Rahn-Otto: Arbiter (Step 3)

Project: Otto Rahn, Invisible Radiations of Organisms (Protoplasma-Monographien IX, 1936; physics chapter by Sidney W. Barnes). Inputs weighed: physicist.md, mathematician.md, contrarian.md, claims.md, facts.json, deep/, sites.md, research/. Tags: [read] from the extracted book; [read: page image p.N] rendered by me (PDF page; book page = PDF page minus 14); [calc] my own arithmetic (Python, scratch only); [web] with URL; [memory] with confidence.

Units used: 1 A (angstrom) = 0.1 nm. Photon energy E = hc/lambda, i.e. E(eV) = 12398 / lambda(A). 1 eV per molecule = 23.06 kcal/mol = 96.5 kJ/mol. "PE" is probable error, 0.6745 standard deviations.

Source gap note

The book is the only document in the project folder (230 PDF pages, text extracted, 9 pages OCR'd). The deep read covers PDF pp.44-67 only. Claims 4-7 and parts of claim 1 rest on digests plus passages that the three reviewers read themselves in the extracted text, and on page images (pp.42, 48, 97, 126, 157 by me or the reviewers). Everything outside the book is known only through the Investigator's notes (research/files/ is empty): Rahn 1934 (J. Bacteriol.; CSH Symposium), Rahn 1937 (Tabulae Biologicae), Hollaender and Claus 1937 (NRC Bulletin 100), Quickenden 1974 and 1989, Wainwright 1997 and 1998, the Soviet "cancer quencher" dissertations, Gerlach 1939, Stempell 1932. No reviewer and not the arbiter has read any of those in the original. Wherever a verdict leans on them it says so.

Corrected claim list

Changes to claims.md (the claims stand; wording is corrected):

1. Claim 1: add the intensity range. Rahn gives 100-1000 quanta/cm2/s for the strongest sender [read: p.137-138]; the 1937 Nature item defines mitogenetic rays as 1900-2500 A at 10-1000 quanta/cm2/s [web, via Investigator: https://www.nature.com/articles/1401007c0]; Rajewsky's physical estimate was 10-100 at 2300 A [read: p.107].
2. Claim 2: add "as shown by a biological detector (yeast, bacteria)". Neutralisation and NaCl dissolving rest on one group (Wolff and Ras), with no photoelectric measurement and no error estimate [read: deep p.47, p.64-65]. Oxidations are the case with counter reports (Frank and Rodionow, Audubert and van Doormal, Gerlach, Barth).
3. Claim 3: "unresolved" describes Rahn's position. Say instead: Rahn's premise (one quantum per reacting molecule) and his reabsorption-chain rescue both fail on his own numbers; he also offers a detector-side route (a single quantum triggers a cell process the cell powers, [read: deep p.62]); the origin "is by no means clear" [read: p.55].
4. Claim 4: the nerve speed is 30 +/- 3 m/s, not 30.3 m/s (facts.json summary). The text layer reads "303"; the mathematician and contrarian each read the rendered page [read: page image p.126]. The onion-root speed (about 30 m/s) has no stated error.
5. Claim 5: add the exceptions the book states. Gesenius names pernicious anaemia, leukaemia, carcinoma and severe sepsis as conditions with absent blood radiation; "occasional exceptions have been observed"; teratomas never show it; sarcoma, hypernephroma, glioma and myoma show normal blood radiation, which Rahn calls "strange" [read: p.193-195, via contrarian and mathematician].
6. Claim 6: "Barnes and Rahn 1933" is not the physicist S. W. Barnes. The book thanks "Mrs. Margaret N. Barnes" and lists "Rahn and M. N. Barnes, 1933, criteria of death in yeast, J. Gen. Physiol. 16:579" [read: preface and author index, per physicist and contrarian]. Probably M. N. Barnes; the text does not say outright which Barnes did the oxy-cholesterol work. Her being a Cornell student is the Investigator's lead from a Macht citation [web, via Investigator], not in the book.
7. Claim 7: replace "the physical counter evidence ... and its critics" with "the physical evidence from photoelectron counters (Rajewsky; Frank and Rodionow; Siebert and Seffert) and its critics (Lorenz's three error sources)". "54 failures in 373" needs its context: Gollsheva (Gurwitsch's laboratory); failures were recognised because all associates using the same yeast culture on the same day got negatives [read: p.107, via contrarian and mathematician].
8. facts.json mainstream_conflict: visible and near-UV chemiluminescence is ordinary; 190-260 nm emission from reactions releasing under 1 eV per event is not.
9. (QE revision) Claim 1: add the internal wavelength dispute. Reiter and Gabor (1928) put the effective band at 3200-3500 A (maximum 3400 A, second maximum 2800 A, "antagonistic" 2900-3200 A); Gurwitsch's school found 1900-2500 A and repeated 3340-3400 A with no effect (Frank 1929; Chariton, Frank and Kannegiesser 1930; Table 18) [read: p.73-76]. Rahn calls the Reiter-Gabor results "never accounted for in a really satisfactory way" [read: p.74]. Also add Rahn's own concessions: the percentage "induction effect" "has no quantitative value whatever" and polarisation makes "all previous measurements of intensities ... practically meaningless" [read: p.128].
10. (QE revision) Claim 4: add the nerve and brain work: Kalendaroff's nerve spectra (with Rahn's own calibration-shift footnote), Anna Gurwitsch's optic-nerve, optic-lobe and colour experiments (Table 49: three isolated heads, two living animals; Rahn notes colour intensities and wavelengths are "not mentioned"), and Latmanisowa's parabiosis (more than 40 repeats; controls named: gastric juice without protein, and the unexposed part of the same nerve; no quartz/glass screen named) [read: p.168-174].
11. (QE revision) Claim 6: the fingertip kill times are 5 minutes (fingertip over an open glass ring, no glass or quartz between finger and yeast) and 15 minutes (through 2 mm quartz sealed with lanolin) [read: page image p.111]. Tadpole limb: net gain 10.6 +/- 1.4% (22 glass and 33 quartz tadpoles) and 12.1 +/- 1.8% (44 and 51) [read: p.182-183]. Oxy-cholesterol Table 53a: on day 1, batches I, II and V dead, IV "mostly dead", III normal (stimulated); batches I and II were not read on days 2-3 [read: page image p.200].
12. (QE revision) Claim 7: add Rahn's rebuttal of Kreuchen and Bateman (one of their series said to equal 140 single Gurwitsch experiments: "That is a mistake" because failure days cannot be identified in advance) and the ten recorded detector failures [read: p.107-109].

Numbers recomputed where reviewers differed or did not check each other

Corrected claims (final wording)

1. Living cells, in a particular physiological state, emit extremely weak light; the book's version has it in the far ultraviolet, 1900-2600 A (4.8-6.5 eV), at 100-1000 quanta/cm2/s from the strongest senders, and says it speeds division in other cells that absorb it.
2. The book attributes the light to ordinary biochemistry (oxidation, glycolysis, proteolysis, nucleic acid and phosphate cleavage, neutralisation, dissolving NaCl), each said to have its own narrow-band spectrum, mostly read with yeast or bacteria as detector and, for oxidations, with counters.
3. The energy accounting: one 2020 A quantum per sucrose molecule would be 420 cal/g, 43 times Rubner's measured 9.7 cal/g. Rahn proposes an activation-chain reabsorption and a detector-side single-quantum trigger, and says the origin "is by no means clear".
4. Irradiated solutions and tissues re-emit light ("secondary radiation"), reported to travel along onion roots at about 30 m/s (no stated error) and along frog sciatic nerve at 30 +/- 3 m/s, which Rahn calls very suggestive of a part in nerve conduction, not proof.
5. Blood of healthy people radiates and blood of cancer patients usually does not, malignant tumour tissue radiates strongly, so blood radiation may serve as a cancer test (exceptions admitted in the book); cancer may be a radiation disorder (Protti, with Roffo's UV rat tumours cited as a possible link).
6. The radiation influences development and health (tadpole limbs and wounds, sea-urchin larvae, yeast forms, crowding effect), and a harmful form from fingertips, eyes, menstrual blood and oxy-cholesterol kills a sensitive yeast through quartz.
7. Failed replications do not disprove the effect, because the detectors respond only in a particular physiological stage and fail for days or weeks (54 failures in 373 Gurwitsch blood experiments); biological error is large; physical counter evidence exists and has three named error sources (Lorenz).

Claim-by-claim arbitration

Claim 1. Cells emit short UV that speeds division in other cells

Agreement. All three: weak photon emission from living matter is real (ultra-weak photon emission, UPE, from the 1950s on); Rahn's intensity (100-1000 quanta/cm2/s) is the same order as modern UPE figures; the mathematics that exists is right (single-quantum hit probability 4.2e-4 per second per yeast cell at 1000 quanta/cm2/s, 40 minutes for a mean of one hit; at that point 63% of cells are hit, not "every cell", as the mathematician and contrarian note, and Rahn himself drops the single-quantum idea [read: p.138]); the biological detector gives no absolute intensity [read: deep p.59]; no blinded, protocol-matched, independent replication exists.

Disagreement. The strong form. Physicist: Tested, not supported (NRC 1937, Quickenden 1989). Mathematician and contrarian: Never properly tested, with the negative tests contested.

Stronger argument. The contrarian's, with the physicist's evidence kept in. The NRC study (Hollaender and Claus 1937) was funded, two years long, used biological and physical detectors, and reported "all the attempts have yielded negative results" [web, Investigator: https://www.nature.com/articles/1401007c0; abstract only; I could not open it, the site redirected to a login]. That is the heaviest single item against the strong form. But the proponents' objection is specific and checkable, and I confirmed the wording on the open-access page: Hollaender and Claus used "agar cultures 15-19 h old (they later shifted to cultures 39 h old)", "an inorganic salt solution" of NaCl, KCl, CaCl2 and water as medium ("No other worker tried such a specific medium for induction"), and "constant stirring" and "rapid twirling" of the recipient culture; of Quickenden et al. the authors write that "the authors mentioned the presence of day light as a necessary condition for MGE in early works, but preferred to violate this recommendation"; and of Wainwright et al. 1997 that they "obtained a good effect. Unfortunately, they could not make it reproducible" [web: https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00241/full; first 100,000 of 185,000 characters read by a fetch tool]. Those authors (Volodyaev and Beloussov, Moscow, Gurwitsch tradition) defend the field; their account is a party's characterization, and it is specific. Neither side's protocol was agreed beforehand. I did not find a 1989 Quickenden abstract (my search returned none), so what that paper found is known only from the proponents' paper and Naumova 2018's reference list.

Positive evidence stays in the record: Gurwitsch school (dozens of papers 1923-1934), Paul 1933 (onion roots, steel-needle control), Julius 1935 (56 pairs, 1.92 +/- 0.15 against 1.01 +/- 0.08; about 3.6 sigma if PEs), Schwemmle's analysis of about 200 onion-root results (error +/-10% at 500 mitoses; pure counting statistics gives sqrt(2/250) = 9% [calc]), Rahn's own group (Ferguson and Rahn; Tuthill and Rahn), and the partial 1997 UK result. The book's own limits on that evidence: Schwemmle classified roots as induced or not before computing the error, and "did not consider it proved that the effect is caused by mitogenetic rays" [read: p.72]; the controls "are not perfect" [read: p.203; tag corrected from book p.189]; only Table 15 (one unpublished run) has a same-time quartz-against-glass control among the deep-read pages. Against the proponents: Taylor and Harvey (photographic plates, 90 days, no effect, [read: p.104]); Lasnitzki and Klee-Rawidowicz, Doljanski (fibroblasts, no effect, [read: p.97]); Kreuchen and Bateman; Nakaidzumi and Schreiber (via Investigator).

Dismissed without testing? No: it was tested by both sides. Was it properly settled? No: the only neutral, well-resourced negative is contested on protocol, and the positive results are unblinded and mostly lack error estimates.

What would settle it. One experiment: a pre-registered, blinded, two-laboratory test. The proponents fix beforehand the culture age, medium, detector layer, induction time, and a physical or calibrated positive control (a UV-C source at 10-1000 quanta/cm2/s through quartz) that declares a "responsive day"; exposed and sham are assigned blind, with quartz, glass and paraffin-insulated glass screens; analysis by day. Alongside it, a modern UV-sensitive photon counter (quartz optics, photomultiplier) on lag-phase yeast and onion roots, 190-300 nm, with blanks, for the emission half.

Final verdict (split). - Weak form, ultra-weak light from living matter (near-UV to near-IR): Established. - Emission at 190-260 nm at 10-1000 quanta/cm2/s: Never properly tested (no neutral 190-260 nm spectrum on record; modern UPE work does not establish UV-C emission [memory, medium]; Quickenden and Que Hee 1974 is the precedent, not read). - Division-accelerating effect as a reproducible phenomenon: Never properly tested under a protocol both sides accepted, with a contested negative (NRC 1937) and contested partial positives on record. Not dismissed without testing. A reader should not read this as support: the physicist's "Tested, not supported" is the better description of the NRC test taken on its own terms, and the proponents' objection to it is a party's account that has not been put to a test.

Claim 2. Ordinary biochemistry emits it; narrow-band fingerprints

Agreement. Oxidations emitting weak light is established chemistry (Haber's pyrogallol and phosphorus [read: deep p.47]). Neutralisation and NaCl dissolving have no energy for a 5-6 eV photon (physicist, mathematician, contrarian all compute it). Fingerprints were read only with biological detectors, 10 A scans only where 50 A scans were positive, no blank (null) spectrum, no stated false-positive rate. Rahn says "we are not dealing with true spectral lines, but with relatively broad regions" [read: p.53].

Disagreement. Physicist: strong form Refuted by calculation (neutralisation, NaCl, hydrolysis). Mathematician: oxidation Plausible; neutral/hydrolysis one-quantum-per-event Refuted. Contrarian: neutralisation and NaCl Never properly tested (one group, no error bars); fingerprints Never properly tested.

Stronger argument. A per-event calculation refutes a specific mechanism: that the 0.58 eV released by one neutralisation (or the roughly zero net for NaCl dissolving, or 0.14 eV for sucrose hydrolysis) is emitted as a 6 eV photon. Energy conservation per transition makes that impossible; this is Lorenz's objection and the physicist is right on it. It does not refute the report that the bacterial detector responded when those operations were done in front of a quartz plate; it refutes the stated source. The contrarian is right that the observation itself was never checked physically. Both are kept: the mechanism is refuted by calculation; the observation is untested; alternative explanations (heat through a thin quartz plate, other transmitted agents, electrical effects such as Magrou and Reiss proposed for sea-urchin larvae, and biological noise) are not excluded in the pages read. In Table 14 the controls begin at about 7,000 cells/cc against 21,750-35,500 for NaCl (different inocula), and palmitic acid shows +21% yet counts as "no effect" (physicist, [read: deep p.64-65]).

Fingerprint statistics (mathematician), checked: five or more positive strips out of 60 would occur by chance 3% of the time at a 3% per-strip false-positive rate, 18% at 5%, 73% at 10%, 96% at 15% [calc]. So the fingerprint claim is meaningful only if the same strips reappear in independent runs; the book does not show that. Several Russian workers (Kannegiesser; Braunstein and Potozky; Ponomareva) agree on the 2200-2350 A oxidation region, but they are one school with one detector.

Counter evidence (Table 10): six of nine rows exceed 2 standard errors, four exceed 3 (strongest 6.1, the dark FeSO4 + K2Cr2O7 row), the last is null (0.4). Barth's 24 of 25 runs above control (p = 7.7e-7 if independent). Lorenz's three artefacts (water vapour condensing on the quartz, charge on the quartz, induction-coil stimulation of muscle) are stated by Rahn, who concedes none of the physical measurements is "absolutely certain" to have excluded them [read: p.105-106]. Counter detection limit about 100-500 quanta/cm2/s means the signals are near the floor of the method (see recomputed table). Hollaender and Claus 1937 report negative physical-detector results; Gurwitsch and Gurwitsch 1939 (Nature 143:1022, "Ultra-Violet Chemi-Luminescence") and Audubert 1938 (Angew. Chem. 51:153) report positive ones [web, titles via Investigator]; not read.

What would settle it. A photon-counting spectrometer on the named reactions: FeSO4 + dichromate, pyrogallol oxidation, sucrose + invertase, NaOH + HCl, NaCl dissolving, each with a blank, 190-300 nm, quartz window, vapour trap and heat control; then a blind repeat of one fingerprint (blood glycolysis, 5 of 60 strips) in three independent runs to see whether the same strips recur.

Final verdict (split). - Weak form, some oxidations and radical reactions emit faint near-UV or visible light: Established. - Oxidations emitting 190-260 nm light at the book's intensities: Plausible, untested (several 1930s counter reports with unresolved artefact objections; NRC negative). - Neutralisation, NaCl dissolution and sucrose or protein hydrolysis as one-photon-per-event sources of 190-260 nm light: Refuted by calculation (0.04-0.6 eV per event against 4.8-6.5 eV). The detector responses reported for them: Never properly tested (one group, biological detector, no error bars, uncontrolled alternatives). - Reaction-specific narrow-band fingerprints: Never properly tested (no false-positive rate, no physical spectrum, no independent replication on record).

Claim 3. The energy accounting

Agreement. Rahn's arithmetic is right; 43, not "one-fiftieth". His activation-chain rescue fails on his own numbers (2768 A). Measured activation energies are lower still: acid hydrolysis of sucrose about 26 kcal/mol (109 kJ/mol), invertase about 7 kcal/mol [web, search snippets: https://pmc.ncbi.nlm.nih.gov/articles/PMC2664835 and https://www.researchsquare.com/article/rs-662668/v1; textbook-style pages gave 62.2 kJ/mol, so acid values vary]. In every case far below 141.6 kcal/mol. Activation energy is borrowed and returned, not net energy.

Disagreement. Physicist: Lorenz's objection Established; the chain rescue Refuted by calculation; the paradox "resolved" against the strong form of claim 2. Mathematician: the paradox itself is Refuted as a real conflict (it exists only at quantum yield 1) and the chain Internally inconsistent. Contrarian: the quantum-per-molecule premise and chain Refuted by calculation; rare-event weak emission Plausible, untested.

Stronger argument. All three are partly right and the verdicts attach to different statements. The mathematician and contrarian add the point the physicist's summary under-weights: the total energy budget costs nothing at the observed intensities (1e-12 of the heat power; yield 2.6e-14 per reacting molecule [calc]). The physicist's point is the per-event constraint, which the mathematician also holds. The combination is what stands: total budget fine; per-event energy for a hydrolysis photon unexplained; Rahn's remedy fails. Rahn's text does not say the yield is 1; Lorenz's objection (as Rahn presents it, "each reacting molecule must emit at least one quantum" [read: p.55]) does. A rare-channel emission would not be "sucrose hydrolysis emitting"; it would be a side reaction, radical recombination or impurity, which removes the claimed specificity (physicist). Rahn's detector-side route (one quantum triggers a process the cell powers) addresses a different problem and is untested.

What would settle it. Nothing new is needed for the premise: the per-event arithmetic decides. For the rare-channel version, measure the absolute photon yield per reacting molecule of sucrose + invertase with a UV photon counter and compare with 2.6e-14.

Final verdict (split). - Each reacting molecule emits a 2020 A quantum (Lorenz's premise, the source of "43 times"): Refuted by calculation. Lorenz's budget objection as stated: Established. - Rahn's reabsorption chain with wall leak: Refuted by calculation on his own numbers (and Internally inconsistent: it needs 138 kcal/mol where he assumes 100). - Rare-event weak emission at the book's intensities from a reaction like sucrose hydrolysis: Plausible, untested (budget satisfied; per-event energy source not identified). - Rahn's statement that the origin is "by no means clear": accurate.

Claim 4. Secondary radiation, 30 m/s, nerve conduction

Agreement. The disc arithmetic is correct. 30 m/s is ten million times slower than light, so it is a relay, not free flight. The root figure is a difference of two coarse angular windows; the nerve figure is 30 +/- 3. Frog sciatic nerve conducts at about 20-40 m/s [memory, medium-high], so the match is real arithmetic but weak evidence. The action potential (Hodgkin and Huxley 1952 [memory, high]) explains conduction without light. Rahn says the propagation claims have only one published experimental proof (de Korosi 1934), that "it is not at all certain" that secondary-radiation intensity increases "in the physical sense", and that it has not been possible to produce muscle contraction by irradiating a nerve [read: deep p.59; p.173-174, per contrarian].

Disagreement. Physicist: speeds Never properly tested; nerve role Historically superseded. Mathematician: delays Plausible, untested; light-relay explanation Never properly tested; nerve role Historically superseded. Contrarian: Never properly tested.

Stronger argument. Mathematician and contrarian: the measurement does not distinguish "light relays the impulse" from "the impulse emits light" or from any chemical or electrical wave carried along the tissue. That is the central point and the physicist's sentence that many processes fall near 1-100 m/s says the same. Parabiosis (nerve blocked) after 2 hours' irradiation "repeated more than 40 times" [read: p.173-174] and Lapitzki's mercury-lamp result (same effect in minutes) are what a UV effect on nerve would do whatever the mechanism. Energy: the secondary photons (2450-2500 A, 5.0 eV) exceed the primary (3230 A, 3.8 eV); the re-emitter would supply 1.2 eV per photon with no budget given, and the beam crosses a glass trough that would absorb the 2450 A output [read: deep p.57]. The "27 times" amplification is a time-to-effect (270 s against 10 s), not a measured photon gain.

What would settle it. A pulsed UV source and photomultiplier at two points along a root or nerve at known temperature, to show a threshold-limited relay with delays independent of the instrument's window width; and a test of whether blocking the impulse (tetrodotoxin or cold) removes the light, which separates "impulse emits" from "light conducts" [memory, medium: standard pharmacology; not in the reviews].

Final verdict (split). - The measured speeds (about 30 m/s in root, +/-30% at best; 30 +/- 3 m/s in nerve) as a property of a UV relay: Never properly tested (one laboratory, biological detection, no independent replication; not dismissed on refutation). - Secondary radiation takes part in nerve conduction: Historically superseded but instructive (the ionic mechanism is directly measured; the optical hypothesis was reasonable in 1934 and never given a decisive test).

Claim 5. Blood radiation and cancer

Agreement. All three: Never properly tested as a diagnostic; no sensitivity or specificity numbers in the book; no blinded or prospective trial on record; exceptions admitted (sarcoma, hypernephroma, glioma and myoma show normal blood radiation; Rahn's gloss "benign" for these is medically wrong [memory, high]); the book itself concedes the blood-to-tumour explanation conflicts with unchanged glycolysis in cancer patients; "radiation disorder" is not testable as stated. UV light does cause skin cancer, through ordinary UV, and Roffo's rat and skin data are consistent with that (UV in 4 months in all of 150 rats; 52% of 700 in sunlight at 10-11 months; none under tungsten [read: p.197; tag corrected from book p.183]); Rahn says sunlight lacks the mitogenetic range and the link "speaks very much against it" [read: p.197-198; tag corrected from book pp.183-184].

Disagreement. None on verdicts. The mathematician adds that Gesenius's rule "healthy blood never fails; if a failure occurs, test the yeast or the apparatus" [read: p.99] means a null is read as disease or detector failure by the same observer, safe only if blinded; no blinding is described. The illustrative calculation (70% sensitivity, 90% specificity, 1% prevalence gives about 7% positive predictive value [calc]) is an illustration, not Rahn's data. The book's Roffo skin-site percentages sum to 100.2% (contrarian [calc]): immaterial.

Soviet continuation (kept in the record). The "cancer quencher" work (Lydia Gurwitsch and Salkind 1920; Pesochensky's 1942 dissertation; The teaching of the cancer quencher 1947; dissertations 1950 and 1955; A. A. Gurwitsch 1968, 1974) and Naumova et al. 2018 (about 190 references, who report "very promising results" and a peptide quencher, and describe the research as "interrupted and almost abandoned") [web, via Investigator: https://jbpe.ssau.ru/index.php/JBPE/article/view/3308]. These authors are in the Gurwitsch tradition (an interested party); none of these were read by anyone here. No neutral test is on record in either direction.

What would settle it. A prospective blinded trial of coded blood samples (cancer, benign disease, healthy) with the detector protocol fixed in advance and receiver operating characteristic analysis; with a counter-based arm using Siebert and Seffert's design (several hundred normal bloods, KCN-blood control) and a modern photon counter.

Final verdict (split). - Blood radiation as a cancer diagnostic: Never properly tested (advocates' clinical series; not tested by neutral parties; not shown to have been tested and failed; whether it was dismissed without testing in the West is not documented in anything read; it simply was not taken up). - Cancer as a radiation disorder: Not testable as stated. - Ultraviolet light causes skin cancer (Roffo): Established, for ordinary UV, not shown for the mitogenetic band.

Claim 6. Development, health, harmful human radiation

Agreement. Never properly tested for the developmental effects (small unblinded series from one school; tadpole forelimb about 4-10%; tail wounds 26-33%; 56-pair and 500-tadpole series); no replication outside the school in the record. The sea-urchin larva result is the book's own cautionary tale: Magrou and Reiss found two glass layers with paraffin between them transmit the effect and proposed an electrical mechanism, so the quartz-passes-glass-blocks test did not discriminate UV from electricity there [read: p.100-101]. Oxy-cholesterol: four of five batches killed Mycoderma through quartz on day 1; the fifth is explained after the fact as delayed by a day while the product loses power; no glass-covered comparison in Table 53a (contrarian [read: p.199-200; tag corrected from book pp.185-186]); Rahn himself says "it cannot be considered proved that the effect is of physical origin" [read: p.198; tag corrected from book pp.184-185]. Human radiation: the book is more careful than the press. It occurred "only very rarely"; the most pronounced case was a man recovering from herpes zoster (kill through quartz in 15 minutes for about 6 months); one hypothyroid patient tested once; "only one especially sensitive species of yeast could be killed"; the radiation "does not reach further than a few inches"; the "evil eye" publicity was not wanted by the authors [read: p.110-111, via contrarian]. The 1932 press account (New York Times, 25 June 1932; Papers Past reprints of the Science Service wire; "in five minutes", right hand stronger, "preliminary stages") [web, via Investigator] is newspaper wording of a talk, not Rahn's text; the headline "Evil Eye is Found to be Based on Fact" overstates the book. Rahn's finger-over-yeast result is 5 minutes with the fingertip held over the drops in an open glass ring and 15 minutes through a 2 mm quartz plate [read: page image p.111; the text layer's "1.5 minutes" is an OCR error, withdrawn here after the QE and the new deep read]. That order (slower through quartz) is what a radiation-or-vapour story predicts; the earlier remark that a faster kill through quartz is anomalous is withdrawn.

Disagreement. Harmful human and oxy-cholesterol radiation: physicist adds Refuted by calculation at the book's intensities; mathematician and contrarian say Never properly tested.

Stronger argument. The calculation is valid and conditional. At the book's own mitogenetic fluxes, 100-1000 quanta/cm2/s at 2020-2300 A, 15 minutes delivers 8e-10 to 9e-9 J/m2, against a UV-C lethal dose for yeast of order tens to hundreds of J/m2 [calc; web as in the table, search summaries only]: a factor of 10^10 to 10^11. So if the fingertip effect were radiation of the book's intensity, it could not kill. The calculation cannot reach the further statement that skin emits something far stronger; Rahn gives no measured flux for fingertips. The refutation therefore applies to "mitogenetic radiation at the book's intensity kills yeast" and not to "something from fingers kills Mycoderma through quartz", which is a separate chemical or thermal question that was not neutrally tested (the quoted sentence "most experiments lacked chemical exclusion from volatile substances" [read: p.175] concerns Christiansen's menstrual-blood work, not the fingertip tests; for the fingertip quartz arrangement Fig. 38 shows the quartz plate sealed with lanolin to a glass ring [read: page image p.111], so vapour is blocked if the seal held, while the open-ring 5-minute result excludes nothing; see QE revision).

What would settle it. Repeat the fingertip and oxy-cholesterol experiments with a sealed quartz window, a sealed glass control and a paraffin-insulated glass control, a heat control, a vapour trap and blind scoring; measure the absolute UV flux at the window with a photon counter. For the developmental claims, a blinded replication of Blacher's tadpole limb and wound experiments with quartz and glass lids.

Final verdict (split). - Mitogenetic UV speeds tadpole limb growth and wound healing and links neighbouring cells; sea-urchin larva distortion through quartz: Never properly tested (one school; tadpole limb gain about 11-12% net of an operation control, the best-designed biological series in the book, unreplicated; mechanism disputed by the original authors for the larvae). - Harmful radiation from fingertips, saliva, menstrual blood and oxy-cholesterol as UV radiation of the book's intensity: Refuted by calculation. - That something from skin or oxidised cholesterol kills a sensitive yeast through quartz: Never properly tested (small series, post-hoc potency window, no tabulated glass control for oxy-cholesterol or fingertips; vapour excluded only by a lanolin seal whose integrity is not tested; heat and moisture through quartz not excluded). The press version ("evil eye proved") is not supported by Rahn's own text.

Claim 7. Failed replications do not disprove the effect

Agreement. Biological detectors depend on culture age, medium, temperature and season (textbook microbiology; Rahn 1934, J. Bacteriol. 28:153-158 states exactly this, [web, via Investigator: https://pmc.ncbi.nlm.nih.gov/articles/PMC533662/]; fetch blocked for the reviewers; not read). Rahn's statistical point is sound: same-day replicates share causes of error, so the effective sample is the number of days. 54/373 = 14.5% [calc], about the 15% yeast-bud increase Gurwitsch accepted as the error threshold [read: p.204]. Failures were recognised after the fact by same-day negatives in the same laboratory, so the exclusion depends on the outcome unless a positive control is fixed in advance. The book's own admissions (Heinemann's complete loss of reaction, Wolff and Ras losing days, Rahn's own long negative periods, "investigators avoid discussing failures due to physicist skepticism" [read: p.107-109]) are as much a reason outsiders could not trust the positives as the negatives were reasons to doubt them.

Disagreement. Wording only. Physicist: state-sensitive detectors Established in general microbiology, Plausible, untested for this effect. Contrarian: Plausible, untested as an explanation for the NRC null; strong form Not testable. Mathematician: Not testable in general form; Never properly tested in the specific form.

What would settle it. An independent laboratory runs the protocol Rahn stated (24 h raisin agar, 30 C [read: p.79-82]) with the validity criterion declared before each run (a known sender producing a prespecified effect), blind exposure, results reported by day.

Final verdict (split). - Detector state sensitivity as a general fact: Established (general microbiology). - As the explanation for the NRC and other null results: Plausible, untested. - As a blanket defence ("no negative can count"): Not testable as stated.

Absence-of-evidence analysis: the record goes quiet after 1937-1939

The anomaly. About 600 papers on mitogenetic rays appeared by 1937 [web, Investigator: Nature 140:1007]. The same year the NRC published its two-year negative bulletin. Western publication then falls away (Gerlach 1939, "Existiert die mitogenetische Strahlung?", Naturwissenschaften; Koyenuma 1947 "On the existence question of so-called mitogenic radiation", Biochem. Z. 318:1-14 [web, via Investigator]). Soviet work continues to at least 1955 and A. A. Gurwitsch's reviews of 1968, 1974 and 1988 (Experientia 44:545-550). Photon-counting studies of UPE resume with Colli and Facchini 1954, Quickenden and Que Hee 1974 and Popp from 1976, under different names. Rahn's own publications on the subject also stop after 1937 (Tabulae Biologicae 14) as far as the Investigator found; he published Microbes of Merit in 1945 and died 26 September 1957 [web, via Investigator: https://hdl.handle.net/1813/18707, not read]. A 1944 printing of the book exists in HathiTrust, which says the 1936 book stayed in demand.

Competing explanations, with evidence for and against and the record that would decide.

1. The NRC negative was accepted as decisive. For: NRC standing, two years' work, both detector types, the Nature item, Gerlach's 1939 title. Against: the specific protocol objections above, which show the study used conditions the positive laboratories had not; supporters never agreed it was decisive. Deciding records: the bulletin text and its protocols; Rahn's reply (Tabulae Biologicae 14, 1937); reviews of the 1944 printing.
2. The biological detector was seen as unreliable. For: Rahn's own admission (14.5% failures, unexplained days or weeks, Heinemann's lost reaction). Against: the proponents say the reliability could be managed with strict protocol (Volodyaev and Beloussov credit Rahn with optimising the bacterial protocol). Deciding record: day-level raw data from any laboratory.
3. Instrument change. For: photomultipliers made direct measurement of weak light possible and the field relabelled itself UPE; Volodyaev and Beloussov date "standard and irrefutable registration" to the photomultiplier era [web: Frontiers 2015]. Against: the UV-C division-stimulating effect itself was not carried over; UPE work measures visible and near-UV. Deciding record: a 190-260 nm spectrum of yeast and onion root with a modern counter.
4. Language, war and political division. For: Russian-language literature, the 1930s purges, war, Cold War; Soviet continuation. Against: no direct evidence that Soviet results were blocked from Western journals; some Western work (Colli, Quickenden) knew the literature. Deciding record: translation and citation analysis, interviews and archives.
5. Suppression of results or funding. For: nothing found in the record. Rahn complains that English papers were few and mostly negative, and that failures were not published "because there is still considerable doubt among physicists" [read: p.9-11; p.107-108]. That is a stated bias toward under-publication of negatives and of positives alike, within the field, not documented external suppression. Against: the NRC study was funded and published; Rahn's book was published in a German monograph series and reprinted. Silence is not contradiction, and the people involved worked partly outside institutions; the record of dissent is thin by nature. Deciding record: Cornell Rahn papers and the Dept. of Bacteriology records (Rare and Manuscript Collections), Science Service records (Smithsonian), NRC correspondence, Hollaender's papers.
6. Interest faded because the field stayed unproductive without a decisive test (a variant of 1 and 2).

Those are not mutually exclusive; 1-3 are best supported by what is on record, and 4 is a live contributor. 5 is not excluded and not evidenced; the records that would decide are named and none has been obtained.

What would settle it (summary)

Records not yet obtained that would help: Hollaender and Claus 1937 (NRC Bulletin 100); Rahn 1934 (J. Bacteriol.; CSH Symp. 2:226); Rahn 1937 (Tabulae Biologicae); the 1944 printing; Quickenden 1989 and Wainwright 1997 original text; Gerlach 1939; Naumova 2018 and the Soviet dissertations; the Cornell memorial statement and Rahn papers; the Science Service wire of 1932; Siebert and Seffert, Frank and Rodionow, Barth and Lorenz in the original.

Open questions

1. Did the NRC study use the detector conditions the positive laboratories require? The proponents say no; the bulletin has not been read here.
2. Is there any 190-260 nm emission from yeast, onion root or blood at 10-1000 quanta/cm2/s? Modern UPE work is visible and near-UV; no UV-C spectrum is on record in what was read.
3. Which of the two printed counter-limit numbers in Barnes's chapter (1e-9 erg/cm2/s or 500 quanta/cm2/s) is right, and what did the cited counters actually achieve?
4. What produced the detector responses to neutralisation and NaCl dissolving, if not a 6 eV photon from the reaction (heat through quartz, vapours, noise)?
5. Does the nerve-speed match reflect an optical relay or light emitted by the travelling impulse?
6. [Withdrawn after QE: the fingertip kill was 5 minutes over an open ring and 15 minutes through quartz, so there is no "faster through quartz" anomaly.] Open instead: did the lanolin-sealed quartz chamber stay sealed, and was heat or moisture transmitted through the 2 mm quartz?
7. What did the Gesenius, Heinemann and Siebert cancer series count as true and false positives and negatives?
8. Is the "Barnes and Rahn 1933" co-author M. N. Barnes, and what was her role?
9. Did Rahn change his view after 1937?

Web verification log (budget 8)

1. Yeast UV-C lethal dose [web, search summary: https://eom.ifa.md/m/journal/download/1837 ; https://www.nature.com/articles/211310b0.pdf]: values reported but the tool's unit conversions were wrong (corrected above); the safe statement is "tens to hundreds of J/m2". Retagged from [memory, medium] to [web, summary only].
2. Quickenden 1989 [web search x3]: no abstract found. A Nature 216:169 (1967) item appeared in results (https://link.springer.com/article/10.1038/216169a0); I did not open it, so I do not attribute findings to it. Result of Quickenden 1989 remains unconfirmed by me; the proponents' account (daylight condition violated) is confirmed on the open-access page.
3. Enthalpies: neutralisation -57 to -58 kJ/mol confirmed (https://en.wikipedia.org/wiki/Enthalpy_of_neutralization); NaCl dissolving: search summary gave +5.8 kJ/mol average with variation; stated as "about zero to slightly endothermic".
4. Volodyaev and Beloussov 2015 [web: https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00241/full]: quotes on Hollaender and Claus, Quickenden and Wainwright confirmed (first 100,000 characters).
5. Nature 140:1007 [web: https://www.nature.com/articles/1401007c0]: fetch redirected to a login page; relies on the Investigator's reading of the abstract (research/2026-10-06-nature.md).
6. Not re-checked (medium confidence, used as context only): frog nerve conduction speed 20-40 m/s [memory, medium-high]; tetrodotoxin as a conduction blocker [memory, medium]; modern UPE not reporting 190-260 nm emission [memory, medium].

Page images read by the arbiter: PDF pp.42, 48, 97, 157, in /tmp/rahn-arb/ (scratch).

---

QE revision (Step 4b remediation)

Date: 2026-10-06. Inputs weighed: qe.md (verdict MINOR), the eight new deep reads appended to deep/Otto-Rahn-Invisible-Radiations-97d8ea.md (PDF pp.68-91, 92-103, 104-116, 117-132, 133-146, 160-174, 175-191, 191-200), and the three reviews as before. Tags as in the header; [calc] is my own arithmetic (Python, scratch only). Result: no headline verdict changes scale value. Wording of several sub-verdicts changes, one factual error is withdrawn, new sub-claims receive verdicts, and material the QE listed as missing is restored below. The book is still the only primary document; every experiment below is Rahn (1936) reporting other people's work unless stated, and the places, laboratories, replicate counts and the meaning of "+/-" are often not given. Rahn is a proponent and an interested party (his laboratory made some of the tested claims); the Gurwitsch school (Leningrad) is the source of most positive biological results.

1. Errors withdrawn or corrected

2. Pertinent material the QE found missing, now weighed

All are in the record whatever they favour. Verdicts use the Step 2 scale. "Source" is Rahn's account unless stated.

A. Reiter and Gabor: a different band (claim 1). Reiter and Gabor (1928, Germany; first names and institution not given) found the effective rays at 3200-3500 A, a sharp maximum at 3400 A and a smaller one near 2800 A, none "near 2000 A", and "antagonistic" 2900-3200 A rays that block the active band even at one-tenth the intensity; filters and glass transmission (8 mm Jena glass, 5 mm common glass, gelatin) also pointed above 3000 A [read: p.73-74]. Gabor is the later Nobel physicist [memory, high]; the book does not say so. Rahn explains only their spectrograph experiments (older root parts irradiated, effect appears at the meristem) and writes that the rest "have never been accounted for in a really satisfactory way" [read: p.74]. The Gurwitsch school repeated at 3340-3400 A and found nothing: Frank (1929, three agreeing runs, no radiation above 2400 A from tetanised muscle); Chariton, Frank and Kannegiesser (1930, no effect beyond 2600 A); and Table 18 (3340 A at relative intensities 0.2-800: values -1, +5, +2, 0, -1; 3380 A: -7, +8, -2) against Gurwitsch's +20 to +30 at 2190-2350 A [read: page image p.75]. [calc] Table 18's right-hand values sit inside Schwemmle's quoted +/-10% noise at 500 mitoses (mitoses counted not stated). The testers are the rival school using the same kind of biological detector; the Reiter-Gabor glass-transmission, 2800 A and antagonism results were not answered. Reiter and Gabor's cytology (Table 37: alpha-1 newly born 21/16/5, alpha-2 ripe 64/56/21, alpha-3 dividing 15/11/3, beta resting 0/17/71 percent at the tip, 50 and 80 cells up; rows sum to 100 [calc]) and their claim that the shaded side of the root is itself affected are reported with Gurwitsch's objection that the opposite-side decrease "has been observed occasionally, but only in about half of all his ... and also of Rossmann's experiments" [read: p.143-144]. Gabor's parthenogenesis counts: 3340 A on salamander eggs, 2 then 1 egg developed (groups of about 5; one reached 32 cells), controls and other wavelengths died; frog eggs, 3 of "a few hundred" started dividing [read: p.183-184]; no denominator for the salamander series, no replication on record here. Verdicts: band 3200-3500 A as the mitogenetic band: Tested, not supported (by the rival school only; same detector class; no error limits in Table 18); inhibitory 2900-3200 A band: Never properly tested; parthenogenesis by 3340 A: Never properly tested (3 plus 3 eggs, denominators missing); the conflict itself is an unresolved dispute inside the field over the defining property of claim 1, which the primer should state.

B. Intermittent irradiation (claim 1, claim 4). Table 31 (Gurwitsch 1932, yeast-on-yeast, about 3000 rpm): with 800 or 400 or 100 interruptions per second, 12.5-13 s of actual light gave 22-51% effect against 6-8 minutes needed continuously, "about 1/30th" [read: page image p.118-119]. At 50 per second, 15-16 s was "usually insufficient" and 30 s was needed (27-29%). Zoglina's half-disk: five values near zero (mean -4.6 [calc]) and five at +28 to +64 (mean 42.4 [calc]). A regular disk worked and an irregular disk with the same total window width (75 degrees) did not; with intermittent light the effect carried 15 cm against 3-4 cm continuous (Gurwitsch 1932, pp.256-263; no data in Rahn) [read: p.121-122]. [calc] Internal problems: rows have 1-4 values and no spread; several window durations do not match 3000 rpm (a 10 degree window open 0.00056 s against 0.0005 or 0.0007 printed); the 13 s row (42%) conflicts with the text's "15-16 s insufficient"; Zoglina's "10 rpm" cannot give "1/20 second" intervals (that needs 600 rpm); Rahn's "1:9" light-to-dark ratio holds for the 16-window disk (40 degrees open, 320 closed, ratio 1:8) but not for the single 10 degree window (1:35). Rahn concedes that the Warburg flashing-light analogy "is not at all certain", that Gurwitsch found a 30-fold gain where the dark/light ratio was at most 1:9, and that threshold times may not measure intensity [read: p.121; p.128]. No continuous-exposure data are tabulated. Verdict: a regular pulse train lowers the threshold exposure about 30-fold and an irregular one does not: Never properly tested* (one school, via one monograph, qualitative for the regular-against-irregular test, internal inconsistencies, no independent replication). It is a testable, anomalous claim that a plain dose explanation cannot absorb; it is not support for the effect until the 6-8 minute continuous baseline and the regular/irregular pair are published with counts.

C. Daylight dependence of senders (Potozky 1930; claim 1, claim 7). Table 33 (Baron method; percent increase in buds): daylight-grown sender against any detector, exposed in daylight, 20-35% (means 24.0, n=6, and 27.0, n=8); every other condition averages -5.9 to +6.5% with spread about +/-10 [read: page image p.122]. [calc] Welch t for the daylight/daylight daylight-exposed column against the same pair exposed in the dark (means 26.9 and 0.1): about 7.1. Yeast grown dark regained the power after about 2 hours in daylight (no data given). Rahn: this "accounts probably for a number of negative results" [read: p.122-123]. Limits: the controls' light treatment is not stated; no UV-blocking screen (glass) between blocks; n = 4-8 per cell; one school; the bud method itself "fails to indicate the true growth rate" (see G). This is the book's own source for the proponents' objection that Quickenden et al. "preferred to violate" the daylight condition [web: Frontiers 2015, as in claim 1]; the condition was stated in 1930 and so is not a post-hoc rescue, but it was not tested with a calibrated positive control either. Verdict: daylight as a necessary condition for yeast senders: Never properly tested (internal factorial design, one school). It lowers the weight of Quickenden 1989 as a refutation, as the claim 1 text already says, and it does not reverse the NRC verdict reading there.

D. Polarisation, gradual approach, spreading, exhaustion (claim 1, claim 4). Polarisation: Wolff and Ras 1934a, "apparently" polarised rays are "enormously" more effective; no numbers, no polariser, no replicates [read: p.127-128]; Rahn's inference that all earlier intensity measurements "have become practically meaningless" is his and stands as his concession. Never properly tested. Gradual approach: two yeast blocks moved from 6 cm to contact over 5-8 minutes gave no induction while blocks placed at the final distance gave +40-50%; a 3-minute approach gave the regular effect; an elliptical disk did the same; words and ranges only [read: p.131-132]. Never properly tested. Lawn spreading: 65% in a 0.1 mm irradiated slit, 100% at 1 mm, falling to 25% at 10 mm; one series; the peak lies outside the irradiated zone and nobody comments [read: p.124]. Sender exhaustion and recovery (37% to 2% to 0% to 1.8%, single values; Rahn: 30 minutes may be too short) [read: p.127]. Salkind's "periodicity": the repeat experiments in Table 35 often have opposite signs at the same exposure (arc, 18 s: +48 and -8; albumin, 10 min: +77 and -38) and Rahn writes "No definite periodicity could be found" [read: page image p.131]. Verdict for periodic maxima: Tested, not supported by the book's own table (Rahn's concession), though Gurwitsch and Wolff and Ras are said to have seen several maxima.

E. Non-biological detectors (claims 1 and 2). Stempell's Liesegang rings and peroxide decomposition (onion allyl-mustard oil disturbed the rings chemically; Stempell later claimed disturbance with chemistry excluded; Stempell himself holds that the effect is usually "radiation and chemical effect of a 'gas'" [read: p.102-103]); Heinemann's gold-sol nephelometer (single runs: blood about 26 units at 15 minutes against 2-3 for controls; dissolving NaCl 21 against drifting controls; no repeats, no units, no statement that controls ran concurrently [read: page image p.104]); Lepeschkin 1933 (dying yeast darkened AgBr in sealed quartz tubes, fogged Eastman Speedway plates except under filter paper; dead and living yeast did not; wavelength 1800-2300 A by absorption; number of experiments not given [read: p.113-114]); Suchow and Suchowa 1934 (egg white coagulated by alcohol over AgBr: quartz vessel's suspension darkened sooner in 25 of 25 experiments, glass vessel as control [read: p.114-115]; one-sided chance probability about 3e-8 if each outcome were 50:50 [calc]; blinding, vessel matching and the size of the effect not stated). Negative: Taylor and Harvey (90 days, plates, no effect [read: p.104]). Rahn's own reading: "None of them are absolutely convincing, and Gurwitsch has refused them all" [read: p.104]; he disputes Lepeschkin's claim that mitogenetic rays are necrobiotic. Caveats that apply: Lepeschkin's filter-paper strip blocks light but only slows dissolved substances, so it does not exclude chemistry cleanly; a gold sol is made unstable with NaCl and is flocculated by salt in ordinary colloid chemistry [memory, high], so the NaCl-dissolving response has non-radiative explanations that the single run does not exclude. Verdicts: photochemical or physico-chemical detectors register UV from dying or reacting systems: Never properly tested (several single-group positives with quartz/glass logic and one 25/25 series; one negative; no independent replication; chemical exclusion incomplete); Heinemann's NaCl-dissolving response as radiation: Never properly tested and the one-photon-per-event mechanism stays Refuted by calculation (claim 2).

F. Physical counter evidence in Table 30 (claim 1). [read: page image p.105; excess over combined error, [calc]] Rajewsky: onion root 1.3, onion pulp 4.1 and 4.3, mouse carcinoma 8.8, onion root 2.3; Frank and Rodionow: frog muscle 9.6 and 3.7, frog heart 2.5 and 3.3, muscle pulp 2.5 (percent excess 10-29% for Rajewsky; 21-233% for Frank and Rodionow). So six of ten rows exceed 3 combined errors, three lie between 2 and 3, one is 1.3. Controls are not described (bare counter, inert sample or closed shutter unknown); no sample sizes. [calc] Rajewsky's printed +/-0.3 on 39.8 would need about 440 ten-minute intervals if it were pure Poisson counting error, so the +/- is something else or the runs were very long; undefined. Photoelectric yields (Kreuchen and Bateman 1934, Table 30a): Rajewsky about 1e2 quanta per electron at 2650 A; Schreiber 2e4; Grey and Ouellet 6e3; Frank and Rodionow 2e3; Lorenz 3e3; Kreuchen 2-5e4 (up to 1e4-1e5 later) [read: page image p.106]. See section 1 for what this does and does not show. Rajewsky's flux, 1e-10 to 1e-9 erg/cm2/s (10-100 quanta/cm2/s at 2300 A), compares with Barnes's own printed counter limit of about 1e-9 erg/cm2/s: 1e-10 erg at 2300 A is 12 photons and 1e-9 is 116 [calc], so Rajewsky's estimate lies at or below the floor of the method the same chapter states (see the recomputed table above). Rahn's evaluation: "In none of the physical measurements ... is it absolutely certain that these errors have been excluded" and counter results expressed as percent of background are "utterly meaningless from a quantitative viewpoint" because stray radiation changes with shielding [read: page image p.106; p.128]. Siebert and Seffert (1934): several hundred normal bloods raised the count; carcinoma blood did not; a KCN substitution for normal blood removed the increase (no counts, no number of carcinoma samples, no statement whether the same sample was poisoned; KCN also changes pH and releases HCN vapour) [read: page image p.106]. Rahn calls the KCN comparison "most convincing"; it addresses Lorenz's vapour and charge artefacts better than an empty-counter control. Kreuchen and Bateman (1934) are named by Rahn as negative reviewers whose series he says is equivalent "to 140 single experiments by Gurwitsch"; he answers "That is a mistake" [read: p.109]; their own data were not read. Verdict: unchanged; the counter evidence (positive, with unresolved artefact objections and unclear yield) remains Plausible, untested for oxidation emission and Never properly tested as a neutral blind measurement for biological emission. Hollaender and Claus 1937 remain the controlled negative on record, contested on protocol.

G. Rahn's own methodological concessions (all claims). (1) The induction effect "has no definite meaning" and Table 26 shows why: the same culture gives 183% from counts and 27% from generation time; the two controls differ by 17% in generation time (47.6 and 56.0 min) and 37% in 6-hour count [read: page image p.93; [calc] generation times reproduce to rounding and induction 14, 25, 27, -1 reproduces exactly]. The 15 and 30 minute effects (25-27%) are therefore about 1.5 times the control spread, and the 7.5 minute dose is slightly negative and the 60 minute dose weak. (2) The bud-percentage method gives depressions (-18 to -52%) where cell counts show increases (22 to 123%) in the same runs (Salkind, Table 34); Rahn: it "fails to indicate the true growth rate" [read: page image p.130]; this affects every bud-method table (Table 31, Table 33, Tuthill-Rahn). (3) Yeast duplicate plates differ by 40% commonly, up to 103%; Nadsonia duplicates up to 237% (Schreiber's data) [read: p.93]. (4) Onion roots "have no real controls"; Motshejewa found that pressure and friction raise mitoses, that Gurwitsch's good results can be explained by pressure in the glass tube, phototropic curving, selection of "good" roots for important experiments and omission of sections with decreased mitoses, and obtained negatives with yeast and blood senders; Rahn: "considered by many critics to be the final proof against mitogenetic radiation" and answers with Margarete Paul (1933): untouched, straight roots, no tube, dark room, mitoses larger on the exposed half, a steel needle caused downward turning but no mitotic increase [read: p.76-77]. Rahn's own caveat on Paul: "The number of examples given is not large enough to draw many more conclusions" [read: p.77]; Rahn does not say whether Paul answers selection and section omission. (5) Gurwitsch's protocol recommends that the counter "should not know which slide is under the microscope" [read: p.80]; whether this was done in the published series is not stated, so "blinding not stated" holds for every table, while blind counting was recommended by the school. (6) Siebert's no-effect controls give standard deviations 2.6-4.6 against mean increases of 7.9-13.4 points, ratios 2.4-5.2 [calc from p.85]; his Table 19 (10 runs) and Table 20 (24 runs, resting against excited frog muscle, excited above resting in all 24; two-sided sign test about 1e-7 [calc]) are positive biological series with a quartz plate and, for muscle, a KCN-stopped chemical control [read: p.78-79]. Weighing: these concessions lower the evidential weight of percentage tables across the field and raise the weight of the few designs with within-subject controls (Julius, Blacher's limb and wound series, Siebert's runs, the Table 28 cornea pairs). They do not by themselves refute the effect.

H. Better-controlled biological series, now stated (claims 1 and 6). Julius 1935 stands as in the table above (1.01 +/- 0.08 confirmed [read: page image p.97]). Corneal epithelium (L. Gurwitsch and Anikin 1928): spectral line 2030 A for 3-4 minutes on one eye, the other eye the control, in triton, frog and rat, five animals each; exposed above control in all 15 paired comparisons, +40 to +300% in triton and frog (rat percentages unreadable on the scan; two rat rows used a yeast sender); two-sided sign test about 6e-5 [calc]; counts between animals vary more than tenfold, so only the pairs mean anything [read: page image p.98]. Sea-urchin eggs (Ziripolo 1930, Table 27): unchanged eggs 71-98% in controls against 8-35% irradiated; baselines differ widely; one control row has no irradiated entry [read: page image p.95]. Magrou's Table 29: 221 no-source controls, 0% abnormal larvae; microorganism sources through quartz 78% (118 experiments) and through glass 33% (9); chemical sources (oxidised glucose) through quartz 85% (54) and through glass 0% (20); Berthelot's culture medium alone 100% through quartz (18) and 0% glass (6); [calc] for P. tumefaciens, 57 of 76 abnormal through quartz against 3 of 7 through glass, Fisher exact two-sided p about 0.09, so seven glass runs cannot separate the two [read: page image p.101]. The experimenters' own alternative (Magrou and Reiss: glass-paraffin-glass transmits; a metallic connection restores normal larvae; needs a large redox-potential difference) is "tentative" and the Magrous do not prefer it [read: p.100]. A non-living culture medium acting through quartz and not through glass fits either UV or an electrical effect. Other detector series: Table 21 (Tuthill and Rahn): a fresh 24-hour parent gave 25.0 and 29.2% buds at ages 0 and 0.5 h against 3.6 and 3.5% in controls and no effect afterwards; a 6-day parent responded only at 2-2.5 h; the printed table has small arithmetic discrepancies (724 against a recomputed 734; -17 against -14; 62 against 59.7), no replicates and no named sender [read: page image p.82; [calc] by the deep reader]. Table 36 (Ferguson and Rahn): a 3-day culture irradiated for 30 minutes and transplanted 2 hours later still grew 3.4-4.2 times as fast at 6 h (6.4 at 8 h immediate), no replicates [read: page image p.140]. Verdict: unchanged ("Never properly tested" for the division-accelerating effect under a protocol both sides accept); these are the specific series that a blinded two-laboratory test would replicate, and the settle-it paragraph in claim 1 should name Julius, Blacher and the cornea pairs as the designs to copy.

I. Single-quantum paragraph, as Rahn gives it (claim 3). The calculation stands (4.2e-4 per second per cell, about 40 minutes for a mean of one hit) and so does the arbiter's 63% correction. Add Rahn's own data and retreat: percent buds after 2 hours at 0, 10, 20, 40 minutes of exposure were 21.0, 25, 36.5, 24.5 through quartz and 22.5, 23, 37.5, 27.5 direct, so only 20 minutes responds and the 40-minute dose, where more cells should be hit, shows almost nothing; no error limits; "direct" and "through quartz" rows look alike and no glass row is given [read: page image p.137]. At 20 minutes the expected hits per cell are about 0.5, so about 40% of cells (Poisson) [calc]. Rahn then writes that, because media near cells become secondary senders, "the assumption of a single quantum acting upon the cell has become unnecessary, even improbable", and raises the paradox that fermenting yeast cells themselves emit 1900-2170 A yet respond to the same wavelengths from outside; his escape (old yeast, stored cold, fermented within 10 minutes) shows "the mitogenetic effect does not occur merely through the increase in energy content of the cell" [read: p.137-139]. Miss Ferguson's two-sided irradiation (unpublished; strong effect, no cancelling; Rahn argues against Gurwitsch's interference idea) is prose only [read: p.142]. Verdicts: unchanged.

J. Claim 5 detail (blood and cancer). Case numbers are absent from every Gesenius, Heinemann and Klenitzky statement quoted [read: p.193-195]. Gesenius's exceptions are in words only ("occasional", "infrequently", "rarely"); Heinemann is quoted that a positive effect "excluded the possibility of a tumor with certainty"; Siebert (1930) reports that of 35 cancer patients "the majority" showed no blood or urine radiation, with exceptions after recent X-ray treatment or isamine blue [read: p.79]. Gesenius says radiation stays absent after tumour removal; Klenitzky (1934) says it returned when every trace of tumour was removed; Rahn states both without reconciling them [read: p.195]. Braunstein and Heyfetz's rats (Table 53): groups of 3-7 (normal 14); induction +33.4 normal, +26 day 5, +13 day 6, +7 days 8 and 10, -4 day 13, +5 day 16; sugar decrease at 1.5 h varies 38-62 mg without trend; same-rat or separate-group design not stated [read: page image p.194]. Table 22 (three blood samples, healthy +108, carcinoma -54, tonsillitis 0): the "actual increase" column (82 and -42) cannot be reproduced from the printed figures [calc, read: page image p.87]. Protti: two lots of 12 mice, tumours grew faster in the starved lot whose blood radiation fell from about 50 to about 10; cytophotolysis (yeast destroys neoplastic cells in vitro, also through a quartz plate; no counts, no glass control, no group sizes); yeast injections liquefied rat sarcoma, heated yeast (60, unit not printed) had no effect; effects dose-dependent in opposite directions in mouse adenocarcinoma [read: p.196]. Protti's young-blood "rejuvenation" (Table 48, ten cases; clinical rating tracks the size of the radiation rise; no control injections; criteria undefined; the two "none/little" cases began with the weakest young blood) [read: page image p.166]; cytagenin (Seyderhelm): raised blood radiation in healthy people and, after 1-2 weeks, in 70-80 year olds ("How long these experiments were continued, Heinemann did not state") [read: p.165-166]. Starved-rat blood needed 2% glucose, far above blood level, to radiate (values 2.5, 5, 2 without; 21, 44, 36, 39 with) [read: p.163]; Table 47 (blood radiation falls to about zero at end of a working day in all nine sessions and returns incompletely in some) is the book's own warning that a blood test depends on the donor's recent work [read: page image p.164]. Rahn's reading that Roffo's sunlight lacks the mitogenetic band stands. Verdicts: blood radiation as a cancer diagnostic: unchanged, Never properly tested. Protti's cytophotolysis and tumour-by-yeast results: Never properly tested. Young-blood rejuvenation: Never properly tested (ten cases, no control). Rahn's own phrase "best results obtained" for Heinemann attaches to an account with no numbers.

K. Claim 4 detail (nerve and brain). Kalendaroff (1932): frog sciatic nerve radiates "distinctly"; minimal exposure about 6 minutes resting and 2-3 minutes excited; spectra differ with kind of irritation; each 10 A strip "is the result of at least three determinations"; Rahn notes that missing glycolytic lines at 1930-40 and 1950-60 and a 2410-20 line "may be due to a slight shift in the calibration of the instrument", but 2000-10 and 2370-80 "cannot be explained by the same error" [read: p.169-171]; the nerve spectra lack the phosphatase lines [read: p.170-171]. Anna Gurwitsch (1932, 1934a,b): light on the frog eye makes optic lobes radiate ("results were always positive while controls without illumination consistently showed no effect"; numbers of animals not given); the lobes' spectrum contains glycolytic lines "even when radiation has been brought about by ... pyrogallic acid" which Rahn reads as "the release of an unknown independent chemical reaction in the optic lobes", not mere secondary radiation; colour-dependent spectra, with "the intensity of the colored lights ... and the wave lengths ... not mentioned" [read: p.171-172]; light-intensity series at 6 and 18 cm (9:1 by inverse square; effect 22.5 at 5 s against 16 at 30 s) and Table 49 (isolated heads stay at +21.8, +12.0, +14.6 after darkening; living animals fall to -2.0 and -2.5) [read: page image p.173]. Frank: sartorius muscle radiating from a spot 20 mm away with greater intensity [read: p.168]. Latmanisowa: parabiosis of the sciatic nerve by protein-digestion radiation after about 2 hours, "repeated more than 40 times with the same success"; controls named: gastric juice without protein, and the unexposed part of the same nerve; recovery 2 hours after removal; the parabiotic nerve radiates more than a normal one; Lapitzki obtained the effect from a mercury lamp in minutes; Rahn: it "has not been possible as yet to produce muscle contraction ... by irradiating a nerve" [read: p.174]. Fig. 40 (nerve exhaustion and recovery) is one run per panel [read: page image p.125]. Speed figures rest on positive angle ranges only (2.5 cm: 25-50 degrees; 5 cm: 40-70; 15 degrees = 0.00083 s = about 30 m/s) with no counts [read: p.125-126; arithmetic reproduced in the earlier table]. Verdicts: the optic-lobe response and spectra as light-evoked UV emission: Never properly tested; nerve spectra as biochemical fingerprints: Never properly tested (same as claim 2 fingerprints, with Rahn's calibration caveat); parabiosis by digestion radiation: Never properly tested (best-controlled item in this section by Rahn's account, but no quartz/glass screen is named, so a chemical or thermal route is not excluded). The "impulse emits light" form is Plausible, untested as ordinary ultra-weak emission from active tissue [memory, medium] but not at 190-260 nm at the book's intensities. The nerve-conduction role stays Historically superseded but instructive.

L. Peripheral items (restored). Acs (1933): B. anthracis retarded 22-136% by Ps. pyocyanea in 14 experiments; reversed roles 42% and 48% in two; B. mycoides stimulated 42% and 50%; yeast retarded staphylococci in 5 and streptococci in 1; no quartz/glass control, no wavelength, no exclusion of diffusible antibiotics; Rahn: such selectivity "cannot be explained by our present knowledge" [read: p.186]. Never properly tested. Kostoff and Kendall (1932): 120 tomato plants inoculated with B. tumefaciens, 7 shoots arose in the tumour, 5 rooted, 1 tetraploid; no control or baseline; link to radiation by association only [read: p.183]. Never properly tested. Nelson and Brooks (1933): infrared 8000-12000 A lowered fertilisation in 9 of 9 experiments (18.5-87.1%), temperature difference at most 0.06 C [read: p.115]; two-sided sign test 0.004 [calc]; Never properly tested (one group, infrared, outside the mitogenetic claim). Zwaardemaker: radioactive substitutes for potassium restarted frog hearts in 34 experiments [read: p.115-116]; not tied to mitogenetic radiation by Rahn; no verdict (a reported result from one author, not checked against other sources). Roffo skin-site percentages sum to 100.20% (immaterial). Sedum and Bromley salamander tail: Bromley's printed table runs 6.2, 2.1, 29.6, 41.1, 39.9, 48.4, 56.4% over 1-24 h and does not itself show the "second maximum at 5-6 days" Rahn describes [read: page image p.188; the text layer's 50.4 is 56.4 on the image]. Blacher's tail-wound series (Table 51): 146 irradiated and 143 control animals in the nine printed rows, over 500 in all; within-dish effect +11 to +54% (mean about 23) irradiated and -10 to +9 (mean about -1.4) controls; the irradiated-at-once under-side gain +33% is reproduced from the printed rows (+31, +32, +35%), the "26% less" on the upper side and the 48-hour "true stimulation" are not clearly reproducible from the printed subset [calc, read: page image p.190]; Table 50 (Pelobates tail-tip pulp, yeast detector) gives double maxima at 12 and 36 hours on a table with many empty cells and a different double maximum from Bromley's and from the metamorphosis blood data (stages IIIb and Vb, ratio of threshold times 20 and 10 against 1/6 at IIIa). Verdict for the wound-healing series: Never properly tested (one school, 500 tadpoles, within-animal design, no independent replication; "double maximum" thinly supported). Magrou's mesenchyme histology stays under claim 6 as before.

3. Verdict changes

No verdict moves along the scale. The table lists the changes in wording or additions.

4. Retention checklist

Restored above, with source: Reiter and Gabor and the Gabor connection (A); intermittent irradiation (B); Potozky's daylight result tied to the Quickenden objection (C); polarisation, approach, spreading (D); non-biological detectors (E); Motshejewa and Rahn's caveat on Paul (G); Siebert and Seffert and Kreuchen and Bateman (F); Anna Gurwitsch, Kalendaroff, Latmanisowa (K); Protti, cytagenin and Table 48 (J); Acs, Kostoff and Kendall, Nelson and Brooks, Zwaardemaker (L). Still not in the arbiter and left to the primer: Stempell 1932 (further reading), the Secret Life of Plants (Tompkins and Bird 1973) and T. G. Hieronymus's "eloptic" devices as a later popular account, Rahn's Mathematics in Bacteriology (1939) and Microbes of Merit (1945) as records of his later view, and the identity of the "Barnes and Rahn 1933" co-author (probably M. N. Barnes; Fig. 37's Ferguson and Mrs. Barnes as Rahn's laboratory workers are named in the book [read: p.110; p.178]). These leads are not read in the original by anyone here.

5. Open questions added or changed

Replace open question 6 (withdrawn above). Add: (10) What did Rajewsky's counters have as window area, and what was the real photoelectric yield (1e2 against 2e3)? (11) Do regular pulse trains lower the effective threshold 30-fold when a physical UV-C source of known flux replaces the biological sender? (12) Does daylight conditioning of a yeast sender change its measured UV-C emission? (13) Did Gurwitsch's counters count blind, as his protocol recommends? (14) What do Kreuchen and Bateman's 1934 data show, read in the original?

6. Web verification log (this revision)

1. UPE spectral range [web, search summary: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10899412/]: 350-1300 nm; page not opened. Budget used this revision: 1 call. Page images read for this revision: none by me; all page-image figures are the deep readers' renders, and every [calc] above is arithmetic on printed numbers, not a re-reading of the page.