Rutherford-Ernest: Science Contrarian review (Step 2)

Reviewer: Science Contrarian, 2026-10-09. Standard: judge by experiment, not consensus or reputation; build the strongest honest case; apply the same rigour to the inventor as to his critics. Mallove is a model for standards, not a voice.

Tags: [read] = in the project documents or the Investigator's saved files (I read the passages quoted; deep reads checked against them); [web] = URL given; [calc] = my own arithmetic; [memory] = recalled, with confidence.

0. What kind of project this is, and what I read

Rutherford is not an outsider. His central claims were mostly accepted within a decade, so the contrarian question here is different from the usual one. It is (i) which of his claims were tested and how, (ii) which of his claims did not survive and why, and (iii) which anomalies near his work (Curie-Danne temperature effect, Dewar's liquid-hydrogen heat, Bumstead's X-ray heating, Le Bon's universal radioactivity, the gravity test) were treated fairly. The author is an interested party in every claim here (the disintegration theory, the nuclear atom); the 1930 book's Ch. VIII reports Chadwick's own experiments with Chadwick as co-author [read: deep/Radiations...md section 3.3].

Read myself (not only the deep reads): the facts.json list and claims.md; all seven deep-read files (the 1911 paper file in full; the 1904, 1906, 1913 and 1930 files by section, with the energy, scattering, detonator, Bumstead, Sun and Earth sections in full); the Investigator's sites.md and research notes; and the primary passages in the saved files: Bakerian Lecture 1904 pp.198-199 and 210 (page images OCR'd by me), Bumstead 1906 (opening, secondary-ray section, Conclusions), Rutherford 1919 Science text, Wikisource St Louis address, Bronson 1905 abstract as recorded in the notes. Not read (and so not relied on): Collected Papers (lending only), the body of Rutherford-Compton 1919 (paywalled), Compton 1920, Curie-Danne CR 138, Badash's letters, the Goodwin thesis, and 1930 book Ch. IX-XI and 1913 Ch. I-X (digest only). The 1913 deep read of the alpha-ray chapter (pp.131-192) is still missing.

Web checks made (6, within budget): Harper's dates; Bumstead; Emery 1972 and chemical effects on decay; Rutherford-Compton gravity; Jenkins-Fischbach dispute; Blackett 1925. URLs are in the text.

Corrections to Step 1

1. claims.md, claim 4, "equals the atomic number (79) by Chadwick's 1920 measurements": Chadwick 1920 measured copper, silver and platinum, not gold: Cu 29.3e, Ag 46.3e, Pt 77.4e for Z = 29, 47, 78, "estimated accuracy between 1 and 2 per cent" [read: 1930 book, deep/Radiations...md 3.3]. Gold within 2% of 79 is Rutherford and Chadwick 1925 [read: same file, 4.2]. Also: the 1911 paper does not say "nucleus" (the word appears only in the web page's running header) and does not establish a positive sign; Rutherford assumed positive "for convenience" and wrote "it has not so far been found possible to obtain definite evidence" [read: deep/structureatom 9]. The claim as worded should say so.
2. claims.md, claim 5, "about hundredfold": the factor is explicit only in the 1905 Harper's article ("about one hundred times the estimate founded on the condensation theory", about 5000 million years) [read: deep/RadiumTheCause... line 70]. The 1913 book says only "much longer" [read: deep/Radioactive-Substances... section 3.6]. The deep reader notes that 5000 Myr is 100 x Helmholtz's 40 Myr, a scaling and not an independent calculation. Also claim 5 says radioactive heat "can account for" the Earth's heat: the 1913 book itself shows that crust-level radium spread through the whole Earth gives about 200 times the observed conduction loss and so needs a thin shell or depth decrease [read: deep/Radioactive-Substances... 3.5]. The Harper's 1905 figure (5e-14 g/g "corresponds" to soil) glosses over crust versus whole Earth [read: deep/RadiumTheCause... 6.1].
3. claims.md, claim 2, "about three-quarters from the emanation": consistent with Harper's 1905 and the Bakerian Lecture (75 per cent) [read: Bakerian p.197, OCR by me], but the 1904 book prints "more than two-thirds" and the 42 cal/h it quotes is per gram of radium bromide, not per gram of radium as printed (72 cal/h per gram of radium) [read: deep/radio-activity 5]. The store "3.7 x 10^9 cal per gram" is from the 1913 book, not the 1902-1906 works. Add the 1904 energy-budget shortfall: Rutherford's own alpha estimate was 50 cal/h against 100 observed, and he said "the energy of the expelled alpha particles does not account for the whole" [read: deep/radio-activity 6, p.180]. This is not in claims.md and bears on claim 2 (see there).
4. claims.md, claim 7, "no known physical or chemical agent changes" the rate: the 1904 Bakerian Lecture says the opposite for radium C. Rutherford, after Curie and Danne, wrote "I have repeated the experiments of Curie and Danne and obtained very similar results", and "This is the first case where temperature has been shown to exert an appreciable influence on the rate of change of any radioactive product" (constant about 1.4 times normal at 1100 C) [read: Bakerian p.198, p.210, OCR of page images]. Bronson (1905, in Rutherford's lab) and Bumstead (1906 footnote) then attributed it to B and C volatilizing at different temperatures [read: sites.md leads]. The Investigator's correction is right and I add the detail that Rutherford himself had replicated the anomaly before it was explained away. Also: claim 7 mixes two claims (ordinary matter transforms slowly; a trigger might exist). They should be judged separately (below).
5. claims.md, claim 6, "nitrogen ... to protons": in the 1919 Science paper Rutherford wrote that the long-range particles are "probably charged atoms of hydrogen or atoms of mass 2", "the evidence so far obtained is not sufficient to distinguish definitely whether these are H atoms or atoms of mass 2, 3, or 4" and that the residual nucleus was not seen [read: research/pages/archive-jstor-1642276.txt]. Identification of the product as oxygen-17 came from Blackett's cloud-chamber photographs (1925) [web: https://www.nndc.bnl.gov/discovery/abstracts/8/17.pdf, via search; see also https://en.wikipedia.org/wiki/Patrick_Blackett].
6. facts.json, Harper's article dates: "March 1905 [memory, medium]" should be February 1905; Harper's archive lists the article under 1905/02 and Rutherford's bibliography gives "Harper's Mag Feb 390-396 1905" [web: https://harpers.org/archive/1905/02/radium-the-cause-of-the-earths-heat ; https://rutherford.org.nz/bkflerov.htm]. The other article is archived under 1904/01, confirming January 1904 [web: https://harpers.org/archive/1904/01/disintegration-of-the-radioactive-elements].
7. Errors in Rutherford's own texts that Step 1 recorded but that I list here so they travel with the verdicts: 1911 b = 3.4e-12 cm is 6% above his own constants (3.19e-12); p1 = 0.0004 should be about 0.0007; three of the eight tabulated z/A^1.5 quotients (tin, copper, iron) do not follow from the printed numbers; the Crowther silver row looks copied from platinum (probable transcription error) [read: deep/structureatom 4, 6, 7]. 1930 book: the cosec^4 table shows a 35-50% rise of N/cosec^4 at 30-15 degrees in the first set (silver 18-21 up to 29-31; gold 28-31 up to 35-40) that the book does not discuss; Schonland's table has a factor-10 inconsistency; the footnote year 1921 should be 1911 [read: deep/Radiations... 2.1, 7]. Harper's 1904: "one per second in every million billion" (lambda about 1e-15 s^-1) and "a million years for a thousandth part" are not mutually consistent (the second implies about 3e-17 s^-1), and both are far from the true thorium rate of about 1.6e-18 s^-1 [read: deep/Disintegration... 8].

Per-claim review (questions a-d, then verdict)

Claim 1. Atoms spontaneously disintegrate, at a rate no known agent changes, following a probability (exponential) law

a. Mathematics. Internally consistent. The decay law I = I0 exp(-lambda t), the recovery law 1 - exp(-lambda t) with "the same constant", and the successive-transformation algebra of 1913 Ch. XI are correct as far as the deep reads checked them (with the OCR digit errors 3300 cal, 1/132 h, 3.77 d, 35 min corrected from page images) [read: deep/Radioactive-Transformations 1.5, 8; deep/Radioactive-Substances A3-A7]. One internal inconsistency in the popular article (rate numbers, above) which Rutherford himself marked "can only be calculated roughly".

b. Experiments. Thorium X separated by ammonia and regrowth of the thorium activity in about a month, with the same 4-day constant for decay and recovery; uranium X (Crookes, Becquerel) the same; emanation half-life 3.77-3.99 days across four laboratories [read: deep/radio-activity 8b; deep/Radioactive-Transformations 1.1]. Temperature independence: emanation constant between 450 C and -180 C [read: Wikisource St Louis address]; Becquerel and Curie at liquid-air temperature [read: deep/radio-activity 7]. The strong form ("no agent changes any rate") is not exactly true by modern measurement: electron-capture rates depend on the electron environment at the 0.2 per cent to about 1 per cent level (Be-7: environment differences of order 0.2%; Be-7 in C60 cages 0.8% lower half-life) [web: https://www.obscure.org/physics-faq/ParticleAndNuclear/decay_rates.html and search results on Emery, Annu. Rev. Nucl. Sci. 22, 165 (1972), https://marvin.byu.edu/beryllium7/references/DecayRate/AnnuRevNuclSci-22-165.pdf]. For the alpha and beta emitters Rutherford studied, no accepted change has been shown. Fully stripped ions can have very different rates (Bosch et al. 1996) [memory, medium; not web-checked]. Rutherford's own words were "any known agency" and "not been found possible", which is a statement about evidence, not a law.

c. Tested? Yes, repeatedly, by Rutherford's group and by Curie, Becquerel, Dewar, Gates, Bronson, Bumstead. The one positive result, the Curie-Danne 1904 temperature effect on radium C (decay constant 1.4 times normal after about 1100 C), was replicated by Rutherford himself in 1904 and then explained in 1905 by Bronson, who found "very little, if any, effect" between 700 and 1100 C and showed that B (more volatile) and C separate when heated [read: Bakerian p.198; sites.md lead 2, Bronson]. This is "tested and the anomaly explained", not "dismissed without testing". What I could not find: whether Curie and Danne accepted the volatility explanation; the files do not hold their reply. Modern small effects: the Jenkins-Fischbach claim that beta-decay rates vary with the Earth-Sun distance (solar neutrinos) is a descendant of the question. It is contested: Kossert and Nahle (PTB) and Norman report no oscillation or instrumental artefacts; Cooper tested alpha-powered RTGs, which Jenkins's group says are outside their claim; Sturrock and colleagues dispute the null results and have a stake in the claim [web: search summary of https://arxiv.org/abs/1208.4357, https://arxiv.org/abs/1210.3334, https://arxiv.org/pdf/1007.3318, https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2020.584101/pdf]. I did not read these papers; that line is open, not established either way.

d. If false/true: the law itself would need nothing from mainstream; the "no agent" form already has footnotes in mainstream physics (electron capture, bound-state beta decay).

Strongest honest case for the sceptic of "rates are constant": Rutherford's own 1904 result, a replicated laboratory measurement of a temperature effect, shows how quickly a decay-rate anomaly can be accepted and then dissolved by a mixed-source artefact. That argues for caution about both reading small decay-rate variations as physics and dismissing them as artefacts without a control that separates the species.

Verdict: split. Spontaneous transformation with an exponential law and the Th X, U X, emanation chains: Established. "No known agent changes the rate", for the alpha and beta emitters in the book: Established to the precision of the day and since. Strong form ("no agent at all"): Tested, not supported at the sub-percent level (electron capture). The 1904 Curie-Danne temperature effect on radium C: Tested, not supported (explained by volatility, one laboratory's replication plus one contrary study). Modern claims of cosmic modulation of decay rates: Never properly tested by a neutral, agreed protocol (contested, partly tested).

Claim 2. The energy (about 100 cal/h per gram of radium; about three-quarters from the emanation and products; store about 3.7e9 cal/g) is the atom's own internal energy, vastly larger than chemical, not taken from the medium

a. Mathematics. Consistent. Checked: 14 cal/h per g of the 1/6-radium barium preparation x 6 gives about 100 [read: deep/radio-activity 4]; 876,000 cal per year is about 225 times the 3900 cal needed to dissociate 1 g of water, which is more striking than his "nearly as much per day" (62%) [read: same]; 3300 cal for 1 mm3 emanation, 9900 with products [read: deep/Radioactive-Transformations 1.5]. Slips: the 42 cal/h per gram printed for radium is per gram of the bromide; the 30 cal/day for gas from radium looks 40% high; thorium and radium beta fractions 1/22 and 1/14 are lower than his own table gives [read: deep/radio-activity 2a, 5, 7].

b. Experiments. Heat emission measured by Curie and Laborde (thermocouple and Bunsen ice calorimeter), by Runge and Precht, by Rutherford and Barnes (differential air calorimeter), 65 cal/h per g of radium bromide, 110 per g of radium [read: deep/radio-activity 4, 5]. Heat tracks the alpha activity (falls to about 25-30% when the emanation is removed and recovers in a month), not beta or gamma [read: same, 5]. External-source tests: heat inside an ice calorimeter; unchanged under several inches of lead (Runge and Precht: no difference to about 5% with a lead screen) [read: Wikisource St Louis address; deep/radio-activity 6]. The energy budget: in 1904 Rutherford's alpha count gave 50 cal/h against 100 observed and he proposed extra "electrical disturbances in the atom" to make up the other half [read: deep/radio-activity 6, p.180]. That could have been read as a missing-energy anomaly. With modern alpha energies (4.87, 5.59, 6.11, 7.83, 5.41 MeV for the chain to lead) the alphas alone give 3.0e9 cal per gram over the life [calc], about 123 cal/h in equilibrium [calc, from deep/radio-activity 6 using [memory, high] energies], i.e. the whole observed heat. His shortfall came from taking the alpha charge as e instead of 2e and low alpha energy (5.9e-6 erg, about 3.7 MeV). His count of alphas (method 3: 1.4e11 per second per gram) matches the modern 1.46e11 [read: deep/radio-activity 3, 6]. Store: radium to lead, mass defect (226.0254 - 205.9745 - 5 x 4.0026 = 0.0379 u = 35.3 MeV per atom) gives 3.6e9 cal/g [calc, atomic masses [memory, high]], against his 3.7e9 [read: deep/Radioactive-Substances 3.6]. The agreement is close (3%) and independent of any 1913 theory.

c. Tested? The internal-versus-external question was tested by the ice-calorimeter and lead-shield experiments above and by the sealed-vessel invariance. Those exclude a source that depends on the ambient air or on ordinary rays stopped by lead; they cannot exclude every conceivable medium that passes through lead without altering the radium's activity. The reply is that an external supply would also have to give the new elements, the helium and the exponential law, and none of the "borrowed energy" proposals did. Dewar reported (British Association 1903) that the heat of a radium tube was "distinctly greater in liquid hydrogen"; Rutherford: "requires confirmation" [read: deep/radio-activity 4]. My check of the printed numbers (73 c.c./min of gas from 0.7 g radium bromide in liquid hydrogen) gives about 42 cal/h, or 60 cal/h per g of bromide, which matches Rutherford and Barnes's 65 for a preparation not yet at equilibrium [calc, [memory, high] latent heat]. No excess is needed; I found no later record of the Dewar claim being confirmed or withdrawn. The Investigator's St Louis (1904) address shows Rutherford stated the "transformer of borrowed energy" view fairly and argued against it from experiments [web: https://en.wikisource.org/wiki/Popular_Science_Monthly/Volume_67/May_1905/Present_Problems_in_Radioactivity].

d. If true: nothing changes; it is the mainstream. Mass-energy equivalence (1905, not used by Rutherford in these texts) later supplied the bookkeeping.

On Le Bon (interested party, priority claims for intra-atomic energy since 1897): he quotes Rutherford as "in doubt" whether the energy is borrowed from the atom or from external sources (Archives, Geneva, 1905, "p.53") [web: https://www.rexresearch.com/lebonmat/lebonmat.htm]. The French original (p.55 of the scan) has the sentence, and the same address continues with the argument against the external source; the English text words it "whether the view is taken that the energy is derived from the atom itself or from external sources" [read: sites.md lead 1; research/2026-10-09-archive-org.md item 5]. So Le Bon's quotation is real and shorn of its surrounding argument. This matters for the Moray follow-ups, which copied Le Bon.

Verdict: Established (internal atomic energy, magnitude about 100 cal/h per gram, store about 3.6e9 cal/g). The 1904 factor-of-two shortfall: Refuted by calculation as a real anomaly (it is his input numbers). Dewar's liquid-hydrogen excess: Never properly tested (reported once; the printed numbers show no excess; no confirmation found).

Claim 3. Alpha particles are helium atoms at about a tenth of c; beta fast electrons; gamma penetrating radiation; helium and lead final products; uranium to ionium to radium (Ra/U = 3.4e-7)

a. Mathematics. Consistent. Energy ratio alpha/beta 83 = (v1^2/(e/m1))/(v2^2/(e/m2)) [read: deep/radio-activity 2c]. 3.4e-7: (1600 y / 4.47e9 y) x 226/238 = 3.40e-7 [calc, half-lives [memory, high]], which matches Rutherford's value in 1913. Boltwood's table: I/III reproduces IV at 215-228 over a 250-fold range of uranium content [read: deep/Radioactive-Substances A10].

b. Experiments. Ra/U constancy in minerals (Boltwood); activity balance 4.64 of uranium (McCoy, Boltwood); Geiger-Rutherford counting 2.37e4 alphas/s/g of uranium; ionium (Boltwood 1906); lead as end product (Boltwood 1907; age by Pb/U) [read: deep/Radioactive-Substances 2.3, A9]. The 1913 book records a dissenting datum: Russell and Rossi's arc spectrum of the thorium-ionium preparation showed only thorium lines where about 1 per cent ionium would have been visible if its life were 100,000 years [read: deep/Radioactive-Substances A9 item 8]. That is a tested, anomalous result that Rutherford reported honestly and whose resolution (ionium is Th-230, inseparable chemically from thorium, half-life 75,000 y [memory, high]) is not in the files. Step 1 should keep it in the record. Errors in 1903-1904: alpha mass "twice hydrogen" (charge taken as e), speed 20,000 miles/s (about 1.5-2 times too high), radium's life "not much more than a thousand years" (mean life about 2300 y), thorium "a million years" for a thousandth (about 1e7) [read: deep/Disintegration... 2, 8, 12]. All are acknowledged by later work, not hidden.

c. Tested? Yes, by many hands. Rutherford and Royds 1909 showed alpha = helium (spectroscopic) [memory, high; in neither the folder nor my web checks; the 1913 alpha chapter deep read is missing, so the project holds no primary text for this].

d. If false: nothing remains of radiochemistry or radiometric dating.

Verdict: Established.

Claim 4. Nuclear atom: small, massive, highly charged centre; cosec^4(phi/2) law; central charge of gold about 100e in 1911, atomic number by Chadwick; Coulomb law to about 3e-12 cm

a. Mathematics. Internally consistent. All derivation steps in the 1911 paper check by SymPy (b = 2p cot theta, equations 1-5, recoil quadratic, 3 pi b / 8R, constants 5.54 and 15.4, -0.72, 0.96 pi) [read: deep/structureatom 10]. Faults in the numbers listed in Corrections 7. One physics flaw he flagged himself: the Crowther comparison; he admitted that Crowther's aluminium beta data favoured compound scattering and answered with Marsden's partial data and "further experiments ... are desirable" [read: deep/structureatom 7(d)].

b. Experiments. Geiger and Marsden (1909): about 1 in 8000 (platinum), 1 in 20,000 deflected through ~90 degrees; Geiger 1910 small-angle; 1930 book: cosec^4 over a factor of 250,000 in rate, thickness proportionality for four metals, scattering per atom proportional to A^2 (Ze = A e/2 within about 20%), NV^4 constant to +/-12% over a ninefold range; Chadwick's nuclear charges within 1.5% of Z; the Coulomb law down to 3.2e-12 cm (gold), 2e-12 (silver), 1.2e-12 (copper), no departure from copper to uranium [read: deep/Radiations... 2, 3, 4.4-4.5]. Order-of-magnitude check of the 1-in-20,000 figure: eq. 3 gives 1 in 53,000, within a factor 2.5 that Rutherford did not claim [calc, deep/structureatom 4]. With modern Z = 79, b for 2.09e9 cm/s alphas is 2.5e-12 cm, not 3.4e-12 [calc].

c. Tested? Yes: Geiger and Marsden 1913 and Chadwick 1920, Rutherford and Chadwick 1925, Rose 1926 (1.2 to 8 degrees, plural scattering below about 2.5 degrees), cloud-chamber work in the 1930 book [read: deep/Radiations... 4]. The authors of the confirming experiments are Rutherford's own group; the independent confirmation (Moseley 1913-14, Bohr's atom) is by other routes [memory, high]. Residual points to keep: the 35-50% excess of N/cosec^4 at 15-30 degrees in the first (angle) data set is not commented on in the book (finite-beam width and plural scattering are likely, my reading); the 1930 book's Chadwick-Mercier gold implies an effective charge about 3 below 79 (not "about two") [read: deep/Radiations... 6.5, 7].

d. If false: nothing in atomic physics survives. As a point of history, the "nuclear" model won because it made a specific, risky prediction (a fourth-power law in angle and velocity) and not because it was elegant; the competing J. J. Thomson plum-pudding model made a different one and failed.

Verdict: split. Nuclear atom with a Coulomb centre down to about 3e-12 cm, and the cosec^4 scattering law: Established. Central charge of gold about 100e (N proportional to A/2), nucleus as 49 helium units (Rutherford's "may be only a coincidence"), 1913's "59 helium atoms" for uranium: Historically superseded but instructive (Z = 79, 92; Moseley). Sign of the charge in the 1911 paper: assumed, not established (Not testable as stated in that paper; later fixed by other experiments).

Claim 5. Radioactive heat can account for the Earth's heat, so Kelvin's cooling limit does not apply; speculatively a release of atomic energy in ordinary matter in the Sun could lengthen its life about a hundredfold

a. Mathematics. The Earth arithmetic reproduces: 270 million tons of radium for the conductive loss (rock conductivity about 1.7 W/m/K assumed; 1.0 or 2.5 would give 157 or 394 Mt), 4.6e-14 g/g for a uniform Earth, 1913's 39 cal/cm2/yr outflow and 20 km shell, shell temperature 300 C [read: deep/RadiumTheCause... 4; deep/Radioactive-Substances 3.5, [calc] by the deep reader]. Crust radium 2e-12 g/g is 20-40 times the uniform threshold, so a uniform Earth with crust composition gives about 200 times the observed flow [read: same]. The 1913 book therefore concludes radioactive matter must decrease with depth (Strutt's thin-shell). The Sun: 2.5 ppm radium would power the Sun (calculated by deep reader 1.66e-6 for exact match) but would be spent in a few thousand years [calc]. The 1905 "5000 million years" is 100 x Helmholtz's 40 Myr, which is a scaling, not a derivation [read: deep/RadiumTheCause... 71]. Darwin's gravitational reservoir (3/5)GM^2/a = 2.7e7 cal/g reproduces [calc], and 3.7e9/2.7e7 = 137, so "about a hundredfold" is what the 1913 numbers imply if radium-like energy is released from all matter.

b. Experiments. Heat per gram of radium (above); crust radium and thorium (Strutt, Joly, Elster and Geitel); Pb/U and He/U ages above Kelvin's limit (Boltwood, Joly) [read: deep/Radioactive-Substances 3.5, 2.3]. The 1913 text says radioactive heat "does not supply any definite evidence as to the age of the earth". Modern: the Earth's radiogenic share of its 47 TW heat flow is large but not the whole, with a primordial component [memory, medium; not web-checked]. Kelvin's limit failed for two reasons (radioactive heating, and mantle convection which England 1895 and Perry raised, a point the 1905 article does not raise) [memory, medium].

c. Tested? The Earth's radioactive budget was measured in the field (soil, rock, sea); the Sun was never tested by Rutherford. The "ordinary matter in the Sun transforms like radium" idea was a conjecture; the right idea (nuclear energy powers stars) was confirmed in a different form: hydrogen fusion releases about 0.7% of mc^2, 1.5e11 cal per gram [calc], about 40 times radium's total per gram, which is the true "similar amount in ordinary elements" at the high end. The mechanism (slow transformation analogous to radium) was not the one that operates [memory, high, Bethe 1939]. Rutherford's 1933 remark that expectations of power from splitting atoms were "moonshine" [web: https://paperspast.natlib.govt.nz/newspapers/LDC19330913.2.68, via Investigator] is relevant to the same energy store: in 1913 he said the store is there; in 1933 he doubted it could be harnessed by the then-known methods. Szilard's chain-reaction idea came within days of that statement and fission in 1938 [memory, high].

d. If true: Kelvin's limit is void, not a revision of physics but of Kelvin's premise (no internal heat source).

Verdict: split. Kelvin's cooling limit does not bind (radioactive heating supplies a large part of the flux, with convection as a second reason): Established. "Radioactive heat accounts for the whole" with a uniform Earth: Refuted by calculation (about 200 times too much; the 1913 book itself shows it). Radium as the Sun's power: Refuted by calculation (spent in thousands of years). A hundredfold lengthening by radium-like release in ordinary matter: Historically superseded but instructive (fusion supplied the true source).

Claim 6. Artificial disintegration of the nucleus by alpha bombardment (nitrogen to protons, 1919), with departures from Coulomb's law at close approach to light nuclei

a. Mathematics. The 1919 paper is a plain account with no derivation. Nitrogen is 3 helium (mass 4) + 2 hydrogen or 1 of mass 2 in his guess [read: 1919 Science]. The recoil kinematics that apply to light targets are in 1911 §4 (alpha on aluminium, rho = 0.86, range 7 cm to 4.5 cm) and were checked [read: deep/structureatom 5].

b. Experiments. 1919: scintillations well beyond the range of the recoil N and O atoms (stopped by 9 cm of air) in dried chemically prepared nitrogen, "equivalent to the number of H atoms produced by the mixture of hydrogen at 6 cm. pressure with oxygen" [read: 1919 Science text]. Blackett 1925: about 23,000 cloud-chamber photographs, 8 forks, the alpha is captured, the product has mass 17 and charge 8 [web: https://www.nndc.bnl.gov/discovery/abstracts/8/17.pdf ; https://en.wikipedia.org/wiki/Patrick_Blackett, as reported by the search summary]. One advocacy publication (New Energy Times, which advocates low-energy nuclear reaction claims and so has a stake in challenging authority) argues the Rutherford credit is a "myth" and that Blackett did the work that identified the process [web: https://newenergytimes.com/v2/sr/Rutherford-Blackett/Rutherford-Blackett.shtml; I did not read it, only the search summary]. The compatible standard reading is that Rutherford discovered the reaction and Blackett identified it. The 1930 book (Ch. IX-X, digest only) gives the departures from the Coulomb law for light elements; I did not read those chapters.

c. Tested? Yes, replicated by other laboratories and by the cloud-chamber method of a different observer (Blackett). Rutherford announced no independent identification in 1919 and said so.

d. If true: nothing in mainstream physics changes; this is mainstream.

Verdict: Established (artificial transmutation of nitrogen by alpha particles; product identified 1925). Departure from Coulomb at close approach with light nuclei: Established, with the caveat that my reading of that part is by digest only.

Claim 7. Ordinary matter may be slowly transforming; disintegration might be triggered by an unknown external force (a "detonator") while the energy stays the atom's own; by 1930 no change in decay rate had been achieved

a. Mathematics. Where there is arithmetic it is consistent: Bumstead's argument that if X-rays "explode" atoms "as dynamite", heat exceeds ray energy [read: Bumstead 1906]. The 1906 hypothesis for "undetectable" ordinary-matter radioactivity (alphas below about 8e8 cm/s would not ionize) is a coherent kinematic idea [read: deep/Radioactive-Transformations section 6]. It is superseded by the Geiger-Nuttall law (1911) and Gamow tunnelling (1928): slow alphas imply astronomically long lives [memory, high].

b. Experiments. (1) Ordinary matter: Campbell and Wood 1906-07 found weak beta activity in potassium and rubidium (real: K-40, Rb-87) [read: deep/Radioactive-Substances 2.1-2.2 and 1906 deep read; memory, high]. Rutherford himself allowed the residual might be "slight radioactive impurity" [read: deep/Disintegration... 11]. Other "ordinary matter" radioactivity was partly impurity and cosmic rays [memory, medium]. (2) Triggers: the detonator passage appears in the St Louis address 1904, Radio-activity 2nd ed. 1905 sec. 270 (adding gravity as the one uncontrolled force), and the 1906 book [read: sites.md leads]; he states in all that "the energy ... is not derived from the detonator, but from the substance on which it acts". (3) Bumstead's measurement: lead strips heated "approximately twice as much" as zinc when X-rays were equally absorbed; the only hypothesis he found "not in more or less direct conflict with experimental facts" was that X-rays break up atoms of certain elements [read: Bumstead 1906, Conclusions]. He examined secondary rays at length and argued that the escaping ones could not account for it [read: same, pp.19-24]. (4) Gravity: Schuster proposed testing whether gravity alters decay; Rutherford and Compton developed a method for emanation over about 100 days to compare samples at different places on the Earth; "The outbreak of the war interfered with this plan" [web: https://www.nature.com/articles/104412c0, abstract]. Result in the Nature letter body and in Compton 1920 not read by anyone in this project. (5) Bakerian Lecture 1904 temperature effect (Claim 1).

c. Tested? (1) Tested; the general form failed, the K and Rb cases succeeded. (2) The "detonator" form is exactly what was later found: neutron-induced fission (1938-39) and nuclear reactions by projectiles are external triggers of a transformation whose energy comes from the nucleus [memory, high]. Rutherford's formal statement was prescient; his practical forecast in 1933 was not. In this sense "no agent can change the rate" is wrong for induced reactions, but they are new events (a collision), not a change in the spontaneous rate. (3) Bumstead's X-ray result: I found no replication and no refutation in the files and none in my web search (the search did not link Barkla's characteristic X-rays to Bumstead's heating data, and I will not assert such a link) [web: search on Bumstead; Wikipedia page https://en.wikipedia.org/wiki/Henry_A._Bumstead]. The standard explanation would be secondary electron and fluorescence differences between lead and zinc [memory, low]. I note, as a reviewer, that heating equals absorbed energy minus escaping secondary radiation, so lead (which fluoresces) would be expected to heat less, not twice as much, unless the beam energy and geometry matter; I cannot resolve that from the sources, and flag it as open. It was therefore "not tested again" in the record I have, and quietly dropped: Rutherford's 1906 book calls it suggestive and "far-reaching" while saying more metals and sources are needed; the 1913 book states "No evidence has so far been obtained that any substance can be made radio-active by bombardment" [read: deep/Radioactive-Substances 1, item 7]. (4) Gravity: Never properly tested in the files; the experiment was prepared by Rutherford's laboratory and interrupted by war. (5) See Claim 1.

d. If any of the strong forms were true (a field that alters spontaneous decay, or ordinary matter at uranium-like rates), nuclear decay theory (tunnelling, weak interaction) would need a new external coupling. Modern tests of the electron-capture environment effect (0.2-1%) show the coupling is small and known.

Verdict: split. Ordinary matter transforming at radium-like rates: Tested, not supported (isolated real cases K-40, Rb-87). External trigger changing the spontaneous rate (temperature, pressure, fields, radiation, gravity): Tested, not supported at the precision of 1930, with the sub-percent electron-capture exception (modern); gravity: Never properly tested in the files (planned, interrupted). Detonator hypothesis (external agent can precipitate a transformation without supplying the energy): Established in a different form (induced reactions, 1919 onward and 1938). Bumstead's X-ray heating anomaly: Never properly tested (one measurement, no replication or refutation found, quietly dropped).

How the anomalies near Rutherford were treated (the rejection record)

• Le Bon, quinine sulphate and universal radioactivity: tested in Rutherford's laboratory by Fanny Gates (Phys. Rev. 18, 135, 1904): ionization appeared only during temperature change, was absorbed in 2-3 mm of air and stopped by thin aluminium, so she concluded it differed from radioactivity; Le Bon, who has a priority stake, accepts that the cause is chemical but disputes her ultraviolet explanation, citing a 1906 note [web: http://cwp.library.ucla.edu/articles/gates/gates2.html ; https://www.rexresearch.com/lebonmat/lebonmat.htm]. The test was a real experiment, but by a pupil of the person whose theory was at stake. I have not seen the 1906 note.
• Gates 1903: heat does not destroy "excited radioactivity" (it volatilizes it); this answers Le Bon's "metal made radioactive loses it at white heat", which he lists as unexplained [web: http://cwp.library.ucla.edu/articles/gates/gates1.html].
• Curie-Danne: replicated, explained (Claim 1). Dewar: reported and not followed up (Claim 2). Bumstead: not followed up (Claim 7). Russell and Rossi's missing ionium lines: reported, not followed up in the project (Claim 3).
• Rutherford's own judgement: the 1933 "moonshine" cable is a single newspaper report of a lecture (Levin Daily Chronicle, a cable from London); I have not seen the lecture text. The 1921 Richardson statement shows others expected power [web: Papers Past links in research/2026-10-09-paperspast.md].

Investigator's findings kept in this review

Own work not in the folder: St Louis 1904 address (English and French), Radio-activity 2nd ed. (1905) sec. 270 with gravity, Bakerian Lecture 1904 (temperature effect), Rutherford-Compton 1919 (gravity test), Science 1919, Franklin Institute 1924 address (satellite alpha-emission picture, "four or five of the swiftest alpha particles" to break a helium nucleus), Die Radioaktivitat 1907, Collected Papers, "Some Cosmical Aspects of Radioactivity" (1907). Supporting: Gates 1903 and 1904, Bronson 1905, Bumstead 1906. Related work: Le Bon (interested party), Lucien Poincare 1907 (all bodies "a kind of explosive decomposing with extreme slowness"), Emery 1972, Goodwin 2012 thesis, Jenkins-Fischbach. Historical record: Cambridge MS Add.7653 (item-level not searchable), McGill Rutherford-Eve letters (Cohen 1988), Rutherford-Boltwood letters (Badash 1969), Todd 2014 on Rutherford's radium, Papers Past 1904/1921/1933, Oregon State collection. All are kept as leads: those that change a verdict are Bakerian p.198/p.210 (Claim 1), Bumstead (Claim 7), the St Louis context for Le Bon (Claim 2), Rutherford-Compton (Claim 7). The remainder I did not need to alter the verdicts and are not dismissed.

What would settle what is open

• Claim 1: Curie-Danne's published answer to Bronson, and a modern replication of heating of a radium C-bearing wire with separate B/C counting.
• Claim 2: none needed; (Dewar) a repeat of radium in liquid hydrogen with an independent calorimeter.
• Claim 7: the body of Rutherford and Compton 1919 and Compton's 1920 Phil. Mag. paper (gravity); a modern equal-absorption lead/zinc calorimeter test in monochromatic X-rays (Bumstead); the Bumstead correspondence if any.
• Claim 4: a scan of Phil. Mag. 21 p.681 and p.685 to settle whether the table inconsistencies are the journal's or the web transcriber's.
• 1913 alpha chapter pp.131-192: re-run the lost deep read before anyone relies on the project for alpha identity.