Niépce's bitumen heliograph coating

The coating that made the first photograph is a glass half filled with powdered tar, oil dripped on until the tar will take no more, and enough oil again to stand three lines deep over the top. That is the whole of the composition, it is the whole of what Nicéphore Niépce wrote down, and no source in the two centuries since has supplied a number for any part of it.

He wrote it out on 5 December 1829, in duplicate and for one reader, as a supplement to the partnership agreement he signed with Daguerre nine days later. Daguerre printed it in 1839, six years after Niépce was dead, with his own footnotes correcting his late partner in the margins. It is the earliest exact description of a photographic process that exists — Eder’s judgement, not this course’s — and its composition is fixed by a vessel nobody ever measured.

This formula's source publishes no quantities. "I about half fill a wine-glass with this pulverised bitumen. I pour upon it, drop by drop, the essential oil of lavender, until the bitumen is completely saturated. I afterwards add as much more of the essential oil as causes the whole to stand about three lines above the mixture, which is then covered and submitted to a gentle heat until the essential oil is fully impregnated with the colouring matter of the bitumen. If this varnish is not of the required consistency, it is to be allowed to evaporate slowly, without heat, in a shallow dish." The French of 1839 says "un verre", a glass, and gives no capacity for it; no mass, volume, percentage or ratio appears in the notice or in any later printing of it.

The composition is fixed by the vessel — the source names a vessel, or a part of one, and no amount. By the vessel, twice, with a saturation between the two. The bitumen is fixed as a proportion of a glass — half of it, filled with the dry powder — and the glass is never named, sized or described. The first oil is fixed against the powder, dropped on until it will absorb no more, which is a real end point anybody with a dropper can reach. The second oil is fixed against the vessel again, poured until it stands about three lines above the mixture, and the ligne is a real unit of about 2.26 millimetres by Eder's own footnote, so this is a genuine measurement of depth over a cross-section that is never stated. The composition is then closed by a consistency: the mixture is digested in a gentle heat until the oil has taken up the colouring matter, and evaporated in a dish if it is thinner than the work wants. What that leaves undetermined is the ratio of bitumen to oil, and therefore the thickness, the compactness and the sensitivity of the coating — which is the whole of the formula's behaviour, since Niépce himself says the varnish must be thick enough to lay a compact and as thin as possible layer, because such a layer resists the solvent better and is more sensitive to light. Two glasses of different proportions, filled and topped by these instructions, give two different varnishes. A modern reader can repeat every operation and cannot repeat the mixture.

The amounts below are the source's own words, not measurements, and the table is not a recipe. 12 documents were read for a published quantity and none states one; they are listed in this page's sources.

Ingredient	As the source states it	Form the source specifies
Bitumen of Judea	“about half a glass, filled with the pulverised bitumen”	Latin-labelled in the trade as bitumen judaicum; French "bitume de Judée", English asphaltum. Ground to a powder before use. The French says "un verre"; the wine-glass of the English literature is Memes's rendering of 1839
Oil of lavender	“poured on drop by drop until the bitumen will absorb no more, then as much again as stands about three lines above the mixture”	Essential oil of lavender, the distilled oil and not the residue. Hunt adds that the English oil was too expensive for the purpose and that redrawn French oil answered as well or better, which is a statement about supply and not about strength
Water	none — this preparation contains no water	There is no water in this preparation. The vehicle is the essential oil of lavender itself, and water is not merely absent but excluded: damp alters the varnish and finally decomposes it, so the mixture is covered while it digests, the coated plate is dried over hot iron under paper from which the moisture has been driven, and Niépce holds a light metal disc in front of his mouth while coating so that his breath cannot condense on the plate. Water enters the process only after development, as the lukewarm stream poured down an inclined board to carry off the last of the solvent.

☠Level D: this page describes a preparation and gives no procedure

Two organic solvents, worked in quantity and warmed. Oil of lavender has no GHS classification in any document this course has read — no PubChem compound record, no ILO-WHO card, no entry in HSE’s EH40 — which is an absence of data and not a finding about the oil. Bitumen of Judea has no classification either, and the nearest classified material to it, petroleum asphalt, carries the statement that its fumes are possibly carcinogenic to humans. White petroleum oil, the bulk of the developing solvent, has no entry in this encyclopaedia at all. Both of the substances that do have entries are Level D, and where the historical purpose is an etched printing plate, nitric acid joins them.

No procedure appears on this page. There is no apparatus, no drying time, no temperature and no quantity — and the quantities are not withheld, they were never published.

The coating a reader can actually make, with every figure sourced, is the classic cyanotype sensitiser at Level B. It is a different chemistry, and it is the nearest thing in this course to what this page describes: a light-sensitive layer, spread by hand, that turns a shadow into an object.

The Part I lesson owns the history and the argument about what Niépce achieved; the process atlas entry owns the workflow and the surviving plates; bitumen of Judea owns the material. This page owns the notice as a formula: what went in the glass, what fixed the composition in the absence of a balance, and what a reader can and cannot take from one paragraph of French.

Purpose

To make a layer that light will render insoluble, so that a solvent can wash away everything the light did not reach. That is the entire mechanism of the process and it is the opposite of every silver formula in this book. Nothing is reduced, nothing is developed in the chemical sense, and nothing is amplified. The varnish is a photoresist, two generations before anyone had a use for the word, and the picture is the part of the coating that survived.

Niépce’s own definition is wider than that, and his commercial purpose narrower. The notice opens by saying that heliography consists in the automatic reproduction, by the action of light and with their gradations of tone from black to white, of the images obtained in the camera obscura. What the applications section is actually about is a printing plate: the varnish goes on stone, metal or glass without any change of handling, and on copper the hardened bitumen is a resist for an acid etch, so the picture can be inked and run through a press. The two purposes came apart in 1827, when he gave up etching his camera views because a continuous-tone image has no way of instructing an acid. The camera view is the half that survives in every history, and it is the half he was least able to use.

Recommended uses

None practical. This is a historical-study entry, and the honest answer to “what is it for” is reading, dating and understanding.

Three things it is genuinely good for in a modern course.

• Reading the 1829 notice. Anybody who meets bitume de Judée, trois lignes, huile de pétrole blanche or “half a wine-glass” in a history can find out here what the words were attached to and which of them is a translator’s.
• Meeting a composition fixed by a vessel. Schulze fixes his mixture by saturation; Niépce fixes his by how full a glass is. Both are real, repeatable operations, and neither is a number. It is worth meeting both before reading Kodak, where every line is a weight.
• Seeing where the resist line begins. Everything in photogravure, photoglyphic engraving, photolithography and semiconductor manufacture descends from the operation this paragraph describes. Change the resin, change the solvent, keep the logic.

When another formula is preferable

Always, for anything a reader intends to make.

For a light-sensitive coating with every quantity published: the classic cyanotype sensitiser. Level B, two salts, a brush and a photogram at the end of it. It is a print-out iron process rather than a hardening one, so the chemistry is not the same; what is the same is the act this page cannot teach — laying a sensitive layer on a support by hand and getting a picture off it.

For the nearest relatives in kind: gum bichromate and carbon printing, where light hardens a colloid instead of a tar. Both have atlas entries in this course and neither has a formulary entry yet, so there is nothing here to send a reader to for quantities; the atlas entries say what the processes are and what they need.

For the other formula in this book whose source publishes no quantities: Schulze’s scotophorus mixture of 1727, which is the same problem a century earlier and fixed by a different operation.

For the oldest photographic formula a reader can actually mix: Wedgwood and Davy’s silver nitrate on leather, 1802 — one part of silver nitrate to about ten of water, which is a ratio, and a ratio is a quantity.

For the same document read as a formula: Daguerre’s 1839 process outline, printed in the very volume that carries this notice, and the entry that shows what this course does when a source gives proportions and vessels rather than weights: it scales them and says so.

Mixing

Niépce’s paragraph is four operations, and the order is not negotiable because each one is defined against what the last one left.

One: the powder. Half fill a glass with pulverised bitumen of Judea. The French is “un verre”, without qualification; every English printing read for this page before Eder’s makes it a wine-glass.

Two: the oil, to the absorption limit. Drop essential oil of lavender onto the powder, drop by drop, until the bitumen will absorb no more of it — “et qu’il en soit seulement bien pénétré”, until it is just well penetrated by it. This is the step that is done by watching rather than measuring, and it has a real end point: a saturated powder stops taking up liquid and the next drop stands.

Three: the oil, to the vessel. Add as much more oil again as will make it stand about three lines above the mixture. Eder footnotes the ligne for his reader as 2.256 mm, so three of them is a little under seven millimetres, and this is a real measurement of a depth. It is not a quantity of oil, because a depth over an unstated cross-section is not a volume.

Four: the heat, and the consistency. Cover the glass and leave it in a gentle warmth until the added oil has taken up the colouring matter of the bitumen. If the varnish is then thinner than the work needs, pour it into a shallow dish and let it evaporate in the air, without heat, keeping the damp off it.

!What fixes the composition, since the numbers do not

The vessel does, twice, with a saturation in between — and the vessel is the term that makes this formula unscalable.

Both terms of the composition are stated against the same glass. The bitumen is half of it. The free oil is three lines standing over what is in it. So the proportion of free oil to bitumen is determined by how tall the glass is, not by how big it is: for a straight-sided vessel, doubling the bore doubles both the powder and the oil and changes nothing, while doubling the height halves the proportion of free oil. A wine glass is not straight-sided, so the taper of the bowl enters as well. That arithmetic is the course’s, not Niépce’s, and it is offered to show what the instruction actually determines rather than to reconstruct a recipe from it.

The saturation step is real and is the one part of this a modern reader could hit exactly: a powder stops absorbing at a definite point and anybody can find it. What it fixes is only the bound oil. The free oil, which is what makes the mixture pourable and sets how thin a layer it will lay, is fixed against the glass.

And the composition is closed by a judgement. The varnish is used at a consistency thick enough to form a compact and as thin as possible coating — Niépce’s own words in the Observations — because such a layer resists the solvent better and is more sensitive to light. That sentence is the nearest thing in the notice to a specification, and it specifies a feel.

What none of it fixes is the ratio of bitumen to oil, and that ratio is the coating’s thickness, its compactness, its resistance to the developing solvent and its speed. Which is to say: it is the formula.

The coating, which is part of the preparation. A small quantity of the varnish is applied cold to a highly polished plated-silver tablet with a light roll of very soft skin, giving a fine vermilion colour and a very thin, even layer. The plate then goes on hot iron wrapped in several folds of paper from which the moisture has been driven, and comes off when the varnish has stopped simmering, to cool and finish drying in a gentle temperature away from damp air. Niépce interrupts his own directions to say that this precaution matters most while the varnish is being applied, and that a light disc of metal on a short stem, held in the mouth, will catch and condense the breath.

Behaviour

It is not sensitive in any sense a photographer would recognise. There is no latent image to develop and no gain: every photon that contributes does its own work on its own molecule, so what light did is all there is. Ware puts the speed at less than 10⁻⁶ ISO. Niépce’s own remedy was time — hours in contact, days in the camera — and the reason the process died is that no amount of skill buys back a gain the chemistry does not have.

Nothing is visible after exposure. Niépce says the impression remains unperceived, and Daguerre footnotes him to insist on the distinction: if the image were altogether imperceptible there would be no result at all, so a faint appearance of the light’s action has to be there for the plate to be worth developing.

The exposed part becomes insoluble, and that is the whole of the change. Where the light fell, the varnish stays; where it did not, the solvent takes it away and bares the metal. The process is therefore negative-working in the printing sense — the coating survives where the light was strongest — and positive-appearing on a bright plate, because the surviving coating is pale and matt and the bared metal is dark.

Damp attacks it at every stage. Moisture alters the varnish and finally decomposes it; a plate coated in a damp room sheds the varnish in sheets the moment it goes into the solvent. Niépce says plainly that this is why he stopped using the camera in the bad season.

The developing solvent is exhausted by what it dissolves. One part of oil of lavender by volume to ten of white petroleum oil, milky when mixed and perfectly clear after two or three days. It will serve several times in succession and loses its dissolving power as it approaches saturation with the asphaltum, which is visible: it turns opaque and very dark brown.

Image characteristics

Colour. The fresh varnish on polished plated silver is vermilion. The finished picture is pale, matt, hardened bitumen against darker bared metal, and it must be lit and angled correctly to read at all — the same viewing behaviour a daguerreotype has, and for a related reason: the light and shade are a difference between two surfaces rather than a pattern of densities.

Gradation, of a kind, and Niépce knew it was thin. His two experiments on glass are the whole of what the notice offers on tone. In the first, the light having acted with less intensity, the varnish was removed in proportion to it, the gradations showed better by transmitted light, and the plate reproduced up to a point the known effects of the diorama. In the second, the action having been more intense, the lightest parts were not attacked by the solvent at all and remained transparent, so the difference of tone came only from the relative thickness of the more or less opaque layers.

The colour effect he could not explain. Viewed by reflection in a mirror, on the varnished side and at a certain angle, the second plate seemed to take on the local colours of some of the objects. Niépce wondered whether it could be attached to Newton’s theory of the coloured rings — supposing that a given prismatic ray, the green for instance, acting on the substance of the varnish and combining with it, gives it just the degree of solubility that leaves a layer reflecting green. Daguerre’s footnote says he has often observed the same colouration and has never been able to consider it a result of the coloured rays in the camera. Neither explanation is endorsed here and neither is corrected. The observation is recorded because it is one of the earliest colour observations in the literature — the Encyclopaedia Britannica gives Seebeck’s of 1810 as the first — and because both men reported it and left it open.

Permanence: solved, and trivially. The material light did not act on is removed, so nothing photosensitive is left in the object. That is the one problem of early photography this process did not have, and the surviving Niépce plates examined by the Getty Conservation Institute with the National Media Museum in 2010 are the evidence for it.

The mechanism

There is no equation on this page, and that is a finding rather than an omission.

Bitumen of Judea is a natural mixture, not a compound. Ware, a chemist, describes its molecular structure as so complex and variable as to constitute a chemist’s worst nightmare. There is no molecular formula for it in this course’s encyclopaedia entry, no molar mass and no CAS number — not because nobody looked, but because there is nothing single to assign them to. A reaction equation would need a reactant.

What can be stated is the phenomenology, and it is solid:

• Light makes the exposed bitumen insoluble in the solvent that dissolves unexposed bitumen. This has been demonstrated every time the process has worked, from 1822 to Marignier’s modern repetitions, and the 2010 infrared analyses find hardened bitumen where the image is.
• Photo-induced cross-linking of the large aromatic molecules is the general shape of the explanation — it is the subject of Marignier’s paper titled “Asphalt as the World’s First Photopolymer” — and this course has not read the peer-reviewed mechanism papers and will not paraphrase a title as though it were a mechanism.
• Oil of lavender changes nothing chemically here. It dissolves the bitumen to make the layer and, diluted with white petroleum oil, dissolves the unhardened bitumen to reveal the picture. The word develop covers two unrelated operations, and this is the second one: a solvent developer, which amplifies nothing.

⚗Why this formula has no water, and why that is not a technicality

Every other preparation in this formulary is a solution of something in water, and the water is usually the largest number on the page. This one has none.

The vehicle is the oil, the solvent is oil and a petroleum distillate, and water appears in the process exactly once — the lukewarm stream poured down the washing board to carry off the last of the solvent adhering to the varnish. Water in the preparation is not merely absent, it is the enemy. Damp alters the varnish and finally decomposes it. A plate coated in a damp room loses the varnish in layers when it meets the solvent. Niépce dries his paper before he dries his plate on it, and holds a metal disc in front of his mouth so that his own breath cannot reach the surface.

water.none exists in the schema for this case, and it requires the page to say what carries the preparation instead. Without that requirement an empty water row would read as a field somebody forgot, which is the misreading the whole unquantified mechanism was built to prevent.

Function of every ingredient

Bitumen of Judea — asphaltum, the tar that is the picture. It is the only light-sensitive substance in the formula and it is also the image substance: what you see on the finished plate is the bitumen that light hardened. Niépce calls it the material that has succeeded best with him and that contributes most immediately to the effect, which is a careful way of saying he had tried others. It arrives as a dark, brittle solid and is ground to a powder so that the oil can get at it; the powder is what half fills the glass. More of it, relative to the oil, gives a thicker and more compact layer — which by Niépce’s own account resists the solvent better and is more sensitive — and past some point, on the course’s reading rather than his, a varnish too stiff to lay thin and even; the evaporation step exists to move a varnish the other way. The amount is unstated, and the consequence of the amount being unstated is that the coating’s speed is unstated too.

Two things about the material are worth carrying away from the encyclopaedia entry. It is a variable natural mixture rather than a compound, so two batches are not necessarily the same material, and no period source distinguishes them. And it has no notified GHS classification, which is an absence of assessment and not a statement about it; the nearest classified material, petroleum asphalt, is recorded on ILO-WHO card 0612 as being absorbed by inhalation of fume, as irritating to eyes and respiratory tract, and as having fumes that are possibly carcinogenic to humans — which is an inference about this material and not a classification of it.

Oil of lavender — the vehicle, and half the developer. Its first job is to make the tar spreadable: dropped onto the powder it is absorbed until the powder will take no more, and the excess above that is what turns a wetted solid into a varnish that will lay a thin even film off a leather pad. Its second job ends on the hot iron: the plate is warmed under dried paper until the varnish is no longer tacky, which is the end point Niépce gives, and the course reads that as the oil leaving the layer — he states when to stop and not why. Its third job comes after the exposure, mixed one part to ten with white petroleum oil, when it dissolves the bitumen light did not harden. The same substance therefore builds the layer and takes it away, and the only thing that decides which it does is whether the light got there first.

More oil gives a thinner, more fluid varnish and a less compact coating, which Niépce says is both less resistant to the solvent and less sensitive. Less oil gives a varnish that will not spread. The notice fixes the amount by a depth in a glass and by a consistency in the hand, and there the matter rests.

Hunt’s footnote to his translation is the only practical remark added to this ingredient in any of the printings read, and it is about the market rather than the chemistry: English oil of lavender was too expensive for the purpose, and an article sold as redrawn French oil of lavender was very much cheaper and answered as well, if not better. The encyclopaedia entry notes the modern form of the same problem — spike lavender oil and lavandin oil are kept as separate registry records, so “oil of lavender” in 1829 need not be what a supplier sells under that name now.

White petroleum oil is the third substance in Niépce’s process and it is not in the table above. It is ten of the eleven parts of the developing solvent, it has no entry in this course’s chemical encyclopaedia, and an ingredient line has to name a substance the encyclopaedia holds. The choice was between leaving the solvent out of the structured data and entering it as though it were oil of lavender alone; the second would have published a solvent with ten elevenths of it missing. So the solvent is stated here, in full and with its ratio, and the object records the coating only. A reader must not take the table above for the whole process, and this paragraph is the place the page says so.

Interactions

Bitumen against its own solvent, which is the entire process. Unexposed bitumen dissolves in oil of lavender and in petroleum distillates; exposed bitumen does not. Every other interaction on this page is a nuisance around that one.

Solvent strength against the image. Niépce is explicit that the proportions of the solvent cannot be fixed with certainty, because the solvent has to be adapted to the result wanted, and that in all cases it had better be too weak than too strong. Daguerre’s footnote gives both failure modes: too weak and the image does not come out strongly, too strong and it is completely destroyed. This is a published statement that a formula’s own author could not specify it, sitting a few lines away from the only ratio in the notice.

Water against the varnish. Covered under The mechanism and under Storage: damp alters the varnish before use and detaches it from the plate during development.

Potassium sulphide against the varnish. Niépce’s first attempt to blacken the bared silver, so that the picture might have a full range from black to white, was a liquid sulphide of potash. He reports that concentrated it attacks the varnish, and diluted with water it only reddens the metal. The double defect forced him to give it up, and the record of that failure is why the next paragraph of the notice is about iodine.

Iodine against the bared silver. The substance he turned to instead, because it vaporises at ordinary temperatures. A few grains in a groove at the bottom of an open box, the plate stood against the opposite wall, a glass over the top to watch through: the effect comes on more slowly and much more surely. Daguerre’s footnote on this passage is one of the sharper things in the volume — that Niépce’s use of iodine to blacken his plates proves he did not know the property this substance has, in contact with silver, of being decomposed by light, since here he proposes it as a means of fixing his proofs. The step that led to the daguerreotype is in this notice, being used for something else.

Ethanol against the varnish, deliberately. After the iodine has done its work the varnish can be taken off with spirit of wine, leaving no trace of the original impression. What is left is blackened silver where the coating had been washed away and bare polished silver where the light had hardened it — which is the whole reason for the step, since the picture had until then been a difference between matt tar and bright metal.

Nitric acid against the bared metal, where the object is a printing plate. Not part of this formula, and named because it is what the formula was for. The hardened bitumen is a resist; the acid bites where the resist is gone.

Variants

Niépce publishes two variants of the coating in the same eight pages, and a third change that is not to the coating at all. None of them carries a quantity either.

• With wax, for engraving on copper. A small quantity of wax dissolved in essential oil of lavender may be added to the varnish without inconvenience. Daguerre’s footnote is worth having: the wax would have neutralised the decomposition of the bitumen in the camera, where the light arrives much weakened, but it was no obstacle for the engraving copies, which Niépce exposed three or four hours to direct sun. “Une petite quantité” is the whole of the specification.
• In Dippel’s animal oil, for glass. The last paragraph of the notice is a new varnish — bitumen of Judea dissolved in Dippel’s animal oil and left to evaporate at atmospheric temperature to the degree of consistency required. Niépce made it because the ordinary varnish adheres badly to glass and is damaged in the washing, and reports it as more unctuous, more tenacious and more strongly coloured, exposable as soon as it is applied because the animal oil is volatile and dries fast. Daguerre’s footnote says that this very property further reduces the resources of the process as regards the lights of the picture. The composition is again fixed by a consistency, and Dippel’s oil has no entry in this encyclopaedia.
• On glass with a black backing. Not a change to the varnish but to the exposure, and Niépce adds it in the Additions as an improvement he did not want to omit: a sheet of black paper under the glass plate and a cardboard border between the varnished face and the engraving, which has been glued to the border and stretched tight. The image appears much livelier than on a white ground, which helps the speed, and the varnish is no longer rubbed by direct contact with the engraving — a mishap hard to avoid in warm weather even when the coating is quite dry.

What Daguerre substituted, in the same volume. The residue left by evaporating essential oil of lavender, dissolved in alcohol and laid on in an extremely thin wash. Hunt records that Daguerre found it more susceptible of change than the bitumen, and that instead of washing the plate with Niépce’s solvent — which often took all the varnish off — he exposed it to the vapour of petroleum. The Musée Nicéphore Niépce’s chronology puts the two partners onto that material in 1832; it becomes the physautotype, and it is not this formula.

Later relatives, in this formulary. None yet in kind: this is the only light-hardened coating the formulary holds. The nearest by date is Daguerre’s 1839 outline, printed in the same book, and the nearest by problem is Schulze’s mixture, which is the other entry here whose source states composition by an operation.

Safety

The hazard is two organic solvents worked in quantity and warmed, and the difficulty is that almost nothing has been assessed.

Oil of lavender — Level D, for absence of data rather than presence of hazard, and its encyclopaedia entry sets the search out with its citations. PubChem holds no compound record for it. The ILO-WHO card index holds none under lavender or Lavandula. HSE’s EH40 has no entry for it or for essential oils as a class, and EH40 states that absence from its list does not indicate that a substance is without risk. The only authority available would be the supplier’s safety data sheet for the specific oil in hand, and this course has not read one. Nothing in that paragraph is a reassurance and none of it should be read as one.

Bitumen of Judea — Level D, on the same grounds. No PubChem compound record and no GHS classification, and ILO-WHO card 0612 — which is for petroleum asphalt, a related material and not this one — has an empty classification section. What that card does record is not nothing: the substance is absorbed into the body by inhalation of fume, it irritates the eyes and respiratory tract, the heated material may cause burns, and its fumes are possibly carcinogenic to humans. Applying any of that to a natural Judean asphalt is an inference, and the encyclopaedia entry marks it as one.

White petroleum oil — not classified here at all, because there is no entry. The course will not transfer a hazard assessment from a modern petroleum fraction to a nineteenth-century article sold under that name, and it will not leave the gap unnamed either. It is ten of the eleven parts of the developing bath.

Nitric acid — Level C, where the object is an etched plate. Its harmonised classification under the European CLP regulation, as reported by PubChem, carries the signal word Danger with H272, may intensify fire; H314, causes severe skin burns and eye damage; and H330, fatal if inhaled. The etch is not part of this formula and is named because it is what the formula was made for.

Three period details from the notice, since a reader may be tempted.

• The varnish is warmed twice and the plate is dried on hot iron. Warming a volatile essential oil in an open vessel, indoors, is the operation this page describes and does not offer.
• The developing bath is an open tinplate dish an inch deep, longer and wider than the plate and holding enough solvent to cover it. Niépce works with his face over it, watching by reflected light for the image to unfold, for as long as the development takes.
• A metal disc held in the mouth is Niépce’s control for breath moisture. It is a control for the varnish, not for the operator.

The Level D policy governs what this course does with a process like this one and is cited rather than restated. There is nothing on this page a reader should attempt.

Storage

The source publishes no keeping figures, so this entry carries none. What it publishes instead is a set of conditions, and they are more use than a shelf life would be.

The varnish keeps only away from damp. Moisture alters it and ends by decomposing it. The precaution, Niépce says, is doubly necessary in winter and in rainy weather, and this is the reason he gives for not using the camera during the bad season at all.

The coated plate keeps away from damp and from light. The Additions open with the instruction that when the varnished plate is taken away to dry it must be protected not only from moisture but from any contact with light. Hunt’s English moves an equivalent sentence to after the washing, where it becomes advice about keeping the finished picture; the French has it before the exposure, where it is advice about keeping the plate sensitive. The course records the French placement and notes the divergence rather than choosing.

The developing solvent keeps and is reused until it is saturated. It will serve several times in succession and loses its dissolving power only as it approaches saturation with the asphaltum, which announces itself by the liquid going opaque and very dark. On what happens next the printings disagree: the French as scanned says the exhausted solvent cannot be distilled and made as good as before, while Eder’s English says it may be. The clause turns on a single word in a poor scan and the two English translations that would settle it, Memes’s and Hunt’s, both omit the sentence, so the course records the disagreement and resolves nothing.

The finished plate keeps. There is no sensitive material left in it once the unexposed bitumen has been washed off, which is the one permanence problem of early photography this process did not have.

Incompatibilities

Water and damp air, at every stage before development. The strongest incompatibility in the formula, and the one Niépce writes most about. A varnish that has taken up moisture is altered and eventually decomposed; a coating laid down in a damp room detaches from the plate in layers the moment it is immersed in the solvent.

Concentrated potassium sulphide, which attacks the varnish, and dilute potassium sulphide, which does not attack the varnish and does not blacken the silver either. Both of those are Niépce’s own findings and both are failures he published.

A solvent that is too strong, which removes the image completely rather than developing it. There is no recovery from this and the source says so.

Direct contact with the engraving, in warm weather. Even when the coating is quite dry, the pressure of the paper against the varnish damages it — which is what the cardboard border in the Additions is for.

Glass, weakly. Not an incompatibility of the chemistry but of the adhesion: the varnish holds less well to a smooth glass surface than to plated silver, so images on glass are more often damaged in the washing. Niépce’s Dippel-oil varnish is his answer to it.

Waste

Nothing here is prepared, so nothing here is disposed of. For a reader who meets the residue in another context: a used heliographic developer is an organic solvent mixture carrying dissolved asphalt. It goes by an organic-solvent route — not the drain, and not evaporation in a room — and the oil of lavender and bitumen of Judea entries carry what the course knows about it, which is not much, including the GOV.UK household guidance they cite: hazardous waste from a home goes to the council’s collection service. Local regulations govern.

Where the historical purpose was pursued to the end there is a second waste stream, spent nitric acid carrying dissolved copper, and its encyclopaedia entry owns it.

Troubleshooting

Every fault below is Niépce’s or Daguerre’s, published in the notice, and each is recorded as history.

The image does not come out strongly. The solvent is too weak. Daguerre’s footnote.

The image is completely destroyed in the bath. The solvent is too strong. The same footnote, and the reason Niépce’s rule is to err weak.

The solvent has stopped working. It is saturated with asphaltum, which you can see: it has turned opaque and dark brown.

The varnish comes off the plate in layers as soon as it is immersed. Damp, absorbed by the varnish or by the room. This is the fault that stopped Niépce using the camera in winter.

The coating will not lay thin and even. The varnish is too fluid. It goes into a shallow dish to evaporate, without heat and protected from moisture, until it is thick enough to form a compact and as thin a coating as possible.

The picture on glass is damaged by the washing. The varnish adheres less well to glass than to plated silver. The Dippel-oil varnish was made to fix this.

The engraving marks the coating. Warm weather and direct contact. Interpose the cardboard border and glue the engraving to it.

Blackening with sulphide of potash either eats the varnish or does nothing. Concentrated it attacks the varnish; diluted it only reddens the metal. Use the iodine box instead — which is what Niépce did, and where the daguerreotype starts.

A camera view will not etch into a printing plate. Daguerre’s footnote is blunt about it: the image is made by the greater or lesser thickness of the varnish according as light attacked it, so the acid cannot act on the metal in the same proportion, and Niépce never made an engraving from a plate exposed in the camera. This is the failure that ended the project.

Experiments

The course does not prepare this coating and every experiment below is reading or arithmetic.

• Read the four texts side by side. The French of 1839, Memes’s English of the same year, Hunt’s of 1841 and Eder’s of 1945. Watch “un verre” become a wine-glass in Memes and stay one for seventy-two years; watch “n’en absorbe plus” become “completely saturated” in Hunt 1841 and go back to “can absorb no more” in Hunt 1844; watch the ten parts of white petroleum become six in Eder. This is what source criticism looks like on a document everybody has quoted and few have collated.
• Work out what the instruction actually determines. Take two vessels of different proportions, half fill each with any dry powder, and add water to stand three lines — about 6.8 mm — above the surface. Measure the ratio of free liquid to solid in each. You will have shown, without any bitumen, why this formula cannot be scaled, and you will have done the only quantitative work this page permits.
• Look at the plate. The Harry Ransom Center’s photographs of View from the Window at Le Gras, and its account of why the object has to be lit and angled to be seen at all. Then read Niépce on the second glass trial, where the tones come only from the relative thickness of the layer.
• Follow the line forward. Photogravure and photoglyphic engraving in this course’s atlas; then, if you want the full distance, the British patent abridgments for photography of 1877 to 1883, which are full of specifications that are recognisably this paragraph with the resin and the solvent swapped.

🧪The one this course would set instead

Coat a sheet with the classic cyanotype sensitiser, lay an object on it, put it in the sun, and wash it in water.

You will have done, at Level B and with every quantity on the page, the thing this formula does: laid a light-sensitive layer by hand, let a shadow fall on it, and removed what the light did not change. The differences are the interesting part. Your coating is water-borne and Niépce’s excludes water altogether; your wash is the developer and his is a solvent that would take the whole image off if it were a little too strong; and your exposure is minutes where his was days.