Physicist review: Otto Rahn, Invisible Radiations of Organisms (1936)

Reviewer: PhD physicist (Step 2). Project: Rahn-Otto. Source: one book (Protoplasma-Monographien IX, 1936; physics chapter by Sidney W. Barnes). Tags: [read] from the extracted book or my own reading of the deep read; [read: deep] taken from deep/ (which cites PDF pages; book page = PDF page minus 14); [calc] my own arithmetic; [web] with URL; [memory] with confidence. The Investigator's items are tagged [web] as the Investigator recorded them (I did not re-open them unless I say so).

Source gap carried over from Step 1: the deep read covers only PDF pp.44-67. Claims 1, 4, 5, 6 and 7 rest on the digests plus the passages I checked in the extracted text (PDF pp.~126, 138, 174, 199-206). Rahn's own 1934, 1937 and later papers, the 1937 negative report (Hollaender and Claus), and the Soviet material were not read by anyone; they are known only through the Investigator's notes of abstracts and citations. Where a verdict leans on these, it says so.

Units used. 1 A (angstrom) = 0.1 nm. Photon energy E = hc/lambda, with h Planck's constant, c the speed of light, lambda the wavelength. In convenient units E (eV) = 12398 / lambda (A). One eV is the energy an electron gains across 1 volt; 1 eV per molecule = 23.06 kcal/mol = 96.5 kJ/mol. [calc]

---

0. What "mitogenetic radiation" is, and the mainstream view stated fairly

Alexander Gurwitsch (Russia, 1923) reported that an onion root tip pointed at another onion root tip makes cells in the second divide faster, even across a quartz plate (transparent to short ultraviolet) but not a glass plate. He named the cause "mitogenetic rays" (mitosis = cell division). Rahn's book is an English-language advocacy summary, mostly of the Gurwitsch school's work, with some experiments by Rahn and his Cornell co-workers [read].

The mainstream position, stated as specifically as I can:

1. Living tissue does emit extremely faint light, called ultra-weak photon emission (UPE) or "biophotons". It is chemiluminescence: electronically excited molecules formed in oxidative metabolism and oxidative stress. Reported fluxes are of order units to hundreds (metabolism) and hundreds to thousands (stress) of photons per cm2 per second, in the near-UV, visible and near-infrared, roughly 350-1300 nm [web: https://www.frontiersin.org/articles/10.3389/fchem.2023.1140128/pdf and the search summary of Cifra and Pospisil 2014; the figures are the review authors' summary as returned by a search tool, not read from the article]. Rahn's estimate for the strongest "mitogenetic" sender, 100-1000 quanta per cm2 per second [read: p.137-138, extracted digest], is in the same intensity range. So the existence of weak emission from living matter is not what mainstream rejects.
2. What mainstream rejected (after the 1937 NRC study, below) is the specific package: (a) the emission is in the far ultraviolet, 190-260 nm (1900-2600 A), photon energies 4.8-6.5 eV; (b) it speeds cell division in other cells at 1-1000 photons per cm2 per second; (c) it is detected by yeast, bacteria and onion roots in a narrow physiological window; (d) it carries chemical "fingerprints", conducts along nerves and tells cancer from health. Modern UPE work does not report 190-260 nm biological emission at that intensity as an established fact (see claim 1, [memory, medium], not checked on the web).
3. Heat, not light, is the ordinary fate of biochemical energy. A 6 eV photon is a large lump of energy compared with what one metabolic reaction step releases (about 0.1-0.6 eV for hydrolysis and neutralisation, about 1-3 eV for strongly exothermic oxidations [memory, high]). This is the core physical difficulty and Rahn states it himself (claim 3).

---

1. Claims, questions a-d, verdicts

Claim 1. Living cells emit short UV (about 1900-2600 A, "mitogenetic radiation") that speeds up division in other cells that absorb it.

a. Mathematics. There is little mathematics. The one calculation I could check is the single-quantum estimate: a yeast cell has probability 0.00042 of being hit by a quantum per second at 1000 quanta per cm2 per second (an effective target area of 4.2 x 10^-7 cm2, about 42 square micrometres), so one hit per cell takes about 40 minutes [read: p.137-138 digest; calc: 1/(4.2e-7 x 1000) = 2380 s = 39.7 min]. That is arithmetically right and is a fair size for a yeast cell. Rahn then notes that the strongest effect appeared at 20 minutes, which argues the effect is not one quantum per cell, and later abandons the single-quantum assumption in favour of secondary radiation amplification [read: p.138 digest]. Internally consistent, but each step is hypothetical.

b. Consistency with established experiment. - Consistent: ordinary UV does speed or slow growth, and UV damage is wavelength dependent (Rahn's own Figs. 27-28, [read: deep p.62-64]); quartz passes and glass blocks short UV (Gerlach, [read: p.48]); chemiluminescence and UPE exist (above). - Not consistent, or unestablished: (i) the biological detector gives no absolute photon count. Rahn says so: "the intensity of radiation has been ascertained only biologically", reciprocity (intensity x time = constant) fails at low intensity [read: deep p.59]. (ii) The effect appears only in a window of intensity and of dose (Fig. 28; Table 12, where each concentration shows one positive exposure time flanked by zeros, [read: deep p.58, p.63]). (iii) Far-UV photons (6 eV) are absorbed within micrometres in water, protein and nucleic acid (nucleic acids absorb strongly near 2600 A, proteins near 2800 A and below 2200 A [memory, high]). Rahn himself records that 0.5 mm of standard broth transmits only above 2500 A, and that ultraviolet "will not pass the skin of man or animals" [read: digest p.89-90; deep p.66-67]. A cell would have to be near the emitter, with the sender and detector effectively in contact. This is compatible with Gurwitsch's contact experiments but conflicts with several-centimetre claims (p.124, 10 mm spreading, 25-87% increase). - Controls: Table 15 (quartz against glass over an oxidising reaction, Ferguson, unpublished, one run) is the only experiment in the deep-read pages with a simultaneous UV-blocking control [read: deep p.65-66]. Quartz and glass dishes also differ in heat conduction and in how they are placed; the effect appears at 3-4 h and the quartz culture is behind at 1-2 h (140 vs 154, 216 vs 253). One unreplicated run cannot carry the claim.

c. Tested? By whom? Replicated? - Supporters' tests: the Gurwitsch school (dozens of papers, 1923-1934), Reiter and Gabor, Siebert, Wolff and Ras, Frank and Rodionow, Magrou, Paul (1933, onion roots in a dark moist chamber, steel-needle control, [read: digest p.74-75]), Julius (1935, 56 pairs, stimulation 0.91 +/- 0.17, 5.3 times its probable error [read: digest p.97]), and Rahn's own group (Ferguson and Rahn 1933; Tuthill and Rahn 1933) [read]. Rahn's own caveat is that the 54 failures in 373 Gurwitsch experiments (14%) were attributed to poor yeast culture [read: digest p.108]. - Opposing tests: Taylor and Harvey (1931-32, no photographic effect after 90 days' exposure, [read: digest p.105]); Moissejewa/Motshejewa (onion roots, mechanical pressure raises mitoses; she proposed selection bias, [read: digest p.74-75]); Kreuchen and Bateman (1932); Nakaidzumi and Schreiber (1931) [web, via Investigator: Naumova 2018 reference list]. - The decisive Western test: Hollaender and Claus, An experimental study of the problem of mitogenetic radiation, National Research Council Bulletin 100 (1937): two years' work, biological and physical detectors, "all the attempts have yielded negative results"; the Nature news item notes about 600 papers had been published and defines mitogenetic rays as 1900-2500 A at 10-1000 quanta per cm2 per second [web, Investigator: https://www.nature.com/articles/1401007c0, abstract only]. I did not read the bulletin itself. - Later: Quickenden measured weak photon emission from yeast (1974) but did not reproduce the division-stimulating effect (1989 attempt) [web, Investigator citing Volodyaev and Beloussov 2015, via a model summary of the article (see Investigator's caveat); my own search for the 1989 paper's abstract returned nothing, so the result of the 1989 paper itself is unconfirmed by me]. Wainwright, Killham, Russell and Grayston (1997, Microbiology 143:1-3) title their paper "Partial evidence for the existence of mitogenetic radiation" [web: https://pubmed.ncbi.nlm.nih.gov/33711853/, record only, no abstract]; Volodyaev and Beloussov (2015) say they obtained "a good effect" but could not make it reproducible [web: Frontiers 2015, summary by a fetch model]. - Interested-party note: Volodyaev and Beloussov (Moscow State University, Gurwitsch-tradition authors) argue the negative studies used untested conditions (culture age, medium, detector layer thickness, induction time) and call the effect "undeservedly forgotten" [Investigator's characterisation of their paper]. This is the supporters' own reading; it is a specific, testable objection and I report it beside the NRC result. Rahn is himself an advocate who co-authored experiments.

Honest weight: this was tested (many times, by both sides) and a neutral, well-resourced test (NRC, Hollaender and Claus 1937) was negative. A recorded counter-argument says the negative tests did not use the detector conditions the positive laboratories found necessary. I found no blinded, pre-registered replication by an independent laboratory using Rahn's published protocol, so the record is not clean either way.

d. If true, what changes? Not much of the physics: cell-to-cell signalling by faint UV is compatible with quantum mechanics and photochemistry. What would change: (1) cell biology would have to include an optical channel at 190-260 nm between cells, which means a source at 5-6.5 eV (see claim 3), and a detector that responds to about one photon per cell; (2) UV absorption in cell media would have to allow centimetre transport (waveguiding or relay chains); (3) the standard assumption that division is controlled chemically and by contact would need a second, optical, route. Quantum yields of photochemistry are normally below one; a one-photon trigger needs an amplifier inside the cell (Rahn's "release of a complex mechanism", [read: deep p.62]), which is plausible in principle (photoreactivation, vision) but not demonstrated here.

Verdict (split). - Weak form, living matter emits ultra-weak light (near-UV to near-IR) from oxidative chemistry: Established [web as above; modern UPE literature]. - Strong form, 190-260 nm emission from cells, at 10-1000 quanta per cm2 per second, that stimulates division in other cells: Tested, not supported (NRC 1937; Quickenden 1989), with the dispute over detector conditions recorded above, and with positive results by Gurwitsch-school laboratories, Paul, Julius and a partial 1997 UK result also on the record. It was not dismissed without testing, but it was also never settled by a controlled, independent, protocol-matched replication.

---

Claim 2. The light comes from ordinary biochemistry (oxidation, glycolysis, proteolysis, phosphate cleavage, neutralisation, dissolving NaCl), each with a narrow-band spectral "fingerprint", read out with yeast or bacteria and, more weakly, with photoelectron counters.

a. Mathematics. No derivations; data tables. My checks of [read: deep p.48-54]: - Table 10 (counter, exposed versus control), combining the printed errors as independent: excess over the combined error for the nine rows is 3.2, 4.5, 1.5, 2.2, 1.6, 2.7, 3.6, 6.1 and 0.4 standard errors [calc]. Six of nine exceed 2; one (the last, 12 against 11) shows nothing. The text does not say how the errors were derived. These are small counts (10-40 per interval). - Barth, proteolysis: 24 of 25 runs above control. A sign test gives p = 7.7 x 10^-7 against chance [calc]; this is a strong statistical signal of something, but not of its spectrum or of its being light from the reaction rather than an artefact (Lorenz's three error sources, below). - Table 11 (50 A strips, yeast cell counts): error limit stated as +/-15 (units not given), seven reactions. Taking +15 as threshold, all seven are positive at 2200-2350 A, which Rahn himself says matches Kannegiesser's earlier range, so the oxidation "general spectrum" is partly one range [read: deep p.50]. - "Fingerprint" claim: Fig. 24 gives 5 of 60 ten-angstrom strips for glycolysis, 8 for phosphagen, 7 for proteolysis, 6-7 for amylase, maltase and sucrase. The deep read notes that 10 A scans were done only where 50 A scans had been positive, and Rahn warns "we are not dealing with true spectral lines, but with relatively broad regions" [read: p.51, p.53]. A few strips scattered across a range of 60 with "positive" meaning a growth increase above an unstated threshold has a high false-positive and false-match rate; no null-spectrum (blank run) is shown. So the fingerprint claim has no stated statistical calibration.

b. Consistent with established experiment? - Chemiluminescence from oxidations is real and old (Haber, visible light from pyrogallol oxidation, phosphorus, etc., [read: deep p.47]) and consistent with UPE. - UV at 2000-2600 A from simple reactions is a different matter. A per-event energy check [calc, constants as above; reaction enthalpies from [memory, high]]: - NaOH + HCl neutralisation: about 55.8 kJ/mol, 0.58 eV per event. A 1960-1990 A photon is 6.2-6.3 eV, about 11 times larger. - Dissolving NaCl in water: slightly endothermic, about +3.9 kJ/mol, 0.04 eV per formula unit. The reaction absorbs heat; there is no net energy to emit. Yet Rahn reports a 28-45% growth increase "only during the act of dissolving" (Table 14, [read: deep p.64-65]). - Sucrose hydrolysis: 9.7 cal/g x 342 g/mol = 3.3 kcal/mol = 0.14 eV per molecule, against 6.14 eV for 2020 A. A factor of 43. - Strongly exothermic inorganic oxidations (for example Fe2+ with dichromate) release a few eV per electron-transfer step at most; photons of 5-6 eV need unusual, rare high-energy events (for example radical recombination or multi-step coupling). So the energy of one chemical event does not suffice for the observed wavelengths in the neutralisation, dissolution and hydrolysis cases (except by fantastically rare multi-event pooling). The claim as a general statement, "all chemical reactions emit part of their energy as short UV" [read: p.47], is wrong for reactions of this size. The detector may be responding to something else. - Alternative explanations for Table 13/14/15 results that the book does not exclude in the pages I read: thermal effects through a thin quartz plate (an exothermic neutralisation warms the plate; dissolving salt cools it); other transmitted agents (vapour is blocked only if the plate is sealed); electrical effects (Magrou and Reiss proposed an electrical explanation for sea urchin effects, [read: digest p.100-101]); and ordinary biological noise, since no replicate spread is given, the Table 14 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". - Physical (counter) detection: Rajewsky, Frank and Rodionow, Barth, and Siebert and Seffert are cited as positive. (Note: Step 1's phrase "physical counter evidence" in claim 7 means evidence from Geiger-type counters, i.e. positive for the effect, see Corrections.) Lorenz (1933-34) identified three error sources in such counters: water vapour condensing on the quartz window changes resistance; charges on the quartz tube change the counting rate; muscle tetanised in front of the window gives a direct effect [read: digest p.106]. Kreuchen and Bateman found more than 2000 quanta per photoelectron in most cases, i.e. detector efficiency of order 10^-3 or less [read: digest p.106-107]. Siebert and Seffert's blood experiments (normal blood increases counts, cancer blood does not, plus a KCN control) were called "the most convincing" by Rahn [read: digest p.106-107]; I could not examine them.

c. Tested? Spectra were obtained by the Gurwitsch school and Wolff and Ras using the same biological detectors; the independent physical spectrum is missing. The Investigator lists later work that measured weak emission from chemical reactions (Gurwitsch and Gurwitsch, Nature 143:1022, 1939; Audubert, Angew. Chem. 51:153, 1938) [web, Investigator, via Naumova 2018 reference list; titles only]. I could not check what they showed. The NRC study of 1937 reports negative results with biological and physical detectors, which would include these reactions, but I have not read it. No replication of the narrow-band fingerprints by an independent laboratory is on record in the material I have.

d. If true: Chemistry would have to have an undiscovered mechanism that concentrates 5-6.5 eV into single photons from reactions releasing 0.1-0.6 eV per event (violating energy conservation per event) or from rare high-energy events never otherwise seen; the thermochemistry of NaCl dissolution would have to have a hidden exothermic channel. Neither is excluded absolutely (rare events exist), but it would be extraordinary.

Verdict (split). - Weak form (some reactions, mostly oxidations and radical processes, emit faint visible and near-UV light): Established. - Strong form (neutralisation, NaCl dissolution, sucrose and protein hydrolysis emit 190-260 nm light with a reaction-specific line spectrum): Refuted by calculation for the cases where the per-event energy (0.04-0.6 eV) is far below the photon energy (5-6.5 eV); no multi-event pooling mechanism is offered or tested. For the oxidations and the claimed fingerprints, Never properly tested by a neutral physical spectrometer, with an NRC negative on record.

---

Claim 3. Energy accounting: one 2020 A quantum per sucrose molecule would be 420 cal/g, about 43 times Rubner's 9.7 cal/g; Rahn offers only a reabsorption (activation chain) explanation and says the origin of the spectra "is by no means clear".

a. Mathematics. Rahn's arithmetic is correct. E(2020 A) = 6.14 eV = 9.8 x 10^-12 erg per photon; N_A/342 molecules per gram (N_A, Avogadro's number); 6.14 eV x 6.02e23/342 g = 418 cal/g [calc; matches deep read]. The ratio 420/9.7 is 43 (Rahn writes "one-fiftieth"). Per mole: 141.5 kcal/mol versus 3.3 kcal/mol. The modern enthalpy of sucrose hydrolysis is of the same order as Rubner's value (about -14 to -15 kJ/mol, about 10 cal/g, [memory, medium]), so Rubner's heat is not the weak point.

The internal problem is in Rahn's proposed resolution (his own text, [read: deep p.55-56, §5c]). In a chain where activation energy is handed from molecule to molecule and a quantum "leaks" at the wall, the leaked quantum can carry at most about (activation energy + reaction heat). With Fricke's "usual" 100 kcal/mol (about 3000 A) plus 3.3 kcal/mol this is lambda >= about 2770 A, longer than the observed 2020-2440 A lines and longer than the detector's upper cut-off (2600-2700 A). To emit 2020 A the activation energy would have to be about 138 kcal/mol, contradicting his own assumption. Measured activation energies of sucrose hydrolysis are much lower: about 109 kJ/mol (26 kcal/mol) for HCl catalysis and about 7 kcal/mol for free invertase [web: search result summary, https://www.researchsquare.com/article/rs-662668/v1 and NCERT-type question pages; I did not open the primary papers; a separate source gives a different figure for acid catalysis, 62.2 kJ/mol, so treat the acid value as uncertain]. In every case the figure is far below 141 kcal/mol, which makes the gap larger. Also, the activation energy is borrowed and returned, not net energy (standard thermochemistry, [memory, high]); a photon removing it from the chain would end the chain.

b. Consistent with experiment? The thermodynamic side is: heat measurements of hydrolysis are accurate to far better than the factor of 43. The photon side is: counter data (Table 10) show only a handful of counts, so the total light intensity is tiny; if one molecule in 10^10 emits a photon the energy budget is not violated in total. What is violated is the per-event energy: one molecule's reaction cannot supply a 6 eV photon from a 0.14 eV reaction. That is Lorenz's objection. A third possibility Rahn does not list is that a tiny fraction of events (side reactions, radical recombination, impurities) emits, at 6 eV from higher-energy channels; the total energy then does not matter but the specificity ("this spectrum belongs to sucrose hydrolysis") would no longer be shown.

c. Tested? The objection (Lorenz 1934) is a calculation, not an experiment. No experiment on record measures an activation-energy value from a mitogenetic spectrum, which Rahn proposed as a use [read: deep p.56]. Rahn's CSH 1934 paper presents the radiation as part of the heat of reaction that "need not necessarily appear as heat" [web, Investigator excerpt: https://symposium.cshlp.org/content/2/226.extract], which is the same view with the same arithmetic problem; I have not read the full paper.

d. If true: Either conservation of energy per elementary event would need a loophole (it does not have one; quantum emission requires energy at least hc/lambda from the transition) or the spectra do not come from the named reactions. The first is not on offer in modern physics.

Verdict. Lorenz's budget objection, as Rahn states it: Established (correct arithmetic, accepted laws). Rahn's resolution (reabsorption chain with wall leak): Refuted by calculation on his own numbers (shows lambda >= about 2770 A, not 2020 A). The claim that the energy accounting is "unresolved" is accurate as a description of Rahn's position; it is resolved by the calculation against the strong form of claim 2.

---

Claim 4. Irradiated solutions and tissues re-emit light ("secondary radiation"), which spreads along onion roots at about 30 m/s and along frog sciatic nerve at 30 +/- 3 m/s, the speed of the nerve impulse, so it may take part in nerve conduction.

a. Mathematics. I checked the numbers in the extracted text [read: PDF p.~126, text lines near 5664-5690]. A. and A. Gurwitsch (1931) used a rotating disc with two windows at 3000 rpm (50 revolutions per second, so one degree of rotation = 5.6 x 10^-5 s). Positive effects only for central angles of 25-50 degrees gave delays of 0.0014-0.0028 s; the stated total for 5 cm of root is 0.00166 s, so the speed is 0.05 m / 0.00166 s = 30.1 m/s [calc]. Arithmetic fine. Note the stated corrections: of 0.00306 s measured at the average angle 55 degrees, 0.00140 s is assigned to "local reactions" and subtracted, an assumption. (In the extracted text the nerve speed appears as "303 meters per second", an OCR dropped-symbol error for 30 +/- 3; the text on a later page, digest p.173-174, reads 30 +/- 3 m/s, and Rahn says it agrees with physiological measurement. I did not view the page image.) Nerve speed 30 +/- 3 m/s is a plausible number for a myelinated frog fibre [memory, medium], so the agreement is real arithmetic.

b. Consistent with established experiment? A signal travelling at 30 m/s is about 10^7 times slower than light: light crosses 5 cm in 1.7 x 10^-10 s [calc], so 0.00166 s is 10^7 times longer than free flight. So the process is not light propagating; it is a chain of absorption, delay and re-emission, or a chemical or excitation wave using light as one step. That makes the speed a property of the biological relay, not of the radiation. A propagation speed equal to nerve speed is a coincidence-level check: many processes (diffusion of small molecules over mm-cm, calcium waves, cardiac conduction) fall near 1-100 m/s. Established electrophysiology explains conduction in nerve by the action potential (voltage-gated Na+ and K+ channels, Hodgkin and Huxley 1952, [memory, high]) with measured speed dependent on fibre diameter and myelination; it does not need light. Also, the chain needs 6 eV photons to cross cell membranes and myelin sheaths; UV absorption lengths in tissue are micrometres (see claim 1b). Rahn himself writes that the propagation and "resonance" claims for secondary radiation have, "as far as the authors have been able to ascertain", only one published experimental proof, de Korosi 1934 [read: deep p.59]; and he warns that the secondary intensity is only known biologically, so "it is not at all certain that the above statements really indicate an increase in intensity of radiation in the physical sense" [read: deep p.59]. The "amplification 27 times" (270 s against 10 s) is that biological measure.

The energy side: secondary photons at 2450-2500 A (5.0 eV) are more energetic than the primary 3230 A (3.8 eV) photon. A re-emitting solution must supply 1.2 eV extra per photon from its chemistry (Rahn invokes a chain) [calc; read: deep p.57]. No budget is given. The experiment geometry is also unclear: the 3220-3240 A beam crosses a glass trough while the secondary is 2450-2500 A, which glass would absorb [read: deep p.57, reader's note].

c. Tested? By Gurwitsch (1931) and Latmanisowa (1932, 1933, 1934) and by A. Gurwitsch's nerve experiments; parabiosis (nerve blocked) after 2 hours' irradiation in "40 or more repeated experiments" per Step 1's key-claims list [read: digest p.173-174 (Latmanisowa 1933, 1934)]. All detection is by yeast or onion-root bioassay. The nerve and the root speeds each come from one laboratory's apparatus. No independent replication is on record in the material I have. Modern work reports light guiding in neural tissue and UPE correlated with neural activity [memory, low; not checked; these concern visible/near-IR UPE, not 2000-2500 A].

d. If true: neurophysiology would gain a parallel optical signalling channel; the conduction story would need a new component. It would not overthrow the action potential, which is directly measured by electrodes.

Verdict (split). - The measured speeds (30 m/s in root, 30 +/- 3 m/s in nerve) as a physical property of a UV relay: Never properly tested (one laboratory's bioassay, no neutral replication; not dismissed because of a refutation, but the whole literature sank with the NRC result). - That secondary radiation takes part in nerve conduction: Historically superseded but instructive (the action potential and ion-channel mechanism, directly measured, explain nerve conduction without light; the optical hypothesis was a reasonable guess in 1934 and was never given a decisive test).

---

Claim 5. Blood of healthy people radiates and blood of cancer patients (and in leukaemia, severe sepsis, old age) does not, while malignant tumours radiate strongly; blood radiation can serve as a cancer test and cancer may be a radiation disorder (Gurwitsch, Gesenius, Heinemann, Protti; Roffo's UV rats cited as a possible link).

a. Mathematics. None; biology and statistics. The book's evidence is clinical series with yeast as detector. Heinemann's yeast method requires the yeast in lag phase and has two yeast strains [read: digest p.84-85]. The book gives "mitogenetic effects more than 3 times the experimental error" for Heinemann's method and treats yeast bud increases below 15% as error [read: digest p.204]. In the Outlook Rahn writes that critics' main complaint is the reporting of only the "induction effect" (relative increase) with no raw data [read: digest p.205], which concedes the problem. Rates of correct diagnosis (sensitivity, specificity, controls, blinding) are not given in the digest passages I read ("positive effect excluded tumours with certainty", "repeatedly prompted tumour search proving tumours", [read: digest p.194-195]); these are anecdotes of selected cases.

b. Consistency with established experiment. - Established: ultraviolet is carcinogenic to skin (UVB, about 280-320 nm, and UVA; sunlight-exposed skin gets most skin cancers [memory, high]). Roffo's rats (all of 150 animals with tumours at 4 months under unfiltered UV, 52% of 700 in sunlight at 10-11 months, none under a tungsten lamp, [read: digest p.197-198]) and his skin-cancer distribution (face 95.6% of 5000 cases) are consistent with this. But the book itself says the author could not tell whether Roffo's UV included the 1800-2600 A range, and ordinary sunlight at the ground does not contain wavelengths below about 290 nm [memory, high]. So Roffo's results show that UV causes cancer, not that mitogenetic rays do. - Not established: blood "radiating" as a cancer marker. Cancer cells are typically glycolytic [memory, high: Warburg effect], while Rahn says blood sugar and glycolysis rate are not greatly altered in cancer patients [read: digest p.195]; the book's own explanation for the missing blood signal is therefore ad hoc (Protti's idea that low blood radiation causes cancer is a separate, testable guess). A blood factor that quenches the signal (the Soviet "cancer quencher") is a plausible chemical story, which the Investigator reports Gurwitsch-school authors developed [web, Investigator via Naumova 2018]. - Mainstream cancer diagnostics did not adopt the blood radiation test; no modern literature I can verify shows a validated version [memory, medium; not checked on the web].

c. Tested? By Gurwitsch's school, Siebert (several hundred normal bloods plus carcinoma patients, counter method, [read: digest p.106-107]), Gesenius (3 years Berlin clinic), Heinemann (Frankfurt hospital, and London), Protti, Klenitzky. Soviet work continued: Pesochensky's 1942 dissertation and later dissertations to 1955, A. A. Gurwitsch's books of 1968 and 1974 [web, Investigator, via Naumova 2018 and PubMed records, titles only, not read]. No double-blind, prospectively controlled trial is in the record I have; Rahn himself writes that "Heinemann after successful cancer diagnosis ... suddenly experienced complete lack of yeast reaction" for a period, the same unexplained detector failure noted for all laboratories [read: digest p.108-109].

d. If true: it would imply a quantifiable optical or chemical signature of blood in cancer, which is not impossible (blood biochemistry in cancer is altered). What would have to change is the claim of UV photons from blood glycolysis at 5 positive 10 A strips (claim 2); a chemical quenching or detector-side explanation would not need any change in physics.

Verdict (split). - Blood radiation as a diagnostic cancer test: Never properly tested (clinical series by advocates; no neutral, blinded trial on record; not shown to be tested and failed; it went out of use in the West along with the whole field after about 1937, i.e. dismissal without testing of the diagnostic claim itself is probable but I did not find a source that says so). - "Cancer may be a radiation disorder" (Protti; mitogenetic rays): Not testable as stated (no defined quantity). The narrower claim that UV light causes skin cancer: Established, but through ordinary UV, not mitogenetic radiation.

---

Claim 6. The radiation shapes development and health: it speeds tadpole wound healing and limb growth, distorts sea urchin larvae and yeast forms, links neighbouring cells, and a harmful form from menstrual blood, fingertips and oxy-cholesterol kills some yeasts through quartz.

a. Mathematics. Effect sizes from the digest, with the caveat that these are digest summaries: tadpole forelimb growth 4.1% +/- 1.1 (33 tadpoles) and 7.4% [read: digest p.182-183]; tail wound healing 26% less (upper, non-irradiated wounds) or 33% more (lower, irradiated at once), with an effect only when irradiation began immediately [read: digest p.190-191]; Julius's 56 pairs, 5.3 times probable error [read: digest p.97]. 4.1/1.1 is 3.7 errors, but the errors are sample-to-sample, with no mention of blinding. In Rahn's oxy-cholesterol experiment, 4 of 5 batches killed Mycoderma the first day through quartz, not the second; the one failure is explained by a delay in separating the product [read: PDF text lines near 8927-8940]. That explanation (unverified, after the fact) could excuse any failure. Note also: the book's Table 15 chemical equation is unbalanced [read: deep p.66], a typesetting slip.

b. Consistent with established experiment? - Developmental effects: ultraviolet at these doses is not known to speed limb growth; the claimed effects are 4-7% (limb) and 26-33% (wound) with small groups and one laboratory (Blacher's, Moscow). I can identify no mainstream result that confirms or opposes; the effect is of the kind that unblinded cohorts and temperature or handling differences can produce. - Magrou and Magrou's sea urchin larvae through quartz but not glass (1928) [read: digest p.100]: Magrou and Reiss (1931) themselves proposed an electrical explanation (an insulator between source and eggs, with a metallic connection preventing the effect) [read: digest p.100-101]. That is an alternative explanation from within the field, which Rahn reports but which was never resolved in the book. - Harmful human radiation (fingertips, nose, eyes, saliva, menstrual blood): the 1932 AAAS talk (New York Times 25 June 1932: eye, nose and fingertip radiation killed yeast; right hand stronger; variation between people; saliva of one woman stimulated while her daughter's and husband's killed) [web, Investigator: https://www.nytimes.com/1932/06/25/archives/evil-eye-is-found-to-be-based-on-fact-radiations-from-human-body.html, a newspaper account, not Rahn's text]; Papers Past 1932 reprints (killed "in five minutes", chest does not radiate, rays work through quartz, "preliminary stages") [web: https://paperspast.natlib.govt.nz/newspapers/HAWST19320910.2.120]. In the book: the effect is rare (many persons tested; most pronounced in a man recovering from herpes zoster, who killed yeast through quartz in 15 minutes for about 6 months; one test of one hypothyroid patient; one author during a sinus infection) and acts mainly on one species, Saccharomyces mycoderma punctisporus, "other yeasts showed slight retardation but never complete killing" [read: digest p.111-112]. A killing UV would have to be at far-UV dose rates of kJ/m2 to kill yeast in 5-15 minutes [memory, medium: lethal UV-C dose for yeast is of the order tens to hundreds of J/m2]; at 100-1000 quanta per cm2 per second (about 1e-15 W/cm2 at 6 eV [calc]) the delivered dose in 15 minutes is about 1e-12 J/cm2, i.e. 1e-8 J/m2, around ten orders of magnitude below the lethal dose [calc, with the lethal dose a medium-confidence memory value]. So even granting the emission, the quoted mitogenetic fluxes cannot kill; a "harmful radiation" at the fingertip would have to be many orders more intense than anything in the rest of the book, or the effect is chemical or thermal. This inconsistency is not noted in the book. With a 2 mm quartz plate between finger and yeast, the 1.5 minutes vs 5 minutes direct contact figure (shorter through quartz) actually favours a non-radiative factor in direct contact and a second factor through quartz, which Rahn does not discuss [read: digest p.110-111]. Gaps in a quartz plate set on a ring allow volatile molecules to pass; the book calls the exclusion of chemicals incomplete in the early experiments ("most experiments lacked chemical exclusion from volatile substances", [read: digest p.175]). - Oxy-cholesterol: Macht and Lubin's "menotoxin" is a chemical substance; Rahn argued it "radiates". Oxidising cholesterol products can chemiluminesce at low levels [memory, low]; the same dose problem applies. Macht (1949) records that "Rahn and his co-workers claimed that oxycholesterol emits destructive ultraviolet rays" [web, Investigator: https://psycnet.apa.org], without further assessment in the notes I have.

c. Tested? By the Blacher group (500+ tadpoles in the wound study), Christiansen (1929), Rahn and Barnes (1933), Ferguson (1932), Macht and Lubin, and by others who found negative results on menstrual toxicity (Sanger 1921, Frank 1921, Poland and Dietl 1924) [read: digest p.109-110]. The Investigator's lead on the identity of "Barnes" is relevant to attribution: the oxy-cholesterol work is credited to "BARNES and RAHN (1933)", and the book's bibliography lists "RAHN and M. N. BARNES, 1933 ... criteria of death in yeast" and thanks "Mrs. Margaret N. Barnes" [read: PDF text near lines 241, 9619-9621, 8927]; this suggests M. N. Barnes, not the physicist Sidney W. Barnes, but the book does not say which. I found no independent replication.

d. If true: biology would need an optical channel that regulates wound healing and growth (not shown by any modern test I can name); the human "evil eye" effect would need a source of UV-C at lethal intensity from skin, i.e. a non-thermal process in skin releasing 6 eV per photon at a dose ten orders above the book's own intensities. This would force a rewrite of dermatological physiology; nothing in modern dermatology suggests it.

Verdict (split). - Mitogenetic UV speeds tadpole limb growth and wound healing, and links neighbouring cells: Never properly tested (small unblinded series from one school; the effect sizes are small; not independently replicated). - Sea urchin larva distortion through quartz: Never properly tested, with a competing explanation from within the field (Magrou and Reiss, electrical). - Harmful human radiation (fingertips, saliva, menstrual blood, oxy-cholesterol) killing yeast through quartz: Never properly tested as a radiation effect; Refuted by calculation if the radiation is the book's 100-1000 quanta per cm2 per second UV (dose about 1e-8 J/m2 against a lethal dose of tens to hundreds of J/m2, [calc + memory medium]). Whether something from fingers kills Mycoderma is a separate, chemical, untested question; Rahn did not neutrally exclude volatiles or secretions.

---

Claim 7. Failed replications do not disprove the effect, because detectors respond only in a particular physiological stage and fail unpredictably for days or weeks (54 failures in 373 Gurwitsch experiments), and because biological error is large and varies between laboratories; Rahn also points out the physical counter evidence (Rajewsky, Frank and Rodionow, Siebert and Seffert) and its critics (Lorenz's three error sources).

a. Mathematics. 54/373 is 14.5% [calc], close to the 15% increase that Gurwitsch took as the experimental error of the yeast bud method [read: digest p.204], i.e. a failure rate about equal to the effect threshold. Schwemmle's ~200 onion-root results: probable error +/-10% for 500 mitoses, all claimed effects outside that range [read: digest p.71-72]; the book also notes error of 10% in some hands and 50% in others [read: digest p.204], so the probability that a positive effect is real depends on whose hands. Rahn's key remark, "parallel experiments the same day provide different evidence than experiments spread over time", correctly states that same-day replicates share a cause of correlated error, so the effective sample is the number of days [read: digest p.109]. That is statistically sound, but it cuts the other way too: if unknown causes make whole days or weeks fail, the proportion of days showing an effect is the relevant count, and the book does not give it. The claim is logically coherent but, as stated, can excuse any negative result.

b. Consistent with established experiment? Biological detectors do depend on culture age, medium, temperature and season; this is textbook microbiology and Rahn's 1934 paper (J. Bacteriol. 28:153-158, "The disagreement in mitogenetic experiments, a problem in bacterial physiology", [web, Investigator: https://pmc.ncbi.nlm.nih.gov/articles/PMC533662/; not read]) states exactly this. Rahn also correctly states that a negative "disproof" requires finding the alternative cause [read: digest p.201-203]. Against: the same flexibility defends any hypothesis. The book's solution of precise protocols (age of yeast, medium, 24-hour raisin agar at 30 C, [read: digest p.79-82]) is in principle a fair answer: a stated protocol and a stated pass criterion. Hollaender and Claus's negative result is at issue because Volodyaev and Beloussov (Gurwitsch tradition, interested party) say their conditions had not been verified; the NRC authors had attempted physical detectors as well [web, Investigator].

c. Tested? The proposition that detector state explains failures has not been tested by a pre-specified protocol and a blinded counterparty, as far as I can tell. It could be: have an independent lab run Rahn's yeast protocol, with the culture age and medium fixed in advance, test and sham exposures blind, day-level analysis.

d. If true: it changes statistical practice, not physics: if an effect exists only in a hard-to-control window, replication needs a stated definition of that window; if that definition can be altered after a failure, the claim is unfalsifiable.

Verdict (split). - As a blanket defence of all negative results: Not testable as stated (any failure can be assigned to physiological state without a defined, checkable criterion). - The weaker claim that these detectors are state-sensitive and specific protocols are needed: Plausible, untested as applied to this effect; in general microbiology it is Established.

---

2. Investigator's findings: weighed and kept

Kept in the record whatever the verdict, with the source party and any stake:

---

3. The silence after 1937: competing explanations (physicist's contribution)

Fact [web, Investigator]: Western publication on mitogenetic rays falls after the 1937 NRC report and Gerlach's 1939 Naturwissenschaften review ("Existiert die mitogenetische Strahlung?"); Soviet work continues to at least 1955 and A. A. Gurwitsch's reviews of 1968 and 1974. Explanations and what would decide: 1. The negative NRC study was accepted as decisive. For: NRC prestige and two-year effort. Against: its detector conditions are disputed by supporters. Deciding record: the bulletin itself and Rahn's response (Tabulae Biologicae 14, 1937). 2. Detector unreliability (Rahn's own claim 7 turned against him). For: the book's own 14.5% failure rate and large error. Deciding record: day-level raw data from any lab. 3. Language and political division (Russian-language papers; 1930s purges and war). For: Soviet continuation. Deciding record: translation and citation analysis. 4. Better instruments (photomultipliers) showed ultra-weak emission only at visible/near-UV wavelengths and no far-UV. For: modern UPE literature [web]. Deciding record: spectra of yeast and onion roots at 190-260 nm with a modern photon counter (the 1974 work of Quickenden is the precedent, not read). I found nothing in the record pointing to suppression of experimental findings; the book itself notes that "investigators avoid discussing failures due to physicist skepticism" [read: digest p.109], which is a stated bias toward under-publication of negatives within the field, not a documented external suppression.

---

4. Corrections to Step 1

1. claims.md claim 7 lists "the physical counter evidence (Rajewsky; Frank and Rodionow; Siebert and Seffert) and its critics". "Counter" here means Geiger or photoelectron counter evidence in favour of the effect, not evidence against it. Rewrite as "the physical evidence from photoelectron counters". [read: digest p.105-107]
2. facts.json and claims.md quote "about 43 times Rubner's 9.7 cal/g" correctly; the book says "one-fiftieth". Verified: 418/9.7 = 43. [calc] No change needed; keep the note that Rahn's text says one-fiftieth.
3. claims.md claim 2 states that neutralisation and NaCl dissolution each "emit" light as part of the ordinary-biochemistry account. In the book that "proof" rests on bacterial and yeast detectors (Wolff and Ras) only; no photoelectric measurement of these two is cited (deep read, p.47, p.64-65). Add "as shown by the bacterial detector". My calculation shows NaCl dissolution is endothermic (+3.9 kJ/mol, about 0.04 eV per formula unit, [memory, high]) so the claim has no energy source in the heat of solution. This is the single clearest internal red flag among the examples.
4. claims.md claim 3 says Rahn "offers only a reabsorption (activation chain) explanation". He also offers a detector-side explanation (a single quantum triggers a cell process the cell itself powers, deep read p.62). The two are different: the first addresses the source side and fails on his own numbers (lambda >= about 2770 A), the second addresses the detector side and is untested. Add the detector-side route.
5. Claim 4 speed: the extracted text on the PDF page for Latmanisowa shows "303 meters per second" (OCR dropped the plus-minus sign); the digest p.126 repeats "303". The correct figure is 30 +/- 3 m/s (digest p.173-174; agrees with Rahn's "30 m/s" for roots). I did not check the page image. If a later step quotes "303", correct it.
6. Barnes: Step 1 attributes "Barnes and Rahn 1933" experiments to the book's physics-chapter author. The book's preface thanks "Mrs. Margaret N. Barnes" and its bibliography lists "Rahn and M. N. Barnes 1933, criteria of death in yeast, J. Gen. Physiol. 16:579" with "see also BARNES". That points to M. N. Barnes as Rahn's collaborator, as the Investigator suggested. The text I read does not decide it. Mark as unresolved, probably M. N. Barnes, and do not assign Sidney W. Barnes' physics authority to the oxy-cholesterol work.
7. facts.json mainstream_conflict gives item (1) "one photon per molecule energetically impossible" (correct) but says the radiation of UV from some reactions "is physically ordinary". Qualify: visible and near-UV chemiluminescence is ordinary; 190-260 nm emission from reactions releasing under 1 eV per event is not.
8. The claims (1) in claims.md mixes two intensities. Rahn's text gives "100-1000 quanta per cm2 per second" for the strongest sender [read: digest p.137]; Rajewsky's physical estimate was 10-100 quanta per cm2 per second at 2300 A and Frank and Rodionow up to 2000 [read: digest p.107]. Both worth stating beside claim 1.

---

5. What one experiment or calculation would settle the main disputes

• Claims 1-2: a modern photon-counting spectrometer (photomultiplier with UV-C sensitivity, quartz optics) on yeast in the lag phase and on onion roots, 190-300 nm, with blanks. If 190-260 nm photons at 10-1000 per cm2 per second appear, claim 1's emission half stands; if not, it falls. Then a blind test of growth stimulation using Rahn's own protocol (24 h raisin agar, 30 C) with day-level analysis.
• Claim 3: no experiment needed; the per-event budget decides. A calculation to confirm: reaction enthalpy per event against hc/lambda for each claimed line.
• Claim 4: measure the time delay in a nerve or root with a photomultiplier and a pulsed UV source; the chain must show a threshold-limited relay, not a speed set by the instrument window.
• Claim 5: a prospective blinded trial of blood samples (cancer, benign, healthy) with the detector protocol fixed in advance, ROC analysis (sensitivity and specificity).
• Claim 6: replicate the oxy-cholesterol or fingertip-yeast experiment with sealed quartz, a glass control, a heat control and a vapour trap.
• Claim 7: independent laboratory, same protocol, pre-defined success criterion, results reported by day.

---

Summary of verdicts

1. Weak (ultra-weak photon emission from living matter): Established. Strong (190-260 nm emission stimulating division): Tested, not supported (NRC 1937; disputed by Gurwitsch-tradition authors; partial 1997 result).
2. Weak (some reactions chemiluminesce): Established. Strong (neutralisation, NaCl dissolving, hydrolysis emit 190-260 nm light with specific spectra): Refuted by calculation (energy per event); fingerprints Never properly tested.
3. Energy objection Established; Rahn's resolution Refuted by calculation on his own numbers.
4. Speeds: Never properly tested; light-conduction in nerve: Historically superseded but instructive.
5. Blood-test: Never properly tested; "cancer is a radiation disorder": Not testable as stated; UV carcinogenesis by ordinary UV: Established.
6. Developmental effects: Never properly tested; harmful human radiation: Never properly tested and Refuted by calculation at the book's own intensities.
7. As blanket defence: Not testable as stated; state-sensitive detectors in general: Plausible, untested for this effect.