Pizzighelli and Hübl's 1886 formulations

William Willis patented the platinotype in 1873 and improved it twice, and his patents are what patents are: enough to claim the ground, not enough to work from. He states no reason for choosing one salt over another, gives no account of what the alternatives did, and — as Pizzighelli and Hübl put it, with some restraint — the results obtained from the substances specified “are altogether wanting”. Ware records the consequence: British writers on platinotype avoided giving explicit instructions for preparing the paper at all, probably for fear of the patents.

Two captains in the photographic department of the Austro-Hungarian military technical administration did the missing work instead. They tested the salts of iron, ranked sixteen candidate developing substances by reducing power, measured the ratio of platinum to iron that gives a black rather than a grey, and published all of it — first in German in 1882, then in French and English in 1883, and as the London book of 1886 from which the quantities below are read. What they published is the first platinum sensitiser anybody outside Willis’s factory could make, and the reason it is still worth reading is not nostalgia. The contrast-control system in every platinum and palladium kit sold today is theirs, essentially unchanged: two bottles of the same iron solution that differ in one substance, traded against each other so that the amount of oxidant moves and nothing else does.

The normal solution of platinum — the metal that becomes the image

Ingredient	Quantity	Form the source specifies
Potassium tetrachloroplatinate(II)	1 g	"1 part of the salt"; the source states parts and no absolute weight, and the pair below is written at 1 g against 6 mL to hold the ratio. Potassium chloroplatinite in the book's language, K2PtCl4, FW 415.11. The authors make it themselves from commercial platinic chloride and give the preparation; they also give two purity tests for a bought salt — it must dissolve completely in about six parts of cold water, and the solution must not be acid
Water	6 mL, added	"Six parts of distilled water." An added volume rather than a make-up volume: the source gives a ratio and never a finished volume, so nothing here can be recorded as "to make". Ware converts the same 1 part in 6 to about 15.6 per cent w/v and 0.376 molar, and notes that Cassell's Cyclopaedia calls it saturated while Paul Anderson prescribes 1 part in 5, about 18.5 per cent — so the strength of this bottle is disputed in the literature and the course prints the ratio the source prints

The normal ferric solution — the only light-sensitive substance in the system

Ingredient	Quantity	Form the source specifies
Ferric oxalate	20 g	Per 100 c.cm. of the finished solution — the source writes the formula as Fe2(C2O4)3 and reaches this strength by analysing a home-made solution and then diluting it, because "ferric oxalate cannot be prepared in a stable form, so as to have a constant composition". The book gives the analysis, the preparation from ferric chloride and caustic soda, and two acceptance tests for a bought solution
Oxalic acid	6–8 %	Crystallised oxalic acid, added until the free acid — including whatever the solution already carried — "amounts to from 6 to 8 per cent. of the ferric oxalate already in the solution". A percentage of the iron salt, not of the solution: at 20 g of ferric oxalate per 100 c.cm. that is 1.2 to 1.6 g per 100 c.cm., which is the course's arithmetic and not the book's wording. The range is the source's own
Water	to make 100 mL	Distilled water. This is a genuine make-up volume: the analysed solution is diluted "with so much distilled water, that every 100 c.cm. of it may contain 20 grammes of ferric oxalate". Ordinary spring water is excluded by name because its lime throws calcium oxalate

The normal chlorate of iron solution — the contrast bottle — the same iron, with an oxidant in it

Ingredient	Quantity	Form the source specifies
The normal ferric solution	100 mL	"A certain quantity of the normal ferric solution." The source names no volume; 100 mL is entered because 100 c.cm. is the basis its chlorate figure is stated against, and because the two solutions are otherwise identical in iron, in acid and in behaviour
Potassium chlorate	0.4 g	"So much chlorate of potash as to cause 100 c.cm. of the liquid to contain 0.4 gramme of the salt" — 0.4 per cent w/v, which is 2 per cent of the weight of ferric oxalate present. Ware's Table 5.1 gives the identical figure for Pizzighelli's 1887 solution C
Water	to make 100 mL	No water is added. The 100 mL is the source's own basis — the chlorate figure is stated per 100 c.cm. of the finished liquid — and not an instruction to dilute. Recording it as a make-up volume is the closest the schema comes to "the chlorate is dissolved in the iron solution and nothing else goes in"

Mixed in the ratio — the normal sensitising solution — very soft, with deep blacks

24 parts The normal solution of platinum + 22 parts The normal ferric solution + 4 parts water

"In the normal sensitizing solution we give the following prescription: 24 c.cm. of platinum solution, 22 c.cm. of ferric solution, 4 c.cm. of water. This works very soft, with deep blacks." The mixture is measured out in a measuring-glass immediately before use, in the quantity the sheet needs and no more.

Mixed in the ratio — rather more brilliancy

24 parts The normal solution of platinum + 18 parts The normal ferric solution + 4 parts The normal chlorate of iron solution + 4 parts water

"In order to produce rather more brilliancy, the following composition may be used: 24 c.cm. of platinum solution, 18 c.cm. of ferric solution, 4 c.cm. of ferric-chlorate solution, 4 c.cm. of water."

The iron still totals 22 parts and the platinum and the water do not move. Four parts of the chlorate bottle put 0.016 g of potassium chlorate into 50 c.cm. of sensitiser, about 0.032 per cent w/v — roughly three times the hundredth of a per cent the authors say is already noticeable

Mixed in the ratio — results corresponding to silver images

24 parts The normal solution of platinum + 14 parts The normal ferric solution + 8 parts The normal chlorate of iron solution + 4 parts water

"To obtain results corresponding to silver images mix: 24 c.cm. of platinum solution, 14 c.cm. of ferric solution, 8 c.cm. of ferric-chlorate solution, 4 c.cm. of water."

The iron is again 22 parts. About 0.064 per cent w/v of chlorate in the finished sensitiser. The description is the useful part: a platinum print at this contrast is meant to look like the albumen and salted paper prints the reader already knew

Mixed in the ratio — very weak negatives, reproductions, engravings

24 parts The normal solution of platinum + 22 parts The normal chlorate of iron solution + 4 parts water

"For very weak negatives, reproductions, engravings, &c.: 24 c.cm. of platinum solution, 22 c.cm. of ferric-chlorate solution, 4 c.cm. of water."

The plain iron bottle is gone and all 22 parts of iron are chlorated, giving about 0.176 per cent w/v in the sensitiser. This is the hardest grade the system reaches and the one the authors' own fault list warns about from the other side — "the prints come out hard; too much chlorate in the sensitizing solution"

☠What this page will not tell you how to do

This is a Level D entry: the chemistry and the history, and no working procedure. Two of the three bottles carry the reasons. Potassium tetrachloroplatinate(II) is a respiratory sensitiser whose governing hazard statement is H334 — may cause allergy or asthma symptoms or breathing difficulties if inhaled — and HSE’s EH40 sets halogeno-platinum compounds at 0.002 mg/m³ as platinum with the Sen notation, against 5 mg/m³ for platinum metal itself. Potassium chlorate carries the harmonised classification H271, strong oxidiser, signal word Danger, and the deciding operation is weighing a dry oxidiser onto a bench covered in paper. The Level D policy carries the argument for why a hazard whose consequence is permanent is classified alongside an acute one. The Safety section below adds two hazards this page has that the modern kits do not: a near-boiling oxalate bath, and a preparation that runs sulphur dioxide gas through a hot platinum solution.

The course’s own route to a noble-metal print is the palladiotype, with the contrast put into the negative instead of into the bottle.

Purpose

To put a measured weight of platinum salt and a measured weight of light-sensitive iron salt onto a sheet of paper, in a ratio chosen so that the iron can reduce all the platinum, at a contrast chosen before the sheet is coated.

Every clause of that does work. The measured weight is the whole point of the publication: Willis worked to a recipe he did not explain, and these two established by experiment what the numbers ought to be and why. The ratio is the finding they were proudest of, and it is arithmetic anybody can check — reproduced below, and correct. And chosen before the sheet is coated is the awkward part of platinum printing that every later system inherits: there is no filter, no graded paper and no second chance, so the only place a contrast decision can live is in the mixture.

The formulations are also a complete account of a working method, which is rarer than it sounds for a nineteenth-century process. The book gives the sizing baths, the coating rate per square centimetre, the drying window in minutes and degrees, the exposure in a named month, the developer, the clearing bath, a seven-item fault list with causes, and the recovery of platinum from spoiled prints. A modern reader can therefore check the system against itself, which is exactly what the Mixing section below does — and one of the checks fails.

Recommended uses

What the source records, as history. The course recommends none of it, and this section exists because reading a period manual or a modern kit sheet requires knowing what the four mixtures were for.

The normal mixture, for an ordinary negative. No chlorate at all. The authors describe it as working “very soft, with deep blacks”, and Ware’s measurements say why that is not a fault: the unmodified platinotype has an exposure scale of about 2.0, which is longer than the density range most negatives carry, so a normal negative lands entirely on the straight part of it and prints with the long grey scale the process is admired for.

The brilliant mixture, for a slightly flat negative. Four parts of the chlorate bottle in twenty-two parts of iron. It is a small change — under a fifth of the iron — and the authors insist it is already at the edge of what the reagent tolerates.

The silver-like mixture, when the print has to look like the prints around it. Eight parts in twenty-two. The description is worth pausing on: in 1882 the comparison print was an albumen or salted paper print, and this mixture exists so that a platinum print could be shown alongside them without looking flat. It is a formula for a social context, not for a negative.

All-chlorate, for very weak negatives, reproductions and engravings. A copy negative of a pencil drawing or a line engraving has almost no mid-tone to lose, so truncating the scale costs nothing and the extra separation is all gain. This is the one case where the modern objection to chlorate does not really bite.

Dilution, for a print that is not meant to have a black. The authors add half or an equal volume of water to any of the four when reproducing pencil drawings, and reduce the platinum again when a print is to be retouched or painted over. Less platinum on the same area is less maximum density; it is the only tone control in the system that does not touch contrast.

When another formula is preferable

• The palladiotype, with the contrast put in the negative. The course’s route to a noble-metal print, and the honest answer to almost everything on this page. The same iron photochemistry and the same coating craft, a metal that is not a respiratory sensitiser and not five times the price, and no oxidiser in the darkroom. Its published developers are Willis’s potassium oxalate developer and the sodium citrate developer, and the clearing sequence is the step that decides whether the print survives.
• Ware’s print-out platino-palladiotype. The descendant of the other Pizzighelli formulation — the 1887 direct-printing one — rebuilt with a well-characterised iron salt. It takes its contrast from humidity and from the metal ratio rather than from an oxidant, it prints out so that test strips become unnecessary, and it is self-masking, which is a better answer to a long-scale negative than chlorate ever was.
• The three-solution drop system, to see what became of this one. Two iron bottles differing in chlorate and a metal bottle, drop-counted: the same idea as the four mixtures above, industrialised. Reading the two entries together shows how little the method changed in 140 years, and the drop-system page carries the modern figures and the suppliers’ disagreements.
• A calibrated negative. Ware’s conclusion, and the course follows it: with modern negative-making, especially by digital negative and a correction curve, contrast-enhancing agents “become unnecessary when a correctly calibrated negative is made”. The chlorate bottle exists because a nineteenth-century printer could adjust the sensitiser and could not adjust the negative.
• The kallitype sensitiser, to learn the craft on. The same iron chemistry with silver instead of platinum, at a fraction of the cost, and the coating, drying, printing and clearing are the same skills. Its three developers — sodium citrate and Rochelle salt among them — are, not coincidentally, three of the substances Pizzighelli and Hübl ranked in 1882.
• The classic cyanotype sensitiser, to learn it more cheaply still, if the point is to understand ultraviolet contact printing rather than noble-metal chemistry.

Mixing

This section gives no mixing procedure. What can be described without giving instructions is the shape of the system and the arithmetic inside it, and in this formula the arithmetic is most of what there is to learn.

Three bottles, two of which are the same solution. The platinum bottle is a simple dissolution and keeps indefinitely. The iron bottle is the difficult one and the book is candid about why: ferric oxalate “cannot be prepared in a stable form, so as to have a constant composition”, so there is no way to weigh out a known quantity of it, and the authors’ answer is to make a solution, submit it to quantitative analysis, and dilute it to a stated strength. The chlorate bottle is the iron bottle with 0.4 g of potassium chlorate per 100 c.cm. dissolved in it and nothing else changed.

The consequence of that third bottle is the design of the whole system, and it is easy to miss. Look down the four mixtures: platinum is 24 parts in every one, water is 4 parts in every one, and plain iron plus chlorated iron is 22 parts in every one. Only the split between the two iron bottles moves. The amount of platinum on the sheet, the amount of iron on the sheet and the volume that has to be spread over it are identical at every grade — so changing contrast changes contrast and nothing else. That is the same invariant the modern drop tables hold, where the metal stays at 0.521 of the mixture across all five grades.

🧮Checking the authors' arithmetic — and it is right

Their theoretical claim is that reducing 1 gramme of potassium chloroplatinite to metallic platinum needs 0.69 gramme of ferrous oxalate, corresponding to 0.9 gramme of ferric oxalate. Modern stoicheiometry says two iron(II) per platinum(II), because platinum(II) takes two electrons:

2 [Fe(C₂O₄)₂]²⁻ + [PtCl₄]²⁻ → 2 [Fe(C₂O₄)₂]⁻ + Pt + 4 Cl⁻

Two iron(II) per atom of platinum

1 g ÷ 415.11 = 2.409 mmol

Platinum salt, K2PtCl4, FW 415.11

2 × 2.409 mmol × 143.86 = 0.693 g

Iron(II) oxalate needed, at two per platinum, FW 143.86

2.409 mmol × 375.76 = 0.905 g

Ferric oxalate that would make it, as Fe2(C2O4)3, FW 375.76

0.693 against their 0.69, and 0.905 against their 0.9. The 1882 calculation is the 2017 calculation, and it uses the anhydrous formula weight — which matters, because Ware lists formula weights for ferric oxalate from 375.76 to 483.84 depending on which hydrate a supplier sends. Their arithmetic is exact and their substance is not, and that gap is the standing weakness of every ferric oxalate recipe ever published.

Now check the normal mixture against their own experimental finding. Taking Ware’s conversion of “1 part in 6” as about 15.6 per cent w/v, 24 c.cm. of platinum solution carries 3.74 g of the salt, and 22 c.cm. of a 20 per cent iron solution carries 4.40 g of ferric oxalate:

4.40 g ÷ 3.74 g = 1.18

Ferric oxalate per gramme of platinum salt in the normal mixture

Their measured range was 1 : 0.9 to 1 : 1.5, and their best results came at 1.0 to 1.2 parts of iron salt to one of platinum. The published mixture sits at 1.18 — the top of their own band, which is where they said to be, because “it will be always better to obtain a slight excess of ferrous oxalate”.

⁇One figure in this book cannot be reconciled with the rest of it

The theoretical half of the book asks for 0.02 to 0.025 gramme of potassium chloroplatinite per 1,000 square centimetres of surface for deep blacks. The practical half coats a sheet of 50 × 66 cm — 3,300 square centimetres — with 10 cubic centimetres of the normal mixture, and a footnote states that this is “according to the proportion of platinum and ferric salt per unit (1,000 sq. centim.) of surface, as obtained on theoretical considerations”. Both figures are printed identically in the French edition of 1883, so this is not a translator’s slip.

24 c.cm. × 0.156 g/c.cm. = 3.74 g in 50 c.cm. = 0.0749 g per c.cm.

Platinum salt in the normal mixture

10 × 0.0749 = 0.749 g per 3,300 cm² = 0.227 g per 1,000 cm²

Coating at 10 c.cm. per 3,300 cm²

That is about ten times the theoretical figure, and the two cannot both be right. Diluting the mixture with an equal volume of water, which the book allows for some subjects, only halves it.

Which of the two is out of step is decidable from outside the book. Ware reads Willis’s patent 1117 of 1880 as putting between 1.7 and 4 grains of potassium chloroplatinite on a square foot, Willis preferring the larger; 4 grains per square foot is 0.279 g per 1,000 cm². Ware’s own modern sensitiser is 0.225 molar in platinum at a specific coating volume of about 20 cm³/m², which is roughly 0.19 g per 1,000 cm². Pizzighelli and Hübl’s coating footnote lands at 0.227, between the two. Their theoretical figure lands an order of magnitude below anything anyone has ever coated a sheet with, and a print made at 0.02 g per 1,000 cm² would have almost no black in it.

The course prints both figures as the book prints them and corrects neither. The reading that reconciles everything is that the decimal point has moved — 0.2 to 0.25 g per 1,000 square centimetres, or equivalently 0.02 to 0.025 g per 100 — but that is a reading, and a reading is not a transcription. docs/FORMULARY-SCHEMA.md already lists “a quantity whose units are wrong in the only printing there is” as an open gap, with Kodak’s G-23 as its example; this is the second instance, and the second time the decision has had to live in prose because there is no field for it.

⁇Ware attributes a strength to these authors that neither printing carries

In §6.2 of the Platinomicon Ware describes his standard 25 per cent w/v ferric oxalate stock as “the same as originally specified by the analysis of Pizzighelli and Hübl”. Both printings of the book specify 20 grammes of ferric oxalate per 100 c.cm., which is 20 per cent w/v, and the footnote Ware attaches to the sentence is about a different formula altogether — the “time-honoured” 15 g in 55 ml recipe, which he calculates at 24 per cent.

The course follows the primary text and records the disagreement rather than choosing. Nothing much rests on it photographically: the two strengths differ by a quarter and the oxalic acid addition differs too, since Ware standardises on 2 g per 100 cc where the book asks for 6 to 8 per cent of the ferric oxalate, or 1.2 to 1.6 g per 100 cc at this strength. What it illustrates is worth more than the number — a figure repeated through a chain of secondary sources drifts, and the only cure is to open the original.

The order matters in one place and the book says so. The iron solution is diluted to strength first and acidulated after, because the amount of free acid to add depends on how much was already there, which the analysis has just measured. Everywhere else the order is unremarkable.

Two purity tests are given for a bought iron solution and both are still good practice. It must not turn blue with red prussiate of potash — potassium ferricyanide, which detects iron(II) and so detects a solution that has already begun to reduce. And boiled after dilution with ten volumes of water it must not go turbid, which detects basic ferric oxalate. For the platinum salt the tests are that one part must dissolve completely in about six parts of cold water, and that the solution must not be acid.

Behaviour

Recorded from the source so that the system can be understood, not so that it can be run.

Almost nothing happens during the exposure, and what happens is the wrong colour. The yellow coating turns brown, and on longer exposure lightens again, so that “often the deepest shadows appear lighter than the darker halftones”. That is partial reversal, familiar to anyone who has over-exposed a modern siderotype, and it is why the authors say platinum printing needs more care in judging progress than silver printing does.

Then the image arrives at once. In the hot oxalate bath “the brown colour changes immediately to a deep black”. There is no visible development to watch and nothing to time; the print is drawn through the bath and it is done. The Getty atlas describes the same moment — the platinotype image “appears almost immediately”.

Temperature is the control the authors actually use. A cooler bath for a print that has been over-exposed; the hot bath for everything else, which is the one they say gave them the best result in ordinary use; and a boiling bath for a print exposed too briefly. The developer is not there to control density — it is there to move iron, and how fast it moves the iron decides how much platinum gets reduced before the salts wash away.

The whites are made in the acid bath, not in the developer. After development the print still holds unreduced iron salts, and until they are out the highlights are not white. Three changes of dilute hydrochloric acid, about ten minutes each, until the bath stops turning yellow. The Getty atlas identifies the failure of this step in surviving prints: platinotypes not completely cleared of ferric salts “exhibit a slight or more pronounced yellowing or reddish tonality of highlights”.

A wet platinum print lies about its tone. “Platinum prints in a wet state appear always more brilliant and lighter than they do when dry. A print, therefore, which while still wet after development seems to be quite right as regards tone, would be too dark when dried.” That is dry-down, described in 1882 and still the commonest reason a first platinum print disappoints.

Damp paper is the enemy of the whole system. Moist paper gives weak and foggy images and is less sensitive; the authors keep sensitised sheets and finished exposures in a calcium chloride box, and put a waterproof cloth behind the paper in the printing frame. Ware records the commercial answer to the same problem — platinotype paper was sold in sealed tins with a desiccant capsule.

Image characteristics

The surface is the identifying feature and it belongs to the paper. There is no binder. The platinum sits in the surface fibres of the sheet, so the print is, in the Getty atlas’s words, “deeply matte”, showing no gloss even when rotated against a light. Everything a reader knows about judging a gelatin print’s surface is inapplicable.

Colour is set by the sizing, and the authors measured it. Gelatine-sized paper gives a bluish-black; arrowroot gives a browner black; and the difference shows more on smooth paper than on rough. The AIC’s process notes confirm the direction from the conservation side and add the other levers: gelatin sizing, oxidising chemicals and more acidic processing all push a platinum print warmer. This formula has all three available, which is why two prints from the same negative on the same paper can differ in colour without anything having gone wrong.

The tonal scale is the reason the process survived its own cost. The AIC records that platinum and palladium give a wider scale of greys and less contrast than silver, “revered for the nuances and subtleties in the prints”. The unmodified sensitiser is the long-scale end of that, and every part of chlorate added trades some of it away.

What chlorate costs, stated by the people who dislike it. Ware records two complaints against the reagent and both are about appearance rather than safety: it is reported to increase image graininess, and the “false sparkle” of lost gradation in emptied highlights is, for the platinotype aesthetic, the wrong kind of brilliance. His own web article puts the mechanism of the objection precisely — an oxidant does not contract the tonal scale uniformly, it truncates the high values.

Permanence is the process’s own argument and it is largely sound. The authors list it second among four advantages and contrast it with the “doubtful permanence” of silver prints. The Getty atlas agrees that platinum is the most chemically stable metal found in nature, and adds the failure mode that is not the metal’s fault: residual iron yellows the highlights, and a platinum print stored in contact with another sheet can catalytically transfer a reddish-brown ghost of itself onto it.

The mechanism

Nothing in this formula is sensitive to light except the iron. The platinum salt is a passenger. The chain runs: ultraviolet reduces iron(III) to iron(II) in the coating; the developer dissolves the iron(II) so it can move; the dissolved iron(II) reduces platinum(II) to platinum metal; the acid bath removes what is left of the iron.

Step one, the photolysis. Döbereiner observed it in 1831. Ultraviolet absorbed near the ligand-to-metal charge transfer band transfers an electron from a coordinated oxalate ligand to the iron(III), reducing it, and the oxidised ligand is lost as carbon dioxide:

hv + Fe₂(C₂O₄)₃ → 2 FeC₂O₄ + 2 CO₂

Ferric oxalate: the photoproduct is iron(II) oxalate

Step two is the one this formula cannot avoid, and it is why there has to be a developer at all. Iron(II) oxalate is nearly insoluble — Ware gives 0.022 g in 100 cc of water — so the iron(II) that light has just made sits where it was made and cannot reach the platinum salt a few micrometres away. Concentrated potassium oxalate dissolves it by complexation:

FeC₂O₄ + C₂O₄²⁻ → [Fe(C₂O₄)₂]²⁻

What the developer is for

That is the whole function of the oxalate bath. It reduces nothing. The authors could not have written it this way in 1882, but they had the observation exactly: hot water alone “will partially reduce the platinum salt”, and the developers they sought were substances “able to increase the action of the ferrous oxalate”. They were looking for a solubiliser and they knew it by its effects.

Step three, the reduction. Two of the soluble iron(II) complexes give up an electron each to one platinum(II), which becomes an atom of metal:

2 [Fe(C₂O₄)₂]²⁻ + [PtCl₄]²⁻ → 2 [Fe(C₂O₄)₂]⁻ + Pt + 4 Cl⁻

The image

Ware gives the redox potential of the iron couple as +0.02 V, which is enough to reduce platinum and palladium and not much else — which is the reason siderotype works with noble metals and not with base ones.

And step three is a race. The moment the sheet enters the developer, the platinum salt starts washing out of the paper as well. Everything the authors say about heat and concentration is about winning that race: “if the reduction is slow, both the platinum and ferric salts might be dissolved before they had time to act on one another”, in which case the reaction “would take place only in the developing liquid, and not on the support of the image” — the platinum reduces in the tray instead of in the print.

⚗What the chlorate does, in their words and in modern ones

Their explanation runs: chlorate added to the sensitiser does nothing while the solution is liquid, but as the sheet dries, unstable ferric chlorate forms and decomposes, oxidising some of the potassium chloroplatinite to the chloroplatinate — a platinum(IV) salt, which “as a platinic salt, alters the character of the image (it makes it harder)”. They add the constraint that decides the whole design: the oxidant must not act in the bottle, because the chloroplatinate is far less soluble and would crystallise out unevenly.

Ware’s account is of the same reagent doing its work one step earlier — chlorate reoxidises some of the iron(II) photoproduct back to iron(III), making it unavailable to reduce platinum and so truncating the exposure scale. The two accounts are not rivals; a strong oxidiser in a drying film containing both an iron(II) salt and a platinum(II) salt can act on either, and both actions raise contrast by removing weakly-exposed image-forming capacity. The course states them as two published mechanisms and does not pick one.

The authors’ sensitivity figure is the practical point either way: “the presence of even one-hundredth per cent. of that substance has a noticeable effect on the image”, and the addition “will, as may naturally be conceived, lower the sensitiveness”. Their four mixtures run from zero to about 0.176 per cent w/v — up to eighteen times a dose they call noticeable.

Why the excess of iron behaves as a restrainer. Too much ferric oxalate gives weak, grey prints, which is not obvious: more of the light-sensitive salt ought to give more image. Their explanation is physical rather than chemical — unreduced ferric salt “settles itself in the interstices between the particles of ferrous and platinum salt, and this remainder acts as a restrainer during the development”, getting in the way of a reaction that has to happen between two solids a few micrometres apart. It is a mechanism a modern reader would call a diffusion barrier, and it is the reason the ratio has a maximum as well as a minimum.

Function of every ingredient

Potassium tetrachloroplatinate(II), 1 part in 6 of distilled water. Potassium chloroplatinite in the book’s language, K₂PtCl₄, FW 415.11. It is the metal that becomes the picture and it is not sensitive to light in any useful degree: the authors note that the change in the platinum salt during an exposure short enough to be practical is “not appreciable by the eye”. Its job is to sit in the paper and wait. More of it means a deeper maximum black — the authors say increase it on porous supports, where part of the image sinks below the surface and is lost — and less of it is their own recommendation for a print that is to be retouched or painted over. What it interacts with is the iron, and only in the stated ratio: too little platinum for the iron present and the excess iron restrains development; too much and there is not enough iron(II) to reduce it, which shows as grey shadows. The authors also record that it can be replaced entirely by palladium — palladium prints come out sepia-brown, tone violet in gold and solarise more readily — but reject it on cost, which is an irony worth noticing, since palladium is now the cheaper metal and the course’s own choice. They tested iridium and got nothing, and rejected gold because oxalate reduces its compounds in the cold.

Ferric oxalate, 20 g per 100 c.cm. — the normal ferric solution. Fe₂(C₂O₄)₃ as the book writes it. This is the only light-sensitive substance present and the only thing an exposure changes. More of it, past about 1.2 parts to one of platinum, gives weak grey prints by the restrainer mechanism above; less of it, below about 0.9, leaves platinum unreduced and gives the same grey shadows for the opposite reason. The substance is also the weak point of the formula and the authors say so: it has no fixed composition, cannot be replaced by another iron salt, and the only honest way to use it is to analyse the solution and dilute to a stated strength. Ware’s modern account vindicates the complaint — ferric oxalate is polymeric, polymorphic, apparently uncrystallisable, its formula weight uncertain between 375.76 and 483.84, and its solutions change with time. Everything Ware’s ammonium system does differently starts from replacing it.

Oxalic acid, 6 to 8 per cent of the ferric oxalate present. About 1.2 to 1.6 g per 100 c.cm. at this strength. It has two jobs and both are about keeping iron(III) in solution and in the right oxidation state. First, excess oxalate suppresses the hydrolysis of iron(III) to basic salts; the authors’ whole objection to a neutral sensitiser is that basic iron salts give the paper “a yellow colour which does not disappear”. Second, acid keeps the finished coating acid, which is the condition they state for both the sensitiser and the developer. Too little and the yellow stain appears and cannot be washed out. Too much is also a fault and they identify it specifically: “excess of acid in the sensitizing solution prevents complete reduction”, and they observed the effect when 8 per cent of oxalic acid was mixed into the sensitiser — which places their 6 to 8 per cent band immediately below a threshold they had actually found. Modern practice is looser: Ware records amounts from 0 to 5 g per 100 cc and standardises on 2 g.

Potassium chlorate, 0.4 g per 100 c.cm. of the iron solution — the normal chlorate of iron solution. The contrast agent, and the reason this entry exists as a system of bottles rather than a single recipe. It is present in the chlorate stock at 0.4 per cent w/v, which is 2 per cent of the weight of ferric oxalate beside it, and it reaches the coated sheet at between zero and about 0.176 per cent w/v depending on the mixture. More of it hardens the image, and past a point makes it hard in the pejorative sense — the fault list’s entry for hard prints gives too much chlorate as one of two causes. Less of it returns the long soft scale. Two costs travel with it and neither is optional: sensitiveness falls, because some of the photoproduct is being destroyed as fast as light makes it; and the image degrades, Ware recording increased graininess and the emptied high values he calls false sparkle. It interacts with the platinum salt directly, by oxidising it to the much less soluble chloroplatinate, which is why it must not act until the film is drying.

The normal ferric solution, used as an ingredient of the chlorate bottle. Worth stating explicitly because it is the structural trick of the formula: the chlorate stock is not a different chemistry, it is the same iron solution with an oxidant dissolved in it. That identity is what lets the two bottles be swapped against each other in any proportion while the iron, the platinum, the acid and the coating volume all stay exactly where they were.

Distilled water, 4 parts in every mixture, and it is not filler. The book’s footnote gives the reason: ordinary spring water contains lime, and calcium ions precipitate calcium oxalate, clouding the sensitiser. The same reaction is the modern rule against calcium carbonate-buffered papers in siderotype. The 4 parts are also the adjustment the authors reach for last — omitted entirely on hard-sized hot-pressed sheets that will not absorb, doubled or trebled when a print is not meant to have a black.

⚗The ingredients of the process that are not ingredients of the sensitiser

Four more substances are needed to make a print with this formula, and none of them is in the mixture. They are described here because a page that explains only what is in the bottle explains half the chemistry.

Gelatin or arrowroot, as a 1 per cent size. Ten grammes in 800 c.cm. of water with 200 c.cm. of alcohol; the gelatine bath also takes 3 g of alum, which hardens the layer so it survives the hot developer, and the alcohol is there to stop air-bells. The purpose is to partly close the pores — not to seal them, because an image lying entirely on the surface washes off, and not to leave them open, because an image that sinks in loses its blacks. Highly sized rough papers get 10 to 20 per cent dextrine instead, which washes out again in processing; the encyclopaedia has no entry for dextrine or for the algeine the authors also name.

Potassium oxalate, saturated in the cold and acidulated with oxalic acid, at 80 °C. The developer, whose mechanism is above. It was chosen from sixteen candidates and the ranking is itself a finding — see Variants.

Hydrochloric acid, 1 part in 80. The clearing bath. It dissolves the remaining iron salts, which are what would otherwise yellow the highlights. Three changes, ten minutes each, to an endpoint you can see.

Ferrous sulphate, nitric acid and ammonia, for the residues. Not part of making a print, but part of the published method, because platinum was too expensive to throw away. See Waste.

Interactions

Acid against basic iron, which is the axis the whole formula turns on. Both solutions must be distinctly acid. Below that, iron(III) hydrolyses, basic salts form in the paper, and the print yellows permanently. The authors’ developer instruction is a maintenance rule for the same fact: evaporation from a hot bath deposits potassium oxalate crystals on the vessel, overheated crystals decompose to potassium carbonate, the bath drifts alkaline, and the remedy is litmus and oxalic acid.

Free acid against reduction, in the other direction. The same acid that keeps the iron in solution, in excess, “prevents complete reduction”. Their observation at 8 per cent is the ceiling on their own 6 to 8 per cent band. Free acid in the developer they describe as much less injurious — which is consistent, since by then the reduction has to happen fast rather than not at all.

Reducing substances against the iron. Formic acid and hyposulphite in either solution reduce ferric oxalate directly and fog the image. This is why the sensitiser and the fixing chemistry of a silver darkroom cannot share a bench, and it is the ancestor of the modern rule about dedicating tools per process.

Oxidising substances against the platinum. Covered above for chlorate. The generalisation the authors draw is the useful part: an oxidant that acts in the bottle is useless because the chloroplatinate it makes is insoluble and cannot distribute itself evenly, while an oxidant that acts only as the film dries is a contrast control. The choice of chlorate is a choice of rate, not of strength.

Sodium chloroplatinate as an alternative to chlorate entirely. A footnote offers it: use the sodium chloroplatinite in place of the potassium salt, dispense with the chlorate, and substitute sodium chloroplatinate directly for part of the platinite. That is the same platinum(IV) the chlorate was being used to manufacture in situ, bought instead. Ware records the same reagent still sold for the same purpose at a very high price and said to degrade the image less — and records the trap, that it cannot be used with any ammonium ion present, because ammonium hexachloroplatinate(IV) is barely soluble.

Iron(II) contamination against everything, and the strangest instruction in the book. A ferric solution that has begun to reduce is detected with red prussiate of potash. The authors’ remedy is to add the chlorate stock — a few cubic centimetres per 100 — to reoxidise the trace of iron(II), then test on paper whether the restoration is complete. It is the clearest possible confirmation of what the chlorate is doing, since here it is used deliberately as an oxidising agent on iron and nothing else.

Ultramarine against hydrochloric acid. Papers blued with ultramarine turn yellow in the clearing bath, and the authors tell the reader to test the paper before buying it; cobalt-blued papers are recommended instead. It is a paper-selection rule that reads as quaint and is not: the modern equivalent is testing a sheet for buffering, for optical brightener and for iron contamination before a batch is committed.

Calcium against oxalate. Spring water clouds the sensitiser with calcium oxalate. The same reaction, in the paper rather than the bottle, is why alkaline-buffered papers are excluded from siderotype.

Variants

The four mixtures on this page are themselves the variants, and the authors present them that way: one formula with a contrast control, not four formulas. The parts table above is the whole variant system.

The developer is where the real alternatives were. Part I ranks sixteen substances by reducing power, tested by cutting a strip of exposed paper and developing pieces of it in each. In their order: sodium acetate, which is energetic and soft but yellows from basic ferric acetate; sodium acetate with oxalic acid, which behaves identically and gives pure whites; potassium citrate; then a group they call indistinguishable — sodium carbonate hot, neutral and acid ammonium citrate, potassium oxalate, Rochelle salt, ammonium benzoate and sodium succinate; then a distinctly inferior group of dilute caustic potash, sodium carbonate and sodium phosphate; and finally hot water, ammonia and oxalic acid, which develop only feebly. Except for the alkalis, all are used hot at 80 °C and concentrated. They discard the citrates as too expensive, discard four more for basic iron salts or feebleness, and recommend three: acetate with oxalic acid, potassium oxalate and Rochelle salt — choosing potassium oxalate despite its higher price because photographers already had it for the ferrous-oxalate negative developer. The encyclopaedia has no entries for ammonium benzoate, sodium succinate or sodium phosphate; the rest are linked above.

That list is also a map of the alternative processes. Two of the three recommended developers survive today as kallitype developers — Rochelle salt and, from the same family, sodium citrate — and the oxalate became the platinum developer, published with a strength by Willis and used ever since.

Pizzighelli’s own 1887 variant is the important descendant and it is a different formula, not a modification of this one. Working alone in Bosnia, he replaced ferric oxalate with sodium iron(III) oxalate, whose photoproduct is already soluble, and got a printing-out platinum image that needed no developer at all — only clearing. Ware’s Table 5.1 shows two of the four stocks carried over unchanged from 1882, including the platinum at 1 part in 6 and the chlorate at 0.4 parts in 100. The Getty atlas dates it to 1887 and traces the line forward: Ware and Malde’s print-out platino-palladiotype of 1986 is explicitly a modern variety of it.

The variants this course records at Level D and does not describe as procedures. Hübl himself described mercury citrate for warm tones, and the Getty atlas records mercury(II) chloride in the developer as the standard route to sepia — the mercuric sepia platinotype has that entry. Willis’s first patent added silver or lead salts to the sensitiser, and the Getty atlas records lead oxalate as a contrast addition recommended by Pizzighelli and Hübl themselves — the lead contrast additions has that one. Neither is in the 1886 formulations above, and neither is a procedure this course publishes.

The modern industrial descendant is the three-solution drop system, whose two iron bottles are these two iron bottles. Ware traces the transmission through Paul Anderson’s American republication of 1937 to Henny and Dudley, Wall and Jordan, Gassan and Crawford, and thence to the kits sold by Photographers’ Formulary and Bostick & Sullivan.

Safety

Level D, on four grounds. The first two are shared with the modern drop system and the argument there is not repeated at length; the second two belong to this formulation and not to the kits.

Platinum, for irreversibility. Potassium tetrachloroplatinate(II) carries H334 — may cause allergy or asthma symptoms or breathing difficulties if inhaled — in 99.5 per cent of the 212 notified reports the encyclopaedia entry aggregates, with H301, H315, H317 and H318 beside it. HSE’s EH40 puts halogeno-platinum compounds, defined in its paragraphs 25 and 26 as coordination compounds in which the platinum atom is directly bonded to halide ions, at 0.002 mg/m³ as platinum with the Sen notation; platinum metal itself is at 5 mg/m³, a limit 2,500 times looser. Ware records where the hazard was first identified, and it is this exact material: platinum allergy was described in 1911 as an occupational disease of photographic factory workers handling platinotype paper. His practical rules are never to touch the surface of platinum-sensitised paper, never to put ungloved fingers in the processing solutions, and not to store large amounts of dried sensitised paper, because dust inhalation is the worst route.

Potassium chlorate, for the availability of a better route. Its page sets out the classification — signal word Danger, H271 as a strong oxidiser under the harmonised CLP entry, with the acute-toxicity statement genuinely disputed among notifiers. EH40 lists no chlorate at all, and its own introduction says absence from the list does not indicate that a substance is without risk. The deciding operation is weighing a dry oxidiser onto a bench carrying paper, cloth and dust, and the contrast it buys is available from a calibrated negative without the graininess.

The hot bath, which the kits do not have. This formulation develops in a near-boiling saturated solution of potassium oxalate — 80 °C as published, exceeded deliberately for under-exposed prints, and used boiling by the authors themselves. Oxalate is systemically poisonous: the encyclopaedia entry carries H302, H312, H315 and H319, and quotes CAMEO’s finding that inhalation of the dust can cause systemic poisoning, with ingestion producing burning pain, vomiting, a weak pulse and possible cardiovascular collapse. EH40 sets oxalic acid at 1 mg/m³ long-term and 2 mg/m³ short-term. A poison at 80 °C in an open tray, splashed by drawing sheets through it by hand, is a different proposition from the same poison cold — the authors’ own footnote about holding the print with bone or horn forceps “to prevent the hands from coming into contact with the liquid” is the period acknowledgement of it.

The preparations, which are laboratory operations and not darkroom ones. The book makes its own platinum salt by passing a strong stream of sulphur dioxide gas through a solution of platinic chloride held at 100 °C, judging the endpoint by spot tests on a watch glass and stopping precisely, because too much gas makes platinous sulphide and too little leaves platinic chloride. It makes its ferric oxalate from 500 g of iron(III) chloride and about 250 g of caustic soda. It recovers platinum by incinerating waste paper and digesting the ash in aqua regia — three parts concentrated hydrochloric acid to one of nitric acid — at 50 to 70 °C. Each of those is a fume-cupboard operation under supervision, and none of them is made safe by being historical.

What is not the reason, and it matters. Neither ferric oxalate nor the gelatine and arrowroot sizing nor the 1-in-80 hydrochloric acid decides this classification. Ferric oxalate is a Level B substance in this course; the sizing baths are food-grade materials and warm water; the clearing bath is a dilute acid of the kind Part II teaches a student to handle in their first week. Remove the platinum, remove the chlorate, put the development in a cold citrate bath and what is left is a Level B process, which is exactly what the palladiotype is and exactly what this course teaches instead.

No control is described above and none is offered. Level D is not a judgement that the hazards here are unmanageable in a laboratory that manages such things every day. It is a judgement that a page written for a reader whose bench is a kitchen table cannot assume the extraction, the balance enclosure, the supervision or the health surveillance that managing them takes.

Storage

What the source records.

The platinum keeps forever and needs nothing. “Both the salt and its solution resist perfectly the action of light and of the atmosphere, and no special precautions are therefore requisite in preserving them.” Modern practice agrees; Bostick & Sullivan say their metal solutions do not oxidise or age. The one caution the book adds belongs to the preparation rather than the store: the freshly precipitated salt is dried in a room the light cannot reach, because a platinum salt wet with alcohol reduces readily in light.

The iron does not keep and must be kept dark. “Both the ferric solutions must be kept away from the light”, in a hyalite flask — dark glass. Any quantity may be made at once and held in stock, which is a claim about convenience rather than stability, and the fault list’s remedy for a spoiled iron solution shows the authors did not entirely trust it. Ware’s modern figure for the same substance is six to nine months in the dark “said by some”, with at least one respected printer making it fresh the night before every session.

The mixed sensitiser does not keep at all, and this is the strongest storage statement in the book. It “will gradually decompose, even though kept carefully in the dark”, so it “should be all used up as soon as possible” and only as much should be mixed as is directly required. The flannel coating pad is changed every quarter of an hour for a related reason: it is exposed to weak daylight while in use, the iron it holds reduces, and a pad carrying iron(II) leaves spots and streaks on the next sheet.

Sensitised paper is stock, but only for weeks. Six to eight and sometimes more; after that it goes both weak and foggy even in the dark, and the recommendation is to coat three or four weeks’ worth at a time. Sheets and exposed prints alike live in a calcium chloride box, whose construction the book illustrates and whose desiccant is checked and re-dried when it goes damp. The encyclopaedia has no entry for calcium chloride.

Sized paper is real stock. Once sized and dried it “will keep for any length of time”, protected from dust and moisture. That is the one part of this workflow a modern printer would recognise without change.

Nothing described here is bought or kept on this course’s account. Recording how a nineteenth-century printer stored a bottle is not advice to acquire one, and the classification above governs the whole entry.

Incompatibilities

Potassium chlorate against combustible material, which is most of its hazard profile. A dry strong oxidiser and a bench of paper, cloth and dust is the combination its classification exists to warn about, and it is worse in a darkroom than in a laboratory because a darkroom is made of paper.

The sensitiser against anything reducing. Formic acid, hyposulphite and their relatives reduce ferric oxalate directly and fog the coating. In a shared darkroom that means thiosulfate fixer, sulphite, bisulphite and metabisulphite — the whole silver-processing side of the room — must not reach the coating bench, the pad, the glass or the water.

Iron solutions against light, which is not usually filed as an incompatibility and behaves like one: the reagent is destroyed by the thing the process uses.

The sensitiser against calcium. Spring water clouds it with calcium oxalate; buffered papers do the same thing inside the sheet. Distilled water and unbuffered paper are both requirements rather than refinements.

Platinum salts against contact with metal, a modern caution rather than a period one: the noble metal plates out on base metals, which is why the suppliers insist on glass and on a dedicated pipette per bottle. The period text has its own version of it — black spots with comet-like tails caused by “particles of metal imbedded in the substance of the paper, causing a reduction of the platinum”.

Ultramarine-blued paper against hydrochloric acid. The paper yellows in the clearing bath. Test before buying.

Waste

Every solution on this page is a noble-metal-bearing or iron-bearing waste and none of it goes to drain unconsidered. The course’s general position is in the disposal guidance: the chemistry is explained here, local regulation governs what may actually be done, and no jurisdiction-specific instruction is given as though it were universal.

The book’s own recovery methods are historically interesting and are given as history. Spent developer, once so loaded with iron that crystals separate or the colour turns dark yellow, is mixed with about a quarter of its volume of saturated ferrous sulphate solution and boiled: metallic platinum drops out and is filtered off, and the filtrate — now iron(II) oxalate — is converted to potassium oxalate exactly as the negative developer of the period was. Waste paper, linen and flannel carrying platinum are collected and incinerated; the ash is digested in aqua regia at 50 to 70 °C, diluted, filtered, and the platinum precipitated with ammonia as ammonium chloroplatinate and ignited to the metal. The recovered metal is boiled in hydrochloric acid to strip residual iron before being taken back to the chloride with fresh aqua regia.

Why that is not a procedure this course offers. Aqua regia digestion and the ignition of platinum salts are specialist laboratory operations, incineration of coated paper is not a domestic possibility, and the whole sequence is being described for a metal this course does not have its students handle. What the passage does teach is the economics: platinum was valuable enough in 1882 that a photographer was expected to burn their failures and get the metal back, and the modern equivalent is that spent noble-metal baths are collected for recovery rather than discarded.

Iron and oxalate. The clearing baths and the spent developer carry dissolved iron and a large quantity of oxalate, which is toxic to aquatic life and is the reason the alternative-process waste plan keeps an iron-bearing stream separate from the silver-bearing one.

The one waste stream that is genuinely benign is the spent sizing bath: dilute gelatine or arrowroot with a little alum and alcohol.

Troubleshooting

The seven faults the authors publish, with their own causes and remedies, read against what is now known. They are given as an aid to reading period prints and period manuals, not as bench guidance for a process this course does not run.

Fog, or a grey deposit in the whites. Four causes are offered: light reaching the paper during sensitising, coating or examination in the frame; drying above 40 °C; a spoiled iron solution carrying iron(II); and over-exposure. Three of the four are the iron being reduced by something that is not the image-forming light, which is the right generalisation and the one to take away.

Weak prints under the developer. Damp paper — “paper once spoiled cannot be made good again” — or paper more than six or eight weeks old, or a weak negative, for which the remedy given is more chlorate. The modern remedy is a better negative.

Vigorous when wet, weak when dry. Paper insufficiently sized, so the image sinks into the substance; or drying after coating that took longer than ten minutes, with the same result. This is dry-down compounded by a real loss of surface image, and it is the fault most likely to mislead a beginner because the print looked right in the tray.

Yellow whites after drying. Insufficient acidity in sensitiser or developer, or insufficient time in the hydrochloric acid, or ultramarine in the paper. The first two are the same fault the Getty atlas identifies in surviving prints as residual iron, and the endpoint the authors give is a test rather than a clock: change the acid until it no longer turns yellow in ten minutes.

Hard prints. Under-exposure, or too much chlorate. The remedy “stands to reason”, says the book, which is the only place in it where the tone is impatient.

Spots and streaks. Dirty brushes, wet fingers on the paper, dirty glass, unclean vessels — and, though it is filed separately, the flannel pad that has been in use more than a quarter of an hour.

Black spots with comet tails. Metal particles embedded in the paper reducing the platinum, or insoluble impurities in the platinum salt; the remedy for the second is to filter the sensitiser. Ware records that the “plague of black spots” is a known affliction of the ferric oxalate development process specifically, and that the ammonium print-out system appears to be immune to it — which is a modern answer to an 1882 fault whose cause was never fully identified at the time.

Experiments

Written for the palladium chemistry the course does teach, not for this formula. Every one of them asks a question this page raises and none of them requires platinum or chlorate.

Does the two-bottle trick actually hold everything else constant? The claim above is that trading chlorated iron against plain iron changes contrast and nothing else. Mix a palladium sensitiser at a fixed metal and iron total, coat a step wedge strip at each of three oxidant levels — using the course’s permitted route, which is contrast from the negative rather than from an oxidant, this becomes a comparison of three negatives against one sensitiser — and measure D-max as well as the exposure scale. If D-max moves, something other than contrast has changed.

How much does a coating volume really vary? The arithmetic in the Mixing section turns entirely on 10 c.cm. per 3,300 cm². Coat six sheets of a known area with a weighed syringe of water-thin liquid, weigh each sheet before and after, and compute the spread. A method that claims a specific coating volume is only as good as the coater’s repeatability, and this is the cheapest measurement in the whole of alternative process work.

What does sizing do to image colour? The authors say gelatine gives a bluish-black and arrowroot a browner black, more visibly on smooth paper than on rough. Size two halves of the same sheet differently, coat both with the same palladium sensitiser from the same syringe, print them under the same negative in the same exposure, and measure. It is one of the very few claims on this page a student can test directly.

Does a wet print really lie by a measurable amount? Measure reflection density on the same patches wet and dry, on paper sized two ways. Dry-down is universally reported and rarely quantified, and a number for your own paper is worth more than a rule of thumb.

Read a print rather than make one. Find a platinotype in a museum’s online collection with a usable image file, and describe what you can and cannot conclude from it: the matte surface, the colour, whether the highlights are yellowed, whether there is a transferred ghost on the facing sheet. Then write down which of your conclusions are diagnostic and which are merely consistent — the Getty atlas’s own warning is that without chemical analysis, platinotype, palladiotype, kallitype and Van Dyke are very hard to tell apart.