Why the cyanide residue argument fails

The central error in the cyanide residue argument is the assumption that a particular history of gassing must have left a particular quantity of chemical residue in the surviving walls. That relationship has to be demonstrated. It cannot be inferred simply from the fact that some delousing chambers have conspicuous blue staining and the sampled crematoria have much less cyanide.

The difference is real and deserves explanation. But a wall is not a gas meter that records every exposure in permanent, directly proportional form. Gas must enter the material, participate in reactions, remain there through subsequent use and destruction, survive decades of environmental change, and finally be recovered by the particular analytical method. A comparison that skips those steps mistakes an observation about present materials for a direct measurement of past events.

Nor does the evidence consist of an absence of cyanide. The 1994 investigation by the Institute of Forensic Research in Kraków reported positive results among samples from each of the five crematorium buildings tested, as well as from the Block 11 cellars. Its results require careful interpretation: they support the presence of cyanide compounds, but do not identify the purpose of an exposure or count its victims. The historical identification of homicidal gassing depends on the combined documentary, testimonial, architectural, and material record.

This distinction supplies the basis of the rebuttal. Low residue concentrations do not establish that killing by gas was impossible; positive concentrations do not establish homicidal purpose by themselves. The question is whether the chemical argument supplies a reliable reason to reject the historical evidence. Its strongest formulations still fail to establish the quantitative prediction on which that rejection depends.

1. The argument that actually needs answering

From Leuchter to Rudolf

Fred Leuchter’s 1988 investigation made the residue comparison widely known. The contrast was striking: low readings in material taken from structures identified as homicidal gas chambers, and a much higher reading from a delousing building. But even Leuchter’s results were not uniformly negative. The Kraków authors report that 14 of his 30 Auschwitz gas-chamber samples contained cyanide, with positive values of 1.1–7.9 mg/kg. His delousing comparison exceeded 1,000 mg/kg. The argument was therefore about the meaning of a difference, not the complete absence of the substance. Markiewicz, Gubała, and Łabędź, 1994, introduction.

Germar Rudolf developed a more elaborate case, discussing the chemistry of iron-cyanide compounds, the properties of masonry, and conditions in the crematorium basements. His later presentation recognises that the absence of reliable residues is not conclusive in isolation. He nevertheless argues that the circumstances should have favoured substantial and persistent deposits. The 2013 paper by Rudolf and Nicholas Kollerstrom also addresses cyanide that need not be visibly blue. A serious answer must therefore engage the chemical comparison, rather than treating the whole argument as a naive inspection of wall colour. Rudolf, 2020, pp. 10–18; Rudolf and Kollerstrom, 2013.

The crucial premise

The strongest version can be expressed in four steps:

1. Known HCN delousing facilities contain substantial persistent cyanide deposits.
2. The alleged homicidal operations would have exposed the relevant walls under conditions producing and preserving comparable deposits.
3. Adequate sampling and analysis would reveal those deposits today.
4. The deposits found are much smaller; therefore the historical account is false or substantially less probable.

The second and third steps do the decisive work. It is not enough to establish that the same chemical was used, that both settings included masonry, or that some conditions in the basements favoured absorption. The argument needs a justified expectation for the amount and distribution of residue that would form, survive, and be measured in the actual specimens.

Each step also has a different subject. An exposure history concerns gas in a room. Formation concerns chemical reactions in material. Preservation concerns what happened to that material afterwards. Measurement concerns what a laboratory can extract and detect. Moving between them requires evidence; calling the result “forensic” does not provide the missing relationships.

A probabilistic argument still needs a sound comparison

Some CODOH participants expressly describe their case as probabilistic. That is a legitimate form for a historical argument. A rebuttal does not succeed merely by showing that Rudolf once acknowledged uncertainty, or by demanding mathematical certainty before any evidence can count. CODOH discussion, July 2025.

The relevant question is comparative: how likely are these results if the historically documented operations occurred, and how likely are they under a specified alternative? A low reading becomes strong negative evidence only if a substantially higher reading was strongly expected. Without that expectation, “there is less than in another building” does not tell us the strength or even the direction of the historical inference.

This is not a rule that every historical argument requires a complete numerical model. Qualitative comparisons can be persuasive when the relevant conditions are sufficiently similar and the causal relationship sufficiently well established. Here, however, differences in duration, materials, preservation, and analytical recovery are precisely what is disputed. Assigning an informal plus or minus to each factor does not establish their combined effect.

2. What a cyanide residue is—and what it is not

Exposure, pigment, and analytical recovery

Hydrogen cyanide, HCN, and the cyanide ion, CN⁻, are chemically related but not interchangeable descriptions of everything a wall may contain. Cyanide can also be incorporated into metal complexes. Prussian blue is an iron-cyanide pigment; it is a product of chemical formation under suitable conditions, not simply HCN gas that has turned blue on contact with a wall.

Several quantities therefore need to be kept separate: the amount of Zyklon B brought into a building, the HCN released into the air, the concentration at the wall over time, the amount taken up by the material, the compounds subsequently formed, and the portion a laboratory method recovers decades later. A change in one does not establish a proportional change in all the others. Green’s discussion and Rudolf’s own reaction scheme both recognise that material conditions affect the chemistry. Green, 2001, pp. 41–50; Rudolf, 2020, pp. 10–12.

Visible colour is another distinct measurement. A specimen can contain cyanide without conspicuous blue staining, and a photograph cannot replace a chemical assay. Equally, identifying blue material as a cyanide-containing pigment would not by itself establish when the exposure occurred, how often it happened, or whether the room was used for killing or pest control.

Why “total cyanide” does not mean total historical exposure

A total-cyanide analysis is relevant because it aims to recover a broader range of cyanide compounds than a deliberately selective method. It is reasonable to ask for it. But “total” describes the analytical target and procedure, not a complete historical record. Material that was removed, compounds that escaped, and surfaces that did not survive are not restored by a more inclusive extraction.

Even an accurate measurement of all cyanide remaining in one specimen would be a measurement of that specimen now. To infer the original dose, one would need to know the fraction retained and how representative the specimen was. To infer purpose, one would need historical evidence. These are limitations of the inference, not reasons to dismiss analytical chemistry.

Why the known delousing rooms are useful but insufficient controls

A positive comparison material from a delousing facility can show that cyanide-containing deposits exist and that a method detects them. It can also help investigate different chemical fractions. It is not automatically a matched control for a demolished crematorium wall.

A negative control serves another purpose: it tests whether comparable material without the exposure of interest gives a signal. Matrix controls address interference from the substance being analysed—for example, plaster rather than a clean laboratory solution. Neither purpose is fulfilled merely by placing one large blue-stained result beside several small results from different structures.

The comparison would be much stronger if the specimens had documented original surfaces, comparable material composition, known exposure and maintenance histories, and matched sampling depths. The absence of those conditions does not make the numbers imaginary. It limits what the difference can establish.

3. Duration, concentration, and the misleading appeal to total consumption

Time at the wall matters

The relevant exposure is a concentration history at a surface, not just the number of operations or the mass supplied to the camp. Long exposure at a sustained concentration and short exposure followed by ventilation can result in different uptake, even when the same nominal amount of material is introduced. They can also permit different amounts of subsequent reaction before conditions change.

Green illustrated the possible scale of the duration difference with approximately 117 cumulative hours for one crematorium chamber and at least 7,200 hours for the BW5a delousing chamber. The first estimate used assumptions about the number and duration of operations; the second used an assumed minimum number of long delousing cycles. These are reconstructions, not readings from surviving exposure instruments or complete operating logs. They should not be applied indiscriminately to every building. Green, 2001, pp. 43–44.

The arithmetic makes the assumptions visible. Green’s first illustration uses about 175 operations at 40 minutes each: 175 × 40 ÷ 60 is approximately 117 hours. His BW5a comparison uses at least 450 cycles of 16 hours: 450 × 16 is 7,200 hours. The number of homicidal operations was itself reconstructed from assumed victim totals and numbers per operation. Neither product is an independently measured exposure history, and the same number of hours at falling and sustained concentrations would not necessarily represent the same exposure.

The figures are useful for a limited reason: they show how a large duration difference can arise without assuming that the gas was harmless or that the instantaneous concentrations were radically different. They do not establish an exact sixtyfold chemical difference. Reaction rates need not be linear in elapsed exposure time, and the estimates themselves depend on uncertain historical inputs.

Minimum lethal concentration is not the concentration actually used

A common shorthand compares a concentration sufficient to kill a person with the higher or longer exposures associated with disinfestation. That comparison cannot establish the actual concentration during a particular homicidal operation. A minimum effect threshold, a recommended fumigation condition, a nominal loading, and a measured room concentration are different quantities.

The rebuttal therefore does not depend on asserting that homicidal use necessarily involved only a tiny fraction of the delousing concentration. The better question is how much HCN was actually released, how it was distributed, how long it remained, and what reached each material. The distinction remains necessary even if the two uses involved concentrations of the same general order.

Similarly, Green’s illustrative gas-loading assumptions should not be presented as measurements of the air in 1943. A nominal loading calculated from the amount introduced is not necessarily attained immediately or uniformly. Continuing release, absorption, ventilation, and local differences affect the time course. Green, 2001, discussion of release and concentration, pp. 3–9, 30–37, 45–50.

Repeated short exposures can accumulate—but not automatically without loss

Repeated exposure is a serious part of Rudolf’s argument. Short events cannot simply be dismissed as chemically negligible because each one was brief. Material retained after one event may remain available during the next, and reactions may continue after the gas is removed.

But accumulation depends on retention. As a conceptual bookkeeping model, suppose each event adds an amount a, while a fraction r of the relevant material survives until the next event. After n equal events, the accumulated amount would be a(1 + r + r² + … + rⁿ⁻¹). If r is near one, accumulation approaches the sum of the inputs. If it is much smaller, earlier contributions largely disappear. Real masonry involves several compounds and changing conditions, so this is an illustration, not a reconstruction of Auschwitz.

The example identifies the missing question. A count of events alone does not establish the residue. Nor does saying “short pulses” establish that no accumulation occurred. One needs evidence about the material retained between events and the proportion eventually converted into persistent compounds.

More Zyklon B in aggregate does not imply more residue per kilogram of wall

An aggregate consumption figure does not specify how the chemical was divided among buildings, purposes, surfaces, and periods. Even a reliable total for one room would not supply its concentration history without further information. A larger room could consume more while exposing a given unit of wall differently; longer retention could matter more than the amount introduced; some chemical could leave before reacting with the masonry.

The inference from total consumption to present residue therefore needs the very exposure and retention model that the comparison often omits. It is not enough to say that mass murder involved many people or that large quantities were purchased. Those facts do not provide a conversion factor between historical supply and modern micrograms per kilogram of sampled masonry.

4. The chemistry of uptake and persistent pigment

Moisture, iron, and alkalinity do not combine by simple vote

Rudolf emphasises conditions such as moisture, fresh plaster, available surface area, and alkalinity. These are relevant. Wet material can behave differently from dry material, and alkaline conditions can alter the balance between HCN and CN⁻. A defensible rebuttal should not deny those points merely because they appear in a denial argument.

The difficulty is the net prediction. A condition favourable to uptake need not have the same effect on every later reaction, and a wall is not chemically uniform. The availability of iron in a useful form, local pH, movement of water, gas transport, and subsequent ageing may affect different stages. Counting four favourable factors and two unfavourable ones does not determine the outcome if their magnitudes, interactions, and timing are unknown.

Rudolf’s later account also allows delayed pigment formation and retention in alkaline cement-rich material. It would be inaccurate to claim that his own mechanism straightforwardly predicts no persistent residues in the crematoria. The sound criticism is narrower and stronger: the mechanism does not, by itself, establish a calibrated quantitative expectation for these specific walls. Rudolf, 2020, pp. 10–18.

What the pH calculation actually says

Using the pKₐ value of 9.31 discussed by Green, the ideal acid–base relation gives the following fraction of the uncomplexed HCN/CN⁻ pair present as cyanide ion:

CN⁻ fraction = 1 / (1 + 10^(pKₐ − pH)).

The change is substantial. But the equation does not tell us the total cyanide concentration, the amount of iron available, the reaction rate, or the final pigment yield. It also does not describe every competing reaction in a real cementitious material. It is one part of the chemistry.

Nor can present-day pH be inserted as though it measured the wall during wartime operation. Carbonation and ageing change material conditions. The surface and interior may also differ. A calculation based on a pH of six or seven is conditional on that pH; it cannot prove what happened at an unknown pH decades earlier. Green’s printed rearrangement contains an algebraic error in its denominator, so the values above follow directly from the acid-dissociation relation rather than reproducing that expression. Green, 2001, pp. 49–50.

Carbon dioxide: a plausible influence, not a universal switch

Exhaled carbon dioxide can change the chemistry of moisture at a surface. The Kraków team explored exposure with HCN alone and with HCN plus CO₂. After a month of airing, its reported average losses of the measured cyanide were 56% and 73%, respectively. Individual materials differed; fresh plaster initially took up more in the combined-gas series. These results demonstrate sensitivity to conditions, not a single uniform effect for every historic wall. Markiewicz and colleagues, Tables V–VI and discussion.

The paper’s illustrative respiratory calculation also requires care. Its stated inputs—3.5% CO₂ in 15–20 litres of exhaled air per person per minute—give 0.525–0.700 litres of CO₂, not 0.950 litres. For 1,000 people over five minutes, that arithmetic yields 2.625–3.500 cubic metres, equivalent to about 0.53–0.70% of a 500-cubic-metre volume before considering other processes.

This checks the stated arithmetic; it is not a physiological or room-mixing reconstruction. It does not establish the actual breathing rate, spatial distribution, duration, or surface chemistry. The experimental result is sufficient to show that co-exposure can matter. It is insufficient to demonstrate a fixed reduction in Prussian-blue production during every homicidal operation.

Cleaning: distinguish floors, bodies, and walls

The testimony cited in Green’s footnote 55 does not describe one identical cleaning routine. It includes Tauber on floor washing, Müller on damp floors, Nyiszli on washing bodies, and Bennahmias on hosing and whitewashing. These observations may be relevant to room conditions, but they are not interchangeable evidence that every wall was thoroughly washed after every operation. Green, 2001, pp. 43–44, note 55.

Cleaning could remove accessible soluble material or alter a surface. That possibility matters when considering accumulation between exposures. Yet a description of hosing does not measure how much cyanide left the pores, how deep the effect extended, or which chemical compounds were removed. Green’s estimates of large dilution factors remain estimates.

The distinction is particularly important for the objection that washing would not reach material absorbed deeper in the wall. That objection cannot be dismissed by invoking surface cleaning alone. One must ask where the material was retained and what subsequently happened to it. Conversely, the possibility of deeper retention does not establish that a large stable pigment deposit necessarily formed there.

The proposed threshold explanation

Green discusses a route to iron-blue formation drawing on the solution-chemistry work of Alich, Haworth, and Johnson. His discussion uses a cyanide concentration of roughly 3.3 × 10⁻⁴ mol/L in explaining a proposed reaction requirement. This is not an experimentally established universal threshold for Auschwitz masonry. It should not be used as a single switch that proves pigment must form above it and cannot form below it in every material. Green, 2001, pp. 45–50.

The useful underlying point is that chemical conversion need not be proportional to nominal exposure. A reaction may depend on concentration, pH, available reactants, and time; competing pathways and later changes may matter. Two different exposure histories can consequently produce different proportions of persistent and more readily extractable compounds.

But explaining a possible mechanism is not the same as demonstrating that it produced the observed distribution. The threshold account cannot be treated as a completed experiment on the historic walls. The residue argument fails because its own predicted equivalence is not established, not because every detail of an alternative chemical reconstruction has already been measured.

5. What the Kraków investigation actually found

The method and its limits

Jan Markiewicz, Wojciech Gubała, and Jerzy Łabędź published their investigation in 1994. They deliberately used an acid microdiffusion procedure that did not break down Prussian blue. The safest description of the reported quantity is therefore cyanide released under the assay’s conditions. Calling it simply “free cyanide,” “water-soluble cyanide,” or all “non-iron cyanide” can imply a specificity that has not been established for every compound present in the material.

The distinction prevents two opposite errors. The assay cannot be used as a measurement of total retained cyanide. But its selectivity does not mean it measures nothing. A laboratory can obtain useful information about one fraction of a substance while leaving another fraction unmeasured. The question is what conclusions the method supports.

The paper reports a lower limit of determinability of 3–4 µg/kg, triplicate analyses, repetition of positive results, calibration standards in each series, and laboratory work performed by a group separate from the sampling team. These procedures provide reasons to take the results seriously. They do not amount to independent replication by another laboratory or complete characterisation of every possible interference. 1994 paper, methods and sampling.

Results from all five crematorium buildings

The table below summarises the published results without selecting only the largest values. A “sample” is a piece of material; its three repeat analyses are not three independent samples. A numeric positive means that at least one positive numerical value was reported. The ranges contain the positive numerical readings, excluding zeros and the separate trace entry.

Crematorium IV sample 42 was recorded as trace / 0 / 0. Some other samples were entirely negative. Thus, among the 33 crematorium samples, 27 had numerical positives, one had the trace entry, and five had only zeros. This describes the specimens tested, not the percentage of all walls that would test positive. Original publication, Tables II–III, pp. 22–23.

The finding is consequently more precise than either common slogan. It is false to say that no cyanide was detected in the crematoria. It is also too broad to say that the study obtained a positive result from every homicidal chamber or every sampled surface. The reported positives occur among the material assigned to each crematorium building tested. Building-level coverage is not a complete room-by-room or surface-by-surface survey.

The size of the positives

The units matter. One milligram per kilogram equals 1,000 micrograms per kilogram. A result of 640 µg/kg is 0.640 mg/kg. Expressing it as a decimal does not make it analytically indistinguishable from zero.

Relative to the paper’s stated 3–4 µg/kg limit, the largest crematorium result is about 160–213 times the limit. Several other higher results are also many times above it. But the smallest numerical positives, 8 µg/kg, are only about two to three times the stated limit. The reassuring ratio for a large positive must not be silently transferred to every small one.

Neither comparison supplies a full uncertainty budget. Analytical credibility still depends on blanks, recovery, specificity, reproducibility, and the way the limit was determined. The modest but decisive point is that the allegation that all the crematorium numbers were printed below the paper’s own stated limit is contradicted by its tables.

What the controls actually were

All eight dwelling-control samples in Table I were negative. Two controls from the earlier screening were negative as well. The full-study table describes the dwelling rooms as probably fumigated once. These are therefore not demonstrably unexposed specimens with a fully known history. 1994 paper, Table I and screening discussion.

Their results still matter. They weaken an explanation in which ordinary camp masonry automatically gives the same positive signal under the assay. But they do not eliminate every material-specific interference, and they do not prove that a single exposure always becomes undetectable. Eight pieces from selected locations cannot establish the behaviour of every brick or plaster surface in the camp.

There were also zero results in known delousing facilities. This is especially important for the logic of a negative finding. A zero from one piece, with one extraction procedure, cannot reliably be equated with a history of no cyanide exposure. Conversely, a positive needs interpretation; neither result abolishes the need to understand the sample.

Screening, sheltered surfaces, and representativeness

The initial 1990 screening found a positive in only one of ten samples from the ruins. The subsequent work sought material sheltered from rain. That design choice is relevant, not something to conceal: the later positives should not be presented as the outcome of random sampling across every surviving surface.

Selecting protected material can be sensible when investigating traces vulnerable to environmental loss. It also changes the question answered. Such sampling may be useful for determining whether any identifiable cyanide remains, while being unsuitable for estimating an unbiased average concentration across the entire building. Confusing those goals exaggerates the result.

The specimens were small amounts of brick, concrete, plaster, or mortar—roughly 1–2 grams in the described sampling—not a universally controlled microscopic surface layer. The fact that the samples were small and often sheltered does not by itself establish a uniform depth or guarantee the original exposed face. Those details remain relevant to quantitative comparison.

The proper conclusion from the study

The reported pattern is evidence of cyanide compounds in historically relevant structures, with negative dwelling controls and substantial variation within the sampled locations. It rebuts the blanket chemical-absence claim. It does not establish the original gas concentration, the number of operations, the purpose of every exposure, or the number of people killed.

This bounded conclusion is not evasive. It preserves what a chemical assay can contribute while leaving historical questions to the wider evidence. Insisting that chemistry alone identify motive or purpose sets it a task that residue analysis cannot perform in either direction.

6. The analytical objections in CODOH’s literature and forums

“Excluding Prussian blue rigged the result”

The exclusion is a genuine limitation if the question is total cyanide burden. A stronger investigation would report complementary assays on matched material. The Kraków authors’ discussion of the possibility of blue paint should not be treated as an established explanation of the delousing stains. Accepting those criticisms does not show that the measured results were fabricated or meaningless.

A selective assay answers a narrower question. It can help establish whether a particular recoverable fraction is present even where a persistent pigment is absent or scarce. It cannot bypass all formation and preservation effects and become a pure detector of historical exposure. Describing it as “the right test” without qualifications is too categorical; describing it as no test at all is equally misleading.

The actual results also undermine the assertion that excluding pigment made every location identical. The dwelling controls were zero, while the crematorium values ranged from zero to 640 µg/kg. The delousing table reached 900 µg/kg and also contained negative samples. Selectivity limits interpretation; it does not erase these differences. 1994 study, Tables I–IV; CODOH discussion, December 2024.

“Sub-ppm results are just noise”

Whether a result is distinguishable from noise depends on the method and the sample, not on whether it is written as a small fraction of one part per million. There is no universal rule that all sub-ppm measurements are meaningless. The appropriate questions concern calibration, background, recovery, uncertainty, and interferences in the actual analytical procedure.

An older sensitivity figure expressed per litre of analytical solution also cannot be compared directly with a result expressed per kilogram of original solid. The calculation requires the mass of solid processed, the relevant solution volumes, concentration or dilution steps, and recovery. Without those factors, the units refer to different denominators. Apparent disagreement may therefore be an invalid conversion rather than evidence of impossible sensitivity.

The published sensitivity is not beyond challenge. An independent laboratory could test it with appropriate materials and controls. But that is different from asserting that the reported positives fall below a limit when, numerically, they do not. Rudolf and Kollerstrom, analytical-method discussion.

“Carbonate explains the low values”

An interference claim needs a specific mechanism, method, material, and direction of effect. Carbonate-related concerns in one total-cyanide procedure cannot simply be transferred to every cyanide assay. Nor is obscuring a small signal the same as generating a false positive. These possibilities must be distinguished before invoking “carbonate” as a universal explanation.

Rudolf and Kollerstrom discuss analytical difficulties and discrepancies between laboratories. Such discrepancies are reasons to investigate low results carefully. They do not establish that every small result from every study is an artifact. A persuasive challenge would show that the relevant interference reproduces the positive readings under the conditions actually used, with appropriate blanks and recovery checks.

The dwelling controls help address some background possibilities, but they do not solve every matrix problem. Both points can be true: the controls make an indiscriminate background explanation less persuasive, and further validation would improve confidence in the smaller measurements.

“The ratio between bound and unbound cyanide is inexplicable”

The ratio argument compares the very large persistent deposits reported in some delousing material with much smaller selectively recovered quantities, then contrasts that pattern with crematorium results. It is reasonable to ask why different forms occur in different proportions. It is not reasonable to assume that every material must have the same conversion efficiency regardless of its exposure, composition, and subsequent history.

There is an additional sampling problem. Dividing a selective-assay result from one investigation by a total-cyanide result from a different specimen does not measure the chemical fraction in either specimen. Even within a building, the published concentrations vary sharply. Different positions, depths, dates, laboratories, and extraction procedures can affect the comparison.

Claims such as “less than one part in a thousand” or “one part in ten” should therefore not be treated as directly measured conversion efficiencies unless the numerator and denominator genuinely describe matched material. A striking cross-study ratio may suggest a question; it cannot supply its own validation. Green’s discussion of nonlinear formation offers possible chemistry, but not a retrospective assay of those unmatched specimens. Green, 2001, pp. 45–50.

“Criticism of Leuchter does not answer Rudolf”

Correct: a weakness in one investigation is not automatically a refutation of every later investigation. Leuchter’s sampling, his laboratory instructions, Rudolf’s specimens, and the Kraków assay must each be examined on their own terms. A rebuttal that stops at Leuchter’s credentials or one defect in his procedure does not answer the more developed chemical argument.

The persistent difficulty for the later argument is nevertheless the same inferential one. Better discussion of chemistry and additional total-cyanide measurements do not establish a reliable expected residue unless the exposure, conversion, preservation, and sampling conditions are adequately connected. The rebuttal concerns that missing connection, not merely the identity of the person making the claim.

7. The church case, sampling depth, and weathering

One fumigation can produce blue staining

The Bavarian church example reproduced by CODOH is relevant. The account describes staining that developed after a single HCN fumigation and continued to emerge over time. It involved particular plaster and building conditions. The case defeats any universal assertion that blue pigment can form only after thousands of exposure hours. A careful rebuttal should accept that consequence. Church report and appended commentary.

But the inverse inference does not follow. If one exposure produced staining in one favourable material, it does not follow that every repeatedly exposed material must show equal or greater staining. A single case establishes a possibility. It does not provide a representative frequency across all fumigations or a validated comparison with the crematorium walls.

Rudolf’s appended historical argument goes beyond the case report when it balances different factors and concludes that the crematorium conditions were as favourable or more favourable overall. That is the step requiring evidence. Differences cannot be declared to cancel one another simply because each can be described qualitatively.

Delayed staining also deserves proper weight. It means the appearance of pigment need not coincide with the period of gas exposure. Subsequent chemical changes may matter. That observation challenges simplistic explanations on both sides: neither immediate absence of blue nor an assumed instantaneous conversion accurately captures every possible pathway.

The ten-micrometre argument is not a universal law

James Roth’s comments about very shallow penetration have often been used to argue that crushing a large sample necessarily dilutes all meaningful surface residue almost to nothing. Sample depth certainly matters. If a residue is concentrated in a thin layer, mixing it with a large mass of clean material lowers the concentration per kilogram.

But a universal ten-micrometre penetration limit is not defensible for all masonry and exposure histories. Green himself questioned its general application, and the literature discusses deposits and cyanide extending below the immediate surface. Rudolf and Kollerstrom’s depth-related results cannot be answered simply by repeating the ten-micrometre figure. Green, 2001, pp. 14–16; Rudolf and Kollerstrom, 2013.

The proper response is to document the profile. Which face was exposed? Was it an original surface? How thick was the specimen? Which layers were analysed? Were concentrations expressed for the whole specimen or a defined layer? These questions explain why sampling geometry matters without assuming the answer in advance.

A large fragment is not intrinsically useless, and a small scraping is not intrinsically representative. Different sampling strategies may answer different questions. Comparisons become especially weak when one study analyses a whole fragment while another analyses a small protected portion, without accounting for their different denominators and spatial distributions.

Stable pigment and weathered ruins are different issues

Prussian blue can be persistent. It is therefore inadequate to dismiss the lack of heavy staining with the casual claim that rain must have dissolved all the pigment. Evidence about loss of a more readily recovered cyanide fraction does not demonstrate removal of a mature pigment deposit.

The Kraków water-flushing experiment reported removal of 82.5% and 90.7% of the measured cyanide from laboratory-fumigated plaster. That result concerns the fraction measured under its conditions. It is not an experiment showing that decades of rain would erase any quantity of Prussian blue from any historic wall. 1994 paper, laboratory experiments.

The opposite absolute is also wrong. Chemical stability does not guarantee that the pigment-bearing material itself survives. Demolition, erosion, removal of plaster, and loss of an original surface can remove evidence without chemically dissolving every pigment particle. A protected, standing delousing wall and a broken, exposed fragment do not automatically have identical preservation histories.

Weathering is therefore one part of the explanation to investigate, not a universal escape clause. It is necessary to distinguish compounds that may be lost, surfaces that may be physically missing, and protected original material that remains available for testing. The argument from present absence becomes strong only where the expected formation and subsequent preservation are both adequately established.

8. Could ordinary delousing explain the positive residues?

Chemistry alone cannot distinguish a homicidal exposure from pest control. This is a real limitation, and it should be stated without qualification. Cyanide is not a chemical label for murder. The interpretation of a positive result depends on where the specimen came from and what independent evidence exists for the use of the room.

The particular appeal to a camp-wide fumigation in 1942 is inadequate as a blanket explanation for later operational facilities at Birkenau. The crematoria there entered service in 1943. An earlier camp event does not establish that the later rooms, in their later configuration and with their later surfaces, received the exposure being invoked. Auschwitz Museum, gas-chamber chronology.

That chronological point does not exclude every possible later non-homicidal exposure. CODOH’s more recent discussion raises the broader possibility of pest control rather than resting entirely on one dated event. The answer must address that broader formulation, not pretend that disproving the 1942 explanation disproves all alternatives. CODOH discussion, July 2025.

An alternative explanation nevertheless needs evidence of its own. The general presence of lice in a camp does not document when a particular crematorium room was fumigated, how often, or which surviving surfaces were exposed. A plausible possible source is not yet an established historical source.

There is also a distinction between explaining the chemistry and explaining the building’s use. Even if some residue could be attributed to ordinary disinfestation, that would not dispose of testimony about killing, gas-tight fittings, introduction devices, construction records, and the wider extermination evidence. A non-homicidal explanation of one measurement must not be silently expanded into a non-homicidal explanation of every other observation.

The point cuts both ways. A positive result should not be described as a standalone proof of homicidal use; an alternative chemical source should not be described as a standalone disproof. Purpose is established by the convergence of evidence tied to the particular building and period.

9. Roof openings, ventilation, and gas-tight doors

The chemical argument is often joined to architectural claims. These need separate examination because a conclusion about a laboratory assay does not settle the layout or operation of a building. They also feed back into the chemistry: assumptions about introduction devices, removal of the carrier, and ventilation affect estimates of exposure duration.

The claim that no introduction openings existed

The strongest version of this objection is more specific than “I cannot see four holes in a photograph.” Carlo Mattogno disputes the identification of surviving openings, interpretations of photographs, differences in witness descriptions, and the assignment of inventory entries. Those are arguments about particular pieces of evidence. Mattogno, 2004.

There is, however, a substantial published physical investigation to answer. Daniel Keren, Jamie McCarthy, and Harry Mazal identified three of four proposed introduction openings in the remains of Crematorium II and a probable location for the fourth under rubble. Their case draws on the structure of the concrete, reinforcement, traces of waterproofing, photographs, and testimony. It does not claim the survival of four undamaged apertures or four intact original introduction columns. Keren, McCarthy, and Mazal, 2004, especially pp. 68–76.

That distinction matters. The slogan “no holes” presents the physical record as empty. The published investigation describes particular features and gives reasons for identifying them. Rejecting it requires engaging those features and reasons. Simply noting that the roof was damaged, or that some openings are difficult to recognise, does not show that the proposed identifications are wrong.

Conversely, citing the paper does not automatically resolve every disputed dimension or documentary label. The strength of a physical identification depends on the combined observations and the plausibility of alternatives such as later damage. A disagreement about one detail must be assessed at that level, rather than inflated into the disappearance of all the other evidence.

The Museum also reproduces an inventory referring to four wire-mesh devices and wooden covers, and Tauber’s 1945 account describes nested mesh structures with a removable inner component. The inventory’s assignment and the details of the testimony are among the contested points, but they are evidence to evaluate alongside the ruins. They are not generated by the cyanide assay. Museum’s documents and testimony concerning the openings.

This is why exposure calculations should not silently assume that no removal apparatus existed. That assumption imports a disputed architectural conclusion into a chemical argument, then risks using the chemical argument to reinforce the same conclusion. Equally, the evidence for an apparatus should not be turned into a precise measured release history for every operation. The historically supported arrangement and an exact quantitative reconstruction are different levels of claim.

Ventilation: equipment, ideal models, and actual operation

CODOH’s stronger ventilation arguments acknowledge fans but dispute their capacity, the rate of clearance, continued release from the carrier, and the plausibility of reported timings. Answering only that ventilation equipment existed misses the technical objection. Mattogno, “The Crematories of Auschwitz,” ventilation discussion.

One elementary issue can be clarified mathematically. In an ideal well-mixed space, with a constant ventilation rate and no continuing source, each successive air exchange removes the same fraction of what remains. It does not remove a fixed fraction of the original amount. After one ideal exchange, about 37% remains; after two, about 14%; after five, about 0.7%. This is the exponential dilution relation discussed by Green. Green, 2001, pp. 3–9 and 30–37.

The model corrects a basic misunderstanding about air exchanges. It does not supply a universal entry time. Real results depend on continuing emission, obstructions, imperfect mixing, local stagnant regions, adsorption and release from surfaces, and the difference between rated and delivered airflow. A nominal fan specification is not a measurement at every point in an occupied or obstructed room.

The appropriate conclusion is therefore neither “a fan proves every reported timing” nor “imperfect ventilation proves operation was impossible.” One must identify the specific timing claim, specify the assumed source behaviour and airflow, and test the inference under realistic uncertainty. An ideal model can illuminate feasibility while leaving particular witness estimates open to examination.

This also prevents a circular treatment of the cyanide residue. If a critic assumes prolonged emission because the carrier supposedly could not be removed, that premise must be supported independently. If a defender assumes immediate and complete clearance, that assumption also needs support. The quantitative chemistry cannot be stronger than the historical and engineering inputs used to construct it.

Wooden doors do not make gas containment impossible

The assertion that a gas chamber would necessarily require a metal door is contradicted by the use of wooden gas-tight doors in disinfestation facilities. The relevant properties concern construction, seals, fittings, and closure, not merely whether the principal material was wood.

The Auschwitz Museum presents construction correspondence concerning gas-tight doors for the Birkenau crematoria, photographs of surviving or recovered doors, and details of layered construction and seals. These are concrete answers to the blanket material objection. They do not imply that every ordinary wooden door is gas-tight, or that every displayed item is untouched since wartime. Museum, door construction and contemporary records.

The related claim that bodies would prevent an inward-opening door from being opened must be checked against the relevant plan. The Museum’s published plans show outward opening from the gas chambers. One cannot establish an architectural impossibility by assigning the wrong opening direction to the room being discussed. Museum, plans and opening direction.

These answers are limited in the appropriate way. They rebut broad claims about impossible materials and door orientation. A dispute about the dimensions, date, or identity of a particular fitting would still need a particular answer. General technical language should not substitute for checking the actual object and document.

10. Cremation, death records, reconstruction, and Majdanek

Oven capacity is not the same as total disposal capacity

Cremation arguments raise questions about throughput, fuel, operating time, interruptions, loading, and disposal outside the ovens. They are independent of the residue assay and cannot be settled by declaring that cyanide was detected. Mattogno’s technical discussion is a source for the objections, but its conclusions must be assessed against the full historical disposal system. Mattogno, “The Crematories of Auschwitz”.

The SS construction office’s report of 28 June 1943 stated a combined capacity of 4,756 bodies per 24 hours for the five crematoria. This is contemporary evidence of an asserted capacity. It is not a daily record proving that all installations operated at that rate continuously. Museum researchers distinguish the document’s historical importance from the separate question of exact practical performance. Museum researchers on the capacity document.

It would therefore be unsound to multiply that nominal capacity by every calendar day and present the product as a measured death toll. The inverse shortcut is also unsound: taking a rate from a different cremation practice, assuming it is a universal maximum, and applying it to all wartime operations without examining the differences.

Most decisively for an ovens-only ceiling, the disposal system was not confined to the ovens. The clandestine photographs made by Sonderkommando prisoners in 1944 document open-air burning in the vicinity of Crematorium V. The Museum preserves the images and the history of their transmission outside the camp. Museum, clandestine documentation; historical photographs.

A calculation that excludes a documented disposal route cannot establish a maximum for all disposal. At the same time, the existence of open-air burning is not a complete numerical model of fuel consumption or daily throughput. A serious reconstruction must consider operating periods, interruptions, the character of the fuel records, and each disposal route. The simple exclusion argument fails because its accounting is incomplete, not because every practical parameter is known exactly.

The death books are a defined record series, not a census of all victims

An administrative total is meaningful only after identifying whom the administration entered in that record and which volumes survive. The Auschwitz archive describes 46 surviving death-book volumes with almost 69,000 recorded deaths of registered prisoners, covering 29 July 1941 to 31 December 1943. People murdered on arrival without registration were not entered into that series. Auschwitz archive, description of the death books.

Using those entries as the total number of Auschwitz victims commits a coverage error. It substitutes the population recorded by one surviving series for everyone who died or was murdered there. It also ignores the series’ date limits. The reliability of individual entries does not repair a mismatch between the register’s scope and the question being asked.

This is a general archival principle, not a special exemption created for Auschwitz. An admissions register cannot automatically count people never admitted to the registered population. A surviving partial series cannot automatically count missing periods. To establish a complete total, one must integrate the relevant records and evidence of what happened outside that administrative process.

CODOH’s use of registration totals, including Irving’s “Battleship Auschwitz” presentation, must be assessed against that scope. The existence of authentic records does not validate an inference that their compiler never intended them to support. CODOH text.

Reconstruction is a question of dating evidence, not proof of invention

Crematorium I underwent wartime alterations and postwar reconstruction. The Museum openly explains that ovens and a chimney were reconstructed using original components and that roof openings previously closed were reopened. Visitors should not treat every visible feature as an untouched survivor of a single wartime moment. Museum, Crematorium I.

That fact has a clear evidential consequence: a reconstructed feature cannot be used as though its present appearance alone independently proved the original arrangement. Its evidential value depends on the documentation supporting the reconstruction and on the surviving original fabric.

It does not follow that the building’s earlier homicidal use was invented. Reconstruction and historical existence are different questions. A feature may be reconstructed on good evidence, reconstructed imperfectly, or remain disputed without every event associated with the building becoming fictitious. The proper inquiry identifies what is original, what changed during the war, what was restored afterwards, and which independent sources describe each phase.

This also matters when shifting between sites. An issue concerning the presentation of Crematorium I does not automatically dispose of the destroyed Birkenau facilities, their construction records, or testimony about their operation. The same standards of provenance and dating should be applied to each structure separately.

Majdanek and the danger of assigning function by colour

Rudolf’s commentary invokes blue-stained facilities at Majdanek while assigning them a disinfestation function. The State Museum at Majdanek, by contrast, identifies homicidal gas chambers and describes the use of both Zyklon B and carbon monoxide. The blanket assertion that homicidal gassing did not occur there is not supported by the Museum’s account. CODOH commentary; Majdanek Museum, extermination; Museum timeline.

The proper lesson for the cyanide debate is methodological. Colour alone cannot establish the function of a room. A particular structure may require a history distinguishing construction phases, different rooms, and different periods of use. One must not infer “only delousing” merely from the existence of pigment, any more than one should infer “homicidal” from pigment alone.

Nor does a blue deposit at another camp supply a universal expected concentration for Auschwitz masonry. It may establish that certain exposure and material conditions can leave persistent compounds. Transferring that observation to a different building still requires the comparison that the residue argument has not adequately established.

11. How the evidence fits together

The historical case does not depend on making every witness exact, every administrative label transparent, or every chemical calculation complete. It depends on evaluating the relationship between sources: contemporary construction records, physical remains, testimony about operation, clandestine prisoner documentation, photographs, and records of deportation and registration.

The Sonderkommando evidence is particularly relevant to the assertion that the case is only a collection of later stories. The Museum documents clandestine writings buried by prisoners near the crematoria and photographs smuggled out during the war. These sources have their own provenance and limitations, but they are not simply repetitions of a postwar chemical hypothesis. Museum, writings and photographs created by Sonderkommando prisoners.

Convergence must not be confused with counting every retelling as independent confirmation. Two authors may rely on the same witness; an architectural interpretation may be influenced by testimony; a later account may repeat an earlier one. A sound assessment identifies those dependencies. It also recognises when genuinely different kinds of evidence bear on the same event.

This standard answers a recurring move in the denial literature: finding one uncertain detail and treating it as if it were the sole support for the whole history. A questionable dilution estimate does not remove a construction document. A disputed feature of a roof does not make a death register complete. An error in a CO₂ calculation does not turn measured positive cyanide values into zeros. Each correction changes the claim it actually bears on.

The same discipline applies to the rebuttal. A valid criticism of an ovens-only capacity argument does not establish every numerical estimate of cremation. A positive assay does not authenticate every witness detail. A museum’s reconstruction does not become original fabric through repetition. Keeping these distinctions clear makes the historical argument stronger because it rests on what the evidence supports, rather than requiring every convenient assertion to be true.

12. Why the residue argument does not overturn the historical record

The strongest cyanide argument asks a legitimate question: why do the surviving residues differ so sharply? Its failure lies in presenting an inadequately calibrated comparison as a decisive answer about historical use. The observed difference does not itself specify how much gas reached each wall, which reactions occurred, how much material was lost, or how the sampled portions represent the former rooms.

Several conclusions follow from the evidence examined here:

• There is no blanket absence of cyanide. The Kraków study reported positive numerical results among samples from every crematorium building it tested, while also reporting zeros and a trace result. Its data cannot honestly be reduced to “nothing was found.”
• Visible pigment and historical exposure are not equivalent. The church case shows that one exposure can produce staining under particular conditions. It does not show that every repeatedly exposed wall must develop the same deposits.
• Neither assay type supplies a direct historical dosimeter. Selective analysis omits part of the retained cyanide; broader analysis still measures what survives in the specimen, not everything that ever entered the room.
• A probabilistic claim needs a justified expectation. Unknown formation, retention, and preservation rates cannot be replaced by an unsupported assurance that all the relevant factors must cancel in one direction.
• Historical purpose requires historical evidence. Ordinary delousing is a possible source of cyanide, but invoking it does not explain away the independent evidence for the operation of the homicidal facilities.
• The related technical objections have to confront their own evidence. Surviving plans, gas-tight fittings, the physical roof investigation, clandestine photographs, and the defined coverage of administrative registers cannot be dismissed by a wall-colour comparison.

Further chemical work could improve the comparison. Useful additions would include independently replicated measurements, well-documented original surfaces and depth profiles, matched specimens tested by complementary assays, material-specific controls, and explicit uncertainty estimates. A model would need to account for changing conditions and preservation, not merely select favourable assumptions until it reproduces a preferred conclusion. These are standards for stronger evidence, not a claim that nothing can be known until every chemical process has been reconstructed.

The remaining uncertainty concerns how precisely modern residues can reconstruct past exposure and subsequent material history. It does not establish uncertainty about whether the Nazis murdered people with gas at Auschwitz. That historical conclusion rests on a much wider and mutually supporting record. The cyanide residue argument has not supplied the validated prediction necessary to overturn it.

A difference in surviving residues is a fact to investigate. It is not a demonstration that the murders did not occur.

Sources

The links throughout the essay connect particular claims to their supporting evidence. The principal chemical studies and historical investigations are collected here for convenient reference. CODOH publications are cited as sources for the arguments being answered, not as authorities for their historical conclusions.

• Jan Markiewicz, Wojciech Gubała, and Jerzy Łabędź (1994). “A Study of the Cyanide Compounds Content in the Walls of the Gas Chambers in the Former Auschwitz and Birkenau Concentration Camps,” Z Zagadnień Nauk Sądowych 30, 17–27. Full-text transcription; scanned publication. The scan supplies the authoritative table layout and sample numbering.
• Richard J. Green (2001). Expert report addressing Rudolf’s claims in the Irving appeal proceedings. Report. Includes the exposure estimates, chemical discussion, ventilation calculations, and cleaning-testimony references discussed above.
• Germar Rudolf (2020). The Chemistry of Auschwitz, fifth-edition presentation script. Text.
• Germar Rudolf and Nicholas Kollerstrom (2013). “Differential Exposure of Brickwork to Hydrogen Cyanide during World War Two.” Article.
• “Wood Preservation Through Fumigation With Hydrogen Cyanide.” Case report reproduced with Rudolf’s commentary. The reported building case and the appended comparison with Auschwitz are distinct evidential claims.
• Daniel Keren, Jamie McCarthy, and Harry W. Mazal (2004). “The Ruins of the Gas Chambers: A Forensic Investigation of Crematoriums at Auschwitz I and Auschwitz-Birkenau,” Holocaust and Genocide Studies 18(1), 68–103. Study.
• Carlo Mattogno. “The Openings for the Introduction of Zyklon B—Part 1” and “The Crematories of Auschwitz”. Sources for the architectural, ventilation, and cremation objections.
• CODOH forum discussions. “The Prevarications of Markiewicz,” December 2024 and “Markiewicz Report in 1994,” July 2025. References here concern the accessible opening pages, not every contribution in those discussions.
• Auschwitz-Birkenau State Museum. Gas-chamber chronology, introduction openings and contemporary inventory, gas-tight doors, door orientation, death-book collection, Crematorium I, and Sonderkommando documentation.
• State Museum at Majdanek. Extermination and timeline.