Mendeleev-D: Science Contrarian review (Step 2)

Project: Mendeleev's "An Attempt Towards a Chemical Conception of the Ether" (written October 1902; English translation 1904) and Stewart's 2019 review of his predictions. Reviewer role: judge by experimental evidence, build the strongest honest case, examine how the idea was handled, and apply the same rigour to the inventor. Source tags: [read] from the project documents or Investigator files, [web] with URL, [memory] with confidence.

No source gap: the essay and the Stewart article have full text layers (Step 1). I add that I have also read the Investigator's 1905 Russian scan myself (see Corrections).

0. How the case stands, in short

Mendeleev was a leading chemist in 1902, not an outsider. He used a method that had worked (periodic extrapolation) on a problem it could not reach (a medium for light). The essay is candid that it is "a hypothesis" and "an expression of a series of thoughts" [read: p.39, p.58-59]. Two parts of it came out right (a zero group of inert gases; the idea that radioactive "emanation" is something material). Its centre, an ether that is a chemical element lighter than hydrogen, was never experimentally confirmed, and the measurements that bear on the ether family (Michelson-Morley and successors) went against it. The essay does not mention those experiments at all (my text search of the English text for "Michelson", "Maxwell", "Hertz", "Lorentz" finds nothing) [read: _text search of A-Chemical-Conception-Of-The-Ether-37bf6d.txt]. That is the main weakness to put to the inventor.

Important to the reader: this was dismissed rather than tested. No experiment was designed to detect Mendeleev's particular gas, because the essay proposes none. What was tested is the ether family it belongs to.

1. Corrections to Step 1

1. Coronium identification (facts.json, Stewart entry). Step 1 writes that the green line was "identified as ionised iron (by Gruenwald)". Stewart's text says the line was "eventually identified as due to a highly ionised form of iron and named 'coronium' by Gruenwald", with footnote 12 citing A. Gruenwald, Chemical News 56 (1887), p.232, a paper on water-vapour spectra [read: mendeleev-predictions p.8, note 12]. That reference cannot be the identification of a coronal line (it is dated 1887, before it was resolved). The identification of the green line as Fe XIV (Fe13+) is credited elsewhere to Grotrian and Edlen, announced 1939-1941 [web: https://en.wikipedia.org/wiki/Coronium]. The modern wavelength is 530.3 nm, against 531.7 nm in Mendeleev's text (a small difference between 1869-era and modern scales) [web: same]. Step 1's "(by Gruenwald)" should not be repeated; Stewart's attribution looks like an error in the article, flagged here as my reading of his footnote.
2. "Eleven" predictions. facts.json says "Five of Mendeleev's eleven briefly listed predictions". Stewart's abstract says "eleven other predictions", but the body names ten: five successes (Tc 100, Re 190, Po 212, Fr 220, Pa 235) and five in the lanthanide dead zone (140, 145, 155, 170, 175) [read: mendeleev-predictions p.4-5]. Scerri's scorecard (eight successes, five failures) also adds up to ten with the big three counted as three. The count of eleven is Stewart's abstract only; I could not find the eleventh in his text. Claim 6 should carry the discrepancy.
3. Where x and y are printed (ru.wikipedia dispute). The Investigator left the 1905 Russian table unchecked and recorded that ru.wikipedia says x and y are absent from the tables. I rendered page 25 of the 1905 scan: the table prints a row "0" with x in the zero group, and a row "1" with y beside hydrogen (H 1.008), both without atomic weights, and no names ("newtonium", "coronium") in the cells [read: page image, Mendeleev-1905-Popytka-StPetersburg-Frolova-scan.pdf p.25]. The accompanying text (p.22) says "under the signs x and y already stand expected elements with atomic weights less than hydrogen's" [read: p.22]. So the Wikipedia statement is wrong for the essay's own table (agreeing with the Investigator's inference). It remains unchecked for the 8th edition of Osnovy khimii (1906).
4. Radium atomic weight. The 1904 English table gives Rd 224 [read: deep read]; the 1905 Russian table prints Rd 225 [read: 1905 scan p.25]. A small revision between editions, consistent with the Investigator's finding that the 1905 text was revised.
5. Space temperature and coefficient. Step 1 states that the English "a = 0.000367" is as printed. The Russian footnote 24 gives 0.00367 (Mendeleev and Kayander) [read: ru-wikisource-popytka text, footnote 24]. I recomputed: 1843 x sqrt(2 x (1 - 80 x 0.00367)) = 2,191 m/s, so x = 4.8e6/v^2 holds with 0.00367. Step 1 is right to call it a misprint.
6. Names. The Investigator's corrections (Kirpichev for "Kirpitchnikoff"; Rogovsky for "Rostovsky"; Nasini, Andreoli, Salvadori 1898) are consistent with my reading of the Russian text and the 1905 scan (the Russian text lists "Nasini, Andreoli i Salvadori, 1898" [read: 1905 scan p.24]). Use the Russian forms.
7. Deep-read item left open. The deep read tags p.57 with a question mark on "130 times". I read the English text near pp.50-52 and p.56 myself: it gives the ether-atom speed as 2,250 km/s against 300,000 km/s for light (ratio 133 by my arithmetic), and p.56 says light is "a certain transverse vibration of ether" [read: pp.51, 56]. That ratio matters for claim 1 (see below).
8. Step 1 confirmations. I recomputed the essay's numbers and they are right: 4.8e6/v^2 gives x = 0.038 (Earth, 11,190 m/s), 1.3e-5 (Sun, 608 km/s), 9.6e-7 (2,240 km/s), 5.3e-11 (c); the finite-difference extrapolation of the ratios 1.56, 2.15, 9.5 gives 23.61, i.e. He:x = 23.6 and x = 0.17 as stated. The two PDFs differ only in page rotation (deep read), and I accept that finding.

2. Claim-by-claim review

Claim 1: The ether is a real chemical element, a monatomic gas with weight that permeates all bodies and cannot combine

a. Mathematics. The essay has no field equations for the ether, so there is little mathematics to be inconsistent. There is one internal tension that is not mathematical but logical: the gas is "incapable of chemical combination" [read: p.22-23], yet it permeates solids "as hydrogen through palladium", which Mendeleev himself explains as formation of indefinite compounds [read: pp.19-21], and it "dissolves in or accumulates about" uranium, thorium and the Sun [read: p.54]. A gas that does not combine cannot permeate by compound formation; he partly answers this by pointing out that helium and argon dissolve in liquids (p.54), which is a physical, not chemical, process. That answer does not rescue permeation of metals.

b. Established experiments. Against the strong form: - Ether-wind experiments: the Michelson-Morley experiment and successors found no motion of Earth through a medium; modern rotating optical-cavity tests limit anisotropy of c to about 1e-17 [web: https://arxiv.org/pdf/1002.1284]. A gas ether with thermal speeds of 2,250 km/s would be expected to show something at the 30 km/s of Earth's orbital speed unless dragged along perfectly, which a gas permeating a vacuum cannot be. - Kinetic theory itself (my calculation). A gas supports sound at roughly its thermal speed. Mendeleev gives an ether-atom speed of 2,250 km/s and light 130-odd times faster. A perfect gas cannot carry a transverse wave at 133 times its own molecular speed; ordinary gases carry no transverse waves at all (no shear stiffness). With Kelvin's density of about 1e-16 g per cubic metre (as Mendeleev quotes it) and atoms of 1e-6 u, the number density is about 6e13 per cubic metre and the pressure is 1.6e-7 Pa at 193 K; that is a nearly collisionless gas for any plausible atomic cross-section, so it has no fluid waves either (my arithmetic; the cross-section assumption is mine). The essay itself raises this ratio only in passing. - No experiment was ever aimed at the weight-of-ether claim. Kelvin's figure is quoted by Mendeleev [read: p.10]; I did not find Kelvin's original, so I treat the number as Mendeleev's report. Against the weak form (some very light, weakly interacting particle population that gravitates and passes through matter): nothing in the experiments excludes it. Modern physics does contain things in that spirit: the vacuum as a medium with properties (Einstein 1922 and Dirac 1951 on "aether" are cited by Lente) [web: Lente 2019, https://link.springer.com/article/10.1007/s40828-019-0092-5, via research/2026-10-06-springer-lente.md]. Mendeleev's mass range (5.3e-11 to 1e-6 u, i.e. about 0.05 eV to 930 eV) overlaps the mass scale of cosmological neutrinos (upper limit below about 1 eV [memory, medium]); the Habr article draws that parallel [web: https://habr.com/ru/articles/653695/, via the Investigator]. I treat that as numerology: Mendeleev's x is a chemical gas serving as the medium for light, while neutrinos are fermions that do not carry light, and nobody derived the range from particle physics.

c. Testing. The ether family was tested hard (1881 onward). Mendeleev's own gas was never the subject of a test; no one tried to measure an element x; the essay proposes no test ("I have no intention of raising [the problems]", p.52). Contemporary reception: the Investigator found the essay widely printed in his lifetime (Russian journal 1903; German Prometheus 1903 and Naturwissenschaftliche Rundschau 1904; English 1904; textbook appendix 1905) [read: sites.md "Leads for review" section 4]. Later reactions found are historical commentary: Bensaude-Vincent's 1982 paper title calls it "un essai malheureux" (an unfortunate attempt) [web: via Google, research/2026-10-06-google.md]; its text was not obtained. So the reception record is: published widely, then left alone, rather than refuted point by point. I found no record of a controlled attempt to detect his gas, and no record of blocked funding or publication; I also found no negative experiment reanalysed or omitted. This differs from cold-fusion-type disputes.

d. What would change if true. Special relativity's rejection of a preferred medium would need a dragged or Lorentz-covariant gas; the periodic table would need elements with atomic number below 1; atomic number as nuclear charge (Moseley 1913) would have to be reinterpreted [memory, high]. Photon propagation would need a mechanism not given by Maxwell's equations plus a gas.

Verdict (split). - Strong form (a gas of an element that carries light, with a medium rest frame): Tested, not supported (ether-drift tests of the family) and Refuted by calculation for the gas-as-light-carrier version (wave speed 133 times the molecular speed; no shear). - Weak form (a sub-hydrogen, inert, very light gas of unknown role): Never properly tested and not dismissed by experiment; the essay's own proposal gives no testable consequence, so in practice it is Not testable as stated.

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

a. Mathematics. The zero group is a qualitative table placement, not a derivation. The extrapolations of y and x are number games on ratios: y at most 0.4 from "He:y at least 10, probably greater" (Li:H = 6.97, Ne:He = 4.98, and so on) [read: p.36], and x at most 0.17 from three ratios (Xe:Kr 1.56, Kr:Ar 2.15, Ar:He 9.5) giving He:x = 23.6 [read: p.42]. I checked: a quadratic through three data points is exactly determined, so the fit has no residual and cannot test anything; the same three numbers with a different rule (geometric, say) would give another answer, and Lente's reconstruction reaches the same conclusion by log-plot and finite differences [web: Lente 2019]. Internally the arithmetic is consistent. As an argument it is under-constrained. Also, x = 0.17 contradicts the later x of one millionth; Mendeleev treats 0.17 as only the maximum (a bound), so this is not a contradiction.

b. Experiments. For the zero group: confirmed. The noble gases He, Ne, Ar, Kr, Xe sit between the halogens and alkali metals, and radon (Z = 86) completes the column [memory, high]. Mendeleev credits Errera (Belgian Academy, 5 March 1900) and Ramsay (Berlin, 19 March 1900) for the placement [read: 1905 scan footnote 13-14; web: Wikisource text via the Investigator]; the essay is a late adoption of their move, and he admits "such a zero group could not possibly have been foreseen" in 1869 [read: p.33]. For elements lighter than hydrogen: none exist; atomic number 1 is the lowest, and the neutron (1932) is a particle of mass about 1 u, not a gas of 1e-6 u [memory, high]. For coronium: the 1869 green line is Fe XIV (Grotrian, Edlen, 1939-1941) [web: https://en.wikipedia.org/wiki/Coronium]. The claim that coronium is an element lighter than hydrogen was tested by spectroscopy and failed. Credit to be given: in 1902 the line was unidentified, other proposed light elements (nebulium, 1898) were also later explained as ionised known elements [web: Habr article via Investigator, medium], so Mendeleev's guess was a natural one for the time. It was also not a "prediction" from properties (Stewart) [read: p.9].

c. Tests. The zero group: tested and confirmed (discoveries of Ne, Kr, Xe in 1898 and radon 1900-1902). x and y: no test possible beyond the spectroscopic one (y), and none was designed for x. The 1905 table printed x and y without weights [read: 1905 scan p.25]; whether the 1906 edition printed them remains an open historical record.

d. If true. The periodic law would be bounded below by something other than H; atomic number would lose its meaning as nuclear charge.

The "removed from the table" claim (absence-of-evidence analysis). Russian advocates (Atsyukovsky; Rodionov) say the zero row was later deleted or falsified [web: Investigator's research/2026-10-06-google.md; both are advocates of an ether and their statement is theirs]. Competing explanations: (a) the idea was dropped by its author's successors after relativity (1905), Moseley (1913) and the Bohr atom (1913) removed any place for elements below Z = 1, as Lente and ru.wikipedia say; (b) deliberate removal. Evidence for (a): the idea was a speculation that Mendeleev himself labelled a hypothesis; the zero group (noble gases) stayed in every table and the zero series left; no measurement ever supported x. Evidence for (b): none found in the files; the claim rests on the advocates' statements and an unverified 1906 table. The deciding record is the 1906 8th edition and the 1927 9th edition and Kedrov's 1958 commentary; none has been obtained. Both explanations are open until then, but (a) needs no extra assumption.

Verdict (split). Zero group of noble gases: Established. Coronium as y: Tested, not supported (Fe XIV). Zero series with ether element x: Historically superseded but instructive (the extrapolation is a three-point fit that tested nothing).

Claim 3: x has atomic weight about one millionth of hydrogen and speed about 2,250 km/s

a. Mathematics. Internally consistent once the printed coefficient is read as 0.00367 (Corrections 5 and 8). Verified: v = 1843 sqrt(2(1 + at)/x) with t = -80 C gives v = 2,191/sqrt(x); my independent check using the rms speed of a gas of mass 1e-6 u at 193 K gives 2,194 km/s (rms) or 2,021 km/s (mean), within the essay's tolerance. The bounds follow: 0.038, 1.3e-5, 9.6e-7, 5.3e-11. The step from a range of four orders of magnitude (5.3e-11 to 9.6e-7) to "nearly one millionth" is a choice: he rejects the lower end "because it would in some measure answer to a revival of the emission theory of light" [read: p.51]. That is an argument from theoretical preference, not from the data, and it should be called out as such.

b. Consistency with experiment. - Space temperature. Mendeleev takes -80 C because "no one has found it less than -150 C or above -40 C" [read: p.44]. The Russian footnote says the temperature may differ by region and that he doubts absolute zero [read: Wikisource footnotes 22-23]. The cosmic microwave background is 2.7 K (discovered 1965) [memory, high], and interstellar gas ranges from tens of K to millions of K; at 2.7 K the same atoms would move at about 260 km/s by my arithmetic, which changes the lower bound on x by a factor of 70 (v^2 scales with T) and puts the speed below the essay's maximum-mass bound for stars. - The escape-velocity criterion is a sharp cut-off on the mean speed. Real retention uses the Maxwell tail (Jeans escape). Mendeleev himself notes that Stoney and Rogovsky argued that H and He were absent from the atmosphere, and "very soon" both were found [read: 1905 scan p.24, footnote 11]; his own example shows that a mean-speed criterion misleads. - Consistency with the rest of the essay (internal): the mean speed of 2,250 km/s exceeds the Sun's escape speed of 608 km/s [read: p.46] by a factor of 3.7, so by his own numbers the ether is not held by the Sun, yet the essay argues the Sun accumulates ether "just as the gases of the atmosphere accumulate in a drop of water" [read: p.54]. Gravitational capture is impossible for the bulk of the gas; accumulation then rests on the unexplained "attractive capacity... between gravity and chemical affinity", which contradicts "no combination". An x of 1e-6 u is also about 1/550 of an electron's mass (1.8e-3 electron masses, my arithmetic) [memory, high: electron mass 9.109e-31 kg], a scale then known only through Thomson's and Kaufmann's electron measurements, which Mendeleev rejected as a division of the atom.

c. Tested? No observation can distinguish 1e-6 from other values in his range; no measurement was made. No replication is meaningful.

d. If true. A massive sea of particles at a few thousand km/s would require explaining why it exerts no detectable drag, pressure or heating (with Kelvin's density, the numbers are small, but that is by assuming the density).

Verdict. Not testable as stated (the number is an assumption-driven choice inside a wide bound; the arithmetic is correct). The sub-claim that the mean speed lets the ether leave the Earth and Sun but be held by very massive stars is Internally inconsistent with the essay's own accumulation argument (claim 4).

Claim 4: Radioactivity, the Curie ZnS demonstration and Dewar's phosphorescence support the ether

a. Mathematics. None offered; qualitative. The internal tension (an inert gas that "dissolves in" heavy atoms) is noted under claims 1 and 3.

b. Experiments. The facts are real; the reading is not. The Curie flask experiment [read: pp.55-56]: opening the tube between a radium solution and zinc sulphide makes the sulphide glow. Mendeleev reads this as "an emissive flow of something material" and, more precisely, as ether atoms entering and leaving. The mainstream reading in 1902 was already forming: radioactive substances give off a "radioactive emanation", a gas (Rutherford for thorium 1899; Dorn for radium 1900; Rutherford and Soddy concluded by 1902 that the emanation is an inert gas, and later that it is a new element) [web: https://iupac.org/publications/ci/2011/3301/8_radvanyi.html]. That gas is radon, Z = 86, the heaviest noble gas. So Mendeleev's observation that something material flows is correct, and the material is a zero-group gas, as his table would predict, but a heavy one, not a light one. This is the strongest honest point in his favour, and it is also the reason his ether reading fails: the gas that excites the sulphide is radioactive and decays (Rutherford-Soddy, 1902-1903, radioactivity as transmutation) [web: same]. Dewar's observation that phosphorescence brightens at liquid-air temperature is an established property of phosphors (reduced non-radiative thermal quenching) [memory, medium]; nothing in it needs ether.

c. Tests. The radon-transport experiment was repeated and quantified within a few years (half-life, mass, spectrum). I did not find a test specific to "ether accumulates on U and Th"; no measurement was proposed. Dismissed in the sense of superseded by a fuller explanation, not by a test of Mendeleev's version.

d. If true. Nuclear decay (alpha emission, later Gamow's tunnelling) would need to be replaced or supplemented by ether adsorption [memory, high].

Verdict. The observations (Curie flask, radioactive emanation, low-temperature phosphorescence) are Established as phenomena. Ether accumulation as their cause: Historically superseded but instructive (the "material flow" was radon).

Claim 5: No primary substance; atoms are indivisible and not transmutable; only the ether is a unity

a. Mathematics. None. Logically, "atoms are indivisible and therefore electrons cannot be parts of atoms" [read: p.14] assumes what it sets out to prove.

b. Experiments. Three parts. - Dismissal of Emmens (silver to gold) and Fittica (phosphorus to arsenic) [read: p.17]: both were claims by others; Mendeleev says they "always proved unfounded". This part is consistent with later experience [memory, medium; I did not check Emmens or Fittica separately]. - Atoms are indivisible and elements cannot change: contradicted. The electron as a constituent of atoms (Thomson 1897) was known; radioactive transmutation was published by Rutherford and Soddy in 1902-1903 [web: https://iupac.org/publications/ci/2011/3301/8_radvanyi.html]; artificial transmutation followed (Rutherford 1919) [memory, high]. Mendeleev's essay (October 1902) saw radioactivity first-hand (he visited Becquerel and the Curies) but did not accept transmutation, and called electrons "not so much metaphysical as metachemical" [read: p.14, p.25, p.53]. - Opposition to Prout's hypothesis (all elements built of hydrogen): Lente calls it a "highly personal" motive for placing elements lighter than hydrogen [web: Lente 2019]. Note that the proton and the nucleus are, in modern terms, a form of Prout's idea [memory, high].

c. Tests. Transmutation and atomic structure were tested exhaustively after 1902 and not supported in Mendeleev's form.

d. If true. Nuclear physics would need to be rewritten.

Verdict (split). Rejection of Emmens and Fittica: Established. Indivisible, non-transmutable atoms and no common constituent: Tested, not supported. Unity of the ether substance: Not testable as stated.

Claim 6 (secondary, from Stewart 2019): Mendeleev's earlier predictions largely succeeded; ether and coronium were not predictions from element properties

a. Mathematics. Stewart's accounting is not wholly consistent (abstract "eleven", body ten; Corrections 2). The classification depends on the author's own rule that a prediction counts "if it falls on a higher analogue of the element on which it is based" [read: p.4]; that rule is Stewart's, and it is the rule that lets Stewart treat the five lanthanide-zone failures as one mistake. Scerri counts them as five. Both counts are shown in the article.

b. Experiments. Gallium (1875), scandium (1879), germanium (1886) matched in detail [read: Stewart p.2; memory, high for the discoverers]. Technetium (100 against the later 98), rhenium (190 against 186.2), polonium (212 against 209), francium (220 against 223) and protactinium (235 against 231) are within a few percent [read: p.5]. Technetium was found in 1937 as an artificial element [memory, high], which Stewart does not discuss. These are real and independent confirmations, made by other people.

c. Tests. Confirmed by independent discovery. This is the strongest evidence in the project, and it applies to the periodic law, not to the ether.

d. If true / relevance. It shows that Mendeleev's interpolation worked and his extrapolation (below the first element) did not. Interpolation is constrained by neighbours on both sides; extrapolation beyond the end of the table is not.

Verdict (split). Ga, Sc, Ge: Established. The five minor successes: Established (with the criteria dependence above). The five lanthanide-zone predictions: Tested, not supported (one structural mistake, per Stewart; five per Scerri). Ether and coronium as predictions: not predictions by Stewart's reading; see claims 1 and 2.

3. The strongest honest case for Mendeleev's ether paper

1. A leading experimentalist in gases (his 1874 work on departures from Boyle's law at low pressure is confirmed by later work [read: p.11]) proposed a specific, quantitative model of the ether, with stated bounds and an explicit admission that it is hypothesis ("if my conception be proved false in its basis, it will prevent others from repeating it" [read: p.58-59]).
2. The zero group came out right, and he accepted it against his own earlier view of argon as N3 [read: pp.26-28]; he admits his earlier error in print.
3. Radioactive "emanation" is a material gas; his intuition that something material flows from the radioactive substance was correct, and the substance is an inert gas of the very group he added.
4. The essay was widely published and not suppressed [read: sites.md]; there is no evidence of blocked funding or of omitted data. The negative result came from the ether family's experiments and from nuclear physics, not from a bad test of this gas.
5. The Russian 1905 text adds ideas the English lacks (ether as a mixture of several gases; the name "newtonium"; his doubt of absolute zero) [read: Wikisource footnotes; 1905 scan footnote 18]; the English 1904 text in the project is thus an incomplete picture of his position. The Investigator found that the German abstract of 1904 already uses "Newtonium" [web: https://archive.org/details/naturwissenschaf19brau, via research/2026-10-06-internet-archive.md].
6. Dirac (1951) and Einstein (1922) revisited the word "aether", and quantum fields give the vacuum properties [web: Lente 2019]. This is no support for Mendeleev's gas.

4. Where the inventor is weak (same rigour)

• Ignores the ether-drift experiments and electromagnetic theory of light (nothing in the essay on Michelson, Maxwell, Hertz, Lorentz).
• Takes the speed of light as 133 times the molecular speed without addressing how a gas carries transverse waves.
• Picks the final mass from a four-order range by theoretical preference (emission theory).
• Uses a three-point exact fit as an extrapolation (x = 0.17); the German edition says "second-order parabola" [web: Lente 2019]. A fit with no free parameter predicts nothing.
• Assumes -80 C for space without measurement; his own footnote concedes regional differences.
• Treats the Sun as accumulating ether while his own numbers say the Sun cannot hold it.
• Rejected electrons and transmutation as the evidence arrived (1897-1903).
• Mistakes in the translation: the printed 0.000367 (Corrections 5), "Kirpitchnikoff" for Kirpichev (Corrections 6).

5. Investigator findings weighed (kept whatever the verdict)

• Russian original and 1905 edition (Wikisource; scan on ether-wind.narod.ru): fuller statement of the claims; newtonium, mixture of gases, Errera and Ramsay priority, Yarkovsky's ether-density pamphlet (1901), doubt of absolute zero, coefficient 0.00367, ether and solar energy. Weigh as the inventor's own later text.
• German translations (Prometheus 1903, Naturwiss. Rundschau 1904) and Principles of Chemistry appendix 1905: show wide circulation; the German adds the "parabola" method. Prometheus Nr 736 is missing from archive.org; one instalment is unread.
• Yarkovsky 1901 (ether density), Errera 1900 and Ramsay 1900: supporting information by others for the weight of the ether (Yarkovsky: not located) and for the zero group (Errera and Ramsay: located only as Mendeleev's citation). No replication or test of a sub-hydrogen element found.
• Lente 2019: a mainstream chemist's reconstruction (geometric series for y; parabola for x) and his judgement that the predictions failed. His characterization; I checked the arithmetic.
• Antropoff 1926 "element zero" (neutronium), nebulium 1898, protofluorine (Habr; ZAAC 2026): related sub-hydrogen proposals; they are different ideas, and none was confirmed. Not read in the original.
• Atsyukovsky and Rodionov on a "genuine 1906 table" and removal: advocates' claims, recorded as theirs. Not supported by the primary scan I read for 1905; the 1906 edition is the deciding record, not obtained.
• ru.wikipedia (x and y not printed): contradicted by the 1905 scan (Corrections 3).
• Blocked or unreached: rusneb.ru (VPN block page), Bensaude-Vincent 1982 text, the Mendeleev Museum-Archive, RGB catalogue, the Kedrov 1958 commentary, the 1906 edition. These are the records that would settle the dispute about how and when x and y were withdrawn.

6. Web checks done (budget 6; 3 used)

• Coronium = Fe XIV, Grotrian and Edlen: confirmed [web: https://en.wikipedia.org/wiki/Coronium].
• Emanation and Rutherford-Soddy transmutation, 1899-1903: confirmed [web: https://iupac.org/publications/ci/2011/3301/8_radvanyi.html].
• Modern ether-drift limit about 1e-17: confirmed [web: https://arxiv.org/pdf/1002.1284]. Not web-checked, left as [memory]: neutrino mass scale (medium), Emmens and Fittica outcomes (medium), Dewar phosphorescence mechanism (medium), CMB temperature and electron mass (high).

7. What would settle the open points

• The 1906 8th edition of Osnovy khimii and the 1927 9th edition: whether x and y were printed and what editors said when they vanished.
• A physical test for the weak form: any measured sub-eV, weakly interacting relic population is a mass-range coincidence and would not vindicate the chemical-gas model, which needs light to travel in the gas.
• Bensaude-Vincent 1982 and Kedrov 1958: how contemporaries received the essay.

8. Summary table