Wedgwood and Davy's silver nitrate on leather

Every formula in this book is descended from this one, and it is one salt and some water. What makes it worth a page is not the ingredient but the ratio, the support and the sentence that ends the paper it appeared in: nothing but a method of preventing the unshaded parts of the delineation from being coloured by exposure to the day is wanting, to render the process as useful as it is elegant. The process had one missing part and its authors knew exactly which one.

Ingredient	Quantity	Form the source specifies
Silver nitrate	100 g	Davy's "one part of nitrate to about ten of water", read as parts by weight and scaled by the course to a kilogram of water. He states no weight, no volume and no batch size, and the "about" is his.
Water	1000 mL, added	Ten parts of water to one of nitrate. This is an added proportion and not a make-up volume, so the finished solution is more than a litre by an amount no source states, and no per cent w/v can honestly be given for it. Taking a kilogram of water as a litre is the course's scaling.

☠Why this page carries no procedure, and why the reason is not hazard

This is a historical-study entry. The reason has nothing to do with danger: silver nitrate at this strength is weaker than the bath the course publishes for salted paper and is handled at the same Level B. There are three other reasons and they are all evidential.

The support is half the formula and cannot be specified. Davy states plainly that on leather the solution “is more readily acted upon than when paper is used”, and gives no other reason to prefer leather. The best available explanation — Michael Gray’s, reported by Mike Ware — is that the white kidskin of the period was tawed, cured with mineral salts rather than tanned with tannic acid, and that residual chloride in the skin formed silver chloride in place. If that is right, then the chloride content of a particular skin, which nobody has ever measured and which the trade itself varied from dresser to dresser, is the active variable in the formula. Modern white leather is not tawed kidskin. A procedure written against it would be a procedure for something else.

The process is defined by not working. Its authors published it as an unsolved problem. Writing a procedure would be issuing instructions for making an object that erases itself in daylight, and calling that a lab exercise misrepresents both the history and the chemistry.

No specimen is known to survive, so there is nothing to check against. The two pictures long circulated as Wedgwood’s were shown in 1864 not to be his, and the more famous of the two turned out to be a Talbot.

The course’s actionable routes to this chemistry are elsewhere and they are properly taught. The unfixed silver photogram reproduces the observation on modern paper, safely and in an afternoon. The plain salting solution with the salted paper sensitiser and a thiosulphate fixer reproduces the chemistry, in the form that survived. Read this page for what those bottles are descended from, and for the one number in it that most accounts of the 1802 paper have lost.

Purpose

To put silver ions into an organic sheet at a strength high enough to blacken visibly in minutes of sunlight, and low enough not to wreck the sheet.

Both halves of that are Davy’s, in one sentence, and it is the earliest formulation criterion in photography:

With regard to the preparation of the solution, I have found the best proportions those of one part of nitrate to about ten of water. In this case, the quantity of the salt applied to the leather or paper will be sufficient to enable it to become tinged, without affecting its composition or injuring its texture.

Read it as a specification and it has an upper bound and a lower bound. The lower bound is tinged — enough silver in the sheet to give a visible tone. The upper bound is without affecting its composition, or injuring its texture — and silver nitrate is corrosive and an oxidiser, so a leather or paper loaded past a certain point is stiffened, discoloured overall, and no longer a support. Davy found a working window and published its middle. That is what a formula is.

There is a second purpose which is this course’s rather than Wedgwood’s. This is the control experiment for the whole of Part I. Every subsequent nineteenth-century silver process is a modification of this one, and the modifications are only legible against it: Talbot’s photogenic drawing paper adds a halide deliberately and in deficiency; the calotype changes the halide and stops printing out; the gallo-nitrate adds a developer; hypo adds a fixer. Take all four away and you are holding this.

Recommended uses

Davy publishes an application list, and it is worth having in his order because it is a list of the things a shadow-printing material can do.

• Copying paintings on glass. A painted glass panel laid on the sensitised surface in sunlight gives “distinct tints of brown or black, sensibly differing in intensity according to the shades of the picture”. This is the use in the title of the paper, and the one for which he specifically recommends leather.
• Profiles. A cast shadow leaves the covered part white and darkens everything around it. This is the photogram, and it is the operation the course runs on modern paper.
• Objects partly opaque and partly transparent. His examples are the best sentence in the paper for a modern reader: “the woody fibres of leaves, and the wings of insects may be pretty accurately represented by means of it”, with the direct solar light passed through them onto prepared leather. That is contact printing from a natural negative, thirty-seven years before Talbot showed his lace to the Royal Society.
• Solar-microscope images, on paper. Davy’s own, and the only projected image the process ever recorded. The condition he attaches — the paper close to the lens — is the whole reason it worked.

And one use he reports as unsatisfactory rather than impossible: copying prints by transmitted sunlight. The unshaded parts copy slowly, but the light coming through the printed areas is “seldom so definite as to form a distinct resemblance”. A modern reader recognises the complaint, though the explanation below is the course’s and not Davy’s: paper is a diffuser, the light scatters sideways under the ink, and the mid-tones smear.

The use it will not serve is the one it was built for. The camera obscura defeated it, and Davy says so without hedging.

Today the recommended use of this formula is reading it. Anything you would actually want to make with silver on a sheet is made better, and permanently, with the salted paper chemistry that descends from it.

When another formula is preferable

Effectively always, and it is worth being specific about which one and why, because the four successors each fix a different fault.

You want a print that still exists next month. Use the plain salting solution and the salted paper sensitiser, fixed with plain hypo or the printing-out thiosulphate fixer. The fault being corrected is the one the 1802 paper itself names: there is no fixer here, and washing and varnishing were both tried and both failed. Thiosulphate dissolves the residual silver halide instead of merely covering it up.

You want more density and a shorter exposure from the same two elements. Use Talbot’s photogenic drawing paper. The fault being corrected is that Wedgwood’s sheet made its silver chloride by accident and in whatever quantity the skin happened to supply. Talbot introduced the chloride himself, deliberately short of the silver, and Reilly puts the resulting rule at about six times more nitrate than chloride — a 2 to 4 per cent salting bath against a 12 per cent silver bath. Deliberate control of the halide-to-silver ratio is the single largest step between 1802 and 1839.

You want an exposure a camera can give. Use a developing-out process — the calotype with the gallo-nitrate of silver. The fault is thermodynamic and no amount of formulation fixes it: a printing-out image is built photon by photon and, on Reilly’s figure, can cost up to a hundred thousand times the light energy of a developed one. Wedgwood was asking a print-out material to work at camera illuminances, and it cannot.

You want to see the effect for yourself this afternoon. Use the unfixed silver photogram, which does the whole of Wedgwood’s experiment — the shadow, the darkening, the fading of the highlights, the twin kept in the dark — on modern paper with no chemistry at all.

Mixing

Everything Davy publishes about making the solution is the ratio and the brush. One part of nitrate to about ten of water; and, from the safety paragraph, “in employing them for moistening paper or leather, it is necessary to use a pencil of hair, or a brush”. There is no vessel, no temperature, no order of addition (there is only one solid), no filtration and no keeping instruction. A single soluble salt in water is about as simple as a formula gets, and the absence of a mixing procedure is not a gap in the record: there was nothing to record.

What does need working through is what the ratio means, because “one part to about ten” is not a concentration and turning it into one takes three decisions that are the course’s and not Davy’s.

🧮What one part in ten actually says, and what it refuses to say

Decision one: the parts are parts by weight. Davy does not say so. The course reads them that way because a crystalline solid is weighed rather than measured out by volume, and because his own next clause is about “the quantity of the salt applied to the leather or paper”. If that reading is wrong the strength is wrong; nothing else on this page depends on it.

Decision two: the scale. A ratio has no batch size. Taking one kilogram of water as the reference, and one litre of water as one kilogram:

1 part : about 10 parts by weight = 100 g of silver nitrate to 1000 g of water

Davy's proportion, scaled

Decision three, and it is a refusal. Ten parts of water is not “water to make”. Dissolving 100 g of a solid into a kilogram of water gives a solution of more than a litre, by an amount that depends on a density no source here publishes. So this page states no grams per litre and no per cent w/v, which is why the schema records the water as added rather than toMake. What can be stated honestly is the composition by weight and the molality:

100 g in 1100 g = 9.1 % w/w silver nitrate

By weight of the finished solution

100 g ÷ 169.873 g/mol = 0.589 mol per kilogram of water

Molality, which is what a parts-by-weight recipe actually gives you

That distinction between molality and molarity is not pedantry here, it is the difference between an eighteenth-century recipe and a nineteenth-century one. Davy’s proportion is reproducible with a balance alone. Kodak’s “water to make 1000 c.c.” — the convention the 1928 primer states explicitly, that a make-up volume “insures dilution to a definite volume, thus yielding a known concentration of chemicals each time” — needs a graduated vessel and gives a concentration per unit volume, which is what a bath’s chemical behaviour actually depends on. Every formula in the rest of this formulary is written the second way. This one is written the first.

Now the comparison, and mind the units. Ware reads Talbot’s sensitiser as 18 to 23 per cent w/v and Reilly specifies 12 per cent w/v for modern salted paper. Davy’s is about 9 per cent by weight, and those are not the same measure. The direction of the difference is worth getting right: a solution of a dense salt weighs more than water, so 100 mL of it holds more than 100 g, and a given per cent w/w therefore corresponds to a larger figure expressed as per cent w/v. Davy’s bath is thus a little stronger than the bare number 9.1 suggests when set beside the other two — which narrows the gap to Reilly and does not close the gap to Talbot. How much larger cannot be said here, because the course holds no source for the density of this solution. Stated as an ordering and no further: Davy’s bath sits somewhat below the modern salted-paper bath and well below Talbot’s, at roughly half of Talbot’s strength.

That is a more interesting result than it looks. The one number Wedgwood and Davy published is approximately right by modern practice. Nothing else about the process was.

The support, which Davy names and the schema cannot hold. The formula is white leather or white paper, and it is not indifferent between them: “For copying paintings on glass the solution should be applied on leather; and in this case it is more readily acted upon than when paper is used.” Ware identifies the material as white kidskin, and reports Michael Gray’s explanation that a skin finished that finely would have been tawed — cured with mineral salts — rather than tanned with tannic acid. The 1911 Encyclopaedia Britannica describes the trade practice in its own terms: leather is tanned (combined with tannin), tawed (prepared with mineral salts), or chamoised, and a tawing paste is “a mixture of alum, salt, flour, egg yolk and water”, one worked example being 9 lb of alum and 5 lb of salt to 100 lb of skin, with the caution that quantities “diverge widely, every dresser having his own recipe”. Glove kid and calf kid are named as treated on those lines.

So the support arrives already carrying chloride, in an amount the tanner chose and nobody recorded. The mechanism section takes up what that does. What it means for mixing is blunt: the most important variable in this formula was set in a tannery, and the one person who could have measured it was the one person who did not know it mattered.

Behaviour

In the dark, nothing. “White paper, or white leather, moistened with solution of nitrate of silver, undergoes no change when kept in a dark place.” That is the control, stated first, and it is the sentence that makes the rest an experiment about light rather than about air or heat — the distinction Schulze established in 1727 and which popular accounts still get wrong.

In daylight, a print-out. The sheet “speedily changes colour, and after passing through different shades of grey and brown, becomes at length nearly black”, and the rate scales with the light: two or three minutes in direct sun for the full effect, several hours in the shade. No development, no intermediate invisible stage, no chemistry after the exposure. What you see is what the light built.

And it is strongly blue-sensitive. Davy’s instrument was a piece of coloured glass and his result is right: red rays “have very little action upon it”, yellow and green “are more efficacious”, and blue and violet “produce the most decided and powerful effects”. He footnotes Scheele and Senebier for the same ordering, and Ritter, Bockmann and Wollaston for the invisible rays beyond the violet. That ordering is why a blue-sensitive material can be handled under orange or red light at all, and it is the reason his own candle-lamp observation, below, works.

Leather beats paper, and moist beats dry. The second of those comes out of his muriate comparison: silver chloride and silver nitrate “were both more readily acted upon when moist than when dry, a fact long ago known”. It is one of the most consequential throwaway lines in the paper, and section The mechanism says why.

The image, once made, is fast. “After the colour has been once fixed upon the leather or paper, it cannot be removed by the application of water, or water and soap, and it is in a high degree permanent.”

✕The word 'fixed' in that sentence does not mean what a photographer means

This is the single most misread line in the 1802 paper. Davy is saying that once the colour has been produced on the sheet, it will not wash off. He is not saying the picture has been fixed. Two paragraphs later he says the exact opposite about the picture as a whole, in the sentence the paper is remembered for.

The distinction is worth holding onto because it is the correct one chemically. The image was permanent — it is metallic silver, bonded into the fibres, and neither water nor soap touches it. The ground was not, because it was still loaded with unchanged light-sensitive salt. A photograph is the difference between those two areas, and a difference is spent as soon as the two areas meet. The problem of permanence owns that argument in full and plots what a week of room light does to it.

In the camera, nothing at all. Not “faint” — nothing, in any moderate time. This is the behaviour that mattered most to Wedgwood, and it is the one the paper reports as a failure.

Image characteristics

Tone and colour. Grey, through brown, to nearly black. The colour sequence is not arbitrary: it is the plasmon colour of finely divided photolytic silver, which shifts as the particles grow, and Ware gives about 10 nm as the ceiling on particle size for silver formed inside a pure silver halide crystal. A modern reader who has watched a printing-out paper go through exactly those colours in sunlight has seen the same physics.

Polarity. Negative, always. Everything the process made was a shadow-image: the covered part white, the exposed part dark. Nobody printed one through to a positive, and with no fixer nobody could have. That the process was inherently negative and nobody exploited it is one of the quiet tragedies of the paper — the negative–positive idea is Talbot’s, and it is a separate invention from the sensitive sheet.

Contrast and separation. High where the modulation was a hard-edged shadow, poor where it was a continuous tone. The glass-painting copies worked; the print copies did not, because the mid-tones did not separate. That combination — clean line, mushy tone — is the signature of a low-contrast material being asked to record a low-contrast subject through a scattering support.

Speed. Two or three minutes in full sun is the whole of its exposure scale, and it is fast enough for a photogram and hopelessly slow for anything else. Reilly’s figure gives the scale of the gap: a printing-out image can need up to a hundred thousand times the light energy of a developed one. Nothing that costs that much light will ever work behind a lens of 1800.

Surface. Two supports, two objects. On paper, the sensitiser sinks into the fibres and the picture is in the sheet. On kidskin the silver sits in a dense, smooth protein surface that takes a much finer line — which is, independently of the chemistry, part of why the leather results were the ones worth showing. Nobody can compare the two objects, because no example of either is known to survive.

The mechanism

The chain runs: silver ion → an electron from the support or from a chloride crystal → metallic silver → a visible difference between two areas of one sheet → a difference that closes. Four links, and Wedgwood and Davy had good evidence about the first, the third and the fourth.

Silver nitrate by itself is not a photographic material

This is the fact that makes the formula intelligible, and it is counter-intuitive enough that the entire nineteenth century talked as though the opposite were true. Ware states it flatly: silver nitrate “is not light-sensitive per se, but becomes so in the presence of organic matter, being reduced to black finely-divided silver metal by its light-induced oxidation of the organic substance, which can be paper or its incorporated sizing agent.”

The redox bookkeeping is short. Making metallic silver from Ag⁺ needs an electron:

Ag⁺ + e⁻ → Ag

The whole of the image-forming reaction, and the whole of the problem

and in a bottle of silver nitrate solution there is nothing to supply it. Nitrate is already fully oxidised. Water can be oxidised, but not easily enough at the potential the silver couple offers. So a bottle of the pure solution in the sun does very little on its own. CAMEO’s datasheet lists the substance as blackening on exposure to light or organic material, and Ware’s account is why those two belong in the same sentence: the organic material is the reagent, and without it the light has nothing to work with.

Put the same solution into paper or skin and the situation changes, because now there is something oxidisable in molecular contact with every silver ion: cellulose, the mill’s sizing, or in leather the collagen and the flour and egg of the tawing paste. Light drives the electron across, and black silver appears. The support is a reagent. Wedgwood’s formula has one ingredient in the bottle and two in the sheet.

Why leather beat paper

Davy observed the difference and offered no explanation. The explanation the course carries is Ware’s, and Ware attributes it to the photohistorian Michael Gray: the fine white kidskin of the period was tawed, not tanned — cured with a mineral-salt paste rather than with tannic acid or (later) chromium salts — and the residual chloride ions in the skin would react with the silver nitrate to form silver chloride in place.

Ag⁺ + Cl⁻ → AgCl

What the leather did to the sensitiser, without anybody asking it to

That single reaction changes the material completely, because silver chloride is a crystal, and plain silver nitrate on cellulose is not. Inside a silver halide crystal, absorbing a photon promotes an electron into a conduction band where it can move, and the silver ions of the lattice are mobile enough to meet it. That is the machinery photography is built on and it does not exist in a solution of silver nitrate soaking a fibre. Ware’s account of the print-out mechanism turns on it: photolysis liberates halogen,

Light + 2 AgCl → 2 Ag + Cl₂

Ware's net photolysis, and the reaction that has to be kept going forwards

and what happens to that halogen decides whether the image builds or unbuilds. With no acceptor present, the chlorine simply reoxidises the silver it just made, and Ware records that photolysis inside a pure crystal therefore stalls at an optical density “in the order of 0.02” — a barely perceptible darkening, nowhere near a picture.

⚗Why Wedgwood's sheet did not stall, and why Talbot's did better

An acceptor has to take the halogen out of circulation. In a Talbot sensitiser, Ware identifies the candidates as excess silver(I) ions, water, the cellulose and the sizing agent, and concludes that water and silver(I) ions together are the major halogen acceptor, contributing the greater part of the image silver. Gelatin, he notes, is not an effective scavenger of halogen at print-out exposure levels, so the sizing is not essential to the photochemistry.

Wedgwood’s sheet had both of those acceptors, and had them for a reason nobody planned. The chloride came from the leather in trace amounts; the silver nitrate came from the brush in gross excess; so the halide was in enormous deficiency and free silver nitrate was lying everywhere around each crystal. Ware’s picture of “sensitized” silver halide — a crystal carrying adsorbed Ag⁺, its excess positive charge pulling photoelectrons to the surface where silver specks can grow free of the lattice — is exactly the state a trace of chloride in a strong silver bath produces. Mark that as a reading: it is the course applying Ware’s mechanism, which he wrote about Talbot’s sensitiser, to a sheet nobody has ever analysed. It follows from Gray’s explanation of the leather and it falls with it.

And that is why Davy’s second observation is so painful in hindsight. Both salts were “more readily acted upon when moist than when dry”, he writes. Water is half the acceptor. He measured the effect of the mechanism’s central reagent and filed it as a curiosity.

So what was Talbot’s improvement, if Wedgwood already had a sensitized silver chloride with an acceptor? Quantity. The chloride in a tawed skin is a contaminant. Talbot put in as much chloride as the sheet could carry while still leaving a large silver excess — Reilly’s six-to-one — and got a sheet with orders of magnitude more silver chloride in it, which is orders of magnitude more image. Wedgwood was running Talbot’s chemistry at a concentration set by a tanner.

Davy came within one sentence of the improvement and did not see it

The paper contains the answer to its own central problem, in the passage after the proportions:

In comparing the effects produced by light upon muriate of silver with those produced upon the nitrate, it seemed evident that the muriate was the most susceptible… The nitrate, however, from its solubility in water, possesses an advantage over the muriate; though leather or paper may, without much difficulty, be impregnated with this last substance, either by diffusing it through water, and applying it in this form, or by immersing paper moistened with the solution of the nitrate in very diluted muriatic acid.

Take that apart. He has established, by direct comparison, that silver chloride is the more sensitive salt. He has identified the reason silver nitrate is used anyway — it dissolves, and silver chloride does not, which is precisely Reilly’s point that silver chloride can never be dissolved and coated but must always be formed in situ by double replacement. And he then gives two working methods for getting silver chloride into a sheet, the second of which is the two-step in-situ formation that the whole of salted-paper photography would be built on. Eder makes the claim explicitly: this publication contains the first account of producing silver chloride paper by successive applications of silver nitrate and chloride solutions on leather and paper, and it served as the starting point for Talbot’s later method.

What Davy did not do was combine his own three findings — that the muriate is more sensitive, that it can be made in the sheet, and that the sheet gives its best result on the substrate that (unknown to him) was quietly making the muriate for him. He treated the chloride route as a footnote to the nitrate route. Talbot treated it as the process.

⚗A detail worth noticing: Davy's chloride route runs backwards

Davy’s second method is silver first, then halide: moisten the paper with silver nitrate, then immerse it in very dilute muriatic acid. Talbot’s is halide first, then silver.

The order is not cosmetic, and the reasoning here is the course’s rather than either author’s. Immerse a silver-loaded sheet in a chloride bath and the chloride is in excess at the surface where the two meet, so the crystals form jacketed in chloride rather than in silver — Ware’s “fixed” silver halide rather than his “sensitized” silver halide — and the free silver nitrate that was supposed to be the halogen acceptor is progressively consumed by the bath. Dip a chloride-loaded sheet in a strong silver bath instead, and the silver is the species in excess everywhere, which is the state that prints out.

Davy gives no strength for the acid beyond “very diluted”, so there is nothing to enter in the table and no way to check the argument against what he actually did. It stays a reading.

And why the picture died anyway

Nothing above helps with the ending, because the ending is not about the image. Wherever the shadow had fallen, the sheet was still fully loaded with unchanged light-sensitive material, and a room lit for human eyes still delivers photons at the wavelengths that material absorbs. So the highlights go on darkening. The shadows cannot outrun them, because print-out silver is self-limiting and they are already near their ceiling. The gap closes from one side only.

Davy tested two remedies and published both failures. A thin coat of fine varnish “has not destroyed their susceptibility of becoming coloured” — of course not; a transparent varnish is transparent to the actinic wavelengths too, and encapsulation is not deactivation. Repeated washings left “sufficient of the active part of the saline matter” adhering to the white parts to darken them in sun — of course again, because water removes soluble silver nitrate and leaves silver chloride, which is among the least soluble things in the darkroom.

His own explanation of why the washing failed is wrong in its detail and right in its shape:

it is probable that both in the case of the nitrate and muriate of silver, a portion of the metallic oxyde abandons its acid to enter into union with the animal or vegetable substance, so as to form with it an insoluble compound.

There is no silver–collagen compound. What there is, is metallic silver, physically locked into the fibres — and Scheele had already shown as much, in an experiment that had been in print for more than twenty years. Eder’s verdict, quoted by Litchfield, is that Wedgwood and Davy “had forgotten, or not known, Scheele’s important discovery that white chloride of silver is completely dissolved in ammonia, but that when darkened by light it leaves behind a deposit of silver”. Read that sentence for what it contains: it identifies the black product and hands over a fixer, in one operation. The full story belongs to Silver salts and light.

Davy’s structural claim — that something insoluble had formed in union with the support — is nonetheless correct, and the sentence after it is the one that should have finished the job: “it is not improbable, but that substances may be found capable of destroying this compound either by simple or complicated affinities.” He was looking for a solvent, and looking for it in the wrong place. The compound he proposed to destroy was the image; the thing that needed dissolving was the residue around it. He never came back to the question. The accounting of every fixer the period did try, and why each one failed, belongs to the problem of permanence.

Function of every ingredient

Silver nitrate, 100 g to a kilogram of water — Davy’s one part in about ten.

What it is. The lunar caustic of the alchemists: colourless crystals, AgNO₃, molar mass 169.873, made by dissolving silver metal in nitric acid, and the only silver salt in ordinary photographic use that dissolves freely in water. Every silver photographic process begins with it for that single reason. The chemical’s own page carries the properties, the hazards and the history.

Why it is here. Because there is no other way to get silver ions into a sheet of anything. Silver chloride, bromide and iodide are the sensitive salts, and all three are insoluble; Reilly states the consequence as a rule — the light-sensitive substance is always formed in situ, by double replacement inside the material, because it cannot be dissolved and coated. Silver nitrate is the delivery vehicle for that reaction, and in this formula it is also, accidentally, one of the two reagents.

What it does. On its own, nothing photographic: nitrate is fully oxidised and cannot supply the electron that Ag⁺ → Ag requires. In contact with the support it becomes sensitive, because light drives the oxidation of the cellulose or the collagen and the silver takes the electron. In leather, part of it is converted to silver chloride by chloride already in the skin, and that fraction is the properly photographic part of the material — a crystal in which photoelectrons move, silver ions migrate, and specks of metal grow at the surface. The rest of it, which is nearly all of it, stays as free silver nitrate and works as the halogen acceptor that keeps the photolysis running forwards.

What follows on the sheet. Greys through browns to nearly black, in two or three minutes of direct sun.

More. More silver means more free silver ions, a higher acceptor concentration, more chloride recycled, more image silver, and a darker maximum — which is exactly the direction Talbot pushed, to 18–23 per cent w/v on Ware’s reading. It also means a stiffer, more discoloured, more corroded support: Davy’s stated ceiling is a quantity “sufficient to enable it to become tinged, without affecting its composition or injuring its texture”, and past that the sheet is damaged. And it means a sheet loaded with an oxidiser, which matters when it is dried at a fire.

Less. Fewer silver ions, less metal, a fainter and slower image. There is also a floor that Davy could not have known about. Weaken the bath far enough that the silver no longer greatly exceeds the chloride the skin supplies, and the halide stops being in deficiency: the crystals stop carrying adsorbed silver, the free nitrate that was acting as halogen acceptor runs short, and the sheet stops printing out. Talbot mapped that same failure from the other end, by adding too much salt, and his own page sets out what he found.

What it interacts with. The support, which is its electron donor. Chloride, which converts it to the sensitive salt and, in excess, deactivates the sheet. Water, which is half the halogen acceptor and also the reason the sensitiser works better moist. Light. Skin, cloth, and everything else it touches.

Water, ten parts to one, and it is not a diluent.

It does three separate jobs and only the first is obvious.

It is the solvent, and the proportion in which it is given is what makes this an eighteenth-century formula rather than a nineteenth-century one: ten parts of water added gives a molality, not a molarity, and it is reproducible with a balance and no glassware. That is a real virtue and it is why the ratio has survived three transcriptions intact.

It is the thing whose amount sets the strength, so more water is a weaker sheet — less silver delivered per unit area, a fainter image, and eventually a sheet that will not tinge at all. Less water is a stronger sheet, up to Davy’s stated ceiling, beyond which it injures the support.

And it is a reagent in the finished dry sheet. Ware’s account puts water and silver(I) ions together as the major halogen acceptor for a print-out material, and notes that cellulose paper holds around 8 per cent w/w of water at ordinary humidity. Davy observed the consequence directly and did not recognise it: both salts were “more readily acted upon when moist than when dry”. A sheet used damp prints faster than the same sheet used dry, and the reason is not that the silver is more mobile but that there is more acceptor available to take up the chlorine.

The support — white leather, or white paper — which is an ingredient and which the schema cannot hold.

It has no line in the table because a formula’s ingredient list is for things you measure into a vessel, and the enum in material: has no value for leather. It is nonetheless half the formula.

On paper, it is the electron donor and nothing else: cellulose plus whatever the mill used as internal sizing, oxidised by light so that silver ions can be reduced. That works, slowly, and it is the case Davy calls less readily acted upon.

On white kidskin, if Gray’s explanation is right, it is both the electron donor and the source of the halide. The 1911 Britannica gives the tawing paste as alum, salt, flour, egg yolk and water; the chloride from that salt, retained in the skin, converts part of the applied silver nitrate to silver chloride where it lies. The skin therefore supplies the crystal that the formula never mentions.

Two consequences follow and both are practical. First, the formula’s key variable was uncontrolled: the Britannica says outright that tawing quantities “diverge widely, every dresser having his own recipe”, so two skins from two dressers are two different sensitivities. Second, the formula cannot be reproduced today by buying white leather, because modern white leather is not tawed kidskin and carries no comparable chloride. Anyone repeating the experiment on modern leather is running the paper case with extra steps.

⁇Is the tawing explanation right?

What is not in dispute. Davy reports leather as more readily acted upon than paper. Silver chloride is more light-sensitive than silver nitrate on cellulose — Davy says so himself, from a direct comparison. Tawed leather is made with a paste containing common salt.

What is an interpretation. That the specific reason Wedgwood’s leather outperformed his paper was residual chloride forming silver chloride in place. That is Michael Gray’s explanation, reported by Mike Ware, and the course carries it because it is the only mechanistic account on offer and because it is consistent with everything else in the paper. It has not been tested against a specimen, and no specimen is known to exist to test it against.

Where the sources do not quite line up. Ware’s report of Gray names “a solution containing aluminium chloride among several other reagents”. The 1911 Britannica’s account of the trade names alum (potassium aluminium sulfate) and common salt, not aluminium chloride. In a paste containing both, the distinction is largely nominal — aluminium ions and chloride ions are present either way, and it is the chloride that matters to the argument — but the two descriptions are not the same statement, and this page records that rather than smoothing it over.

The rival explanation, from 1854, and it is wrong for an interesting reason. The Rev. J. B. Reade told Robert Hunt that he was aware Wedgwood “found leather more sensitive than paper”, and that “it is highly probable that the tanning process, which might cause the silver solution to be more readily acted upon when applied to the leather, suggested my application of the tanning solution to paper”. Reade assumed leather’s advantage was tannin. On Gray’s account he was wrong: the fine white kidskin was tawed, not tanned, and contained no significant tannin at all. But acting on the wrong premise, he put a tannin infusion onto silver-nitrate paper and got a strikingly better material — because gallic and tannic acids are developing agents, working by an entirely different mechanism from the one he was trying to copy. That line runs directly into the gallo-nitrate of silver and the priority quarrel around it.

Interactions

Silver nitrate with the support. The only interaction inside the formula, and it is the whole formula. Two versions of it run at once on leather: a slow photoredox between silver ions and the organic substrate, and a fast one inside whatever silver chloride the skin’s own chloride produced. Change the support and you change the process, which is exactly why this entry cannot be given as a procedure.

Silver nitrate with chloride, deliberately. Everything after 1839 is this interaction under control. Add chloride short of the silver and you get a sensitised, printing-out material. Add chloride in excess of the silver and you get the opposite — a chloride-jacketed crystal that repels photoelectrons and cannot shed its halogen, which is why a strong salt bath was Talbot’s usual fixer and why silver nitrate and sodium chloride are listed as incompatible below. The same two substances sensitise or desensitise according to which one is in excess, and nothing in the 1802 paper indicates that either author saw this.

Silver nitrate with water, at two different points. In the bottle it sets the strength. In the dry sheet it is half the halogen acceptor. The two roles pull in the same direction for once: a wetter sheet at exposure prints faster.

The image with water and soap. Davy’s own test, and the result is a real distinction rather than a null one: the silver is not removed, the residual salt is not removed either, but for different reasons — the first because it is metal locked into fibre, the second because silver chloride is insoluble. Washing separates neither. It only removes the free silver nitrate, which is the one component the sheet could most afford to keep.

Silver nitrate with varnish. Not a chemical interaction at all, which is the point. A physical barrier over a photochemical problem does nothing, because the barrier is transparent at exactly the wavelengths that cause the trouble. The attempt is worth recording because it is the intuitive move and it is instructive that it failed.

Silver nitrate with skin. Davy’s warning is the first published safety instruction in photography: these salts “produce a stain of some permanence, even when momentarily applied to the skin, and in employing them for moistening paper or leather, it is necessary to use a pencil of hair, or a brush.” The stain is the same reaction as the picture, running on the same kind of substrate.

Variants

The muriate route, published in the same paper. Davy gives two ways to get silver chloride into the sheet — diffusing it through water and applying it as a suspension, or moistening the paper with the nitrate solution and then immersing it in very diluted muriatic acid. Neither carries a strength, so neither can be an entry of its own; the first is the method Reilly explains cannot work well, since silver chloride cannot be dissolved and coated; the second is the two-step in-situ formation, run in the reverse order from the one that would prevail. Eder identifies this passage as the first published account of making a silver chloride paper by successive applications, and as the starting point for Talbot.

On paper instead of leather. The same solution, explicitly the worse result, and the version most often reproduced in secondary accounts because “paper” is easier to picture than “white kidskin”. It is a different material with a different mechanism, and the paper says so.

Reade’s tanned paper, 1836–37. A descendant rather than a variant: leather was hard to replenish, so Reade applied a tannin solution to paper instead, reasoning from Wedgwood’s leather. He dates the work himself to “1836 and following years”; Litchfield’s account of the solar-microscope photography puts it at 1837, and the course gives both rather than choosing. It made a far better material for a reason he had not anticipated, and it leads to the gallo-nitrate of silver. The contested callout above sets out why his premise was probably wrong and his result right anyway.

Talbot’s photogenic drawing paper, 1839. The successor, and the change is not the ingredient list — it is still silver nitrate, water and a chloride — but the control. The halide is introduced deliberately, before the silver, and kept far short of chemical equivalence. That one decision buys orders of magnitude more silver chloride, a large and deliberate excess of free silver as acceptor, and the first sheet fast enough for a camera.

The salted paper sensitiser with the plain salting solution, which is the modern form. Reilly specifies a 2 to 4 per cent salting bath and a 12 per cent silver bath, and puts the ratio at about six times more nitrate than chloride. Set that beside Davy’s roughly 9 per cent by weight and one honest observation survives: the silver strength barely moved in a hundred and eighty years. Everything that changed is on the other side of the sheet — a controlled halide, a chosen binder, and a fixer.

There is no status: variant entry derived from this formula, and there will not be. A variant under Rule 6 has to say what was changed against a base and what follows; changing anything here would mean supplying quantities the source never gave for a substrate that no longer exists, which is a new formula wearing a borrowed name. Reconstructions belong to the person who publishes them, under their own name and with their own evidence.

Safety

Level B, and the level is set by the one substance in the bottle.

Silver nitrate carries the harmonised European classification H272 (may intensify fire; oxidiser), H314 (causes severe skin burns and eye damage), H400 and H410 (very toxic to aquatic life, with long-lasting effects), signal word Danger. The aggregated notifications held by PubChem agree, and the airborne limits are unusually low: HSE’s EH40 gives soluble silver compounds, as Ag, a long-term limit of 0.01 mg/m³. Repeated exposure causes argyria, which is permanent. The chemical’s own page carries the full classification and the silver nitrate handling SOP carries the regime: sealed splash goggles rather than glasses, nitrile gloves, an apron, a dedicated spatula, weighing into a tared glass or plastic vessel and never onto paper, and eyewash within reach before the cap comes off.

What is not a hazard here, and why. This is a single dilute salt in water. There is no acid, no alkali, no sulphide, no cyanide, no volatile solvent and no reaction that generates a gas, so nothing in the making of the solution requires ventilation as a control and nothing about it is an inhalation risk once it is in solution — the exposure limit above applies to the dust of the solid and to mists, which is a weighing hazard, not a brushing one. It is not flammable. At Davy’s strength it is weaker than the modern salted-paper bath. None of that makes it a mild substance: the course treats any silver nitrate solution as corrosive, because it has no source that establishes a concentration below which the classification stops applying, and because the practical failure mode — a splash in an eye — does not care about the difference between 9 and 12 per cent.

Two hazards peculiar to this formula.

The first is the stain, which Davy identified and which is the reason for his brush. It appears hours after the splash, it is metallic silver in the outer skin, and it goes when the skin does. The stain itself is cosmetic; a splash large enough to burn is not, and first aid governs there. Either way it is the most reliable evidence you will get that your gloves failed.

The second is an oxidiser on an organic sheet, dried at a fire. Silver nitrate is not itself combustible but it accelerates the burning of combustible material, and a sheet of paper or leather loaded with it and held near a flame is the textbook case. Every practitioner of the period dried at the fire because there was no other heat source. Nobody should copy that, and nothing in this course asks anyone to.

And the exposure. Direct sunlight was the light source, for minutes to hours, with the sheet in the open. The ordinary sun-protection reasoning applies to the operator, not to the material; where the course uses a UV source instead, the UV unit operation SOP governs.

Storage

The solution. Nothing is published, by anyone, about how long this bath keeps. What is established elsewhere and applies to any silver nitrate solution: CAMEO’s datasheet records the substance blackening on exposure to light or to organic material, so amber or dark brown glass, capped, cool and out of the light; never a container that has held food or ever will; labelled with the strength, the date and the hazard following the labelling SOP. Note that “the strength” is awkward for this formula in particular, because the honest label is 100 g AgNO₃ per kg water, not a per cent w/v.

The sensitised sheet. Davy publishes no keeping time and gives no instruction for storing an unexposed sheet, which is itself informative: the sheet was made and used, and the paper reads throughout as describing a bench operation rather than a stock of prepared material.

The finished picture, which is the only storage question the source really answers. Three instructions, all of them Davy’s, and all of them correct:

The copy of a painting, or the profile immediately after being taken, must be kept in an obscure place. It may indeed be examined in the shade, but in this case the exposure should be only for a few minutes; by the light of candles or lamps, as commonly employed, it is not sensibly affected.

That is a light budget and a safelight, in 1802, arrived at entirely empirically. Both are right for the reason the course teaches: the material absorbs at the blue and violet end, a candle or oil lamp is a low-temperature thermal source that emits very little there, and the useful question about any illuminant is never how bright it is but whether the material absorbs what it emits. Part XVI’s safelight tables are the same question with numbers attached.

Long-term, there is nothing to report and that is the finding. No Wedgwood specimen is known to survive. The two pictures reproduced as his through the nineteenth century were disposed of in 1864 — the “Breakfast Table” shown to be a Talbot photograph of about 1841 whose negative and other prints were extant at Lacock, and the “Savoyard Piper” never evidenced at all. So the permanence of a Wedgwood picture cannot be measured, only reasoned about; and the reasoning says a sheet that was never fixed or even stabilised is in a worse position than the chloride-stabilised photogenic drawings whose conservation data the permanence lesson carries.

Incompatibilities

The list belongs to silver nitrate and it is unusually consequential for a substance this simple.

Ammonia and alkali together. An ammoniacal silver solution treated with a strong alkali can deposit silver nitride, also called fulminating silver, which detonates on contact and sometimes when wet. This is not hypothetical and it is the reason the course carries an ammoniacal silver quench SOP. Nothing in this formula calls for ammonia — but ammonia was the fixer Wedgwood and Davy needed and did not use, so a reader following that thread has a real chance of putting the two together.

Alcohols, acetylene, phosphorus and tin(II) chloride, all of which CAMEO records as giving explosive or shock-sensitive products with silver nitrate.

Organic material in general, because it is an oxidiser. Absorbents, wipes, gloves and cloths that have taken up a silver nitrate spill are not left in an open bin to dry.

Chlorides, which is the one incompatibility this formula deliberately exploits and which is therefore worth stating precisely. Add sodium chloride or hydrochloric acid to a silver nitrate solution in a vessel and you get a useless curd of silver chloride and a ruined bath. Bring the same two substances together inside a sheet and you get a photograph. The difference between those two outcomes is the whole of nineteenth-century silver printing.

Thiosulphate, sulphide and any developer, from the ordinary contamination standpoint: a trace of fixer in a silver bath destroys it, and the traffic runs the other way too.

Waste

Silver-bearing, and treated as an asset rather than as an effluent. The classification is unambiguous — very toxic to aquatic life with long-lasting effects — and the regulated concentrations are small: Kodak’s own literature gives a mean municipal limit around 1.2 mg/L of silver, and one spilled gram of silver nitrate carries about 0.64 g of silver, enough to push half a cubic metre of water past it.

The recovery method is the incompatibility above, used on purpose: add a chloride to the collected solution, let the insoluble silver salt settle, decant the clear liquid and send the solid to a refiner. The silver-bearing waste SOP sets out the procedure and the chemical’s page the reasoning. Smelting and re-nitrating the recovered solid are not home operations.

Two things specific to this formula. A sensitised offcut is silver-bearing solid waste — leather, paper or anything else that has taken the solution — and goes to the same stream rather than to the bin. And a finished, unfixed picture is still loaded with light-sensitive salt, so it is not inert: it is a piece of silver-bearing material that will go on darkening wherever it is put.

Disposal is governed by local regulation everywhere, and the rules differ between jurisdictions and change. The disposal page states the general principles; check what applies where you are.

Troubleshooting

There is no procedure here to troubleshoot, so what follows is the other kind: the failure modes the source itself reports, and the failures of reading that this paper attracts more than any other document in Part I.

The failures the paper publishes.

Symptom	Cause, as the source gives it
The picture fades until it is gone	The ground is still fully loaded with light-sensitive salt and goes on darkening; the shadows are already near the print-out ceiling and cannot outrun it
Varnishing does not stop it	A transparent coating is transparent at the actinic wavelengths; encapsulation is not deactivation
Repeated washing does not stop it	Water removes free silver nitrate and leaves the silver chloride, which is essentially insoluble
Nothing appears in a camera obscura	Print-out needs light energy the camera image cannot supply in any moderate time
A copied print comes out mushy	The transmitted lights from the shaded parts are “seldom so definite as to form a distinct resemblance”; paper scatters and the mid-tones do not separate
The sheet is stiff, discoloured or damaged	The solution was stronger than the ceiling Davy states — enough to tinge, not enough to affect the composition or injure the texture

The failures of reading.

What you will meet	What the evidence says
“The 1802 paper contains no quantities”	It contains one, and it is a formula. Hunt’s much-reproduced 1854 transcription elides it behind a row of asterisks, and most later accounts descend from Hunt
“The colour was fixed on the leather” means the picture was fixed	It means the colour had been produced and would not wash off. The picture was never fixed, and the same paper says so two paragraphs later
“1802 is the date of the experiments”	It is the date of the publication. No paper in that volume of the Journals carries a date, and the experimental work is undated. Ware puts the first contact images around 1797 and the photograms around 1800; Litchfield points to a letter of November 1800 about coloured glasses and solar-microscope parts, and to a Watt letter of about 1790–91 mentioning “Silver Pictures”. The permanence lesson sets out what the course does and does not assert
“Wedgwood and Davy worked on it together”	The title says invented by one and observed by the other. Litchfield finds no evidence that Davy did more than the solar-microscope work and the write-up, and Eder is emphatic that the credit for the invention is Wedgwood’s alone
“The Savoyard Piper is the first photograph”	It is not, and neither is the “Breakfast Table at Etruria Hall”. Both attributions are Eliza Meteyard’s; the second was shown in January 1864 to be a Talbot photograph of about 1841 from Lacock Abbey, with the negative extant, and no evidence was ever produced for the first
“The paper was published and the world ignored it”	Litchfield’s correction is sharper and more useful: the Journals of the Royal Institution was a subscriber bulletin that ran to a single volume, there is nothing to show the account was ever read at a meeting, and it can hardly be said to have been published in any effective sense
“They could not fix it because hypo had not been discovered”	Chaussier had synthesised thiosulphate in 1799, three years earlier. Ware’s point is stronger: thiosulphate is not an essential prerequisite at all, since halide-stabilised and water-washed silver images survive to the present. What was missing was knowledge, not chemistry
“It failed because silver nitrate is a poor sensitiser”	Silver nitrate is not a sensitiser at all on its own. It becomes one only in contact with something oxidisable, and it becomes a good one only when a halide crystal is present. That is the finding the formula is built on and the one its authors did not have
Reproducing it on modern leather and getting nothing much	Modern white leather is chrome- or vegetable-tanned, not tawed with an alum-and-salt paste. Without the chloride you are running the paper case, and the paper case is the slow one

Experiments

Every experiment below is run with chemistry the course publishes properly, not with this formula. That is the point: what is worth testing here are the claims, and the claims can be tested on materials a reader can actually obtain.

1. Davy’s colour ordering, with modern filters. He ranked red, yellow, green, blue and violet by their effect on the sheet using nothing but coloured glass. Repeat the ranking on modern printing-out material with gel filters over a step of identical exposures, and record the order. The unfixed silver photogram sets the procedure out in full. Predict the order first, from the absorption of silver chloride, then measure it. Hypothesis: the order will be Davy’s. Control: an unfiltered strip.

2. The chloride question, on paper you can buy. Gray’s explanation says the leather worked because it contained chloride. That is testable without leather. Coat two sheets of the same paper with the salted paper sensitiser; pre-treat one with the plain salting solution and leave the other unsalted, so that one sheet has silver chloride in it and the other has silver nitrate on cellulose alone. Expose both together under the same object. Hypothesis: the salted sheet prints out far faster and far darker. The variable: presence of a halide. What it tells you: how much of Wedgwood’s leather advantage a trace of chloride could plausibly account for.

3. Wet against dry. Davy reports both salts as more readily acted upon moist than dry, and Ware explains why. Expose two identically sensitised sheets side by side, one damp and one dried, and compare the density reached in the same time. The variable: water in the sheet at exposure, which is half the halogen acceptor. Record relative humidity, because it sets the “dry” condition.

4. The order-of-addition question. The reading in the mechanism section says Davy’s chloride route — silver first, then dilute acid — should print out worse than Talbot’s halide-first order, because it consumes the free silver that acts as acceptor. It is a reading, and readings should be tested. Using only the course’s published salted-paper chemistry, prepare one sheet salted-then-silvered and one silvered-then-salted at matched strengths, and compare maximum density and speed. Hypothesis: the halide-first sheet is faster and denser. Note what the second sheet looks like immediately after the second bath; the appearance of the crystals is part of the evidence.

5. Do the arithmetic that the schema refused to do. This page states no per cent w/v for Davy’s bath, on the grounds that no final volume is given. Find out what it would take to state one: what additional measurement is needed, why a handbook of 1802 would not have had it, and what the number would be if the solution’s density were, say, 1.07 g/mL — a figure invented for the exercise, because the course holds no measured density for this solution. Then explain in two sentences why the course still would not print it.

6. Read four printings of the same document. The 1802 account exists here in Davy’s collected works of 1839, Hunt’s transcription of 1854, Litchfield’s reprint of 1903 and Eder’s in translation. Read all four and list what each one drops or changes. You will find at least this: Hunt loses the formula entirely; the collected works and the 1903 reprint disagree on a word in the fixing sentence, where one scanned copy reads “not commonly affected” and the others “not sensibly affected”; and the same two disagree on whether it is the “metallic oxyde” or the “metallic acid” that abandons its acid — of which only the first is chemically possible. None of those is a large matter on its own. Together they are the reason a page like this one cites the printing it actually read rather than “the 1802 paper”, and the reason the sources page exists.