Rahn-Otto: Quality engineering (Step 4b)

Project: Otto Rahn, Invisible Radiations of Organisms (Protoplasma-Monographien IX, Borntraeger, Berlin, 1936; physics chapter by Sidney W. Barnes). QE date: 2026-10-06. Inputs checked: _text/manifest.json, all 9 digests, deep/ (1 file), facts.json, claims.md, sites.md, research/ (12 notes; research/files/ is empty), physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py. Tags: [read: p.N] = extracted text, PDF page N (book page = N - 14); [read: page image p.N] = page I rendered myself (scratch /tmp/rahn-qe); [calc] = my arithmetic.

Verdict: MINOR

The central verdicts survive. The energy argument (claim 3) and the per-event arguments (claim 2) were checked in a deep read, by three reviewers, by the arbiter and by example.py, and the numbers hold. But I found one factual error that runs through four files (the fingertip-through-quartz time; see Accuracy, check 5). I also found a page-tag convention mixed up in the contrarian review and copied into the arbiter and primer, and a coverage shortfall: claims 4-7 have no deep read. Several pertinent experiments in the book never reached claims.md, the reviews, the arbiter or the primer. None of this is likely to move a headline verdict. The fixes are to update facts, the arbiter and the primer, after the deep reads listed at the end.

---

1. Coverage

Coverage verdict: acceptable. No skipped text.

2. Shortcuts

• One deep read only (PDF pp.44-67). claims.md says plainly that "claims 4-7 rest on the digests alone and need deep reads in later steps". Those deep reads were never run (deep/ has one file). By the STEP 4b rule this is a defect: claim 4 (secondary radiation and nerve speed), claim 5 (blood and cancer), claim 6 (development and harmful radiation) and claim 7 (failures, physical counters) are central claims. Each one rests on experimental results that no deep read has covered.
• Mitigation, which is real but partial. The reviewers read primary passages themselves and rendered page images. The contrarian lists PDF pp.72-73, 104-111, 126-127, 137-138, 173-174, 181-182, 192-200 and 202-205. The mathematician read page images pp.97, 126, 143/157 and 186. The arbiter read pp.42, 48, 97 and 157. That is why the 30 +/- 3 m/s figure, the Julius error and the Table 45 problems were caught. But these reads were targeted spot checks, not section reads. Chapters V and VI (pp.117-146) and most of Chapter VII (pp.147-191) were never read beyond the digest. The one error the digest introduced that nobody caught (check 5) sits in a passage the reviewers cite as read.
• Sampling. No key document was sampled in part beyond the above. The book is the only document.

3. Accuracy spot checks (against the extracted text and page images)

4. Missed material (from the digests and text; absent from claims.md, the arbiter and the primer)

Material in the book that bears on the claims but is absent from claims.md, the arbiter and the primer (and mostly from the reviews):

1. Reiter and Gabor's conflicting wavelength [read: digest p.72-74; text p.75]. Reiter and Gabor (1928, Siemens) found the effective rays at 3200-3500 A, with a maximum at 3400 A, a second maximum at 2800 A, and "antagonistic" rays at 2900-3200 A. That contradicts the Gurwitsch band (1900-2500 A). Rahn says their results "have never been satisfactorily explained" and attributes them to an apparatus error. Gabor's parthenogenesis results at 3340 A (3 of a few hundred frog eggs) belong here too [read: digest p.184]. This is a dispute inside the field over the defining property of claim 1, and it involves Dennis Gabor (later a Nobel physicist). The arbiter and primer mention neither the dispute nor Gabor, though the Investigator listed Reiter and Gabor as related work.
2. Intermittent irradiation [read: digest p.117-121]. Pulsed light at 100-800 interruptions per second gave the effect in 12.5-13 s against 6-8 min continuous, about 1/30 of the time. The effect then carried over 15 cm against 3-4 cm. Only a regularly spaced disc worked, not an irregular one. Warburg's photosynthesis analogy did not fit. facts.json lists it as a key claim. No review, the arbiter or the primer weighs it. It is an anomalous, testable claim of exactly the kind the contrarian role exists to examine.
3. Daylight dependence of senders (Potozky 1930) [read: digest p.122-123]. Yeast grown in the dark did not radiate; it regained the power after 2 hours in daylight. This is the book's own evidence for the proponents' objection to Quickenden 1989 ("daylight as a necessary condition"). The arbiter quotes that objection from Volodyaev and Beloussov but does not connect it to the book.
4. Polarisation (Wolff and Ras 1934a) [read: deep p.59; digest p.127]. Secondary radiation was said to be polarised, and polarised rays far more effective. The deep read records this. No review weighs it as a claim.
5. Gradual approach prevents induction [read: digest p.131-132]. Blocks moved together over 5-8 minutes gave no effect, while blocks set at the final position gave +40-50%. An accommodation effect, never discussed.
6. Non-biological detectors [read: digest p.102-104, 113-115]. These are Stempell's Liesegang-ring and H2O2-decomposition detectors, Heinemann's gold-sol flocculation measured with a photoelectric nephelometer (blood 3-26 units against control 0-2.5), Lepeschkin's "necrobiotic rays" blackening AgBr and photographic plates (dying yeast only), and Suchow and Suchowa's AgBr test (25 experiments, quartz vessel darkened sooner than glass). The primer's lesson "a biological detector cannot calibrate itself" and its table row "mostly a biological detector" are fair. But these chemical and photographic detectors are pertinent evidence on claims 1-2 and appear nowhere in the reviews.
7. Anna Gurwitsch's optic-nerve and brain experiments [read: digest p.171-173]. Light on the eye made the optic lobes radiate. Spectra changed with electrical irritation of the sciatic nerve and with the colour of the light (green, red, blue). Proposed for brain-function localisation. These bear on claim 4 and are absent from every review (the grep hits for "optic" are "optics").
8. Protti's "cytophotolysis" and the cytagenin and young-blood "rejuvenation" claims [read: digest p.165-167, 195-197]. Yeast destroyed tumour cells in vitro, also through quartz. Yeast injections liquefied rat sarcoma, and heated yeast had no effect. Seyderhelm's cytagenin raised blood radiation. Protti claimed clinical rejuvenation after injecting young blood. Claim 5 names Protti, but the primer never mentions him and the arbiter gives him one clause.
9. Acs (1933) antagonistic radiation [read: digest p.186]. Ps. pyocyanea retarded B. anthracis by 22-136% in 14 experiments, and yeast retarded staphylococci and streptococci. Kostoff and Kendall (1932) reported one tetraploid shoot from tumour tissue [read: digest p.183]. Nelson and Brooks (1933) found infrared reduced sea-urchin fertilisation in 9 of 9 experiments [read: digest p.115-116]. Zwaardemaker restarted frog hearts with radioactive substitutes for potassium (34 experiments) [read: digest p.116-117]. These are peripheral, but pertinent, and listed in facts.json key claims. The reviews never mention them.
10. Siebert and Seffert (1934) [read: digest p.106-107]. Several hundred normal blood samples raised counter rates; carcinoma blood and KCN blood did not. Rahn calls this the most convincing physical evidence. The arbiter keeps it; the primer evidence table (rows 2b and 5a) drops it, so the strongest physical item for claim 5 is missing from the document the reader studies.
11. Rahn's own caveat on Paul (1933): "The number of examples given is not large enough to draw many more conclusions" [read: p.77]. The primer and arbiter cite Paul as positive evidence without this caveat. (SOURCE HONESTY cuts both ways.)

5. Retention check

Question: was pertinent material from the sources or from the Investigator left out of the arbiter or primer because a reviewer judged it negatively?

• No case where a reviewer's negative verdict caused the omission. The arbiter keeps the positive evidence explicitly ("Positive evidence stays in the record": Gurwitsch school, Paul, Julius, Schwemmle, Rahn's group, Wainwright 1997). It reports the proponents' protocol objections to Hollaender and Claus as a party's characterisation, analyses the post-1937 silence with six competing explanations, and keeps the Soviet "cancer quencher" line. The primer carries all of this.
• Omissions that are not verdict-driven but should still be fixed (Keep pertinent information):
• Investigator leads dropped from the arbiter and primer: The Secret Life of Plants (Tompkins and Bird 1973) linking Rahn to T. G. Hieronymus's "eloptic" devices; Stempell's 1932 book of nearly the same title (named in the arbiter, missing from the primer's further reading); Reiter and Gabor 1928 and Dennis Gabor; Koyenuma 1947 (arbiter only); Rahn's Mathematics in Bacteriology (1939) and Microbes of Merit (1945) as records of his later view (arbiter mentions only Microbes of Merit); the Barnes identity (M. N. Barnes as the probable oxy-cholesterol co-worker; arbiter only, not in the primer).
• Pertinent negative evidence dropped: Motshejewa's critique of the onion-root method (pressure and friction raise mitoses; selection of roots; omitted sections) [read: p.76]. The physicist keeps it; the arbiter and primer do not, though the primer gives Paul's rebuttal.
• Pertinent positive evidence dropped from the primer: Siebert and Seffert (item 10 above); Anna Gurwitsch's brain experiments (item 7); Protti (item 8).
• Conclusion: no Retention defect of the verdict-driven kind. There are seven omissions of pertinent material (listed above) to restore in the arbiter "QE revision" section and the primer.

6. Process

• Source tags present and mostly honest. [web] items give URLs. Investigator items are marked "via Investigator" and "not read". Every reviewer says the outside documents were not read. [memory] carries confidence levels. Good.
• Page-tag convention mixed (defect). Every file declares "[read: p.N] = PDF page". The contrarian review cites several passages by book page while using that label: "p.188" (Outlook, actually PDF p.202), "p.189" (controls "not perfect", PDF p.203), "p.183" and "p.183-184" (Roffo, PDF p.197), "p.184-185" ("cannot be considered proved", PDF p.198), "p.185-186" (oxy-cholesterol and Table 53a, PDF pp.199-200). The arbiter and primer copy these tags unchanged (arbiter claim 1, claim 5, claim 6; primer table rows 1c, 5c, 6c and section 7). The mathematician also uses book pages in places ("p.112" for the 0.0014 s, "p.~92" for Gollshewa, "p.~180" for Gesenius). Other tags are PDF pages and correct (p.107, p.110-111, p.126, p.137, p.175, p.195). A reader opening the PDF at "p.183" finds the wrong page. Fix: convert every book-page tag to PDF page (add 14) in the arbiter and primer.
• Web checks. The physicist, mathematician and arbiter logged their web checks (the arbiter used 8, the mathematician 6). One [memory, medium] fact that a verdict depends on was never web-checked: "modern UPE work does not report 190-260 nm emission at mitogenetic intensities". It underlies the arbiter's verdict 1b (Never properly tested) and the primer's textbook table. It should be checked (a Cifra and Pospisil 2014 or similar review), or the verdict should say it rests on [memory]. The Quickenden 1989 result remains unconfirmed (the arbiter says so honestly).
• Verdict scale. Applied consistently, with split verdicts stated. Two small irregularities. The physicist and arbiter call Lorenz's objection "Established", which applies the scale to a counter-argument rather than a claim; it is harmless but should be worded as "the calculation is correct". The arbiter gives claim 3b two verdicts at once ("Refuted by calculation (and Internally inconsistent)"); choose one, or say which part each covers.
• Primer against STEP 4 and the reader profile. All nine sections are present in order, plus answers. Symbols and acronyms are defined (UPE, NRC, OCR, PE, eV, E_a, Φ). There is a numerical worked example and example.py. The textbooks have chapter numbers only where safe (Lakowicz ch.1). No new verdicts. The reader-level physics is well judged. Defects: the fingertip row (6c) carries the 1.5-minute error, some page tags are wrong, and items 7, 8 and 10 of Missed material are absent.
• example.py. Runs cleanly with the project venv (exit 0). It reproduces every number in primer section 5: 414 cal/g, ratio 42.6, 43 and 11 events per photon, 2767 A, 138.2 kcal/mol, the Fricke checks, yield 2.6e-14, 40% and 64% hit fractions, and doses of 8.9e-10 to 7.8e-9 J/m2.
• Deferred Step 1 corrections not yet applied (expected at remediation): facts.json "30.3 m/s"; facts.json and claims.md "Barnes and Rahn 1933" to "M. N. Barnes (probable)"; facts.json "about 500 quanta/cm2/s" counter limit (the printed figures do not convert; about 100 quanta/cm2/s at 2000 A per 1e-9 erg); claims.md claim 7 wording ("physical evidence from photoelectron counters"); claim 2 qualifier ("as shown by a biological detector").

7. Remediation actions

Deep reads (at most 8, most important first; slug Otto-Rahn-Invisible-Radiations-97d8ea)

1. PDF pp.104-116, Opus. Physical detection, unaccounted failures, injurious human radiation, necrobiotic rays, infrared and beta radiation. Questions: What exactly did Rajewsky, Frank and Rodionow, Barth, and Siebert and Seffert measure (counts, controls, KCN substitution, number of bloods)? How does Rahn present Lorenz's three error sources and Kreuchen and Bateman's yields? For each failure report (Gollshewa, Wolff and Ras, Heinemann, Siebert, Rahn), what was recorded? For the fingertip experiments, give the Fig. 38 geometry (is there glass between finger and yeast in the "glass ring" set-up?), the kill times (5 min and 15 min, confirm on page images), the number of people tested, and the controls. What did Lepeschkin and Suchow and Suchowa's AgBr and photographic results show (counts, controls)? Render page images for tables and figures.
2. PDF pp.117-132, Opus. Chapter V. Questions: The intermittent-irradiation data (Table 31, Zoglina, regular against irregular disc, 15 cm against 3-4 cm). The daylight dependence of senders (Potozky). The secondary-radiation spreading data (10 mm, 25-87%) and exhaustion. Polarisation (Wolff and Ras 1934a). Overdose depression and Salkind's periodicity. The gradual-approach experiment. For each: numbers, controls, replicates, and what Rahn concedes.
3. PDF pp.160-174, Sonnet. Tissues, blood, nerve and brain. Questions: blood-radiation methods and numbers (Potozky and Zoglina starved rats, Wassiljeff, Golischewa's rabbit-blood spectrum, NaF and KCN tests); Protti, cytagenin and Seyderhelm; Kalendaroff's nerve spectra; Anna Gurwitsch's optic-nerve, optic-lobe and colour experiments (Table 49); Latmanisowa's parabiosis (more than 40 repeats, controls) and Lapitzki.
4. PDF pp.191-200, Sonnet. Cancer and oxy-cholesterol. Questions: every clinical number Gesenius, Heinemann, Braunstein and Heyfetz, and Klenitzky give (cases, exceptions, follow-up); Protti's cytophotolysis and yeast-injection experiments (group sizes, heated-yeast control); Roffo; Table 53a (oxy-cholesterol batches, any glass control). Render page images for the tables.
5. PDF pp.68-91, Sonnet. Detector methods. Questions: the Reiter-Gabor 3200-3500 A results and Rahn's explanation (Table 18, Fig. 31); Schwemmle's analysis (Fig. 30); Motshejewa against Paul, including Rahn's caveat on Paul; the Tuthill-Rahn yeast-bud protocol and Table 21; Heinemann's blood protocol; Gesenius' fermentation counts (102 experiments; the printed numbers do not sum).
6. PDF pp.133-146, Sonnet. Chapter VI. Questions: Rahn's single-quantum and lag-phase arguments with their numbers (Tuthill and Rahn, Ferguson and Rahn Table 36, Rahn 1928 and Rahn and Barnes 1932 fermentation); the mitogenetic-field and interference argument and Miss Ferguson's two-sided irradiation; the Reiter-Gabor cytology of the onion root (Table 37).
7. PDF pp.175-191, Sonnet. Morphology, metamorphosis, parthenogenesis, symbiosis and antagonism, wound healing. Questions: Blacher's tadpole limb and wound series (quartz against glass figures, group sizes, the 6.5% +/- 0.8 operation control); Magrou's mesenchyme histology; the Reiter-Gabor 3340 A parthenogenesis counts; Acs's antagonism data; Kostoff and Kendall.
8. PDF pp.92-103, Sonnet. Bacterial, egg, tissue-culture, cornea and physico-chemical detectors. Questions: Ferguson and Rahn's generation-time method (Table 26); Julius (confirm 1.01 +/- 0.08); Gesenius' respiration and fermentation; Magrou and Reiss's electrical alternative (detail); Liesegang rings; H2O2; Heinemann's gold-sol flocculation numbers.

What each later file must change

• facts.json: nerve speed 30 +/- 3 m/s (not 30.3); M. N. Barnes as the probable oxy-cholesterol co-worker; correct the counter-limit conversion (printed 1e-9 erg/cm2/s and 500 quanta/cm2/s are inconsistent for UV); "fingertip killed yeast in 5 min held over a glass ring, 15 min through 2 mm quartz" [read: page image p.111]; update how_read after the deep reads; add Reiter-Gabor, intermittent irradiation, daylight dependence, polarisation, the non-biological detectors and the Anna Gurwitsch nerve and brain work to summary or key_claims, with page tags.
• claims.md: apply the arbiter's corrected wording. Add to claim 1 the internal wavelength dispute (Reiter and Gabor 3200-3500 A). Consider whether intermittent-irradiation potency (about 30-fold) belongs in claim 1 or 4. Replace "rest on the digests alone" once the deep reads exist.
• arbiter.md, "QE revision" section: (a) withdraw the "faster kill through quartz" point in claim 6 and open question 6; restate the times as 5 min (finger over a glass ring) and 15 min (through 2 mm quartz); no verdict change expected. (b) Convert the book-page tags to PDF pages (add 14): p.183 to 197, p.183-184 to 197-198, p.184-185 to 198, p.185-186 to 199-200, p.189 to 203. (c) Weigh the missed items (Reiter-Gabor wavelength, intermittent irradiation, daylight dependence against the Quickenden objection, polarisation, non-biological detectors, Anna Gurwitsch, Protti and cytagenin, Acs) and state whether any verdict changes. The expectation is no change in headline verdicts, but claim 1's band definition and claim 4's evidence base should mention them. (d) Restore Motshejewa (negative) and Rahn's caveat on Paul beside Paul. (e) Web-check, or retag as [memory], the "no 190-260 nm UPE" statement behind verdict 1b. (f) Make the claim 3b verdict single, or say which part each verdict covers.
• primer.md: fix evidence-table row 6c (5 min and 15 min; drop the implied anomaly); fix the page tags as above; add Siebert and Seffert to rows 2b and 5a, Protti to 5b, Reiter-Gabor's conflicting band to section 4 or row 1b, Anna Gurwitsch to row 4a, and the non-biological detectors to the "How the light was measured" row; add Rahn's caveat on Paul and Motshejewa's critique to row 1c; add to further reading Stempell 1932, Reiter and Gabor 1928, the Secret Life of Plants and Hieronymus lead (marked as a later popular account), and Rahn's Mathematics in Bacteriology and Microbes of Merit; mention the M. N. Barnes attribution in row 6c. Re-run example.py if any number changes (none expected).
• QE recheck (Sonnet): confirm items in sections 3 (check 5), 4, 5 and 6 are fixed and append a "Recheck" section here.

Recheck (Step 4b, 2026-10-06)

Method: grep and read of the current arbiter.md, primer.md, facts.json, claims.md and deep/ against each defect above. [calc]/[read] tags as before. I did not re-render page images; the 15-minute figure is accepted from my own earlier image read and from the deep reader's confirmation recorded in arbiter.md.

Remaining issues: (1) arbiter "Established" applied to Lorenz's objection; (2) facts.json summary still says "Barnes and Rahn 1933". Both are wording-level and neither moves a verdict.

Final: OPEN (2 issues)