Arbiter: Crookes, "On Radiant Matter" (Sheffield, 22 August 1879)

Role: Arbiter (Step 3). I read facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, sites.md and the research notes. Tags: [read] = the project's document or review files; [web] = URL given; [memory] = recall with confidence; [calc] = my own computation (numpy, project venv). Final weighing stays with the human reader; nothing the reviewers or the Investigator found has been dropped.

Source gap note. There is no source gap for the lecture itself: one document, 23 pages, read in full by the deep reader with page images for the three OCR errors [read]. But the project holds only the lecture. The following rest on secondary or unread sources: (1) Crookes's technical papers (Bakerian Lecture 1879, "Contributions to molecular physics in high vacua"), which should hold the pressures, apparatus and the "laws of magnetic rotation" the lecture lacks; (2) Stokes's referee reports RR/8/151 and RR/8/152 (page images online, unread by every role; my fetch of RR/8/152 failed again with a certificate error); (3) Starke 1900 and Thomson 1903 on the paddle wheel, known only through a Harvard demonstration page and Wikipedia (my fetch of the Harvard page returned HTTP 403, so I rely on the Investigator's note of it); (4) the later tests of the charge and mass (Schuster, Perrin, Thomson 1897), known through Wikipedia and Busch 2023; (5) Crookes's later "radiant matter now passes as electrons", known from an obituary quotation, original place and date untraced. Where a verdict below leans on these, it is said.

Corrected claim list

Step 1 is accurate. Corrections and refinements I adopt from the reviews:

1. Claim 1 (pressure). The "one millionth of an atmosphere" is the best-phosphorescence regime, not a single threshold. Mechanical effects start near 0.5 mm Hg (about 660 millionths atm) and the dark space reaches the glass there; the highest vacuum is 1/20,000,000 atm [read: deep read; contrarian 1.1]. The regime is defined by mean free path comparable to the vessel.
2. Claim 4 bundles three things: (i) the stream is negatively electrified, (ii) it is made of gas molecules, (iii) impact causes phosphorescence. They have different verdicts and are split below.
3. Claim 6 bundles two things: gas independence of the effects, and the molecule count. Split below.
4. Printed values. Page-image values stand: ruby line 6,895; highest vacuum 1/20,000,000 atm; 10^24 and 10^18 molecules [read: page images via deep read]. The text layer still shows "6,595" and "2 millionth"; those are OCR slips.
5. Self-consistency check of 1/20,000,000. 0.01 inch / 3 miles = 5.26 x 10^-8 = 1/19.0 million [calc]. Two reviewers got this independently; it supports the page-image reading.
6. Claim 4 is split a further time (QE revision). Fig. 18 gives an observation (the first ray moves away from the second cathode's ray when the second cathode is switched on) and Crookes's interpretation (beam-on-beam repulsion of charged molecules). They no longer share one verdict; see "QE revision".
7. Claim 7 added (QE revision). At the highest vacua the tube stops conducting; a trace of vapour restores conduction (pp.7-8; Morgan 1785 footnote). The original arbiter pass did not treat it.
8. Claim 5, hot-ring mechanism, re-rated (QE revision). The "Never properly tested" mechanism verdict is replaced by a split; see "QE revision".
9. Pressure table (p.7 footnote), now read from the page image: 1 millionth atm = 0.00076 mm Hg; 1 mm Hg = 1315.789 millionths; internally consistent [read: page image p.7, via deep read].

Where the reviewers' numbers differ, recomputed

Web checks (budget 8; 7 used)

• Ruby R lines: R1 694.3 nm, R2 692.8 to 692.9 nm. Confirmed [web: https://fluomin.org/uk/fiche.php?id=870 and https://duffy.princeton.edu/laboratory/ruby-fluorescence-measurement, from search summaries]. Retag from [memory, high/medium] to [web]. Crookes's 6,895 (if Angstroms) is about 0.5 to 0.7 % low, which fits a hand spectroscope reading.
• Loschmidt number: 2.65 x 10^19 per cm^3 (about 2.69 x 10^19 at 0 C) [web: https://en.wikipedia.org/wiki/Johann_Josef_Loschmidt, search summary]. Confirmed. Loschmidt's 1865 estimate was the first of its kind, and Stoney published his own estimate in 1868 [web: same search]. The numerical value of Stoney's 1868 figure (the lecture's 10^21) was not retrievable on the web; I take it from the lecture footnote [read].
• Cathode-fall pressure-length product: check failed to return a number (sources found, tables not legible: https://arxiv.org/pdf/physics/0409047, a von Engel and Steenbeck normal-fall model; text only said that the thickness is set by a constant dependent on cathode and gas, in cm Torr, and that "with a pd value in the order of 1 Torr cm" a normal phase does not exist). The 0.2 to 0.5 Torr cm figure stays [memory, medium] and I do not let a verdict depend on its exact value.
• Plasma credit, Hertz null result, Schuster, Perrin and Thomson e/m: taken from the Investigator's Wikipedia "Cathode ray" and Busch 2023 notes; I did not re-fetch.
• Starke 1900 and Thomson 1903: Harvard page fetch failed (HTTP 403). The Investigator's note is the evidence; unchanged. Retag stays [web, via research note; primary texts unread].
• Crookes's "What I then called 'Radiant Matter' now passes as 'Electrons'": search did not find the quotation in any source I could open (I found general statements that Crookes used "radiant matter" in 1879 and that Thomson identified electrons about twenty years later; https://rinconeducativo.org/en/?p=21630). The Dyer 1920 obituary is the only cited source. Check partly failed; carried as one documented obituary quotation, no second source found yet, and not a reason to discard it.
• Stokes RR/8/152: failed (certificate error, same as two reviewers). Open.
• Plucker and the hinged cross: not re-checked. Plucker died in 1868 [memory, high], which fits neither a 1869 date nor the lecture's credit to Gimingham's workmanship [read: p.12]; carried as an unresolved attribution that does not affect any verdict.

Corrected claim list: final verdicts

All three reviewers use the same split structure, and I agree with the split. The recurring shape: the observation is Established, the explanation (carriers are residual gas molecules) was superseded. Per-claim details follow.

Claim 1. At about 10^-6 atm the residual gas is a "fourth state of matter" (Faraday's radiant matter)

• Agree: all three. The collisionless (free-molecular, Knudsen) regime is real and Crookes's quantitative description of it is right: lambda = 6.6 cm at 10^-6 atm against a 13.5 cm bulb [calc]. "State" is a label with no phase boundary; rarefied gas changes continuously with lambda proportional to 1/p. The stream he studied was not the residual gas but electrons from the cathode (claim 3 and 4).
• Disagree: only on emphasis. The contrarian credits the plasma reading as "not baseless" because the discharge in his tubes is partly ionised; the physicist says the modern plasma (free charges, collective behaviour, named by Langmuir 1928) is a different defining property and the beam is not the plasma.
• Stronger argument: the physicist's and mathematician's. Crookes's own definition is a collisionless neutral gas [read: p.2-4]; the Knudsen number (lambda/L greater than 1) and the Debye length with ionisation fraction are different parameters. The modern plasma attribution is retrospective reception, kept in the record.
• Verdict, split: weak form (a collisionless regime exists below the pressure where lambda exceeds the vessel, with new phenomena): Established. Strong form (a distinct fourth state composed of the residual gas molecules): Historically superseded but instructive. Plasma credit: later reception, not Crookes's definition. Crookes himself later reinterpreted the stream as electrons (obituary quotation, untraced).
• What would settle it: nothing experimental remains. The residual question is historical: the original place and date of Crookes's "now passes as Electrons" statement (his presidential addresses, 1891 and 1898, are the likely places [memory, low]; to be searched).
• Keep in the record: Theosophy (Blavatsky) took the lecture as support for finer states of matter; the closing "border-land where matter and force seem to merge" [read: p.23] has no measurable content. Puluj (1883) called the fourth state "transcendental" with an inclination towards spiritualism, as Busch reports; Puluj had a rival theory (charged electrode particles) and a priority stake.

Claim 2. The dark space round the negative pole is the mean free path of the residual molecules

• Agree: the scaling (the dark space grows as the vacuum improves, roughly as 1/p) is correct and a good instinct. The identification is not: the dark space is the cathode fall region, set by the distance electrons need to gain ionising energy in a strong field, so it depends on voltage and gas as well as pressure.
• Disagree: the mathematician gives "Refuted by calculation" (about 100 times off, on a [memory, medium] cathode-fall product that failed its web check); the physicist and contrarian say "Historically superseded but instructive" and decline a number because Crookes gave no pressure for his one-inch dark space [read: p.3].
• Stronger argument: the physicist's. The dark-space number is not stated in the lecture, so there is nothing quantitative of Crookes's to refute; the mathematician's calculation compares different species (electrons in a field against neutral molecules) on a remembered constant that I could not verify. On the remembered constant the ratio is 46 to 100 free paths at 0.5 Torr [calc], and on Crookes's own "about an inch" at an unstated "good exhaustion" the identification can be neither confirmed nor refuted from the lecture. The contrarian's "centimetres" at 0.5 mm Hg is Crookes's own report of the dark space reaching the glass [read: p.15], which, if correct, is larger than a normal-glow estimate and would fit an abnormal glow at high coil voltage; this is a side point.
• Verdict, split: scaling observation: Established. Identification of the dark space with the molecular mean free path: Historically superseded but instructive (not "Refuted by calculation", because the calculation needs a constant that I could not confirm on the web and Crookes gave no pressure).
• What would settle it: measured dark-space thickness against pressure and voltage in the same tube. The Bakerian Lecture and "Contributions" papers (Phil. Trans. 1879, on archive.org per sites.md) may contain such pairs; or look up the tabulated normal cathode fall (cm Torr) in a standard gas-discharge text (Raizer, or von Engel) and compare with the mean free path.

Claim 3. Straight lines normal to the surface, independent of the anode, sharp shadows, magnetic deflection like projectiles

• Agree: all three. The observations are textbook cathode-ray physics. Priority is shared: Hittorf (1869) had straight lines, shadows and glass fluorescence; Goldstein (1876) named "cathode rays" and found emission normal to the cathode (both via Busch 2023 [web: https://pmc.ncbi.nlm.nih.gov/articles/PMC10311274/]). Crookes's contribution is systematic public demonstration plus engineering (cup cathode, hinged cross).
• Interpretation: carriers are gas molecules. The mathematician and physicist both compute that a molecular carrier bends about 226 times less than an electron: r = 3.4 cm vs 7.6 m for 10 kV in 0.01 T [calc, assumed field]. The contrarian does not compute but agrees on the verdict.
• Disagree: the physicist labels the molecule identification "Refuted by calculation"; the mathematician and contrarian label it "Historically superseded but instructive".
• Stronger argument: all three agree on the physics. The label matters. "Refuted by calculation" requires accepted laws applied correctly with no experiment: the Lorentz force applied to a gas ion of the beam energy does that, but the calculation uses an assumed 10 kV and 0.01 T because the lecture gives neither, and the real refutation is Thomson's 1897 e/m measurement (an experiment, secondary source). The radius ratio sqrt(m_ion/m_e) is independent of voltage and field if the charge is one elementary charge and the energy equal, so the factor survives the assumption; but the "molecules, singly charged, same energy" premise is mine, not Crookes's (he gives no charge or energy). I therefore use the milder scale entry and report the calculation beside it.
• Verdict, split: observations (straight lines, normal emission, shadows, magnetic deflection): Established (shared priority with Hittorf and Goldstein; replicated by every later worker, including a modern Harvard tube). Carriers are residual gas molecules: Historically superseded but instructive, with the radius-ratio calculation as supporting evidence.
• Keep: "tired glass" (exposed area gives less light afterwards [read: p.11-12]) is a real reproducible effect and nothing in the record disputes it. Crookes says the law of the magnetic deflection would be "of great interest" [read: p.17]; the law r = mv/(qB) was found later.
• What would settle it (already settled in practice): measure e/m from deflection in crossed electric and magnetic fields, as Thomson did in 1897; the reader can do it with a school cathode-ray tube.

Claim 4. Negatively electrified, parallel streams repel (so not current carriers), impact makes glass, diamond, ruby and others phosphoresce

• Agree: all three. Negative charge: Established (Schuster, Perrin 1895, Thomson 1897, all via Wikipedia and Busch 2023). Repulsion argument is sound: for parallel beams the magnetic attraction is a fraction (v/c)^2 of the electric repulsion; at 10 kV v/c = 0.20 and the fraction is 3.9 %, so the net force is repulsive [calc: physicist and mathematician, agreed]. Space-charge neutralisation by residual-gas ions would reduce the repulsion; the lecture gives no angle, voltage or pressure for Fig. 18, so it cannot be checked quantitatively. Phosphorescence is cathodoluminescence.
• Ruby line: 6,895 (units unstated) is within about 0.5 to 0.7 % of the R lines [web check above].
• Rejection history (contrarian, kept): Hertz (1883) found no electrostatic deflection; his vacuum was not good enough (surface charge masked the field), later shown, as a Wikipedia summary reports [web: https://en.wikipedia.org/wiki/Cathode_ray]. A real null result, wrong in cause, and on the very thing (vacuum) Crookes had specialised in.
• Verdict, three parts: negative electrification of the stream and mutual repulsion: Established. Phosphorescence/cathodoluminescence of glass, diamond, ruby: Established (and the method was extended into radiant-matter spectroscopy 1883-87, per the Investigator). "The charged bodies are molecules of the residual gas": Historically superseded but instructive.
• What would settle it: nothing outstanding. The Bakerian Lecture 1879 may give Fig. 18's numbers.

Claim 5. The stream carries strong mechanical force and heat

• Agree: heating at the focus is Established (energy budget: 10 W at 10 kV and 1 mA, against about 1.6 W per mm^2 radiated at 2300 K [calc, assumed beam; lecture gives no temperature or current], so melting is feasible; electron-beam melting is standard industrial practice). The wheel and vanes do turn (modern 40 kV tube). Mechanism disagrees with Crookes: the momentum of a 10 kV, 1 mA electron beam is about 3.4 x 10^-7 N [calc], about four to five orders below the weight of a gram-scale wheel; radiometric force at 0.1 Pa and a 50 K difference is about 8 x 10^-7 N per cm^2 [calc, physicist]. Starke (1900, experiment) concluded that thermal effects exceeded impulse; Thomson (1903, p.501) attributed the motion mainly to the radiometric effect and, per Wikipedia, estimated about one revolution per minute from electron momentum [web: Harvard page, via research note; Wikipedia Crookes tube; primary texts not read].
• Disagree: on the hot-ring demonstration. The physicist reads it as supporting the thermal mechanism; the contrarian adds that it is also, ambiguously, the earliest observation in the record of charged emission from a hot wire (later thermionic emission; Edison effect 1880, Richardson [memory, medium]) and says its mechanism was "never properly tested here". The mathematician reads it as thermal (free-molecular radiometric force, lambda = 6.6 cm, far above the vane-to-ring gap).
• Stronger argument: all are compatible; the contrarian's point is the more careful one. The lecture gives no pressure, ring temperature or bias, so it cannot separate radiometric force from emission-driven effects, and a hot red-hot platinum ring gives little thermionic emission [memory, medium]. The mechanism has been tested for the wheel (Starke), not for the ring.
• Verdict, split: heating and melting at the focus: Established. The wheel and vanes turn: Established as an observation. Impact momentum of the stream as the main cause of the wheel's motion: Tested, not supported (Starke 1900, Thomson 1903, secondary reports; my momentum number agrees but rests on an assumed current). The hot-ring vane spin: observation Established; mechanism Never properly tested in this record (thermal drive is the accepted explanation for radiometers [memory, high]).
• What would settle it: Starke 1900 (Ann. Phys. 3, 101) for the thermal-versus-impulse data, and Thomson 1903 p.501 for the calculation. For the ring, run the contrarian's experiment: a heated, unbiased ring against a biased one in a tube at a known pressure, with the vane current measured.

Claim 6. Effects the same for hydrogen, carbonic acid and air; about 10^18 molecules remain at 10^-6 atm in a 13.5 cm bulb

• Gas independence. Agree: the observation (phenomena identical, onset pressures differ [read: p.21]) is what one expects if the carriers come from the cathode and not from the gas; Crookes's own observation points toward the later view (the contrarian's remark: hydrogen of mass 2 and CO2 of mass 44 should deflect very differently if molecules were the carriers). Nobody gave data; the lecture reports it qualitatively, in one source (the lecture), and later work (Thomson 1897) found the same universality (reported at medium confidence from memory by the contrarian, high by the physicist; not web-checked).
• Count. Agree: arithmetic correct from Stoney's input (printed 1,288252,350000,000000 for the bulb at one millionth; 1.29 x 10^18) [read: page image p.22; calc]. The input is 37 to 40 times too high, so the correct number is 3.2 to 3.5 x 10^16 (see recomputation table), and the filling time is about 10 to 11 million years instead of 408 million. Crookes himself says "figures when they mount so high cease to have any meaning" [read: p.23]. The rhetorical point survives.
• Disagree: the contrarian rates "same behaviour across gases" both Established and "Plausible, untested as a quantitative statement"; and the counts "Established as arithmetic from inputs". The physicist and mathematician say "Refuted by calculation" for the figure.
• Stronger argument: the physicist and mathematician. The conclusion drawn from the number (a huge number of molecules remain) survives; the number itself does not.
• Verdict, split: gas independence of the beam phenomena: Established as an observation (reported qualitatively by Crookes; consistent with Thomson 1897). The figures 10^18 molecules and 408 million years: Refuted by calculation (accepted number density; show the calculation above). The counting method and the claim that a millionth of an atmosphere still holds an enormous number of molecules: Established (about 3 x 10^16).
• What would settle it: nothing needed beyond the calculation; to put the number on an independent footing, retrieve Stoney's 1868 paper (Phil. Mag. 36, 141, as cited in the lecture footnote) for his stated figure.

What would settle it: summary

Open questions

1. What did Stokes find unclear in 1879, and which "further clarifying experiment" did he suggest?
2. What do the unpublished numbers in the notebooks (vols. 2 and 5) say about the pressure, voltage and fields behind each demonstration?
3. When and where did Crookes say that radiant matter "now passes as electrons"? Was it a concession, or the claim that the electron was his radiant matter all along?
4. Did the photographic-fogging remark Busch quotes appear in some printing of the lecture or in a different paper?
5. How large is the thermal versus impulse share of the vane force in a well-characterised modern tube? Starke's data would answer.
6. Is the hot-ring effect partly thermionic? The Edison effect of 1880 is a year after the lecture; the lecture predates a clean separation.
7. Gimingham's dates (1869-1881 against 1870-1882) and whether his ledger (IET SC MSS 285) records the apparatus built for the lecture.
8. The Plucker attribution of the hinged cross, as against the lecture's credit to Gimingham.

Absence of evidence: where the record goes dark

The Investigator found no gap in the main line of the record, and I agree: Crookes published the technical work (Phil. Trans. 1879, refereed favourably by Stokes), the exchange continued in Nature (1880), the method carried on to 1887 as radiant-matter spectroscopy, and the interpretation was settled by a test (Thomson 1897, e/m), not by decree [web: via Investigator; sites.md]. I looked for the anomalies the prompt asks for.

1. The laboratory notebooks (MS/0410, five volumes, July 1875 to July 1883, including the notes for the Sheffield lecture) are "currently unavailable". The Science Museum's catalogue gives the reason: the volumes "register slight radiation" [web: https://collection.sciencemuseumgroup.org.uk/documents/aa110067393/laboratory-notebooks-of-william-crookes, via research note]. - Competing explanations: (a) radioactive contamination of the paper from the radium and uranium work Crookes did from the 1890s, the stated reason and, to a lab that used uranium glass and radioactive preparations, a physically ordinary one; (b) curatorial caution about handling; (c) concealment, which nothing in the record suggests. - For (a): the catalogue states it; Crookes worked with uranium and radium compounds later [memory, high]; the museum applies this to the notebook set openly. Against: none found. For (b), (c): no evidence either way; nothing indicates withholding of the lecture itself or of any paper. - What would decide: a radiological assessment or a licensed digitisation (the museum may hold scans); a request to the Science Museum Group; the Gimingham ledger at the IET (SC MSS 285), which is independent. - Weight: the working record behind the lecture's numbers is therefore not reachable, a real limit on what can be checked in claims 2, 5 and 6, but a documented, mundane reason is on record.

2. Stokes's referee reports are online but unread. Not a gap in the record, a gap in the review. They are the most useful unread document in the project.

3. Crookes's own retraction. The strongest statement of his revised view comes from an obituary, not from a dated, located primary source. Competing explanations: (a) he said it, in a later address, and the obituarists reused it; (b) the obituarists compressed several statements; (c) misattribution. For (a): one documented obituary quotation (Dyer 1920, J. Chem. Soc., via the Investigator), no second source found. Against: I found no contradicting source. Deciding record: his 1891 and 1898 addresses and the Royal Society obituary notice.

4. Newspapers and reception. Papers Past was blocked by extension permissions, Chronicling America, Theosophy Wiki and Sage by bot checks; Trove, the British Newspaper Archive, Royal Society Publishing, JSTOR and Cambridge Core were not reached (time limit). Silence from these sites is not contradiction. Nothing suggests suppression, since the lecture was reprinted in Popular Science Monthly.

Overall judgement

The lecture is a set of demonstrations that nearly all hold up today. What did not hold up was the explanation of the carriers (residual gas molecules) and, in detail, the mechanism of the paddle wheel, the dark-space identification and the molecule count. The "fourth state" is a label that was later reassigned to plasma. The decisive evidence against the molecular picture is experimental (Thomson 1897, electron e/m) and supported by a calculation (deflection radius about 226 times that of an electron); the other quantitative items rest on assumptions that the reviewers stated.

Final verdicts, in brief

QE revision

Role: Arbiter, second pass, after qe.md (verdict MINOR) and the new deep reads in deep/ (Fig. 18 pp.18-19; pp.7-8 Morgan footnote, pressure table and potash tube; pp.14-15 hot-ring apparatus; p.9 bulb A). Tags as above. Web budget for this pass: 4 calls used. Nothing from the earlier arbiter text is removed; where a verdict changes, the old verdict is stated beside the new.

What the new deep reads establish [read]

• Fig. 18 (pp.18-19). Crookes gives no distance, pressure, voltage, coil setting or angle. The text says only "two negative terminals (a, b) close together at one end"; the first ray "jumps up from its first position, d f, to d g"; the lower ray's deflection is asserted ("if time permitted I could show you"), not shown. In the engraving the discs a and b sit one above the other, about a third of the tube's inner height apart, and the screen starts after a gap from them; the displacement appears to grow along the screen from the slit d (a fan), but the deep reader flags the engraving's geometry as unreliable and the gap hides anything that happens near the cathodes. No test is described with b present but not emitting. [read: pp.18-19, page image p.18]
• Non-conducting vacuum (pp.7-8). Four statements: the spark "refuses to pass" past the optimum of about a millionth of an atmosphere; the Fig. 5 potash tube is at first "non-conducting"; warming the potash to release "a trace of aqueous vapor" makes conduction start "instantly"; the last stage (described, not shown) is "the vacuum becoming non-conducting". Morgan's 1785 footnote (Phil. Trans. vol. lxxv, p.272) reports "neither the smallest ray of light, nor the slightest charge" in a boiled-mercury gauge, and says conduction also stops when air is dense enough, so he describes a conducting band between two non-conducting extremes. Crookes quotes it without comment. The citation was not checked against the original volume. [read: page image p.7, p.8]
• Hot ring (pp.14-15). Only "exhausted to a very high point", "a small galvanic battery", "red-hot"; no ring-to-vane distance, pressure, temperature, rotation sense or control. The vanes are four clear mica squares at 45 degrees. With the coil disconnected the vanes "spin as fast as they did when the induction-coil was at work". Crookes reads this as a hot wire setting radiant matter in motion. [read: p.15]
• Bulb A (p.9). Verbatim text of the low-against-high-vacuum contrast is now recovered: in low vacuum the glow follows the positive pole; in high vacuum it "scarcely matters". [read: page image p.9]

My recomputation of the QE numbers [calc, project venv]

Assumptions are mine, not Crookes's; he gives none.

The QE's central deflection number is confirmed. One sentence in qe.md needs qualifying: "about 100 to 1,000 times too small" holds at 10 kV only. Beam-on-beam deflection scales as I / V^(3/2) (line charge falls as 1/v, and the transit time as 1/v), so at 1 kV it is about 30 times larger, 2 mm at 1 mA and 21 mm at 10 mA, which is a visible jump. Crookes gave no voltage. A 1 kV beam is not implausible for his early induction-coil tubes, though I have no source for the voltage he used [memory, low]. The QE's second cathode field estimate was not computed by the QE; I did a crude one (cathode-fall field 10^4 to 10^5 V/m acting for 1 to 3 cm on a 10 keV beam gives 1.5 to 45 mm at 30 cm), which only shows that a nearby cathode's field is of the right size. It depends on field geometry I do not have.

Web checks (4 used; budget 8)

• Thomson 1897, e/m independent of gas and of cathode metal: confirmed. Thomson repeated the measurement with different gases and different cathode metals and obtained the same value within experimental accuracy [web: https://www.chemteam.info/AtomicStructure/Disc-of-Electron-Intro.html; paper: J. J. Thomson, Phil. Mag. 5th series 44, 293-316 (1897); https://web.mit.edu/8.13/8.13c/references-fall/relativisticdynamics/thomson-cathode-rays-1897.pdf, listed in the search, not opened]. Claim 6a stays Established and the [memory, medium] tag is retired.
• Edison effect date: confirmed. Edison rediscovered the effect on 13 February 1880 and in 1883 found current across the vacuum to a glowing filament; Richardson's law is from 1901 [web: https://wikipedia.com/wiki/Thermion and the other results of the same search, from search summaries]. So the hot-ring spin of 1879 predates the first thermionic report by a few months.
• Paschen low-pressure branch: partly confirmed. Sources describe the left (low-pressure) branch of the Paschen curve, separated from the right branch by a minimum in the breakdown voltage as a function of pressure times gap [web: https://w3.pppl.gov/~ikaganov/pdf_files/Three_Regimes.pdf, search summary]. The same source says that at high voltage and low pressure additional processes (fast ions and fast neutrals) matter, so the summary of the mechanism as simply "too few molecules to ionise" is the textbook account but not the whole modern picture. I did not fetch the page.
• Goldstein 1880 on deflection of cathode rays by a neighbouring cathode: check failed. The search returned only Goldstein's magnetic deflection (1880) and Crookes's charged-molecule picture, and nothing on neighbouring-cathode deflection. The QE's recollection stays [memory, medium] and unverified; no verdict below rests on it.
• Platinum work function (5.6 eV): not retrieved (search did not state it). The conclusion does not depend on it: even 4.5 eV gives a negligible current.

Claim 4: Fig. 18 repulsion (verdict changes)

• Previous verdict. "Negative electrification of the stream and mutual repulsion: Established."
• Disagreement now. The three reviewers accepted beam-on-beam repulsion as the explanation. The QE shows that nobody estimated its size. The deep read adds that the record contains no test separating beam-on-beam force from the field of the second, nearby cathode.
• Stronger argument. Both sides have limits. For a free, parallel, 10 kV beam of 1 to 10 mA, the space-charge force gives 0.07 to 0.7 mm over 30 cm and falls further with ion neutralisation (the physicist and mathematician say that is likely in a tube with residual gas); that is too small for a jump an audience sees. But the force grows as V^(-3/2), so at about 1 kV it is a few mm to 20 mm, and Crookes states no voltage, current, pressure or separation. The alternative (the second cathode's field deflects the first beam near its source) is physically reasonable and fits the deflection direction (away from b) and the fan shape suggested by the labels, but it is a hypothesis: the web check of Goldstein failed and the engraving is unreliable. So the calculation weakens Crookes's reading without refuting it, and the alternative is not demonstrated.
• New verdict, split. (i) The first ray is displaced away from the second when the second cathode is switched on: Established as a reported observation (single source, the lecture, shown live to an audience; the lower ray's deflection is asserted only). (ii) The cause is mutual repulsion of the two beams as like-charged streams: Plausible, untested, with the note that at 10 kV and 1 to 10 mA the calculation falls short by a large factor, so the claim needs a lower voltage or a higher current to work. (iii) The cause is the electrostatic field of the second cathode: Plausible, untested. (iv) Negative charge of the stream: Established, unchanged; it rests on Perrin 1895 and Thomson 1897 independently, not on Fig. 18. (v) Phosphorescence: Established, unchanged. (vi) Carriers are gas molecules: Historically superseded but instructive, unchanged.
• What would settle it. The Fig. 18 tube, or a modern equivalent, with the second terminal present but shielded or not emitting; with its potential varied independently; and with beam current and accelerating voltage measured. Crookes's Phil. Trans. 1879 "Contributions" paper may give the numbers (not downloaded).
• Keep in the record. Hertz 1883 found no electrostatic deflection in his tube (reported as a vacuum problem in Wikipedia's account, unchanged from the earlier text). Parallel beams do repel in principle, with the magnetic attraction only (v/c)^2 of the electric force, 3.8 % at 10 kV: this physics is correct and stays; the open question is the size of the effect in Fig. 18.

Claim 5: the hot-ring vanes (verdict changes)

• Previous verdict. "Hot-ring vane spin: observation Established; mechanism Never properly tested."
• Disagreement now. The contrarian offered thermionic emission as a possible second mechanism; the physicist and mathematician read it as thermal. The QE shows Richardson's law excludes thermionic emission at red heat.
• Stronger argument. The calculation. At about 900 to 1100 K the current density is 10^-20 to 10^-14 A/m^2; the wire has a surface of order 10^-5 m^2; the current is below 10^-14 A. Lowering the work function to 4.5 eV does not rescue it (3.5 x 10^-9 A/m^2 at 1100 K). The ring was also run unbiased (coil disconnected, battery heating only) [read: p.15], so no field was pulling anything off it. The contrarian's point was reasonable on 1879 knowledge (the Edison effect is 1880, a year later) and is kept. One caveat: "red-hot" is Crookes's phrase and no temperature is given; above about 1500 K the figure rises toward 10^-7 A/m^2, still negligible; and the full Richardson current is also reduced by surface contamination or increased by oxide layers, which I cannot assess.
• Radiometric reading. The thermal drive is the accepted explanation of the light-mill effect (radiometer) [memory, high, as earlier], and the apparatus is a radiometer-type fly (mica vanes at 45 degrees near a hot body at about 10^-6 atm, where the mean free path is 6.6 cm). But the lecture gives no control (no cold ring, shield or reversed fly), no pressure, no distance, no direction of turning. I therefore do not call the thermal mechanism measured here; it is the accepted physics applied to an unmeasured demonstration. A direct check is still the unbiased-versus-biased ring.
• New verdict, split. Vanes spin when the unbiased ring is red-hot: Established (observation; single source, the lecture; later radiometer work is consistent). Thermal (radiometric) drive: Established as accepted radiometer physics, with the apparatus unmeasured. Thermionic contribution at red heat: Refuted by calculation. Crookes's own reading (the ring emits radiant matter that moves the vanes): Tested, not supported only for the impact-momentum version via Starke 1900 and Thomson 1903 (secondary reports, as in the first pass); for the ring specifically nothing in the record tests it directly, so it stays in the record as his stated interpretation.
• Open question 6 above (is the hot-ring effect partly thermionic?) is answered for red heat by calculation and remains open only for a ring much hotter than red heat.

New claim 7: the highest vacuum does not conduct (new verdict)

• Claim. At the very highest exhaustion the tube stops conducting and the spark refuses to pass; a trace of vapour restores conduction at once (pp.7-8). Crookes demonstrated the onset (potash tube, Fig. 5), described but did not show the final stage, and quoted Morgan 1785 as precedent.
• Reviewers. None considered it (the QE found no mention in any review). So there is no disagreement to arbitrate; this is a gap the QE caught.
• Assessment. This is a demonstration of something real and consistent with accepted discharge physics: breakdown needs enough gas for electrons to produce ionising collisions, so at very low pressure the voltage needed rises again, the left branch of the Paschen curve [web partial, see above; memory, high for the textbook statement]. The lecture states it qualitatively; no pressure is given for the failure point beyond "about a millionth of an atmosphere" as the best-phosphorescence optimum. Morgan's own extract is both experiment and interpretation ("the particles of air may be so far separated from each other as no longer to be able to transmit the electric fluid"); that interpretation is a historical one. Morgan's vacuum is a boiled-mercury Torricellian vacuum with mercury vapour present [memory, medium] and a frictional machine as source, a different setup from Crookes's, so the two results agree in kind but are not the same measurement. Crookes's own paper "On Electrical Insulation in High Vacua" (1878; archive.org philtrans02850338, per the Investigator) is unread and probably gives the numbers.
• Bearing on claim 1. It bounds the radiant-matter regime from above in vacuum: it cannot be reached by pumping alone with this discharge, and Crookes got his best vacuum partly with absorbents (below).
• Verdict. Observation (no conduction at the highest vacuum; conduction restored by a trace of vapour): Established. Explanation (too few molecules to sustain the discharge): Plausible, untested in this record (consistent with accepted discharge physics; the Paschen web check was partial and the Phil. Trans. paper is unread). I do not use "Established" for the explanation because I did not retrieve a source that states it for this kind of tube.
• What would settle it. Crookes 1878 and the 1879 Bakerian Lecture; or modern low-pressure breakdown curves for air in a long tube.

Other points from the QE adopted without a verdict change

• Chemical absorbents (p.21). Crookes reached his highest vacuum by putting phosphoric anhydride, potash, palladium and carbon in the tube, "to much higher degrees than I can get by mere pumping" [read: p.21, via deep read]. It is an early form of chemical gettering [memory, medium on the term's history]. It bears on how the 1/20,000,000 atm figure was reached; the gauge and method for that figure are not described in the lecture, so the figure's measurement basis remains unknown.
• Low-against-high vacuum contrast (pp.9-10, 17). Verbatim now available. In low vacuum the discharge follows the positive pole and, when deflected by a magnet, rises and returns to the pole; in high vacuum the position of the positive pole "scarcely matters" and the deflected stream "did not recover itself" [read: page image p.9; p.10, p.17]. Observation: Established. It supports claim 3's independence from the anode.
• "Three hundred trillions a second" (pp.22-23). In Crookes's long-scale words this is 3 x 10^20 per second; his bulb count divided by it gives 71.6 minutes, which matches "before you leave this room" [calc]. The figure is self-consistent; whether the rate itself is physically right is not testable from the lecture. I do not change claim 6; the count's input is still 37 to 40 times too high.
• Long-scale number words. "Quadrillion" is 10^24 and "trillion" 10^18 in 1879 British usage [read: page image p.22; memory, high for the usage]; the digest's 10^15 and 10^12 were misreadings.
• Accuracy corrections in the record. The printed pages of the journal are 13-24 and 157-166, not 13-25. The focusing cup is not credited to Gimingham in the lecture; he is credited with the hinged cross (Fig. 9), the paddle-wheel railway (Fig. 11) and the magnetic wheel (Fig. 17) [QE A1, A2, checked by the QE against the text; I did not re-read these pages]. The primer's "far emptier than anyone before him" and "wrong only about which particles" are overstatements the primer author should qualify; Morgan's 1785 vacuum is a counterexample for the first.
• Web budget. Reviewers left medium-confidence items unchecked (QE PR1). I checked Thomson 1897 and Edison 1880 (both confirmed); Goldstein failed; the platinum work function was not needed.

Updated "what would settle it" rows

Absence of evidence

No change. The QE's remediation item I1 (download Stokes RR/8/151-152, the Bakerian Lecture, the 1879 and 1878 papers, Starke 1900 with the owner's approval) stands; these are the records that would settle claims 2, 4, 5 and 7.

Final verdicts after the QE revision, in brief