Cyanide fixing and toning

For about forty years the working darkroom had two fixers, and the faster of them was a poison that kills by swallowing, by breathing and through unbroken skin. Hardwich’s 1864 equipment list for a portrait photographer going out to work reads, without comment or emphasis, “Cyanide of Potassium for cleaning fingers and fixing”. The same book tells you to remove silver stains by rubbing a moistened lump of it on your hands and leaving it there for a little while.

This page exists so that a reader who meets that sentence in a reprint, or a milky-tan collodion negative in a collection, understands what was happening chemically and why nobody does it now. It also exists to take apart its own title, because the phrase cyanide toning is a trap that the nineteenth century’s own nomenclature laid, and the answer to “what was the cyanide toning bath?” is that in almost every case there wasn’t one.

☠Level D: this page is not a procedure

Potassium cyanide and sodium cyanide are classified for historical study only, and the Level D policy governs what this page may contain. It may give the chemistry, the historical practice in outline, the image characteristics and the hazards. It may not give a procedure, and it does not: no mixing sequence, no fixing routine, no immersion time and no temperature appears below as an instruction.

The composition is published for one reason, and it is the same reason IN-1 publishes its: a formula that cannot be read cannot be identified, and identification is the only use a reader has for this entry. Seven further published strengths appear under Variants as a comparison, because the spread between them is itself the evidence for how the bath behaved. The course asks nobody to weigh any of it. No protective equipment makes a cyanide bath reasonable in a room that also contains an acid stop bath, and none is offered.

The fixers this course does publish are the plain hypo bath for plates, the weak thiosulfate bath for printed-out papers and Reilly’s alkaline bath for albumen.

Ingredient	Quantity	Form the source specifies
Potassium cyanide	2 %	Printed by Reilly as "a 2 % solution of potassium cyanide". Hardwich records that the commercial salt of the period was contaminated with potassium carbonate, in some cases to more than half its weight, so a period 2 per cent is a nominal figure and not an assay
Water	to make 1000 mL	Reilly states a strength and no volume. A litre is written here because the schema requires a water statement and a percentage needs something to be a percentage of; the formula is the 2 per cent. No temperature is published, by Reilly or by any of the period sources.

Purpose

To dissolve unexposed silver halide faster and more completely than thiosulfate does. That is the whole of it, and both halves of the sentence matter.

Faster is why the wet-plate trade adopted it. Eder’s history dates the publication to 1853 and credits Gaudin, and gives the reason in the same breath: it acts more rapidly and contributes to the clearing up of the negative. Kodak’s 1928 primer puts the chemical case behind that preference — potassium cyanide was employed for fixing wet collodion plates, which being made from silver iodide are not easily fixed in hypo. Towler puts the same fact the other way round — silver chloride and silver bromide are soluble to a greater extent than silver iodide in hyposulphite of soda — and Hardwich puts a number on it: one part by weight of silver iodide requires about twenty-four parts by weight of hyposulphite of soda in a cold solution. A wet plate is mostly silver iodide.

More completely is the property Davanne and Girard were testing in 1859, and it is why this entry’s quantity is theirs rather than a plate-maker’s. Their question was not how to fix an albumen print — that was long settled — but whether a fixed print was actually free of silver. Reilly’s account is that a 2 per cent cyanide solution removed all traces of silver from albumen prints where strong hypo solutions did not, and that the residue hypo leaves behind is the origin of the highlight yellowing that disfigures most surviving albumen prints. The bath answered the question and was useless as an answer to the problem, for reasons the next sections give.

There is a third purpose the sources record and it belongs here because it is the one the title of this page points at: a change in the colour of the finished image. Archer’s remark in 1854 is that pictures fixed with cyanide of potassium have a slightly whiter tone. Towler is more definite — the bath produces in the ambrotype and the melainotype a whiteness in the silver film which cannot be effected with hyposulphite, and for that reason it is regarded by many photographers as the fixing agent peculiarly adapted for collodion positives by reflected light. On a positive viewed against black varnish, whiter highlights are the picture. That is a tonal effect of fixing, not toning; the difference is worked out under Variants.

Recommended uses

These are the uses the sources record, set down as history. The course recommends none of them.

Wet collodion negatives, where Hardwich allows it for a bromo-iodised plate developed with iron and Towler prints a bath for it, both with warnings attached. Eder records that it was still in use in the twentieth century, especially with half-tone negatives, and the British Journal Photographic Almanac was still printing a fixing solution for it in 1906.

Collodion positives — ambrotypes and tintypes — for the whiter image. This is the use every period source agrees on, and the conservation record agrees too: the AIC’s ambrotype entry describes the developed plate as fixed in hypo or potassium cyanide, because both are found in surviving objects.

Field work, where water was short. Vogel’s rule is practical and revealing: in the atelier, where an abundance of water is constantly at hand, hyposulphite; where the supply of water is limited, when travelling, cyanide. Hardwich gives the chemistry behind that — the double cyanide, unlike the double thiosulfate, is not decomposed by dilution with water, so it washes out instead of breaking down in the film. A plate fixed in cyanide needed about a minute’s washing where a hypo-fixed one needed five, by Cassell’s ferrotype figures.

Once, on albumen paper, as an analytical reagent. Davanne and Girard’s 2 per cent bath was an experiment about residual silver, not a working process, and its result was a diagnosis rather than a method.

Nothing whatever in a modern darkroom. Princeton’s environmental health guidance lists cyanide among the older intensifiers and reducers now discarded, and this course teaches no process that uses it.

When another formula is preferable

Always, without a case to argue. The alternatives are not compromises.

• The plain hypo fixing bath does the same job on the same materials. Reilly’s judgement, after listing every substance with fixing or stabilising action — ammonia, potassium cyanide, strong chloride solutions, thiocyanates, thiourea, sodium sulfite, sodium and ammonium thiosulfate — is that of all of them sodium thiosulfate has the fewest drawbacks, and that it has been in almost exclusive use since the earliest days of photography. That is not squeamishness; it is a list of eight candidates with one winner.
• The weak thiosulfate bath for printed-out papers and Reilly’s alkaline fixing bath are the two the course teaches for salted paper and albumen. Every period source that considered cyanide for prints rejected it: Hardwich says it is not adapted for fixing positive proofs upon chloride of silver, Vogel that it is not admissible as it affects the pictures very much, Abney that hyposulphite is used because cyanide attacks the organic oxide formed by light.
• TF-2 or Kodak F-24 where a modern plain or alkaline bath is wanted for film and plates.
• A thiocyanate, if what you actually want is the historical alternative rather than a modern one. Towler prints Meynier’s ammonium sulphocyanide bath at 1 drachm to 12 ounces alongside his cyanide formula, and records the claim that it is as powerful without the poisonous properties — then adds, drily, that Meynier “must have made a mistake as to this latter property”. Towler was wrong to be so sceptical: ammonium thiocyanate and potassium thiocyanate are not in the same hazard class as the cyanides at all, which is exactly the confusion this page has to break.
• For toning, every gold and platinum bath in this formulary. The gold-thiocyanate toner, the gold-borax toner and the platinum toner are what nineteenth-century printing-out actually used, and none of them contains a cyanide.
• Against residual silver specifically, the answer is not a stronger solvent. It is fresh thiosulfate in two baths and a real wash — Reilly’s own remedy — plus the hypo and silver tests that show whether it worked.

Mixing

This section gives no mixing procedure, and the policy is not the only reason there is nothing much to give. The bath is one salt in water. There is no order of addition, no second solution, no temperature and no ripening. What is worth setting out is why a formula that simple was nonetheless impossible to state exactly, because that is the most interesting thing about it.

The commercial salt was not the salt. Hardwich is explicit: potassium cyanide as sold occurred in fused lumps contaminated with a large percentage of potassium carbonate, amounting in some cases to more than half its weight. His conclusion follows immediately — “the percentage of carbonate of potash in commercial cyanide of potassium is so variable that no exact directions can be given for the formula”. A quantity in this entry, or in any period printing of it, is therefore a nominal weight of an assay that nobody performed. Two workers following the same figure could be working at strengths a factor of two apart, and the carbonate they were unknowingly adding raised the pH of the bath as well.

So the strength was set by a behaviour, not by a number. Hardwich’s instruction is to use it rather dilute, “of such a strength that the plate is cleared gradually in from half a minute to a minute”. Read that as a piece of experimental design and it is rather good: the worker titrates an unknown reagent against the only assay available, which is the plate itself, and the criterion is that clearing should be slow enough to watch. Archer had reached the same place ten years earlier from the other direction, printing 4 grains to the ounce as the ceiling — “a cyanide of potassium fixing solution should not be of greater strength than the following” — and warning that it must not be left on the plate one moment longer than is necessary.

Which is why the published strengths spread over a factor of four, from Archer’s 0.9 per cent to Vogel’s 4 per cent, with no author disagreeing with any other about the chemistry. They are not rival formulas. They are the same instruction — dilute enough to watch — expressed by seven people working with seven different tins of an impure salt. The table is under Variants.

The odour was a mixing problem in its own right. Hardwich records that many workers used a vertical bath rather than a dish, in order to escape the pungent odour evolved by the salt — and then warns that a vertical bath forces a change in the preceding step, because plates must be washed completely free of iron before they go in or the iron decomposes the cyanide and throws a blue deposit through the whole tank. A ventilation control that creates a contamination hazard is a pattern worth recognising; it recurs whenever a solution is moved from a dish into a tank.

Behaviour

Recorded from the sources so the formula can be understood, not so that it can be run.

It does not stop at the halide. This is the single fact that governs everything else. Hardwich: unless the solution is tolerably dilute, it attacks the image, converting it superficially into cyanide of silver and then dissolving it as the double cyanide of potassium and silver. Archer, in 1854: it possesses great power over the iodide of silver, but it has one great defect — unless used with great caution it would dissolve the deposited image as well. His comparison with hypo is exact and is worth keeping: the difference between the two in this respect is only one of degree. Every fixer dissolves the picture; this one is fast enough that you can watch it happen.

The half-tones go first. Not the shadows, not the highlights — the middle. Hardwich says formula No. 2 cannot be used safely on an iodised film developed with pyrogallol alone because of its great tendency to weaken the half-tones, and that even in very dilute proportions it will weaken them if left on the plate. Vogel says the same: it dissolves the silver of the picture and destroys the delicate half-tones when it is not quickly removed by washing. The mechanism section explains why the finest silver goes first.

Light makes it worse. Hardwich records that the action of light favours the conversion of the image into cyanide of silver, and that the loss is especially bad when a plate is fixed outdoors with the sun on it. A fixer with a photochemical sensitivity of its own is unlike anything in a modern darkroom.

No time is published anywhere in the corpus, and the field is left empty on purpose. Hardwich’s half-a-minute-to-a-minute is a criterion for choosing the strength, not a time to hold; Archer gives an instruction to stop at once and no duration; Reilly reports Davanne and Girard’s strength and no time at all. A Level D page would not print a time even if one existed, and none does.

It was reused until it stopped working. Hardwich for positives: the cyanide is used over and over again until exhausted. For negatives: it need not be thrown away till it ceases to dissolve the iodide readily. Both are recorded as published facts about the bath rather than as advice.

The keeping is contested, and by two first-rank sources.

⁇Weeks, or days? Hardwich and Vogel disagree about the same solution

Hardwich, 1864: a solution of cyanide of potassium decomposes slowly on keeping, but it will usually retain its solvent power for several weeks.

Vogel, 1875: the solution of cyanide of potassium decomposes rapidly, and is transformed into potassium formate — set against hypo, which in the same paragraph “will keep for several days”.

Both are stated as fact by authors of standing, eleven years apart, about the same bath. They are not straightforwardly reconcilable, and the course does not pick a winner. Two observations can be made without picking one. Vogel names a product, and the hydrolysis of a cyanide to a formate and ammonia is real chemistry rather than a guess, so his is a mechanism and not just an impression. Hardwich’s “solvent power” and Vogel’s “decomposes” may not be the same test: a bath can lose a large fraction of its cyanide and still clear a plate, because it was in a large excess to begin with. The one thing both agree on is that the solution is a decomposing one, which is more than can be said of a hypo bath and is a reason the period made it fresh and small.

The course’s own position is the one that does not depend on the answer: neither figure is a measurement, both come from an era with no assay of the starting material, and no reader of this course will ever have a bottle to keep.

On some materials it wrecked the binder rather than the image. Towler’s tannin dry plates are fixed in hyposulphite and not cyanide “because it is apt to loosen the film”. Hardwich says hypo is better than cyanide in all cases where albumen is employed. The bath’s alkalinity — remember the carbonate — is the obvious suspect, but no source in this corpus says so, and the course does not put words in their mouths.

Image characteristics

Lighter, and milkier. The Image Permanence Institute’s booklet on negatives sets out the difference as a table of process variations, which is exactly how a conservator uses it: fixing in hypo gives a darker, brownish grey image colour; fixing in cyanide gives a lighter, milky-tan one. That is a signature you can look for, and the reason it exists is on the same page — both fixers stayed in use for negatives throughout the collodion era, so the choice was a workshop preference and is legible in the object.

On a positive, whiter is better, which is why the trade paid the price. Towler’s account is that cyanide is not only a solvent of the silver salts but also a reducing agent, and produces in the ambrotype and the melainotype a whiteness in the silver film that hyposulphite cannot give. Archer’s 1854 note is milder — a slightly whiter tone — and eleven years earlier. On a collodion positive the image is the highlight: the plate is backed with black, the shadows are the clear glass, and every bit of brightness in the picture is scattered light from finely divided silver. A treatment that makes that silver whiter makes the whole picture better, and there was no other way to get it.

On a print it destroys the picture. Davanne and Girard’s two drawbacks, as Reilly gives them, are that cyanide attacked and severely bleached the silver image and that it was highly poisonous, and the first is enough on its own. Reilly’s account of why printing-out images are so vulnerable is that the image is composed of very small, very highly dispersed particles of metallic silver whose surface area is very large relative to their mass, so that a large part of the total silver is on the surface and readily accessible to a destructive chemical agent. There is nothing for a fast solvent to bite on except the picture.

The half-tone loss is the failure mode to recognise. A plate over-fixed in cyanide is not uniformly thin — it has lost its middle, so it reads as harsh: clear shadows, surviving highlights, and a gap where the modelling was. That is the opposite signature to an exhausted hypo bath, which leaves a milky, unfixed base and takes nothing off the image at all.

Nothing is added and nothing is stained. Unlike the mercury and chromium treatments in this formulary, the cyanide bath deposits nothing; it only removes. The image after fixing is still silver, and there is no cyanide signature left behind in a properly washed plate for an analytical technique to find.

The mechanism

Fixing is complex formation, and cyanide forms one of the most stable complexes silver has. The potassium cyanide page carries the ligand chemistry and this page does not restate it; what belongs here is what the formula does that thiosulfate does not.

⚗Two ions, one complex, and why it does not fall apart in the wash

Two cyanide ions lift a silver ion out of the lattice and hold it as the linear dicyanoargentate ion:

AgI + 2 CN⁻ → [Ag(CN)₂]⁻ + I⁻

Fixing a wet plate: the silver iodide of the collodion layer goes into solution as dicyanoargentate

Hardwich’s two statements about this complex are the ones that matter, and they pull in opposite directions. The first is that solution of cyanide of potassium is a most energetic agent in dissolving the insoluble silver salts, far more so than the hyposulphite of soda. The second is that the resulting double salt, unlike the double sulphocyanide, is not decomposed by dilution with water.

Put together they explain both the appeal and the danger. A complex that forms readily and does not dissociate on dilution is a fixer that clears quickly and then washes clean, which is why a cyanide-fixed plate needed a minute’s washing where a hypo-fixed one needed five, and why the bath was the travelling photographer’s. But stability towards dilution is stability, full stop — and the same complex will happily form from silver that is not in the lattice at all.

Why the image goes, and why the half-tones go first. Hardwich describes the attack in two stages: the image is converted superficially into cyanide of silver, and that is then dissolved as a double cyanide of potassium and silver. The second stage is the same complexation as above, one step along:

AgCN + CN⁻ → [Ag(CN)₂]⁻

Hardwich's second stage, in symbols: silver cyanide dissolving in excess cyanide as the double salt

Silver cyanide has no page in this course’s chemical encyclopaedia and is not in its register of planned entries, so the reader cannot follow that intermediate to a page of its own. The same is true of potassium dicyanoaurate, which appears under Variants. Neither gap is an oversight the page can repair — a chemical entry is a research task of its own — and both are stated here rather than papered over with a link that goes nowhere.

The first stage is the one this course will not write as an equation. Converting metallic silver into silver cyanide is an oxidation, and it needs an oxidant — dissolved air is the obvious candidate, and cyanide leaching of noble metals in industry works exactly that way, which Ware notes in passing. But no source in this corpus states that reaction for this bath, and a plausible balanced equation for a mechanism nobody in the citations has written down is precisely what Rule 1 forbids. So the page stops where Hardwich stops: a surface conversion, then dissolution.

What the two-stage picture does explain, without needing the equation, is the half-tone observation. A surface reaction removes silver in proportion to surface area, not to mass. The finely divided silver of a half-tone has an enormous surface for its little mass; the dense, coarser silver of a shadow has much less. So the same bath for the same time takes a large fraction of the half-tone and a small fraction of the shadow, and the picture loses its middle. The same reasoning explains why a printed-out image — colloidal silver, smaller particles again — is bleached outright, and why a developed-out plate survives at all.

⚗The albumen result: what Davanne and Girard actually found

Their experiment is often reported as “cyanide fixes better”, which is not what it shows.

Reilly’s account of the underlying chemistry is that silver bonds during sensitisation to sulfur-containing side groups on the albumen protein, and is held so tightly there that treatment in hypo is not sufficient to remove it. That silver is not a halide crystal waiting to be complexed; it is a silver-protein compound, chemically bound in the binder, and it sits in the highlights as much as anywhere else.

So the 1859 result is a statement about ligand strength: cyanide’s affinity for silver is high enough to strip silver off a protein side chain, and thiosulfate’s is not. That is the same property, read from the other end, that lets it strip silver out of an image. One bath cannot be selective in one direction only. The residue it removed is the origin of the highlight yellowing that Reilly calls so prevalent in albumen prints that it often serves as an important clue in their identification, and cyanide would have prevented it — at the cost of the picture.

Function of every ingredient

Potassium cyanide, 2 g in 100 mL of bath — 20 g to the litre in the strength Reilly reports. It is the entire formula and it does four things, of which only the first is wanted.

What it is. KCN, the potassium salt of hydrocyanic acid, made in the period by heating potassium ferrocyanide to a red heat in an iron vessel — Archer and Towler both describe the process, and Hardwich adds what came out of it, fused lumps carrying up to half their weight of potassium carbonate.

Why it is there. It is the ligand. Two cyanide ions per silver ion, forming a complex stable enough to pull silver out of a crystal lattice and out of a protein, and stable enough not to fall apart in the rinse.

What it does photographically. Clears the plate — fast. Leaves the collodion positive whiter than hypo can, which was the trade’s reason for using it. Leaves a lighter, milky-tan negative that a conservator can still identify a century and a half later.

What happens with more. The image goes, starting with the half-tones, and the loss accelerates in light. Every period source that gives a strength gives it as a ceiling rather than as a target: Archer’s “should not be of greater strength than”, Hardwich’s “best to use it rather dilute”, Towler’s “it has to be used in a dilute condition, and to be watched very closely”. No source anywhere in this corpus recommends a stronger cyanide bath for anything.

What happens with less. The plate clears more slowly and, at the strengths in question, that is mostly an advantage — Hardwich’s whole rule is that clearing should be gradual enough to watch. There is no lower limit in the sources, because a slow cyanide bath is still faster than hypo on silver iodide.

What it interacts with. Acids, which is the fact that ends the argument — see Incompatibilities. Iron developer residues, which throw a blue deposit. Light. And silver nitrate, with which it makes the silver-cyanide solution of Monckhoven’s intensifier, the other formula in this course that contains it.

Sodium cyanide instead. Hardwich states that sodium cyanide will answer the purpose equally well, and Kodak’s IN-1 offers either salt at the same weight. No source in this corpus gives a quantity for sodium cyanide in a fixing bath, so none is recorded in the object above; its own notified classification is the same H300, H310 and H330 as the potassium salt, and nothing about the substitution changes anything on this page.

Water. Hardwich writes “common water” and Towler “rain-water”, which is the period’s way of saying that nothing in the bath is sensitive to hardness — there is no alkali to precipitate and no metal ion to complex, unlike the hypo baths of the same books. No temperature is published in any printing. The volume in the object above is the schema’s, not Reilly’s; the formula is the percentage.

Interactions

With acid: the interaction that closes the subject. Any acid liberates hydrogen cyanide from a cyanide salt or its solution, at once and at room temperature:

KCN + H⁺ → HCN + K⁺

The reaction that makes a cyanide bath incompatible with a darkroom: the hazard statement of every safety data sheet, written out

CAMEO’s datasheet and the PubChem aggregate both carry this as a defining property of the substance, and the Level D policy singles it out as the decisive fact for the cyanides specifically: a darkroom is a room with an acid stop bath in it. It is also, historically, a room with acetic acid in the developer — every collodion iron developer in Towler, Hardwich and Cassell carries several per cent of it — and a room where hydrochloric acid was used to clean plates and remove mercury salts. The nineteenth-century darkroom put a cyanide bath a splash away from three different acids.

With iron developer: Prussian blue. Hardwich’s warning is practical and its chemistry is worth naming. Wash the plate free of iron before putting on the cyanide, or a blue deposit will often be formed. Cyanide and iron(II) make hexacyanoferrate, which with iron(III) makes Prussian blue — the same pigment the cyanotype is made of, arriving here as a contaminant. In a dish it spoils one plate; in the vertical bath the period used for ventilation, it spoils the tank.

With thiosulfate. Both fixers were on the same bench, and both books warn about the crossover from the other direction: Towler’s blackening bath must be followed by hypo and not by cyanide, because cyanide reduces the silver film to white again. Where the two are mixed the course has no source for what happens and does not speculate; it is enough that one of the two liberates a fatal gas on acidification and the other does not.

With light. Hardwich states that light favours the conversion of the image to silver cyanide, so the bath’s attack on the picture is photochemically accelerated. Fixing in the open, with the sun on the plate, was the recorded way to lose a negative.

With the binder. Collodion tolerates it and gelatin-bound tannin plates do not — Towler’s plates loosen from the glass — and albumen is worse still, since the bath attacks the silver the protein is holding. The pattern is that the more organic silver-binding there is in the layer, the less a cyanide bath can be trusted.

With gold. Cyanide and gold(I) make the dicyanoaurate ion, which is a real and important compound — it is the electrolyte of electro-gilding, as Ware records and as Towler’s own recipe for a gilding bath confirms — and it is the reason a reader might expect a cyanide gold toner to exist. It did not. See below.

Variants

The eight published strengths, and what their spread means

Every figure below is the metric equivalent of what the source printed. The imperial originals are given so the conversion can be checked; grains and drachms are apothecaries’ weight, in which Cassell states that formulae are made up, so a drachm is 60 grains and a grain is 64.8 mg. British sources are read against the imperial fluid ounce of 28.41 mL and Towler, published in New York, against the US fluid ounce of 29.57 mL.

Source	Year	For	As printed	Metric	Per cent w/v
Archer, The Collodion Process on Glass	1854	collodion, as a ceiling	4 gr to 1 oz	9.1 g/L	0.91
Davanne and Girard, via Reilly	1859	albumen prints, in an experiment	“a 2 % solution”	20 g/L	2.0
Hardwich, Manual of Photographic Chemistry	1864	glass positives	10 gr to 1 oz	22.8 g/L	2.3
Hardwich, same book	1864	collodion negatives	15 gr to 1 oz	34.2 g/L	3.4
Towler, The Silver Sunbeam	1864	collodion positives	1 drachm to 4 oz	32.9 g/L	3.3
Towler, same book	1864	collodion negatives	1 drachm to 5 oz	26.3 g/L	2.6
Vogel, Handbook of the Practice and Art	1875	collodion	1 part in 25 of water	about 40 g/L	4.0
British Journal Photographic Almanac	1906	collodion, half-tone	25–30 g to 1,000 cc	25–30 g/L	2.5–3.0

Cassell’s 1911 ferrotype instruction is deliberately left out of the table: it says to keep the salt as a saturated solution and dilute it with double its volume of water for use, and a saturated solution of potassium cyanide would be extraordinarily strong — the ILO-WHO safety card summarised on the potassium cyanide page gives its water solubility as 716 g/L at 25 °C — which after a threefold dilution would still put the working bath an order of magnitude above everything above it. The course does not believe that is what the workshop did, cannot resolve whether “saturated” is being used loosely, and therefore records the instruction without computing a strength from it. A number that can be calculated is not the same as a number that can be trusted.

Three things are visible in the table and none of them is a disagreement about chemistry. The spread — Vogel’s 4.0 per cent against Archer’s 0.91, a factor of four and a half — is what you get when six authors specify a reagent whose purity nobody could measure. The two authors who print two baths each, Hardwich and Towler, disagree with each other about which way round they go: Hardwich’s negative bath is stronger than his positive one, Towler’s is weaker. And the twentieth-century printing sits in the middle of the nineteenth-century range, which is what you would expect once the salt itself became reliable.

Sodium cyanide

Hardwich records that sodium cyanide answers equally well, and Kodak’s IN-1 offers either at the same weight. There is no separate entry and no separate quantity, because no source in this corpus prints one for a fixing bath.

The toning question, which is mostly a nomenclature question

The title of this page promises toning and the honest answer is that the corpus does not contain a cyanide toning bath for printing-out papers. Three different things get called one, and telling them apart is worth more to a reader than a formula would be.

“Sulphocyanide” is thiocyanate, and it is a different substance. This is the trap. Cassell’s entry for potassium sulphocyanide gives the synonyms potassium thiocyanate, sulphocyanate and rhodanide, the formula KCNS, and the statement that it is chiefly used in the sulphocyanide toning bath. Reilly dates that class of toner to 1867 and explains its chemistry — thiocyanate reduces gold(III) to gold(I), so the substitution of gold for silver is more complete and the image runs colder — and notes in the same paragraph that thiocyanate was formerly called sulphocyanide in the old style nomenclature. Kodak’s 1928 primer files the sulphocyanides as a derivative of the cyanides, formed with sulfur, and mentions in the same sentence that ammonium sulphocyanide is used in gold toning baths. So a reader meeting “sulphocyanide toning bath” in a nineteenth-century manual is reading about the ancestor of the course’s own gold-thiocyanate toner — not about cyanide. The distinction is not pedantic: one anion is fatal by three routes and the other is not, and the two names differ by four letters.

Hunt’s gold-cyanide compound was a sensitiser, not a toner. Ware’s history identifies the “protocyanide of potassium and gold” of Hunt’s 1844 Researches on Light as potassium dicyanoaurate(I), and records that it was made at the time by Himly’s method from fulminating gold — “a dangerously sensitive explosive” — and potassium cyanide. Ware’s verdict on the preparation is that it “offers a challenging combination of hazards, not recommended for the amateur chemist today”, and his verdict on the result is that since the proportions of gold and silver in the image were never reported it cannot be considered a true gold print. It was a paper you coated and printed on, in other words, not a bath you toned in. The same compound’s real nineteenth-century trade was electro-gilding: Ware describes gold potassium cyanide as the usual electroplating electrolyte because it lays down a smooth coherent film where plain gold chloride gives a deposit that is too crystalline, and Towler’s own book prints a gilding electrolyte made by dissolving gold terchloride to saturation in saturated potassium cyanide. Gold and cyanide belong together in the plating shop, and did not come into the toning tray.

Towler’s silver-cyanide blackening bath is the one genuine cyanide treatment for image colour, and it is not for prints. Working on collodion positives to be viewed by transmitted light, he notes that iron development leaves the shadows greyish or silver-white, agreeable by reflected light but too grey against the lamp, and states the object plainly: “to communicate to them a rich black hue”. His sequence is a saturated mercury(II) chloride solution until the film is black, a wash, and then a saturated solution of silver cyanide in potassium cyanide — his Formula No. 1, made from 100 grains of the salt in 2 ounces of rain-water with a 50-grain-per-ounce silver nitrate solution run in until the precipitate just dissolves — after which “the image assumes an intense black hue”. Formula No. 2 substitutes copper nitrate for the silver. Both baths, he says, can be used over and over again, and the plate is refixed afterwards in hypo and not in cyanide, because cyanide would reduce the silver to a white film again. He closes by noting that the same treatment may be used as an intensifier — which is exactly what it became, since this is Monckhoven’s solution, the blackener of Kodak IN-1, where the course records it in full.

That is a cyanide toner: a bath containing a cyanide, applied to a finished image, for the sole purpose of changing its colour. It is also a mercury process with a silver-cyanide second bath, which is to say it is doubly beyond anything this course would put in a reader’s hands, and it never had anything to do with printing-out papers.

One more red herring, for completeness. Cassell’s Eburneum entry says of a collodion transparency “fix with cyanide, well wash, and tone with gold”. Fixing with cyanide and toning with gold are two separate baths in that sentence, and the second is an ordinary gold toner.

Safety

Level D, and the classification is not a judgement call. The safety classification rubric names this process in its own words — Level D covers “mercury vapour development, cyanide fixing or toning, uranium and cadmium compounds, and any process whose historical practice is documented mainly through the harm it caused” — so this page is the entry the rubric was written against. The Level D policy owns what may go on it. The figures below are quoted as a measure of how much control the substance is judged to need, never as thresholds to work to.

The notified classification. PubChem aggregates 575 reports across 30 notifications to the ECHA C&L Inventory for potassium cyanide: signal word Danger, with H300 fatal if swallowed, H310 fatal in contact with skin and H330 fatal if inhaled each appearing in 99.8 per cent of reports, alongside H372 for organ damage through repeated exposure, H370, H318, H315, H290 and the aquatic statements H400 and H410. No notifier reports that it meets no criteria. Sodium cyanide’s record is the same three statements. Fatal by all three routes, with no dissent across 575 reports, is a profile nothing else in this formulary approaches.

The occupational limits. HSE’s EH40 gives cyanides other than hydrogen cyanide, cyanogen and cyanogen chloride a workplace exposure limit of 5 mg/m³ as CN over eight hours with the Sk skin notation, and lists potassium cyanide by name with the same notation; hydrogen cyanide is at 0.9 ppm over eight hours and 4.5 ppm over fifteen minutes, also with Sk. NIOSH sets a ceiling of 5 mg/m³ as CN over ten minutes and an immediately-dangerous-to-life level of 25 mg/m³, requires that skin and eye contact be prevented, that contaminated clothing be removed when wet and changed daily, that eyewash and quick-drench facilities be provided, and that respiratory protection be air-supplied. Its first-aid column reads, in full: eye, irrigate immediately; skin, soap wash immediately; breathing, respiratory support; swallow, medical attention immediately. Kodak’s own 1928 primer, which was not a safety document, put it more briefly: a few grains of cyanide swallowed will cause death.

The acid incompatibility is the reason no control list is offered. Everything above describes a substance that can be handled under laboratory control. What removes the process from a domestic darkroom is that the same room contains an acid stop bath, an acetic-acid developer or a cleaning acid, and the consequence of one splash in the wrong direction is hydrogen cyanide in the air of the room. There is no glove, no apron and no fan that makes that acceptable, and the course does not pretend otherwise.

🕰What the period actually did, which is the point of teaching this at all

Hardwich’s list of apparatus and chemicals for a photographer going out to do portraits includes the line “Cyanide of Potassium for cleaning fingers and fixing”, between the sulphate of iron and the developing glasses. His appendix recommends removing silver stains from the hands by rubbing them with a moistened lump of the salt and leaving it on the skin for a little time. Both are printed without alarm, in a book that elsewhere says the substance is highly poisonous and must be used with caution.

Towler, in the same year, saw further. He describes a glass plate-holder he made so as not to hold the negative in his hand during intensification, “in order to avoid all the poisonous effects that might arise from contact of such substances with the broken skin or wounds”; he notes that stains can be removed with cyanide but that doing so after every negative “might entail upon him incurable ulcers”; and he records that “the health of operators is much impaired, and especially in those large printing establishments, where a number of females are employed” turning prints by hand in the toning and fixing baths. That is an occupational-health observation, made in 1864, from inside the trade.

The distance between those two accounts is not a difference of knowledge — both men knew the salt was lethal — but a difference in what they thought skin contact meant. Reading them together is the best short lesson this course can offer on why a hazard assessment is made from routes of exposure rather than from a general sense of danger.

Nothing here is a control list. No PPE, ventilation or handling regime is given, because the course gives no procedure and there is nothing to protect a reader from except the page.

Storage

Not stocked. No reader of this course has a reason to hold a cyanide, and the storage section of a Level D entry exists to say so and to explain what the period was dealing with.

The solid was hygroscopic and slowly self-destructing. Archer records that in contact with a moist atmosphere it is slightly decomposed and smells of prussic acid — which is to say the lump in the bottle was quietly generating hydrogen cyanide in its own headspace. NIOSH records that it absorbs moisture from the air to form a syrup; Hardwich says the same in period language. The smell was the period’s warning and it is not a reliable one: the ILO-WHO safety card, summarised on the potassium cyanide page, records the dry salt as odourless, so the prussic-acid smell Archer describes is a symptom of decomposition that has already happened rather than a detector of the hazard. Odour is recorded here as history, never as a control.

The solution kept badly, and the two accounts of how badly are set out under Behaviour. Whichever is right, the bath was made small and often, and the pungency of the salt drove the trade towards deep tanks with lids rather than open dishes — Hardwich’s vertical bath.

Anyone who finds some. Old cyanide turns up in inherited darkroom stock, in school and university chemistry cupboards and in the residue of commercial studios. It is not a darkroom disposal question and should not be treated as one: it is a hazardous-waste question for the local authority or a licensed contractor, and the container should not be opened to check. The unlabelled container SOP sets out the general rule.

Incompatibilities

The incompatibility matrix carries the cyanide pairs with the rest; the ones that belong to this formula specifically are these.

• Any acid. Hydrogen cyanide, immediately. This is the pair that decides the classification, and it covers the stop bath, the acetic acid of a collodion developer, hydrochloric acid used for cleaning, and the carbon dioxide dissolved in ordinary water to a much smaller degree. NIOSH lists acids, strong oxidisers, alkaloids, chloral hydrate and iodine among its incompatibilities.
• Iron salts. A blue deposit through the bath, per Hardwich, and a plate that cannot be cleaned. The practical consequence in the period was a mandatory intermediate wash.
• Iodine. Hardwich records that adding a little iodine to an aqueous cyanide solution greatly increases its solvent power on metallic silver — the “iodo-cyanide” of the period, printed by Wall as a reducer and by Hardwich as an energetic silver-stain remover. That is a deliberate combination rather than an accident, and it is recorded here because it shows how narrow the margin was: the same bath, plus a trace of a common darkroom reagent, becomes an image destroyer.
• Silver nitrate. Precipitates silver cyanide, which redissolves in excess cyanide. Monckhoven’s intensifier is built on exactly that; a silver bath contaminating a cyanide bath is the same reaction happening where nobody wants it.
• Strong oxidisers, per NIOSH and CAMEO.
• Everything organic that binds silver — albumen above all, and gelatin. Not an incompatibility in the safety sense, but the same warning applies: the bath will take silver out of the binder as well as out of the crystal.

Waste

No procedure is given here, so no bath is made, and the only cyanide waste this page can generate is an object somebody already owns. That is worth stating rather than assuming.

What the period did, and why it is instructive. Hardwich’s appendix treats a spent cyanide solution for silver recovery by diluting it largely with water and adding sulphide of potassium — a substance the course’s encyclopaedia does not cover, the nearest entry being sodium sulfide — until no further precipitation of black silver sulphide ensues. Abney’s residues chapter puts the old hypo baths and the cyanide fixing solutions in the same tub and treats them the same way. Both are silver-recovery procedures, not detoxification procedures: they take the silver out and leave the cyanide in the liquid. The period was recovering the valuable thing and pouring the dangerous thing away, which is the exact inversion of a modern waste hierarchy and is the reason these accounts are printed here as history and not as method.

The one modern rule that follows from all of it: never acidify. Any treatment of a cyanide-bearing liquid that lowers its pH releases hydrogen cyanide, and that includes the acid additions that are routine in silver recovery from thiosulfate baths.

Waste classification. England’s WM3 guidance lists photographic fixer solutions at 09 01 04*, an absolute hazardous entry, and a cyanide-bearing solution would in any case carry its own hazardous properties from the cyanide. A householder’s route is the local authority’s hazardous-waste service rather than the drain. Local regulation governs and this course cannot tell you what it says where you are; the disposal ruling and the general chemical waste SOP set out how the course handles that, and ordinary silver-bearing darkroom waste is covered by the silver-bearing waste SOP, which does not apply to anything on this page.

Troubleshooting

The faults the sources record, given as identification evidence rather than as fixes. Nothing here is a corrective action, because nothing here is a procedure.

What was seen	What the sources say caused it
Half-tones weak, the image harsh, shadows and highlights surviving	The bath too strong, or left on too long. Hardwich, on pyrogallol-developed iodised films, and Vogel on the delicate half-tones
The whole image whitening and then dissolving	Metallic silver converted superficially to silver cyanide and dissolved as the double salt. Hardwich
Loss much worse than usual, on a plate fixed outdoors	Light favours the conversion of the image to silver cyanide. Hardwich
A blue deposit through the bath and on the plate	Iron developer carried into the cyanide. Hardwich’s warning, and the reason for the intermediate wash
The collodion film loosening from the glass	Cyanide on a tannin dry plate. Towler, who fixes those in hypo instead
An albumen print bleached and thin after fixing	Cyanide on a printing-out paper. Davanne and Girard, via Reilly; Vogel’s “not admissible”; Abney’s attack on the organic oxide
A negative that clears fast and then keeps clearing	The bath’s solvent power on silver salts is far greater than hypo’s, and there is no self-limiting step. Hardwich
A print or plate that will not tone as expected afterwards	Not addressed by any source in this corpus. The course does not know and does not guess

For a reader identifying an object rather than diagnosing a failure, the useful signature is the image colour: milky-tan on a cyanide-fixed collodion negative against brownish grey on a hypo-fixed one, per the Image Permanence Institute’s table. It is not conclusive on its own — development chemistry shifts the same axis, pyrogallol giving a darker and more burnt-umber image than iron — but it narrows the question, and the AIC’s ambrotype entry is a reminder that both fixers are found in surviving plates.

Experiments

There is no experiment on this page, and there will not be one. A Level D entry that ended with something to try would have given a procedure by the back door.

What a reader can do instead, all of it away from the bench:

Read the strength table as data. Seven authors, one instruction, a factor of four. Work out for yourself what Hardwich’s carbonate contamination does to every figure in the column, and then ask what it does to the comparison between them — whether a spread that wide is evidence about the chemistry at all, or only about the salt.

Test the nomenclature trap on a real source. Find a nineteenth-century toning formula that says “sulphocyanide” and work out, from the formula alone, whether the author meant a thiocyanate. Cassell, Vogel, Abney and Wall all print several. Then look at what the course’s own gold-thiocyanate toner contains, and satisfy yourself that the ancestor and the descendant are the same bath.

Compare the two fixers on the same axis. Take the halide solubility argument from Kodak’s primer — silver iodide is hard to fix in hypo — and set it beside the plain hypo bath and the iodide hypo check. The question worth answering is why the slower reagent won, and the answer is not about speed.

Look at plates. The milky-tan signature is visible on collodion negatives in many collections, and seeing one is worth more than any amount of reading about image colour. Handle nothing: a cyanide-fixed plate, properly washed, carries no cyanide, but the collodion and its varnish are fragile and the glass may be decaying.

What a properly equipped laboratory could still settle, and this course cannot: whether Hardwich’s weeks or Vogel’s days is right for a made-up bath, and by what measure; what the carbonate content of period commercial cyanide actually was, from analysis of surviving stock; and whether the half-tone loss really scales with particle surface area, which is the mechanism this page argues for from Hardwich’s description and cannot demonstrate.