Mendeleev's ether as a chemical element: a primer

Sources. No source gap. The project folder holds two copies of the same book, the 1904 Longmans English translation (by George Kamensky) of Dmitri Mendeleev's essay An Attempt Towards a Chemical Conception of the Ether, written in October 1902. Both copies have a full text layer; they differ only in how six pages are rotated in the PDF [read: deep read]. The folder also holds Philip J. Stewart, "Mendeleev's predictions: success and failure", Foundations of Chemistry 21 (2019) 3-9, a modern secondary review. The Investigator found the Russian original (Wikisource), a scan of the revised 1905 Russian edition, two German translations (1903, 1904) and a 1905 English reprint; these add material that the 1904 English lacks and are used here as the inventor's own later text. Not obtained: the 1906 8th edition of Mendeleev's textbook Osnovy khimii (Principles of Chemistry), the 1927 9th edition and Kedrov's 1958 commentary, which decide the question "were x and y later removed from the table?" (section 6).

Tags. [read] from the essay or Stewart (page numbers are PDF pages of the 1904 copy; printed page = PDF page minus 8); [web] with URL (most found by the reviewers or the Investigator); [memory] from training, with confidence; [calc] computed in example.py, which you can run yourself:

.venv/bin/python review/Mendeleev-D/example.py

This primer adds no verdicts. Verdicts are the arbiter's (arbiter.md).

---

1. The idea in one paragraph

In 1902 Dmitri Mendeleev, the Russian chemist who built the periodic table and famously predicted gallium, scandium and germanium before anyone found them, turned his table on the biggest open question of physics: what is the ether, the invisible medium that was thought to fill space and carry light? Most physicists treated the ether as something beyond ordinary matter. Mendeleev wanted it brought into chemistry. He argued that it is a real chemical element: a gas like helium or argon, so inert that it never combines with anything, so light and fast that it slips through all solid bodies and cannot be held by the gravity of any planet or star. To give it a home he added two things to the periodic table: a "zero group" for the newly found inert gases (helium, neon, argon, krypton, xenon), and, in a new row before (above) hydrogen, a "zero series" of elements lighter than hydrogen, holding the ether element he called x (named "newtonium" in the Russian text), with a second unknown element y between x and hydrogen, perhaps the "coronium" seen only as a green line in the Sun's corona. From the kinetic theory of gases and the escape speeds of the Earth, the Sun and massive stars he estimated that an ether atom weighs about one millionth of a hydrogen atom and moves at about 2,250 km/s. He saw support in radioactivity, which he had just watched in the Curies' laboratory. He called the essay "a series of thoughts" and "a hypothesis" whose "one object is to bring this question to the fore" [read: pp.39, 58-59]. The zero group was right and is today's Group 18. The ether element was not, and this primer shows why, using only the kinetic theory Mendeleev himself used.

---

2. What you need to know first

The kinetic theory of gases. A gas is a crowd of molecules flying in straight lines and bouncing off each other and the walls. The pressure on a wall is the momentum they deliver per second per unit area. Working this out gives P = (1/3) ρ ⟨v²⟩, where P is pressure (pascals, Pa), ρ (rho) is the mass density (kg/m³), and ⟨v²⟩ is the average of the squared molecular speed. Combining with the ideal-gas law gives the root-mean-square (rms) speed v_rms = √⟨v²⟩ = √(3kT/m), where k = 1.380649 × 10⁻²³ J/K is Boltzmann's constant, T the absolute temperature in kelvin and m the mass of one molecule. Two consequences carry this whole project: at a fixed temperature, lighter molecules move faster (v ∝ 1/√m), and at a fixed mass, hotter gas moves faster (v ∝ √T). Read: Schroeder, An Introduction to Thermal Physics, chapter 1 (the microscopic model of an ideal gas); Halliday, Resnick and Walker, Fundamentals of Physics, the chapter on the kinetic theory of gases.

The Maxwell speed distribution and the Boltzmann factor. The molecules do not all move at v_rms; their speeds spread over a bell-like curve (the Maxwell distribution) with a long tail of fast molecules. The mean speed is √(8/3π) ≈ 0.92 of the rms speed. A related rule, the Boltzmann factor, says that in a gas at temperature T the density at a place where a molecule has extra potential energy E is reduced by e^(−E/kT). So gravity concentrates a gas near a planet only when the gravitational energy of a molecule is comparable to kT. Read: Schroeder, chapter 6 (Boltzmann statistics, the Maxwell speed distribution).

Escape speed. A body thrown up from a planet of mass M and radius R escapes for good if its kinetic energy beats its gravitational binding: ½mv² ≥ GMm/R, so v_esc = √(2GM/R), where G = 6.674 × 10⁻¹¹ N m²/kg² is Newton's gravitational constant. For Earth this is 11.2 km/s. A planet keeps an atmosphere of a gas only if that gas's typical molecular speed is well below v_esc (in practice, below about one sixth of it, because the fast tail of the Maxwell distribution leaks away over billions of years: this is "Jeans escape") [memory, high]. Read: Taylor, Classical Mechanics, the chapter on gravitation or central forces; any astronomy text on planetary atmospheres.

Waves in a medium: sound versus shear. A wave travels through a medium because neighbouring bits push or pull on each other. Two kinds are possible. A longitudinal (compression) wave, like sound, moves the medium back and forth along the direction of travel; any gas, liquid or solid carries it. In a gas its speed is v_s = √(γP/ρ) = √(γkT/m), where γ is the ratio of specific heats (5/3 for a monatomic gas such as helium). Compare this with v_rms = √(3kT/m): the sound speed in a monatomic gas is √(5/9) ≈ 0.745 of the rms molecular speed. A sound wave cannot outrun the molecules that carry it. A transverse (shear) wave moves the medium sideways, across the direction of travel, like a wave on a rope. It needs a medium that resists being sheared, a solid with a shear modulus G_s, and travels at √(G_s/ρ). A gas or liquid has no shear stiffness, so it cannot carry transverse waves at all. Read: Halliday, Resnick and Walker, the chapters on waves and on sound waves; for shear waves in solids, any text on elasticity or seismology.

Light is a transverse wave. Light can be polarized (Malus, 1808; Fresnel and Young, 1810s-1820s), which means its vibration is sideways to its direction of travel [memory, high]. Maxwell's equations (1865) describe light as transverse oscillating electric and magnetic fields moving at c = 1/√(μ₀ε₀) ≈ 3.00 × 10⁸ m/s, where μ₀ and ε₀ are the magnetic and electric constants of empty space. Read: Griffiths, Introduction to Electrodynamics, chapter 9 (electromagnetic waves).

The periodic table, then and now. Mendeleev's 1869-1871 table ordered elements by atomic weight (the mass of an atom relative to hydrogen, H = 1) and grouped elements with similar chemistry in columns. Gaps in it let him predict new elements by interpolation, reading properties off the neighbours above, below and beside the gap. Since Moseley (1913) the table is ordered by atomic number Z, the electric charge of the nucleus in units of the proton's charge. Hydrogen has Z = 1; there is no chemical element below it [memory, high]. Read: Eric Scerri, The Periodic Table: Its Story and Its Significance (Oxford University Press).

The ether in 1902. From Fresnel to Maxwell, physicists believed light needed a medium to wave in, as sound needs air. Because light is transverse, that medium had to behave like an elastic solid, yet planets move through it without any drag: the "elastic-solid ether" puzzle. The Michelson-Morley experiment (1887) looked for Earth's motion through the ether and found none. Einstein's special relativity (1905, three years after this essay) dropped the medium altogether [memory, high]. Read: Whittaker, A History of the Theories of Aether and Electricity (the classic history, readable without tensors); any modern physics text, the chapter on special relativity.

---

3. The textbook view

3.1 How fast are the atoms of a gas?

From section 2: v_rms = √(3kT/m). In chemists' units, with the gas density ρ and pressure P, v_rms = √(3P/ρ). Mendeleev starts the same way: "a litre of hydrogen at 0° and 760 mm weighs about 0.09 grms", which gives "1843 metres a second" [read: p.43]. Check: with P = 101,325 Pa and ρ = 0.0899 kg/m³, √(3P/ρ) = 1,839 m/s [calc]. He calls it the "mean velocity"; it is the rms speed (the true mean is 1,694 m/s) [calc]. That is a label slip, not an error in the number.

3.2 Which gases can a body hold?

v_esc = √(2GM/R). Earth 11.2 km/s, Sun 618 km/s with modern values [memory, high]. A gas whose atoms move well above a body's escape speed escapes from it, and gravity raises its density near the body only by the Boltzmann factor e^(GMm/RkT).

3.3 What can a gas carry?

Only longitudinal waves, at the sound speed √(γkT/m), a little below the atoms' own rms speed. No transverse waves. This was the core of the 1800s ether problem: a medium for light had to be stiffer than steel to shearing, yet offer no resistance to the planets [memory, high].

3.4 What is the modern answer about the ether?

There is no medium in the 1902 sense. Ether-drift tests from Michelson-Morley (1887) onward found no preferred rest frame; a modern rotating-cavity test limits any direction-dependence of the speed of light to about 1 part in 10¹⁷ [web: https://arxiv.org/pdf/1002.1284, Herrmann et al. 2009]. Modern physics does give empty space physical properties (Einstein's 1920 Leiden lecture on "ether and relativity", Dirac's 1951 "Is there an aether?", the quantum vacuum and the Higgs field), but none of these is a gas of atoms with a mass and a temperature [memory, high; Lente 2019 cites Einstein and Dirac, via the Investigator].

3.5 The noble gases and the start of the table

Argon (1894), helium on Earth (1895), and neon, krypton and xenon (1898) were found by Ramsay, Rayleigh and Travers [read: p.21; memory, high for dates]. Errera (Belgian Academy, 5 March 1900) and Ramsay (19 March 1900) proposed a separate group for them between the halogens and the alkali metals; Mendeleev credits both [read: 1905 Russian scan, footnotes 13-14, via the Contrarian and Investigator]. Radon was added later; the column is today's Group 18. The table starts at Z = 1. The neutron (Chadwick 1932) has zero charge but weighs about as much as a hydrogen atom; the electron (5.5 × 10⁻⁴ of the atomic mass unit u) and the neutrino (below about 0.45 eV of mass-energy, about 5 × 10⁻¹⁰ u) are lighter than hydrogen, but they are not atoms and do not form a chemical element [web: https://www.kit.edu/kit/english/pi_2025_029_astroparticle-physics-neutrinos-weigh-less-than-0-45-electronvolts.php for the neutrino; memory, high for the rest].

---

4. Where Mendeleev departs

First, what is right in his starting points. Mendeleev was not an outsider. He was the leading chemist of his country and an experienced experimenter on gases: in 1874, with M. L. Kirpichev (spelled "Kirpitchnikoff" in the English), he measured departures from Boyle's law at low pressure, which he says Ramsay confirmed, and he says this work was aimed at the ether question [read: p.11; Russian text via Investigator]. That work was part of a long, funded programme. Stewart (2019) writes that from 1872 Mendeleev "threw himself into research into the gas laws, under a new Commission for Research in the Elasticity of Gases. His ambition was to establish the nature of the 'luminiferous aether'", and that the work went on "with diminishing intensity until he resigned in January 1881" [read: Stewart J p.6 / PDF p.4; "J" is the journal page]. So the 1902 essay returns to a question he had worked on for about ten years. Stewart adds that this interest in the ether "led naturally into concern with spiritualism": 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 on such matters". Mendeleev took the sceptic's side, and he left after about a year when it became clear that some members were themselves spiritualists [read: Stewart J p.6, citing Gordin 2004, pp. 81-111]. The link between the ether programme and the spiritualism commission is Stewart's characterization; he gives no Mendeleev document that makes it. In the essay itself Mendeleev says that calling the ether a gas makes "all mystical, spiritual ideas about ether disappear", and that it "does not require the recognition of a peculiar fourth state beyond the human understanding (Crookes)" [read: p.22]. His standing at the time was high: Stewart says he was nominated for the 1906 Nobel Prize in Chemistry, which went to Henri Moissan "by 5 votes to 4" [read: Stewart J p.8 / PDF p.6; Stewart gives no source for the vote, and the Nobel records were not checked]. His insistence that anything real must have weight is the Newtonian principle every physicist then accepted [read: p.10]. His kinetic theory is the standard one, and every number in his calculation follows from his formulas (one misprint aside) [calc; all three reviewers agree]. His periodic method had a famous record of success by interpolation. And he was right that the inert gases needed a group of their own, against his own earlier view (1895) that argon was a triple nitrogen molecule N₃, an error he admits in print [read: pp.27-28].

The departures, side by side:

The exact point where the argument breaks is the third row. Mendeleev chose a gas because only a gas could pass through solid bodies like hydrogen through palladium [read: pp.19-21]. But he still needed the ether to carry light, and the only waves a gas carries are slower than its atoms and longitudinal. His own "130 times" sentence states the gap without closing it. He says himself that it is "very difficult to form any opinion on this still dim province" [read: p.57]. The essay never mentions Michelson, Maxwell, Hertz or Lorentz [read: text search by the Contrarian].

Two further departures concern the table. His estimates of y and x come from extrapolating ratios of atomic weights in neighbouring rows, below the first row of the table. For y he takes the series of ratios Cl:F 1.86, S:O 2.00, P:N 2.21, Si:C 2.37, Al:B 2.45, Mg:Be 2.67, Na:Li 3.28, Ne:He 4.98, and the one known ratio of the next level, Li:H 6.97, and concludes He:y must be "at least 10", so y ≤ 0.4 [read: pp.35-36]. For x he takes Xe:Kr 1.56, Kr:Ar 2.15, Ar:He 9.5 and gets He:x = 23.6, so x ≤ 0.17 [read: p.42]. The 23.6 is exactly the next value of the quadratic through those three points (the German translation says "from a second-order parabola") [calc; web: Lente 2019, via the Investigator]. Mendeleev knew this was different from his 1871 successes: then he interpolated and "risked nothing, now I do" [read: pp.24-26]. And he had "a vague notion that hydrogen might be preceded by some elements of less atomic weights" in 1869, but "feared to injure the first impression of the periodic law" [read: p.33].

He was not the only one thinking along these lines. Two contemporaries entertained related ideas:

• Werner, 1905. Alfred Werner (1866-1919; Nobel Prize in Chemistry 1913, per Stewart) published a periodic table in 1905 that, Stewart writes, "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: Stewart J p.8 / PDF p.6]. The figure Stewart reproduces shows four dotted spaces around hydrogen, but split 3 + 1 (three on the top row, one before He), not Stewart's 2 + 2 [read: page image Stewart PDF p.6, deep read]. Werner's spaces are blanks with no atomic weights and no physical argument attached.
• Dewar, 1902. Mendeleev writes that while he was making his calculations, "my friend Professor Dewar" sent him his presidential address to the Belfast meeting of the British Association, in which Dewar "expresses the thought that the highest regions of the atmosphere, which are the seat of the aurora borealis, must be considered to be the province of hydrogen and of the argon analogues". Mendeleev adds: "This is 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: p.51]. The first sentence reports Dewar; the second is Mendeleev's own gloss. Dewar's address itself was not obtained, so what Dewar said in his own words has not been checked.

So in 1902-05 elements lighter than hydrogen were not unthinkable to leading chemists. The arbiter weighs this as context: it makes the zero series less of an outlier, but it does not change the verdicts, because the wave-speed problem (above) and the floor Z = 1 (section 3.5) are untouched by it.

Mendeleev's own reason for rejecting a "unity of matter" (claim 5). Many in 1902 hoped that all elements were built from one primary substance (hydrogen, or electrons). Mendeleev rejected this, and his stated reason was philosophical: "when I am told that the doctrine of unity in the material of which the elements are built up responds to an aspiration for unity in all things, I can only reply that at the root of all things a distinction must be made between matter, force, and mind; that it is simpler to admit the germs of individuality in the material elements than elsewhere". He concludes: "I see no object in following the doctrine of the unity of matter, while I clearly see the necessity of recognising the unity of the substance of the ether" [read: p.40]. The Russian and German texts open with the same matter, force and spirit "trinity" [web: via Investigator]. This is the framing behind his statement that atoms are "individuals" and electrons "metachemical" (table above, row "What atoms are").

---

5. Worked example: follow Mendeleev's own numbers, then ask one question of them

All numbers below are reproduced by example.py (Parts A to G). Try each step on a calculator first.

Step 1. His speed formula

At fixed temperature, v ∝ 1/√(density). A monatomic gas of atomic weight x has density x/2 relative to hydrogen gas (whose molecule, H₂, weighs 2). Temperature enters through the gas-law factor (1 + at), with t in °C and a = 1/273 = 0.00367. So Mendeleev writes [read: p.43]:

v = 1843 √(2(1 + at)/x) m/s (formula I)

At t = −80 °C: 1 + at = 1 − 80 × 0.00367 = 0.706, so v = 1843 × √(2 × 0.706) / √x = 2191/√x, and squaring, x = 4,800,000/v² (formula II, with v in m/s) [read: p.44; calc]. The English text prints "a = 0.000367"; that value would give 2568/√x, not 2191, so it is a misprint. The Russian footnote 24 gives 0.00367 [Russian text via Investigator; calc].

Cross-check with modern constants: an atom of mass 1.008 u at 193 K has √(3kT/m) = 2,186 m/s [calc]. His constant is right.

Step 2. His escape speeds, and the bounds on x

His criterion: the ether must escape the gravity of every body, "as otherwise it would accumulate about the largest mass and not fill all space" [read: p.42].

• Earth: he gets GM from g R² with g = 9.807 m/s² and R = 6,373,000 m, so GM = 3.983 × 10¹⁴ m³/s² (his "398.10¹²"). v_esc = √(2gR) = 11,180 m/s (he says 11,190). Formula II: x < 4.8 × 10⁶ / 11,180² = 0.038 [read: pp.45-46; calc]. Hydrogen and helium stay in the atmosphere; an ether must be at least 26 times lighter than hydrogen.
• Sun: 325,000 Earth masses, 109.5 Earth radii. v_esc scales as √(M/R): 11,180 × √(325,000/109.5) = 609,000 m/s (he says 608,300). x ≤ 1.3 × 10⁻⁵ [read: p.46; calc].
• A star of 50 solar masses with the Sun's mean density: then R ∝ M^(1/3), so v_esc ∝ M^(1/3): 609,000 × 50^(1/3) = 2.24 × 10⁶ m/s. x ≤ 9.5 × 10⁻⁷ [read: pp.48-50; calc].
• A body that could hold particles moving at the speed of light would need (3 × 10⁸ / 608,000)³ = 493³ ≈ 1.2 × 10⁸ solar masses. Particles at c give x = 4.8 × 10⁶ / (3 × 10⁸)² = 5.3 × 10⁻¹¹ [read: p.50; calc].

So by his own method x lies between 5.3 × 10⁻¹¹ and 9.5 × 10⁻⁷, a range of about 18,000 times [calc].

Step 3. His choice

He rejects the lower end "because it would in some measure answer to a revival of the emission theory of light" and adopts "an atomic weight nearly one millionth that of hydrogen" and "a velocity of about 2,250 kilometres per second" [read: p.51]. Check: formula II gives 2,191 km/s at x = 10⁻⁶, and x = 9.5 × 10⁻⁷ at 2,250 km/s [calc]. So his answer is the upper edge of his allowed range (the 50-Sun bound), chosen by a judgement about theories of light, not by calculation. The arithmetic up to here is correct; all three reviewers agree.

Step 4. The question: can this gas carry light?

Use the same kinetic theory. One ether atom has mass m = 10⁻⁶ × 1.008 × 1.66054 × 10⁻²⁷ kg = 1.67 × 10⁻³³ kg (about 1/550 of an electron's mass) [calc]. At T = 193 K:

• rms speed √(3kT/m) = 2.19 × 10⁶ m/s
• sound speed √(5kT/3m) = 1.63 × 10⁶ m/s (0.745 × rms) [calc]

Light travels at 3.00 × 10⁸ m/s. That is 133 times the atoms' speed (Mendeleev's "130 times", p.57) and 184 times the fastest wave this gas can carry [calc]. And that fastest wave is longitudinal, while light is transverse.

Run the question backwards: what x would make a sound wave in a 193 K gas travel at c? x = γkT/(c² m_H) = 3 × 10⁻¹¹ [calc]. That is the same order as the lower limit Mendeleev rejected because it would make the ether an "emission theory". So the model needs two incompatible things: atoms slow enough (x ≈ 10⁻⁶) to avoid the emission theory, and a medium fast enough (x ≈ 10⁻¹¹) to carry light. Even then the waves would be longitudinal. This is why the arbiter's verdict on the strong form of the claim is "Refuted by calculation": the deciding law, the wave speed in a gas, is the same kinetic theory he uses in formula I.

Step 5. Two more checks using only his own premises

Can the Sun hold the gas he says it gathers? His adopted speed is 2,250 / 609 = 3.7 times the Sun's escape speed [calc]. By the criterion of Step 2 the Sun cannot hold it, yet on p.54 he writes that the ether "must naturally accumulate from all parts of the universe towards the medium or mass of the sun" [read: p.54], after writing on p.42 that it must not "accumulate about the largest mass". How much would the Sun's gravity concentrate it? The Boltzmann factor at the solar surface is e^(GMm/RkT) = e^(0.12) = 1.13, a 13 per cent rise [calc, modern solar values]. Accumulation would need some non-gravitational attraction, which he describes as "an embryo of such a motion" short of combination [read: p.54], with no quantity attached.

What if the gas takes the temperature of what it passes through? He says the ether is in "rapid interchange" with matter [read: p.58] and permeates all bodies. For a fixed x = 10⁻⁶, v ∝ √T [calc]:

In a hot star an ether that exchanged energy with matter would move faster than light, the very objection he raised against the lower end of his range. The core temperature was not known in 1902, so this is not a charge that he ignored a known number; it shows that x ≈ 10⁻⁶ holds only for a gas that never shares heat with matter. And the temperature of space is now known to be 2.7 K, which at fixed speed would make x about 70 times smaller (1.3 × 10⁻⁸) [calc]. The number moves by orders of magnitude as soon as one assumption changes. This is why the arbiter calls the claim "x ≈ 10⁻⁶, v ≈ 2,250 km/s as properties of a real ether" Internally inconsistent and, in practice, Not testable as stated.

Step 6. Two side checks (Parts F and G of example.py)

Weight. Mendeleev quotes Lord Kelvin: a cubic metre of ether weighs "about and not less than 0.000,000,000,000,000,1 grm", i.e. 10⁻¹⁶ g = 10⁻¹⁹ kg [read: p.10, page image]. The total gravitating matter near the Sun, measured from the motions of stars (the Oort limit), is about 0.1 solar mass per cubic parsec = 6.8 × 10⁻²¹ kg/m³ [web: https://arxiv.org/pdf/astro-ph/0401341, via the arbiter; calc]. Kelvin's floor is 15 times all the matter the stars feel. If the ether gravitates like ordinary matter, that density is excluded. The arbiter rates this medium confidence: Kelvin's own derivation was not found, and Mendeleev himself notes that hydrogen or steam at tiny pressure would be as thin [read: p.10].

The extrapolation. Through Xe:Kr 1.56, Kr:Ar 2.15, Ar:He 9.5, the quadratic gives the next ratio 23.61 (x = 0.169); a straight line through the last two gives 16.9 (x = 0.24); repeating the last growth factor gives 42 (x = 0.095); a quadratic in logarithms gives 135 (x = 0.030) [calc; the last from the mathematician]. Three points fix a quadratic exactly, so the fit has nothing left over to test it. And the final x ≈ 10⁻⁶ does not use this bound at all.

What the example teaches. Mendeleev's arithmetic is sound. What fails is the job the gas was given: a gas cannot carry a wave faster than its own atoms, and it cannot carry a transverse wave at all. That conclusion needs nothing from relativity, and nothing that was unknown in 1902.

---

6. Evidence table

Verdicts are the arbiter's final verdicts (arbiter.md section 9).

Notes kept in the record: - Stewart's abstract says "eleven other predictions"; his body names ten (six in his 1871 quotation, four "scattered through Mendeleev's works"), and the eleventh is not named. His own final tally is "eight successful predictions, one multiple error and two irrelevancies" (the irrelevancies are the ether and coronium) [read: Stewart J pp.3-9 / PDF pp.1-7, deep read]. His "six more blanks between hydrogen and helium" is not in the 1904 table, which shows only x and y; his own Fig. 2, a reproduction of that table, shows the same [read: page image p.34; page image Stewart PDF p.5]. His sentence that the green coronal line was "eventually identified as due to a highly ionised form of iron and named 'coronium' by Gruenwald" cites note 12, an 1887 Chemical News paper on water-vapour spectra whose title does not mention the corona; the iron identification is Grotrian and Edlén's (1939-42) [read: Stewart J p.8 / PDF p.6, note 12; web: https://en.wikipedia.org/wiki/Coronium]. Smaller slips in Stewart, kept so you can weigh him: eka-niobium is 146 in one place and 145 in another; germanium is dated 1886 and 1885; and he says gold "was missing" from the 1904 table, while his Fig. 2 prints Au = 197.2 [read: Stewart J pp.3-7, deep read]. - The Russian text adds: the name "newtonium" for x "in honour of the immortal Newton"; that the ether may be a mixture of several very light gases (this does not help with the wave problem: a mixture of gases is still a gas); a possible light halogen between H and He; and a question whether ether piling up around the Sun could supply its energy. The mathematician's energy budget answers that last: the ether in the Sun's volume at Kelvin's density would supply 7 × 10⁻⁸ s of sunlight; powering the Sun for 10⁹ years would need about 30 solar masses of ether [Russian text via Investigator; calc, mathematician, checked by arbiter]. - Mendeleev's credit (from the Contrarian, kept by the arbiter): something material does flow from radium, and it is a zero-group gas as his table would predict, though a heavy one, not a light one. - A caution from the essay itself: Mendeleev notes that Stoney (1894-98) and Rogovsky (1899; "Rostovsky" in the English) argued that hydrogen and helium move fast enough to leave the Earth's atmosphere [read: p.37], and in a footnote of the Russian text that both were "very soon" found in it [read: 1905 scan p.24, footnote 11, via the Contrarian]. A sharp cut-off at the escape speed is too crude: what a planet keeps depends on the fast tail of the Maxwell distribution and on time (section 2, "Escape speed"). The same crude criterion sets the bounds on x in section 5, Step 2. - Related later proposals (leads, not support). The idea of an element below hydrogen recurred after Mendeleev. The Investigator found these; the arbiter lists them as the same family of ideas, not as evidence for the ether: - Nebulium (1898): unexplained emission lines in nebulae, attributed in the source to Margaret Huggins, given an assumed atomic weight of about 2.74 [web: https://habr.com/ru/articles/653695/, via research/2026-10-06-habr.md]. The lines were later explained as known elements in ionised states [memory, high: Bowen 1927, doubly ionised oxygen]. - Protofluorine: a very light halogen above fluorine, which matches Mendeleev's own Russian footnote on a light halogen between H and He [web: same Habr article]. - Neutronium (1926): Andreas von Antropoff's "element zero", described as an element of about 0.1 of hydrogen's weight that hardly combines and penetrates everything [web: same Habr article, a 2022 popular article by O. Sivchenko]. A paper titled "100 Years of Element Zero" on Antropoff appeared in Zeitschrift für anorganische und allgemeine Chemie (ZAAC) in June 2026; it is known by title only (not read) [web: via research/2026-10-06-google.md]. - Caution on the Habr source: it swaps Mendeleev's labels (it calls x coronium and y newtonium; the Russian text has x = newtonium), and its comparison of newtonium with the neutron and neutronium with the neutrino is the author's own interpretation.

Where the record goes dark: what happened to x and y?

The 1904 English table prints x (row 0) and y (row 1, beside hydrogen) in the zero group, without atomic weights [read: page image p.34]; the revised 1905 Russian edition prints the same rows [read: page image p.25 of the 1905 scan, by the Contrarian]. So ru.wikipedia's statement that x and y were absent from the essay's tables is contradicted by both primary scans. What the 1906 textbook table showed, and when x and y left the table, is unknown: those editions were not obtained. The competing explanations, as the arbiter sets them out:

• (a) Abandoned as physics moved on. Mendeleev died in February 1907 [read: Stewart], before Moseley and Bohr (1913) gave a physical reason for the table to start at hydrogen; relativity (1905) removed the need for a medium. He had labelled the idea a hypothesis. Lente and ru.wikipedia take this view. No document has been seen that records anyone deciding it.
• (b) Deliberate removal or falsification, as stated by two modern advocates of an ether: V. A. Atsyukovsky (the zero row "was later removed") and V. G. Rodionov (a "genuine 1906 table" with a zero row). These are their statements; neither text was obtained [web: via research/2026-10-06-google.md]. Against suppression of the essay itself: it was printed at least five times in his lifetime (Russian 1903 and 1905, German 1903 and 1904, English 1904 and 1905).
• (c) Mendeleev himself reworked the rows in 1905-06. He did revise the 1905 text (radium 224 became 225, footnotes added); nothing found about the 1906 table.
• (d) Editorial removal in posthumous editions (1927 and later). No evidence either way.

Records that would decide it: the 1906 8th and 1927 9th editions of Osnovy khimii, Kedrov's 1958 commentary, the 1905 brochure in the Polytechnic Museum (Moscow), and the Mendeleev Museum-Archive in St Petersburg. Until they are read, (a) needs the fewest extra assumptions, but (b), (c) and (d) are open, not refuted.

Contemporary reception is thin in the files: no point-by-point refutation and no experiment aimed at the gas were found; a 1904 Nature review and a Scientific American article are known by title only, and Bensaude-Vincent's 1982 paper calls it "un essai malheureux" (an unfortunate attempt) by its title [web: via arbiter and Investigator; texts not obtained]. No record of blocked funding, blocked publication or omitted data was found: the pattern is "widely printed, then left alone".

---

7. What a learner should take away

Solid. - The kinetic theory in the essay is correct, and every number follows from it. Lighter means faster; escape speed decides what a planet keeps. You can redo every step with a calculator. - The zero group of noble gases was right. Mendeleev accepted it against his own earlier view of argon, and credited Errera and Ramsay. - His instinct that radioactivity involves "something material" flowing was right. It was radon, a gas of the very group he added. - His earlier interpolations (gallium, scandium, germanium and five later elements within a few per cent) remain among the best predictions in the history of science.

Wrong, and why it is instructive. - A medium cannot carry a wave faster than its own particles, and a gas cannot carry a transverse wave at all. Mendeleev's ether has atoms at 2,250 km/s and has to carry light at 300,000 km/s, transversely. His own "130 times" sentence states the gap. This is the same obstacle that defeated every material ether of the 1800s; making the ether a chemical element does not remove it. - A method that works by interpolation can fail by extrapolation. Between known neighbours a pattern is pinned on both sides. Below the first row, a curve through three points can give almost any answer, and three points fix a quadratic exactly, so the fit tests nothing. - A bound is not a value. The essay's own method allows x anywhere across a factor of 18,000; "one millionth" is the edge of that range, picked to avoid the emission theory of light. And the number shifts by orders of magnitude with the assumed temperature. - Check a hypothesis against all of its own premises at once. The gas must escape every star (p.42) and must accumulate on the Sun (p.54); it must have "its own temperature" and exchange freely with matter. Each premise is plausible alone; together they conflict. - Definitions cannot settle facts. "Atoms are individuals, so electrons cannot be parts of them" assumes what it sets out to show. The electron (1897) and transmutation (1902-03) decided it.

Genuinely open. - Historical: when and by whom x and y were dropped from the table (section 6). The 1906 and 1927 editions decide it. - Historical: Kelvin's derivation of 10⁻¹⁶ g/m³, Yarkovsky's 1901 estimate of the ether's density (cited by Mendeleev, not located), and how Nature, the German press and Russian chemists received the essay in 1903-05. - Physical, weak form only: nothing rules out some very light, weakly interacting particle population in space (neutrinos are one), but such a thing would not be Mendeleev's ether, because his ether's purpose was to carry light.

---

8. Check yourself

1. Mendeleev's formula II is x = 4,800,000/v². What escape speed would a white dwarf like Sirius B (mass 1.018 solar masses, radius 5,635 km) have, and what bound on x would his own criterion then give?
2. Why can a gas carry sound but not a transverse wave? What property of a solid makes the difference?
3. If the temperature of space is 2.7 K instead of 193 K, and the ether atoms must still move at 2,250 km/s, by what factor does x change, and in which direction?
4. Mendeleev predicted germanium's atomic weight within 1 per cent, but his estimate x ≤ 0.17 can be argued to be anything from 0.03 to 0.24. What is the methodological difference between the two predictions?
5. In the Curie two-flask demonstration, opening a tube between a radium solution and zinc sulphide made the sulphide glow. Mendeleev saw "an emissive flow of something material". In what sense was he right, and what was the material?

---

9. Further reading

The inventor's own work - D. Mendeleeff, An Attempt Towards a Chemical Conception of the Ether, tr. G. Kamensky (Longmans, Green, 1904). Free at the Internet Archive: https://archive.org/details/cu31924012371096 [read: p.4]. - Mendeleev, Popytka khimicheskogo ponimaniya mirovogo efira (Russian; 1903 journal text on ru.wikisource; revised St Petersburg edition 1905, scan at ether-wind.narod.ru/Mendeleev_1905/). Contains the "newtonium" name and the footnotes missing from the English [web: via Investigator]. - Mendeleev, The Principles of Chemistry, 3rd English edition (1905), vol. 2, Appendix III (the ether essay as a textbook appendix) [web: archive.org, via Investigator]. - German translations: Prometheus XV (1903), Nr 735, 737, 738, tr. S. Tschulok; Naturwissenschaftliche Rundschau 19 (1904) [web: archive.org, via Investigator].

Modern assessments of the essay - Philip J. Stewart, "Mendeleev's predictions: success and failure", Foundations of Chemistry 21 (2019) 3-9, https://doi.org/10.1007/s10698-018-9312-0 (in the project folder). - Gábor Lente, "Where Mendeleev was wrong: predicted elements that have never been found", ChemTexts 5:17 (2019), open access, https://link.springer.com/article/10.1007/s40828-019-0092-5 (reconstructs the x and y extrapolations) [web: via Investigator]. - Eric Scerri, The Periodic Table: Its Story and Its Significance (Oxford University Press): the history of the table, the noble gases and the predictions.

Related later proposals (leads; not read in the review) - A. von Antropoff's 1926 "element zero" (neutronium), and the June 2026 ZAAC paper "100 Years of Element Zero" (title only) [web: via Investigator]. - O. Sivchenko, "Из чего состоит мировой эфир. Последняя теория Менделеева" (What the world ether is made of: Mendeleev's last theory), Habr, 2022, https://habr.com/ru/articles/653695/ (popular; mislabels x and y) [web: via Investigator]. - A. Werner, "Beitrag zum Aufbau des periodischen Systems", Berichte der deutschen chemischen Gesellschaft 38 (1905) 914-921 (the 1905 table with spaces around hydrogen; known here through Stewart) [read: Stewart note 13].

Mainstream physics - Schroeder, An Introduction to Thermal Physics: kinetic theory, the Boltzmann factor, the Maxwell distribution. - Halliday, Resnick and Walker, Fundamentals of Physics: kinetic theory, waves, sound. - Griffiths, Introduction to Electrodynamics, chapter 9: light as a transverse electromagnetic wave. - E. T. Whittaker, A History of the Theories of Aether and Electricity: the elastic-solid ether and why it failed. - A. Einstein, "Ether and the Theory of Relativity" (Leiden lecture, 1920); P. A. M. Dirac, "Is there an aether?", Nature 168 (1951): what a modern "ether" can and cannot mean [memory, high for these citations].

---

Answers

1. v_esc = √(2GM/R) = √(2 × 6.674 × 10⁻¹¹ × 1.018 × 1.989 × 10³⁰ / 5.635 × 10⁶) ≈ 6.9 × 10⁶ m/s (6,900 km/s), three times his adopted ether speed. Then x ≤ 4.8 × 10⁶ / (6.9 × 10⁶)² ≈ 1.0 × 10⁻⁷, ten times below his "one millionth" [calc, as in the arbiter's table; Sirius B data: https://en.wikipedia.org/wiki/Sirius_B]. Sirius B's density was not known in 1902, so this is supersession, not a mistake by 1902 standards; but it shows the escape argument has no stopping point short of c.
2. A wave travels because displaced bits of the medium push or pull their neighbours back. A gas resists being compressed (pressure rises), so a compression wave travels. It does not resist being sheared: slide one layer of gas sideways and nothing pulls it back except viscosity, which damps the motion instead of restoring it (the linearised equation is a diffusion equation, not a wave equation). A solid has a shear modulus, a restoring force against sideways deformation, so it carries transverse waves at √(G_s/ρ).
3. From v_rms = √(3kT/m), at fixed v the mass m (and so x) is proportional to T. 2.7/193 ≈ 1/70, so x falls about 70 times, to about 1.3 × 10⁻⁸. Equivalently, an x = 10⁻⁶ atom at 2.7 K moves at only 260 km/s.
4. Germanium was interpolated: the gap sat between known elements on every side (silicon above, tin below, gallium and arsenic beside), so many properties were pinned by neighbours. The ether was extrapolated below the first row, from three ratios of a trend that jumps sharply at helium (2.15 to 9.5). Three points fix a quadratic exactly; other equally reasonable rules give other answers. Mendeleev said so himself: in 1871 he "risked nothing, now I do".
5. Something material did flow: radium emanation, a radioactive gas (now radon, Z = 86) that diffuses through the tube; its alpha particles make zinc sulphide scintillate (Crookes's spinthariscope, 1903). It is a zero-group gas, as his table would place it, but the heaviest of them, not a light ether, and it decays, which is transmutation, the thing he denied. (The identification of this exact flask with radon is [memory, medium]; Mendeleev does not name the emanation.)