# Fact-check and revision record

Changes made before publication

**Verdict: the draft’s central conclusion is defensible, but it required substantial qualification and several factual or numerical corrections.** The published article preserves the argument that cyanide-residue differences do not establish the impossibility of homicidal gassing. It removes claims that chemistry alone proves purpose, that proposed mechanisms have been fully validated, or that every objection has been conclusively excluded.

*Source reviewed: Cyanide_Residue_Argument_Rebuttal_2.pdf, 11 pages. Review completed 1 October 2026. [Read the revised article](cyanide-residue) · [Additional arguments](additional-arguments).*

## Corrections to facts and arithmetic

| Draft location and claim | Finding and published correction |
| --- | --- |
| pp. 1–2, 7: cyanide in “every homicidal chamber” | **Overbroad.** Positives occur among samples from each of the five crematorium buildings and Block 11. Some samples are zero; a building-level result is not a result for every individual room. |
| pp. 1–2: “unexposed” controls | **Incorrect description.** Table I describes dwelling rooms probably fumigated once. The published article retains that uncertainty. |
| p. 7: table of principal positives | **Accurate selected figures, incomplete coverage.** Crematorium III was omitted. The revised table includes it and shows all sample counts and positive-reading ranges, including zero and trace limitations. |
| p. 7: positives “60 to 200 times” the limit | **True only of selected higher results, with rounding.** It does not describe all positives. The minimum numeric positive, 8 µg/kg, is only about 2–3 times the stated limit. |
| p. 6: about 1% CN⁻ at pH 6–7 | **Numerically inaccurate.** For the given pKₐ 9.31, the ideal fractions are approximately 0.049% at pH 6 and 0.487% at pH 7; at pH 10 they are 83.0%. |
| p. 5: 950 litres of CO₂ per minute from 1,000 people | **Repeats an inconsistency in the study.** Its stated 3.5% and 15–20 litres/minute imply 525–700 litres/minute for 1,000 people. The revised article shows the arithmetic and does not treat it as a physiological reconstruction. |
| p. 8: newest thread found was December 2024 | **Outdated.** An accessible thread beginning 30 July 2025 contains relevant objections and probabilistic formulations. |
| p. 9: controls “number ten samples” | **Needs separation by investigation.** Eight are in Table I of the full study; two additional controls are reported from the earlier screening. |

The residue figures were checked against the [scanned 1994 paper, Tables I–III](https://codoh.com/media/files/downloads/MarkiewiczEtAl1994-ocr.pdf), not just a search extract. Its [HTML transcription](https://phdn.org/archives/holocaust-history.org/auschwitz/chemistry/iffr/report.shtml) duplicates sample number 17 where the scan lists 16 and 17. The pH values were recalculated from the equation; the source [Green report, pp. 49–50](https://phdn.org/archives/holocaust-history.org/irving-david/rudolf/affweb.pdf) itself prints an incorrect rearranged denominator. The later discussion is available in the [2025 forum thread](https://www.codohforum.com/viewtopic.php?t=468).

## Corrections to scientific interpretation

| Draft claim | Assessment and revision |
| --- | --- |
| “Measuring non-iron cyanide isolates exposure” | **Overstated.** An assay measures what its conditions release and detect. Formation, survival, extraction, contamination, and matrix effects remain relevant. A residue is not a direct measurement of historical exposure. |
| Excluding Prussian blue is simply “the right test” | **Too categorical.** The method is selective and potentially useful; total cyanide is also relevant. Complementary measurements on matched samples would be stronger. |
| Prussian blue is stable, therefore weathering “cannot” explain its absence | **Too absolute.** Persistence limits a simple rain-washout explanation. It does not exclude physical erosion or removal of the material containing the pigment. |
| Rudolf’s own mechanism “predicts the opposite” | **Not established as stated.** His later text includes delayed formation and persistent alkalinity in cement-rich materials. The correction focuses on the lack of validated quantitative predictions. |
| All conversion factors favour delousing | **Unsupported generalisation.** Interacting conditions cannot be scored conclusively without measured inputs and a validated model. |
| Rare reports of blue staining establish a base rate | **Not a statistical demonstration.** The reviewed reports do not supply a representative denominator or systematic observations of all fumigations. |
| 117 versus 7,200 exposure hours | **Attributed estimates, not measured totals.** Retained as Green’s illustration with its dependence on assumptions. |
| Several witnesses establish wall hosing after every use | **Overgeneralises the citations.** Green’s note includes floor washing, body washing, damp floors, and a separate hosing/whitewashing account. These do not establish one uniform wall-treatment procedure. |
| Washing reduced cyanide by a specified number of orders of magnitude | **Model assumption.** The amount and depth of any actual dilution are not established by the cited testimony. |
| CO₂ co-exposure necessarily suppressed pigment formation | **Not demonstrated by the pilot tests.** The study measured a selective cyanide fraction, with material-dependent results, not a complete reconstruction of pigment formation in the historic walls. |
| Only the carbonating surface could be reactive | **Too restrictive.** Depth profiles, later carbonation, transport, and delayed reactions need consideration. |
| 3.3 × 10⁻⁴ mol/L is a universal masonry threshold | **Unverified extrapolation.** Presented only as part of Green’s discussion of a particular chemical route. The Alich experimental full text was not independently checked. |
| Different studies give exact bound/unbound fractions | **Not established for matched specimens.** Ratios across laboratories and different samples are not measured conversion fractions of the same material. |
| A positive Zyklon carrier sample confirms recovery | **Limited control.** It demonstrates a positive result on a cyanide-containing material, not a quantified recovery percentage from every masonry matrix. |
| All small total-cyanide positives reflect carbonate noise | **Too strong.** Possible assay interference does not prove every trace result false. The discussion now distinguishes interference, detection, and historical interpretation. |

The main checks used [Green’s 2001 report](https://phdn.org/archives/holocaust-history.org/irving-david/rudolf/affweb.pdf), [Rudolf’s 2020 script](https://holocausthandbooks.com/wp-content/uploads/ChemistryAuschwitz-5th-Script-2020.10.20.pdf), the [1994 experimental paper](https://phdn.org/archives/holocaust-history.org/auschwitz/chemistry/iffr/report.shtml), and the [2013 Rudolf–Kollerstrom article](https://codoh.com/library/document/differential-exposure-of-brickwork-to-hydrogen/). Checking a claim against an advocate’s own publication verifies attribution, not independent experimental validity.

## Corrections to the argument’s logic and historical scope

- **A concession of uncertainty is not a concession that an argument has no evidential weight.** The revised version addresses both absolute and probabilistic forms.
- **The 1942 chronology rules out a particular explanation, not all possible later disinfestation.** Residue alone cannot identify purpose.
- **Negative controls improve the case; they do not exclude every possible interference.** Different matrices and preservation histories matter.
- **Roth’s ten-micrometre claim cannot be used as a universal rule.** The draft already recognised this; the revision retains the concession consistently.
- **Minimum lethal concentrations do not establish actual historical concentrations.** The draft recognised this, but sometimes then treated assumed concentrations as known. The revision distinguishes nominal loading and concentration through time.
- **A simplified ventilation result is conditional.** It cannot establish an unconditional entry time or neglect continued emission and imperfect mixing.
- **The 2004 roof investigation reported three identifiable openings in Crematorium II, with a proposed location of the fourth.** The companion article states this precisely and acknowledges that individual architectural objections require their own audit.
- **No universal claim is made that “no experiment or model has been produced.”** The narrower conclusion concerns the absence of a validated site-specific prediction in the sources actually reviewed.

The roof findings were checked in [Keren, McCarthy, and Mazal, 2004, *Holocaust and Genocide Studies* 18(1), 68–103](https://phdn.org/negation/HGS-2004spr-v18n1-Auschwitz_Gas_Chambers_Study.pdf). Some CODOH catalog entries incorrectly give volume 9; the publication itself gives volume 18.

## What remains unresolved

This review did not conduct new chemical analysis, independently validate the 1994 assay, obtain the full Alich experimental text, or re-investigate the ruins. It did not establish precise wartime wall pH, concentrations, cleaning regimes, or total exposure histories. Those gaps are explicitly reflected in the conclusions.

CODOH library articles and accessible forum pages provided substantially better coverage than the original draft’s search snippets. Some later forum pages were blocked, so this is a documented, focused survey, not an assertion that every post or every later publication has been checked.