Cyanide in the Historical Darkroom, and the Ferricyanide That Is Not Cyanide

Hardwich’s 1864 manual lists what a portrait photographer should pack before going out to work. Among the plates and the bottles, without emphasis and without comment, is a line that reads “cyanide of potassium for cleaning fingers and fixing”. One salt, two jobs, and the second of them is washing your hands with it. Five items earlier in the same list, between the pyrogallic acid and the citric acid, is eight ounces of glacial acetic acid.

That line, and the acid five items above it, are the subject of this page. Neither was written by a reckless man. Hardwich is careful, and four sentences after he introduces the salt he calls it highly poisonous and says it must be used with caution. He was somebody who knew the substance was dangerous, knew exactly what it did to silver, and had no way of knowing what it does inside a cell, because the mechanism this page sets out was not available to him or to anybody else in his century. Holding both halves of that at once — the chemistry that made the choice rational and the biochemistry that makes it unthinkable — is most of what this page is for.

What this page contains, and where the rest of it lives

☠Level D: there is no procedure here, and none is withheld by accident

This lesson is Level D under the Level D policy, which governs what it may contain. It gives the chemistry as reactions and mechanism, the historical practice in outline, the image the practice produced, and the hazards with their sources and dates. It gives no quantity, no concentration, no bath, no time, no sequence and no substitute recipe, and it does not describe how any cyanide solution was compounded.

The prohibition covers the tone as well as the headings. There is no sentence anywhere below of the form “if you must do this”. A partial procedure reads as permission and omits the controls, which is worse than nothing at all.

Four other pages hold parts of this subject, and this one does not repeat them.

• Potassium cyanide’s encyclopaedia entry carries the full hazard record — the aggregated notifications, the exposure limits, the solubility, the incompatibilities and the waste classification — with every source and the date it was read.
• The formulary entry for cyanide fixing and toning carries the documentary evidence: which manuals printed which strengths, what a cyanide-fixed plate looks like to a conservator, and the demonstration that the nineteenth century’s so-called cyanide toning bath was in almost every case a thiocyanate.
• Part XX owns cyanide as it acts on a finished silver gelatin print, alongside uranium, mercury and chromium(VI).
• Part XI owns fixing itself, and did the arithmetic this page starts from.

What is left, and what this page owns, is cyanide as a working practice in a room — why it was chosen, what it did to the people who chose it, and the boundary between it and a salt with the same syllables in its name that sits on this course’s own shelf.

Why it fixed better, in numbers

Fixing is complex formation. A ligand takes a silver ion out of the crystal lattice and holds it in solution, and the whole question is how hard it holds. Thiosulfate holds one way; cyanide holds another.

AgBr + 2 S₂O₃²⁻ ⇌ [Ag(S₂O₃)₂]³⁻ + Br⁻

The fixer this course teaches: two thiosulfate ions per silver

AgI + 2 CN⁻ → [Ag(CN)₂]⁻ + I⁻

The fixer the wet-plate darkroom often preferred: two cyanide ions per silver

The two reactions have the same shape, and the difference between them is one number. OpenStax’s table of formation constants gives the dicyanoargentate ion a K_f of 1 × 10²¹. The bis(thiosulfato)argentate ion, whose constant Part III established and Part XI uses throughout, has a K_f of 4.7 × 10¹³. Cyanide’s grip on silver is about twenty million times tighter.

Part XI showed what to do with that. Dissolving a halide and complexing the silver share the silver ion, so the two equilibria add and their constants multiply:

K = K_sp × K_f

The combined constant for fixing

Run it for both ligands across the three halides and the entire history of the wet plate falls out of the table.

The combined fixing constant, thiosulfate against cyanide

• Cyanide, Kf = 1 × 10²¹
• Thiosulfate, Kf = 4.7 × 10¹³

Show the numbers behind this plot

Two series of three points each, plotted against a horizontal axis carrying silver chloride at position one, silver bromide at position two and silver iodide at position three. The vertical axis is the base-ten logarithm of the combined equilibrium constant for dissolving that halide in that ligand. The thiosulfate series, drawn lower, runs from plus 3.88 at silver chloride through plus 1.38 at silver bromide to minus 2.15 at silver iodide, crossing the horizontal guide line at zero between bromide and iodide. The cyanide series, drawn above it, runs from plus 11.20 through plus 8.70 to plus 5.18 and never approaches the guide line at all. The vertical gap between the two series is the same 7.3 units at every halide, because it is nothing but the ratio of the two formation constants. The teaching point marked on the plot is that thiosulfate's line crosses zero between bromide and iodide while cyanide's stays five orders of magnitude above it at the worst halide, so a wet collodion plate made of silver iodide is a plate that thiosulfate is thermodynamically unable to clear and cyanide clears easily.

Series	Silver halide, in order of decreasing solubility product	log₁₀ of the combined constant K
Cyanide, Kf = 1 × 10²¹	1.00	11.20
Cyanide, Kf = 1 × 10²¹	2.00	8.70
Cyanide, Kf = 1 × 10²¹	3.00	5.18
Thiosulfate, Kf = 4.7 × 10¹³	1.00	3.88
Thiosulfate, Kf = 4.7 × 10¹³	2.00	1.38
Thiosulfate, Kf = 4.7 × 10¹³	3.00	-2.15

Computed, not measured. Each point is the product of the solubility product Part XI adopted and the formation constant OpenStax Appendix K gives for that ligand; the horizontal positions are labels rather than a quantity. The derivation is the course's own, following Part XI's method. The curve is drawn to show the shape, not measured from a real material. Your own materials will differ, and measuring them is what the sensitometry part of the course is for.

A collodion plate is silver iodide, sometimes with bromide added later in the period, poured as a thin layer in nitrocellulose. Kodak’s 1928 primer states the consequence in a dozen words, with none of this arithmetic behind it: potassium cyanide was used for fixing wet collodion plates, “which, being made from silver iodide, are not easily fixed in hypo”. A combined constant of 0.0071 means the equilibrium sits on the side of the solid. A combined constant near 1.5 × 10⁵ means it does not. The wet-plate photographer who reached for the cyanide bottle was not being cavalier; he had a material one reagent could not touch and another that could.

⚗The second property, which decided it for a photographer working from a cart

Speed was not the whole of it. Hardwich adds a property that no modern fixer has and that mattered enormously to somebody sensitising, exposing, developing and fixing a plate before the collodion dried: the soluble double salt cyanide forms, unlike the double sulphocyanide, is not decomposed by dilution with water.

Read that as a washing problem and it is decisive. A thiosulfate complex is an equilibrium that runs backwards when you dilute it, which is exactly why Part XII spends a whole part on getting the last of it out of a fibre base. A dicyanoargentate ion put into rinse water stays a dicyanoargentate ion and leaves. The cyanide-fixed plate washed in a fraction of the time.

The salt it made has a name and a registry number. The World Health Organization’s assessment of cyanides lists potassium silver cyanide, KAg(CN)₂, CAS 501-61-6, in its own table of cyanide compounds. That is the point worth carrying: the used bath was not a spent fixer in the ordinary sense. It was a cyanide solution that now also contained silver, and it stayed one.

The same avidity is why it could not be trusted. A ligand strong enough to strip silver out of a crystal lattice is strong enough to strip it out of the picture, and Hardwich describes the attack in two stages — the image converted superficially to silver cyanide, then dissolved as the double salt. Wall’s 1912 dictionary gives the working photographer’s version half a century later: cyanide is more powerful than hypo, “but its action on the image is so great as to deteriorate the half-tones occasionally”. One bath cannot be selective in one direction only, and the formulary entry works through why the half-tones go first.

Where else it turns up in the literature

A reader going through period manuals will meet the salt in three roles besides fixing, and will also meet a fourth that turns out not to be there. All four are worth naming, because a formula without the word fixer over it can still be a cyanide formula, and because one widely repeated association does not survive being checked.

As a stain remover, used on the hands. This is the use that reads worst now, and all three of the sources are matter-of-fact about it. Hardwich packs it “for cleaning fingers and fixing”. Towler’s Silver Sunbeam gives it for removing silver nitrate and pyrogallate stains, and — in the same breath — warns that doing so “might entail upon the operator incurable ulcers”. Wall’s 1912 dictionary still carries an entry for cyanogen soap: a soap containing potassium cyanide, manufactured and sold for taking silver stains off the hands.

As a reducer, usually with iodine. Hardwich records that adding a little iodine greatly increases cyanide’s solvent power on metallic silver, giving the colourless liquid the period called iodo-cyanide of potassium. Wall’s 1924 formulary prints an iodo-cyanide reducer under the heading “extremely poisonous, but clean acting”, and a mercury-and-cyanide reducer credited to Eder under “an extremely poisonous but non-staining reducer”. Cassell’s 1911 cyclopaedia reaches for the same mixture to destroy the photographic image entirely beneath ink work, calling iodine and potassium cyanide, “both of which are poisonous”, the best solvent for the job because it acts quickly and without stain.

As the high-contrast step of a published intensifier. Kodak Limited’s IN-1 mercury intensifier bleaches with mercury(II) chloride and then offers four blackening baths of increasing strength. The strongest is silver nitrate dissolved in cyanide — Monckhoven’s bath, which raises the highlights while the cyanide very slightly cuts the shadows, and so builds contrast from both ends at once. It was still in print in 1949.

Not, on the evidence, in the gold toning baths. This lesson’s manifest entry originally named gold toning among cyanide’s uses. It has been corrected rather than the page written to it, because Part XX tested that against the sources and found it does not hold: Towler’s six gold formulas are gold chloride neutralised with carbonate or chalk, Wall’s working baths are alkaline, thiocyanate or combined, and cyanide appears in none of them. This page inherits that correction rather than re-arguing it, and adds only the reason it is easy to make: the salt whose name looks like cyanide in a gold toner is ammonium thiocyanate, which period manuals spell sulphocyanide. Two different ligands, two different levels, one confusing century of nomenclature. The gold thiocyanate toner is a formula this course publishes.

What it does to a person

The photographic literature of the period says poison and stops. That is not a criticism of it — the mechanism was not available — but it means the reason for the classification has to come from somewhere else. The World Health Organization’s international assessment of cyanides, prepared for the International Programme on Chemical Safety and approved in 2003, states it in one sentence, and every consequence below follows from that sentence.

…free cyanide (hydrogen cyanide or cyanide ion) has high acute toxicity due to its primary toxic effect of inhibiting cytochrome oxidase (by binding haem iron), the terminal enzyme of the mitochondrial electron transport chain.

Why the effect is fast, and why breathing does not help

1. 1Cyanide reaches the bloodstreamNIOSH lists four exposure routes for the salt - inhalation, skin absorption, ingestion, and skin or eye contact - and the WHO assessment records that symptoms can begin within seconds of inhaling the gas and within minutes of swallowing the salt
2. 2It binds the haem iron of cytochrome c oxidaseThe terminal enzyme of the mitochondrial electron transport chain - the last step, where electrons are finally handed to oxygen. It is the same ligand doing the same kind of thing it did to silver: the assessment records that cyanide inhibits around forty enzymes, mostly metalloenzymes built on iron, copper or molybdenum
3. 3The chain stops, so the cell cannot use oxygenOxygen is still in the lungs and still on the haemoglobin. What has failed is the ability to spend it. The assessment calls the resulting state histotoxic anoxia - tissue suffocation with the oxygen supply intact
4. 4The tissues most dependent on oxidative metabolism fail firstThe central nervous system is named as particularly vulnerable because of its high dependency on oxidative metabolism and its limited anaerobic capacity. Death is believed to follow from central nervous system depression caused by inhibition of brain cytochrome oxidase

Three things about that chain matter for reading the historical record.

It explains the speed. There is no accumulating dose, no organ to be damaged first, no metabolic activation step to wait for. The molecule reaches a haem group and the last reaction of aerobic respiration stops. PubChem’s ChEBI description records the same fact in ontology terms: potassium cyanide has a role as an inhibitor of EC 1.9.3.1, which is cytochrome c oxidase’s enzyme number.

It explains the symptom list. NIOSH’s entry — asphyxia, weakness, headache, confusion, breathing first faster and then slow and gasping, with the cardiovascular system, central nervous system and blood as target organs — is what tissue suffocation looks like from outside. It is also, almost word for word, what the 1911 photographic antidote table quoted later on this page describes as “insensibility, slow gasping respiration, dilated pupils”. Two documents nearly a century apart, one mechanism, and only the later one knows what it is describing.

It is not the only enzyme affected. The assessment adds that cyanide can inhibit around forty enzymes, mostly metalloenzymes containing iron, copper or molybdenum. Cytochrome oxidase is the one that kills quickly; the rest are part of why the chronic picture exists at all.

⚗The detoxification pathway runs on thiosulfate, and this is not a first-aid instruction

The body’s main defence against cyanide is an enzyme called rhodanese, which transfers a sulfane sulfur from thiosulfate to the cyanide ion, making thiocyanate — a far less toxic species that the kidneys excrete. About 80 per cent of a cyanide dose leaves that way, and the assessment is explicit about what limits it: the rate-limiting step is the amount of thiosulfate, because the body’s supply of sulfur-containing substrates is small. Spontaneous detoxification in humans runs at roughly one microgram per kilogram of body weight per minute.

Look at what that means for this course. The species the body uses to dispose of cyanide is thiosulfate — the same ion, by name, that replaced cyanide as the photographic fixer for reasons entirely to do with silver. Whether the two roles rest on the same property of the ion is not something the course can state, because no source it holds says which sulfur atom of thiosulfate binds silver, and Part III gives none. What is established is the coincidence itself, and it is worth holding on to.

It is not an antidote you can improvise, and nothing here should be read as one. Clinical cyanide treatment is intravenous, dose-calculated and administered by people trained for it; HSE’s statutory first-aid guidance lists management of a casualty suffering from cyanide poisoning as additional training identified by a needs assessment, alongside oxygen administration where hydrogen cyanide exposure is a risk — which is to say a competence that a workplace has to acquire deliberately and that no darkroom possesses. The connection above is chemistry. It is not treatment, and a bottle of hypo is not a medical countermeasure.

Through skin, over years. The route that matters most for reading Hardwich and Towler is the dermal one. The assessment gives a rabbit dermal LD50 of 0.343 mmol/kg for potassium cyanide on intact skin, and states that dermal toxicity is markedly greater on abraded skin, which enhances penetration. Now re-read the period practice: the hands being cleaned with the salt are the hands of somebody who handles glass plates all day. Towler’s “incurable ulcers” and his observation that the health of operators is much impaired, “especially in those large printing establishments where a number of females are employed in this department, who by continual manipulation in the toning and fixing baths are frequently in a suffering condition”, describe exactly the exposure the toxicology predicts would be worst.

And the chronic tail. Thiocyanate, the detoxification product, prevents the uptake of iodine and acts as a goitrogen. In a cross-sectional study of thirty-six electroplating workers in Egypt, twenty of them, 56 per cent, showed thyroid enlargement. That is not photography, and the assessment itself notes that the workers were exposed to several other things as well, so the contribution of the cyanide is difficult to discern. It is quoted here for one purpose only: it shows that the chronic effect is real and has a named mechanism, which is what turns Towler’s “frequently in a suffering condition” from an anecdote into an observation with a plausible cause.

The reaction that turns a solid into a gas

One equation ends the argument about whether this substance can live in a darkroom, and it is not a subtle one.

KCN + H⁺ → HCN + K⁺

Any acid, at room temperature, at once

NIOSH classifies potassium cyanide a noncombustible solid and then adds that contact with acids releases highly flammable hydrogen cyanide; its incompatibility line names acids, acid salts, chlorates and nitrates. A modern darkroom is a room containing an acetic acid stop bath and an acid fixer within arm’s reach of the sink. A collodion darkroom was worse: the silver bath was held acid with nitric acid, the iron developer carried several per cent of acetic acid, and hydrochloric acid was used for cleaning.

And the evidence for that is in the packing list this page opened with. Hardwich’s outfit for a day’s work in the field puts eight ounces of glacial acetic acid, half an ounce of citric acid and a lump of potassium cyanide into the same basket, to be handled by the same wet hands under a tent. It is not a hypothetical arrangement reconstructed by a modern author. It is what one careful, well-informed practitioner told his readers to carry, in print, in 1864.

⚗The same reaction happens in the stomach, which is why swallowing is not a different hazard

The assessment makes a point the photographic literature never had the vocabulary for. When a simple cyanide salt is swallowed, the free cyanide ion meets “the highly acidic medium of the stomach”, and essentially all of it becomes hydrogen cyanide, which is then quickly absorbed. And because hydrogen cyanide has a pK_a of 9.22, at the pH of blood and tissue the substance is distributed as the molecule rather than as the free ion — so the form to which exposure occurred, salt or acid, makes no difference to distribution, metabolism or excretion.

The routes still differ in how quickly the dose arrives — the assessment notes that gastrointestinal absorption is slower than pulmonary, so onset is delayed and the severity reduced — but they do not differ in what arrives. The sink and the stomach are the same reaction in different containers.

Two numbers put the exposure in scale. NIOSH sets its recommended limit as a ceiling of 5 mg/m³ over ten minutes rather than as an eight-hour average — a deliberate choice, because an exposure that can kill inside a minute is not the kind you average — and gives 25 mg/m³ as immediately dangerous to life or health. HSE’s EH40 gives hydrogen cyanide 0.9 ppm long-term and 4.5 ppm over fifteen minutes, with the Sk notation for skin absorption. The WHO assessment’s human estimates, which it labels itself as crude averages, run from slight effects at 20 to 40 mg/m³ to immediately fatal at 300.

Reading a period warning as evidence

The most instructive document this page has is not a manual from the collodion decade. It is Kodak Limited’s Chemicals and Formulae of 1949, a mainstream trade handbook a hundred and ten years after Daguerre, printing the silver-and-cyanide blackening bath of IN-1 and then printing this beneath it:

Warning: Cyanide is a deadly poison and should be handled with extreme care. It reacts with acid to form poisonous hydrogen cyanide gas. When discarding a solution containing cyanide, always run water to flush it out of the sink quickly. Cyanide solutions should never be used in poorly ventilated rooms.

Three separate things can be read off that block, and keeping them apart is the whole exercise.

What it establishes as fact. The acid reaction was understood, stated correctly, and printed by the manufacturer beside its own formula. Nobody in 1949 was unaware.

What it establishes about publication. The formula was printed anyway. A warning is not a withdrawal. The course has read the whole volume, and — this is the course’s own observation about the document rather than a claim any source makes — the cyanide block is one of only two cautions of any kind in it, the other being a note that silver nitrate stains the skin. In a formulary containing mercury, dichromate, uranium and concentrated acids, cyanide is one of two substances the publisher thought worth stopping for. That tells you where the line of ordinary practice sat, and that cyanide was just outside it and still on the page.

What it gets wrong, judged from now. The third sentence is a disposal instruction, and it is the one nobody quotes. Run water to flush it out of the sink quickly directs a cyanide solution into the drain — the same drain that in any working darkroom receives the acid stop bath. Under the Environment Agency’s current waste classification, salts of hydrogen cyanide appear in the table of substances that give a waste the hazardous property of releasing an acute toxic gas, at a threshold of 0.2 per cent, and spent fixer solutions have their own absolute hazardous entry, 09 01 04*. The advice is not a lapse of care by the standards of 1949. It is evidence of how much the standard has moved.

⁇Period assertion against modern toxicology: the 1911 antidote table

Cassell’s 1911 Cyclopaedia of Photography reprints a table of “the chief photographic poisons and their antidotes”, credited to J. V. Elsden and taken from the British Journal of Photography Almanac, which a Dr R. J. Hillier is quoted as regarding as reliable and accurate. Its potassium cyanide rows read, in full:

• Taken internally. Three grains fatal. Symptoms — insensibility, slow gasping respiration, dilated pupils, spasmodic closure of the jaws. Antidote — “no certain remedy; cold affusion over the head and neck most efficacious.”
• Applied to wounds and abrasions of the skin. Symptom — a smarting sensation. Remedy — sulphate of iron applied immediately.

The symptom lists are accurate: they are histotoxic anoxia described by careful observers. The responses are not. Cold water poured over the head is not a treatment for cytochrome oxidase inhibition, and the table says as much in its own words by admitting there is no certain remedy. The dermal row is worse in a subtler way: “a smarting sensation” is offered as the characteristic symptom of a route whose aggregated modern classification is fatal in contact with skin. The remedy beside it is not chemically empty — iron and cyanide form the extremely stable complexes the next section of this page is about — but it is a substance applied to the outside of the skin against a poison that is already inside the person. The course evaluates it as historical evidence and not as treatment.

What the course takes from the table is not medical. It is that a competent, cross-checked, widely circulated photographic reference could set out an accurate description of the poisoning and an ineffective response to it, in the same row, and that this was the state of the art the operators were working under. The gap between recognising harm and being able to do anything about it is the honest subject of this part.

What the record supports about actual harm, and what it does not

The story that photographers were routinely killed by their fixing baths circulates widely, and this course has tried to trace it. What can be cited, and what cannot, are different lists.

What is documented. Towler’s 1864 testimony of ulceration and impaired health among printing-room workers, quoted above. Kodak’s own primer stating flatly, in the 1924 printing, that a trace of cyanide swallowed will cause death, and in 1928 that a few grains will. The Pharmacy Act 1868, under which — as Cassell’s 1911 legal note records — “cyanide of potassium and all metallic cyanides and their preparations” sat in Part 1 of the poisons schedule, the strictest part, so that the purchaser had to be known to the seller, sign a register naming the quantity and the purpose, and receive it labelled Poison. A legislature does not put a substance in the first schedule for a hypothetical. Princeton’s environmental health guidance for photography, and the same text as reproduced by other university safety offices, stating that cases of cyanide poisoning have occurred through treating Farmer’s reducer with acid — the ferricyanide side of this page rather than the cyanide side. And the WHO assessment’s account of a silver-reclaiming facility in the United States that had been closed after a worker died of cyanide poisoning, whose thirty-six traceable former employees were surveyed in 1983.

What could not be traced. A named individual poisoned by a photographic cyanide fixing bath, in any source this course holds. The Princeton statement gives no case citation and the course has not found the underlying report. The silver-reclaiming plant is adjacent to photography but the assessment does not say what its silver came from, and the course does not assume it was photographic. No inquest, no case series, no occupational study of photographers.

What follows. The honest summary is that the historical record establishes a substance treated in law as a first-schedule poison, sold under a register, used daily on bare hands, and reported at the time to be harming the people who handled it — and that it does not, in this corpus, produce the roll of named dead the folklore implies. That absence changes nothing about the classification, and it is worth being clear why. As the Level D policy puts it, a level answers one question: can a reader assemble the controls that make this reasonable? For a salt that is fatal by three routes and becomes a gas on contact with the acid two trays away, the answer is no, and it would still be no if the nineteenth century had never lost a single photographer.

The ferricyanide question

Towler’s Silver Sunbeam describes how to lay out a collodion darkroom, and lists what stands together on the shelf beneath the corner of the bench: iron developer, pyrogallic acid, cyanide of potassium, hyposulphite of soda, iodine, nitrate of silver, bichloride of mercury and sulphide of potassium. Every bottle legibly labelled, always in the same place, always corked. Then a housekeeping note, which turns out to be the most useful sentence on this page:

Care must be taken here also not to interchange dishes; for the cyanide of potassium decomposes the iron-salt into what soon becomes Prussian blue by oxidation of the iron, and thus renders it a difficult task to clean the dish afterward.

A dish used for the iron developer and then for the cyanide fixer goes blue. Towler records it as a cleaning problem and gets the chemistry substantially right — iron and cyanide meeting, then oxidation. The intermediate is a hexacyanoferrate, and that identification is the course’s inference from the constants below rather than a statement Towler makes: free cyanide meeting iron(II) has somewhere very stable to go, and once the iron is oxidised the product is the pigment the cyanotype is made of. The nineteenth-century darkroom was demonstrating the boundary this section is about, in a dish, as a nuisance.

That boundary is the part of the page with consequences for a reader’s own darkroom, because potassium ferricyanide is on this course’s shelf. It is in the rehalogenating bleach that starts every sepia toner, in Farmer’s reducer, in the iron-blue toner, and in the classic cyanotype sensitiser — all at Level A or B.

It is not free cyanide, and the difference is a coordination complex. The salt is potassium hexacyanoferrate(III): six cyanide ligands bonded to a single iron(III) centre in an octahedron, the whole thing carrying a 3− charge and moving through solution as one ion.

Fe³⁺ + 6 CN⁻ ⇌ [Fe(CN)₆]³⁻

The complex the bleach actually contains

OpenStax gives that formation a K_f of 2 × 10⁴³, and the iron(II) analogue — ferrocyanide, the product of every bleaching reaction — 1.5 × 10³⁵. Set those beside the dicyanoargentate ion’s 1 × 10²¹ and the scale of the difference is plain, with one caveat that is itself worth teaching: the constants are not directly comparable, because one describes six ligands binding and the other two, so the numbers carry different units and the comparison is an order-of-magnitude impression rather than an equation. What it impresses is real. Cyanide bound to iron(III) in this arrangement is not on the point of leaving.

The World Health Organization’s assessment supplies the field evidence. At former manufactured gas plant sites, iron-complexed cyanides — dominated by the ferrocyanide ion — make up over 97 per cent of the total cyanide in the soil, and the assessment names them as such precisely to distinguish them from “the highly toxic free cyanide forms”. The same document lists potassium ferricyanide’s uses and puts photography second on the list, after blueprints.

And the law drew the same line twice, a century apart. Cassell’s 1911 note on the Pharmacy Act records that it was an open question whether “all metallic cyanides and their preparations” covered the ferri-, ferro- and sulphocyanides, and that “as a matter of fact, however, it is held that these preparations are not to be classed as scheduled poisons”. The Environment Agency’s current waste classification guidance carries the same exclusion in modern language: the entry for substances that make a waste release an acute toxic gas reads hydrogen cyanide, salts of (with the exception of complex cyanides such as ferrocyanides, ferricyanides and mercuric oxycyanide) — and writes out the reaction it is worried about, sodium cyanide meeting a proton, exactly as this page did two sections ago.

Where the boundary breaks

It is a real distinction and it is not an unconditional one. Four sources say four different-sized things about it, and a reader deciding how to behave needs all four rather than whichever one they met first.

Source	What it states	What kind of claim it is
Wall’s dictionary, 1912	Double cyanides such as the sulphocyanides and ferrocyanides, “although apparently not actually poisonous themselves, should be regarded with extreme caution, as simple cyanides may be very readily produced from them under unexpected conditions”	A period practitioner’s principle, stated before the mechanism was available, and correct in shape
International Chemical Safety Card 1132	The solid decomposes on heating, producing toxic gases including hydrogen cyanide; it reacts with acids, which generates a toxic hazard; store it separated from acids and dry. The card also records that the decomposition temperature is unknown in the literature	A Tier 1 hazard statement with no threshold attached, because the threshold has not been established
Princeton’s guidance for photography	Potassium ferricyanide will release hydrogen cyanide if heated, if hot acid is added, or if exposed to strong ultraviolet light, and cases of cyanide poisoning have occurred through treating Farmer’s reducer with acid	An institutional safety statement naming three specific conditions, and an assertion of harm without a case citation
Ware, Cyanomicon, 2020	The heating has to be very strong and the acid very concentrated to liberate much hydrogen cyanide, and neither is done in cyanotype; the traditional process emits no significant, perceptible or harmful amount during normal exposure and processing	A chartered chemist’s conditional qualification, made for one named process rather than for the salt in general

The course states all four and reasons from them rather than choosing. Ware’s qualification is the one people quote when they want the rule relaxed, and it deserves to be read exactly: it is not “there is no risk”, it is these two conditions create the risk and this procedure creates neither. That sentence only holds while the conditions hold, and a darkroom is a place where they can stop holding without anybody deciding anything — a bleach tray that meets the stop bath, a bottle left on a sunlit windowsill, a waste container that receives both.

So the rule the course carries in Parts XX, XXI and here is absolute in the places where a mistake is possible and honest about the places where it is not:

• A ferricyanide bath never meets the stop bath, and never meets concentrated acid. This is the one condition an institutional source ties to poisonings that actually happened. Some published toning baths are themselves mildly acid — the iron-blue toner is, and Part XX says so — so the line the course draws is between a formulated bath a source publishes and a concentrated acid arriving in a tray by accident.
• It is not heated, and it is not stored anywhere warm.
• It is not left in the sun or under a strong ultraviolet source. The cyanotype is the apparent exception and is not one: what the ultraviolet reaches there is a dried coating in the course of an exposure Ware has assessed, not a bottle of solution.
• Its waste is its own stream, kept apart from acid waste, and taken to a licensed route.

The course writes no balanced equation for that decomposition, on this page or on any other, because it has not read one in a source that meets its standard. The incompatibility matrix records the pair and the same gap.

Why thiosulfate won, and stayed

It is tempting to say hypo replaced cyanide because cyanide was dangerous. The record does not support that as the mechanism, and the real answer teaches more.

The material changed first. Cyanide’s decisive advantage was a plate made of silver iodide, and no gelatin emulsion is. A film is silver bromide, or bromide carrying a few mole per cent of iodide in solid solution rather than as a separate phase; a contact paper is silver chloride. The moment the dry plate displaced the wet one, the one job only cyanide could do stopped existing. Look back at the plot: thiosulfate’s combined constant for silver bromide is 24 — an unremarkable number, comfortably greater than one, entirely sufficient. Reaching for 5 × 10⁸ buys nothing.

Weaker binding turned out to be the feature. This is the inversion worth carrying away. A fixer must dissolve the halide and leave the image, and those two demands point in opposite directions. Thiosulfate’s grip is strong enough for silver bromide and too weak to strip metallic silver, which is precisely the selectivity cyanide cannot offer at any concentration. The property that made cyanide the better reagent on paper made it the worse one in a tray.

Not everything went the right way, and the course says so. Cyanide-fixed plates washed faster, because the complex does not fall apart on dilution. Thiosulfate’s does, and Part XII exists because of it — residual thiosulfate, residual silver, hypo clearing agents, wash tests. The modern fixer is genuinely worse on that axis, and the trade was accepted.

And the failure modes are not comparable. A cyanide bath meeting an acid produces a gas with a fifteen-minute limit of 4.5 ppm and a mechanism that stops aerobic respiration. A thiosulfate bath meeting an acid decomposes to sulfur dioxide and colloidal sulfur — an irritant gas that the incompatibility page treats seriously and that many asthmatics are particularly sensitive to. Both are hazards and both have controls. The difference is that the failure of a thiosulfate bath is measured in ruined negatives, a bad smell and a bad afternoon, and the failure of a cyanide bath was measured differently. That asymmetry, not the toxicity of the salt on its own, is what the classification is about.

The bottle in the inherited darkroom

This is the one place where a Level D page has to help a reader do something, and what it has to help them do is stop.

An old darkroom, a school chemistry cupboard, a jobbing photographer’s estate: any of these can contain a container of cyanide, and the period sources describe what it will look like. Archer and Towler both give the manufacture — ferrocyanide of potassium heated to a red heat in an iron vessel — and Hardwich records what came out of it: fused lumps of considerable size, usually contaminated with a large percentage of potassium carbonate, in some cases more than half the weight. It will not look like a laboratory reagent. It may be a grey-white lump in a stoppered jar with a handwritten label.

It may also have changed. NIOSH records that the solid absorbs moisture from the air to form a syrup. Hardwich states that a made-up solution decomposes on keeping, “changing in colour and evolving the odour of prussic acid”, and Archer says the same of the solid left in a moist atmosphere. An old cyanide container may hold gas as well as salt.

⚠What to do, stated as the kind of authority to contact rather than as a procedure

Do not open it. Do not smell it. NIOSH gives the salt a faint almond-like odour, and the international safety card records that the dry solid is odourless — so smell is not a test, and attempting one is an inhalation exposure. Do not add anything to it, and above all not vinegar, descaler, or any other acid, which is the single reaction this page has been about. Do not pour it away, and do not decant it into another container to make it look tidier.

Do not assume from the label. An unlabelled or illegibly labelled bottle in a photographic context could be cyanide, could be ferricyanide, could be hypo. Those three demand entirely different handling and the difference is invisible.

Contact the authority that handles hazardous waste where you live, and describe what you have found rather than moving it. In England and Wales GOV.UK maintains a finder for council household hazardous waste services, and that page is itself scoped to two of the four UK nations — which is exactly why this course never states a disposal route as universal. A business or an institution holding the same material is under a separate duty-of-care regime again, with its own signposts for Northern Ireland, Scotland and Wales.

Where the container is damaged, leaking, or simply cannot be identified, the course has found no published guidance addressed to a householder in that situation, and says so rather than inventing one. What it will say, as its own reasoning from the chemistry above rather than as a sourced instruction, is that an unidentified container that may be evolving a gas is no longer a waste-collection question, and that the people to reach are the ones who attend chemical incidents.

Check your local regulations; they govern. The course names the kind of authority and deliberately gives no steps, because every step it could give would be a procedure for handling a cyanide salt.

The same find with a ferricyanide bottle is a different problem, and a reader who has followed this page can now say why: red crystals of a stable coordination complex, classified far below the simple salt, kept away from acid and heat, and disposed of through the ordinary hazardous-waste route for a photographic chemical. Knowing which of the two you are holding — and knowing that you cannot tell by looking — is the practical skill this page delivers.

Recap

The wet-plate photographer chose cyanide for a reason that is quantitative and correct: silver iodide’s combined fixing constant in thiosulfate is 0.0071 and in cyanide about 1.5 × 10⁵, and the complex it forms does not fall apart in the wash. The same avidity that made it the better fixer made it unable to leave the image alone, and the ligand that lifts silver out of a lattice also binds the haem iron of cytochrome c oxidase and stops a cell spending oxygen it already has. Any acid converts the salt to hydrogen cyanide at once, in a sink or in a stomach, which is why a room with a stop bath in it can never be a room with cyanide in it. The period documents show the hazard was known, the formula was published anyway, and the remedies offered were not remedies. And the hexacyanoferrates this course does use really are a different substance — six cyanides locked to iron, excluded by name from the regulator’s own list of acute-toxic-gas releasers. That exclusion holds under conditions, and heat, concentrated acid and strong ultraviolet are the conditions that break it, which is why the one rule a modern darkroom keeps about them is absolute rather than proportionate.

Check your understanding

Question 1. Using the course's constants — Ksp for silver iodide 1.5 × 10⁻¹⁶, Kf for the bis(thiosulfato)argentate ion 4.7 × 10¹³ and Kf for the dicyanoargentate ion 1 × 10²¹ — what does the arithmetic say about a wet collodion plate?

• Both ligands dissolve it readily; cyanide is merely faster in practice
• Thiosulfate gives a combined constant below one, so it cannot dissolve a separate silver iodide phase, while cyanide gives about 1.5 × 10⁵ and can
• Neither ligand can dissolve silver iodide; the plate was cleared mechanically
• The solubility products do not enter the calculation, because fixing is a kinetic rather than an equilibrium problem

Show the answer and why

Answer: Thiosulfate gives a combined constant below one, so it cannot dissolve a separate silver iodide phase, while cyanide gives about 1.5 × 10⁵ and can

Multiply and the two answers are 7.1 × 10⁻³ and 1.5 × 10⁵ — seven orders of magnitude apart, and on opposite sides of one. A constant below one means the equilibrium sits on the side of the solid. Kodak's 1928 primer reached the same conclusion by observation, saying that wet collodion plates, being made from silver iodide, are not easily fixed in hypo. The gap is nothing but the ratio of the two formation constants, so it is the same 7.3 orders of magnitude at every halide; what changes is only whether that lands above or below the line.

Question 2. Why does a cyanide-fixed plate need less washing than a hypo-fixed one, and why is that the same property that makes the spent bath a permanent cyanide waste?

• Cyanide baths are used more dilute, so less enters the collodion in the first place
• The dicyanoargentate complex is not decomposed by dilution with water, so it leaves in the rinse instead of re-depositing silver — and for the same reason it stays intact in whatever it is poured into
• The collodion layer is non-absorbent, so no fixer enters it at all and washing is symbolic
• Cyanide oxidises to a volatile compound during washing and removes itself

Show the answer and why

Answer: The dicyanoargentate complex is not decomposed by dilution with water, so it leaves in the rinse instead of re-depositing silver — and for the same reason it stays intact in whatever it is poured into

Hardwich states it directly: the soluble double salt, unlike the double sulphocyanide, is not decomposed by dilution with water. A thiosulfate complex is an equilibrium that runs backwards on dilution, which is the whole subject of Part XII. The stability cuts both ways — the WHO assessment lists potassium silver cyanide as a cyanide compound in its own right, so the used bath was not a spent fixer but a cyanide solution that had acquired some silver.

Question 3. A reader finds the 1911 photographic antidote table giving "cold affusion over the head and neck" for swallowed potassium cyanide, and concludes that the period sources are unreliable and should not be cited. What is wrong with that conclusion?

• Nothing; a source that gets a safety fact wrong should be dropped entirely
• Cold affusion is in fact an accepted treatment and the reader has misread the table
• The table's symptom description is accurate observation and its remedy is not, and separating what a period source establishes from what it merely believed is the skill the whole part teaches
• The table is about hydrogen cyanide gas rather than the salt, so the two cases are unrelated

Show the answer and why

Answer: The table's symptom description is accurate observation and its remedy is not, and separating what a period source establishes from what it merely believed is the skill the whole part teaches

Insensibility, slow gasping respiration and dilated pupils is a correct clinical description of histotoxic anoxia by people who had no idea what histotoxic anoxia was. The response beside it is ineffective, and the table admits as much by saying there is no certain remedy. Rule 7 asks for exactly this separation: established chemistry, historical evidence and period belief are three different things, and a source can supply all three in one row.

Question 4. Potassium ferricyanide is used at Level A and B in this course while potassium cyanide is Level D. Which statement best captures the chemical basis of that difference, including its limits?

• Ferricyanide contains no cyanide, so the comparison does not arise
• Six cyanide ligands are bound to iron(III) in a complex whose formation constant is 2 × 10⁴³, and the Environment Agency excludes complex cyanides by name from its acute-toxic-gas list — but heat, concentrated acid and strong ultraviolet can still liberate hydrogen cyanide
• Ferricyanide is safe because it is used in dilute solution, whereas cyanide was used concentrated
• The difference is regulatory convention rather than chemistry, since both release cyanide equally readily

Show the answer and why

Answer: Six cyanide ligands are bound to iron(III) in a complex whose formation constant is 2 × 10⁴³, and the Environment Agency excludes complex cyanides by name from its acute-toxic-gas list — but heat, concentrated acid and strong ultraviolet can still liberate hydrogen cyanide

Both halves matter. The complex is real and the exclusion is written into current British waste guidance exactly as the Pharmacy Act's interpreters wrote it in 1911. But the international safety card records decomposition on heating and reaction with acids, Princeton names the three conditions and records poisonings from acidified Farmer's reducer, and Ware's qualification is conditional rather than absolute — it says these two conditions create the risk and this procedure creates neither. A rule that depends on conditions holding is made absolute where a mistake is possible.

Question 5. The body detoxifies cyanide mainly by rhodanese transferring a sulfane sulfur from thiosulfate to make thiocyanate. What does that fact license a darkroom worker to do?

• Keep a bottle of hypo solution as a first-aid measure wherever ferricyanide is used
• Nothing operational at all: it explains the mechanism and the rate limit, while clinical treatment is intravenous and HSE lists cyanide first aid as separately assessed additional training
• Add thiosulfate to a ferricyanide waste bottle to neutralise any cyanide it might release
• Treat a splash by rinsing with fixer rather than water

Show the answer and why

Answer: Nothing operational at all: it explains the mechanism and the rate limit, while clinical treatment is intravenous and HSE lists cyanide first aid as separately assessed additional training

The chemistry is genuine and worth knowing — the rate-limiting step really is the supply of thiosulfate, and spontaneous detoxification runs at about a microgram per kilogram per minute, which is why the pathway saturates. None of that converts a darkroom chemical into a medicine. Option three is worse than useless: Farmer's reducer already mixes ferricyanide with thiosulfate and the mixture is a reagent, not a treatment, which is why it is made immediately before use and discarded.

Check my answersClear