Mendeleev-D: Arbiter (Step 3)

Project: Mendeleev, "An Attempt Towards a Chemical Conception of the Ether" (written October 1902; English translation by Kamensky, Longmans 1904), plus Stewart (2019), a modern review of Mendeleev's predictions. Inputs weighed: facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, deep/, sites.md, research/ (including research/files/). Tags: [read] from a project document or a page image I rendered; [calc] my own computation (venv Python, reproduced in section 6); [web] with URL; [memory] with confidence. Page numbers: "p.N" in the reviews is the PDF page number of the 1904 copy (printed page = PDF page - 8), though a few cites mix the two. Example: the Kelvin density passage is on PDF p.10 (printed p.2), not "p.10 printed" [read: page image PDF p.10].

Source gap note

No source gap for the essay: the 1904 English text has a full text layer, and I re-read the periodic table (PDF p.34) and the Kelvin passage (PDF p.10) from page images. Stewart (2019) is a secondary source and is read in full text. Three things rest on documents the reviewers did not read in full: - The Russian original and the 1905 edition are known to the reviewers through the Investigator's notes (Wikisource and a scan). The Contrarian rendered p.25 of the 1905 scan; the others used footnotes as reported. - The 1906 8th edition of Osnovy khimii (Principles of Chemistry), the 1927 9th edition and Kedrov's 1958 commentary were never obtained. They are the deciding records for the "removed from the table" question (section 4). - Conclusions that rest partly on [memory] alone are marked below (Dewar's phosphorescence mechanism; neutron-star escape speeds; the CMB temperature and electron mass, all high confidence). No final verdict turns on a medium or low [memory] item alone.

Overall in one paragraph

Mendeleev, a leading chemist, proposed that the ether (the hypothetical medium that carries light) is a monatomic inert gas, about one millionth the atomic weight of hydrogen, placed in a new "zero series" of the periodic table. He called it "a series of thoughts" and "a hypothesis" and said its aim was to bring the question to the fore [read: PDF pp.39, 58-59; QE revision corrected p.47 to p.39]. The three reviewers agree on almost everything that can be checked: his arithmetic is right (one misprint), his zero group of noble gases was right, his sub-hydrogen elements have no support, and his own numbers cannot make a gas carry light. They differ mainly on which verdict label fits and on how much the experiments that sank the ether (Michelson-Morley and successors) bear on his particular gas. The record shows the essay was widely printed and not suppressed; no one ran a test aimed at his gas, because the essay proposes none.

1. Reconciling numbers where reviewers differ

Everything else the mathematician and Contrarian recomputed (escape speeds 11.2, 608 and 2,240 km/s; bounds 0.038, 1.3e-5, 9.6e-7, 5.3e-11; the 23.6 extrapolation; the misprint a = 0.000367 for 0.00367) matches the physicist and the essay to about 0.1 to 3 per cent. I did not redo these.

2. Corrected claim list (final verdicts)

Scale: Established; Plausible, untested; Never properly tested; Tested, not supported; Refuted by calculation; Internally inconsistent; Not testable as stated; Historically superseded but instructive.

Claim 1. The ether is a real chemical element: a monatomic gas, with weight, permeating all bodies and unable to combine

Agreement. All three: no mainstream support; ether-drift experiments (Michelson-Morley 1887; modern bound on anisotropy of c about 1e-17 [web: https://arxiv.org/pdf/1002.1284, Herrmann et al. 2009]) found no preferred medium frame; the essay never mentions those experiments; a gas cannot carry transverse light waves at about 130 times its own molecular speed; the 1902 question was reasonable. Disagreement. - Label for the strong form: Physicist "Refuted by calculation" plus "Tested, not supported"; Mathematician "Internally inconsistent"; Contrarian "Tested, not supported" and "Refuted by calculation" for the gas-as-light-carrier version. - Whether the ether-drift tests count as tests of Mendeleev's gas: Physicist yes; Contrarian says they tested the ether family, and no one tested his gas. - Weak form (some very light inert gas of unknown role): Contrarian "Never properly tested / Not testable as stated"; the others say nothing separate. - Kelvin's density floor: Physicist says it is refuted by the local gravitating-matter density; the others did not compute it. Which argument is stronger. The calculation is the strongest argument, and it uses only accepted kinetic theory, the same kinetic theory Mendeleev uses in his own equation I. With his numbers the sound speed is 1.63e6 m/s, c / 184, and a gas has no shear stiffness, so no transverse wave at all (linearised transverse motion is diffusion, not a wave; mathematician section 2.3 item 5). So "Refuted by calculation" fits better than "Internally inconsistent": the contradiction needs a law from outside the essay (the gas wave speed), even if the essay's own "130 times" remark is the admission that exposes it. The Contrarian is right that Michelson-Morley and successors tested the ether family, not a gas of atoms at 193 K; so I do not rest the verdict on them. I record them as "Tested, not supported" for the light-carrying medium with a preferred rest frame, which is what Mendeleev needs. On the weight sub-claim: Kelvin's floor of 1e-19 kg/m^3 is 15 times the total local dynamical matter density, and a gas that gravitates as ordinary matter would add to that density whether or not it is bound to the Galaxy. So if "weight" means ordinary gravitating mass at Kelvin's density, it is excluded. Caveats: Kelvin's own derivation was not located (the figure is as Mendeleev quotes it); Mendeleev himself notes that steam or hydrogen at tiny pressures would have as small a density [read: PDF p.10]; and the claim holds only for a gravitating ether. Confidence: medium. A search summary I saw gave the Kelvin figure ambiguously (as "0.1 x 10^-16 g"); the printed essay reads 1e-16 g by the digit grouping [read: page image PDF p.10]. Final verdict (split). - Strong form (a gas of atoms with an element's mass, with a thermal speed, carrying light): Refuted by calculation (wave speed 1/184 of c; waves would be longitudinal); and Tested, not supported (ether-drift tests of a light-carrying medium). - "The ether has weight, at least Kelvin's 1e-16 g per cubic metre": Refuted by calculation against local matter density, medium confidence, if the ether gravitates like ordinary matter. - Weak form (a very light weakly interacting inert gas with no stated role): Not testable as stated; Never properly tested. Nothing in the experiments excludes it, and nothing supports it. - As a 1902 question ("what is the ether made of, and does it have weight?"): Historically superseded but instructive. What would settle it. The wave-speed calculation already decides the strong form. For the weak form, an observed sub-eV, weakly interacting relic population would be a mass-range coincidence and would not vindicate a gas that carries light; a test would need Mendeleev to state a measurable consequence, which he did not. For the weight claim: Kelvin's original derivation, and a dynamical limit that allows for an unbound gas. Kept in the record: the Russian text says the ether may be a mixture of several limiting gases [read via Investigator: ru.wikisource]; the physicist notes this does not repair the wave problem (a mixture of gases is still a gas). The 1905 Principles of Chemistry appendix says giving the ether "a chemical dress" serves "the cause of unity in natural philosophy" [web: archive.org, via Investigator]. Yarkovsky (1901, ether density), cited by Mendeleev as an independent estimate, was not located, so nothing is said of its content.

Claim 2. A zero group of inert gases, and a zero series holding x, with y (possibly coronium) between x and hydrogen

Agreement. All three: the zero group (He, Ne, Ar, Kr, Xe, later Rn) is correct and now Group 18; Mendeleev credits Errera (5 March 1900) and Ramsay (19 March 1900) and admits his own earlier error in treating argon as N3 [read: PDF pp.21, 27-28, 33; 1905 scan footnotes 13-14]; the green coronal line is Fe XIV (Grotrian and Edlen, 1939-1942) [web: https://en.wikipedia.org/wiki/Coronium]; no element lighter than hydrogen exists; the extrapolations are curve fits. Disagreement. Label for the zero series: Physicist and Contrarian "Historically superseded but instructive"; Mathematician "Never properly tested; numbers Not testable as stated". Which argument is stronger. The mathematician's sensitivity table decides the numbers. Three points fix a quadratic exactly (23.61), so the fit has no residual and tests nothing. Other plausible extrapolations through the same points give x = 0.24, 0.095 or 0.030, a spread of 8 in x [calc, mathematician; I checked the 23.61 and the 0.17]. The y estimate (at most 0.4) rests on one data point (Li:H = 6.97) and an assumed jump. The 0.17 bound is never active in the final answer: x of about 1e-6 comes from the escape-speed argument alone, six orders lower (mathematician section 2.2). So the numerical placement is Not testable as stated. The existence question (a chemical element below Z = 1) is a different matter: atomic number is nuclear charge, and no species below Z = 1 exists, so the idea was superseded, not tested. The three reviewers' labels are compatible once the two questions are separated. Mendeleev himself distinguishes this from his 1871 successes (interpolation between known neighbours); here he extrapolates and "risks" something [read: PDF pp.24-26, 33]. Final verdict (split). - Zero group of noble gases: Established. - Zero series with ether x, as a place in the table: Historically superseded but instructive. - The numerical estimates of y (at most 0.4) and x (at most 0.17): Not testable as stated (the answer depends on the extrapolation chosen). - y = coronium: Tested, not supported (Fe XIV). What would settle it. Hold-back test of the extrapolation method on groups where the answer is known (predict an element left out of the data); if the method fails there, 23.6 is unreliable. For the table dispute see section 4. Corrections: Mendeleev's wavelength 531.7 nm is 530.3 nm in the modern value [web: https://en.wikipedia.org/wiki/Coronium]. The Stewart identification attributed to "Gruenwald" is not an iron identification: Stewart's footnote 12 is Gruenwald, Chemical News 56 (1887), on water-vapour spectra [read: _text Stewart, p.8 and note 12]; Fe XIV is Grotrian and Edlen. Stewart's own sentence is ambiguous. Note also coronium reports in volcanic gas (Nasini, Andreoli, Salvadori 1898), per the Russian text via the Investigator.

Claim 3. x has atomic weight about one millionth of hydrogen and moves at about 2,250 km/s (kinetic theory at -80 degC, escape velocities)

Agreement. All three: the arithmetic is right; the printed "a = 0.000367" is a misprint for 0.00367 (confirmed by the Russian footnote 24); the final x sits at the upper end of the range his own method allows (50-solar-mass star: 9.6e-7), the lower end (5.3e-11, v = c) was set aside because it "would revive the emission theory of light" [read: PDF pp.50-51]; the assumed single temperature is unmeasured; the Maxwell tail makes the mean-speed escape cut-off crude. Disagreement. Verdict label: Physicist "Refuted by calculation" (premises); Mathematician "Internally inconsistent" for the framework; Contrarian "Not testable as stated", with the accumulation argument "Internally inconsistent". Which argument is stronger. I take the Contrarian's internal finding together with the mathematician's. Using only the essay's own premises: (i) mean speed 2,250 km/s is 3.7 times the Sun's escape speed (608 km/s), so by his own criterion the Sun cannot hold the ether, yet he says the Sun and uranium accumulate it (PDF p.54); (ii) he says the ether exchanges with matter (PDF p.58) and permeates bodies, yet fixes its temperature at -80 degC irrespective of where it is; thermalisation at stellar temperature would take x = 1e-6 to above c in a star's core, the very limit he rejects (2.08 c [calc], with a modern temperature). Those are contradictions inside his framework. The physicist's strongest points (temperature at 2.7 K gives 70 times smaller x; white dwarfs push x to 1e-7 or lower) use measurements made after 1902 and are better described as supersession than as calculation from accepted law. They do show the number is not robust: it moves by orders of magnitude as soon as the assumptions change, which is why the label "Refuted by calculation" overstates it. The arithmetic itself is not at fault. Final verdict (split). - The arithmetic from the stated premises: the arithmetic is correct (finding, not a scale verdict: the scale's "Established" means supported by experiment and accepted theory). [QE revision: qualifier] - x about 1e-6 and v about 2,250 km/s as properties of a real ether: Internally inconsistent (accumulation against escape speed; fixed temperature against his own exchange of ether with matter), and in practice Not testable as stated (the number is a choice inside a range of five orders of magnitude). What would settle it. A calculation: apply his own escape criterion to Sirius B (x <= 1e-7), and thermalisation to the Sun. Neither needs an experiment. No experiment can fix x unless the ether is detected. Kept in the record: the Russian text doubts absolute zero and promises a separate article on it (footnote 23, via Investigator); I found no record of that article. Mendeleev's speculation (Russian text) that ether piling up around the Sun might supply its energy: the mathematician's energy budget gives 6e31 kg of ether (about 30 solar masses) to power the Sun for 1e9 years, and 7e-8 s of solar output from the ether in the Sun's volume at Kelvin's density [calc, mathematician; I checked 1.4e8 kg, 2.6e19 J and the 33 solar masses]. It is phrased as a question, so I treat it as a hypothesis, not a claim.

Claim 4. Radioactivity, the Curie zinc-sulphide demonstration and Dewar's low-temperature phosphorescence support the ether

Agreement. All three: the phenomena are real; the ether reading adds no measured quantity; the "something material" that flows is radium emanation, a radioactive inert gas (now radon, Z = 86) that decays, which explains the glow by alpha particles hitting zinc sulphide. Disagreement. Label: Physicist "Historically superseded but instructive" and "Never properly tested as an ether claim"; Mathematician "Not testable as stated"; Contrarian "Historically superseded but instructive". Web checks. Rutherford saw a gaseous emanation from thorium in 1899; Dorn observed a radium emanation in 1900 but his paper did not call it a gas or an element; Rutherford and Soddy concluded by early 1902 that it behaved as an inert gas; Crookes showed the spinthariscope on 15 May 1903 and the sparks are single alpha particles on zinc sulphide [web: https://iupac.org/publications/ci/2011/3301/8_radvanyi.html; https://en.wikipedia.org/wiki/Friedrich_Ernst_Dorn; https://lapada.org/art-and-antiques/a-cased-william-crookes-spinthariscope/; the Dorn and Rutherford details are from search summaries, not fetched]. So the mainstream reading was already forming when Mendeleev wrote (October 1902), and the essay does not mention emanation. The link between this exact Curie flask and radon is still [memory, medium]: Mendeleev does not name the emanation. Credit to Mendeleev (Contrarian, kept): something material does flow, and it is a zero-group gas as his table would predict, though a heavy one, not a light one. Which argument is stronger. All three agree on substance. The mathematician's point that no measurable quantity is offered is true and I keep it. The superseded label fits better for the reading, because the explanation was replaced within a year by the Rutherford-Soddy and Crookes accounts. Final verdict (split). - The phenomena (Curie flask glow, radioactive emanation, brighter low-temperature phosphorescence): Established. - Ether accumulation on U, Th and Ra as their cause: Historically superseded but instructive; as offered it also Not testable as stated (no solubility, flux or energy predicted), and Never properly tested as an ether claim. Dewar's mechanism (reduced thermal quenching) is [memory, medium] and unchecked; no verdict depends on it. What would settle it. Nothing further is needed beyond the existing radon work; the historical record to read is Mendeleev's own source for the Curie demonstration (his 1902 visits to the Curies and Becquerel).

Claim 5. No single primary substance; atoms are not divisible into electrons or transmutable; only the ether is a unity

Agreement. All three: atoms are divisible (electron 1897; nuclei; transmutation 1902-03, 1919); "atoms are individuals, so electrons are metachemical" is circular [read: PDF pp.13-15, 25]; Mendeleev saw radioactivity first-hand and did not accept transmutation; Lente's "highly personal" motive (opposition to Prout's hypothesis) is Lente's characterisation. Disagreement. Emmens and Fittica: Physicist [memory, medium]; Mathematician "not assessed"; Contrarian "Established" (also [memory, medium]). "No primary substance": Physicist "Plausible, untested", others "superseded". Web checks done by me. Fittica announced in 1900 that amorphous phosphorus heated with ammonium nitrate gave arsenic; Noelting and Feuerstein repeated it with arsenic-free phosphorus and got no arsenic, and commercial phosphorus carries arsenic as a contaminant [web: https://natuurtijdschriften.nl/pub/563096/ADN1901050001002.pdf, from a search summary, not opened]. Emmens (argentaurum, silver to gold, 1897) was reported as exposed as a fraud and "by 1901 nowhere to be traced" [web: https://www.historytoday.com/archive/alchemy-new-light-dark-art and https://huxley.media/en/disclosures-in-science-gold-from-silver-and-stock-market-games/, search summaries, not opened]. Emmens's story included a US Assay Office purchase; I have not seen the primary documents. So Mendeleev's dismissal of these two particular claims is supported by the later record, with the evidence for Emmens thinner than for Fittica. Mendeleev himself cites them as examples of "many such instances, but they have always proved unfounded" [read: PDF p.17 (_text line 456); QE revision corrected from p.9]. Which argument is stronger. The Physicist's modern point is useful: "no single primary substance" is partly right in modern terms (quarks and leptons, and nuclei built of two nucleon types, not from H atoms), but not in the sense he meant (atoms as unalterable individuals). Final verdict (split). - Dismissal of Fittica's phosphorus-to-arsenic claim: Tested, not supported (the claim he dismissed; the test is Noelting and Feuerstein's, from a secondary web summary). Emmens's silver-to-gold claim: dismissed as a fraud in the later record, source thin (secondary web summaries); I do not label it with a scale verdict beyond "not supported by what I found". - Atoms indivisible and not transmutable: Tested, not supported (superseded). - "No single primary substance": Historically superseded but instructive (a partial modern echo: several elementary particles). - Ether as the only unity: see Claim 1. What would settle it. Already settled for atoms. For Emmens: the US Assay Office records and the contemporary press of 1897-1901.

Claim 6 (secondary, Stewart 2019). Earlier predictions largely succeeded; ether and coronium were not predictions

Agreement. All three: gallium (1875), scandium (1879) and germanium (1886) matched in detail; Tc, Re, Po, Fr and Pa within a few per cent; the lanthanide-zone failures are real; interpolation worked and extrapolation below hydrogen did not. This is the strongest evidence in the project, and it supports the periodic law, not the ether. Disagreement. Counting: Stewart one multiple mistake against Scerri five failures; the Contrarian points out that "eleven" is only in Stewart's abstract. Final verdict (split). - Ga, Sc, Ge: Established. - Tc (100), Re (190 vs 186.2), Po (212 vs 209), Fr (220 vs 223), Pa (235 vs 231): Established, as Stewart counts them (his criterion: a prediction succeeds if it lands on a higher analogue of the element on which it is based and has appropriate properties [read: _text Stewart, p.4]). - The five lanthanide-zone predictions (140, 145, 155, 170, 175): Tested, not supported (one structural mistake per Stewart; five per Scerri). - Ether and coronium as "predictions": Not testable as stated, meaning "not a table-based prediction in Stewart's sense" (he calls them two irrelevancies). Their separate fate is under Claim 2. [QE revision: qualifier made explicit] Corrections: the count should read "ten named, eleven in the abstract"; Stewart's "six more blanks between hydrogen and helium" is not in the 1904 table (section 1). [QE revision: see items 1-6 of the new section for further Stewart corrections.]

3. Disagreements that carry weight

1. Strong form of Claim 1: Refuted by calculation, not "Internally inconsistent". Reason above: the decisive law comes from outside the essay.
2. Claim 3: Internally inconsistent, not "Refuted by calculation". The decisive points are in the essay (accumulation against escape speed; fixed temperature against ether-matter exchange). The modern temperature and white-dwarf points show fragility but use post-1902 data.
3. Claim 5, Emmens and Fittica: I did the check the reviewers left as [memory, medium]; Fittica is supported by a secondary web summary; Emmens is thinner.
4. The Contrarian's observation that ether-drift tests concern the family, not Mendeleev's gas, is accepted; the physicist's "tested, not supported" is retained only for the light-carrying medium.
5. The Contrarian correctly adds that there is no record of blocked funding or publication or of reanalysed negative data in this case; this differs from cold-fusion-type disputes. The reception evidence found is: widely printed in his lifetime, then left alone.

4. Absence-of-evidence analysis: where the record goes dark

A. What happened to x and y in the later tables

Where the record goes dark. The 1904 English table prints x (series 0) and y (series 1) in the zero group, with no atomic weights [read: page image PDF p.34]; the 1905 Russian scan prints rows 0 and 1 the same way, without names (Contrarian, page image p.25 of the 1905 scan [read]); the text says "under the signs x and y already stand expected elements with atomic weights less than hydrogen's" (Contrarian, 1905 scan p.22 [read]). Whether the 1906 8th edition of Osnovy khimii printed them, and when they vanished from the editions, is unknown: none of the 1906, 1927 or 1958 documents was obtained. Competing explanations. (a) Abandonment as the hypothesis lost ground. Mendeleev died in February 1907 [read: Stewart], before Moseley (1913) and Bohr (1913) gave a physical reason to start the table at hydrogen; relativity (1905) removed the need for a medium. Evidence for: the essay is labelled a hypothesis by its author; the zero group stayed; no measurement ever supported x; ru.wikipedia and Lente take this view (ru.wikipedia is also wrong that x and y are missing from the essay's own tables). Evidence against or missing: no document seen records anyone deciding this. (b) Deliberate removal or falsification, as claimed by Atsyukovsky ("later removed") and Rodionov (a "genuine 1906 table" with a zero row), both advocates of an ether whose statements are their own [web: via Investigator's research/2026-10-06-google.md]. Evidence for: the advocates' statements only; their texts were not obtained by any reviewer. Evidence against: the essay was printed in at least five forms in Mendeleev's lifetime (Russian 1903, German 1903 and 1904, English 1904 and 1905), which is hard to reconcile with suppression of the essay itself; no removal document found. (c) Mendeleev himself dropped or reworked the rows in 1905-1906 (a third option, not in the reviews). Evidence: the Russian 1905 text was revised from the 1903 text (Rd 224 became 225; footnotes added, per Contrarian and Investigator), so he did change the text; nothing found on the table in 1906. (d) Editorial removal in later (posthumous) editions, for science or other reasons: no evidence either way; the 9th edition (1927) and later Soviet editions would show it. What would decide it. The 8th edition (1906) table pages (RGB/NEB, the Investigator's rusneb.ru was blocked); the 9th edition (1927) and later editions with any editors' notes; Kedrov's 1958 commentary; the Polytechnic Museum 1905 brochure (in Moscow per ru.wikipedia). Until these are seen, (a) needs no extra assumption but (b), (c) and (d) are open, not refuted.

B. A hypothesis that stops without follow-up

Mendeleev proposed no experiment, and no follow-up by him on the ether (the separate article on absolute zero, any ether measurement) has been found; he died in 1907. Competing explanations: (a) he saw the hypothesis as a question to leave to others (the essay's own stated aim); (b) the new physics (electron, radioactivity, relativity) overtook it quickly; (c) ill health and other work (not evidenced in the files); (d) unpublished papers in the Mendeleev Museum-Archive in St Petersburg, not found online. Records that would decide: the Museum-Archive catalogue (his manuscript, correspondence with Ramsay and Dewar), Kedrov 1958, Bensaude-Vincent 1982 (BJHS 15:183, "un essai malheureux", cited via Google only, text not obtained). Silence in these sources is not contradiction; the essay's wide printing means it was not hidden.

C. Contemporary scientific reception

The Contrarian found no point-by-point refutation and no negative experiment aimed at the gas. Two leads from my own search, not opened: a Nature review (Nature 69, 558, 1904, items 069558a0 and 069558b0) and a Scientific American article "An Attempt at a Chemical Conception of Universal Ether" [web: https://www.nature.com/articles/069558a0.pdf; https://www.scientificamerican.com/article/an-attempt-at-a-chemical-conception]. Titles only; content unverified. They would show how the English-language press received the essay.

D. Smaller dark spots (kept as open)

• Yarkovsky 1901 (ether density), not located.
• Errera 1900 primary text, not located (credited only through Mendeleev).
• Prometheus Nr 736 (one German instalment) missing.
• Antropoff 1926 "element zero" and the June 2026 ZAAC paper, and the Habr/protofluorine ideas: titles only.
• The 1904 table "six more blanks" (Stewart): unsupported by the figure.

5. Open questions

1. What do the 1906, 1927 and 1958 editions show about x and y, and with what editorial note?
2. Did Mendeleev publish the promised article on absolute zero, or any later ether work?
3. Is there a dynamical limit on an unbound, fast, weakly interacting gas that bears on Kelvin's density, as opposed to the local matter density? (The weight claim is medium confidence for that reason.)
4. How did English-language and German-language science receive the essay in 1903-1905 (Nature 1904, Rundschau 1904, Prometheus 1903)?
5. What was Kelvin's derivation of 1e-16 g/m^3?
6. Is Stewart's "six blanks" from Mendeleev's text, the 1906 table or Werner's 1905 table? (Werner "included spaces for two elements lighter than hydrogen and two between hydrogen and helium" [read: Stewart p.8].)

6. What would settle it (compact list)

Reproduction (all [calc] in this file): k = 1.380649e-23 J/K; u = 1.66054e-27 kg; atom mass m = 1e-6 x 1.008 u; v_rms = sqrt(3kT/m); sound speed sqrt(5kT/3m); escape speed sqrt(2GM/R) with G = 6.674e-11, Sirius B M = 1.018 x 1.989e30 kg, R = 5.635e6 m; Oort 0.1 x 1.989e30 kg / (3.0857e16 m)^3; x = 4.8e6 / v^2 (v in m/s).

7. Corrections to Step 1 (facts.json and claims.md)

1. Stewart entry: "identified as ionised iron (by Gruenwald)" is wrong as a credit; the identification is Grotrian and Edlen (1939-1942). Stewart's own sentence and footnote 12 are ambiguous.
2. Stewart entry: "eleven briefly listed predictions" should read "ten named in the body (the abstract says eleven)". "Six more blanks between hydrogen and helium" is not visible in the 1904 table; mark as Stewart's statement, unsupported by his Fig. 2.
3. Essay entry: "mean molecular velocity 1843 m/s" is the root-mean-square speed (the mean speed is 1694 m/s).
4. claims.md claim 3: add that the adopted x is the upper edge of its own range (9.6e-7); that it depends on one assumed temperature (x scales with T, 70 times smaller at 2.7 K); that it does not use the x <= 0.17 periodic bound; and that the mean speed exceeds the Sun's escape speed.
5. claims.md claim 2: add that x <= 0.17 is a three-point quadratic extrapolation (He:x = 23.61), reproducible several ways with an 8-fold spread in x; and that y is shown in series 1 beside hydrogen, x in series 0, without weights.
6. Name corrections from the Russian text: Kirpichev (not "Kirpitchnikoff"), Rogovsky (not "Rostovsky"), Nasini, Andreoli, Salvadori 1898 (not Anderlini 1893).
7. Claim 5 note: Emmens and Fittica entries should carry the web-check result of section 2 (Claim 5).
8. The wavelength of the coronal line is 530.3 nm modern (531.7 nm as printed).

8. Web checks done (budget 8; 6 used)

Not checked on the web (no verdict turns on them): Dewar's phosphorescence mechanism [memory, medium]; neutron-star escape speeds [memory, high]; CMB temperature and electron mass [memory, high]; Jeans escape criterion [memory, high]; Einstein 1920 and Dirac 1951 on "aether" [memory, high, and Lente 2019].

9. Summary table

10. QE revision

Inputs: qe.md (MINOR; no central verdict to change) and the new deep read deep/mendeleev-predictions-bdcaef.md (Stewart 2019, pages 1-7 read in full; "J" = journal page, "PDF" = PDF page, PDF = J - 2). I also checked the essay's extracted text (_text/A-Chemical-Conception-Of-The-Ether-37bf6d.txt) for the QE's citation corrections. No new web calls were made; the web checks of section 8 stand. Tags: [read] from the document; [web] the Investigator's note, URL given; [memory] with confidence.

10.1 Verdict changes

None of the six claims changes verdict. Three scale-use qualifiers are made explicit (QE section 6) and are edited in place in section 2 and the table in section 9: 1. Claim 3, arithmetic: "Established as consistent" becomes "the arithmetic is correct", a finding, not a scale verdict. 2. Claim 6, ether and coronium: "Not testable as stated" means "not a table-based prediction in Stewart's sense". 3. Emmens stays without a scale verdict (as before; the QE accepts this).

10.2 Corrections to the arbiter from the deep read and the QE

1. Citation fixes (checked against the essay text): "a hypothesis" is on PDF p.39 (_text line 1297), not p.47; the Emmens and Fittica passage is on PDF p.17 (_text lines 456-457), not p.9. Both edited in place. Stewart page references below use J pages with PDF pages beside them; earlier mixed uses are to be read that way.
2. Stewart's tally, as he states it [read: J pp.6, 8-9 / PDF pp.4, 6-7]: eight successes (the three big predictions plus five lesser ones), one multiple error (the five lanthanide-zone failures counted as one) and two irrelevancies (aether, coronium); 8 + 5 + 2 = 15 in Scerri's count (half). Stewart's abstract says "eleven other predictions", but the body names ten (six in the 1871 quotation, four scattered); the eleventh is not named. This confirms section 1.
3. Internal slips in Stewart (kept so the reader can weigh him): eka-niobium is En = 146 in his 1871 quotation (J p.4) but 145 in his failure list (J p.6); germanium is dated 1886 (J p.3) and 1885 (J p.6); he says gold "was missing" from the 1904 table (J p.7), but his Fig. 2 prints "Gold Au=197.2" in row 11 [read: page image PDF p.5]. These do not touch the arbiter's claim 6 verdicts, which rest on the atomic weights.
4. Coronium: Stewart's sentence (J p.8) says the green line was "eventually identified as due to a highly ionised form of iron and named 'coronium' by Gruenwald" and cites note 12, an 1887 Chemical News paper titled "On Remarkable Relations Between the Spectrum of Watery Vapour and the Line Spectra of Hydrogen and Oxygen". The sentence does not say who thought coronium explained the line, and the cited title does not show that the paper named coronium. The Fe XIV identification is Grotrian and Edlen (1939-42) [web: https://en.wikipedia.org/wiki/Coronium, section 2]. This sharpens the existing correction in Claim 2 and section 7 item 1; it does not change it.
5. "Six more blanks between hydrogen and helium": the deep reader re-read Stewart's Fig. 2 and found only x (row 0) and y (row 1) above helium, as I found in section 1. Stewart's Werner figure (Fig. 3) is a different matter, below.
6. The matter, force and mind passage: the QE places Mendeleev's own reason for rejecting a unity of matter at essay PDF pp.40-41 ("a distinction must be made between matter, force, and mind"; _text line 1326). It belongs to Claim 5 below.

10.3 Material the QE found missing, now in the record

Stated as what each source says, with its source; no verdict depends on them.

• Contemporaries who entertained sub-hydrogen elements (bears on Claim 2). Stewart: in his 1905 table Werner "included spaces for two elements lighter than hydrogen and two elements between hydrogen and helium, but this just shows that such ideas were around at the time and were not an innovation of Mendeleev's" [read: J p.8 / PDF p.6]. What the figure shows: the deep reader counts four dotted placeholders around hydrogen as printed, three on the top row and one before He, so the split is 3 + 1 and not Stewart's 2 + 2 [read: page image PDF p.6]. Werner won the 1913 Nobel Prize in Chemistry, per Stewart. This makes the zero series less of an outlier in 1902-05 than Claim 2's verdict might suggest. It does not change the verdict: Werner's spaces are blanks with no numbers and no physical argument, and nothing lighter than hydrogen exists (Z = 1 is the floor).
• Dewar's Belfast address, 1902 (bears on Claim 1). Mendeleev cites it as kindred support: Dewar thought "the highest regions of the atmosphere ... must be considered to be the province of hydrogen and of the argon analogues", which Mendeleev reads as "only a few steps from the yet more distant regions of space, and from the necessity of recognising the existence of a still lighter gas capable of permeating and filling space" [read: essay PDF p.51, _text lines 1738-1749]. The second half is Mendeleev's gloss; what Dewar said himself I have not seen (the address was not obtained). Weighed: a leading physical chemist's remark on the upper atmosphere, offered by Mendeleev as support. It does not touch the wave-speed refutation.
• Related later "element zero" ideas (Investigator, priority 3; section 4D previously gave titles only). Antropoff's 1926 "neutronium" is described in Sivchenko (2022, popular article) as an element of about 0.1 of hydrogen's weight that hardly combines and penetrates everything; "nebulium" (nebular lines of 1898, attributed there to Margaret Huggins, assumed weight about 2.74); and "protofluorine", a light halogen that matches Mendeleev's own footnote 17 [web: https://habr.com/ru/articles/653695/, via research/2026-10-06-habr.md]. Sivchenko compares newtonium to the neutron and neutronium to the neutrino; that is his interpretation. He also reverses Mendeleev's own labels (x = newtonium, y = coronium per the Russian text), so the article is not a source for the labels. The June 2026 ZAAC paper "100 Years of Element Zero" is known by title only. The mass-range coincidence with neutrinos (section 1) is weighed there; these later proposals are listed as the same family of ideas, not as support.
• Mendeleev's funded ether programme and the spiritualism commission (bears on Claims 1 and 3). Stewart says Mendeleev threw himself, from 1872, into gas research under a Commission for Research in the Elasticity of Gases, "His ambition was to establish the nature of the 'luminiferous aether'", resigning in January 1881 [read: J p.6 / PDF p.4]. He adds that this interest "led naturally into concern with spiritualism", and that in May 1875, "at Mendeleev's instigation", the Russian Physical Society set up the Commission for the Investigation of Mediumistic Phenomena "with the object of re-establishing science as the source of truth", which he left after about a year when it became clear that some members were themselves spiritualists (source: Gordin 2004, pp. 81-111, as Stewart cites it) [read: J p.6]. The link between the two is Stewart's characterisation; the text gives no Mendeleev document for it. Mendeleev's own 1874 low-pressure Boyle's-law work is covered by section 2, Claim 1's context.
• The 1906 Nobel vote. Stewart: nominated for 1906, the prize going to Moissan "by 5 votes to 4"; no citation and no committee named [read: J p.8 / PDF p.6]. [memory, medium: the Nobel committee records, not checked.] It shows how the scientific world rated Mendeleev in these years; it carries no weight for any claim.
• Crookes's "fourth state": Mendeleev says his gas ether does not require the recognition of a peculiar fourth state [read: essay PDF p.22; _text line 620]. Context for Claim 1 only.
• Contemporary warnings: Stoney and Rogovsky argued that H and He escape the Earth's atmosphere, and Mendeleev notes both were found there soon after [read, via Contrarian and QE]. It is a ready worked caution on the mean-speed escape criterion used in Claim 3.

10.4 Claim 5: Mendeleev's own stated reason

Mendeleev's rejection of a unity of matter rests, in his own words, on a distinction "between matter, force, and mind" [read: essay PDF pp.40-41]. The three reviewers and I treated "atoms are individuals, so electrons are metachemical" as circular (Claim 5). That stands. The QE point is that the reader should see the stated framing, which the earlier arbiter text left out. The verdict is unchanged.

10.5 Changes still owed by other files (not done here)

• facts.json (Stewart and essay entries) and claims.md: QE R1-R3, including the root-mean-square wording and the Stewart corrections of section 7 and 10.2 above.
• primer.md: QE R5, including the Werner, Dewar, ether-programme and Nobel context above and the "below the hydrogen row" slip.
• Investigator reads (QE R6, R7): the 1905 Principles of Chemistry appendix, the 1905 Russian table p.25, Lente's PDF, Prometheus 737, and the 1906/1927/1958 records, which decide Section 4A. Until they are read, section 4A and the open questions stand unchanged.