Gold Toning and the Permanence of a Printed-Out Silver Image

A fixed salted paper print is a yellowish or reddish brown that most people do not much like, and it is, chemically, about the most vulnerable object this course produces. Those are the same fact. The image is nanoparticle silver — 10 to 100 nm, with a hundred times the surface area per unit mass of a developed print, and about a tenth as much silver in total — and both the colour and the fragility follow directly from that. Anything you do to the particle changes both.

Which is why this page will not separate them. Gold toning is taught here as one operation with two results, and the page’s job is to say what the chemistry actually is, what each toner family does differently, and — the harder question — what the evidence for the permanence claim really consists of.

This page gives no procedure. It sets the rules for the toning done in the printing lab, and it is placed before that lab because you should know what gold does to a nanoparticle before you spend any.

The mechanism: gold arrives as an ion and leaves as metal

Toning a silver image with a noble metal is a displacement. The image silver is the reducing agent; the gold ion is the oxidant; metallic gold ends up on the print and silver ions end up in the bath. Ware states the accounting in one sentence: in the process of reducing the gold salt to deposited gold metal, a chemically equivalent amount of metallic silver is dissolved out of the image.

Everything else follows from what “chemically equivalent” means, because silver is only ever oxidised to silver(I) and gold arrives in one of two oxidation states.

3 Ag + Au³⁺ → 3 Ag⁺ + Au

Gold(III): three silver atoms lost for every gold atom deposited

Ag + Au⁺ → Ag⁺ + Au

Gold(I): one for one

All useful gold toners employ gold(I). That is Ware’s conclusion and it is the single most useful fact about the family. Reilly, coming at it from the conservation side and describing the actual appearance of the failure, says what the gold(III) route costs a print: in an acid bath such as a simple gold chloride solution, one atom of gold replaces three of silver, “a case in which toning action would lag far behind bleaching of the silver”, and the result “would be a flat, lifeless image with a reddish color”. The bath eats the print faster than it colours it.

Part XX established this arithmetic and the redox reasoning behind it on a developed silver gelatin image and should be read for the ligand chemistry, the potentials and the disproportionation of gold(I) that makes a complex necessary. This page does not repeat it. What it does is take the same chemistry to a different substrate — and the substrate turns out to change almost everything about the outcome.

⚗The awkward part: gold(I) is not what you can buy

Gold(I) is not a commonly stable oxidation state in water. It disproportionates:

3 Au⁺ → 2 Au + Au³⁺

Gold(I) in water falls apart into the metal and gold(III)

so a toner cannot simply be a bottle of a gold(I) salt. Every published gold toner is a recipe for making gold(I) in the tray and holding it there, and Ware sorts the whole literature — about nine variations, on his count — into two ways of doing it.

Reduce and stabilise with a sulfur-containing ligand. Thiosulfate, thiocyanate or thiourea both reduce the gold(III) and hold the resulting gold(I) as a complex whose formation constant is large enough to keep it there. This is fast, it is what the thiocyanate toner does, and it is why the ingredient panel of almost every gold toner in print has a sulfur compound in it.

Or dissolve the gold(III) in a mild alkaline buffer and wait. Acetate, borate, carbonate, phosphate or even chalk will do it: under those conditions the gold(III) reduces itself to gold(I) by oxidising water, over as much as 24 hours. That is what the borax bath and the acetate bath are. Reilly’s practical test for whether a bath has got there is a colour: the stock gold solution keeps its yellow while it is in the relatively inactive acidic state and becomes colourless when contact with alkali has moved it to the active one, and that decolorisation is the best guide to the state of the bath. His sodium acetate bath generally needs 24 hours to ripen, which is Ware’s figure arrived at from the other direction.

And there is a ceiling. Ware: if the solution is made too alkaline it becomes ineffective. Reilly says the same thing operationally — a too-alkaline bath tones more quickly and then loses its activity much faster, and you can end up with a bath that still contains a great deal of gold and will no longer tone anything at all.

Deposition or displacement? A real disagreement, and it matters

The two standard accounts do not say the same thing about where the gold goes, and the difference predicts different densities.

⁇Does the gold coat the silver particle, or replace it?

Ware’s account is coating. Gold “tends to coat the silver nanoparticles, protecting them and modifying their colour”. On that picture the particle acquires a shell: its size goes up slightly, its surface chemistry changes completely, and the silver that dissolves comes from underneath and from the outside layer.

Reilly’s account is replacement, with growth. He describes the colour change as being produced “by changing the size and shape of the image particles through replacement of silver atoms by gold or platinum atoms”, and says that this enlargement of the metal aggregates is what makes the print appear colder and more neutral.

Why it matters. If gold coats, the particle is protected everywhere and the density need not fall much; if gold replaces atom for atom from gold(I), the total metal content is roughly conserved and the enlargement Reilly describes has to come from aggregation rather than from addition. The two predict different density outcomes for the same weight of gold, and neither author publishes a density measurement to settle it.

What the course uses. Both, labelled. Ware’s coating picture for the protective argument, because it is the mechanism that makes protection intelligible — a shell of a metal that resists sulfur and oxidation, over a core that does not. Reilly’s replacement-and-enlargement picture for the colour argument, because particle size is what sets a plasmon resonance and his account connects directly to the physics the chemistry lesson sets out. Where the two cannot both be right, the course says so rather than choosing, and the measurement that would settle it — reflection density and particle sizing on the same print before and after a known weight of gold — is one nobody read for this course has published.

And one consequence is not in dispute at all. Reilly: in either gold or platinum toning, the substitution process creates silver chloride as a by-product, so toning must always be followed by a fixing step to remove it. That is the reason the sequence in the printing lab is tone, then wash, then fix — and the reason Reilly’s alternative of toning after fixing works “with equal success” only if you fix again, since the toning would otherwise re-sensitise the print.

Gold at the particle, before and after

1. 1Nanoparticle silver, about 10 nm — the whole surface exposed; a hundred times the area of developed silver per unit mass
2. 2Gold(I) complex arriving from the bath — held in solution by a sulfur ligand, or made in situ by an alkaline buffer over hours
3. 3Silver leaving as the ion — one per gold from gold(I), three per gold from gold(III); it becomes silver chloride in the sheet
4. 4Gold at the surface — the colour shifts colder, and the exposed metal now resists sulfur and oxidation

Drawn to show the accounting rather than a measured structure. Whether the gold forms a shell over the silver or replaces it atom for atom is exactly what the two standard sources disagree about, which is why the drawing shows gold at and near the surface without committing to a thickness.

What the deposition does optically

The physics is the physics of the chemistry lesson: a metal particle smaller than the wavelength of light absorbs by a surface plasmon resonance whose position depends on the particle’s size, its shape, how aggregated it is, and the refractive index around it. Toning changes at least the first three.

Reilly names all four factors as the things that set image colour and then says which the toner moves: the process modifies the colour by changing the size and shape of the image particles, and the enlargement of the aggregates makes the print appear colder — more neutral. In practice the sequence a printer watches is red-brown, through purple, to a cool purple-brown, blue-grey or, at the extreme, something near black.

Three details worth carrying to the tray.

Different toners stop at different colours, and it is chemistry rather than time. Reilly: a thiocyanate toner “achieves a more complete substitution of gold for silver and thus produces a colder image tone, generally deep purple tending to black”. The alkaline baths are gentler and land warmer. The choice of toner sets the range available; the time sets where in that range you stop.

A thiocyanate toner bleaches before it tones, and this alarms people the first time. Reilly’s explanation is that thiocyanates are solvents of silver chloride, so a print introduced into a thiocyanate bath at first bleaches and then intensifies as the gold is deposited. Wait.

And you cannot judge the colour wet. Photographers’ Formulary put it plainly for their own gold-borax bath: it is difficult to judge the end result of the toning until the print is fixed, washed and dried, and the cold tone increases on fixing and drying. Bostick & Sullivan say the same about theirs. Toning is judged by comparison with a print you took to a known point, not by what is in the tray. Reilly’s own recommendation for a beginner is the most useful piece of advice on this page: leave one or several prints completely untoned as a basis of comparison.

The four families, distinguished chemically

Tradition sorts gold toners by their name. Chemistry sorts them by how the gold(I) is made and what holds it, which is a better sort because it predicts the rate, the colour and the risk.

Family	How gold(I) is reached	Rate	Colour it goes to	The characteristic risk
Gold with borax and the other mild alkalis	Gold(III) oxidises water in a mild alkaline buffer, over hours	Slow to start; the bath must ripen	Purple-brown through to cooler	Goes inactive spontaneously after a few hours, with gold still in it and no warning but the cessation of toning
Gold with sodium acetate	The same route; the commonest alkaline bath of the albumen period	Needs about 24 hours to ripen before first use	Rich browns and purple-browns	Same, except that an acetate bath is reusable if strengthened with more gold stock
Gold with thiocyanate	The thiocyanate reduces gold(III) and holds gold(I) as a complex	Fast; ready when mixed	Deep purple tending to black — the coldest of the four	Consumes more gold than the alkaline baths; and a thiocyanate must never meet an acid
Gold with thiourea	The thiourea reduces and stabilises gold(I), usually with an acid	Fast	Blue on fine-grained developed papers; the classic blue toner	Thiourea’s own classification, which is why this course does not publish the procedure — see below
(and the historical fifth) sel d’or, gold with thiosulfate	Gold(III) plus thiosulfate gives the gold(I) complex Fordos and Gélis’ salt	Slow on albumen; it was the first method used on paper	Cool brown, purple or bluish black	Simultaneously a fixer and a sulfiding agent; abandoned by 1860 and discussed under History below

Why a bath’s pH is a variable in its own right. Three separate things depend on it and they pull in different directions.

Whether the gold is gold(I) at all. In an alkaline buffer the pH is what drives the reduction; Reilly’s yellow-to-colourless test is a pH-driven speciation change that you can see.

How fast the bath works and how long it lives. Too alkaline and it tones quickly and dies quickly. Reilly’s warning is that too active a toner, whether from too much gold or too high a pH, does not produce pleasant tones and is more difficult to control, and that a leisurely pace lets each print get individual attention.

And whether the bath attacks the image. An acidic gold(III) bath is the case Reilly and Ware both describe: it bleaches three silver atoms for every gold atom it deposits. Le Gray’s second-edition method was exactly this — 0.1 per cent gold chloride further acidified with hydrochloric acid — and Ware records that it required heavy overexposure of the print because so much image silver was inevitably dissolved. Thomas Sutton’s 1855 criticism was that Le Gray must have overexposed his prints to “a prodigious intensity” to make it work at all.

☠Two toners this course does not put in your hands, and why

Thiourea. It is chemically the best of the gold(I) reducing ligands for a blue tone on a fine-grained image, Ware gives the classic formulations, and the course still does not publish a procedure. The reason is the classification. Under Regulation (EC) No 1272/2008, thiourea carries H351, suspected of causing cancer, and H361, suspected of damaging fertility or the unborn child, along with H302 and H411, with the GHS07, GHS08 and GHS09 pictograms; the thiourea page classifies it Level C, which is a fume cupboard, controlled waste and supervision. The rubric takes a formula’s level from the highest of its chemicals, so a thiourea gold toner is a Level C bath and this course does not give Level C procedures to a home reader.

Platinum, for the same structural reason. Potassium tetrachloroplatinate(II) is Level C in this course’s encyclopaedia, so the platinum toner for printing-out papers is published with status: historical-study and no procedure at all. It is chemistry and history here, and the section below treats it that way.

Neither exclusion is a judgement about the chemistry. Both are the classification rubric applied consistently, and in both cases the course teaches what the substance does and where it sat historically without telling anybody how to use it.

Platinum, taught as chemistry and history

The same reasoning with a different metal, and the differences are instructive.

It runs in acid, not alkali. Reilly states the main technical difference between the two plainly: gold toning goes on best in an alkaline environment, platinum toning requires a neutral or acidic one. That single inversion is why the two cannot share a bath and why a print toned in both is washed between them.

The salt matters more than it does for gold. Platinum has two accessible oxidation states in this context and they behave completely differently. Platinic chloride — platinum(IV) — was what M. De Caranza recommended in 1856, the first published mention of platinum toning; Reilly says very little notice was taken, because such a toner has only a very slight toning action and a strong tendency to bleach the silver image, so all platinic formulae are restricted at best to matte papers and are totally ineffective on albumen. Platinous salts — platinum(II), chiefly potassium chloroplatinite — are by contrast very active toners when combined with acids, and have a smaller tendency to attack the image.

Its history is downstream of the platinotype. Potassium chloroplatinite was a fairly obscure substance until 1879, when Willis began to market the platinotype and made it an article of commerce. In 1886 J. Reynolds discovered that it was a very energetic toner of silver prints, giving brown and brownish-black tones instead of gold’s purples and blue-blacks; Stieglitz published one of the pioneering formulae in 1889, simply nitric acid, potassium chloroplatinite and water. The practice peaked between 1895 and 1925, when matte collodion and matte gelatin printing-out papers were popular: matte gelatin papers were often toned with platinum alone and are therefore brown, while matte collodion papers were generally toned with both gold and platinum, which produced the familiar olive-black of studio portraits of that era.

Its failure modes are its own. Yellowish highlights if the toning goes too long or the solution is too strong or too acidic. And a bath that has been robbed of activity by contamination: Reilly notes that any trace of silver nitrate or sodium thiosulfate alters the platinum to an irreducible condition, which is why prints for a platinum toner are well washed, treated in a 5 per cent sodium chloride solution for two minutes and washed again before they go in.

And the reason it was economic at all is worth remembering when you read the cost section below: in the early 1890s, when the price of platinum soared, it turned out to be cheaper to tone a silver print with platinum than to make a real platinotype, and several manufacturers sold matte salted papers under names like “silver-platinum paper” on exactly that logic.

The history, and where the sources disagree

The course does not give this part a page of its own, but three dates recur in later arguments and two of them are contested.

Gold on paper came from the daguerreotype. Fizeau’s gilding bath of 1840 — Reilly’s date, and Ware gives 23 March 1840 for Arago’s announcement with Fizeau publishing in August — was adapted to paper prints by P. F. Mathieu in 1847, in a pamphlet Reilly calls Auto-Photographie and Ware Autophotographie. Both agree the technique did not catch on until Le Gray publicised it after 1850, and the Getty Atlas dates the general adoption of gold toning to after 1850 with further acceleration after 1855.

Talbot appears never to have used it. Ware states there is no evidence that Talbot ever applied gold toning to his prints — remarkable, he adds, given the impermanence already besetting the Reading prints by the mid-1840s — and no evidence that the Scottish photographers used it either. The survival of Hill and Adamson’s 3,000-odd salt prints of 1844 to 1847, against the fate of the 10,000 Reading prints, is therefore not an argument for gold. Ware attributes it to the care they took over fixing and very thorough washing.

⁇Two dates in the standard accounts that do not agree

Alkaline gold toning. Reilly credits James Waterhouse, “sometime around 1855”, with proposing a bath of gold chloride and one or two mild alkalies. Ware dates a sodium carbonate bath to Waterhouse of Halifax in 1858, adds Maxwell Lyte’s trisodium phosphate in 1859, and argues that Humbert de Molard’s chalk-neutralised gold bath of 1851 anticipated the whole class, since a saturated calcium carbonate solution sits at about pH 9.4 and would have reduced the gold to gold(I) over 24 hours — a conjecture he marks as one, supported by the reported decolorisation of the bath.

Thiocyanate toning. Reilly: “In 1867 another important class of gold toning baths was discovered, this time based on the combination of gold chloride with thiocyanates.” Ware: ammonium thiocyanate “had first been used on print-out papers by Meynier in 1863 and by Liesegang in 1868”.

What the course uses. The decade, not the year, for both — alkaline gold toning belongs to the 1850s and thiocyanate toning to the 1860s — and it names the disagreement rather than picking a side. Neither source is a primary document here; both are secondary accounts citing period literature, and this course has not read the originals. Where a date is load-bearing for a claim about permanence, the claim is stated in terms of the practice rather than the year.

What is not contested is the 1855 committee and what it concluded. The Photographic Society of London appointed its committee on 21 May 1855 to consider the fading of positive photographic pictures upon paper, and it reported on 21 November. Reilly quotes the finding: “the most ordinary cause of fading may be traced to the presence of sulphur, the source of which may be intrinsic from hyposulphite left in the print, or extrinsic from the atmosphere, and in either case the action is much more rapid in the presence of moisture.” Ware adds that gold-toned prints were deemed more resistant, and that the recommendation that gold in some form should be used was not unanimous. Two controls came out of it — thorough washing, and gold toning — and they are still the two.

The permanence argument, in full

Here is the claim in its strong form: an untoned printed-out silver print is more vulnerable to sulfiding and to oxidation than a developed silver gelatin print, and a gold-toned one is measurably less so. Take it in three parts, because they are supported by very different amounts of evidence.

Part one: why printed-out silver is unusually vulnerable. Established.

This part is not in doubt and the reasons are physical as much as chemical.

There is very little of it. Ware’s figure for a whole-plate salt print is about 3.3 mg of silver in total, at a coating weight of about 0.1 g/m² — roughly a tenth of a modern silver-gelatin print. It takes very little hostile impurity to react with that much silver.

Nearly all of it is surface. At a particle size perhaps a hundredth of a modern print’s, the surface area for the same mass is about a hundred times larger. Reilly makes the same argument in words: the small particle size means a very large surface area relative to mass, so a large portion of the total mass is on the surface and readily accessible to destructive chemical agents.

There is no binder in the way. Ware’s comparison with albumen is direct: a print in which the colloidal silver is protected by a vehicle has much more restricted access of atmospheric oxygen to the image silver than a salt print, which is quite permeable to gases. And Reilly’s washing chapter makes the same structural point about the other direction of attack: in a modern paper a substratum of baryta and gelatin separates the image from the base, and in a salted paper it does not, so the base paper can become a reservoir of image-threatening substances.

And sulfiding costs more density here than anywhere else. This is the quantitative core of the argument and it is Ware’s. Nanoparticle silver has an extinction coefficient of about 16,000 dm³ mol⁻¹ cm⁻¹ at its absorption maximum; nanoparticle silver sulfide has been estimated at about 560. So complete conversion of the image to silver sulfide drops the maximum optical density by a factor of nearly thirty.

🧮Why the same reaction tones a modern print and destroys a salt print

Part XX uses sulfide deliberately, as a toner, and the print gets richer. Here the same reaction is the principal failure mode. Two numbers explain the difference.

560 ÷ 16,000 ≈ 0.035, so the density falls by a factor of about 30

Density lost on complete sulfiding, from the extinction coefficients

In a modern print that hardly shows. The silver is ten times more plentiful and its particles are far larger, so there is density to spare and the covering power of the coarse sulfide is adequate. Silver sulfide, Ware notes, is a good stable pigment in sufficient concentration — which is exactly what a sulfide-toned bromide print has.

In a salt print there is nothing to spare. Three milligrams of silver at a density of, say, 1.4 becomes silver sulfide at a thirtieth of the covering power, and what is left is the pale putty colour of a faded nineteenth-century print. Ware’s own explanation of why sulfide should look yellow rather than black completes it: at nanoparticle thickness you see only the attenuated absorption of silver sulfide, whose band is centred in the ultraviolet and only tails into the blue.

And a little sulfide is not the same as a lot. Ware records Henglein’s observation that dilute sulfide first shifts a silver sol’s absorption towards the green, turning it from yellow to brown — which is why a trace of sulfiding enriches an image — and that further exposure replaces the band altogether with the feeble spectrum of colloidal silver sulfide. A little sulfiding protects and enriches; complete sulfiding destroys. That paradox is the whole of the “old hypo” story below.

Part two: why gold should help. Well founded, and partly mechanistic.

Reilly’s statement of the mechanism is the clearest: gold and platinum react much less readily with sulfur and are much more difficult to oxidise than silver, so a layer of gold or platinum on a silver particle will tend to shield the silver inside from attack, especially from oxidising agents. Ware’s version says the same in the language of nanoparticles: gold tends to coat the silver nanoparticles, protecting them and modifying their colour.

Two supporting observations from the historical record are worth having because they are dose arguments rather than mechanism arguments.

More gold is better than less. Reilly notes that the alkaline gold toners deposited more gold than the earlier methods and that this contributed to albumen prints’ resistance to oxidative fading.

And more silver is better than less. In a passage about albumen paper that generalises directly to this part, Reilly observes that prints of the 1860s and 1870s were salted more heavily than their successors of the 1880s and 1890s, deposited more image silver, and have a slightly better average record of durability. That is an argument for printing a salt print strongly rather than thinly, and it bears directly on the exposure decision in the printing lab.

Part three: what has actually been measured. Much less than you would expect.

!Classify every permanence statement before you repeat it

The evidence divides into four kinds and they are not interchangeable.

Nineteenth-century testimony. Sidebotham, writing in 1861: the prints he toned with gold chloride in 1852 were “as beautiful as on the day they were printed”, while of those toned with old hypo “few are in existence, and none are in their original state”. That is a careful observer reporting a real comparison over nine years, and it is not a controlled experiment.

Aggregate survival. The Reading prints against the Hill and Adamson prints; the state of the albumen prints Reilly examined. Real evidence about outcomes, confounded by everything — different papers, waters, washing, storage and use.

Measured light-fading on real objects. This exists and it is thin. Ware reports John McElhone’s quantitative exhibition study, which included a salted paper print by Benjamin Turner and a lightly albumenised print by Baldus of 1855: after a total exposure of 30 kilolux hours under UF-1 filtration at under 100 lux, the densities of both had fallen by about 0.02 — a change Ware notes is similar in magnitude to the precision of the densitometer used, so not deemed significant. Read as an upper bound on the change, it implies a threshold exposure lifetime of nearly 300 hours of gallery display. Ware’s own recommendation for thiosulfate-fixed material is that 100 kilolux hours is an acceptable risk. For comparison, chloride-stabilised photogenic drawings, which are a different object, have a threshold exposure of the order of 200 lux hours — about four hours at 50 lux.

Accelerated ageing. Reilly and his co-workers did exactly this work by incubating test specimens at high temperature and various relative humidities — on albumen prints. It produced two important results: the Maillard protein-sugar reaction is a further contributor to highlight yellowing, and the instability of albumen prints is inherent rather than due to inadequate processing, greatly accelerated at high relative humidity. Pretzel and Martin’s study of Lady Hawarden’s albumen prints found average threshold exposure lifetimes of the order of one or two years — about 10,000 hours — with nearly as much change in dark-stored as in exhibited material, so light exposure alone was not the significant factor.

And here is the gap this page cannot close. All of that accelerated-ageing work is on albumen, whose chemistry differs from a plain salt print in exactly the respect that matters: egg protein carries sulfur-containing side groups that bind silver, which a cellulose sheet does not. Ware, at the end of his albumen section, writes that it might be hoped that an equally thoroughgoing study will soon be made of salted paper prints also — which is a specialist saying, in 2019, that it has not been. Reilly said the corresponding thing in 1980: no research into these problems had been done since the days of Haddon and Grundy.

So: this course has found no accelerated-ageing study of toned against untoned salted paper prints. The permanence recommendation on this page rests on the mechanism, on the dose arguments, on nineteenth-century comparative testimony and on the aggregate survival record — which together are a good case, and are not a measured one. Say that when you repeat it.

The other half of permanence, which no toner fixes

Gold addresses the image silver. It does nothing at all about what is left in the paper, and on these papers that is the larger threat.

Residual thiosulfate. The 1855 committee named it first, and the mechanism is the sulfiding above, supplied from inside instead of from the air. Reilly’s structural argument is why it is worse here: the image is in intimate contact with the fibres, so the base paper becomes the reservoir. And his one honest admission is the one to carry: experimental evidence on the washing of albumen and salted paper prints is almost nonexistent, so his own recommended times are extrapolated from gelatin papers on a stated assumption.

Residual silver compounds. Not the image and not removable by the fixer — the chain of nineteenth-century experiments that established this is set out in the chemistry lesson, ending with Haddon and Grundy’s measurement that a thoroughly fixed and washed but unexposed albumen print still held nearly 5 per cent of the silver put on it. That figure is for albumen and should not be transferred to a plain salted paper, whose retention mechanism is different and for which no source read for this course reports a measured figure.

Which is why the sequence in the printing lab is what it is: first wash, tone, wash, fix in two fresh baths, wash, sulfite washing aid, final wash — and then the residual silver and thiosulfate tests, because the one thing this page can say with confidence is that the wash times are not measured on these papers and the print therefore has to be tested rather than assumed.

The permanence chain for a printed-out print, and where you can intervene

1. 1The image silver itself: 3 mg, nanoparticulate, a hundred times the surface area of a developed imageIntervention: print strongly rather than thinly, and gold tone. Both are dose arguments — more silver and more gold each associate with better survival in Reilly's record of albumen prints
2. 2Residual silver compounds left in the sheet by the fixerIntervention: fresh fixer, two baths, and the time the source gives rather than longer. Measured for albumen at nearly 5 per cent of applied silver; unmeasured for plain salted paper
3. 3Residual thiosulfate in the paper fibresIntervention: do not over-fix, wash, use a sulfite washing aid, wash again, and test. There is no baryta layer between this and the image
4. 4The mount and its adhesive, attacking from behindIntervention: do not mount unless the print needs it. Reilly says most salted papers can simply be matted; where mounting is unavoidable, fresh pure starch on a rag board
5. 5The atmosphere: sulfur compounds, and oxidising gases such as ozoneIntervention: the enclosure, and where the print is kept. Reilly names automobile exhaust and copier ozone specifically
6. 6Temperature and relative humidity, which set the rate of everything aboveIntervention: 18 to 20 °C at 35 to 45 per cent in Reilly's recommendation; the AIC give 18 to 30 °C and 30 to 50 per cent for these prints. IPI's Preservation Index exists because this is a continuum rather than a threshold

Six links, and the ones a printer controls are the first three. The last three belong to whoever keeps the print, which is the argument for writing the processing record on the back of the mount.

Mounting, adhesives and enclosures

This is where the advice for a salt print diverges from the advice for everything else in the course, and there are three divergences worth naming.

Most salt prints do not need mounting at all. Reilly’s rule: because they are made on relatively heavy paper and carry little binder, most salted papers can be simply matted and do not require mounting, and mounting is a process that should be employed only when necessary, not as a matter of course. That is the opposite of the albumen case, where a thin sheet under a heavy albumen layer curls hard enough to tear itself in half, and where about 95 per cent of nineteenth-century prints were mounted.

Where mounting is unavoidable, the adhesive is fresh pure starch. Reilly reports the consensus of photographic conservators: pure fresh starch has proven itself the best mountant, removable with the least difficulty and not attacking the prints. The nineteenth century also used gelatin, gum arabic, dextrines and albumen itself, and Reilly records that impure “glue” was known as a cause of trouble even in the 1850s, alongside rancid flour paste and India rubber solution.

And the board is the documented killer. A typical nineteenth-century mount was thin good paper over a core loaded with lignin; the decomposition products migrate through the top sheet and attack the photograph, causing staining, brittleness and accelerated fading and yellowing, with foxing often on top of that. Reilly notes that the danger is especially acute for prints on thin stock because there is so little barrier between the silver image and the mount — which for a salt print is the same argument as the missing baryta layer. Removing prints from defective mounts and remounting them on appropriate material with a safe adhesive is, he writes, the only preservation treatment for these prints that has proven itself in practice.

On enclosures, the course defers to the specialists. IPI’s guide to ISO 18902 “photo-safe” testing pairs each layer of a photograph with the reactants that attack it — oxidising agents, reducing agents, high alkali, acids, lignin — and the damage each produces, including image fade, silver mirroring and yellowing, and describes the Photographic Activity Test that is itself an international standard, ISO 18916. That test is the thing to look for on a product, and its existence is the reason this page names no brand.

On the storage environment, two published ranges, and they are not the same.

Source	Temperature	Relative humidity
Reilly, for albumen and salted papers	18 to 20 °C	35 to 45 per cent
AIC Photographic Materials Group, for photogenic drawings, salted paper prints and calotypes	18 to 30 °C, stable, to avoid embrittlement	30 to 50 per cent

Reilly’s is the tighter and older recommendation, and his stated reason is that temperature and humidity govern the rate of all possible destructive reactions; the AIC’s is wider and foregrounds stability rather than a set point. IPI’s preservation metrics explain why both can be right: the effects of the environment follow a continuum, there is no clean line dividing good storage from bad, and their Preservation Index expresses the difference in years — calibrated so that 20 °C at 45 per cent gives a PI of 50 years for a typical “preservation problem object”, with a PI of 100 years meaning that object would take twice as long to reach the same state. Neither range is a threshold. Both are places on a slope.

Cost, honestly

Gold is the most expensive thing in this cluster after platinum, and the arithmetic is worth doing before you mix a litre of anything.

🧮Gold per print, from three published strengths — and the price the course cannot give you

The strengths, all of them sourced. Reilly’s rule for porous papers — arrowroot, plain salted — is 0.1 to 0.2 g of gold chloride per litre, against 0.4 to 0.5 g/L for glossy albumen, because porous papers tone quickly and would over-tone in the stronger bath. The two kits land inside that: Bostick & Sullivan’s working bath is 50 mL of a 0.2 per cent gold solution in a litre, which is

0.2 % × (50 ÷ 1000) = 0.010 % w/v = 0.10 g of gold chloride per litre

Bostick & Sullivan working bath

and Photographers’ Formulary’s gold-borax bath is 8 mL of a 1 per cent stock in 500 mL, which is

1 % × (8 ÷ 500) = 0.016 % w/v = 0.16 g of gold chloride per litre

Photographers' Formulary Solution D

Per print. Bostick & Sullivan publish a consumption figure directly: assume each print uses up approximately 5 mL of the 0.2 per cent gold solution, which is

5 mL × 0.002 g/mL = 0.010 g of gold chloride per print

Gold chloride consumed per print, from the supplier's own replenishment rule

so their 500 mL bottle of solution B, containing 1 g of gold chloride, tones on the order of a hundred prints.

And here the arithmetic stops. src/data/prices.json carries no price for gold chloride. It does name the gap, and it names it in this page’s own terms, with the figure above written into the register beside it. What would settle it is a dated UK listing stating the pack size and the form beside the price, because 1 g of the dry salt and 10 mL of a 1 per cent solution are sold under the same words and are not the same quantity. Until then this page gives grams per print and no pounds per print, and the consumables tables in the printing lab report the row as unpriced.

Four ways to use less gold, all of them from the sources rather than from thrift.

Trim the print’s edges before toning. Photographers’ Formulary say it in one line, and it is the cheapest saving available: the toner works on the area you put in the tray.

Use the strength the paper needs, not the strength albumen needs. Reilly’s 0.1 to 0.2 g/L for a matte salted paper is a quarter of the albumen figure, and a print in the stronger bath does not tone better, it over-tones faster.

Choose the bath for its gold economy where the colour allows. Reilly notes that thiocyanate toners consume more gold than the alkaline varieties. If a warmer, alkaline-toned print is the one you want, it is also the cheaper one.

And replenish rather than remake, where the bath allows it. Bostick & Sullivan give both routes for their thiocyanate bath: add 5 mL of gold stock after each print to hold full strength, which suits toning to a cool blue endpoint, or watch the bath and add 25 mL when toning slows, which suits partial toning for warmer prints. Reilly’s caution applies to the alkaline baths instead: most are one-use and become inactive spontaneously after a few hours, with no clue other than the cessation of toning, and an acetate bath is the exception that can be used repeatedly if strengthened.

How this page relates to Part XX

Part XX’s gold toner lesson covers gold toning of a developed silver gelatin image with the Kodak formulations, and it owns the redox arithmetic, the ligand chemistry and the potentials. This page covers gold toning of photolytic silver in a paper with no binder. The mechanism is the same; almost every consequence is different.

	Developed silver gelatin (Part XX)	Printed-out silver (here)
Image silver	Micron-sized filaments, roughly ten times the mass	Nanoparticles of 10 to 100 nm, about 3 mg in a whole plate
What gold does to the colour	Very little, and it may be imperceptible — Ware says the filamentary silver is far larger than nanoparticle silver, appears neutral black, and offers little scope for shifting its colour	A great deal: red-brown through purple to blue-grey, because the plasmon resonance moves with the particle
Why you would do it	Protection rather than toning — Ware’s own words for the modern practice, and the reason GP-1 exists at 0.01 per cent gold with 1 per cent thiocyanate	Both at once, and they cannot be separated
Where the print is protected from	Mostly oxidation and handling	The atmosphere directly: no binder, a hundredfold surface area, and sulfur is the documented attacker
The blue-tone case	Real, on fine-grained warm-tone chlorobromide papers, and it needs thiourea	Not applicable: the particles are already small, and the toner is chosen for the warm-to-cool range instead

Both accounts of the mechanism agree, and the agreement is worth stating explicitly because the two pages were written from different sources: gold(III) costs three silver atoms per gold atom, gold(I) costs one, all useful toners therefore deliver gold(I), and gold(I) has to be made and held by a ligand or by an alkaline buffer. Part XX derives that from Ware’s redox potentials; this page reaches it from Reilly’s account of what an acid gold bath does to a print. Same conclusion, two routes.

Recap

Colour and permanence are one subject here, because both follow from the same nanoparticle silver: 10 to 100 nm, about a tenth the silver of a modern print, a hundred times the surface area, and no binder between it and the air.

Gold(I), always. From gold(III) a print loses three silver atoms per gold atom and bleaches faster than it tones. Gold(I) is one for one, is not stable in water on its own, and must be made and held either by a sulfur ligand — thiosulfate, thiocyanate, thiourea — or by a mild alkaline buffer that reduces the gold over as much as a day.

pH is a variable, not a detail. It decides the speciation, the rate, the life of the bath and whether the bath attacks the image.

Two toners are excluded by the rubric rather than by the chemistry: thiourea, for H351 and H361, and platinum, whose salt is Level C. Both are taught here and neither is given as a procedure.

Sulfiding costs a salt print about thirty times its density, from the two extinction coefficients, which is why the same reaction that tones a bromide print destroys this one — and why a trace of sulfide enriches while complete sulfiding does not.

The permanence case is good and it is not measured. Mechanism, dose, testimony and survival support it; no accelerated-ageing study of toned against untoned salted paper was found for this course, and the specialists say so themselves.

And the half gold cannot fix is the residue — thiosulfate and silver compounds in a paper with no baryta layer, on wash times that Reilly says are extrapolated rather than measured. Which is why the print is tested rather than trusted.

Check your understanding

Question 1. A print is gold toned. What happens to the image silver, and where does it go?

• Nothing: the gold deposits on top and the silver is unchanged
• A chemically equivalent amount of silver is oxidised to silver(I) and dissolves out of the image; in the sheet it re-precipitates as silver chloride, which is why toning must be followed by fixing
• The silver is converted to silver sulfide, which is what gives the colder colour
• The silver migrates to the surface of the print and is washed away in the following bath

Show the answer and why

Answer: A chemically equivalent amount of silver is oxidised to silver(I) and dissolves out of the image; in the sheet it re-precipitates as silver chloride, which is why toning must be followed by fixing

Ware states the accounting: in reducing the gold salt to metal, a chemically equivalent amount of metallic silver is dissolved out of the image. Since silver is only oxidised to silver(I), gold(III) costs three silver atoms per gold atom deposited and gold(I) costs one - which is why all useful gold toners employ gold(I). Reilly adds the consequence that decides the processing order: the substitution creates silver chloride as a by-product, so a toned print must always be fixed afterwards, or that chloride will re-sensitise it.

Question 2. Why is a simple acidic solution of gold chloride a poor toner for a salt print?

• It is too slow to be practical
• It delivers gold(III), so three silver atoms are bleached for every gold atom deposited, and the toning lags far behind the bleaching
• It cannot penetrate the paper fibres
• It reacts with the paper sizing before it reaches the image

Show the answer and why

Answer: It delivers gold(III), so three silver atoms are bleached for every gold atom deposited, and the toning lags far behind the bleaching

Reilly describes the outcome exactly: in an acid bath one atom of gold replaces three atoms of silver, a case in which toning action would lag far behind bleaching of the silver, and the result would be a flat, lifeless image with a reddish colour. Ware records that Le Gray published precisely this method - 0.1 per cent gold chloride further acidified with hydrochloric acid - and that it required heavy overexposure of the print to compensate, which Thomas Sutton criticised in 1855 by observing that Le Gray must have overprinted to a prodigious intensity.

Question 3. Two identical prints are made and one is gold toned. Both are stored in the same album for twenty years. State what you expect to differ, and classify your statement.

• The toned print will be measurably better preserved; this is established by accelerated ageing studies on salted paper
• The toned print is expected to have faded and yellowed less, on a mechanism that is well founded and on nineteenth-century comparative testimony - but no accelerated ageing study of toned against untoned salted paper was found for this course, so the statement is inference and testimony rather than measurement
• No difference is expected, because permanence depends only on washing
• The untoned print will be better preserved, because the toner dissolves image silver

Show the answer and why

Answer: The toned print is expected to have faded and yellowed less, on a mechanism that is well founded and on nineteenth-century comparative testimony - but no accelerated ageing study of toned against untoned salted paper was found for this course, so the statement is inference and testimony rather than measurement

The distinction is the point of the question. The mechanism is well founded: gold and platinum react much less readily with sulfur and are harder to oxidise, so a coating of gold shields the silver inside. The testimony is real: Sidebotham in 1861 reported his 1852 gold-toned prints as beautiful as the day they were printed while few old-hypo prints survived at all. But the accelerated ageing work Reilly and his co-workers did was on albumen, whose sulfur-containing egg protein makes it a different object, and Ware writes at the end of that section that it might be hoped an equally thoroughgoing study will soon be made of salted paper prints also - which is a specialist saying it has not been.

Question 4. At the strengths the sources publish, how much gold chloride does one print consume, and what does that make the cost?

• About 0.01 g, and the course cannot give a cost because no dated price for gold chloride exists in the price file
• About 0.1 g, at roughly £1 per print
• About 1 g, which is why gold toning is rarely done
• It cannot be estimated because consumption depends entirely on the print

Show the answer and why

Answer: About 0.01 g, and the course cannot give a cost because no dated price for gold chloride exists in the price file

Bostick and Sullivan publish the consumption directly in their replenishment instruction: assume each print uses up approximately 5 mL of the 0.2 per cent gold chloride solution, which is 0.010 g of gold chloride. Their 500 mL bottle therefore holds about a hundred prints' worth. Reilly's bath strength for a matte salted paper is 0.1 to 0.2 g per litre, and both kits sit inside that range. The price is the missing term: src/data/prices.json carries no figure for gold chloride and does not list it among its named gaps, so this part names it as a new one.

Question 5. Why does gold toning change the colour of a printed-out silver print dramatically and a modern bromide print hardly at all?

• Because bromide papers are hardened in the fixer and the toner cannot penetrate
• Because the developed silver of a bromide paper is filamentary and much larger than nanoparticle silver, so it already appears neutral black and offers little scope for shifting its colour
• Because gold does not deposit on developed silver at all
• Because bromide papers contain a supercoat that blocks the gold

Show the answer and why

Answer: Because the developed silver of a bromide paper is filamentary and much larger than nanoparticle silver, so it already appears neutral black and offers little scope for shifting its colour

Ware makes the point directly, and it is why the modern use of gold on silver-gelatin paper is described as protection rather than toning: the silver formed in developed bromide papers has a filamentary structure much larger than nanoparticle silver and appears neutral black, so there is little colour to shift. Kodak GP-1, at 0.01 per cent gold chloride in 1 per cent sodium thiocyanate, exists for the archival benefit rather than the colour. On a printed-out print the same reaction moves a plasmon resonance and the colour changes from red-brown to purple to blue-grey.

Question 6. You have mixed a gold-borax bath from the published formula and it will not tone a print. Which explanations are consistent with the sources?

• The bath has not ripened: an alkaline bath reduces gold(III) to gold(I) by oxidising water over as much as 24 hours, and Reilly's test is that the yellow stock becomes colourless when it is ready
• The bath was made too alkaline, which makes it tone quickly and then lose activity, so it can contain a great deal of gold and no longer tone at all
• The bath is exhausted: most alkaline baths are one-use and go inactive spontaneously after a few hours, with no clue but the cessation of toning
• A trace of fixer has got into it, which Reilly says ruins a toning solution

Show the answer and why

Answer: The bath has not ripened: an alkaline bath reduces gold(III) to gold(I) by oxidising water over as much as 24 hours, and Reilly's test is that the yellow stock becomes colourless when it is ready, The bath was made too alkaline, which makes it tone quickly and then lose activity, so it can contain a great deal of gold and no longer tone at all, The bath is exhausted: most alkaline baths are one-use and go inactive spontaneously after a few hours, with no clue but the cessation of toning, A trace of fixer has got into it, which Reilly says ruins a toning solution

All four are in the sources, which is what makes a dead alkaline gold bath annoying to diagnose. Ware gives the reduction of gold(III) by water in a mild alkaline buffer over as much as 24 hours, and the failure of an over-alkaline solution. Reilly gives the yellow-to-colourless test as the best guide to the state of the bath, the warning that a too-alkaline bath tones fast and dies fast, the observation that most alkaline baths are one-use with no indication of exhaustion but the cessation of toning, and the flat statement that the toner solution is ruined by even a trace of fixer.

Question 7. Why does this course teach the platinum toner for printing-out papers but publish no procedure for it?

• Because no reliable formula has been published
• Because potassium tetrachloroplatinate(II) is classified Level C in this course, and the rubric takes a formula's level from the highest of its chemicals, so the bath is a Level C procedure that a home reader is not given
• Because platinum toning does not work on salted paper
• Because the platinum price makes it impractical

Show the answer and why

Answer: Because potassium tetrachloroplatinate(II) is classified Level C in this course, and the rubric takes a formula's level from the highest of its chemicals, so the bath is a Level C procedure that a home reader is not given

It is the classification rubric applied consistently rather than a judgement about the chemistry. Reilly publishes a workable formula and the course carries it in the formulary with status historical-study; what it does not do is give steps. The same reasoning excludes thiourea gold toners, since thiourea carries H351 and H361 and its encyclopaedia page is Level C. In both cases the course teaches what the substance does and where it sat historically - platinum toning peaked between 1895 and 1925, gave matte gelatin papers their brown and matte collodion papers their olive-black when combined with gold - without telling anybody how to use it.

Check my answersClear