Otto Rahn and mitogenetic radiation: a primer

Source gap. The project folder holds one document: Otto Rahn, Invisible Radiations of Organisms (Protoplasma-Monographien IX, Berlin: Borntraeger, 1936), with a first chapter on the physics of radiation by Sidney W. Barnes. Its text was fully extracted (230 PDF pages, 9 of them by optical character recognition, OCR). Deep reads now cover PDF pp.44-146, 160-200: chemical sources, spectra and the energy argument (pp.44-67); the detector methods (pp.68-103); physical detection, failures and human radiation (pp.104-116); secondary radiation (pp.117-132); the single-quantum argument (pp.133-146); tissues, blood, nerve and brain (pp.160-174); development, wounds and cancer (pp.175-200). The first eight of these were added after the quality check (QE, Step 4b), and the arbiter weighed them in a "QE revision" section. PDF pp.147-159 and the physics chapter before p.44 rest on the digests plus the reviewers' own reading. The reviewers, the arbiter and the deep readers rendered key pages as images (PDF pp.42, 48, 75, 97, 105-106, 111, 118, 122, 126, 137, 157, 200 and others). Everything outside the book is known here only from the Investigator's notes on abstracts, catalogue records and one open-access review: Rahn's 1934 and 1937 papers, the 1937 National Research Council (NRC) negative study by Hollaender and Claus, Quickenden's work (1974, 1989), Wainwright et al. (1997, 1998), and the Soviet "cancer quencher" dissertations. Nobody in this review has read any of those in the original. Where a statement leans on them, it says so.

Tags. [read: p.N] = from the book, PDF page N (book page = N − 14; so PDF p.55 is printed page 41). [read: page image p.N] = read from a rendered page. [web] = Internet source with URL (mostly found by the Investigator; "via Investigator" means I did not reopen it). [memory] = recalled from training, with confidence. [calc] = computed in example.py or by the arbiter. Verdicts are the arbiter's (arbiter.md); this primer adds none.

Units used throughout. 1 Å (ångström) = 10⁻¹⁰ m = 0.1 nm. 1 eV (electron-volt) is the energy an electron gains crossing 1 volt, 1.602 × 10⁻¹⁹ J. 1 cal (calorie) = 4.184 J; 1 kcal/mol = 1000 cal per mole. 1 eV per molecule = 23.06 kcal/mol = 96.5 kJ/mol. 1 erg = 10⁻⁷ J (Rahn works in ergs and calories).

---

1. The idea in one paragraph

In 1923 the Russian biologist Alexander Gurwitsch pointed the tip of one onion root at the growing zone of another and found more dividing cells on the side that faced the first root [read: p.68-69]. He decided that living cells give off an invisible radiation that makes other cells divide, and called it "mitogenetic" (mitosis = cell division). Over the next twelve years his school and others built a large literature on it: the rays were said to be extremely weak ultraviolet light, shorter in wavelength than anything in sunlight; to pass through quartz but not glass; to come from ordinary chemical reactions such as oxidation, sugar fermentation and protein digestion, each with its own spectral "fingerprint"; to travel along nerves; and to be missing from the blood of cancer patients. Otto Rahn (1881-1957), Professor of Bacteriology at Cornell University from 1927 to 1949 [web, via Investigator: https://hdl.handle.net/1813/18707 and https://www.weberrarebooks.com], wrote this 1936 book to put the case before English-speaking scientists. He was an advocate who also did experiments with his students, but he was unusually candid: he printed the energy objection against the theory himself and admitted that his detectors stopped working for weeks at a time. One year after the book, a two-year NRC study reported that all its attempts to find the rays had failed, and Western interest faded. Today it is accepted that living tissue does emit very faint light ("ultra-weak photon emission"), but in the visible and near ultraviolet, and no neutral test has shown that it makes other cells divide. This primer teaches the physics you need to judge the claim yourself, and its worked example is the energy calculation that Rahn himself could not solve.

---

2. What you need to know first

Photons and photon energy. Light of wavelength λ (lambda) comes in packets, photons, each carrying energy E = hc/λ, where h = 6.626 × 10⁻³⁴ J·s is Planck's constant and c = 2.998 × 10⁸ m/s is the speed of light. A handy form: E [eV] = 12 398 / λ [Å]. Visible green light (5000 Å) is 2.5 eV per photon; the mitogenetic range, 1900-2600 Å, is 4.8-6.5 eV per photon. Absorption and emission happen one photon at a time, so what matters for a single molecule is the energy of one photon, not the total brightness. Read: Krane, Modern Physics, the chapter on the particle-like properties of electromagnetic radiation; or Serway, Moses and Moyer, Modern Physics.

The ultraviolet bands and what blocks them. Ultraviolet (UV) is divided into UV-A (3150-4000 Å), UV-B (2800-3150 Å) and UV-C (1000-2800 Å). The Earth's ozone layer removes UV-C, so sunlight at ground level contains essentially none [memory, high]. Ordinary window glass absorbs below roughly 3000-3500 Å; fused quartz transmits down to about 1600-2000 Å [memory, high]. Gerlach, cited in the book, found the chemical "rays" came out of quartz vessels but not glass ones, and concluded they were shorter than 3500 Å [read: p.48]. The oxygen in air absorbs strongly below about 2000 Å, and Rahn notes that radiation "below 1900 A will be readily absorbed even by air" [read: p.54]. Water, proteins and nucleic acids absorb UV-C within micrometres to millimetres; proteins absorb near 2800 Å and below 2200 Å, nucleic acids near 2600 Å [memory, high]. This matters because a mitogenetic photon has to get out of one cell and into another. Read: Hecht, Optics, the sections on absorption and dispersion.

Heat of reaction and activation energy. When a chemical reaction runs, the difference in energy between products and reactants is released (or absorbed) as heat: the heat (enthalpy) of reaction, ΔH. To start, molecules must first climb an energy hill, the activation energy E_a, and they get it back when they roll down the other side. So E_a is borrowed and returned; only ΔH is net energy given out. The fraction of molecules with thermal energy above a threshold E at absolute temperature T is about e^(−E/kT), where k = 1.381 × 10⁻²³ J/K is Boltzmann's constant (kT ≈ 0.025 eV at room temperature). That factor is the Boltzmann factor, and it is why reaction rates follow the Arrhenius law, rate ∝ e^(−E_a/RT) (R = N_A k is the gas constant, N_A = 6.022 × 10²³ per mole is Avogadro's number). Read: Atkins and de Paula, Physical Chemistry, the chapters on thermochemistry and on the rates of chemical reactions; Schroeder, An Introduction to Thermal Physics, the chapter on Boltzmann statistics.

Chemiluminescence, photochemistry, quantum yield. Some reactions make light directly instead of heat: a reaction step leaves a product molecule in an excited electronic state, which then emits a photon. Glow sticks and fireflies are examples. The photon energy is limited by the energy released in that one step. The reverse process is photochemistry: a molecule absorbs a photon and reacts. The quantum yield Φ (phi) is the number of molecules that react per photon absorbed, or, for emission, the number of photons emitted per reacting molecule. In a chain reaction one absorbed photon can start a sequence in which each step regenerates a reactive fragment, so Φ can be enormous: hydrogen and chlorine react with up to about 10⁶ molecules of HCl per photon, an example Rahn cites [read: p.60-61]. Read: Atkins and de Paula, Physical Chemistry, the photochemistry sections; Lakowicz, Principles of Fluorescence Spectroscopy, chapter 1, for excited states and emission.

Fluorescence and the Stokes shift. A molecule that absorbs a photon and re-emits light (fluorescence) normally emits at a longer wavelength (lower energy) than it absorbed, because some energy is lost as heat in between. This is the Stokes shift. Emission at a shorter wavelength than the light absorbed needs an extra energy source. This rule matters for Rahn's "secondary radiation" (section 4). Read: Lakowicz, chapter 1.

Counting statistics. If events (photon counts, dividing cells) occur at random at a steady average rate, the number N counted in a fixed time follows the Poisson distribution, with standard deviation √N. The probability that a target receives at least one hit when the mean number of hits is m is 1 − e^(−m). Older papers quote the "probable error" (PE), which is 0.6745 standard deviations: so "5.3 times the probable error" is 3.6 standard deviations. When you test many independent strips of a spectrum, some will look positive by chance (the multiple-comparisons problem). Read: John R. Taylor, An Introduction to Error Analysis, the chapters on the Poisson distribution and on significance tests.

The cell cycle and the nerve impulse. Cells divide by mitosis after passing through a sequence of controlled stages; the triggers that are known are chemical (growth factors, nutrients, contact) [memory, high]. UV-C hitting a cell usually damages DNA and causes the cell to pause division and repair, not speed up [memory, medium]. Nerve impulses travel as action potentials: a wave of opening and closing sodium and potassium channels in the membrane, described by Hodgkin and Huxley in 1952 [memory, high]. Speeds range from about 1 to 100 m/s depending on fibre diameter and myelin. Read: Alberts et al., Molecular Biology of the Cell, the cell-cycle chapter; Nicholls et al., From Neuron to Brain, the chapters on the action potential.

---

3. The textbook view

Five pieces of standard physics and chemistry carry the whole argument.

(a) Photon energy.

E = hc/λ or E [eV] = 12 398 / λ [Å]

• E = energy of one photon; h = Planck's constant; c = speed of light; λ = wavelength.

Per mole of photons (an "einstein"), multiply by N_A: a 2020 Å photon is 6.14 eV = 9.83 × 10⁻¹² erg = 141.5 kcal/mol [calc].

(b) Energy conservation per event. In one emission event, a molecule can emit a photon only of energy less than or equal to the energy available to it in that event:

E_photon ≤ E_available (per molecule, per event)

For a chemiluminescent step, E_available is roughly the energy released by that single step, ΔH per molecule, plus a little thermal energy (a few kT). There are known ways to pool energy from two events into one photon. The best known is the "dimol" emission of singlet oxygen: two excited O₂ molecules, about 0.98 eV each, collide and emit one red photon of about 1.95 eV (634 nm) [memory, high]. That shows pooling of two events is possible; pooling 40 is not a known process.

(c) Energy conservation in total. Over many events, the total radiant power cannot exceed the total power the reaction releases:

P_light = Φ × (reaction rate in molecules/s) × E_photon ≤ ΔH × (reaction rate)

• P_light = radiant power (W); Φ = photons emitted per reacting molecule (quantum yield).

This is a much weaker condition than (b) when Φ is tiny.

(d) Thermal activation. The fraction of molecules with energy above E is ≈ e^(−E/kT). Activation energy is not net energy: after the reaction it is returned to the surroundings as heat, so the net heat measured in a calorimeter is ΔH.

(e) Absorption and dose. A beam of flux F photons per cm² per second, each of energy E_photon, carries intensity I = F × E_photon (W/cm²). A target of area A receives on average m = F A t photons in time t. The energy delivered per unit area, the dose, is I × t (J/m² after converting cm² to m²). UV-C kills yeast and bacteria at doses of order tens to hundreds of J/m² [web, search summaries as checked by the arbiter: https://eom.ifa.md/m/journal/download/1837 ; https://www.nature.com/articles/211310b0.pdf; sensitivity varies widely with strain].

And the modern experimental fact. Living tissue does emit extremely faint light, now called ultra-weak photon emission (UPE) or "biophotons". It is chemiluminescence from electronically excited molecules formed in oxidative metabolism and oxidative stress (radical reactions involving oxygen, where single steps can release several eV). Reported fluxes are of order units to hundreds of photons per cm² per second at rest and up to thousands under stress, in the near-UV, visible and near-infrared, roughly 350-1300 nm [web, via physicist: https://www.frontiersin.org/articles/10.3389/fchem.2023.1140128/pdf, figures as returned by a search tool]. Photomultiplier measurements of it began with Colli and Facchini (1954) [web: https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00241/full]. The standard review (Cifra and Pospíšil 2014) is summarised as reporting photons "at near UVA, visible, and near IR spectral ranges from 350 to 1300 nm" [web, search summary checked by the arbiter: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10899412/; page not opened]. So mainstream UPE reports start near 350 nm. That is silence at 190-260 nm, not a measured absence.

---

4. Where Rahn departs

Before the points of departure, the real physics Rahn and Gurwitsch stood on. Much of it was sound.

• Quartz-versus-glass is a real optical test. Quartz passes UV-C and glass blocks it, so an effect that crosses quartz but not glass points to UV (or to anything else that differs between the two plates: heat conduction, gas tightness, electric charge). The idea of a screen test was good physics.
• Chemiluminescence was established. Haber had shown that slow oxidation of phosphorus, H₂ + O₂ and pyrogallol give off light that is not heat glow, "part of the original energy of reaction is liberated in the form of visible light" [read: p.47]. It was reasonable to ask whether some reactions also emit UV.
• Weak light from living matter turned out to be real. Rahn's intensity for the strongest senders, "about 100 to 1000 quanta per cm² per second" [read: p.137], is the same order as modern UPE figures. On intensity, the old school and the modern measurements agree.
• Chain reactions with huge quantum yields exist. Rahn's idea that one photon could trigger a large cellular response (the cell supplies the energy, the photon only pulls the trigger) [read: p.62, p.138] has real parallels: a dark-adapted rod cell in the eye responds to a single photon, using a dedicated pigment and an amplifying chemical cascade [memory, high].
• Living detectors really are state-dependent. Bacterial and yeast growth depends on culture age, medium, temperature and history; that is textbook microbiology (arbiter, claim 7).
• Some of the counter data are statistically non-trivial. In the book's Table 10 (Frank and Rodionow, Geiger-Müller counter), six of nine rows exceed twice their combined error and four exceed three times; the dark FeSO₄ + K₂Cr₂O₇ row (27 ± 1.3 against 17 ± 1.0 counts) is 6.1 standard errors [read: page image p.48; calc by arbiter]. Barth found more photoelectrons from proteolysis than from the control in 24 of 25 runs [read: p.48], a one-sided chance probability of 7.7 × 10⁻⁷ if the runs were independent and scored without bias [calc]. Julius (1935) reported 56 pairs with a ratio 1.92 ± 0.15 against 1.01 ± 0.08 [read: page image p.97], about 3.6 standard deviations if the ± are probable errors [calc]. In Table 30 (Rajewsky; Frank and Rodionow, counters on onion root, mouse carcinoma, frog muscle and heart), six of ten rows exceed three times their combined error, though the controls and sample sizes are not described [read: page image p.105; calc by arbiter]. Siebert and Seffert (1934) got higher counts from several hundred normal blood samples and none from carcinoma blood. Their counter rose with normal blood and did not rise when blood with cyanide (KCN) added was used instead. Rahn calls this the "most convincing" physical experiment [read: p.106]. The book gives no counts, no number of carcinoma samples, and does not say whether the same sample was poisoned. KCN also changes pH and gives off HCN vapour (arbiter).

Now the departures, side by side.

The exact point of departure is the second row. Everything else (fingerprints, secondary radiation, nerve conduction, cancer) depends on reactions producing 5-6 eV photons from steps that release a fraction of an eV. The worked example tests that point with Rahn's own numbers.

The first row was disputed inside the field too. Reiter and Gabor (1928, Germany; the book does not give their institution) found the effective rays not at 1900-2600 Å but at 3200-3500 Å. They reported a sharp maximum at 3400 Å, a smaller one near 2800 Å, nothing "near 2000 Å", and "antagonistic" rays at 2900-3200 Å that blocked the active band even at one-tenth of its intensity. Glass and filter tests also pointed above 3000 Å [read: p.73-74]. The Gabor here is Dennis Gabor, the physicist who later won the Nobel Prize for holography [memory, high; the book does not say so; the Investigator's lead in sites.md names him]. The Gurwitsch school repeated at 3340-3400 Å and found nothing (Frank 1929; Chariton, Frank and Kannegiesser 1930; Table 18, values between −7 and +8 against +20 to +30 at 2190-2350 Å) [read: page image p.75]. Rahn offers an explanation only for their spectrograph experiments (older parts of the root were irradiated and the effect appeared at the growing tip) and writes that the rest "have never been accounted for in a really satisfactory way" [read: p.74]. Gabor also reported that 3340 Å made unfertilised salamander and frog eggs start dividing (2 then 1 salamander egg in groups of about 5; 3 frog eggs of "a few hundred") [read: p.183-184]. So the defining property of the radiation, its wavelength, was never agreed among the people who believed in it. Both sides used the same kind of biological detector.

Two further points of departure are about method rather than physics, and the book itself supplies the evidence for both.

First, the spectral strips were scanned coarse-then-fine: "if the general spectrum has been investigated by the above-mentioned coarser methods, the negative regions need not be examined" [read: p.51], and no blank (null) spectrum with a false-positive rate is shown. Second, failed runs were identified afterwards: in Gollsheva's 373 blood experiments in Gurwitsch's laboratory, 54 failures were put down to "poor quality of the yeast culture", recognised because "all other associates using the same culture on the same day obtained negative results" [read: p.107]. Rahn's point (same-day results share causes, so the real sample size is the number of days) is statistically sound; but a failure recognised by its outcome, without a positive control fixed in advance, cannot be told apart from a true null.

Rahn himself states how weak the measures were. The percentage "induction effect" "has no quantitative value whatever" [read: p.128]. In his Table 26 the same culture gives 183 % from cell counts and 27 % from generation time, and the two controls differ from each other by 17 % in generation time [read: page image p.93]. The bud-percentage method shows decreases (−18 to −52 %) in runs where cell counts show increases (22 to 123 %), so it "fails to indicate the true growth rate" [read: page image p.130]. Duplicate yeast plates commonly differ by 40 % [read: p.93]. If polarised rays are far more effective, as Wolff and Ras reported, then "all previous measurements of intensities ... [are] practically meaningless" [read: p.128]. The arbiter's reading: these concessions lower the weight of every percentage table and raise the weight of the few designs with a control inside the same animal or the same run (Julius's pairs, Blacher's tadpoles, the cornea pairs, Siebert's runs). They do not by themselves refute the effect.

---

5. Worked example: can a sugar-splitting reaction emit a 2020 Å photon?

This is Rahn's own test case [read: p.55-56]. In the book's Fig. 24, the shortest band assigned to sucrose hydrolysis by the enzyme invertase (sucrase) starts at 2020 Å [read: page image p.51, deep read]. Sucrose (table sugar, C₁₂H₂₂O₁₁, molar mass M = 342.3 g/mol) splits into glucose and fructose. Rubner (1913) measured the heat in silver-lined Dewar flasks: 9.7 cal per gram [read: p.55]. A modern value, about −14 to −15 kJ/mol, is close to this [memory, medium], so the heat is not the weak point.

example.py reproduces every number below. Run: .venv/bin/python review/Rahn-Otto/example.py.

Step A: the photon's energy.

E = hc/λ = (6.626 × 10⁻³⁴ J·s)(2.998 × 10⁸ m/s) / (2020 × 10⁻¹⁰ m) = 9.83 × 10⁻¹⁹ J = 9.83 × 10⁻¹² erg = 6.14 eV.

Rahn prints 9.8 × 10⁻¹² erg [read: page image p.55]. Per mole: 6.14 eV × 23.06 = 141.5 kcal/mol.

Step B: Rahn's calculation (one photon per molecule). One gram of sucrose holds N_A/M = 6.022 × 10²³ / 342.3 = 1.76 × 10²¹ molecules. If each emits one 2020 Å photon:

1.76 × 10²¹ × 9.83 × 10⁻¹⁹ J = 1730 J = 414 cal per gram.

Rahn gets 420 cal/g (he used N_A = 6.1 × 10²³ and M = 342). Against Rubner's 9.7 cal/g, the ratio is 414/9.7 = 43. Rahn writes "one-fiftieth"; the true figure is about 1/43 [calc]. Rahn: "It seems impossible that appreciable amounts of energy could have been lost by the method used" [read: p.55].

Step C: the same comparison per event. Divide the heat by molecules:

sucrose: 9.7 cal/g × 342.3 g/mol = 3.32 kcal/mol = 0.144 eV per molecule.

So one 6.14 eV photon needs the energy of 43 hydrolysis events gathered into one molecule at one instant. For the other "sources" in the book [calc]:

The neutralisation heat (about 57-58 kJ/mol) was confirmed on the web by the arbiter [web: https://en.wikipedia.org/wiki/Enthalpy_of_neutralization]; for NaCl the arbiter's search gave values from slightly exothermic to about +6 kJ/mol, so "about zero to slightly endothermic" [web, search summary]. Rule (b) says no single event can do this. This is Lorenz's 1934 objection, which Rahn reports fairly.

Step D: Rahn's rescue, checked on his own numbers. Rahn proposes that activation energy is the hidden reservoir [read: p.56]. Quoting Fricke (1934), he takes a "usual" activation energy of 100 000 cal/mol, equal to a photon of about 3000 Å. His picture: a molecule absorbs E_a, hydrolyses, and hands on E_a + ΔH to the next molecule, a chain; at the vessel wall a few of these packets "leak" out as UV, "of a wave length equal or shorter than that required for activation".

The most energy one hydrolysed molecule can hand on is E_a + ΔH:

100 + 3.32 = 103.3 kcal/mol = 4.48 eV → λ = 12 398 / 4.48 = 2767 Å.

That is the shortest wavelength the chain can leak, and it is 750 Å longer than the 2020 Å it is supposed to explain. It is also outside the detector's own range: Rahn says radiation above 2600 Å (2700 Å on another page) gives no mitogenetic effect [read: p.54, p.64]. To reach 2020 Å, E_a would have to be 141.5 − 3.3 = 138.2 kcal/mol [calc], which contradicts the 100 kcal/mol he assumed. Measured activation energies for sucrose hydrolysis are far lower: about 26 kcal/mol with acid and about 7 kcal/mol with invertase [web, search snippets checked by the arbiter: https://pmc.ncbi.nlm.nih.gov/articles/PMC2664835 ; values vary between sources]. And step (d) of the textbook view applies: activation energy is returned, so a packet that leaks out at the wall ends its chain; it is not free energy.

The Fricke numbers themselves check out. kT at 293 K = 9.7 × 10⁻²² cal per molecule (Fricke: about 10⁻²¹); 100 kcal/mol is 1.66 × 10⁻¹⁹ cal per molecule (Fricke: 1.5 × 10⁻¹⁹); the Boltzmann factor for 10⁻¹⁹ cal is e^(−103) = 1.2 × 10⁻⁴⁵ (Fricke: 10⁻⁴³) [calc]. The physics Rahn quoted was right; the chain built on it does not reach the number.

Step E: the total budget, done properly. Step B assumed every molecule emits. What if only a few do? Take Rahn's strongest intensity, F = 1000 photons/cm²/s, at 2300 Å (5.39 eV = 8.64 × 10⁻¹⁹ J):

I = 1000 × 8.64 × 10⁻¹⁹ J = 8.6 × 10⁻¹⁶ W/cm².

Suppose 1 g of sucrose is hydrolysed in one hour inside a sphere of radius 1 cm (area 4π × 1² = 12.6 cm²):

light out: 8.6 × 10⁻¹⁶ × 12.6 = 1.1 × 10⁻¹⁴ W heat: 9.7 cal × 4.184 J/cal / 3600 s = 1.13 × 10⁻² W light / heat = 1.0 × 10⁻¹²

molecules reacting per second: 1.76 × 10²¹ / 3600 = 4.9 × 10¹⁷ photons per second: 1000 × 12.6 = 1.26 × 10⁴ quantum yield Φ = 1.26 × 10⁴ / 4.9 × 10¹⁷ = 2.6 × 10⁻¹⁴ [calc].

So at the intensities actually claimed, the light would cost about a millionth of a millionth of the reaction's heat. The total budget (c) is satisfied for any Φ below 1/43 = 2.3 %. The 43-fold "paradox" exists only at Φ = 1, which was Lorenz's premise ("each reacting molecule must emit at least one quantum" [read: p.55]), not something measured.

But the per-event rule (b) is untouched. Each photon that is emitted, however rare, still needs 6 eV in one place at one time, and a single hydrolysis step supplies 0.14 eV. A rare UV-C photon would have to come from some other, more energetic step (a radical reaction, an oxidation, an impurity), which is the first of Rahn's own two alternatives: "the spectra do not originate from the reactions for which they are considered specific, but from some quantitatively unimportant side reaction" [read: p.55]. That would remove the specificity on which the "fingerprints" rest.

Step F: the detector side. Rahn estimated how often a yeast cell, about 6 × 7 μm (area A = 4.2 × 10⁻⁷ cm²), would be hit at 1000 photons/cm²/s [read: p.137]:

hit rate = F A = 1000 × 4.2 × 10⁻⁷ = 4.2 × 10⁻⁴ per second; mean of one hit after 1/(4.2 × 10⁻⁴) s = 40 min.

His arithmetic is right. He then reads the peak effect at 20 minutes as evidence that one quantum per cell is enough. With Poisson statistics, the mean number of hits at 20 minutes is 0.50, so 1 − e^(−0.50) = 40 % of cells have been hit at least once; at 40 minutes, 64 % [calc]. His own data do not follow the hit count: buds after 2 hours were 21.0, 25, 36.5 and 24.5 % for 0, 10, 20 and 40 minutes through quartz (22.5, 23, 37.5, 27.5 % direct), so only 20 minutes responds, and the 40-minute dose, where more cells were hit, shows almost nothing. No error limits and no glass row are given [read: page image p.137]. He drops the single-quantum idea on the next page anyway, in favour of the culture medium itself becoming a "secondary sender" [read: p.138].

Dose: at 100-1000 photons/cm²/s for 15 minutes the energy delivered is

100 × 9.83 × 10⁻¹⁹ J × 10⁴ cm²/m² × 900 s = 8.9 × 10⁻¹⁰ J/m² up to 7.8 × 10⁻⁹ J/m² (1000 at 2300 Å).

A lethal UV-C dose for yeast is of order 100 J/m² (130-440 J/m² in the sources the arbiter checked) [web, as in section 3(e)], so the gap is a factor of 10¹⁰ to 10¹¹ [calc]. That number is the basis of the arbiter's verdict that the book's "harmful" human and oxy-cholesterol radiation, if it were UV of the book's intensity, could not kill yeast (claim 6). It does not rule out that something else from skin or oxidised cholesterol (vapour, heat) affected the yeast. The calculation does not use the kill times, so it does not depend on them.

What the example shows. 1. Rahn's arithmetic is correct to within rounding. 2. "One 2020 Å photon per sucrose molecule" is impossible by a factor of 43: refuted by calculation. 3. Rahn's activation-chain rescue fails on his own numbers (2767 Å, not 2020 Å). 4. Rare emission at the claimed intensity costs nothing in total energy; the open problem is the 6 eV needed in each single event. 5. At the claimed intensities the dose to a detector cell is about one photon per cell per 40 minutes, so any effect would have to be a single-photon trigger with an amplifier inside the cell: possible in principle (vision), never demonstrated for yeast.

---

6. Evidence table

Verdicts are the arbiter's (Step 3, including its "QE revision" section, which changed wording but moved no verdict along the scale), using the review scale. "Party" means the source has a stake (an advocate's own laboratory, or authors defending the tradition).

---

7. What a learner should take away

What is solid. - Living things do glow, very faintly. The intensity Rahn reported for his strongest senders (100-1000 photons/cm²/s) is about what photomultipliers measure today. The Gurwitsch school was early to a real phenomenon, even if its wavelength and its biological role were not confirmed. - Rahn's arithmetic, and the Fricke activation-energy figures he quotes, are correct. His statement that the origin of the spectra "is by no means clear" was accurate. - The quartz-versus-glass screen, the rotating-disc timing and the strip spectrograph are clever experimental ideas; Rahn and Barnes explain them clearly. - Several individual data sets in the book (Table 10's and Table 30's counter excesses, Barth's 24 of 25, Julius's 56 pairs, the 15 cornea pairs, Suchow and Suchowa's 25 of 25, Blacher's tadpole limbs, Siebert and Seffert's blood counts) are statistically significant on their face. Their weakness is design (blinding, artefacts, controls) and lack of independent replication, not arithmetic.

What is wrong, and why the mistake is instructive. - Per-event energy is the binding constraint, not total energy. The famous "43 times" paradox disappears if only a tiny fraction of molecules emit (step E). What does not disappear is that each emitted photon needs 6 eV in one place, while each reaction step supplies 0.14 eV. The lesson generalises: whenever someone claims a high-energy quantum from a low-energy process, ask for the energy per event, not the energy per gram. (The same reasoning applies to photoelectric thresholds, to the Stokes shift, and to every "over-unity" device that claims to concentrate ambient energy.) - A rescue must be checked against its own numbers. Rahn's chain mechanism was offered in good faith and sounds plausible; one line of arithmetic (step D) shows it gives 2767 Å, outside his own detector's window. - A biological detector cannot calibrate itself. When intensity, wavelength and existence are all read from whether yeast grew a little faster, and the detector "fails" on some days with the failures recognised afterwards, every result can be explained. Rahn saw this himself ("It is not at all certain, therefore, that the above statements really indicate an increase in intensity of radiation in the physical sense" [read: p.59]). - Multiple comparisons. Scanning 60 strips, many reactions and many exposure times, and keeping the positives, will produce "lines" by chance unless the same lines recur in independent runs. - Check the page, not the text layer. The OCR text of PDF p.111 reads "1.5minutes" where the printed page says 15 minutes. That one dropped space made the fingertip kill look faster through quartz than over an open ring, and two reviewers (physicist and contrarian), the arbiter and the first version of this primer discussed an "anomaly" that the page does not contain. The quality check caught it on the page image. When a surprising number carries an argument, look at the original. - Screens test more than one thing. Quartz and glass differ in more than UV transmission; the sea-urchin larva case, where paraffin-insulated glass also passed the effect, shows the test can mislead [read: p.100-101].

What is genuinely open. - Did the NRC study use the detector conditions the positive laboratories required? The proponents say no; the bulletin has not been read here (open question 1 in the arbiter). - Is there any 190-260 nm emission from yeast, onion root or blood at 10-1000 photons/cm²/s? A modern UV-sensitive photon counter could settle this in a few weeks; no such spectrum is on record in what was read. - What produced the detector responses to neutralisation and salt dissolving, if not 6 eV photons (heat, vapour, charge, chance)? Did the lanolin-sealed quartz chamber of the fingertip test stay sealed, and did heat or moisture pass through the 2 mm of quartz? - Which band, if any, is the active one: Gurwitsch's 1900-2600 Å or Reiter and Gabor's 3200-3500 Å? The dispute was never resolved inside the field. - Do regular light pulses lower the threshold about 30-fold when a physical UV-C source of known flux replaces the biological sender, and does daylight conditioning change a yeast sender's measured UV-C output? Both claims are specific and could be tested today. - What window area and photoelectric yield did Rajewsky's counters really have? Without them the counter evidence cannot be checked against the claimed fluxes. - Why the record goes quiet after 1937-1939. About 600 papers had appeared by 1937 [web, via Investigator: Nature 140:1007]. Western publication then falls away (Gerlach 1939, "Existiert die mitogenetische Strahlung?"; 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, and photon counting resumes under new names in the 1950s-1970s. Rahn's own publications on the subject stop after 1937 as far as the Investigator found; he went on to publish Mathematics in Bacteriology (1939) and Microbes of Merit (1945) [web, HathiTrust catalogue via Investigator], which nobody here has read, so whether he later changed his view is unknown. The arbiter sets out competing explanations: (1) the NRC negative was taken as decisive; (2) the biological detector was judged unreliable (Rahn's own admissions); (3) instrument change, as photomultipliers let the field relabel itself "ultra-weak photon emission"; (4) language, war and Cold War division; (5) suppression of results or funding, for which nothing was found in the record, though it is not excluded; (6) loss of interest without a decisive test. The arbiter finds 1-3 best supported, 4 a live contributor, and 5 neither evidenced nor excluded. The records that would decide (the NRC bulletin and correspondence, Rahn's 1937 Tabulae Biologicae review, the Cornell Rahn papers, the Science Service records of the 1932 wire story, Hollaender's papers) have not been obtained. - Rahn's one-photon trigger (the cell supplies the energy, the photon only releases it) is physically possible and was never tested directly.

What would settle the main question (arbiter): a pre-registered, blinded test in two laboratories, with the proponents fixing beforehand the culture age, medium, induction time and a calibrated UV-C positive control at 10-1000 photons/cm²/s that declares a "responsive day"; quartz, glass and paraffin-insulated glass screens assigned blind; analysis by day. The designs to copy are the book's best-controlled series, where the control sits in the same animal or run: Julius's pairs, Blacher's tadpole limbs and wounds, the cornea pairs and Siebert's runs (arbiter, QE revision). Alongside it, a photomultiplier spectrum of lag-phase yeast and onion roots from 190 to 300 nm, with blanks.

---

8. Check yourself

1. A "mitogenetic" photon at 2500 Å: what is its energy in eV? How many NaOH + HCl neutralisation events (55.8 kJ/mol each) would have to pool their energy to make one?
2. Gurwitsch timed secondary radiation with a disc turning at 3000 revolutions per minute. When the root was lengthened by 2.5 cm, the window of positive effect shifted by about 15 degrees of rotation. (a) What speed does that give? (b) How long would light take to cross 2.5 cm, and what does the comparison tell you about the mechanism?
3. Julius reported a quartz-irradiated ratio of 1.92 ± 0.15 against a control of 1.01 ± 0.08, where ± are probable errors. Compute the difference in units of its combined probable error, then in standard deviations. What would you still want to know before believing the effect?
4. A 10 Å strip scan covers 60 strips. If each strip has a 5 % chance of reading "positive" when there is no radiation at all, what is the probability of seeing 5 or more positive strips by chance? What does this imply about the "5 of 60" glycolysis fingerprint?
5. In the nucleic-acid experiment, 3220-3240 Å light went in and 2450-2500 Å "secondary radiation" came out. Why can this not be ordinary fluorescence, and how much extra energy per photon must the solution supply?

---

9. Further reading

The inventor's and his school's own works - Otto Rahn, Invisible Radiations of Organisms, Protoplasma-Monographien IX, Borntraeger, Berlin, 1936 (with S. W. Barnes's physics chapter); a 1944 printing is listed in HathiTrust [web, via Investigator: https://catalog.hathitrust.org/Search/Home?lookfor=Rahn%2C+Otto&type=author]. The book read for this review. - O. Rahn, "The disagreement in mitogenetic experiments, a problem in bacterial physiology", J. Bacteriol. 28:153-158 (1934), free at https://pmc.ncbi.nlm.nih.gov/articles/PMC533662/ [web, via Investigator; not read]. - O. Rahn, "The physico-chemical basis of biological radiations", Cold Spring Harbor Symp. Quant. Biol. 2:226-240 (1934), excerpt at https://symposium.cshlp.org/content/2/226.extract [web, via Investigator]. - O. Rahn, "Mitogenetic radiation", Tabulae Biologicae 14:1-35 (1937), written after the book, possibly after the NRC study [web, cited by Naumova et al. 2018; not located]. - A. G. Gurwitsch and L. D. Gurwitsch, the school's summaries (1934, cited by Rahn as containing the error data [read: p.54]) and "Ultra-violet chemi-luminescence", Nature 143:1022 (1939) [web, title via Investigator]. - Cornell memorial statement for Otto Rahn, 1881-1957: https://hdl.handle.net/1813/18707 [web, via Investigator; not read]. - O. Rahn, Mathematics in Bacteriology (1939) and Microbes of Merit (1945), his later books, possible records of his view after the NRC study [web, HathiTrust catalogue via Investigator; not read].

Related work of the same period (not Rahn's school) - W. Stempell, Die unsichtbare Strahlung der Lebewesen (Jena, 1932), a German book with nearly the same title, using Liesegang rings and hydrogen peroxide as detectors [web, via Investigator; not read]. - Reiter and Gabor, Zellteilung und Strahlung (Siemens, 1928), the rival 3200-3500 Å results [web, via Investigator; not read]; the Gabor is Dennis Gabor [memory, high; named in sites.md].

A later popular account - P. Tompkins and C. Bird, The Secret Life of Plants (1973), which mentions Rahn and his book and links him to T. G. Hieronymus's "eloptic" radiation devices [web, via Investigator: archive.org listing seen in a search, not opened]. A popular book written decades later, not a primary source; what it says Rahn did has not been checked.

The critics and the later record - A. Hollaender and W. Claus (Claus's initials not confirmed here), An Experimental Study of the Problem of Mitogenetic Radiation, National Research Council Bulletin 100 (1937); news summary in Nature 140:1007 (1937), https://www.nature.com/articles/1401007c0 [web, via Investigator]. - A. Hollaender, "The problem of mitogenetic rays", in B. M. Duggar (ed.), Biological Effects of Radiation (1936), pp. 919-958; Taylor and Harvey, Biol. Bull. 61:280-293 (1931) [web, citations from Naumova et al. 2018 via Investigator; not read]. - W. Gerlach, "Existiert die mitogenetische Strahlung?", Naturwissenschaften (1939) [web, title via Investigator]. - I. Volodyaev and L. V. Beloussov, "Revisiting the mitogenetic effect of ultra-weak photon emission", Frontiers in Physiology 6:241 (2015), open access, https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2015.00241/full. The proponents' case against the negative studies (authors in the Gurwitsch tradition). - Naumova et al., review of 1923-1948 mitogenetic research including the Soviet "cancer quencher" work, J. Biomedical Photonics & Eng. 4(4):040201 (2018), https://jbpe.ssau.ru/index.php/JBPE/article/view/3308 [web, via Investigator]. - M. Wainwright et al., "Partial evidence for the existence of mitogenetic radiation", Microbiology 143:1-3 (1997), https://pubmed.ncbi.nlm.nih.gov/33711853/; M. Wainwright, "Historical and recent evidence for the existence of mitogenetic radiation", Perspect. Biol. Med. 41:565-571 (1998) [web, via Investigator; not read]. - T. I. Quickenden and S. S. Que Hee, weak luminescence from yeast and mitogenetic radiation (1974); Quickenden et al., "An attempt to stimulate cell division in Saccharomyces cerevisiae with weak ultraviolet light", Radiation Research 117:145-157 (1989) [web, citations via Investigator and contrarian; not read].

Mainstream textbooks and reviews - Krane, Modern Physics (photons and E = hc/λ). - Atkins and de Paula, Physical Chemistry (thermochemistry, Arrhenius law, activation energy, photochemistry, quantum yield, chain reactions). - Schroeder, An Introduction to Thermal Physics (Boltzmann factor). - Lakowicz, Principles of Fluorescence Spectroscopy, chapter 1 (excited states, Stokes shift). - John R. Taylor, An Introduction to Error Analysis (Poisson statistics, significance). - M. Cifra and P. Pospíšil, "Ultra-weak photon emission from biological samples: definition, mechanisms, properties, detection and applications", J. Photochem. Photobiol. B 139 (2014) [memory, medium for the exact citation; cited by the physicist via a search summary]. The modern account of biological light. - Irving Langmuir, "Pathological Science" (1953 lecture, published in Physics Today, 1989) [memory, high], for how small, detector-limited effects can persist; read alongside Volodyaev and Beloussov for the opposing view.

---

Answers

1. E = 12 398 / 2500 = 4.96 eV. One neutralisation event releases 55 800 / 6.022 × 10²³ J = 9.27 × 10⁻²⁰ J = 0.578 eV. So 4.96 / 0.578 ≈ 8.6, i.e. about 9 events would have to deliver all their energy to one molecule at once. No known process pools that many.

2. (a) 3000 rpm = 50 revolutions per second, so one degree takes 1/(50 × 360) s = 55.6 μs, and 15 degrees take 0.833 ms. Speed = 0.025 m / 0.000833 s = 30 m/s. (With windows 25-30 degrees wide, reading the shift at the lower edge, middle or upper edge gives 30, 26 or 22.5 m/s, so "about 30 m/s" is good to roughly ±30 % [calc by arbiter].) (b) Light crosses 2.5 cm in 0.025 / 3 × 10⁸ = 8.3 × 10⁻¹¹ s, ten million times faster. Whatever travelled at 30 m/s was not light flying freely; it was a relay of absorption, delay and re-emission, or a chemical or electrical wave. The measurement cannot tell "light carries the signal" from "the signal gives off light" (claim 4).

3. Combined probable error = √(0.15² + 0.08²) = 0.170. Difference = 1.92 − 1.01 = 0.91, so 0.91 / 0.170 = 5.35 probable errors, matching the book's 5.3. In standard deviations: 5.35 × 0.6745 = 3.6 σ [calc]. Before believing it you would want to know: were exposed and control halves assigned and counted blind; were all runs reported, including "failed" days; and what other differences existed between quartz and glass halves (heat, gas exchange). The book does not state blinding (arbiter).

4. The number of chance positives is binomial with n = 60, p = 0.05. P(X ≥ 5) = 1 − Σ_{k=0}^{4} C(60,k)(0.05)^k(0.95)^(60−k) = 0.18 [calc]. At 3 % per strip it is 0.03; at 10 % it is 0.73. So "5 positive strips of 60" means something only if the false-positive rate per strip is small and known, and if the same five strips reappear in independent runs. The book shows neither (arbiter, claim 2).

5. Fluorescence emits at longer wavelength than absorbed (Stokes shift), because energy is lost as heat between absorption and emission. Here the output is shorter. Energy in: 12 398 / 3230 = 3.84 eV; energy out: 12 398 / 2475 = 5.01 eV. The solution must supply about 1.17 eV per photon from its own chemistry, plus the whole energy of any extra photons if the output is "stronger" than the input. Rahn proposes a chain reaction but gives no energy budget [read: p.57]; and Rahn himself warns that the "stronger" output was judged only by how fast yeast responded, not measured physically [read: p.59]. Note also that the experiment used a glass trough, which would absorb 2450-2500 Å light; the geometry is not explained in the text.