Pellet's positive-working process

Every other iron-blue process in this formulary makes its image where the light fell. This one destroys it there. The paper carries iron(III) in a gum film; light turns that iron to iron(II); and the developer — yellow prussiate rather than red — gives an intense blue with the iron(III) that survived and a colourless salt with the iron(II) that did not. Put a black-line tracing over it and you get blue lines on white instead of the white-on-blue of a blueprint, which is why every drawing office in Europe wanted it and why almost none of them could make it work reliably.

Henri Pellet patented the process in 1877 and then wrote his specifications so that nobody could follow them. The quantities below are not his. They are Giuseppe Pizzighelli and Ludwig von Itterheim’s of 1881, the formula that five separate publications print and that the trade came to call “the Pellet process” anyway.

Solution A — the gum

Ingredient	Quantity	Form the source specifies
Gum arabic	20 g	powdered or lump gum, dissolved cold; Lietze and Wall both print it as 20 parts in 100 of water, and the almanac as 200 g per 1000 cubic centimetres
Water	100 mL, added	Water added, not a make-up volume. Ware states plainly that these recipes follow the "old fashioned" method of stating a weight of solute against a stated volume of water, and every printing reads that way: "Gum Arabic, 20 parts. Water, 100 parts."

Solution B — the light-sensitive iron

Ingredient	Quantity	Form the source specifies
Ammonium iron(III) citrate	50 g	the brown variety in every nineteenth-century printing of this formula; Valenta's green salt did not exist until 1897
Water	100 mL, added

Solution C — the hardening and image-forming iron

Ingredient	Quantity	Form the source specifies
Iron(III) chloride (anhydrous)	50 g	Ware, Lietze and Wall all print "ferric chloride" without qualification; the British Journal Photographic Almanac prints "ferric chloride (crystallised)", which is the hexahydrate, and the two differ by two thirds in iron content — see Function of every ingredient
Water	100 mL, added

Developing solution — the developer

Ingredient	Quantity	Form the source specifies
Potassium ferrocyanide (anhydrous)	20 g	yellow prussiate of potash; what a supplier ships is the trihydrate, so 20 g of crystals carries about 17.4 g of the anhydrous salt
Water	100 mL, added

Discharging bath — the clearing bath

Ingredient	Quantity	Form the source specifies
Hydrochloric acid	100 mL	concentrated acid as sold, at the usual 36 to 37 per cent w/w; poured into the water and never the reverse
Water	1000 mL, added	Lietze gives the bath as 1 part of acid to 10 of water; Ware's footnote states the same ratio as 100 cubic centimetres of the concentrated acid added to 1 litre of water, "not vice versa".

Discharging bath with sulphuric acid substituted — Pellet's own lower-fuming acid bath

Ingredient	Quantity	Form the source specifies
Hydrochloric acid	8 mL
Sulfuric acid	3 mL
Water	100 mL, added	Lietze prints this as Pellet's own recommendation, from the Bulletin de la Société française de photographie for 1880, "in order to diminish the formation of muriatic acid vapors". Both acids go into the water, and Lietze adds that the mixing is done in a bottle or jug and not in a tray. Abney prints the same three figures.

Mixed in this order — the sensitiser, mixed at the time of coating

1. Start with 20 parts Solution A — the vessel that receives — the gum solution is what the other two are added to
2. Then add 8 parts Solution B
3. Then add 5 parts Solution C

For use, mix A, 20 volumes; B, 8 volumes; C, 5 volumes — and in the order given, as otherwise the gum will coagulate. The mixture is, at first, liquid, but, in a few minutes, will become viscid, and, after a few hours, it will have the consistency of soft butter. In this condition, it is the most suitable for the preparation of the paper, and maintains its usefulness, for several days, if kept in the dark and in a closed vessel.

Wall gives the same order with the reason stated the other way round — "if A be added to C, the mixture becomes lumpy and useless" — and Ware notes that the published accounts disagree about when in the thickening the mixture is fit to coat. See Mixing.

Used in this order — processing, after the exposure

1. Developing solution — the image appears in a few seconds in a dark blue colour — Brushed on with a soft hair brush in parallel, joining strokes, quickly and without pressing hard. Pizzighelli then removes the developer at once with a strong jet of water; Lietze warns that this can easily be overdone and that the blue will lift if the ferrocyanide has not had time to reach the paper through the film.
2. Water — a few minutes, floated back upward until the lines can be seen from the back — Then the print is turned over, the greenish-blue surface film removed with a soft brush, and the sheet left in clean water until most of the gum film has dissolved.
3. Discharging bath — a few minutes, where the print has been soaked in water first — The gum film scales off, the blue lines that faded in washing darken again, and the ground clears to white. Abney gives about ten minutes for a print taken straight from the wash.
4. Water — not stated — Washed again and hung up to dry. Lietze records that some workers add a little alum to the last rinse to restore the sizing of the paper, which the acid has attacked.

In order to develop the picture, the print is placed upon a drawing board and, by means of a soft hair brush, quickly — without pressing hard — coated in parallel and joining strokes with the yellow prussiate solution. After the image has been developed the print is floated on water, the greenish-blue film removed with a soft brush, and the sheet immersed in clean water until most of the gum film has dissolved. It is then immersed in the discharging bath, washed again in water, and hung up to dry.

Ware's summary of the same sequence is "developed by a solution of potassium ferrocyanide ... which may be brushed on. It is next washed in water and the soluble gum is removed by a soft brush. Finally it is cleared in a solution of dilute hydrochloric acid, washed again and dried." The mixed-acid bath under Solution clearing-mixed replaces the discharging bath step where the hydrogen chloride fume is a problem.

Purpose

To make a positive Prussian blue print — blue lines on a white ground from a black-line original — by a single contact exposure and a development that reverses the polarity of the ordinary cyanotype. No negative is made, no intermediate is needed, and the reversal is done by chemistry rather than by copying twice.

The whole of the difference from a blueprint lies in one substitution. A classic cyanotype develops in water, and the blue forms from the iron(II) that light made, reacting with the ferricyanide already in the coating. Pellet’s paper contains no ferricyanide at all. It goes into a bath of ferrocyanide after the exposure, and ferrocyanide gives its intense blue with iron(III) — the iron that light has not touched.

Recommended uses

Copying black-line tracings, which is what it was built for. Both Pizzighelli and Lietze are explicit that the original should be on tracing paper or tracing cloth in perfectly opaque ink; Lietze adds that the paper must be well glue-sized so the sensitiser stays on the surface. The process was used from 1877 into the 1940s and, in Colonel Waterhouse’s office, for duplicating maps for the Survey of India.

Teaching what a developer actually decides. This is the cleanest demonstration in the whole iron family that polarity is not a property of the sensitiser. The same photoreduction of iron(III) to iron(II) that makes a cyanotype makes a Pellet print; swap the hexacyanoferrate from the (III) oxidation state to the (II) and the image inverts. A student who has coated both papers has understood something that no amount of reading about “positive-working processes” conveys.

Identifying a historical print. The Getty Conservation Institute’s atlas states that Pellet prints are indistinguishable from cyanotypes by microscopy and by X-ray fluorescence, since the image substance is the same Prussian blue; the only proposed instrumental discriminator, detecting the gum arabic by infrared spectroscopy, it calls a theoretical possibility with a very weak signal. Ware gives the one criterion that works by eye: a Pellet print carries a deposit of insolubilised gum holding the pigment, so it shows a binder layer in relief in the shadows and looks more like a blue Woodburytype or a carbon print than like a cyanotype, which is matte and single-layered with the pigment in the paper fibres.

Not for pictorial photography, and not for a first alternative process. The Getty’s assessment is that it was “much more difficult and delicate than the basic cyanotype process”; Ware records that it called for greater skill and tended to be confined to professional blueprinters. Every source read for this page that tried it reports trouble with the whites.

When another formula is preferable

• For almost any photographic purpose at all, the classic cyanotype sensitiser. It is two chemicals instead of five, develops in plain water, has no acid step and no time limit on getting a colourless salt off a sheet before it goes blue.
• For a positive print from a positive today, make a digital or paper negative and print it negative-working. Reversing in the enlarger or the printer is easier than reversing in the chemistry, which is exactly the judgement the drawing offices came to when diazo copying arrived around 1950.
• For the historical article at the start of the line, Herschel’s 1842 cyanotype. Herschel invented the positive-working version too, in the same year, and never made it reproducible; §4.1.4 of Ware’s Cyanomicon explains why, and Pellet’s gum is the answer to the problem it describes.
• For a long exposure scale and a deep maximum density, the New Cyanotype; for contrast chosen by formulation, the Simple cyanotype. Neither is positive-working, and neither pretends to be.

Mixing

Three stocks, kept apart, combined in a stated order at the time of coating. The order is not a convention. Wall states the consequence of getting it wrong in one sentence: “if A be added to C, the mixture becomes lumpy and useless”.

1. Solution A. Dissolve 20 g of gum arabic in 100 mL of water. Powdered gum goes in faster; lump gum wants a day. Filter it — Wall’s account of the closely related Waterhouse formula says to filter the gum solution through a sponge or a pad of cotton, and gum arabic always carries bark and grit.
2. Solution B. Dissolve 50 g of ammonium iron(III) citrate in 100 mL of water.
3. Solution C. Dissolve 50 g of iron(III) chloride in 100 mL of water. This solution is strongly acidic and corrosive and gets hot as it dissolves; add the solid to the water in portions, never the reverse.
4. Mix, in this order and no other: 20 volumes of A, then 8 of B, then 5 of C, adding in small doses with constant stirring, as the almanac puts it. A is the vessel that receives.
5. Wait, or do not wait, depending on which manual you believe. Lietze says the mixture is liquid at first, viscid within minutes, of the consistency of soft butter after a few hours, and best for coating in that last condition. Wall’s 1912 dictionary says the opposite about the same formula: the solution “is thin, then becomes thick and cloudy, and then clear and liquid again, when it is ready for use”. Ware notes the disagreement and does not resolve it, and neither does this page. Coat a test sheet at each stage and settle it for your own gum.
6. Coat well glue-sized hard drawing paper, pinned flat, with a wide flat bristle brush, as uniformly and quickly as possible and not too heavily; as soon as the brush begins to stick, even the coating with a soft blender. Lietze notes that some workers use a velvet-covered board instead, which needs no blending.
7. Dry quickly, in the dark and warm, and keep the sheet pressed flat or rolled with the coated side outwards. Coating may be done in subdued daylight; drying may not.

The developer and the discharging bath are made separately and kept separately. Dissolve 20 g of potassium ferrocyanide in 100 mL of water for the developer. For the discharging bath, add 100 mL of concentrated hydrochloric acid to 1 litre of water — acid into water, in a bottle or a jug and not in a tray, which is Lietze’s own instruction and still the right one.

🧮What the mixed sensitiser actually contains, and how close it is to Pellet's own

Take the volumes as additive, which they very nearly are: 20 + 8 + 5 gives 33 volumes of sensitiser.

20 vol × 20 g/100 mL ÷ 33 vol = 12.1 % w/v

Gum arabic

8 vol × 50 g/100 mL ÷ 33 vol = 12.1 % w/v

Ammonium iron(III) citrate

5 vol × 50 g/100 mL ÷ 33 vol = 7.6 % w/v

Iron(III) chloride

Now set that beside the one honest sentence in Pellet’s American patent. Lietze transcribes it as “The liquor which is commonly employed by me is composed of: first, gum, 10 to 12 per cent.; second, salt of iron (perchloride at 45° Baumé), 8 to 12 per cent.; third, citric, tartaric, or similar acid, 1½ to 3 per cent.”

The gum figures agree almost exactly. Pizzighelli’s 12.1 per cent sits at the top of Pellet’s 10 to 12 per cent band. The iron is harder to compare, because Pellet is quoting a solution of perchloride and Pizzighelli a weight of solid: Lietze states that a 45° Baumé perchloride has a specific gravity of 1.45 and contains 50 per cent of iron(III) chloride, so Pellet’s 8 to 12 per cent of that solution is roughly 4 to 6 per cent of the salt against Pizzighelli’s 7.6 — but Pizzighelli also has 12 per cent of the citrate carrying iron of its own, and Pellet’s patent nowhere says whether his percentages are by weight or by volume. Treat the comparison as showing that the two formulas are the same animal, not as an equivalence. The same sentence of Lietze’s that gives the 50 per cent also says the solution contains 60 per cent of water, which cannot both be true; only the specific gravity and the 50 per cent are used here, and the discrepancy is left standing where the source put it.

Behaviour

It prints out, faintly and in the wrong colours. The coated sheet is dark brown. Under exposure the image appears in yellow on that brown ground, and Lietze’s end point is simply that the yellow picture is plainly visible: 3 to 10 minutes in the sun, 15 minutes or more in diffused light. That is markedly faster than a classic cyanotype, which Ware confirms in general terms.

Judging the exposure is done on a strip, not on the print. Pizzighelli’s method, which Lietze sets out in full, is a working photometer: expose strips of the same paper under a small glass plate and a scrap of the tracing, alongside the frame and in the same light; withdraw one from time to time and develop it; cover the frame while you do. When the strip goes blue only where the light did not act, and nowhere else, the print is ready. It is the discipline of a test strip, a century before enlarging papers made it routine.

Both failures are visible in the developer within seconds. Too short an exposure and blue appears where it should not, sometimes over the whole sheet. Too long and the blue develops very slowly and some lines never appear at all.

Development is a race, and the prize is a white ground. The image appears in a few seconds. What follows is a genuine conflict of instructions between the two people who knew the process best. Pizzighelli says to remove the ferrocyanide at once with a strong jet of water. Lietze answers that this is easily overdone, that the blue must be given time to form through the film and into the surface pores of the paper or it will lift off with the gum, and that a blue liquid will run over the whole sheet in any case — of no consequence to the print if the exposure was right, and distinguishable from the blue spots of under-exposure because those keep growing. He then floats the print face up for a few minutes until the lines can be seen from the back, which tells him the development has gone right through.

The print gets paler before it gets better. After the surface film is brushed away and the gum dissolves in the water bath, the paper carries a uniform light yellow tint and the lines are noticeably lighter than they were. The acid bath is what returns them: the gum scales off, the blue deepens, the yellow goes, and a blue image on white paper is what remains.

Image characteristics

Blue lines on white, and the polarity follows the original. The Getty states the three cases: positive from a positive transparency, negative from a negative, blue lines on a white ground from a black-line drawing or plan.

The blue is the same Prussian blue as a cyanotype’s, which is why no analytical method separates the two prints on the image substance. Lietze describes the developed lines as dark blue, deepening to indigo, then paling in the wash and darkening again in the acid.

The surface is not a cyanotype’s surface. The image sits in a hardened gum layer rather than in the paper fibres. Ware’s identification note is that this produces a binder layer in relief in the shadow regions, so the print resembles a blue Woodburytype or a carbon print, and that it may therefore be harder to tell from those than from the cyanotype it is chemically identical to.

The characteristic fault is a tinted ground, and everyone reports it. Lietze, having transcribed Pellet’s own patent, concludes that he tried it several times “without succeeding in producing good prints. The background was always more or less colored.” Duchochois, quoted by Ware, says of Pellet’s commercial paper that “it was impossible to work satisfactorily”. The mechanism section explains why this is the process’s structural weakness rather than anybody’s carelessness.

Contrast is a line-copying contrast and nobody measured it. No source read for this page gives an exposure scale, a density range or a characteristic curve for Pellet’s process, and none is offered here.

The mechanism

Four things happen, in this order, and the third is the one that decides whether the print is any good.

⚗Why the developer, not the sensitiser, chooses the polarity

One. Light reduces the iron. Ware’s Table III.1 lists the carboxylic acids whose iron(III) complexes are photosensitive, and citric and tartaric acids are both on it. Blue and ultraviolet light drives charge from the ligand to the metal; the iron(III) becomes iron(II) and the ligand undergoes oxidative decarboxylation.

light + iron(III) citrate → iron(II) + oxidised ligand + carbon dioxide

The photoreduction, in the exposed areas only

Two. Iron(III) hardens the gum; iron(II) does not. This is Alphonse Poitevin’s observation of 1863, and Ware names it as the thing Pellet built on: ferric salts cause gum and similar colloids to harden and become insoluble in water, whereas ferrous salts do not. So the exposure has already written a second, invisible image — a pattern of solubility — before any developer touches the sheet.

Three. The ferrocyanide meets two different irons and gives two different products. There are exactly four combinations of a ferric or ferrous salt with a ferricyanide or a ferrocyanide, and Ware’s Table 3.1 sets out all four. The two that matter here are:

4 Fe³⁺ + 3 [Fe(CN)₆]⁴⁻ → Fe₄[Fe(CN)₆]₃

Unexposed: iron(III) gives Prussian blue

iron(II) + hexacyanoferrate(II) → ferrous ferrocyanide, colourless

Exposed: iron(II) gives Prussian white

Prussian blue is, in the table’s words, “highly insoluble; most intensely coloured, and most stable of all four possible products”. Prussian white — ferrous ferrocyanide, which the same table also calls Everitt’s salt, Williamson’s salt or Berlin white — is “insoluble and colourless”. (Standard inorganic chemistry writes it with potassium in the lattice, as K₂Fe[Fe(CN)₆]; that formula is textbook chemistry and is not given by any photographic source read for this page.) That colourless salt in the highlights is the whole process. Compare it with the classic cyanotype, where the ferricyanide in the coating reacts with the photochemically produced iron(II) to give Prussian blue where the light fell. Same photoreduction, opposite image, and the only thing changed is the oxidation state of the iron inside the hexacyanoferrate — and, with it, whether that reagent has to be present before the exposure or after it.

Four. The exposed film comes off, taking the white salt with it. Because the exposed areas were never hardened, the gum there is still soluble; Ware’s account is that it “became re-solubilised”, and that the Prussian white formed in it “could be readily wiped off, thus fixing the image without much disturbing the Prussian blue”. The wash, the soft brush and the acid bath are all one operation doing that work. In the unexposed areas the iron-hardened gum holds on, and so does the blue inside it.

Now the reason the whites are so hard to keep. Ware’s Table 3.1 gives Prussian white one further property: it is “readily oxidised by air (and other oxidants) to Prussian blue”. The colourless salt sitting in the highlights of a freshly developed Pellet print is not a stable product. It is Prussian blue that has not happened yet.

Prussian white + oxygen + water → Prussian blue + hydroxide ions

Ware's equation for the reoxidation

So the operator is in a race. Every second between development and the removal of the exposed film is a second in which the highlights are quietly turning blue, and every failure the literature records — Lietze’s coloured background, Duchochois’s unworkable commercial paper, Pizzighelli’s insistence on a strong jet of water the instant the detail is out — is that race being lost. It also explains the one instruction that looks like fussiness and is not: the soft brush is not cleaning the print, it is removing a reagent before it has time to react.

⁇What does the light do to the film — soften it or seal it?

Two accounts of the same observation, and they are opposites.

Lietze, 1888, reporting the process as worked. He notes that in the unexposed parts the Prussian blue goes right through the film to the paper and adheres, while in the exposed parts the precipitate forms only on the surface of the film. His explanation: “It is supposed that this abnormal behaviour of the ferric gum film is due to the fact that the film loses its penetrability in the light.”

Ware, 2020. The exposed gum is re-solubilised, because the iron there is now iron(II) and iron(II) does not harden gum. The precipitate stays on the surface because the film beneath it is dissolving, not because it has sealed.

What the course can say. Both explain the observation. Poitevin’s 1863 result — which Ware cites and Lietze does not — favours Ware, because it makes a testable statement about ferric and ferrous gum rather than a supposition about penetrability. Lietze’s phrasing (“it is supposed”) shows he knew it was a conjecture. The course treats Ware’s account as the better-founded one and Lietze’s as the period reading, and notes that nobody in the sources read for this page has done the experiment that would settle it: coat two sheets, expose one fully, and measure how much gum each loses in water.

The second anomaly is stranger and remains open. Lietze reports that developing this film with red prussiate — potassium ferricyanide — also gives a positive, “just as with the yellow prussiate, but somewhat slower and of a greenish color”, which turns blue in the acid bath. He says plainly that “any one who is familiar with the reaction of ferrous salts would expect a negative picture”, and he is right to be surprised: iron(II) with ferricyanide is the classic cyanotype reaction and should have printed the exposed areas blue. Ware’s Table 3.1 supplies part of an answer — iron(III) with ferricyanide gives Prussian yellow, a soluble and powerful oxidant that is reduced through green intermediates to Prussian blue, which fits both the greenish colour and its conversion in acid — but why the exposed areas do not go blue is not answered by anything this course holds. It is recorded here as an unexplained observation by a named practitioner, not as chemistry. It does have one consequence for reading the literature: when Abney prints a positive-process developer as a saturated solution of potassium ferricyanide in the same passage where his own explanation says ferrocyanide, that may be a genuine variant reported by a third party rather than the misprint it looks like.

Function of every ingredient

Gum arabic — 20 g in 100 mL, 12.1 per cent in the mixed sensitiser. The ingredient that turned Herschel’s failed experiment into a patentable process, and it does three distinct jobs. First, it is a thickener: it keeps the iron salts on the surface of a well-sized sheet instead of letting them sink into the fibres, which is the plain reason Abney gives — “the gum in this process, and in that of Pizzighelli, is used to prevent the staining of the background”. Second, and this is Pizzighelli’s own point as Lietze transcribes it, the gum “possesses the property of forming, when dry, with iron salts, a substance resembling horn, which does not easily dissolve in water, and which causes the developer to act only superficially, and to but slowly penetrate the film”. Third, that horn-like ferric gum is the binder of the finished image: it is what holds the Prussian blue, and it is why Ware can identify a Pellet print by the relief of its binder layer. Less gum and you get blue spots in the ground; more and the mixture becomes so thick that it cannot be coated uniformly. Lietze states both consequences and concludes that the quantity in the recipe “has proved to be the best”. Note what this ingredient is not: it is not a photosensitive material in its own right in any source read here, and the process is not a gum-dichromate relative.

Ammonium iron(III) citrate — 50 g in 100 mL, 12.1 per cent in the mixed sensitiser. The light-sensitive component. Iron(III) held by a citrate ligand absorbs in the blue and near ultraviolet and is reduced to iron(II) while the citrate is oxidised and decarboxylated; that is the reaction listed for citric acid in Ware’s table of photosensitive iron(III) carboxylates, and it is the same reaction that drives every cyanotype. Every nineteenth-century printing of this formula means the brown salt, because Valenta’s faster green form was not prepared until 1897; a modern worker using the green salt is not using the historical sensitiser and should say so. On its own it is not enough. Lietze is categorical: “Ferric ammonium citrate alone, with or without gum arabic, will give no good results.” The citrate supplies the sensitivity; something else has to supply the blue.

Iron(III) chloride — 50 g in 100 mL, 7.6 per cent in the mixed sensitiser. The other iron, and it does the two jobs the citrate cannot. It is the hardener — the ferric salt that, by Poitevin’s 1863 result, makes the gum insoluble — and it is the bulk of the iron(III) that the ferrocyanide converts to Prussian blue in the unexposed areas. It is also comparatively insensitive to light, which is exactly what you want in an ingredient whose job is to survive the exposure. More of it means a darker print and a slower one, which is Lietze’s own statement: “a larger proportion of ferric chloride diminishes the sensibility to light, but furnishes pictures of darker color”. Less of it and both the blue and the hardening weaken together, which shows up as lines that lift off in the wash.

🧮Which ferric chloride? The answer changes the formula by two thirds

Ware, Lietze and Wall all print “ferric chloride” and stop. The British Journal Photographic Almanac prints “ferric chloride (crystallised)” — the hexahydrate. The difference is not pedantry:

55.845 ÷ 162.2 = 34.4 % iron

Anhydrous FeCl₃

55.845 ÷ 270.29 = 20.7 % iron

FeCl₃·6H₂O

Fifty grams of the anhydrous salt carries 17.2 g of iron; fifty grams of the hexahydrate carries 10.3 g — a factor of 1.67. Which one the formula means matters more than any other unstated detail on this page. Two arguments say it means the crystals. The almanac says so outright. And the anhydrous salt is so hygroscopic that the course’s own encyclopaedia entry calls it “not a stable form to keep in a darkroom, because the jar takes up water and becomes the hexahydrate” — a nineteenth-century drawing office would not have had it. The formula object records the substance as the sources name it, unqualified; this course’s reading is that a modern worker should use the hexahydrate, and should say which was used.

Potassium ferrocyanide — 20 g in 100 mL of water, brushed on after the exposure. The developer, and the ingredient that makes the process positive. Yellow prussiate of potash gives an intense Prussian blue with iron(III) and a colourless Prussian white with iron(II); since the exposure has converted iron(III) to iron(II) exactly where the light fell, the blue lands everywhere the light did not. Swap it for red prussiate and the ordinary cyanotype reaction takes over — or at least it should, and Lietze reports that it does not, which is discussed under The mechanism. A stronger bath develops faster and stains the ground more; a weaker one is slower and cleaner. Pellet’s patent says only that the bath may be “concentrated, or dilute, neutral, slightly acid, or slightly alkaline, hot, or cold”, which is the sort of guidance that made the patent useless; the almanac prints a 10 per cent bath against Lietze’s 20, so the working range in period practice was at least two-fold. One practical point the sources do not make and this course will: what suppliers sell is the trihydrate, so 20 g of crystals is about 17.4 g of the anhydrous salt, a 13 per cent shortfall that sits well inside the range the printings themselves disagree over.

Hydrochloric acid — 100 mL of the concentrate to 1 litre of water, after the wash. Period manuals call this bath the discharging bath rather than a fixer, and the word is accurate: nothing is being made insoluble, things are being taken away. It dissolves the iron salts that never became Prussian blue — Abney’s end point is that the acid “will have removed all iron salts not turned into the blue compound” — it loosens the remaining gum film so that it scales off the sheet, and it deepens the blue that washing had paled. Prussian blue is stable in acid and destroyed by alkali, which is why an acid clear is the right final treatment for this image and a soda bath is not. Too weak or too brief and the ground stays yellow; too long and the sizing of the paper suffers, which is why Lietze records that some workers add a little potassium alum to the final rinse to put the size back.

Sulphuric acid — 3 mL beside 8 mL of hydrochloric acid in 100 mL of water, in the alternative bath. This ingredient exists for one reason and it is a safety reason, which makes it unusual in a nineteenth-century formula. Lietze attributes the substitution to Pellet himself, publishing in the Bulletin de la Société française de photographie in 1880, “in order to diminish the formation of muriatic acid vapors”: sulphuric acid is far less volatile than hydrogen chloride, so replacing part of the hydrochloric acid keeps the bath just as acidic while putting much less acid into the air above the tray. Abney prints the identical proportions. It changes nothing about the image and a great deal about the room.

Water, and the paper. Both are ingredients here in a way they are not in a developer. The water is added in every printing of this formula — “20 parts gum, 100 parts water”, not “made up to 100” — which is the old-fashioned convention Ware flags, and it means these are not quite the percentages they look like. The paper is stipulated: Lietze says well glue-sized hard drawing paper is “absolutely necessary”, and that soft, poorly sized paper lets the mixture soak into the pores and produces blue spots. On this process the sizing of the sheet is as much part of the formula as the gum.

Interactions

The two iron salts divide one job between them, and the division is the formula’s design. The citrate is sensitive and gives a weak blue; the chloride is insensitive and gives a strong one and hardens the gum. Raise the chloride and you buy density and hardening with speed. Drop the chloride to nothing and, by Lietze’s test, you have no process at all. This is the same trade the classic cyanotype sensitiser makes between citrate and ferricyanide, moved one step along: there, the second ingredient is the precipitant; here it is a second source of the ion to be precipitated.

Gum against iron is the process’s real control. The gum needs enough iron(III) to harden and the iron needs enough gum to be held; too much gum and the developer cannot penetrate at all, too little and it penetrates everywhere. Lietze’s two failure modes — blue spots from too little gum, an uncoatable syrup from too much — are the two ends of that one adjustment.

The order of mixing is a colloid interaction, not a chemical nicety. Gum arabic solutions are coagulated by concentrated multivalent electrolyte. Adding A to C means introducing gum into a strong ferric chloride solution, and Wall reports the outcome: lumpy and useless. Adding C to A, in small doses with constant stirring, dilutes the salt as it arrives. This is why the mixture then thickens over minutes to hours: the gum is being progressively cross-linked by iron(III) in the bottle, and the sensitiser is going off from the moment it is made. That is also why it keeps only days.

The acid bath interacts with the developer, and the interaction is the safety control on this page. A hexacyanoferrate and a mineral acid are being used in the same workflow, minutes apart, on the same sheet. See Safety and Incompatibilities: the sequence is deliberate and the reason it is tolerable is dilution and temperature, neither of which is optional.

Alkali is the one thing that destroys the finished print, because Prussian blue is decomposed by it. That is also, usefully, the basis of the period correction technique: Lietze removes blue spots with lime water or a 1 : 25 solution of caustic potassa, and Abney uses a 4 per cent solution of potassium hydroxide for the same purpose. Wall’s dictionary and the almanac give the same trick with potassium oxalate at 4 to 15 per cent, thickened with gum if a fine line is wanted, for writing titles and erasing lines — the oxalate complexes the iron rather than attacking the pigment.

Variants

Pellet’s own liquors, from United States patent 241,713 of 17 May 1881, transcribed by Lietze. The patent offers five example compositions “which can be used singly or mixed”, of which the first and fourth are the clearest: perchloride of iron at 45° Baumé 8 cm³ with oxalic acid 6 g, water to 100 cm³; and citrate of iron or ammoniacal citrate of iron 3 to 5 g, water to 100 cm³. A second type uses citric acid as sodium or potassium citrate; a fifth uses tartaric acid, citric or similar at 6 to 8 g with 6 to 8 cm³ of the perchloride. The one sentence of the patent that is worth more than the rest together is the statement of what Pellet actually used: gum 10 to 12 per cent, perchloride of iron at 45° Baumé 8 to 12 per cent, and citric, tartaric or similar acid 1½ to 3 per cent. The British abridgment of the 1877 patent describes the same latitude without any numbers at all — “an iron salt, an organic acid, and a gummy substance, such as dextrin, gelatine, gum, isinglass, albumen, glycerine” — and Ware’s verdict on all of them is that no clear working method can be extracted, “doubtless, just as he intended”.

Waterhouse’s, or rather Fisch’s, or arguably Poitevin’s. The other clearly formulated version, and the one that keeps better in a hot climate. Gum arabic 26 g in 100 cm³, tartaric acid 27 g in 100 cm³, ferric chloride 100 g in 100 cm³, mixed in the volume ratio 13 : 3 : 2, set aside in the dark for 24 hours, then diluted to a hydrometer reading of specific gravity 1.100. Processing is identical. There is no ammonium ferric citrate in it at all — the photosensitive complex is iron(III) tartrate — which is the cleanest possible demonstration that the citrate is not what makes this process work, only what makes it fast. Waterhouse credits Fisch; Duchochois says the formula is substantially Poitevin’s with Pellet’s gum added; and when Waterhouse had both versions tested against each other in the Survey of India’s photographic office at Calcutta in 1887–8, this one gave the better keeping.

Collache’s, patented in France in 1880. Gum arabic 7 to 10 parts, citric acid 2 to 3, ferric chloride solution at 45° Baumé 4 to 6, water 81 to 87, applied in two coats with the first dry before the second. Lietze reports it as easier to coat than Pizzighelli’s and as giving a lighter, less brilliant blue.

The formulas circulated as “Pellet’s secret”. Because the patent said nothing usable, several mutually inconsistent recipes were published as his. Lietze prints “Joltrain’s Positive Blue Print Process, according to Liesegang” — gum arabic, ferric chloride solution, ferric sulphate and tartaric acid — under the name of the woman who first manufactured the paper. Abney prints a quite different one, also on Liesegang’s authority and also labelled “Pellet’s process (supposed to be a secret)”: sodium chloride 3 parts, ferric chloride 8, tartaric acid 3½, water 100, with 25 parts of powdered gum arabic dissolved in half the water. Neither is corroborated by any other source read for this page, and the disagreement between them is itself evidence that nobody outside Pellet’s works knew the composition.

Differences between printings of the formula that is on this page. The British Journal Photographic Almanac gives the developer as 10 per cent rather than 20, and “fixes” in sulphuric acid at 1 : 25 rather than in the 1 : 10 hydrochloric bath. Wall’s 1912 dictionary gives the same 1 : 10 hydrochloric acid as Lietze; his 1924 book gives “a 1 per cent solution of hydrochloric or sulphuric acid”, which is a very much weaker bath than either. These are not variants of substance, but a reader comparing manuals should know that the acid step is the least standardised part of the process.

Red prussiate as the developer. Lietze’s report that potassium ferricyanide also gives a positive image is set out under The mechanism. It is a variant nobody has explained, and it is not recommended here — but it is the likeliest explanation of the ferricyanide that appears in some manuals’ accounts of this process, and it is a good experiment.

Safety

Level B, and the level comes from the chemicals rather than from the process. Every substance in the formula object is Level B on its own encyclopaedia page — iron(III) chloride, potassium ferrocyanide, hydrochloric acid and sulphuric acid — except gum arabic and ammonium iron(III) citrate, which are Level A. The course takes a formula’s level from the highest of its ingredients, and that is B.

But this page carries one hazard that no single ingredient’s level describes, and it must be stated plainly. The process deliberately brings a print loaded with hexacyanoferrate into a bath of mineral acid, minutes later. That is the pairing the whole cyanotype literature warns against, and the classic cyanotype sensitiser page forbids acid anywhere near the workflow for exactly this reason. Here it is not avoidable: without the discharging bath there is no white ground.

⚠The acid step, and why it is tolerable at these strengths and not at others

What the risk actually is. Hexacyanoferrate(II) is a complex, not a cyanide salt; the cyanide is bound to iron and is not free. Ware’s qualification of the standard warning is the one to hold on to: to get much hydrogen cyanide out of one of these salts, “the heating has to be very strong and the acid very concentrated”. The discharging bath is a room-temperature bath of about 1.2 mol/L hydrochloric acid — a tenth-strength dilution — acting on the small residue of ferrocyanide left on a sheet that has already been washed. That is not the strong-acid, high-temperature case.

The controls that make it so, none of which is optional.

• Wash before the acid, thoroughly. Lietze’s own reason is a working one — pre-soaked prints need only a few minutes in the acid and the bath lasts longer — but the safety consequence is the one that matters: most of the ferrocyanide never reaches the acid tray at all.
• Never concentrate the bath and never warm it. Both of Ware’s conditions are conditions you would have to create deliberately. Do not.
• Ventilate the acid tray, and prefer Pellet’s own mixed bath — 8 parts hydrochloric acid and 3 of sulphuric in 100 of water — which he published in 1880 for precisely this purpose. Hydrogen chloride fume is the hazard you will actually meet on this process; hydrogen cyanide is the one you must not create.
• Never put the ferrocyanide developer into the acid, or the acid into the developer, and never pour spent developer into an acid waste bottle. Keep the two liquids in different vessels, on different parts of the bench, with different measures.
• Dilute acid into water, in a bottle or jug, never in a tray — Lietze’s instruction, still correct, and the reason is that the dilution is strongly exothermic and a tray gives you no depth of water to absorb the heat.

If the process required a concentrated or a warm acid bath, this page would be Level C and would carry no procedure. It does not, and the classification says so honestly rather than being raised for the appearance of caution.

The rest of the safety picture:

• Iron(III) chloride is the most aggressive thing you will handle. It is corrosive, its solutions are strongly acidic by hydrolysis, and dissolving 50 g in 100 mL of water releases noticeable heat. Gloves, eye protection, add the solid to the water in portions. It also stains skin, wood, cloth and every white surface a rust brown that does not come out.
• This is an ultraviolet process, and a sunlit one. Three to ten minutes in direct sun is repeated daily exposure for anyone printing production quantities. Cover the skin and shade the eyes; a UV unit must be enclosed.
• No dust. All three solids are weighed dry. Weigh them without raising dust, and keep every utensil that has touched them away from food.
• The finished print is Prussian blue and gum, and is not hazardous. Prussian blue is Level A.

Storage

Three bottles, never one, and the gum bottle is the one that fails. Lietze gives solution A a few days before it becomes sour and unfit; Wall gives it about a week before it becomes acid. Both are reporting the same thing — an aqueous gum solution is a nutrient medium and it ferments — and both are short. Make the gum fresh, in the quantity the session needs. Solutions B and C keep: Lietze says several weeks in closed vessels, Wall says indefinitely in the dark. The disagreement is not resolvable from the sources and both figures are recorded in the entry rather than averaged.

The mixed sensitiser keeps for several days in the dark in a closed vessel, and is changing the whole time. It thickens because the gum is being cross-linked by iron(III), so a bottle that has stood is not the same coating fluid as one just mixed. Date it.

The coated paper is the weakest link in the whole process. Wall states flatly that “the paper has very poor keeping qualities”. Lietze’s handling rules are precise and are worth following literally: dry it quickly in a dark, warm place; keep it pressed flat, or rolled with the coated surface outward, because it curls the other way; protect it from light and from moisture. Waterhouse’s tests in Calcutta found that the tartaric-acid version of the sensitiser kept better in a hot climate, which tells you what the enemy is.

Store the acid separately from everything else, in a bottle rated for it, on a low shelf, with the ferrocyanide nowhere near.

Incompatibilities

• Alkali, with the finished print, absolutely. Prussian blue is decomposed by alkali to weakly coloured iron compounds. That means no buffered mountboard, no alkaline sizing, no alkaline tap water in the final rinse, no ammonia or household cleaner in the room. The correction techniques that use lime water, caustic potash or soda are deliberate, local applications of exactly this destruction.
• Concentrated acid or heat, with the ferrocyanide. See Safety. The dilute room-temperature bath is the process; anything stronger or warmer is not a stronger version of the process, it is a different and worse chemistry.
• The developer and the acid in the same vessel, at any point, including the waste bottle.
• Iron(III) salts and the developer, other than on the print. They give Prussian blue on contact, which is the reaction the process is for and a permanent stain everywhere else. Dedicate measures, trays and brushes; a pipette that has been in solution C and goes into the ferrocyanide bottle turns the bottle blue.
• Silver, and every silver bath in the darkroom. Ferro- and ferricyanides attack image silver; a shared tray or cylinder will damage a silver print later.
• Soft or poorly sized paper, which Lietze names as a cause of failure rather than a preference.
• Metal trays, tongs and clips. An iron-and-cyanide chemistry finds any trace of iron and prints it.

Waste

Four streams, and they must not meet. The spent developer is a dilute hexacyanoferrate(II) solution; the spent discharging bath is dilute mineral acid carrying dissolved iron salts; the wash water carries gum, iron salts and peptized Prussian blue; and the surplus sensitiser is a concentrated iron(III) solution.

Keep the ferrocyanide bottle and the acid bottle apart, labelled, and closed. This is the disposal consequence of the safety rule, and it is the one thing on this page that a tidy-minded person is most likely to get wrong: pouring two spent baths into the same carboy at the end of a session is precisely the concentration-and-mixing step the process spends its whole procedure avoiding.

Do not drain the ferrocyanide. The aggregated classifications for potassium ferrocyanide are dominated by the long-lasting aquatic hazard, and the course’s chemical page collects it rather than draining it. Neutralising the acid bath is a matter for local rules, not for this page.

Everything here is jurisdictional. Bottle each stream, label it with the substance and the date, and take it to a licensed hazardous waste route. Your local regulations govern, and they differ.

Troubleshooting

• Blue spots in the ground, growing while you watch. Under-exposure. Lietze distinguishes it from the general blue liquid that runs over every sheet during development: that colouration is uniform and static, the spots “spring up and continually increase in number and size”. Expose longer and check on a strip.
• The whole sheet goes blue in the developer. Severe under-exposure, or a sheet that was fogged by being coated or dried in too much light.
• The lines develop slowly and some never appear. Over-exposure. There is no rescue; the iron(III) those lines needed has been reduced.
• A tinted background on an otherwise correct print. Prussian white in the highlights has oxidised to Prussian blue before you got it off. Develop, wash and brush faster; consider Lietze’s method of floating face-up until the lines show through from the back, then turning over and brushing at once.
• The blue lines lift off in the wash. The developer was removed too soon and the Prussian blue formed only in the surface of the gum film, not down to the paper. Lietze’s warning against Pizzighelli’s instant jet of water is exactly this failure.
• Blue spots scattered over the sheet from the start, with a mottled coating. Too little gum, or paper that is too softly sized, letting the sensitiser sink into the fibres.
• The mixture will not coat evenly and drags under the brush. Too much gum, or a mixture left too long. Both of Lietze’s limits are being hit.
• The mixture goes lumpy on mixing. A was added to C instead of C to A.
• The ground stays yellow after the acid bath. Too short a time in the acid, too weak a bath, or a bath exhausted by prints that went in without a pre-soak.
• The paper is limp and cockled after drying. The size has been attacked by the acid; Lietze’s workers added a little potassium alum to the final rinse for this.

Experiments

• Prove that the developer chooses the polarity. Coat two sheets with the same sensitiser and expose them identically under the same tracing. Develop one in 20 per cent potassium ferrocyanide and the other in a classic cyanotype’s plain water bath after adding potassium ferricyanide to the sensitiser. Record which areas go blue in each. This is the single most instructive hour in the iron processes.
• Time the highlight. Develop a print, wash it, and then instead of brushing the exposed film away immediately, brush successive strips of it away at 30 seconds, 2 minutes, 5 minutes and 15 minutes. Measure or visually rank the residual blue in each strip. You are measuring the reoxidation of Prussian white by air, and the result is the reason the process failed commercially.
• Test the two accounts of what light does to the film. Coat two sheets; expose one to full sun until the yellow image would be well past printed, leave the other in the dark. Soak both in water for a fixed time, dry, and weigh. If the exposed sheet loses substantially more gum, Ware’s re-solubilisation account is supported; if it loses less, Lietze’s is. Nothing in the sources read for this page has done this.
• Map the speed-against-density trade. Make five sensitisers varying only solution C, at 3, 4, 5, 6 and 7 volumes against A 20 and B 8. Expose a step wedge on each in the same light and develop identically. Lietze’s claim is that more ferric chloride means less speed and a darker image; find out by how much, and whether the ground stains more.
• Find your gum limits. The same experiment on solution A, at 15, 20, 25 and 30 volumes. Lietze predicts blue spots below and an uncoatable syrup above.
• Repeat Lietze’s anomaly. Develop an exposed sheet in potassium ferricyanide instead of ferrocyanide and record what you get, including the colour before and after the acid bath. If it comes out negative rather than positive, you have found the limit of a report that has stood unexamined since 1888; if positive, you have confirmed an observation nobody has explained.