Tyndall-John: Physicist review (Step 2)

Reviewer: PhD physicist. Project: John Tyndall's radiant-heat, gas-absorption and atmospheric-vapour work (1859-1892). Date: 2026-10-08. Tags: [read] from the project documents or deep reads (page images were checked by the deep readers; I re-read the extracted text for the pink-glow passage, the "nearly eight times" sentence and the "conservation of molecular action" passage); [web] with URL; [memory] with confidence; [calc] my own arithmetic (numpy, run in the project venv).

How I worked, and limits

• Read in full: claims.md, all five deep reads, sites.md and the Investigator's notes (archive.org, Epsilon, Wikisource, Royal Society, viXra, Science History Institute, Protons for Breakfast). Spot-checked three passages in _text/. I did not re-read facts.json entry by entry; I checked its claims against the deep reads.
• The Investigator downloaded no files (research/files/ is empty), so everything from outside the folder rests on the Investigator's notes and quotations, which I treat as [web] second-hand.
• Not read by anyone: Magnus 1861/62, Lecher and Pernter 1880/81, Strachey July 1866, Tyndall's own memoirs (Contributions to Molecular Physics), Foote 1856 text. Verdicts below that lean on them are marked.
• Tyndall is an interested party on his own priority (calorescence, diamond burnt by a terrestrial source, water-vapour absorption, glacier theory). Single-source statements (e.g. "most delicate experiments" on ice stretching, "ten-foot column intercepts 10-15 per cent", "two-inch ether shell intercepts 35 per cent") are quoted and weighed, not dismissed; each has no experimental record in the folder.
• Web checks done (budget ~6): Schmidt et al. 2010 greenhouse attribution; Glen 1955 ice creep; Draper point. Three medium/low-confidence memory facts remain tagged [memory].

Corrections to Step 1

1. claims.md claim 2 says water vapour acts "70-90 times per unit amount" of the dry air carrying it. That is wrong. The 70, 72 and 90 are the absorption of the vapour actually present in the sample (about 0.45-0.5 per cent of the molecules, Heat para 469) compared with the whole air that carries it [read: Heat p.380, 387-388]. Per molecule Tyndall gets 80 x 200 = 16,000 [read: Heat p.401; calc: 80 x 200 = 16,000]. The two figures differ by a factor of about 200 and must not be merged. (The deep read states this correctly; claims.md garbled it.)
2. claims.md claim 1: "air = 1" is not an absolute zero but a floor. Tyndall's air value is the smallest deflection he could measure (about 1 degree of 30, Heat p.380) and he himself says traces of organic impurity can raise it fiftyfold [read: Heat p.381]. All the large ratios (972, 5460, 6030 ...) are therefore lower bounds on the true gas-to-air ratio. They are ratios at equal pressure and for one source (copper plate, or Bunsen plate in Heat); the two tables (1 inch and 30 inches) are consistent: CO2 972 at 1 inch becomes 90 at 30 inches because air's absorption grows about 30-fold with pressure while CO2's saturates [calc: 972 x ~2.8 / 30 = 90].
3. Chlorine 60 and bromine 160 (Heat Ch. X) appear in the list of absorbers. Cl2 and Br2 are homonuclear diatomics with no dipole change, so by modern selection rules they should be infrared-inactive; the values are most likely traces of HCl or similar, or the floor effect in item 2 [memory: high on the selection rule; the cause is my inference]. claims.md should not present "elements are transparent" without this exception, and without ozone, which Tyndall himself found a strong absorber (Heat paras 435-442, headings only) and which is an elemental molecule that absorbs at about 9.6 micrometres [memory: high].
4. claims.md says the invisible/visible ratio "one-ninth" (On Radiation, 1865) "is the same ratio as 11:89". It is not: 11:89 is 1:8.1, consistent with the 8 of 1866 (89/11 = 8.1, deep read 1-full); one-ninth of the dark is 10 per cent of the total, and 1:9 is a slightly different number. A 10 per cent difference, minor, but it shows the 1865 and 1866 figures were not the same measurement.
5. The "sun about 2" (Muller) is repeated in claims.md as Tyndall's solar figure. It is Muller's, quoted by Tyndall [read: 1866 p.3, p.7]. Modern blackbody-sun arithmetic gives infrared:visible near 1.07 at the top of the atmosphere (see claim 3); 2 would be too high [calc; [memory: medium] for the real solar spectrum].
6. sites.md "Corrections": the Strachey dates (March 1850 in Heat; Madras 1841-44 in the 1882 Bakerian) are not a discrepancy I can resolve without Strachey's July 1866 paper. Both come from Tyndall's description of the same paper; two data sets in one paper is plausible. Left open.
7. claims.md claim 4 calls calorescence "the reverse of Stokes's fluorescence rule". That is Tyndall's description. The physicist's reading is in claim 4 below; the wording in claims.md should say "which Tyndall presented as", because Stokes himself disagreed [web: Epsilon letter 2215, per Investigator].
8. Confirmations: nitric oxide 1590, olefiant gas 6030, aniseed 372 are corroborated by the validated Wikisource text [web: https://en.wikisource.org/wiki/On_Radiation_(Rede_Lecture), per Investigator]. The Rumford 107 (F, after one hour) and the 23.5 versus 28 minute corrections are sound as stated in the deep read.

Mainstream view in brief (what each verdict is measured against)

• Infrared absorption by a gas needs a molecular vibration or rotation that changes the electric dipole moment. N2, O2, H2 have none, so they are practically transparent from the near infrared to the far infrared; CO2 (bending mode, 15 micrometres), H2O (6.3 micrometres band and the rotational band beyond about 17 micrometres), CH4, N2O, O3 do. Absorption by a gas is described by lines, broadened by pressure; emission and absorption at a given wavelength are equal for a body in thermal equilibrium (Kirchhoff, 1859-60). [memory: high]
• Greenhouse effect: with sunlight absorbed of about 240 W/m2 the Earth would sit near 255 K without absorbing gases; the observed mean surface is about 288 K [calc: (240/5.67e-8)^0.25 = 255 K; surface value memory high]. In current model attribution about 50 per cent of the clear-and-cloudy greenhouse effect is water vapour, 25 per cent clouds, 20 per cent CO2 [web: https://www.giss.nasa.gov/research/briefs/2010_schmidt_05/ (Schmidt, Ruedy, Miller, Lacis 2010, JGR 115, D20106)], while the authors stress the shares depend on how overlap is allocated, and that water vapour and clouds respond to temperature whereas CO2 largely does not.
• Light and radiant heat are electromagnetic waves of different wavelength (Maxwell 1865; Hertz 1887-88). [memory: high]

---

Claim 1. Elementary gases are practically transparent to radiant heat; compound gases and vapours absorb it strongly (tables of 972, 5460, 6030 times air); radiating power equals absorbing power

a. Mathematics. There is almost no mathematics, only ratios of galvanometer deflections. The deflection-to-heat conversion is a calibration table that is not given in the 1866 paper, and it changes the answer by several points (94.3 per cent raw versus 94.8, 86.8 per cent converted for the iodine cell, reader check in the deep read). Within the gas tables, internal consistency is good: the 1-inch and 30-inch tables reconcile (Correction 2); humid-air absorption is linear in pressure (observed 16, 32, 49, 64, 82, 98 for 5-30 inches against 16, 32, 48, 64, 80, 96, i.e. 98/16 = 6.1 against 30/5 = 6.0 [read: Heat p.388, page image; calc]), as Beer-Lambert predicts when total absorption stays below about 10 per cent. Where absorption is large the data show saturation, as the physics requires: CO2 against a carbon-monoxide flame rises only from 48.0 to 74.3 per cent while pressure rises tenfold (1 to 10 inches), a power-law exponent of about 0.19; olefiant gas 23.2 to 65.5 per cent, exponent about 0.45 [read: Heat digest p.431, not image-checked; calc]. Small internal slips: nitrous oxide is "1800" in the prose and 1860 in the table (same page) [read: On Radiation p.23].

b. Established results. Fully consistent. Modern infrared spectroscopy (the HITRAN database, FTIR) reproduces the pattern: CO2, H2O, NH3, N2O, SO2, CH4, olefiant gas (ethylene) are strong absorbers; N2, O2, H2 are not [memory: high]. The ordering (ammonia, olefiant gas, SO2 thousands of times air; CO, CO2 hundreds) matches line-strength expectations. The reciprocity of emission and absorption is Kirchhoff's law, which Tyndall tested for gases (hot copper ball above a gas jet, Rede lecture) [read: On Radiation p.44-47].

c. Tested? By Tyndall from 1859. The first outside reaction was surprise at the size of the effect (J. D. Forbes to Tyndall, 15 July 1859, asking tube length and percentage) [web: https://epsilon.ac.uk/view/tyndall/letters/Tyndall1627, per Investigator]. Dry-gas results were confirmed by Lecher and Pernter (1880, who dissented only on vapours) [web: https://en.wikisource.org/wiki/Popular_Science_Monthly/Volume_20/March_1882/Sound_and_Radiant_Heat, per Investigator]; by Rontgen with an independent method (pressure rise of the irradiated gas) [web: https://archive.org/details/sim_annalen-der-physik_1884_23_9, per Investigator]; Bell and Tyndall's photoacoustic sounds (1880-81) are the same absorption seen another way. Later band-resolved work (Paschen, Knut Angstrom, Rubens and Aschkinass) is summarised in Very 1900 [web: https://archive.org/details/atmosphericradia00veryrich, per Investigator]. Magnus (1861-62) and Buff disagreed, on vapours (see claim 2) and also on gases: Very reports Magnus's "14.75 per cent from dry air" with glass end plates and rejects it as a fallacy of heated plates and convection [web, per Investigator]. Priority: Eunice Foote (1856, sunlit gas-filled cylinders; CO2 and water vapour heated more) and R. Franz (1855, diathermancy of gases) preceded Tyndall's 1859 work; Jackson (2019) concludes this was simultaneous discovery from New York, London and Berlin and that Tyndall did not cite Foote [web: https://royalsocietypublishing.org/doi/10.1098/rsnr.2018.0066, per Investigator]. Franz found 3.54 per cent apparent absorption for dry air in his tube [web: Very 1900, per Investigator]; that is the floor of Correction 2. Tyndall's first crude trial (air, hydrogen) was negative; the compensated two-pile method then gave "very large effects" [web: https://archive.org/details/jstor-111604, 1859 Note, per Investigator].

d. If true. Nothing needs to change; it is mainstream. The only residual objection: Macdonald (viXra:1504.0165, unrefereed, recorded as his claim) argues that thermopile detection misreads the experiment and that N2 and O2 also absorb because they have Raman-active vibrations [web: https://vixra.org/abs/1504.0165, per Investigator]. Physics reply: Raman activity needs a change in polarisability, infrared absorption needs a change in dipole moment; they are different selection rules, and homonuclear diatomics are infrared-inactive for a single photon. N2 and O2 do show very weak collision-induced and (O2) magnetic-dipole/quadrupole absorption, orders of magnitude below CO2 [memory: medium-high; not checked on the web]. So Macdonald is right that "zero" is not exact, and wrong that it changes the picture. Tyndall's own "1" already allows for a small residual.

Verdict: Established. Qualification: the numerical table values belong to Tyndall's source, tube, rock-salt windows and calibration; they are lower bounds relative to air (Correction 2) and are not line strengths.

---

Claim 2. Atmospheric water vapour, far more than dry air, controls the Earth's heat loss ("blanket"); at least 10 per cent of terrestrial heat is intercepted within ten feet; night cooling, frost, desert and mountain ranges and tropical rain depend on it; CO2 is minor

Split into five parts.

2A. Water vapour absorbs infrared strongly, much more than dry air (70-90 times for the vapour present; 16,000 per molecule). a. The arithmetic is consistent but rests on round numbers: 80 x 200 = 16,000 uses 80 (measured 70-90) and 1 vapour molecule in 200 (0.5 per cent) where the same chapter gives 0.45 per cent. Order of magnitude only [read: Heat p.401]. The ratio's rise (15 times in the first run, then 30, 40 ... 70, 90) is explained by Tyndall as the dry reference becoming purer, which is the correct direction: the denominator approaches the true floor [read: Heat p.380, 401]. b. Consistent with spectroscopy. Water vapour has the 6.3 micrometre band and a rotation band beyond about 17 micrometres. For a 100 degree C source (his boiling-water cube) about 15 per cent of the blackbody power lies at 5.5-7.5 micrometres and about 24 per cent beyond 17 micrometres [calc: Planck integral at 373 K]. Tyndall's 7.9 per cent absorbed in about 0.5 m of room air [read: Heat p.386, path length inferred 21 in by the deep reader; calc: 62/780 = 7.95 per cent] implies an effective attenuation of about 0.16 per metre [calc: -ln(0.921)/0.53 m], reasonable for strong band centres. c. Tested. Tyndall answered seven objections (Heat paras 471-486): brine film on hygroscopic rock-salt (plates removed; open tube 4 ft 3 in, still 45 degrees deflection against 0), impurity (air from Hyde Park, Epsom Downs, Isle of Wight; 70 times each), wall condensation (proportionality table), the tube itself (free-air cylinder test, 5 degrees toward opacity for saturated air, 10 degrees toward transparency for dried) [read: deep read sections 2.3]. That is a careful control set. Independent: Rontgen 1884 (different method) found "moist air absorbs rays of lower refrangibility" and reported many physicists still regarded the matter as open; Tyndall (1882) says Rontgen's vapour results agree with his [web: Investigator archive.org note]. Against: Magnus (1861-62) found no action and blamed vapour-hesion films; Buff likewise; Lecher and Pernter (1880-81) found "moist air a little more transparent than a vacuum" [web: Investigator, quoting Tyndall's 1882 Bakerian]. A result more transparent than a vacuum is itself a warning of a systematic error in their apparatus (adsorbed moisture on walls in the "vacuum" run is one candidate; my inference, no source). Modern gas-phase spectroscopy settles the sign: the water-vapour bands do absorb at 100 degrees C blackbody wavelengths [calc above; memory: high]. Very 1900 defended Tyndall but asked for "further elucidation" of the humid-air proportionality table [web, per Investigator]; I note the table is one of the best-behaved sets in the book. d. Changes: none. Verdict 2A: Established (qualitatively and in order of magnitude); per-molecule 16,000 figure is Plausible, a rounded estimate.

2B. At least 10 per cent of terrestrial heat is intercepted within ten feet of the ground (and a 10-15 per cent column). a. Not a derivation: an extrapolation from a roughly 100 C laboratory source in a 4 ft polished tube ("over 10 per cent") to the Earth (about 288 K) [read: Heat p.387]. The source spectrum matters: at 288 K only 7.5 per cent of the power lies at 5.5-7.5 micrometres and 38 per cent beyond 17 micrometres, 19 per cent in 13-17 micrometres [calc]. Ten feet is 3 m. b. Order-of-magnitude plausible: 3 m of humid air plus 3 m of CO2 at 400 ppm would absorb most of the radiation in the cores of the 6.3 and 15 micrometre bands and part of the far-infrared rotation band. A crude band count (a few tens of per cent of the 7.5 + 19 + 38 per cent bands, nothing in the 8-12 micrometre window at 25 per cent) suggests a few per cent up to about 10 per cent [calc, crude; no line-by-line model was run]. Tyndall's "at least 10 per cent" is therefore at the high end of what I can justify, and "certain" is too strong; the 2 inch ether shell "35 per cent" [read: Rede lecture p.44] has no stated experimental record at all, and the Popular Science Monthly 1882 summary gives 2 per cent for a 1/20 inch gap and 35 per cent at 2 inches for sulphuric ether between rock-salt plates, i.e. it is a laboratory number for ether vapour, not air [web, per Investigator]. c. No independent test of the "within ten feet" number is on record in the folder. Modern spectroscopic calculation (outside the folder) shows the lower atmosphere is optically thick in the water bands over much shorter paths in humid air; I did not compute it. Verdict 2B: Plausible, untested as stated (the rough figure is a lab extrapolation; the effect is real, the number is not established).

2C. Water vapour is the principal controller of surface temperature and the "blanket" explains the diurnal range, nocturnal cooling, deserts and mountains. a. Qualitatively sound. Strachey's series (vapour tension 0.888 down to 0.435 inch against night fall 6.0 up to 16.5 degrees) shows near-monotonic dependence (one reversal, 12.6 then 12.1) [read: Heat p.391, page image]; halving the vapour pressure almost triples the night cooling [calc: 16.5/6.0 = 2.75]. Martins and Bravais: 24 F difference at the Grand Plateau of Mont Blanc against 10 F at Chamonix [read: Heat p.393-394]. A single "representative" series chosen by Tyndall is not an independent test, and confounders (wind, cloud, ground, elevation) are not controlled. b. Consistent with modern physics: dry-desert diurnal ranges, radiative frost, high-altitude cold are all dominated by downwelling longwave from water vapour and clouds. The planetary framing ("the elevation of our planet's temperature is a direct consequence of aqueous vapour"; "were that garment removed, terrestrial life would probably perish" [read: New Fragments p.106]) is right in sign: about 33 K between the 255 K blackbody and the 288 K surface [calc]. The modern attribution of the present-day greenhouse effect is about 50 per cent water vapour, 25 per cent clouds, 20 per cent CO2 [web: GISS 2010 Schmidt]. So "vapour far more than CO2" is the correct statement about the present composition. c. Tested by Strachey (as quoted), Martins and Bravais, and for the modern era by observation and radiative transfer; no controlled neutral test of Tyndall's vapour-screen reading is in the folder. Verdict 2C: Established (qualitative), Plausible for specific meteorological examples (tropical rain by vapour radiating to space; cumulus tops chilled by radiation) which Tyndall states as physical causes alongside others, not as a measurement [read: Heat p.392-393]. Sahara frost, the "Hungarian peasants' cloaks" and Livingstone/Sturt temperature ranges are anecdotal evidence cited from travellers; taken as stated, qualitatively consistent.

2D. The self-limiting feedback and the solar/terrestrial asymmetry (vapour transparent to sunlight, opaque to earth heat). Established, with a quantitative correction. The asymmetry is the heart of the greenhouse effect. Tyndall's example that a water layer one-tenth inch passes the sun's rays but half that stops "no single ray" of terrestrial heat is stronger than his own data (more than 97 per cent cut at about 0.04 inch; opaque at 0.27 inch to a hydrogen flame) [read: On Radiation p.42-43 against Heat OCR para 543]. He also overstates the transparency to the sun: water vapour has near-infrared bands that absorb a measurable share of sunlight (about 5-8 per cent absorbed by atmospheric vapour, memory medium, not web-checked), which is also why Muller's solar infrared share is depressed (Tyndall says this himself). Verdict 2D: Established.

2E. CO2 is "a minor, weaker constituent". a/b. In 1859-68 Tyndall had measured CO2 as a strong absorber per molecule (972 at 1/30 atmosphere) but weak for a copper-plate source against olefiant gas, and strong against the carbon-monoxide flame [read: On Radiation p.47-49]; he gave 0.05 per cent as its abundance [read: Heat p.379]. He never apportions climate effects between vapour and CO2 [read: Heat deep, section 5]. As an assessment of the present budget, CO2 is not "minor" (20 per cent of the greenhouse effect in the GISS attribution), and the structural point Tyndall could not see is that water vapour is a feedback tied to temperature, while CO2 largely is not [web: Schmidt 2010]. The 1859 claim was never "CO2 is unimportant"; the claim attributed to Tyndall here is a silence turned into a ranking. He also never wrote a climate-sensitivity or anthropogenic statement in these pages [read: Heat deep]. Verdict 2E: Historically superseded but instructive.

Overall claim 2 verdict: Established (core: vapour absorbs and controls heat loss; blanket effect). The 10-ft number: Plausible, untested as stated. The CO2-minor reading: Historically superseded but instructive. Priority note (Investigator): Tyndall's 1859 Note had no climate statement; the climate argument appears 1861-65 [web: Jackson 2019]. Tyndall's own letters (Hooker: "Magnus denies the action; but he has not a foot to stand upon") are his characterisation [web: Epsilon letter 1941]. Airy and Herschel approval (Airy: "we should be in a way for beginning Meteorology") is reported through Tyndall's side of the correspondence [web: Epsilon 2029].

---

Claim 3. The invisible radiation of a luminous source is several times its visible radiation (arc 7.7 by spectrum area, 8 by filter; sun about 2); iodine in carbon bisulphide removes the visible light and passes the heat, so the dark rays ignite materials and make platinum and carbon glow

a. Mathematics/method. Area under the heat-versus-pile-position curve is a legitimate energy measure, because a linear thermopile of constant width collects a fixed slice of the spectrum at each position; no wavelength conversion is needed (so there is no hidden dispersion error) [reasoning; reader check of the deep read]. The two methods agree within 4 per cent (7.7 area, 8 by filter) but both depend on the unpublished galvanometer calibration and on the assumption that all 11 per cent absorbed by iodine is visible light ("may be inferred", [read: 1866 p.11]). Tyndall states the iodine does absorb a little beyond the red [read: 1866 p.11]; a few per cent of extra-red absorbed makes the true visible share a bit smaller and the ratio larger, not smaller, so the direction of error does not undermine the conclusion. The "twice" for the sun is Muller's. b. Consistency with experiment and theory. Planck arithmetic for a blackbody gives infrared:visible (0.4-0.76 micrometres) of 18.2 at 2500 K, 7.7 at 3000 K, 4.3 at 3500 K, 2.7 at 4000 K, and 1.07 at 5772 K (the sun's effective temperature) [calc]. The carbon arc's positive crater sits near 3800-4000 K [memory: high], hotter than this equivalent 3000 K; but a Tyndall arc lamp radiates from a large area of cooler incandescent carbon, reflectors and the electrode tips as well, so an effective ratio of 7.7 is plausible (my reading; his lamp, battery and optics are not specified closely enough to compute it). The solar figure of 2 is high: with the sun a 5800 K body the top-of-atmosphere infrared share is only a little above the visible [calc]; the atmosphere removes more infrared, so Muller's 2 was likely inflated by detector or dispersion problems [inference]. The main qualitative point (an incandescent source emits the greater part of its power in the infrared, and a hotter body's share of visible rises) is textbook Planck/Wien. The iodine-bisulphide filter works for the stated reason: it is an opaque solution at visible wavelengths and transparent beyond the red, consistent with the molecular iodine absorption edge [memory: medium]. Table III is interesting physics: the iodine loses 0 per cent from dark sources, 3-4.6 per cent from flames and 10 per cent from the arc, as expected from the amount of visible light in the source [read: 1866 p.10]. Combustion at the dark focus is straightforward power density. For example, to heat a thin platinum leaf to the Draper point (about 798 K, where solids begin to glow dull red [web: https://en.wikipedia.org/wiki/Draper_point]) a leaf radiating from both faces needs about 4.6 W/cm2 [calc: 2 x 5.67e-8 x 798^4 W/m2 = 46 kW/m2]; for white heat near 1500 K about 57 W/cm2 [calc]. With a focus of a few square millimetres, a watt or a few watts suffices, and a lamp running at around a kilowatt delivering several times more infrared than light makes this easy. The unwarmed air at the focus follows because N2 and O2 do not absorb [read: 1866 p.17]. The failure to fuse platinum is explained by Tyndall by loss of the absorbing layer (platinum black burns off, reflectivity of the metal returns) [read: 1866 p.19-20]; the physics is correct, since bright polished platinum reflects most of the infrared and the emissivity drops once the black is lost [memory: medium]. The note that black glass can transmit 63 per cent of the arc's total heat and that black-bulb thermometers are therefore defective [read: 1866 p.24] is physically sound for the glass used but quantitatively unsupported (thickness and detector not given). c. Tested. By Tyndall (exhausted-receiver controls, alum-cell checks, eye tests). The Akin dispute and Stokes's comments show others did make glow with focused sunlight; Tyndall's solar tests only reached a bright red heat from the dark rays alone [read: 1866 p.20-21], which is itself consistent with the solar infrared share being modest. Independent measurement of the arc ratio: none found. d. If true: nothing changes. Safety note, not a physics point: the iodine-filtered focus passes 0.76-1.4 micrometre radiation that the eye's media transmit to the retina, so the undefended-eye test [read: 1866 p.16; footnote "I do not recommend the repetition"] is dangerous, and the observation that "the retina was not sensibly affected" does not mean the eye was safe [memory: high on the eye's transmission window]. Tyndall's three bisulphide fires (including 2 November, year unstated) show the hazard of the filter [read: 1866 p.13].

Verdict: Established. Qualification: the numerical ratios (7.7, 8) are Tyndall's for his lamp and calibration; the sun's 2 is Muller's and is high against modern values.

---

Claim 4. The glow is a transmutation of heat rays into rays of higher refrangibility (calorescence), the reverse of Stokes's fluorescence rule; sub-claim: certain blue glasses make platinum glow pink (a gap in the calorescent spectrum)

Split into three parts.

4A. The observation: dark focused rays make platinum, carbon and others glow visibly, also in vacuo, giving a continuous spectrum when viewed through a prism. a/b/c. Observed, reproducible, and explained as ordinary thermal incandescence: absorbed energy raises the metal's temperature and the hot body emits a Planck spectrum, including visible light once it passes the Draper point [web: Draper point; calc above]. Tyndall's own earlier text says every incandescent solid emits both ranges with growing intensity [read: 1866 p.3, p.7]. Tested by Tyndall, by Akin and Griffith (sunlit platinum, 2 ft mirror, pocket spectroscope, announced 1863), Croll and others [web: Phil. Mag. 29 (1865), per Investigator]; no one disputed the phenomenon. Verdict 4A: Established.

4B. The interpretation: "transmutation" of long-wave rays into short-wave ones, "reverse of Stokes's rule"; each ray returns from the metal "at a quicker rate". a. The paper holds two pictures without reconciling them: thermal emission from a heated body (p.3, p.7), and rays whose "refrangibility is raised" by the atoms of the platinum (p.17-18) [read: deep read 1-full, section 10]. Tyndall's own words to Croll are the second: "you hit the mass with waves of slow recurrence, and ... obtain, in return from the mass, waves of quick recurrence" [web: Epsilon letter 2216, per Investigator]. A single absorbed infrared quantum cannot be converted into one visible quantum (energy is not available), so the picture of "each ray raised" is wrong if taken literally. The correct statement: many absorbed quanta thermalise and the hot lattice emits a distribution whose tail reaches the visible. Energy is conserved, the second law is satisfied (the source at about 3000 K is hotter than the glowing foil at about 1500 K), and no individual photon is up-converted. The "quantum" objection is mine; Tyndall had no quantum theory (1865) but he had the energy-conservation language to say "heating". b. Stokes himself, the authority on fluorescence, wrote to Tyndall on 13 Jan 1865 that heating a solid to glow by concentrated rays "is not a little different from fluorescence", and that "negative fluorescence" as a name was not quite correct [web: https://epsilon.ac.uk/view/tyndall/letters/Tyndall2215, per Investigator]. Stokes's rule (fluorescent emission is at lower frequency than the exciting light) is a statement about photoluminescence; calorescence is not a violation of it, and the rule has genuine exceptions (anti-Stokes processes, upconversion in rare-earth materials) that are unrelated to Tyndall's effect [memory: high]. c. Priority: Akin's "calcescence" (BA 1863; Phil. Mag. 29, 1865, 28-43) and the dispute with Tyndall, who answered with 'On the History of Calorescence' (1865); Sabine and Armstrong's letters supported Tyndall's side, Tait (1868) credits Akin with "Calcescence or Calorescence" [web, per Investigator]. This is a priority dispute among men who agreed on the phenomenon; the physics of which name is correct is settled in Stokes's favour. d. Changes: nothing. Verdict 4B: Historically superseded but instructive (a reasonable language for 1865; replaced by thermal-radiation and quantum accounts).

4C. Pink glow through certain blue glasses; "a gap in the calorescence". a. Tyndall puts the claim in a conditional: "If subsequent experiments should confirm this result, it would follow that there is a gap ..." and adds that further experiments are necessary "to satisfy my own mind" [read: 1866 p.23-24, and confirmed in _text/ page 24]. b. The only control is "every care was taken to avoid subjective illusion"; the reported pink appears when the glow is very faint, in a darkened room, after passing a blue glass [read: 1866 p.23]. A plausible mundane explanation (mine, not Tyndall's, no experiment): a very dim thermal glow seen at near-scotopic levels loses colour discrimination, and a blue glass that transmits red and blue but absorbs green and yellow (the table shows the light blue glass transmitting blue, green and bands of red) would give a magenta-pink tint of a continuous spectrum. A gap in a hot platinum body's emission is not expected from Planck/Kirchhoff, since emissivity of platinum varies smoothly in the visible [memory: medium]. c. Tyndall kept the footnote in the Rede lecture and later reprints (Wikisource, Fragments of Science) [web: per Investigator]; no later confirmation or refutation found. Verdict 4C: Never properly tested (no controlled replication or refutation on record; it was not dismissed by anyone, simply not followed up).

---

Claim 5. Absorption is a property of the individual molecule, selected by resonance between the wave period and the periods of the atoms, by molecular shape and complexity, and is unchanged when the same number of molecules is compressed into a shorter tube or condensed from vapour to liquid

Split into four parts.

5A. Resonance selection of absorption. Qualitatively correct and remarkable for 1865: absorption happens when the radiation frequency matches a natural frequency of the molecule. Tyndall's CO2 test is a good illustration of Kirchhoff's law: copper-plate heat is poorly absorbed by CO2, but the radiation of a carbon-monoxide flame (hot CO2 is the radiator) is strongly absorbed; at one inch CO2 absorbs 48.0 per cent and olefiant gas 23.2 per cent, i.e. 2.07 times as much [read: On Radiation p.48-49; Heat OCR; calc 48.0/23.2]. This is exactly the line-matching behaviour the modern theory predicts [memory: high]. Tyndall's "why elements are transparent" suggestion (their atoms vibrate too fast to match) is replaced by the dipole selection rule: a homonuclear molecule has no dipole, so it cannot absorb a photon in a vibrational transition, while its vibrational frequency is not "too rapid" (N2's fundamental is near 4.2 micrometres) [memory: high]. His "ether entangled in the molecule" (the explanation of why shape matters, [read: On Radiation p.50-51]) has no counterpart: shape matters via symmetry and dipole derivative. "Absorption by the atoms, not the molecule as a whole" is replaced by normal modes of the whole molecule. Verdict: Established (the resonance principle and the selective absorption); the ether mechanism is Historically superseded but instructive.

5B. Conservation of molecular action: absorption depends on the number of molecules in the path, not on how they are packed. a. The experiment is clean in design: the same quantity of vapour in a 38 inch tube and a 10.8 inch tube (38/10.8 = 3.52, matching "three and a half") gives 18.4 per cent and 18.8 per cent (1 inch long tube against 3.5 inches short tube), 25.7 and 25.6 (2 inches against 7), 23.5 and 23.4, 32.1 and 32.6, differences of 0.1-0.5 points [read: New Fragments p.103; calc 38/10.8]. The equivalence is the Beer-Lambert statement that absorbance depends on the column density, which is exact for non-overlapping, unsaturated and pressure-independent lines. b. The modern exception: for gases with a line-structured spectrum (CO2, water) absorption at fixed column increases with total pressure because pressure broadening spreads the line; his own CO2 data show the opposite scaling in a different way (absorption rises only as p^0.19). The ether and hydride of amyl vapours he used absorb in broad, nearly structureless bands, so the law holds there. The law therefore holds well for his chosen vapours, and breaks for others, a point Tyndall's pressure tables hint at but that he does not draw. c. Tested by Tyndall only; "representative" single examples, no repeats or errors given [read: New Fragments p.106 note]. Later photoacoustic spectroscopy confirms that gas absorption is molecular [web: Investigator notes, Bell/Rontgen/Manohar 2016]. Verdict: Plausible, untested beyond Tyndall for general gases; confirmed in the weak-line, broad-band case he used.

5C. Thermal continuity of liquids and vapours (same absorption per molecule in liquid and vapour). a. The ether vapour at 7.2 inches of mercury over 38 inches condenses to a 1 mm liquid layer: I recomputed 9.6e-4 mol/cm2 for the liquid against 9.7e-4 for the vapour (the deep reader's own check, [memory/calc] for densities); that matches. Results: ether, platinum-spiral source, vapour 66.7 per cent, liquid 67.2 per cent [read: New Fragments p.105]. b. For a broadband source and bands that are broad in both phases, integrated absorption per molecule is similar (the oscillator strength is conserved when bands shift and broaden); that is a modern way to express the effect [memory: medium]. It fails spectrally and for strongly hydrogen-bonded liquids: water's liquid bands are shifted and broadened (O-H stretching near 3 micrometres in the liquid against sharp lines at about 2.7 micrometres in the vapour), and liquid water absorbs strongly across the 8-12 micrometre window where the vapour is nearly transparent [memory: high]. Therefore Tyndall's inference "water is the most opaque liquid, so its vapour must be the best absorber" is true in the integrated sense but hides the atmospheric window, which is what lets surface radiation escape; it also stops him from deriving the correct climate-relevant detail (windows) that later spectroscopy supplied. The vapour measured water directly in Heat, so the continuity argument is not the only support. Verdict: Plausible for the tested organic vapours; Tested, supported in the broadband sense only. I assign Plausible, untested for water.

5D. Photoacoustic "sounds from free molecules" (Bell 1880; Tyndall 1881 Bakerian) as evidence that absorbers are molecules not films [web: Investigator]. A modern descendant: photoacoustic spectroscopy is a standard trace-gas method [web: Manohar 2016, via Investigator; memory high]. Verdict: Established.

---

Claim 6. Radiant heat and light are the same ether-wave motion, differing in period (reflection, total reflection, invisible images, polarisation, double refraction, magnetisation of dark heat), and heat is a mode of motion (Rumford, Davy, mechanical equivalent 772 with Mayer's method reconstructed to 771.4)

a. Mathematics. Mayer's reconstructed value 771.4 = 2160/2.8 [calc: 771.43]; the deep read notes it becomes about 769 with unrounded intermediates and about 797 with modern gamma = 1.40 [read: claims.md]. Rumford's heat estimates: 26.58 lb of ice-cold water to boiling (180 F) is 26.58 x 180 = 4784 lb-F; in modern units about 2.2 MJ, not given by Tyndall. The Rumford/Davy "inexhaustible source cannot be material" argument is qualitative and correct as an argument against caloric. b. Modern value of the mechanical equivalent: 778 ft-lbf per Btu (per lb per degree F), about 0.8 per cent above 772, and Joule's 1850 value was 772 [memory: high]. Tyndall's 772 (a Joule value) is therefore accurate to 1 per cent; "771.4" is a reconstruction of Mayer's method, not Mayer's figure (Tyndall says Mayer's first paper gave no details, [read: claims.md]). Light and radiant heat are the same electromagnetic wave: the demonstrations (reflection, total reflection, polarisation, double refraction, Faraday rotation of heat, "150 of the lower degrees of the galvanometer") are all expected, since a polarisation-dependent device acts the same on any wavelength within its transparency range [read: Six Lectures p.206-215]. Tyndall does not mention the speed of light for heat beyond the single sentence "flies through space with the velocity of light" [read: Six Lectures p.203]; Maxwell's identification of light with electromagnetic waves was published in 1865 [memory: high]. c. Tested. Rumford and Davy 1798-1799 (Tyndall repeats them), Joule's paddle-wheel and electrical equivalents (1840s), many later experiments. The Faraday rotation of heat was tried by De la Provostaye and Desains with only two or three degrees; Tyndall claims a much larger effect [read: Six Lectures p.215], single-source in the folder. d. Changes: none. Tyndall's "ether" is a historical commitment, not a physical one now. Mild slips: the 1865 Rede lecture and Heat call the ether the substance in which the focus heat lies ("the ether at the focus, and not the air, is the substance in which the heat is embodied" [read: On Radiation p.30]); in modern terms the heat is in the radiation field and the absorber, not in an ether. Rumford's "frigorific rays" (which Tyndall rejects) are an instructive error (Pictet's mirrors show net radiation exchange with a cold object) [read: New Fragments p.175]. Verdict: Established (identity of heat radiation with light; heat as molecular motion; mechanical equivalent within 1 per cent). The ether language: Historically superseded but instructive.

---

Claim 7. Glacier ice moves without viscosity, because broken ice re-freezes at its contacts (regelation); the Ice Ages show abundant vapour and strong solar heat followed by effective condensation, not a weaker sun (Heat Ch. VI)

Split into three parts.

7A. Regelation is real: two ice surfaces at the melting point freeze together; Faraday (1850) discovered it, the name from Hooker [read: Heat p.208]. Tyndall's demonstrations (pressed-ice cups and spheres) are correct and reproducible. Verdict: Established.

7B. Strong form: glacier ice does not deform viscously (stretching tests "failed to detect this stretching power"); the motion is fracture plus regelation, so ice is not viscous. a. Internal consistency is poor: in the same chapter he concedes that under pressure ice "behaves like a viscous body" and credits Forbes with accounting for "half the facts" [read: Heat p.211-212]. The "most delicate experiments" on stretching are cited to no one and with no numbers [read: Heat p.207-208; he refers readers to his Glaciers of the Alps]. b. Inconsistent with later experiment. Glen (1955) showed that polycrystalline ice creeps continuously under constant stress, with strain rate proportional to roughly stress cubed (n of about 3), temperature-dependent with an Arrhenius factor; this "Glen's law" is the basis of glacier-flow models [web: https://data-ww3.ifremer.fr/BIB/Glen_PRSLA1955.pdf and Journal of Glaciology literature found by search; Glen's mechanistic interpretation was not seen in the search results]. Dislocation creep in the ice crystals and grain-boundary processes are the usual interpretation [memory: high]. Regelation does contribute in glacier sliding over rough beds (Weertman 1957 [memory: medium, not checked on the web]) and in healing of crevasses, so Tyndall's regelation is part of the mechanism, not the main deformation law. Ice is therefore a (non-Newtonian) viscous fluid on glacier time scales; Forbes's view that the two descriptions were "complementary" [web: https://www.sciencehistory.org/stories/magazine/under-pressure/, Forbes 1864, per Investigator] is closer to the mature picture, though that does not mean Forbes's treacle model was right in detail. c. Tested. Not by Tyndall in the folder; tested later (Glen; also borehole and surface measurements). Verdict 7B: Tested, not supported (strong form: ice is not viscous). Weak form (regelation operates and matters): Established. Note on the dispute: the Forbes-Tyndall quarrel included priority accusations (Rendu) and strong language ("utterly annihilated"); I give no physical weight to either [web: SHI, per Investigator]. Tyndall's Chapter VI arguments are his side; Forbes's side is not in the folder.

7C. Ice Age energy argument: glaciers need vapour, hence a hot sun and a good condenser, not a weaker sun. a. His bookkeeping is internally consistent: 1,000 units latent heat; iron specific heat one-tenth gives 10,000 degrees; fusing point 2,000 degrees gives 5 lb of iron per lb of vapour [read: Heat p.215-216; calc 1000 x 10 / 2000 = 5]. Latent heat of vaporisation is about 2.5 MJ/kg (about 1,000 Btu/lb) [memory: high], so the number is right in sign and order. b. The conclusion that a dimmer sun would not make more ice is partly correct (less evaporation) but not the correct modern causal story. Ice ages are mainly orbital (Milankovitch) forcing amplified by albedo and CO2 feedbacks [memory: high; unchecked on the web; the Investigator did not research Croll]. Tyndall writes "I have no new hypothesis to offer", and the "improved condenser" is never specified [read: Heat p.214, 216]. Verdict 7C: Historically superseded but instructive.

---

Weighing the Investigator's findings (kept whatever the verdict)

• Priority for gas absorption: Franz 1855, Foote 1856 preceded Tyndall; Jackson 2019 treats it as a simultaneous discovery and that Tyndall made no reference to Foote [web]. This does not change the physics. Tyndall's own Heat priority statements are his as an interested party.
• Tyndall's 1859 Note: first trial negative, then compensated method "very large effects"; no numbers and no climate [web]. The folder's popular books hold the numbers; his Phil. Trans. memoirs (not in folder) hold the full data. The reviewer cannot say which tube lengths were used in the 972 and 5,460 figures beyond what the lecture states.
• Magnus, Buff, Lecher and Pernter: the objections are real contemporary criticisms by experienced experimenters. The physics now sides with Tyndall on the sign; on method, Magnus's glass plates (secondary radiators) and vapour-hesion are plausible sources of error [web: Very 1900]. Neither side's raw data were seen.
• Rontgen 1884: independent, different method, supports Tyndall on water vapour [web].
• Stokes's letter on calorescence: crucial to claim 4B [web].
• Macdonald (viXra): recorded; physically mistaken on Raman versus infrared activity [web; memory: high].
• Forbes versus Tyndall (glaciers): Forbes's complementary view recorded (claim 7B).
• Protons for Breakfast (de Podesta commentary): not evidence, no replication found.
• Akin's silence after 1865 is a historical-record question for the Contrarian and Arbiter. The physics contribution: the phenomenon is uncontroversial; no one was proved wrong about the observation.
• The Royal Institution tube (698 mm x 85 mm, dated 1861) [web, per Investigator] is a physical artefact of claim 1-2 apparatus; a reconstruction with rock-salt windows would be the cleanest modern replication. I found no published replication of the gas tables with Tyndall's apparatus; the closest tests are the spectroscopic ones above.

Summary table