighed: the 1922 Nobel lecture is the more detailed and in part against Soddy's own interest (he concedes that Fajans's was first complete); the 1920 "same month" does not contradict it. Both are Soddy's own words; neither is a neutral account. The deciding record is the dated 1913 papers of Hevesy, Russell, Fajans and Soddy (Physikalische Zeitschrift, Chemical News, Philosophical Magazine) and the Fajans letter of 19 Oct 1913 and the 1935 Noyes correspondence on priority listed in the Bodleian/CSAC catalogue [web: research/2026-10-08-archiveshub-csac-catalogue.md; letters unread]. I do not adjudicate priority between persons; I say which record would. Final verdict: Established (isotopes, displacement rules, lead-isotope result). Historically superseded but instructive (integer rule; two-isotope lead; Aston table entries). Priority: not a physics verdict; documented as plural and contested among interested parties. Settles it (priority): retrieve the 1913 papers and the Fajans and Noyes letters.

C6. Nuclear atom, nitrogen, nuclear energy

• C6a. Nuclear atom and atomic number: Established. Order-of-magnitude check: nuclear/atomic radius squared = 1e-10 against Soddy's quoted 1 in 1e9 hits [calc, same order].
• Proton radius 1e-16 cm: Historically superseded but instructive. Soddy labels it "merely a suggestion"; the mathematician's "refuted by calculation" is stronger than the evidence warrants, since what refutes it is measurement (0.84e-13 cm) plus the fact that proton mass is not electrostatic.
• C6b. Hydrogen from nitrogen: Established (Rutherford 1919; Blackett 1925 cloud-chamber photographs gave the capture mechanism [web, summary level, check 1]). Soddy's surmise (carbon-12 plus two hydrogen): Refuted by calculation (charge balance 6+2 = 8 versus 7) and by Blackett's experiment. Soddy himself says the data "strongly suggest, though they do not yet rigorously prove".
• C6c. Mass-3 doubly charged particle: Tested, not supported (Rutherford's report, [web, check 3]; follow-up work showed only protons and alpha particles; He-3 was identified in 1939 and is not what Rutherford saw). Soddy reports it in a note added while the book was in the press.
• C6d. Release of nuclear energy: Plausible, untested as of 1920 (the physics was right, the source real). It has since been realised: Cockcroft and Walton 1932 [memory, high], fission 1938-39, chain reaction 1942 [memory, high]. His route (alpha particles accelerated through 2-4 million volts) is Historically superseded but instructive: the classical Coulomb barrier overestimates the energy needed for protons on light nuclei because of tunnelling (Gamow 1928 [memory, high]). The warning ("great energy and great danger") is a social prediction, not a physics claim, and cannot be scored on the scale; it was borne out. The Investigator documents that Soddy warned of atomic war publicly in 1915 (Guston 2012, [web: https://pmc.ncbi.nlm.nih.gov/articles/PMC3456912/]; the 1915 text itself was not found; Sclove 1989 not read) and that H. G. Wells's The World Set Free (1914) is dedicated to the Interpretation of Radium and "owes long passages to the eleventh chapter" [web: https://www.gutenberg.org/cache/epub/1059/pg1059.txt]. The contrarian's point that the leading authority, Rutherford, was reported by The Times in 1933 as calling practical atomic power "moonshine" [web, check 4] is kept in the record, as that paper's report. Settles it: nothing physical is open. Historical: Soddy's 1946 manuscript "Real Alchemy and the Release of Atomic Energy", the 1947 pamphlet and the 1952 and 1954 Lindau lectures (unread) would show how his 1920 warning developed.

C7. The wider picture

• C7a. Earth heat. Radiogenic heat is a major part of the earth's heat budget: Established (about 20 TW from U and Th of about 44 TW surface heat flow by KamLAND and Borexino geoneutrinos [web: Gando et al., via the contrarian]; physicist's range 15-41 TW [web: Wikipedia]). Uniform radium throughout the earth, secular heating, an exploding earth and alternating geological and incandescent ages (both books; Joly's theory): Tested, not supported, and Refuted by calculation: uniform 2e-12 g Ra/g gives about 1,960 TW against 44-47 TW measured [calc]. Soddy writes conditionally ("if the known laws hold") and corrected the sun statement but not this one between 1911 and 1920.
• C7b. The sun. 1920 statement (radium cannot supply it; source probably internal): Established; 9.4e-8 W/g for uranium plus chain needs about two solar masses [calc]. The 1911 "mainspring of the universe": Historically superseded but instructive. Fusion (Eddington 1920 suggestion; Bethe 1938-39 [memory, high]).
• C7c. "Energy and wealth are synonymous." Physical premises (conservation, no perpetual motion, finite fossil stock): Established. As an economic identity: Not testable as stated. His economics (monetary reform, dedicated to Arthur Kitson) is outside the folder and was not weighed. Soddy's energy-based economics is cited as an origin of ecological economics (Daly 1980, Merricks 1996, [web snippets via the Investigator]).
• C7d. Gold and silver in one decay series: Refuted by calculation. The premise (Interp. PDF p.203 [read]: abundances proportional to mean lives, gold changing "many times more rapidly") is the equilibrium law, which holds only inside a chain with a long-lived parent. Gold is rarer, so it would be the shorter-lived daughter, and silver the parent. But decay never raises A (alpha lowers by 4, beta leaves it), and A would have to rise by 88 or 90 [calc]. Reversing the roles makes gold the parent and so longer-lived and more abundant, contradicting the premise. By Soddy's own displacement rules the idea therefore fails under his own premise, the contrarian's softer verdict "Tested, not supported" and the mathematician's "historically superseded" are also true (gold has one stable isotope, silver two; no decay observed) but the calculation is the stronger point and needs no experiment. In fairness: Soddy says "at least a possible view" and "Obviously we are only at the beginning". The contrarian's point that no test was designed for it is right: it was refuted by logic and by stable-isotope data, not by a test of the claim. Soddy's own 1905 request to the British Museum to examine old coins "for chemical reaction" (declined, CSAC item 217 [web, catalogue entry; letter unread]) might be connected; the purpose is not stated, so that link is an inference only.
• C7e. A forgotten race that mastered transmutation: Not testable as stated. [read, Interp. PDF p.230] "in a past possibly so remote that even the very atoms of its civilisation literally have had time to disintegrate": the last clause explains the absence of relics, so no observation could refute it (contrarian). The physicist adds the corollary that a transmuting civilisation would leave isotopic traces (Oklo shows natural reactors leave them [memory, high]); none is known. The passage is in the 1909 first edition, Chapter XI [web: archive.org, via the Investigator], was the passage Wells used, and is read by Morrisson (Modern Alchemy) and the Brill volume Lost Knowledge as part of an early-20th-century link between alchemy and atomic science [web, snippets only; not opened]. Kept as part of the record.

---

5. Corrected claim list: summary of final verdicts

(Superseded where it differs by section 10.5, the QE revision: adds C6e, C8a-e, C9; C6d now leads with Established; C2 and C3 gain inconsistent sentences.)

Verdict-scale note: "Never properly tested" does not apply to any claim here: nothing in these two books was dismissed without testing. C7d and C7e were never experimentally tested as such because they are refuted by logic or are not testable.

---

6. What would settle it

• C1 strict constancy: long-term, multi-site, shielded decay counting (PTB, NIST) against the Sturrock-Fischbach modulation claims.
• C2: nothing open. Reader's check: 24.6 cal/h from radium alone and 123-147 cal/h with products, from decay energies [calc above].
• C3: nothing open for the law. The raw 1904 and later gas-volume data would show whether the 0.6 mm^3 emanation figure was derived from the adopted 2,500-year life (the mathematician's caveat) or measured independently.
• C5 priority: the 1913 papers (Hevesy, Russell, Fajans, Soddy) and the Fajans letter of 19 Oct 1913 and the 1935 Noyes correspondence (CSAC items 217-218 and later).
• C6d development of Soddy's views: the 1946 manuscript, 1947 pamphlet, 1952 and 1954 Lindau lectures, 1956 NYT obituary on his hydrogen-bomb-test views.
• C7c: Soddy's economics books (not in the folder): the monetary-reform half of "energy and wealth" is outside this review.
• C7d: a decay-series chart (done: no route). C7e: no experiment is possible; Oklo-type isotopic anomalies would be the only physical test, none known.
• Kelvin's own 1906 Times letters and Armstrong's side: would let the reader weigh the opposition from its own words, not through Soddy's reply.

7. Open questions

1. How much of the gap between 100 and 133 cal/h in Soddy's own text came from better calorimetry and how much from the 25:75 emanation split? The 1909 edition gives 25 radium plus 75 emanation and products; modern alpha energy gives 24.6 for radium alone, so the split was right and the total was low. Which experimenters supplied 133 is unknown (no source found).
2. Where does "58" come from? Misprint, half-life/mean-life mix-up, or a different input? Only an earlier edition could say. The 1909 edition on archive.org is the obvious check.
3. Is 3.2e6 a mass ratio or an atom ratio in Soddy's reasoning? (5% effect.)
4. Why did Soddy keep Joly's heating-earth picture in 1920 after correcting his view of the sun? A question of his sources, not of physics.
5. What did Soddy intend by the 1905 British Museum coin request?
6. Who suggested he put the "forgotten race" passage in Chapter XI of 1909, and how much did he mean it? He hedges ("Can we not read into them"); a reader must weigh it.

8. Absence-of-evidence analysis

Where the record goes dark (none of it touches the physics claims above).

A. Soddy's wartime papers and late letters. [web: research/2026-10-08-archiveshub-csac-catalogue.md] The Bodleian/CSAC catalogue says the Aberdeen war-work envelopes were "probably ... disturbed, and some of them removed", and that the originals of his 1953-54 letters to Muriel Howorth are lost (only Trenn's 1969 copies of 32 survive). Secret patent 5527/16 (fireless boilers, 1916) was not found in Google Patents (404 on GB191605527A). Competing explanations: 1. Ordinary loss or reorganisation by his literary trustees. For: the catalogue's own wording ("probably"), the long gap between his death (1956) and cataloguing (1974); Soddy's own 1952-56 re-sorting "intending to resume work on certain inventions". Against: nothing found. 2. Removal of classified wartime material. For: the topic (fireless boilers, steam turbines, concentrators for the Admiralty, 1916-17) is war work; wartime secret patents may never have been published. Against: no record of who removed what, or when. 3. Selection by Howorth for her 1958 book. For: only she held the originals. Against: none found; Trenn holds copies of the other 32. 4. Deliberate suppression for political reasons (Soddy was a vocal critic of banks and of private radium supply and, late in life, of the hydrogen bomb tests). For: his controversial public positions. Against: no document, witness or correspondence in the files supports it; silence does not show it, and the same gap is explained by 1-3. Records that would decide: Admiralty Board of Invention and Research files (National Archives, ADM series, not searched); the Patent Office record of 5527/16; the Bodleian box-level accession notes; Howorth's 1958 book and her papers; Trenn's 1969 copies. The Investigator did not reach the National Archives (time limit); this is the first place to look.

B. Unread primary sources that exist but were not obtained. Not silence in the record but gaps in this review: Kelvin's letters (The Times, Aug 1906), the 1913 isotope papers, Soddy and Hitchins 1915-19, Rutherford's 1919 and 1920 papers, Hoenigschmid's lead papers, Sclove 1989, Hull thesis. Each is held by libraries; none is suppressed.

C. The Glasgow and West of Scotland Radium Committee minute book and the Knoefler dispute over mesothorium (GB 25,504, 1910). Soddy is the patent holder and a public campaigner; the German firm's side is absent from the record. This is Soddy's characterisation of an episode in which he had a stake; no second account was found.

---

9. Corrections to earlier documents (for the orchestrator)

1. claims.md claim 2: add (i) 1911 constant is 100 cal/h, stated in the 1909 edition and implied by the 1911 "38 hours" sentence; (ii) "complete change" means radium plus its quick products to Po-214 (about +23% for the full series); (iii) "million times" is per gram of coal plus oxygen (3.6e5 per gram of coal alone); (iv) 133 sits inside the modern 123-147 range.
2. claims.md claim 3: Soddy's lives are 8% (radium) and 16-24% (uranium) above modern; add six betas; add the 58-versus-40 slip (Interp. printed p.243); add the weight-versus-atom assumption.
3. claims.md claim 5: add Hevesy and the order from the 1922 Nobel lecture; the 1920 book names Russell, Fajans, Soddy. Add that the word "isotope" was suggested to Soddy by Margaret Todd (1913) [web]. Ordinary lead is four isotopes. 206.05 has no named experimenter; thorite lead 207.90 is Hoenigschmid's.
4. claims.md claim 6: split into 6a to 6d (see section 3); the mass-3 particle was not confirmed; the hydrogen from nitrogen was.
5. claims.md claim 7: split into 7a to 7e; the gold-silver aside is self-contradictory under its own premise; the 1911 and 1920 books differ on the sun.
6. deep/interpretationof00sodd-f5dd74.md section 4: "roughly 27-28 MeV, 135-140 cal/h" should read "about 24-29 MeV, about 123-147 cal/h".
7. deep/matterenergy...md: the 1911 100 cal/h is no longer an inference alone (38-hours sentence plus 1909 edition). Chapter X was read by the contrarian, pp.240-250 (Joly's "steadily hotter" and "must explode" passage, capital-of-coal argument); facts.json how_read should say so.
8. facts.json Interpretation: add the "58 g" and the "lead an alpha and a beta gives gold" slips as printed errors in the book.
9. physicist.md (my correction, not Step 1): the Claim 5 ionium paragraph says "Soddy's 58 uses the 'forty times longer'" and is "20% different"; the correct statement is that 58 contradicts forty and the book's own table (see section 1). The physicist's statement that the mass-3 particle is "a mis-measured He-3 not identified until 1939" is better put as: an early mistaken hint (He-3 was proposed in 1934 by Oliphant and found in 1939).
10. mathematician.md: the "Needs page-image check" for the gold recipe is resolved: the error is in the book (page 233 as extracted), not in the deep read. The mathematician's "Refuted by calculation" for the proton radius is better stated as "superseded; refuted by measurement" (section 1).
11. contrarian.md section 7c: "gold ... from removing nucleons from bismuth-209 at the Bevalac in 1980" is from a search summary only (yield sources conflict); retained as tagged. Not verdict-bearing.
12. Stale-copy search for the orchestrator: the figures "uranium:radium 3,200,000" and "uranium life 8,000 million years" should name the edition (1911: 3,000,000 and 7,500 million; 1920: 3,200,000 and 8,000 million); also grep site/ and work/rankings/entries/ for "58 grams", "complete change", "two to four million volts", "forgotten race" and "gold and silver" to check wording against this file.
13. Previously noted and still valid (claims.md): Roentgen X-rays are dated "1905" in Matter and Energy p.204 (should be 1895); the error is in the book.

---

10. QE revision (2026-10-08)

Inputs: qe.md (verdict MINOR), the deep reads added since the first arbiter pass (Interpretation PDF pp.49-57 and 106-132; pp.141-146, 169, 195; pp.207-215; pp.216-221; Matter and Energy PDF pp.13-42, 101-118, 236-257), and my own reading of the extracted text of Interp. PDF pp.247-251 (actinium) and of the Investigator's local copy of the 1909 first edition (research/pages/ia-interpretationof00soddrich.txt, from archive.org). Tags as in the header; [calc] is my own Python arithmetic. Web budget: 6 of 8 used (the new check is item 7 below). The QE's deep-read request 3 (actinium, PDF pp.246-256) was NOT run; item 5 rests on the text I read myself, not on a deep-read file, and says so.

Result: no verdict on the original seven claims is reversed. Three printed-text inconsistencies are added to the record (helium, the "thirteen per cent", and the earlier "58 g"), one claim from facts.json that had been dropped is restored (life and the second law), two investigator leads that had been dropped are restored (Ramsay-Cameron; Le Sage as part of the same chapter), and one new historical item is added (the actinium branch). C6d is restated to lead with the present-day verdict.

10.1 Accepted QE findings that change nothing but wording

1. "Nearly a million times" was not generous. I accept QE section 4 item 3 and withdraw the earlier reviewers' and my own implied reading. [read: Interp. PDF p.164] compares radium with "a similar weight of matter undergoing any change known previously"; [read: PDF p.219] gives the coal-alone factor himself as "about 360,000"; [read: PDF p.53] gives 1,400 times coal for a half-grain of radium bromide over sixteen years, with coal weighed against the bromide. [calc] 2.91e9 cal/g divided by the 3,792 cal/g of hydrogen burning to water is 7.7e5, "nearly a million"; against coal plus oxygen 1.3e6; against coal alone 3.6e5. The three figures are consistent. For the half-grain: 32.4 mg of RaBr2 gives about 123 cal against hydrogen and oxygen, so 350,000 cal is about 2,900 times that: Soddy's wider sentence ("any other kind of substance") also holds. C2 is unchanged and now has a defended basis.
2. The 133 cal/h figure still has no named experimenter. [read: Interp. PDF p.52] "Exact experiments have proved that 1 gram ... of radium gives out 133 calories per hour." The Curie-Laborde and Rutherford-Barnes attributions in section 4 C2 are [memory, high] / [web, abstract level], not from this book. The arithmetic of the half-grain illustration is internally consistent (2.5 cal/h x 16 y = 3.5e5 cal; the true Ra fraction is 0.585 where Soddy uses 3/5; 2.52 against 2.58 cal/h). Soddy was the lecturer and owner of the specimen: the figure is presented in service of his own first alternative (internal energy).
3. The "1905" misprint is systematic (QE section 4 item 2; four places in Matter and Energy for 1894-95). Correction 13 below is replaced.
4. Joly, not Kelvin, assumed two parts of radium per million million (digest error; arbiter and primer already attributed it to Joly). [read: Matter and Energy PDF p.240-241] Joly: that amount raises the interior by 1,800 C in 100 million years and "a world so constituted must explode". This is still the passage under C7a.
5. A reading of the 1911 Chapter X sun-shrinkage figure. [read: Matter and Energy PDF p.237] "every thousand years the diameter shrinks about 40 miles". [calc] Kelvin-Helmholtz contraction of a uniform sphere, with half the released gravitational energy radiated, supplying the solar luminosity for 1,000 years, needs about 74 km, 46 miles. Soddy's number is right to 15%. His two objections to shrinkage are weaker than his number: "no ground for believing" it is shrinking is an absence-of-observation argument, and "why should the sun shrink unless it is falling in temperature" is loose (a contracting gas sphere heats up, the virial theorem [memory, high]). His real argument, the time scale, is the sound one (hundreds to thousands of millions of years against Kelvin's tens of millions of years).

10.2 New and restored claims, with verdicts

C8. Life and the second law (restored from facts.json; Matter and Energy PDF pp.101-109, deep read of those pages). Split, because Soddy makes three separate statements and hedges all of them. - C8a. A "molecular intelligence" would vitiate the second law [read: PDF p.104]. This is Maxwell's demon, which Soddy does not name (identification [memory, high]). Verdict: Historically superseded but instructive. As a conditional it is correct and was the standard view among the best exponents; what 1911 lacked, and later work supplied (Szilard 1929, Landauer 1961, Bennett 1982 [memory, high]), is the accounting: the demon's information processing carries an entropy cost that restores the second law. Soddy's own gloss, that the individual molecules of a gas at one temperature differ widely in speed and "there seems no reason why" molecules should not utilise this [read: p.105], is a hedged speculation, not a result. - C8b. Living cells may be molecular sorters that beat the second law ("Have the minute cells of the body the power of taking advantage of the difference in the temperature of the molecules bombarding them ...?" [read: PDF p.106]; "time alone and experiment will answer it decisively"). Verdict: Tested, not supported, with a caveat that no experiment was designed for this exact proposal. What counts against it is the general record: organisms are open systems that export entropy; whole-body and muscle calorimetry balance energy and stay inside the thermodynamic limits; and the physical analysis of Brownian ratchets (the Feynman ratchet and later work) shows a ratchet cannot rectify thermal fluctuations at one temperature [memory, high for the ratchet analysis; medium for the early respiration calorimetry; not web-checked]. Soddy offered it as a legitimate open question and not as a finding ("a fascinating and a legitimate line of inquiry"; efficiency "far too little is known" [read: p.108-109]). - C8c. Food energy changes directly into work without first becoming heat, so the Carnot limit does not govern it [read: PDF p.108]. Verdict: Established, with a misframing. The Carnot formula (T1-T2)/T1 [read: p.104] binds heat engines only; chemical and electrochemical conversion (fuel cells, muscle) is bounded by the free-energy change, and can exceed the Carnot figure for the same temperatures [memory, high]. Soddy is right that "the second law would not then operate" in the heat-engine sense. He is wrong, or at least careless, to put this beside an "opinion" that life processes "cannot possibly obey" the second law: they obey it, through the free-energy bound. Note that his "consensus of opinion" is attributed to unnamed others, and his own text calls the efficiency of the living body unknown. - Side note on his 27% (PDF p.102) [read; paraphrase by the deep reader]: cooling a body from 373 K to 273 K and calling the limit 100/373 = 26.8% gives the same number as the Carnot figure for a 373 K source and a 273 K sink, but that is a coincidence of the treatment. A reversible engine running a finite body of constant heat capacity down to 273 K delivers 14.8% of the heat taken out [calc]. The Carnot form given on p.104 is correct. - Forces as fictions; "tractate" and "pellate" [read: pp.109-116]. Soddy's proposal that "force" is a positional shorthand, not a thing, and that potential energy is "a way of expressing facts, not supposed explanations". Verdict: Not testable as stated (a terminological and philosophical position; it makes no different prediction). Kept for the record: he later argued, and the Investigator notes, a similar energy-first stance in his economics. - Le Sage gravitation [read: pp.116-118]. Soddy says it is "not accepted, though very suggestive", that it "explains beautifully" the inverse square, and that the two explanations (Newton's force and Le Sage's corpuscles) "are on a par" but Le Sage's would be "knowable". His one stated difficulty: screening goes with area, not mass. He proposes no experiment and does not say what other difficulties there are. Verdict on Le Sage models: Refuted by calculation in the form Le Sage gave (Maxwell and Poincare's heating and drag arguments [memory, high]; the area-versus-mass difficulty Soddy names is one of the standard objections); Historically superseded but instructive as to Soddy's use of it, which is methodological: a hypothesis that posits real, ultimately detectable entities is a better kind of explanation than a restatement. Le Sage-type gravity is still discussed in alternative science, so the record is kept and the physics stated fairly.

C6e. Ramsay and Cameron's copper-to-lithium claim (1907), restored (Investigator lead; physicist.md claim 6). Verdict: Tested, not supported (Curie and Gleditsch found lithium traceable to the glass; Soddy's own negative copper and gold results in the 1909 Chemical Society report [web, snippets via the Investigator; originals unread]); the proposed mechanism is Refuted by calculation: 5.5 MeV alpha from radium emanation against copper has a Coulomb barrier of about 12.5 MeV [calc, above], and removing 26 protons is not a nuclear process available at those energies. This is the documented test and failure of a pre-1919 transmutation claim by a Nobel laureate, and it belongs beside C6b and C7d: it was rejected by experiment, not by authority.

C9. The actinium branch (restored; Interp. PDF pp.246-251, text I read; no deep-read file). [read: PDF pp.247-250] Soddy argues that actinium cannot be in the main uranium-polonium series, since "at least five a-particles" would leave no room between uranium and polonium, and since its alpha activity in uranium minerals is "only about one-fifteenth or one-sixteenth" of the uranium series where five-eighths would be expected. He sets two possibilities: actinium is "an entirely separate and independent primary radio-element" (and then "its occurrence always in uranium minerals ... is difficult to understand"), or a branch of uranium, in which case "out of every twelve uranium atoms, eleven go through the main line ... and one goes through the actinium line" (about 8 per cent), a view "now supported by much new evidence", including the branches at thorium C (35 / 65) and radium C (0.03 / 99.97) per cent. [read: PDF p.252, digest only] he reports the 1919 discovery of the parent of actinium by "Cranston and the writer" and by Hahn and Meitner. Modern picture [memory, high; Dempster's 1935 discovery of U-235 and its 0.72% share of natural uranium, [web, summary level] https://en.wikipedia.org/wiki/Arthur_Jeffrey_Dempster and https://en.wikipedia.org/wiki/Uranium-235]: the actinium series descends from U-235, a separate isotope of uranium, not from a branch of U-238; U-238 has no such branch. [calc] U-235 activity is 4.6% of U-238 activity, so the actinium series (seven alphas) gives 4.6% x 7/8 = 4.1% of the uranium series' alpha activity, about 1/25, against Soddy's 1/15 to 1/16, and the "branch fraction" would be 4.4%, against his 8%. So the observation was right to within a factor of 1.6, and his arithmetic from it (about 8 per cent) is high by a factor of 1.8, because the true share is a separate uranium atom. Verdict: Historically superseded but instructive. The lesson for the reader: chemistry cannot tell a separate uranium isotope from a branch of the main line, because the two are chemically identical; Soddy's two alternatives (a separate primary element always found with uranium, or a branch) are each half right, since actinium's parent is a separate primary parent and it is a uranium isotope. The branch idea was a sound way of explaining the data in 1920 and it points toward the isotope concept he had himself named in 1913. - Protactinium priority (Soddy and Cranston; Hahn and Meitner). Soddy is an interested party. [web, summary level, https://en.wikipedia.org/wiki/Protactinium, via search result] states that the two groups discovered Pa-231 independently, that Hahn and Meitner's paper is dated 1918 ("Die Muttersubstanz des Actiniums"), and that Soddy and Cranston had reported indirect evidence "less than three months earlier"; credit is usually given to Hahn and Meitner. I record both and do not adjudicate. The deciding record is the dated papers (Proc. Roy. Soc. A and Physikalische Zeitschrift, 1918) [memory, medium]; not retrieved. It sits beside the isotope priority in C5.

10.3 Changes to C1, C2, C3, C5 and C6d

C1 (exponential decay; constancy). Verdict unchanged: Established. The deep reads add three items of evidence for constancy and one negative test that were not weighed: - [read: Matter and Energy PDF pp.246-248] Soddy states decay as "uninfluenced by external conditions ... qualifies it to be used as an absolute standard of time". He reports a test, second hand: if bombardment by the alpha particles from neighbouring atoms sped up decay, a concentrated radium preparation should decay faster than a diluted one, "but this has been found not to be the case". No experimenter, date, dilution or precision is given. Verdict on the induced-decay-by-alpha-neighbours hypothesis: Tested, not supported (a reported negative; no data in the book; consistent with modern knowledge, since 5 MeV alpha particles are stopped by the Coulomb barrier long before they reach a nucleus [calc, above: 4.2 MeV nitrogen, 12.5 MeV copper]). He then concedes that "the only process known at all likely to be competent to affect the stability of an ordinary atom is the change of a neighbouring radioactive atom". That concession is exactly what Rutherford's 1919 nitrogen experiment realised: an alpha from a radioactive atom acting on a stable atom. The 1911 pessimism ("how little hope there is for a long time") held for the rate of decay and failed for transmutation. - [read: Interp. PDF pp.165-167] Pleochroic halos: Joly and Fletcher (Phil. Mag., 1910) measured ring radii that match the alpha ranges of the uranium, thorium and radium products, formed from "sometimes less than 100 a-particles per year, continuing for several hundred million years". The fixed ring radii argue that alpha energies did not change over that time, which is the stronger of the two constancy inferences; constancy of the rate itself is only implicit, since Soddy gives no independent age for the mica. [calc] 100 a year for 3e8 years is 3e10 alphas, the activity of about 1e-16 g of radium. A later, contested use of halos (Gentry) is not in the book and is not weighed here. - [read: Interp. PDF pp.207-215] The alpha-range table (25 to 69.5 mm in air; 8,800 to 12,400 miles per second, about 5-7% of light speed [calc]). Radium F is the exception to the trend, as the table itself shows.

C2 (radium heat). Verdict unchanged: Established. Add: - The rival external-energy hypothesis in Soddy's own words [read: Interp. PDF pp.106-109, 126-129]: "Either the energy must be derived from within the radium ... or it must be supplied from outside"; radium would be "a transforming mechanism", like an electrical transformer on a high-tension line. Soddy concedes "so vague a view ... cannot be directly disproved". He names no proponent (the Investigator names Kelvin, H. E. Armstrong and L. Poincare, [web, snippets]). His objections: radioactivity cannot be quenched, in a mine hundreds of feet down it is normal, so the external supply would have to pass "without loss hundreds of feet of solid rock"; the emanation, mass for mass far more radioactive than radium, gives three times its parent's heat (33 against 100 cal/h per gram, [read: PDF pp.125-126]); its activity decays by half every four days; and radium's recovery over a month "on the second alternative ... no answer whatever can be given". "Straining at a gnat and swallowing a camel" is rhetoric by an interested party (Soddy co-authored the disintegration theory). I weigh the argument as follows: his first objection (non-quenching) does not by itself exclude a penetrating field, and he partly concedes it, so it rests on incredulity; the decay and recovery of the emanation, and the fact that the heat tracks identified atoms whose mass difference gives the energy by E = mc^2 (not mentioned in this book), are the evidence that decides it. The verdict is not changed. - Tait (1876, quoted by Soddy) limited the earth to "ten or at most fifteen millions of years" and found the sun's chemical fuel good for "5,000 years", both conditional on knowing "all the physical laws". Soddy uses the conditions against Tait, a fair use. Page numbers in Tait's book were not checked [not verified]. - The "thirteen per cent" (Interp. PDF pp.218-220): Internally inconsistent (sentence level; the argument is unaffected). [read: PDF p.218] "uranium evolves ... some thirteen per cent more energy than ... the same weight of radium", giving a pound of oxide "at least one hundred and sixty tons of coal". The 1920 chain (Fig. 34 and the p.199 table) has eight alpha emitters from uranium to lead and five from radium on, so by Soddy's own principle (equal energy per alpha) uranium gives 8/5 x 226/238 = 1.52, about 52 per cent more per gram [calc; the deep read gets the same, and 39% with real alpha energies, [memory, medium]]. I also checked the 1909 first edition: [read: local copy of the 1909 edition, around the passage on "the internal energy of all matter"] it reads "at least some fourteen per cent more energy". That is reproduced by one extra alpha: 6/5 x 226/238 = 1.14 [calc]. So the figure is a carry-over from the 1909 chain (and it was an "at least" figure then), changed by one point and without the "at least" in 1920 when the chain had grown to eight alphas. This settles QE's inference ("looks like a carry-over; the 1909 edition was not checked"). The 160-ton figure [calc: 397 g x 2.9e9 cal/g x 1.13 / 8,000 cal/g = 163 t; unrounded 1.00 gives 144 t] stays a lower bound ("at least") under every version; with the book's own chain it would be about 219 t, and with modern full-chain energy about 250 t (QE) to 300 t (deep read). Add that the 2.9e9 cal/g is the radium-to-Po-214 figure (four alphas), so the radium side of the 13% comparison omits the Po-210 alpha; that is a second reason the percentage is off, in the direction already stated. The "thousand times as much as the uranium itself" is a value claim, not reproduced from the text ([memory, low] on Edwardian coal prices; the deep read could not reproduce it either). - The 1909 first edition also gives 100 cal/h and the 25 / 75 split (QE and section 4 C2 stand).

C3 (equilibrium, lives, uranium to lead). Verdict unchanged (law, chain, growth test: Established; numbers: Historically superseded but instructive). Add a second internally inconsistent sentence, from the helium deep read (Interp. PDF pp.141-146, 169, 195): - "I obtained the same result by direct experiment from the rate of production of helium from uranium" (PDF p.169): Internally inconsistent within the 1920 edition. The facts, all [read]: Soddy worked "for four years" on helium from uranium and thorium, with seven separate large flasks of salts in solution (Fig. 26), calcium absorption of other gases, and a micro spectrum tube; detection limit one millionth of a cubic centimetre (one five-thousand-millionth of a gram) by the D3 line; typical sample "not more than a thousandth part of a cubic millimetre"; the rate from uranium "about two milligrams of helium from a thousand tons of uranium per year", which is 2e-12 of the mass per year, the "one five-hundred-thousand-millionth" [calc, agrees]. The text gives no uranium or thorium masses, no accumulation times, no table of runs and no separate thorium rate, and calls the agreement "almost exactly" on p.142 and "of the same order" on p.145. [calc] The printed prediction equals one alpha per uranium atom with a life of 8e9 years (4/238.5/8e9 = 2.10e-12). But p.195 of the same edition gives uranium two alpha emitters, uranium I and uranium II, which cannot be separated chemically, so a purified salt gives two helium atoms per uranium atom: the prediction is 4.2e-12 per year, and the measured 2e-12 would imply a life of 1.7e10 years, twice the 8e9 of the radium-ratio route. The 1909 text has the same experiment and the same words except "recently succeeded"; I checked the 1909 edition and it has the one-alpha picture [read, local copy: "only one a-particle is expelled from each atom"]. The 1920 preface says the text was "amplified rather than modified". So the helium passage is 1909 text carried into 1920. Modern value [memory, high on the inputs]: U-238 mean life 6.45e9 years, two alphas from U-238 and U-234: 5.2e-12 per year, about 2.6 times Soddy's measurement, which is low by that factor on any reading (helium held in solution or lost in extraction are possible causes; the text does not say). Fair weighing: as a threshold-detection measurement, a factor of two is "of the same order", so the experiment shows the order of magnitude (about 1e10 years) under either picture; it cannot honestly be called an independent confirmation of 8,000 million years. The radium-ratio route (3.2e6 x 2,500 = 8e9, weight ratio corrected by 226/238.5 to 7.6e9) stands, and the 1920 table correctly assigns the 8e9 to uranium I. Soddy was the experimenter and has a stake in his own measurement: the figure is his own account, and no one else is reported to have repeated it in the book. - Misprints and inconsistencies in the same edition [read, page image]: PDF p.237 "uranium I, with its period of 8,000,000 years" (should be 8,000,000,000; the page image shows it is the book's error); uranium II given as "about two million years" (PDF pp.195, 211) and "3,000,000 years (?)" (tables, pp.199, 208). - Mean life of radium C' "1/1,000,000th sec (?)" (PDF p.199, p.208, with the query mark in the book). [calc] Po-214's half-life is 164 microseconds [memory, high], mean life 237 microseconds, so Soddy's guess is 240 times too short, not "within a factor of 100 or so" as the deep read says. The deep read's other modern comparisons stand. - The range-period rule [read: PDF p.210]: log period against log range is straight for each series, three parallel lines, "the reason ... still obscure". This is the Geiger-Nuttall law [memory, high], explained in 1928 by Gamow and by Gurney and Condon as tunnelling [memory, high]. Verdict: Established (as the empirical rule Soddy states); mechanism 1928. Soddy used the line to estimate ionium at 200,000 years before it was measured at 100,000, and uranium II at two to three million years against a modern mean life of about 354,000 years (half-life 245,000, [memory, high]). The deep read's own fit of the book's table gives about 350,000 years for ionium, not 200,000; the line is so steep (period proportional to about range to the -52 [calc, deep read]) that errors of three or more are the expected accuracy. Soddy's "approximate" is the right word. - The 58-versus-40 sentence (section 1) is not in the 1909 edition (ionium is mentioned once in the OCR text there; I did not find the sentence). Open question 2 is therefore narrowed: it is a 1920 addition, and its source cannot be settled from the 1909 edition. Not checked: the 1912 third edition.

C5 (isotopes and priority). Verdicts unchanged. Add: Fleck and van den Broek credits ([read: Interp. PDF pp.276, 279], QE) and the protactinium priority statement above. The word "isotope" and the order of 1913 discoveries remain as in section 4 C5.

C6d (nuclear energy). Restated with the present-day verdict first, as QE correction 5 asks: Established, by routes other than the one Soddy proposed (induced reactions by accelerated protons 1932, fission 1938-39, chain reaction 1942 [memory, high]); in 1920 the verdict was Plausible, untested. His 2-4 million volt alpha route: Historically superseded but instructive. The warning (great energy and great danger) is a social prediction and cannot be scored. The 1911 chapter [read: Matter and Energy PDF pp.245-246] has "a pound of uranium contains and evolves ... the same amount of energy as a hundred tons or more of coal", and "little more than 1/10,000,000,000 part of this is given out every year". [calc] The pound gives about 300 t of coal (51 MeV per atom, [memory, medium]), so "or more" is correct; 1.3e-10 per year with his own life of 7,500 million years, so "little more than 1/10^10" is fair. The yearly release is about 1e3 J per pound, trivial. His diagnosis was that control, not discovery, is the barrier; the chain reaction later overturned "excessively slow" for U-235 [memory, high]. His 1911 statement that attempts to influence radioactivity are "signal failures", "like trying to influence the course of a bullet by blowing at it" is a reported negative result with no experiment named. Chapter X is also where he says radioactivity "is competent to be regarded as the mainspring of the universe" (hedged: "whether it be so or not") and, in the same chapter, that the "forgotten savage who kindled the first artificial fire" is a figure for the discovery of atomic energy; those are in the record under C7b.

C7a (earth heat). Add the Strutt claim as Soddy reports it [read: Matter and Energy PDF p.239]: rock-average radium in a crust only 50 miles thick would supply all the heat lost by the earth. [calc, from earlier section 1] This is the correct order of magnitude for the crust, and the modern budget puts about 20 of 44-47 TW in U and Th [web, section 4]. Soddy's own hedge, that the central argument is "necessarily somewhat hypothetical, for, under the enormous pressures ... even radioactive changes may not run their normal course" [read: p.240], was the right caution and does not rescue the uniform-radium picture (section 4 C7a).

C7c (energy and wealth). Verdict unchanged. The deep read of Chapter I (PDF pp.13-42) gives the exact wording [read: p.38]: "The discovery of a large quantity of gold leaves the world not a penny the richer ... But energy and wealth are synonymous. Energy is the thing of which gold itself is but the guinea stamp, to adapt the simile of Burns." Points to add: (a) the 1911 book contains no monetary-reform content, only the premise; the claim that the 1911 book "already contains" the reform would be wrong; (b) atomic energy is a prospect "physically possible" but "by no means of necessity", so "boundless energy" is conditional in 1911; (c) the capital/income distinction ("augment the natural supplies out of capital, under the delusion that it was supporting itself out of income" [read: p.41]) is Established as physics: fossil fuel is an ancient solar store that is being drawn down; (d) the coal numbers are right: 8,000 cal/g, 1 ton of coal against a 20,000-ton liner lifted 500 feet (3.4e10 J against 3.0e10 J [calc, deep read]). The prices at the pit head "represent the cost of bringing it to the surface", not what it cost Nature: a statement about external cost, still the standard objection to fossil-fuel pricing. The link to the later economics is documented in the book only as the same real-versus-token distinction; Soddy's later books were not read. Also in Chapter X, the Moissan carbide hypothesis for petroleum (reported as Moissan's view, [read: pp.252-253]): Tested, not supported as the main origin of petroleum (biogenic markers dominate, [memory, high]); small abiotic hydrocarbon sources exist [memory, medium]. Kept in the record as pertinent, stated as the French chemist's hypothesis reported by Soddy. Soddy's dismissal of direct solar and tidal engines ("doubtful ... will ever suffice", p.253) is an assertion with no calculation and was overtaken by photovoltaics [memory, high].

C7b (sun). Add 10.1 item 5 (the shrinkage figure is right; two objections are weak). The 1911 "mainspring of the universe" passage is hedged and its qualification is that radioactivity "including ... processes involving atomic transformations" is "competent to be regarded" so; he also says the solar source is "in the individual atoms" without naming radioactive decay. The earlier contrarian-led point (1911 and 1920 differ) stands. The "sun made of pure radium would shine a million times too brightly" [read: Interp. PDF p.56] is a thought experiment, not a claim; [calc] consistent with 9.4e-8 W/g for uranium plus chain in section 1 only as an order of magnitude.

10.4 Digest-only items still not weighed by a deep read (state of the record)

• Interp. PDF pp.246-256 (actinium and the thorium series): read by me for p.247-251 only; pp.252-256 (protactinium, "uranium Y" by Antonoff, the table of complete series) digest-only. QE request 3 remains open.
• Matter and Energy PDF p.141 (atmospheric emanation, "hundreds of tons of pure radium" equivalent), Interp. PDF p.287 (missing elements, rhenium and technetium), p.290 footnote (Merton and Hartley photochemical separation of chlorine isotopes): QE rated these low priority; verdicts for them are not given. [calc by QE, memory low] The atmospheric figure is plausible in order of magnitude.

10.5 Updated verdict table (supersedes section 5 where they differ)

10.6 Corrections to earlier documents (additional; for the orchestrator)

14. Correction 13 above is replaced: Matter and Energy misprints 1894-95 as "1905" in four places (PDF pp.128 twice, 165, 204) [QE section 4 item 2].
15. facts.json, Matter and Energy entry (summary, relevance, mainstream_conflict): the 1911 book names ionium ("an intermediate substance, ionium", PDF p.224); delete "ionium unnamed" and "names it only as an intermediate substance" (QE correction 1). Verified against the deep read.
16. physicist.md Claim 2a, mathematician.md Claim 2a and table 2b, contrarian.md (section 2 item 7, section 5 item 6), deep/interpretationof00sodd-f5dd74.md section 3, primer.md step 6: "nearly a million times" is a comparison with any known chemical change, not with coal alone (see 10.1 item 1). The primer sentence to use is QE correction 3.
17. deep/interpretationof00sodd-f5dd74.md, range section: the Radium C' period guess is 240 times too short ([calc] 237 microseconds mean life against 1 microsecond), not "a factor of 100 or so". Same file, same section: "uranium II half-life 245,000 years" is the modern figure; the deep read's 354,000 mean life is the matching quantity for Soddy's mean-life periods.
18. primer.md and arbiter.md "123-147 cal/h": keep as 123-147 with the note that the upper end is 145 with the primer's Q-value set (28.7 MeV) and 147 with 29.2 MeV including beta Q-values; both are [calc, memory medium]. QE correction 7 is thereby closed without a choice between the two.
19. claims.md: add C8 (life and the second law, Le Sage, forces), C6e and C9 (actinium); apply corrections 1-5 above; claim 2 needs the 13% and the 160-ton lower bound; claim 3 needs the helium inconsistency and the p.237 and U II misprints; claim 5 needs protactinium and Fleck / van den Broek.
20. facts.json: Interpretation entry: add the helium experiment (PDF pp.142-145: seven flasks, four years, 2 mg per thousand tons per year), the "thirteen per cent" and "160 tons" (pp.218-220), the actinium branch (pp.247-250), the protactinium co-discovery (p.252), the Geiger-Nuttall rule (p.210), pleochroic halos (pp.211-215), the p.237 misprint, and the 1909 edition as a second source for the unchanged passages. Matter and Energy entry: add Chapter I (energy and wealth; coal as capital), the second law and life (pp.101-109), Le Sage (pp.116-118), Chapter X (the self-bombardment negative test, the solar shrinkage figure, Moissan, the "10,000 horses" and "slave" argument, the 1/10^10 per year release). how_read for Chapter X: now read in full (pp.236-257).
21. Stale-copy search: also grep site/ and work/rankings/entries/ for "thirteen per cent", "13 per cent", "uranium life" with "helium", "mainspring", "Le Sage", "second law", and "Plausible, untested" in a headline for Soddy (C6d now leads with Established).
22. Deep-read file Interp. section 5 (helium) says it has not checked which review files present the helium measurement as independent confirmation: I checked the arbiter (it did not) and the primer is for the primer author. The orchestrator should grep primer.md and physicist.md for "helium" with "8,000" or "same result".