Mercuric sepia platinotype

For about forty years the answer to “how was this brown platinum print made?” was, in most of the printed literature, “with a hot developer”. It is the wrong answer. A hot bath warms a platinum image only very slightly; what made a sepia platinotype brown was mercury, and the prints still contain it. X-ray fluorescence has since found mercury in essentially every early sepia platinotype that has been looked at, and in a good many that were catalogued as black.

☠Level D: this page is not a procedure

Mercury(II) chloride is classified for historical study only, and the Level D policy governs what this page may contain. It may give the composition, the chemistry, the historical workflow in outline, the image characteristics, the identification evidence and the hazards. It may not give a procedure, and it does not: no coating routine, no exposure, no development time, no clearing sequence and no washing schedule appears below as an instruction. The temperatures in the data are historical observations of what the commercial papers needed, not settings to work to.

The composition is published because a formula that cannot be read cannot be identified, and identification is the main use a reader has for this entry. The course asks nobody to weigh any of it. There is no protective equipment list that makes a heated mercuric chloride bath acceptable in a domestic darkroom, and none is offered.

Mike Ware’s own verdict, from the chemist who reproduced the chemistry and measured it: “the lesson for contemporary platinum printers is clear: do not over-use mercury(II) additives.” The course goes one step further and does not use them at all. The platinum and palladium printing this course does teach is on the three-solution drop system and the potassium oxalate developer pages.

The coating solution, with mercuric chloride in place of the plumbic chloride — the light-sensitive coating, carrying both the iron and the first dose of platinum

Ingredient	Quantity	Form the source specifies
Ferric oxalate	4.54 g	70 grains, at 0.0648 g per grain: 15.96 per cent w/v, or 0.425 molar in Fe2(C2O4)3 at a formula weight of 375.75. Willis adds "with enough oxalic acid to render this ferric oxalate freely soluble" and states no quantity for it, so the oxalic acid is not in this table; see Mixing.
Potassium tetrachloroplatinate(II)	0.972 g	15 grains: 3.42 per cent w/v, or 0.0824 molar at a formula weight of 415.1. Willis writes "potassic chloro-platinite", the platinum(II) salt, and names the iridium and platinum(IV) salts he would substitute for it.
Mercury(II) chloride	0.259 g	4 grains: 0.912 per cent w/v, or 0.0336 molar at a formula weight of 271.5. Both figures are Ware's own conversion of this line of this patent, and they agree with the conversion table in Appendix VIII, which is the arithmetic check that the grain and the fluid ounce have been read the same way here as there.
Water	to make 28.4 mL	One imperial fluid ounce, at Ware's 28.413 cm3. Willis's line is "each fluid ounce of this solution will contain", a make-up volume and therefore a strength; any volume in the same proportions is the same coating solution.

The developing solution — the bath that dissolves the iron(II) image and supplies the second dose of platinum

Ingredient	Quantity	Form the source specifies
Potassium oxalate monohydrate	7.78 g	120 grains: 27.36 per cent w/v, or 1.485 molar at a formula weight of 184.23. The same 120 grains per fluid ounce that Willis printed for the plain platinotype developer in his 1880 patent, and the strength this course records on the potassium oxalate developer page. The two patents word it differently and the difference matters: 1880 says "dissolving 120 grains or more of the salt in one fluid ounce of water", a volume of water added, and 1878 says the solution "contains" that much per fluid ounce, a make-up volume.
Potassium tetrachloroplatinate(II)	0.454 g	7 grains: 1.596 per cent w/v, or 0.0384 molar. Adding a platinum salt to the oxalate bath is what Willis's 1878 claim is actually about; the patent's own words are that the developing solution's use "constitutes the novel and essential feature of my Invention".
Water	to make 28.4 mL	One imperial fluid ounce, at Ware's 28.413 cm3, and again a make-up volume: "Each fluid ounce of this developing solution contains 120 grains of potassic oxalate and 7 grains of potassic chloro-platinite."

Used in this order — the process as the 1878 patent states it

1. The coating solution, with mercuric chloride in place of the plumbic chloride — Coated on the paper and dried, then exposed under the negative "for a sufficient time, which will usually be indicated by the image or picture becoming faintly visible". The patent gives no exposure and no light source.
2. The developing solution — Not stated. The patent says only that the solution "quickly deepens in color or blackens the picture or image produced on the coated paper by the action of light" — The exposed paper is floated face down on the solution, or immersed in it, "preferably hot". Ware puts the working temperature of this route at 140 to 160 F, that is 60 to 71 C. A wash in a weak acid, preferably oxalic, and then in plain water, follows; the patent states no strength or duration for either.

I coat the same with an aqueous solution which I term the coating solution ... I then dry the paper and expose it to light under or behind the negative ... I then float the said paper with its face or coated surface downwards on or immerse it in an aqueous solution, preferably hot, which I term the developing solution ... I then wash the paper in a weak solution of some acid (preferably oxalic acid), and finish by washing it in plain water.

This is a sequence of two treatments rather than a mixture: the two solutions never meet in a vessel, they meet in the paper. That is the structural difference between a platinotype and every developer in this formulary, and it is the reason the mercury can be put in either one of them and give a different result each time.

!What this table is, and what it is not

It is one patent, converted from grains and fluid ounces and otherwise untouched. British Patent No 2800 of 12 July 1878 prints the coating solution with “preferably two grains of plumbic chloride” and then, four paragraphs later, offers the substitution this entry records: “I substitute for the two grains of plumbic chloride 4 grains of mercuric chloride.” That sentence is the origin of the sepia platinotype, and the four grains per fluid ounce is the earliest mercury dose for a platinum sensitiser that any source read for this page prints.

It is not the commercial sepia paper. Willis later found a way of carrying the mercuric chloride in the coating that avoided the split tones of his first attempts, and he never published it. The Platinotype Company’s “Special Sepia Solution” and “Sepia Crystals” were trade secrets; Ware’s judgement that both “likely” contained mercury salts is an inference and is marked as one here.

It is not the only published mercuric sepia formula. Willis’s own patent of November 1887, Wall’s glycerine developer of 1912, Wall’s two sepia sensitisers of 1924 and Hübl’s mercuric citrate of 1902 all print quantities, all differ, and all are set out under Variants.

Purpose

To record the composition of the first published mercuric platinum sensitiser, to explain why a mercury salt turns a neutral platinum image brown, and to make a class of print in the world’s collections identifiable. That is the whole of the entry. It publishes no procedure and asks nobody to carry one out.

There is a second reason it earns its place, and it is a lesson about evidence rather than about chemistry. Between about 1900 and 1930 the printed literature contained three incompatible accounts of what a sepia platinotype was: that it was made with a hot developer; that it contained mercury but that the mercury was not in the image; and that it contained mercury which made it impermanent. The first was wrong, the second was wrong, the third was partly right for the wrong reason, and none of them could be settled by looking at a print. They were settled, eighty years later, by pointing an X-ray fluorescence spectrometer at the prints and counting atoms. Very little else in this course shows so plainly what a non-destructive analytical method is for.

Recommended uses

Reading the period literature without being misled by it. Three quotations, in order, are the argument of this page:

• Willis to the Camera Club, on why the developer route failed: “where the ferrous image is strong a black colour will be developed, for there will not be enough mercuric chloride in contact with it to have the same effect relatively as on the other parts.”
• W. H. Smith, the Platinotype Company’s manager, in 1911: “The Platinotype Company had never advocated the use of mercury in their developers simply because it was not stable, and they would not include a formula in their instructions which would injure the reputation of platinum printing for permanency.”
• Paul Anderson, in 1917, recommending exactly that practice and adding that “it has been stated on the authority of von Hübl that the image in a mercury developed platinum print consists of pure metallic platinum, the warmth of color being due to a difference in the size of the grains of metal deposited”.

Anderson’s book was influential and its second half is wrong. There is mercury in the image, and X-ray analysis has shown so repeatedly. Ware’s remark is that Anderson’s passage “does tend to perpetuate the misconception that there is no mercury in sepia platinotypes, which we now know to be wrong”.

Identifying a print, and telling the two routes apart. This is the strongest practical use of the page, and the evidence is quantitative. In 2015 Lewis and Koseki measured the mercury-to-platinum ratio in finished platinum prints by XRF and found a clear separation at the concentrations historically used: prints made from a mercury-containing sensitiser show Hg:Pt of about 0.2 to 0.3, while prints made with mercury in the developer show 0.8 to 1.1. Prints made both ways sit with the developed group. A single spectrum therefore says not only that a print is a sepia platinotype but roughly how it was made, which is more than any catalogue entry of the period says.

Understanding that mercury is necessary but not sufficient. Stulik and Vo recorded XRF spectra of Gertrude Käsebier’s prints in the J. Paul Getty Museum. Of the prints examined, thirty-five were platinum and two were pure palladium; every platinotype except one black specimen on Japanese paper contained mercury — yet only about thirteen of them are described as brown, nineteen as black or grey-black and two as warm black. Ware’s conclusion is exact and worth carrying: the presence of mercury looks like a necessary condition for a brown platinum image and is demonstrably not a sufficient one, and what else has to be true is not yet known.

Understanding why platinum needs help at all. The mercury is in the formula because pure platinum printing does not work very well. Ware’s account is that plain ferric oxalate with potassium tetrachloroplatinate(II) “often yields a weak, fibrous or ‘grainy’ platinum image”, because the reduction of platinum(II) by iron(II) is slow enough that the exposed sensitiser washes out of the paper before the image has finished forming. Everything Willis added over fifteen years — silver, lead, mercury — was an attempt to speed that reaction up. The colour shift was, at first, a side effect that he then sold as a product.

Conservation. Ware’s argument that a pure platinotype cannot fade — nothing in a normal gallery oxidises, dissolves or reacts with metallic platinum — makes any faded “platinum print” evidence that something else is in it. Some of the badly faded prints in the Library of Congress collection, by Eva Watson-Schütze and Sarah Sears around 1900, are believed from their XRF spectra to be mercury-developed. The AIC’s conservation catalogue states plainly that mercury-developed prints can lose colour over time and appear lighter brown.

Not for making a brown print. There is no version of this that this course teaches. The routes to a warm platinum-group image that do not involve mercury are the subject of the next section.

When another formula is preferable

• For a brown image in a noble metal, palladium. This is the answer the trade itself eventually reached. Palladium prints are naturally brown, and Ware notes that mercury need not be introduced into a palladium sensitiser at all — palladium is much less responsive to it, and mercury(II) tends to make palladium images more neutral rather than less. The Getty’s atlas says the same. The practical route is the three-solution drop system, where the platinum-to-palladium ratio is a colour control with no heavy metal beyond the image metal itself.
• For a warmer platinum image without mercury, a hotter or weaker developer — within limits. The potassium oxalate developer page records the temperature and concentration effects. Be clear about the size of the effect: Ware states that the warming a hot bath alone achieves on ordinary cold-bath paper is “only very slight”. Wall’s own table of paper, developer and temperature says the same thing in the other direction — a cold-bath black paper at 140 F gives a “warm” black, not a sepia.
• For a genuinely brown print by exposure alone, the print-out process. Ware’s print-out platino-palladiotype controls colour with the humidity of the paper at the moment of exposure, over a range the Getty describes as running from neutral black through sepia to Van Dyke brown, with no additive of any kind.
• For a toned rather than a developed brown, Packham’s catechu process. The Getty’s atlas records that James Packham patented a catechu-based toning process in 1894 (British Patent 24,963 of 27 October 1894) and that “it appears that Packham’s process was used as a safer alternative to the mercury-based toning process”. It is a plant tannin bath on a finished print. The course has no encyclopaedia entry for catechu and does not publish the process, but the historical point is worth having: a safer alternative was sought at the time, not only in hindsight.
• Never uranium instead. The other historical brown for platinotypes was uranium toning, and it is Level D for its own reasons; see the uranium toner and the Kodak Ltd uranium intensifier.
• Never this, for any practical purpose. The classification is not a difficulty rating.

Mixing

This section gives no mixing procedure. What can be described without instructing is the shape of the thing, the arithmetic that turns Willis’s units into modern ones, and the one quantity the patent leaves out.

Two bottles that never meet. A platinotype is not a developer formula. The coating solution goes on the paper and dries; the developing solution is a separate bath the exposed paper is floated on or immersed in. The two solutions meet inside the paper and nowhere else. That structural fact is the reason a mercury salt can be put in either one and give a different print each time, and it is why this entry is a sequence of two solutions rather than a mixture of them.

The units. Willis writes in grains and imperial fluid ounces. Converting with the table in Appendix VIII of the Platinomicon — 1 grain = 0.0648 g, 1 imperial fluid ounce = 28.413 cm³, and therefore 1 grain per fluid ounce = 0.228 per cent w/v — gives the figures in the table above. The percentage conversion is worth stating separately because it is the one a reader can check against Ware’s own published conversions of the same patents: he gives 4 grains of mercuric chloride per fluid ounce as 0.912 per cent w/v and 0.0336 molar, and 5 grains as 1.14 per cent w/v and 0.042 molar, and both agree exactly with the factor above.

🧮The same formula per litre, which is the course's arithmetic and not Willis's

Willis’s fluid ounce is 28.4 mL, which nobody mixes. Because both of his statements are make-up concentrations, scaling is a straight multiplication by 35.2, and the result is the same solution:

Solution	Ingredient	Per fluid ounce, as printed	Per litre, computed	Per cent w/v
Coating	Ferric oxalate	70 gr (4.54 g)	159.6 g	15.96
Coating	Potassium tetrachloroplatinate(II)	15 gr (0.972 g)	34.2 g	3.42
Coating	Mercury(II) chloride	4 gr (0.259 g)	9.12 g	0.912
Developer	Potassium oxalate monohydrate	120 gr (7.78 g)	273.6 g	27.36
Developer	Potassium tetrachloroplatinate(II)	7 gr (0.454 g)	15.96 g	1.596

In molar terms the coating is 0.425 M in ferric oxalate, 0.0824 M in platinum(II) and 0.0336 M in mercury(II) — a platinum-to-mercury ratio of about 2.4 to 1 — and the developer is 1.485 M in potassium oxalate and 0.0384 M in platinum(II). The litre column and the molarities are this course’s arithmetic, done so that the doses on this page can be compared with the ones under Variants; the formula is the percentage, and the fluid ounce is Willis’s.

The ingredient with no quantity. Willis writes “70 grains of ferric oxalate (with enough oxalic acid to render this ferric oxalate freely soluble)” and stops. There is no figure, in this patent or in its abridgment. Ferric oxalate is awkward to dissolve and free oxalic acid keeps it in solution, so the omission is a real gap rather than an editorial one. The only number any source read for this page attaches to it is Wall’s, from a different formula in 1924: his normal ferric oxalate stock is “a 20 per cent solution of ferric oxalate with an excess of 1.2 per cent of oxalic acid”. That is a useful order of magnitude and it is not Willis’s, and it has therefore been kept out of the table.

The substance that is missing, and why. The patent’s principal coating solution carries two grains of plumbic chloride per fluid ounce, and the mercuric version replaces it. Lead(II) chloride has no page in this course’s chemical encyclopaedia — lead nitrate, the salt of Willis’s 1873 patent, does — so an ingredient line for it would point nowhere. The lead additions are the business of the separate entry on lead contrast additions; what matters here is that Willis himself printed the two salts as alternatives in one sentence, so recording the mercuric one is reporting the patent rather than editing it.

Nothing on this page was heated in the making. The heat in the historical process arrives at the development stage, applied to a bath containing a mercury salt, and that is the step that both makes the colour and creates the worst of the hazard.

Behaviour

What follows is what the sources record happening, in the order they record it. It is history, and the numbers in it are observations rather than settings.

The bath had to be hot, and it stayed hot long after black platinotype stopped needing to be. Willis’s 1878 developer route required 140 to 160 °F (60 to 71 °C). When he introduced cold development in 1892 it transformed the black papers and would not work for the sepia ones; he told the Camera Club he was “doubtful whether sepia would ever be obtained under a temperature of 150 °F”, and in 1893 wrote that “to make good sepia paper is a heart-breaking problem, so much so indeed that I have rarely had the courage to attack it”. The Platinotype Company was still selling “sepia hot bath paper” in 1906, and in 1911 its sepia papers still wanted 160 to 170 °F.

Mercury in the developer produced double tones. This is the single most-reported defect of the process and both period explanations survive. Willis’s own is a mass-transport argument: where the iron(II) image is strong, the mercuric chloride in contact with it is used up faster than it can arrive, so the shadows develop black while the high values, with plenty of mercury per unit of iron, develop brown. Wall’s 1912 dictionary gives the kinetic version: “the oxalate in the developer is much more rapid in its action than the mercury”, so the platinum reduction outruns the mercury effect in the densest areas. The colours Wall reports are specific — the light tones incline “towards pink” while the darker ones “remain more or less black”.

The cure for double tones was glycerine, and it worked by slowing diffusion. Wall’s remedy is to mix the cold-bath developer with an equal volume of glycerine and add an alcoholic mercuric chloride solution to it, “which retards the action of the oxalate” and gives the mercury time to act evenly. Ware gives the same mechanism in general terms: increasing the viscosity of the developer with an inert water-miscible liquid diminishes the rates of diffusion, which is also the basis of the “glycerine method” of local development that Willis described around 1893 and that Stieglitz and Keiley took up around 1900. The Getty’s atlas records that glycerin development was often used precisely to get a double-tone effect deliberately, with a mercury-containing developer applied locally to the hands or the face of a portrait.

Sepia paper was faster than black paper. Wall’s 1912 dictionary states that sepia paper “is even more sensitive to daylight than the ‘black’ variety, and the exposure is therefore correspondingly less”, and that “increased care must be taken when printing” because faint light affects it.

The clearing bath took some of the warmth away, and under-clearing was the price of keeping it. Anderson records that most writers recommended clearing such prints for not more than two minutes in a single bath of one ounce of hydrochloric acid in three hundred of water, “since it is undeniably the case that a normal clearing bath will remove some of the warmth resulting from the use of mercury”. Wall’s 1924 rule for the mercuric citrate paper is the same shape: acid baths of no more than 0.5 to 1 per cent, and no more than thirty minutes of acid treatment in all. Ware’s comment on Anderson is the important one: he was right that so dilute a bath would leave much residual iron in the paper, and he never asked why a normal clearing bath should damage the image so much if the image were really pure platinum. It is a good example of a correct observation resting on a wrong model. The course’s ordinary clearing practice is on the platinum clearing sequence page.

A mercury-developed image could be dissolved off. F. C. Lambert, commenting on Smith’s 1911 lecture, reported as a novelty that mercury-developed sepia images were soluble in a solution of sodium thiosulphate and potassium ferricyanide — that is, in Farmer’s reducer, a reagent that does nothing whatever to platinum. That single observation, made thirty-five years after the process was introduced and unexplained at the time, is direct chemical evidence that the brown image contained something that was not platinum.

Whether it lasted was disputed, and the dispute has been partly settled. Smith showed severely faded mercury-developed platinotypes at a meeting in 1915. Anderson, in 1917, said flatly that in ten years of his own practice he had not found instability, and attributed the reports to insufficient clearing. Ware’s modern position is that the archival stability of historic platinum images containing mercury “appears to be variable and unaccountable”: some faded, and a growing body of XRF evidence shows that many undegraded sepia platinotypes of about 1900 contain mercury too.

Image characteristics

A platinum print that behaves like a palladium print. Ware’s summary of what mercury(II) does to the finished image is that “it results in an image resembling that of palladium, i.e. having a lower contrast, brown colour and very smooth texture”. All three halves of that matter. The smoothness is the cure for the fibrous, grainy pure-platinum image; the lower contrast is a cost, not a benefit; and the brown is the reason anybody paid for it.

The colour is a dose, and the period sources put it on a scale. Wall’s 1912 six working proportions for a mercuric glycerine developer are the most explicit colour ladder any source read for this page prints, and they are reproduced under Variants with their computed mercury strengths. The ends of the ladder are “a print of a slightly warmer black than the ordinary black platinotype colour” and “a warm sepia”, the latter at what Wall calls “the maximum proportion of B which it is advisable to use”. Wall’s 1924 handbook, working with mercuric citrate in the sensitiser instead, says the opposite thing about optimisation: “the smaller quantities of the mercury solution give the best colours”.

There is more mercury in the print than in the solution. Ware coated a platinum sensitiser containing mercury(II) nitrate at a final 0.34 mol/dm³, a molar ratio of platinum to mercury of 1 to 1, exposed it over a range of times, and measured both metals in the processed images by quantitative XRF. Mercury is precipitated much more readily than platinum: the finished image runs at an Hg:Pt ratio of about 4 at low exposures and about 2 at high ones, from a solution that was 1 to 1. Whatever else the mercury is doing, it is not a trace additive that stays behind in the bath.

The image particles are small, and that is what a brown metal print means. Electron microscopy of a modern platinotype resolves single ellipsoidal platinum nanoparticles of 15 to 25 nm distributed within the surface cellulose fibres. No published measurement of particle size in a mercury-containing print was found for this page; both the Getty’s atlas and Ware agree that the brown colour is a smaller-particle effect, and both are reasoning from optics rather than from a micrograph of a sepia platinotype.

The mechanism

Take it in three steps: what light does, what the developer does, and what the mercury changes.

What light does

Iron(III) oxalate is the only light-sensitive substance in the coating. Absorbing ultraviolet, it loses one oxalate to carbon dioxide and drops to iron(II):

light + Fe₂(C₂O₄)₃ → 2 FeC₂O₄ + 2 CO₂

Photolysis: iron(III) oxalate to insoluble iron(II) oxalate

The product, iron(II) oxalate, is almost insoluble — Ware gives 0.022 g per 100 cm³ — so at the end of the exposure the image is a faint, immobile, brownish deposit that can reduce nothing because it cannot move.

What the developer does

Free oxalate from the potassium oxalate bath dissolves that deposit as a mobile complex, and the mobile iron(II) then reduces platinum(II) to metal:

FeC₂O₄ + C₂O₄²⁻ + 2 H₂O → Fe(C₂O₄)₂(OH₂)₂²⁻

Dissolution: the immobile photoproduct becomes a mobile bisoxalatoferrate(II)

2 Fe(C₂O₄)₂(OH₂)₂²⁻ + PtCl₄²⁻ → 2 Fe(C₂O₄)₂(OH₂)₂⁻ + Pt + 4 Cl⁻

Development: iron(II) reduces platinum(II) to platinum metal, releasing chloride

Two features of that last equation decide everything on this page. The first is that the oxalate complexation collapses the iron redox potential: E°(Fe³⁺/Fe²⁺) is +0.771 V, but the oxalato couple E(Fe(C₂O₄)₃³⁻/Fe(C₂O₄)₂²⁻) is only +0.02 V, which is why oxalate-complexed iron(II) is a strong enough reductant to precipitate a noble metal at all. The second is the chloride on the right-hand side.

What the mercury changes

Platinum(II) is reduced through its aquated form, and there is not much of it. The tetrachloroplatinate anion is not easy to hand electrons to; the aqua-complex is much easier, and the two are in equilibrium:

H₂O + PtCl₄²⁻ ⇌ PtCl₃(OH₂)⁻ + Cl⁻

Aquation, the equilibrium that decides how reducible the platinum is

Ware gives the equilibrium constant at 20 °C as 0.015 for platinum against 0.17 for palladium, so at sensitiser concentrations less than 20 per cent of the platinum is aquated where more than half of the palladium is. Worse, the equilibrium is slow to establish — a half-time of 2.4 hours has been measured — which is why freshly made tetrachloroplatinate solutions are left to mature for a day.

Mercury(II) pulls that equilibrium to the right by taking the chloride away. Mercury(II) binds chloride covalently and very strongly. Ware gives log K₁ = 6.74 and log K₂ = 6.48 for the first two steps, and an overall formation constant of 1.2 × 10¹⁵:

Hg²⁺ + 4 Cl⁻ → HgCl₄²⁻

Chloride scavenging: the tetrachloromercurate(II) complex

H₂O + PtCl₄²⁻ + Hg²⁺ → PtCl₃(OH₂)⁻ + HgCl⁺

The consequence: more of the platinum is in its reducible form

The reaction inhibits itself, and that is what the mercury is fixing. Each platinum atom reduced releases four chloride ions into the paper, which push the aquation equilibrium back the other way. Ware’s evidence for this being real is the reverse experiment: deliberately adding chloride, as an alkali metal chloride or as hydrochloric acid, inhibits platinum image formation. Self-inhibition also explains why platinotype exposures are longer than palladiotype ones despite both using the same iron sensitiser. A chloride scavenger is therefore not a colouring agent at all in the first instance: it is a way of stopping the reaction from poisoning itself.

Mercury(II) is also reduced, and stays in the print. Its potential is high — E°(Hg²⁺/Hg) = +0.854 V against +0.02 V for the oxalato-iron couple — so the iron(II) photoproduct reduces it to metal without difficulty, and the metal co-precipitates with the platinum. Ware notes that mercury does not form an amalgam with platinum, which becomes important below. In the developer route the chemistry is different again, because concentrated potassium oxalate converts the mercury(II) largely to an oxalate complex:

Hg²⁺ + 2 C₂O₄²⁻ → Hg(C₂O₄)₂²⁻

In the oxalate developer, mercury(II) is complexed before it is reduced

That complexation drops the mercury potential to about +0.39 V, which is still comfortably above the iron(II) couple, so reduction still proceeds; it also destabilises mercury(I) relative to mercury(II), because Hg₂²⁺ is not strongly complexed by oxalate for steric reasons, so the intermediate becomes less likely. One practical consequence follows directly: the mercury(II)-oxalate system is photosensitive, which is why the historic manuals insist that a sepia developer be kept in the dark.

⁇Why a mercury-bearing platinum print is brown: two mechanisms, one observation

Everyone agrees that the colour comes from smaller platinum particles and that smaller particles scatter and absorb differently. Nobody agrees on how the mercury produces them.

The Getty Conservation Institute’s atlas says the mercury slows the reaction down: “the mercury chloride acts as a retardant for the development of platinum particles, reducing the particle size and making the final platinotype image appear browner. As such, the mercury treatment is more a physical effect than a toning effect.”

Ware argues the opposite direction of effect from the same starting point: mercury(II) scavenges chloride, which increases the concentration of the reducible aqua-complex and so gives “easier reduction and better dispersion of the platinum within the cellulose fibres”. Faster nucleation with better dispersion also gives smaller particles.

Both accounts end at a smaller particle and a browner print, by opposite routes. Ware is explicit that the question is open: “to date, there has been no clear explanation put forward as to how the presence of mercury so strongly influences the colour of images in nanoparticle platinum, which does not form an amalgam, but the answer probably lies with Mie scattering theory, and a smaller particle size.” He adds a third complication of his own — mercuric chloride is known to add to platinum(II) complexes to give platinum-mercury bonded molecules, and reducing such a heterobimetallic species might give alloyed nanoparticles unlike ordinary platinum black — and invites anyone with an experiment that would falsify the chloride-scavenging hypothesis to produce it.

The course takes no side. It records that the colour is a particle-size effect, that both a retardation account and a promotion account are in the literature, that the two are testable against each other, and that the mercury is measurably in the image whichever is right.

⚗Why 'mercury toning' is the wrong name for it

Both the Getty and the AIC’s conservation catalogue use the phrase and both flag it. A toner acts on a finished image: it converts, coats or replaces an image substance that is already there. This does neither. The mercury is present while the image is being formed, in the coating or in the developing bath, and it changes what gets formed. The Getty’s wording is that the description “is not entirely correct” but “has historical value and was widely used in the photographic literature”, so a reader will keep meeting it. The AIC’s catalogue writes “mercury development (sometimes called ‘toning’)” and lists the mercury toner among things “added to the sensitizer prior to its application to paper, developing solution, or both”, which is a developer by any ordinary definition. The genuine post-process toners for platinotypes — uranium, gold, Packham’s catechu — are a separate family, described by Ware in his section on toning platinotypes.

Function of every ingredient

Ferric oxalate, 70 grains per fluid ounce, 4.54 g in 28.4 mL, 15.96 per cent w/v. Iron(III) oxalate is the light-sensitive substance and the only one; nothing else in the coating responds to light at all. Its job is to become iron(II) in proportion to exposure and then, once the developer has dissolved it, to hand its electrons to platinum. Willis’s 15.96 per cent is not a strange figure: his 1887 patents use 60 grains per fluid ounce, 13.68 per cent, and Wall’s 1924 stock is 20 per cent. More iron gives more reducing power per unit area and a heavier deposit, at the cost of more iron to clear out afterwards, which is the substance that stains a badly-cleared platinum print yellow. Less gives a thin image that the wet processing removes before it has finished forming. Free oxalic acid accompanies it in an unstated amount, purely to keep it dissolved. Note what iron(III) oxalate is not doing here: it is not the developer, and the brown deposit visible after exposure is not the image.

Potassium tetrachloroplatinate(II) in the coating, 15 grains per fluid ounce, 0.972 g in 28.4 mL, 3.42 per cent w/v, 0.0824 molar. This is the image metal, dissolved in the coating so that it is already beside the iron when the developer arrives. Platinum(II) is reduced to platinum metal by iron(II); the metal is the picture, and there is nothing else in a finished platinotype except paper. More platinum gives a denser and more neutral image and costs a great deal more money — this was and remains the most expensive ingredient in photography. Less, in the limit, is Willis’s platinum-in-the-bath process of 1887, which took the platinum out of the coating entirely and put all of it in the developer; that saved money, needed mercury or lead to work at all, and failed commercially because the mixed developer decomposed on standing.

Mercury(II) chloride, 4 grains per fluid ounce, 0.259 g in 28.4 mL, 0.912 per cent w/v, 0.0336 molar. The sepia-giving agent, and the reason this page is Level D. Willis introduced it as a substitute for plumbic chloride, that is, as a second candidate for the job of making platinum reduce properly, and discovered that it did something the lead salt did not: it changed the colour. Chemically it does two things at once. It binds chloride ion — the by-product that inhibits the platinum reaction — as HgCl⁺, HgCl₂, HgCl₃⁻ and HgCl₄²⁻, with an overall formation constant of 1.2 × 10¹⁵, and so keeps more of the platinum in the aquated form that is easy to reduce. And it is itself reduced to metallic mercury by the same iron(II), preferentially: Ware’s XRF measurement shows mercury depositing from a 1:1 solution to give an image two to four times richer in mercury than in platinum. More mercury gives a warmer colour up to a point and then stops being an improvement: W. H. Smith warned in 1911 that an excessive amount added to the developer gives “a disagreeable colour” and an impermanent image; Wall’s 1924 handbook says the smaller quantities give the best colours and that the more mercury salt is used, the weaker the developer should be; Ware’s modern instruction is simply not to over-use it. Less mercury gives a warm black rather than a sepia, which is exactly Wall’s number one proportion. The mercury is also the ingredient that makes the print’s future uncertain, for a physical rather than a chemical reason: elemental mercury is volatile at room temperature, it does not amalgamate with platinum, and so nothing in the image holds it down.

Potassium oxalate monohydrate in the developer, 120 grains per fluid ounce, 7.78 g in 28.4 mL, 27.36 per cent w/v, 1.485 molar. A solvent, not a reducing agent, and this is the single most misunderstood ingredient in alternative-process chemistry. It supplies free oxalate anions that convert the insoluble iron(II) oxalate the light has made into a soluble bisoxalatoferrate(II) complex, so that the iron can move to the platinum. It reduces nothing itself. More oxalate dissolves the photoproduct faster: Wall’s 1924 rule is that “the stronger the developer, the more rapid its action and the softer the print”, and that with more mercury the developer should be weaker, which is a statement about balancing two rates against each other. Less gives a slower, more brilliant, more contrasty print, and eventually one that never fully develops. The full account of this bath, its strengths across fifty years and its temperature behaviour is on its own page, the Willis potassium oxalate developer.

Potassium tetrachloroplatinate(II) in the developer, 7 grains per fluid ounce, 0.454 g in 28.4 mL, 1.596 per cent w/v, 0.0384 molar. The same salt again, in the other bottle, and the actual subject of the 1878 patent’s claim. Willis’s argument was that iron(II) diffusing out of the paper is wasted unless it meets platinum on the way, so a platinum-bearing developer catches reduction that a plain oxalate bath loses. More platinum in the bath means more of the image forms in the developer and less in the coating, which is the direction Willis pushed all the way in 1887 and then retreated from. Less is the ordinary modern practice: today’s oxalate developer contains no platinum at all, and everything is in the coating.

Two substances that appear in the sources and not in the table. Plumbic chloride, lead(II) chloride, is the salt the mercuric chloride replaces in this formula, at two grains per fluid ounce; it does the chloride-scavenging half of the mercury’s job — lead(II) forms PbCl⁺ through PbCl₄²⁻ in the same way — but it is too electropositive to be reduced to the metal by iron(II), so it stays in the paper as a colourless salt and contributes nothing to the image or to its colour. Ware notes that XRF surveys find the lead signal uncorrelated with image density, which is exactly what that prediction requires. The course’s encyclopaedia has no page for it; the closest entry is lead nitrate, the salt of Willis’s 1873 patent, at 40 grains per fluid ounce. And oxalic acid, present in the coating in an amount the patent does not state, purely as a solubilising agent for the ferric oxalate.

Interactions

Mercury(II) with chloride is the whole hypothesis. Everything the mercury does to the platinum reaction is mediated by chloride ion: the mercury takes it out of solution, the platinum reaction puts it back in, and the balance between those two rates is what a mercuric platinum sensitiser actually is. A corollary worth carrying into the darkroom generally: any free chloride is an inhibitor of platinum image formation, whether it arrives as a salt, as hydrochloric acid, or from the reaction itself. That is why hydrochloric acid belongs in the clearing bath and nowhere near the developer.

Mercury(II) with oxalate changes both the potential and the stability. In the concentrated potassium oxalate developer, mercury(II) is largely present as the bisoxalatomercurate(II) complex, which lowers its reduction potential to about +0.39 V — still reducible — and makes the mercury(I) intermediate less likely. It also makes the bath photosensitive, so a mercuric sepia developer is not a solution that can be left on a shelf in the light.

Mercury with temperature works in two directions at once. The hot bath the sepia papers needed speeds the mercury’s contribution enough to compete with the oxalate, which is why the double-tone problem is worse in a cold bath; and heating a mercury(II) salt solution is precisely the operation that makes the hazard on this page unacceptable. The two facts are inseparable, and they are the main reason the process cannot be modernised into something safe: the thing that makes it work is the thing that makes it dangerous.

Mercury with the clearing acid is a conflict of interest. A normal-strength acid clearing bath removes some of the mercury and with it some of the warmth; a weak one preserves the colour and leaves iron in the paper, and residual iron darkens in light and yellows the print. The period literature resolved it by under-clearing, which is the wrong resolution, and Anderson’s own text records the trade-off without recognising it as one.

Mercury with gelatin, and platinum with gelatin. Willis avoided gelatin sizing from an early stage, and Ware’s explanation is chemical: gelatin coordinates strongly to platinum(II) in aqueous solution and renders it less readily reducible, which de-activates the very reaction the mercury is there to promote. Gelatin is also listed among mercuric chloride’s own incompatibilities. Wall’s 1924 recommendation for the mercuric citrate sepia paper is an agar-agar size, and for plain platinotype an arrowroot size. A modern paper with a calcium carbonate buffer is worse still, for a different reason: alkaline buffering decomposes the iron sensitiser outright.

Mercury with palladium runs the other way. Ware and the Getty agree that palladium is much less responsive than platinum to mercury, that palladium images are already brown, and that mercury(II) tends to make them more neutral. So the one place a mercury additive might have seemed most at home is the place it has least to offer.

Mercury with Farmer’s reducer. A mercury-developed image dissolves in thiosulfate and ferricyanide, which platinum does not. This is an interaction with a conservation consequence rather than a darkroom one: a treatment that is harmless to a platinotype is not harmless to a sepia platinotype, and there is no way to tell them apart by looking.

Variants

Every entry below is a different published formula, not a rewording of the one in the table. They are given with their own quantities and their own sources so that a reader meeting any of them can place it.

Willis’s own sequence of patents

Patent	Date	Mercury dose, as printed	Converted	Notes
No 2011	5 June 1873	none; lead(II) nitrate 40 gr/fl oz	9.12 % w/v, 0.275 M	The first platinotype patent. Lead only, applied as a separate coating
No 2800	12 July 1878	mercuric chloride 4 gr/fl oz in the coating, replacing 2 gr of plumbic chloride	0.912 % w/v, 0.0336 M	The formula on this page. Mercuric chloride is also named as an addition to the developing solution, with no quantity
No 1117	15 March 1880	none	—	“Dispenses with the lead salts… avoiding their use”
No 1681	2 February 1887	lead and/or mercury salts, 1 to 3 gr/fl oz, in the ferric oxalate coating	0.228 to 0.684 % w/v; 0.0084 to 0.0252 M as HgCl₂	The patent that says it in as many words: “I find the mercuric salt very useful where a warm tone or effect somewhat resembling that of sepia is desired”
No 16,003	21 November 1887	1 gr/fl oz each of plumbic and mercuric chloride in a 60 gr/fl oz ferric oxalate coating, or 2 or more gr/fl oz of mercuric chloride alone; or 5 gr/fl oz of mercuric chloride in a developer of 90 gr/fl oz potassium oxalate and 9 gr/fl oz chloroplatinite	0.228 % (0.0084 M); 0.456 %+ (0.0168 M+); 1.14 % (0.042 M)	The platinum-in-the-bath process. Mercury or lead is “in all cases essential”; the developer decomposed on standing and Willis withdrew the process in 1892

Two things are worth reading off that table. The first is that Willis went back and forth: lead in 1873, mercury introduced in 1878, both abandoned in 1880, both returned to in 1887 as necessities. The second is that the mercury dose in the developer — 5 grains per fluid ounce, 1.14 per cent — is higher than any dose he ever printed for a coating, which is consistent with what XRF later found: developed prints carry three or four times the mercury-to-platinum ratio of sensitised ones.

Wall’s mercuric glycerine developer, 1912

This is the most completely specified mercuric sepia formula in the period literature, and it is a developer applied to ordinary black paper rather than a sensitiser. Wall’s dictionary prints it as two solutions and six working proportions:

• A — ordinary cold-bath developer, 1 part; glycerol, 1 part.
• B — 10 per cent mercuric chloride solution in alcohol. Wall explains the solvent: “mercury is much more soluble in alcohol than in water”, which the PubChem record bears out — 6.9 g per 100 cm³ in water at 20 °C against 33 g per 100 cm³ in ethanol at 25 °C.

Wall’s number	A : B	Mercuric chloride in the working bath	Wall’s description
1	40 + 1	0.24 % w/v, 0.009 M	“a print of a slightly warmer black than the ordinary black platinotype colour”
2	30 + 1	0.32 % w/v, 0.012 M	between 1 and 3
3	20 + 1	0.48 % w/v, 0.018 M	“a good brown black”
4	20 + 2	0.91 % w/v, 0.033 M	between 3 and 6
5	20 + 3	1.30 % w/v, 0.048 M	between 3 and 6
6	20 + 4	1.67 % w/v, 0.061 M	“a warm sepia”, and “the maximum proportion of B which it is advisable to use”

The percentage and molarity columns are this course’s arithmetic from Wall’s stated 10 per cent stock and his stated ratios; Wall prints the ratios and the colours and no concentrations. Read against the patents, Wall’s number 6 is about half as strong again as Willis’s developer dose of 1.14 per cent, and his number 4 lands, coincidentally, within a hundredth of a per cent of Willis’s 1878 coating dose. The coincidence is arithmetic and not evidence of any connection between the two formulas.

Wall adds three practical observations that make the chemistry visible. Prints made in solutions 1 to 3 come out darker than in a plain developer and those in 4 to 6 lighter, the lightening happening in the acid clearing bath “in direct ratio to the proportion of B solution”, and it can be prevented by returning the print to plain developer for a few minutes first — which is a clear statement that some of the image substance is acid-soluble. A “dirty scum” forms during development, which Wall calls harmless; the course does not adopt his word for anything on a page about mercury. And the bath must never fall below 60 °F. Wall’s verdict, against Smith’s from the year before: “prints produced in this way can be regarded as reasonably permanent.”

Wall’s hot-bath sepia sensitiser, 1924

Mercury in the coating again, this time as a measured volume of a made-up stock rather than as a weight: ferric oxalate solution 6 ccm, potassium chloroplatinite solution 4 ccm, 5 per cent mercuric chloride solution 0.2 to 1 ccm, sodium chloroplatinate solution 2 to 10 drops, water 2 to 4 ccm, developed at 70 °C in a bath of potassium oxalate 100 g, potassium phosphate 50 g, citric acid 20 g and potassium chloride 10 g per 1000 ccm. Ignoring the drops, that puts the mercuric chloride between roughly 0.07 and 0.4 per cent w/v of the mixed sensitiser, or 0.003 to 0.014 molar — this course’s arithmetic — which is well below Willis’s 0.912 per cent. Wall’s related “platinum-in-developer” sensitiser A goes further and uses 0.2 ccm of the same 5 per cent mercury solution in place of the platinum salt altogether, which is Willis’s 1887 idea surviving into the 1920s.

Hübl’s mercuric citrate, 1902, and the cold-developed sepia paper

The one real advance after Willis. Baron Arthur von Hübl, in the second edition of Der Platindruck, described using mercuric citrate rather than mercuric chloride in the sensitiser, and Ware records the decisive advantage: it does not require hot development. Later formulations built on it work at 0.02 to 0.1 molar in mercury, and the salt is easily made from mercuric oxide and citric acid.

Wall’s 1924 handbook prints the recipe in full. The stock is 5 g of yellow mercuric oxide and 25 g of citric acid in 100 ccm of water, heated until dissolved and filtered; the sensitiser is ferric oxalate solution 8 ccm, potassium chloroplatinite solution 4 ccm, mercuric citrate solution 1 to 4 ccm and sodium chloroplatinate solution 2 to 5 drops. Taking mercuric oxide at 216.59 g/mol, the stock is 0.231 molar in mercury and the sensitiser therefore between about 0.018 and 0.058 molar — this course’s arithmetic, and it lands inside the range Ware states independently for Hübl-derived formulations, which is a reasonable check that both are describing the same practice. Wall’s processing rules for it are the ones quoted under Behaviour: development not less than five minutes, acid baths of no more than 0.5 to 1 per cent for no more than thirty minutes in all, an agar-agar size, and smaller mercury doses for better colour.

Neither mercuric citrate nor mercury(II) oxide has a page in this course’s chemical encyclopaedia, so neither name above carries a link.

The Platinotype Company’s secret preparations

Ware records that for best results with the commercial sepia paper the Company recommended adding a small amount of its “Special Sepia Solution” to the usual oxalate developer, or making the developer up from its proprietary “Sepia Crystals”. Both were trade secrets and neither composition was ever published. Ware’s assessment that both “likely contained mercury salts” is an inference from context and not a measurement, and it is recorded here as his inference.

The non-mercury sepia claims

Two other additives were proposed for warm platinum tones and neither was pursued. In 1885 a Professor Borlinetto claimed that adding 10 per cent of a saturated copper chloride solution to a hot (176 °F) potassium oxalate developer gave a “warm sepia-brown tone”. A Dr Jacoby’s sepia developer contained zinc oxalate, of which “the larger the quantity, the warmer the tone”. Ware notes that a number of other chloride-binding metals — zinc(II), cadmium(II), copper(II), thallium(I), tin(II), bismuth(III) — should in principle behave the same way, which is a testable prediction of the chloride-scavenging hypothesis and, so far as the sources read here show, an untested one. Several of those metals carry their own reasons not to try.

Safety

Level D. The course gives no procedure for this formula and lists no protective equipment that would make one acceptable in a home darkroom. What follows is the hazard assessment that produced that classification.

The salt. Mercury(II) chloride carries the harmonised European classification under Regulation (EC) No 1272/2008, signal word Danger, with pictograms GHS05, GHS06, GHS08 and GHS09 and the hazard statements H300 fatal if swallowed, H314 causes severe skin burns and eye damage, H341 suspected of causing genetic defects, H361f suspected of damaging fertility, H372 causes damage to organs through prolonged or repeated exposure, and H400 and H410, very toxic to aquatic life with long lasting effects. The encyclopaedia page records the aggregated industry notifications alongside the harmonised entry, and their agreement. That is a substance that is acutely lethal by mouth, corrosive on contact, a suspected mutagen and reproductive toxicant, an organ toxicant on repeated exposure, and an aquatic poison — six separate grounds, any one of which would be enough.

The exposure limit. HSE’s EH40 lists “mercury and divalent inorganic compounds including mercuric oxide and mercuric chloride (measured as mercury)” at a long-term workplace exposure limit of 0.02 mg/m³ as an eight-hour time-weighted average. It also carries a biological monitoring guidance value in Table 3 — 20 μmol of mercury per mole of creatinine in a random urine sample — which is the sort of control a workplace applies when airborne monitoring alone is not considered sufficient. Neither is a control a domestic darkroom can implement.

Heat is the aggravating factor, and it is not optional in this process. The commercial sepia papers required development at 71 to 77 °C. Warming an open dish of a mercury(II) salt solution raises the vapour above it and enlarges every route of exposure at once. This is the specific reason the mercuric sepia platinotype is Level D rather than Level C: the hazard cannot be engineered out of the process, because the temperature is what makes the process work.

The platinum is not the mild half. The Level D policy page records the workplace exposure limit for soluble platinum salts, which is the tightest limit anywhere in this course and carries a sensitisation notation. A platinotype sensitiser is a hazardous solution before any mercury is added to it, which is one reason the drop system this course does teach is taught at its own classification with its own controls.

The finished print. No source read for this page assesses the hazard of handling a mercury-containing platinotype in a collection, and this page does not offer an assessment. What the sources do say is relevant to the print’s safety rather than the handler’s: elemental mercury in the image is volatile, the image can lose it, and conservation literature treats mercury-developed prints as a distinct and less stable class. Anyone handling such material professionally should be working to their institution’s own guidance, not to a course page.

Eating the historical practice for what it teaches. Willis, Smith, Wall and Anderson were all arguing about colour and permanence. None of the period sources read for this page discusses the toxicology of the substance at all — not the patents, not the dictionary, not the handbook. That silence is itself part of the history, and it is worth noticing before assuming that a nineteenth- century process was worked with any knowledge of what it cost.

Storage

Not stocked. The course keeps no mercury(II) salt and gives no storage instruction for one, and the encyclopaedia page for the salt says the same.

What the historical sources record about storing these solutions is chemistry rather than housekeeping, and it survives here for that reason. All platinotype stock solutions must be kept in the dark — Wall’s instruction in 1924, and the reason is that ferric oxalate is the light-sensitive substance of the process and does not know it is in a bottle rather than on paper. A mercuric sepia developer must be kept in the dark for a second reason: Ware records that the mercury(II)-oxalate system is itself photosensitive, so a stored mercury-containing oxalate developer can change with time, and notes that the historic instruction manuals say so. And a developer containing both a platinum salt and oxalate does not keep at all: Willis’s own warning about his 1887 platinum-in-the- bath developer is that its constituents “undergo a slow mutual decomposition; hence it is necessary to mix them not too long before use”, and in practice platinum oxalate and platinum metal precipitated out and the bath went black. That defect ended a commercial process in four years.

Incompatibilities

The reactivity of mercuric chloride is catalogued on its own encyclopaedia page, read from CAMEO Chemicals and NIOSH; three of the entries in that list bear directly on this formula and are worth taking together.

Reduced iron and iron salts. The list of things mercuric chloride is incompatible with includes iron salts and reduced iron. That is not a warning that happens to apply here: it is the reaction this entire formula is built on. The mercuric sepia platinotype is an intended incompatibility, carried out deliberately, in a paper sheet, at 70 °C.

Sulfides, sulfites and thiosulfate. Mercury(II) has an enormous affinity for sulfur, and the practical consequence for this process is the one Lambert reported in 1911: a mercury-bearing image dissolves in Farmer’s reducer. Any thiosulfate-bearing solution near a mercuric platinum print is attacking the image.

Alkalis, ammonia, carbonates and phosphates. All are listed, and all matter. An alkaline paper buffer destroys the iron chemistry independently; ammonia was the historical blackener for a mercury-bleached silver negative, which is the mercury intensifier and a different process entirely; and phosphate is the salt Willis put into his 1887 developer to render the ferrous image insoluble, and then removed in his next patent because he blamed it for the non-uniform tone.

Chloride, which is an incompatibility of the process rather than of the substance. Free chloride inhibits platinum image formation. Hydrochloric acid belongs in the clearing bath, after development, and nowhere before it.

Waste

This course generates none, because it does not use the process. Anyone who does is generating mercury-bearing effluent, and the position of the sources is unambiguous. Ware, writing about whether to admit mercury into a palladium sensitiser, states that “mercury introduces problems of toxicity and environmental disposal, so all precautions regarding residues and effluent waste must be taken”.

The aggregated hazard classification carries H410, very toxic to aquatic life with long lasting effects, so the aquatic hazard is not a marginal consideration. In England and Wales, a photographic bleach solution falls under the Environment Agency’s WM3 entry 09 01 05*, bleach solutions and bleach fixer solutions, which is an absolute hazardous entry — hazardous regardless of concentration — and mercury waste is separately controlled wherever it arises. GOV.UK’s guidance for households is that hazardous waste goes to a household hazardous waste collection point and never to the bin or the drain.

The course has not found a published domestic treatment that renders a mercury-bearing photographic solution safe to discharge, and does not offer one. Jurisdictions differ and local regulations govern.

Troubleshooting

A reading guide to the faults the period literature reports, not a set of remedies. Each entry is here because it identifies something — in a historical print, or in a historical text — rather than because anybody should be producing it.

Brown high values against black shadows. Double tones, the signature fault of the developer route. Willis’s diagnosis is local depletion of mercury where the iron(II) image is strongest; Wall’s is that the oxalate acts faster than the mercury. Wall’s remedy was glycerine to slow the oxalate down; Willis’s was to abandon the developer route and put the mercury in the coating. A print showing this fault was almost certainly mercury-developed rather than mercury-sensitised, and XRF would be expected to confirm it with an Hg:Pt ratio near 1.

Pink high values. The same fault, further along. Wall reports that in a haphazardly-mercurised developer “the lighter ones incline towards pink”.

A disagreeable colour, and an image that does not last. Smith’s 1911 warning about excessive mercuric chloride in the developer, demonstrated with faded specimens in 1915. Ware’s XRF work on faded Library of Congress prints supports the association, with the mechanism being the volatility of elemental mercury that has nothing to hold it.

A print that lightens in the acid clearing bath. Wall reports this for his stronger mercury proportions, “in direct ratio to the proportion of B solution”, and records that a few minutes in plain developer before the acid prevents it. Chemically it says that part of what has been deposited is acid-soluble — that is, that it is not platinum.

A print that yellows with age. Residual iron, left because the clearing bath was deliberately weakened to preserve the warmth. Anderson describes both halves of the trade-off and does not connect them.

A “platinum print” that has faded at all. Ware’s argument is that this cannot happen to platinum, so it is diagnostic: the image contains something else, and mercury and silver are the two likely candidates — mercury in a sepia platinotype, silver in a Satista print.

Black specks in the image. Recorded by Ware as a plague of platinum printing, attributed to particulate reducing impurities in the paper and to the photosensitivity of a re-used developer. Relevant here because a mercury-bearing developer is more photosensitive, not less.

Experiments

None of these uses mercury. Each isolates one of the variables this page attributes to it, using materials the course does teach, so that the claims above can be tested rather than believed.

1. How much does a hot bath alone actually warm a platinum print? This is the claim the whole page opens by denying. Print a step wedge on the drop system sensitiser, cut it into strips and develop identical strips in the oxalate developer at room temperature, at 40 °C and at 70 °C, clearing and washing all of them identically. Record the colour against a neutral grey reference under a stated light source. Ware’s prediction is that the warming will be “only very slight”; Wall’s own table predicts a “warm black” rather than a sepia at 140 °F. Whichever you find, you will have measured the size of the effect that a century of handbooks mistook for the cause of sepia platinotype.

2. Does chloride inhibit platinum development? The reverse experiment on which Ware rests the chloride-scavenging hypothesis. Prepare identical exposed strips and develop them in oxalate baths to which increasing quantities of potassium chloride have been added. The prediction is a progressive loss of density. This is the cheapest and most direct test of the mechanism on this page, it uses nothing hazardous beyond the platinum chemistry itself, and a null result would be genuinely interesting.

3. Platinum against palladium, at the same exposure. Ware’s account of self-inhibition predicts that platinum needs longer exposures than palladium despite the identical iron chemistry, because platinum is less aquated and because its own reaction releases the inhibitor. Print the same negative on a pure platinum and a pure palladium sensitiser, matched for coating volume, and compare both the exposure needed and the image colour. You are looking for the palladium print to be faster and browner, which is the pair of properties that mercury was added to platinum to imitate.

4. Developer strength against print colour and contrast. Wall’s rule is that the stronger the developer, the more rapid its action and the softer the print. Develop matched strips in oxalate at 10, 20 and 30 per cent w/v and plot the result. This is the variable the trade used before mercury and after it, and it is the one still available.

5. An identification exercise, with no darkroom at all. Take a museum collection that publishes analytical data on its photographs, find the platinum prints for which XRF results are given, and sort them by whether mercury is reported. Then compare that sorting with the colour described in the catalogue entry. The Käsebier survey is the published precedent and its result is the interesting one: mercury in almost every print, brown in only about a third. Write down what you think the missing variable is, and what would test it.

6. Read one patent properly. Willis’s No 2800 of 1878 is about two thousand words long and the whole of this page is in it. Read it, list every quantity it states and every one it does not, and mark the places where the specification hedges (“about”, “preferably”, “I do not confine myself”). Then compare your list with the table at the top of this page. It is the fastest way to learn what a formula from a patent is and is not evidence of.