Double-coated albumen procedure

Float a sheet of paper on albumen, dry it, and float it on albumen again, and you will have a sheet with exactly as much albumen on it as you started with. The second float dissolves the first coat and carries it back into the tray. Everything on this page follows from that one fact and from the way around it, which is to make the first coat insoluble before the second one touches it — and the solution that does it is the only thing double coating adds to the albumen process that has a quantity attached.

Ingredient	Quantity	Form the source specifies
Isopropyl alcohol	70 %	printed by Reilly as "a 70 % solution of isopropyl alcohol", with no statement of whether the percentage is by volume or by weight; the course reads it as the ordinary 70 per cent v/v rubbing alcohol, which CAMEO records is the strength the material is sold at, and says so rather than converting it to a weight the source never gives
Ammonium chloride	2 %	matched to the chloride content of the albumen actually being used; 2 per cent is the illustration Reilly prints, and the albumen solution of this formulary carries 1.5 per cent, which is what a bath serving it would take. Where the albumen was salted with sodium chloride instead, Reilly's rule reads across to sodium chloride, though he prints no figure for that case
Water	to make 1000 mL	A basis, not a volume Reilly states. He gives two percentages and no batch size; a percentage is a statement about a made-up volume, so the litre is the volume the two figures are printed against here and the proportions are what travel. The salt is dissolved in the water first and the alcohol brought in afterwards, and the bath is then made up to volume rather than topped up by adding a measured 300 mL — mixing alcohol and water contracts, and 700 plus 300 does not make 1,000. Nothing is heated at any point: albumen coagulates above 65 °C, and a warm bath would cook the layer it is meant to set.

That table is the coagulating bath and nothing else. The two floats are the albumen solution, used twice, and it has its own entry. The procedure, in Reilly’s order, is this:

Step	What happens	Time
1	Float the sheet on the salted albumen, one side only	1 to 1.5 min
2	Hang to dry, blotting the runoff bead; mark the edge that hung lowest	until dry
3	Pile and flatten under weights	—
4	Pull the sheet slowly through the coagulating bath, wetting both sides	not published
5	Hang to dry	until dry
6	Pile and flatten under weights	—
7	Float on the salted albumen again, hanging the marked edge uppermost	1 to 1.5 min
8	Hang to dry, blotting the bead; flatten	until dry

Steps 4 and 7 are the whole of it. Step 4 is the chemistry; step 7 is the geometry.

Purpose

To put twice as much albumen on the sheet, and to do it evenly. Those are two purposes and the procedure has a separate step for each.

The thickness. Reilly lists what depends on the thickness and glossiness of the albumen coating: the ultimate colour of the prints, their brilliance and “depth”, and the ease of toning and fixing. The first three of those are what the nineteenth-century trade was buying and the fourth is what it was paying. Vogel had put the same axis in one sentence thirty years earlier — “thick albumen producing a glossy surface and brilliant prints, therefore the thickly albumenized paper is called brilliant albumen paper” — and the ladder he is describing runs from a matte sheet made with diluted albumen, through the undiluted albumen solution, to this. Double coating is the top rung, and it is the only rung that cannot be reached by changing the recipe.

The evenness. A sheet hung up to dry has more albumen at the bottom than at the top, because albumen runs downhill while it dries. Reilly says there will always be a slight difference in gloss and thickness between the top half and the bottom half of the sheet, and that it is usually not so pronounced as to spoil a batch. Coat twice with no other precaution and you double that gradient. Coat twice with the marked edge reversed and the two gradients run opposite ways and largely cancel. The instruction costs nothing and is the difference between a procedure and a mess.

The negative you can print. This is the purpose a modern worker actually cares about, and it is downstream of the other two. Reilly’s rule from Chapter Seven is that a glossy paper needs a negative of shorter density range than a matte one, because a transparent binder minimises diffuse reflection and the scattering of light by the paper fibres — the whites look whiter and the shadows look denser for the same amount of reduced silver. Rank the family by required density range and it runs plain salted greatest, matte arrowroot and matte albumen slightly less, glossy albumen lowest. Double-coated albumen is the glossiest of them, and Reilly draws the conclusion in one line: double-coated paper usually produces better prints from thin negatives than single-coated paper does.

To make a paper you can identify a hundred and fifty years later. That is not a purpose a nineteenth-century albumeniser had, but it is a real property of what they made. The Getty Conservation Institute’s atlas gives an objective spectroscopic criterion: in an ATR-FTIR spectrum of a single-coated print of 1855 to 1870, the Amide I peak of the protein near 1640 cm⁻¹ and the shoulder of the paper’s cellulose envelope near 1100 cm⁻¹ have similar intensities, while in a print made after about 1870 with more binder than one coating can deposit, the cellulose shoulder is much weaker than the Amide I peak, because the thicker protein layer attenuates the beam before it reaches the paper. The procedure on this page is what that measurement is measuring.

Recommended uses

Printing a thin negative on albumen. This is the first recommendation because it is Reilly’s own, and because it is the situation a modern worker meets constantly. Albumen paper grew up alongside the long-scale collodion negative, and Reilly records the trade shifting its chloride and silver strengths between 1880 and 1900 to accommodate the lower density ranges of the gelatin dry plate — so the mismatch this page addresses is not a modern discovery but the oldest problem the paper has. A modern film negative developed for a silver gelatin enlargement, or a digital negative made without a correction curve for printing-out papers, is far too short for it. Double coating moves the paper toward that negative rather than the other way round, and it moves it in a direction nothing else on the albumen bench does.

Maximum gloss and depth, where those are the point. If you are making an albumen print because you want the specific surface of an albumen print — a transparent skin holding the silver above the fibres, not among them — a single coat gives you a diluted version of it. The trade knew this: after about 1870 the public preference turned to glossy prints, the industry answered with double-coated paper, and a large portion of the paper sold after 1880 was double-coated.

Reproducing a post-1880 commercial paper rather than a hand-coated 1860s one. These are different materials and they print differently. If the object of the exercise is to understand what a carte de visite or cabinet card of the 1880s actually was, a single-coated sheet is the wrong control.

Where the rawstock is porous. Abney’s note is worth having: Rives and Saxe were the two papers principally used for albumenising, Saxe much more porous and consequently less glossy, and “Saxe paper can be rendered nearly as glossy as Rives by doubly albumenising and rolling”. A modern worker choosing between rawstocks has the same problem, and double coating is one of the two answers to it. The other is mechanical and is under Variants.

As a controlled experiment, which is the use this course would put first. Two sheets from one batch of albumen, one single-coated and one double, printed from one negative on one silver bath, is the cleanest single-variable demonstration available anywhere in the printing-out family of how a binder changes tone reproduction. It is set out under Experiments.

When another formula is preferable

On a first albumen print, always. Double coating adds a whole extra day to the paper making, doubles the albumen consumption, and makes every step after it harder: the sheet curls more, it is harder to manipulate in sensitisation and printing, it tones more slowly and it fixes more slowly. Reilly warns that two coatings may leave a paper so heavily coated that it is brittle and hard to tone. Make the albumen solution work single-coated first. You cannot diagnose a double-coating fault on a bench where the single coat has never been made to behave.

When the negative is dense and contrasty. Then you want the opposite of this page. Go down the gloss ladder, not up: the albumenised salting solution for a half-matte sheet, or Reilly’s arrowroot salting solution or his plain salting solution for a matte one. Each of those asks the negative for more density range, which is what a contrasty negative has to give.

When permanence is the governing concern. Everything that makes a double-coated print handsome makes it harder to process thoroughly. Reilly is explicit that coagulated albumen is resistant to penetration by solutions and that both fixation and washing are made more difficult by the nature of albumen itself; a thicker layer is a stronger version of the same problem. The AIC’s conservation account of albumen names the chemical bonding of image silver to sulfur-containing side groups of the protein among the proposed causes of the yellow highlight staining that is albumen’s characteristic failure, and a double-coated print has more protein to do it with. None of that is a reason never to double coat; it is a reason not to do it on the print you most want to survive, until your fixing and washing are demonstrably right.

When you have no ventilation and no way of making any. A tray of 70 per cent isopropyl alcohol at room temperature is above what this course classifies as a standard home darkroom, for reasons set out under Safety. The heat route under Variants reaches the same end with a dry-mount press or an iron and no solvent at all, and it is the honest alternative rather than a lesser one.

When what you actually want is gloss and not thickness. Two published methods get gloss without a second coat. Reilly’s is thermal: the drying rooms of the nineteenth-century albumen factories were held at 30 to 50 °C, and the higher the drying temperature the glossier the paper, which he offers explicitly as a way to improve the gloss and depth of single-coated albumen. Abney’s is mechanical: rolling, and after 1870 burnishing, which the Getty records as the other half of the reason prints of 1870 to 1890 are glossier than prints of 1850 to 1870. Neither adds albumen, so neither shortens the density range the negative has to have. Which of the three you want depends on whether the problem is the surface or the negative.

Mixing

The bath

Dissolve the chloride in part of the water, add the alcohol, then make up to volume. Reilly gives no mixing order — he gives two percentages and a tray — so this order is the chemistry’s, and the reason is a solubility that Wall’s dictionary happens to record for exactly this pair.

🧮Why the salt goes in first, and how little margin there is

Wall’s 1912 entry for ammonium chloride gives its solubility twice: 1 in 3 of cold water, and 1 in 55 of alcohol. Written as grams per 100 mL those are about 33 g and about 1.8 g. Reilly’s bath asks for 2 g per 100 mL, in a liquid that is 70 per cent alcohol.

2 % w/v = 2 g per 100 mL = 20 g per litre

What the bath asks for

1 part in 55 = 1.8 g per 100 mL

What alcohol alone will hold, from Wall

1 part in 3 = 33 g per 100 mL

What water alone will hold, from Wall

The bath is a mixture of the two, so its capacity lies between those bounds — and the figure asked for sits above what the alcohol half could carry by itself:

700 mL ÷ 55 = 12.7 g

What 700 mL of alcohol alone would hold, at 1 in 55

20 g ÷ 12.7 g ≈ 1.6 times too much

Against the 20 g the bath is asked to carry

20 g × 3 = 60 mL

What the same 20 g needs of water alone, at 1 in 3

The course has found no published solubility for ammonium chloride in 70 per cent isopropanol, and will not interpolate one: mixed solvents do not average, and the true figure could be above or below a proportional estimate. What the three lines above do establish is the practical rule. Tip the salt into the alcohol and you are asking for around half as much again as that liquid could hold on its own, and you will be watching it sit on the bottom. Dissolve it in 250 mL of water first — four times the water it strictly needs, so nowhere near saturation — and then bring the alcohol in, and the salt is already in solution and merely has to stay there.

The same relationship read from the other side is on the isopropyl alcohol page: HSDB records isopropanol as miscible with water in all proportions and insoluble in salt solution. At 2 per cent chloride you are nowhere near the concentration that would separate the two liquids, but it is the same physics, and it is the reason a bath asked to carry a heavy salt load is being asked for something a mixed solvent may not give.

Watch it as the alcohol goes in. If the bath clouds or throws crystals, it cannot carry the chloride at that strength, and the honest response is to record the fact and work at the chloride the bath will hold, saying so, rather than to filter out the difference and pretend the bath is at 2 per cent.

The sequence, then:

1. Weigh the chloride against the albumen, not against Reilly’s example. The rule is that the bath carries whatever the albumen carries. This formulary’s albumen solution is 15 g of ammonium chloride to the litre, which is 1.5 per cent, so the bath for it is 15 g per litre and not 20. Reilly’s 2 per cent is the figure he illustrates the rule with.
2. Dissolve it in about a quarter of the final volume of water, cold. Nothing here is heated at any stage, for the reason that governs the whole albumen bench: albumen coagulates above 65 °C.
3. Add the isopropyl alcohol — 700 mL for a litre of 70 per cent bath — and stir gently.
4. Make up to the final volume with water. Not by adding a measured 300 mL: alcohol and water contract when mixed, so 700 plus 300 is measurably less than 1,000, and a bath made that way is slightly stronger than 70 per cent rather than slightly weaker. Make it up to the mark.
5. Filter if anything is visible in it, and pour it into the tray only when you are ready to work.

✕A percentage of alcohol is not a percentage of anything else

“70 per cent isopropyl alcohol” almost always means 70 per cent by volume, and that is what is sold as rubbing alcohol — CAMEO records that strength as the commercial form. Reilly does not say which he means, and the course reads it as v/v and says so rather than converting silently.

The distinction is not academic. Isopropanol has a relative density of 0.79, so 70 mL of it weighs about 55 g against the 30 g of the water beside it, and 70 per cent by volume is about 65 per cent by weight — that conversion is the course’s arithmetic on the safety card’s density, not a figure any source on this page prints. If you are working from a bottle of 99 per cent isopropanol and diluting it yourself, dilute by volume and make it up to the mark, and you will land where Reilly is.

The procedure

Step 1 — the first float. Exactly as for a single-coated sheet: the tray filled to two thirds or three quarters of an inch, the albumen filtered through muslin immediately before use, 4 mL of a wetting agent per litre stirred in without raising bubbles, the sheet floated on one side only for 1 to 1.5 minutes, timed from the moment all bubbles are broken and the sheet lies flat. Albumen paper cannot be immersed. The albumen solution page carries the full account of that operation and its failure modes.

Step 2 — dry, and mark the sheet. Hang it by two corners along its long edge. Blot the bead of albumen that collects along the bottom edge several times as it dries: left to dry, it forms a thick rind that makes the sheet very difficult to handle in every operation that follows, and the last droplet takes a very long time to dry. Before the sheet leaves the rack, mark the edge that hung lowest. Everything about step 7 depends on knowing which edge that was, and it is unidentifiable an hour later.

Step 3 — flatten. Pile the dry sheets and put them under weights. This is not tidiness: it makes them supple, and a sheet that is not supple cannot be laid on a liquid surface without trapping air.

Step 4 — the coagulating bath. Put the bath in a tray and pull the sheets slowly through it. This is the one operation in the making of the paper in which the sheet is deliberately wetted on both sides. Every other bath up to and including the silver is a float — Reilly says albumen paper must be coated on one side only and cannot be immersed, and warns that silver solution reaching the back of a sheet gives patches of uneven density. Only after the print has been exposed does the paper go into anything as a whole. Hang the sheets to dry.

Step 5 and 6 — dry and flatten again. Same as before, and for the same reason. The alcohol is gone in minutes rather than hours, which is one of the few respects in which this step is easier than the ones around it.

Step 7 — the second float, reversed. Float on the albumen exactly as before, for the same 1 to 1.5 minutes, and hang the sheet with the marked edge at the top. The first coat is thickest at what was its bottom edge; hanging that edge uppermost puts the second coat’s thin end over the first coat’s thick end.

Step 8 — dry and flatten. Blot the bead again. Then the sheet is finished paper: not light-sensitive, keeping very well cool and dry, and ready to sensitise whenever you are.

!Everything up to sensitising is a white-light operation

Nothing on the sheet is light-sensitive until it meets the silver nitrate. Reilly says so and the modern kit’s instructions say so in as many words. Double coating therefore has no safelight requirement and no deadline, which is precisely why it should be done as a batch, days or weeks before the printing session. The step with the deadline is the sensitising: a sensitised sheet remains in good condition for 24 to 48 hours and no longer.

Behaviour

In the tray, it behaves like a solvent and not like a photographic bath. It is thin, it wets paper instantly, it smells strongly, and it evaporates. Those four facts govern everything about handling it.

It weakens while you use it, from both ends. Alcohol leaves an open tray by evaporation, and water leaves the sheets and enters the tray. Both move the strength downward. That direction matters because Reilly’s lower bound is the dangerous one: too dilute a solution “is not strong enough to coagulate the albumen before it partially dissolves into the water”. A bath that has stood open through a long session is therefore not merely less effective — it can be actively dissolving the coat it was supposed to set. Keep the tray covered between sheets, mix only what the session needs, and treat a long pause as a reason to remix rather than to carry on.

🧮The tray is the consumption, not the sheet

A sheet takes almost nothing out of an alcohol bath. What a session consumes is the volume needed to fill a tray deep enough to pull a sheet through, and that is a geometry problem.

Reilly’s advice is to start with a sheet about 10 by 12 inches, which wants a tray of roughly 28 by 35 cm. That tray holds about 98 mL for every millimetre of depth:

0.28 m × 0.35 m = 0.098 m², so 1 mm of depth = 98 mL

The tray's footprint

0.098 m² × 0.010 m = 0.00098 m³ ≈ 980 mL

Filled to 10 mm, the least a sheet can be drawn through

0.098 m² × 0.015 m ≈ 1,470 mL

Filled to 15 mm, which a stiff double-thickness sheet will want

0.098 m² × 0.017 m ≈ 1,670 mL

Reilly's albumen tray, at his two thirds of an inch

So a session needs something between one and one and a half litres of alcohol standing in a tray, and about one and three quarter litres of albumen standing in another, whether it coats one sheet or twenty. That, and not the albumen consumed per print, is the real cost of the procedure, and it is the argument for coating in batches and for using the smallest tray the sheet will fit.

All four figures are the course’s arithmetic on a tray size the reader chooses, not a number from any source. Reilly gives a depth for the albumen tray, two thirds to three quarters of an inch, and none at all for this one.

The sheet comes out of the bath changed and you can feel it. A coagulated albumen layer is harder, less glossy in the wet state, and noticeably less inclined to be wetted by the second float than a raw one — which is the whole point, and which is also why bubbles are more troublesome on the second float than on the first. Reilly’s remedy for bubbles is the same at both floats: sweep the tray surface, add the wetting agent, and start the clock only when the sheet lies flat.

It is a batch operation with no deadline, and it should be treated as one. The finished sheets keep very well cool and dry. A worker who albumenises and coagulates and albumenises again on the morning of a printing session has put three drying periods and a solvent tray in front of the one step that has a two-day clock on it.

The corroboration is broad and the numbers are not. Four independent sources say double-coated albumen paper existed and say why: Wall in 1912 gives the method in a sentence, Abney in 1905 records that it makes a porous stock nearly as glossy as a smooth one, the Getty dates it to 1870 and can detect it spectroscopically, and the modern kit sells a procedure for it. Exactly one source — Reilly — gives a bath with numbers in it. That asymmetry is normal for a trade practice and it is why the provenance note refuses to record Wall as corroboration: he corroborates the procedure and prints a different bath.

Image characteristics

Gloss and “depth”, which are the two things it was for. The Getty’s account of the taste change is worth having straight: in the 1850s the glossiness of albumen prints was criticised in the photographic literature, because the public was used to the matte surface of salt prints; after about 1870 the preference turned the other way, and the industry answered with double-coated paper “known to yield high-gloss albumen photographs”. A double-coated print does not merely reflect more light — it scatters less of the light that gets into it, because more of the path is transparent protein and less of it is paper fibre.

A shorter negative density range, which is the practical consequence. Reilly’s ranking and his one-line conclusion are both above under Purpose. Put concretely: a negative that prints flat and weak on a single-coated albumen sheet may print properly on a double-coated one, and the same negative would be hopeless on plain salted paper.

Less overprinting. Every printing-out paper must leave the frame looking too dark, because it bleaches in toning and fixing. Reilly ranks the amount needed: salted paper most, albumen slightly less, double-coated albumen less still. That is a direct saving of exposure and of judgement, and it is the one respect in which double coating makes the printing easier rather than harder.

Slower toning, and toning that shows every fault in the coating. Two separate statements of Reilly’s, both from the toning section. First, albumen prints need a much stronger and more effective toning solution than salted papers because albumen is less permeable and the silver particles are more protected, and toning is slower even so. Second — and this is the one that catches people — “any variations in the thickness of the albumen coating, especially in the case of double-coated papers, will be immediately apparent in the toning step, because less heavily coated areas will tone more quickly and deeply”. The toner is a thickness map of your coating. If step 7 was done without reversing the marked edge, the print will tell you in the gold bath.

Brittleness and curl. Reilly’s warning is unhedged: two coatings produce papers that are quite glossy “and may even be so heavily coated that they are brittle and hard to tone”, with a greater tendency to curl and a greater difficulty in sensitisation and printing. This is the cost side of the ledger and it is not small.

Crazing, and the fact that the trade dealt with it mechanically. The Getty finds a fine network of surface microcracks in most albumen prints made before 1870, and either a reduced pattern or none at all in prints made after 1870 that were burnished and heat treated. Reilly’s identification appendix describes the same characteristic crackled or crazed texture and names the same two causes for its absence in later prints: burnishing and rolling machines, and the increased use after 1870 of double-coated paper. The two changes arrived together and the evidence cannot fully separate them, which is worth saying plainly rather than attributing the smoother surface to either alone.

Colour, which depends on thickness. Reilly lists the ultimate colour of the prints among the properties that depend on coating thickness, and gives no numbers for it. What the course can say without inventing is the direction and the reason: more binder means more of the image silver is held in transparent protein above the fibres rather than scattered among them, and a print whose silver is concentrated in a thin transparent layer reads deeper and more saturated than the same quantity of silver spread through paper. The gold toner then acts on that silver more slowly, so the colour a given toning time reaches is not the colour the same time reaches on a single-coated sheet.

The mechanism

Why the second coat dissolves the first

Dried albumen is a water-soluble protein film. The albumen solution is mostly water. Lay one on the other for ninety seconds and the water gets underneath the first layer and takes it back into solution — Reilly states the result flatly: without a hardening step “there was no gain in thickness or amount of albumen on the sheet, because the second coating step dissolved off the albumen remaining from the first coating operation”.

This is worth pausing on, because it is the reason the whole procedure exists and it is also the reason it is not obvious. Nothing goes visibly wrong. The sheet floats, the second coat dries, the paper looks like albumen paper. It simply is not any thicker, and you find out at the toning bath or at the densitometer weeks later.

What alcohol does to a protein

Reilly gives the fact in Chapter Two: “Albumen is insoluble in alcohol, and in fact alcohol will coagulate albumen, a property that is useful to obtain multiple coatings of albumen on a single sheet.” He lists three agents that coagulate it — alcohol, temperatures above 65 °C, and contact with salts of metals — and the double-coating procedure uses the first, the alternative routes use the second, and the sensitising bath afterwards uses the third.

⚗How far the course will take the protein chemistry, and where it stops

A folded protein is held in shape by a large number of individually weak interactions, and it stays in solution because its outside is hydrated — surrounded by ordered water. Anything that disrupts that balance can unfold it, and an unfolded protein exposes surfaces that stick to other unfolded proteins. The result is an aggregated network that no longer dissolves. That is what “coagulation” and “denaturation” name, and it is why all three of Reilly’s agents produce the same outcome by different routes: heat supplies enough energy to break the interactions directly, metal ions bridge and neutralise charged groups, and alcohol changes the solvent the protein is sitting in.

What the course will not do is write a reaction equation for it. Albumen is not one substance; Reilly says so in his first sentence about it. Denaturation is not a stoichiometric reaction with a balanced equation, there is no defined product, and any equation printed here would be a decoration rather than a claim. The general account above is standard protein chemistry and is stated as such; everything specific to albumen on this page is Reilly’s observation, and the boundary between the two is drawn here on purpose.

Why exactly 70 per cent

Reilly bounds it from both sides, and the two bounds are the most useful sentences on the subject anyone has published:

• Too strong. “Pure alcohol is too strong and unevenly coagulates the albumen layer.”
• Too weak. “Too dilute alcohol solutions are not strong enough to coagulate the albumen before it partially dissolves into the water.”

Read together, they describe a race. The bath has to get alcohol into the whole thickness of the dried layer before the water in the same bath has dissolved the layer’s surface. Neat alcohol wins the coagulation race so decisively that it sets the outside of the film before the inside is reached, and an unevenly set film is exactly what Reilly reports. A weak bath loses the race the other way: the water does its work first. Somewhere in between, the layer swells enough for the alcohol to penetrate it and is set through before it is lost. Reilly puts that point at 70 per cent and says experience found it.

That reading of the two bounds as a race is the course’s, not Reilly’s. He states the two failure modes and the working strength; the account of why a middling strength should be better than either extreme is an interpretation, offered because it makes the number remembered and testable rather than arbitrary. It also predicts something checkable, which is what makes it worth stating: the optimum should depend on how thick and how dry the first coat is, and a thicker or drier coat should want a slightly weaker bath and a longer dip. That prediction is the first experiment at the foot of this page.

Why the bath carries salt

Reilly’s reason is one clause long: “To prevent leaching out of the chlorides from the albumen, whatever chloride content is present in the albumen itself should also be added to the alcohol solution.”

The first coat is a chloride-loaded protein film. Put it into a chloride-free liquid and the chloride does what any dissolved solute does at a concentration difference: it moves out, down the gradient, into the bath. Match the bath’s chloride to the coat’s and there is no gradient and nothing to drive the movement. The salt in the bath is not there to do anything to the paper; it is there to not take something away from it.

The photographic consequence of getting this wrong is precise, because Reilly quantifies the effect of chloride content elsewhere. A paper of 1 to 1.5 per cent chloride is less sensitive than one at the normal 1.5 to 2.5 per cent and tends to give slightly more contrasty prints from thin negatives, but it does so at the expense of a rich, dense image; his recommendation is to keep the chloride at 1.5 per cent or above. A first coat that has been washed out in a chloride-free alcohol bath is a chloride-poor layer sitting under a normally salted one. What that does to the print is not a uniform loss of density but a stratified one: the halide, and therefore the image, is concentrated in the upper part of the coating.

NH₄Cl + AgNO₃ → AgCl + NH₄NO₃

What the chloride is for, two baths later — the silver chloride is assembled inside the binder

The full account of that double replacement, and of why the free silver nitrate left in the sheet matters as much as the chloride, is on the albumen solution page. What this page adds is that the reaction can only happen where the chloride still is.

From coating thickness to the print

The chain runs like this, and every link in it is sourced:

1. A second coat is retained rather than dissolved, because the first was coagulated. Reilly.
2. The albumen layer is thicker — measurably so, and the Getty’s ATR-FTIR criterion is the measurement: the cellulose signal of the paper is attenuated because the beam has more protein to get through.
3. A thicker transparent binder scatters less light. Reilly’s Chapter Seven: a transparent binder minimises diffuse reflections and the scattering of light by the paper fibres.
4. So the whites look whiter and the shadows look denser for the same amount of reduced silver. The same silver that gives a deep shadow on glossy paper gives a much paler-looking one on a matte surface.
5. So less contrast has to come from the negative. Hence the ranking, hence “better prints from thin negatives”, hence less overprinting.
6. And the same thickness makes the layer less permeable, so toning and fixing are slower and harder — Reilly states the permeability point twice, once in the double-coating section and once in the fixing section, where he attributes the difficulty of fixing and washing albumen prints to the resistance of coagulated albumen to penetration by solutions.

Steps 3 to 5 and step 6 are the same physical fact seen from two sides. You cannot buy the first without the second, and no procedure on this page pretends otherwise.

The other coagulants, and what they tell you

Heat. Albumen coagulates above 65 °C. Steam is at 100 °C; a dry-mount press at the supplier’s 250 °F is at about 121 °C. Both are comfortably over, which is why both work, and the modern kit’s choice of heat over solvent is a hazard trade rather than a chemical compromise.

Six months in a warm loft. Reilly calls this the simplest and most widely used method of the nineteenth century and cites Stiefel’s 1894 manual for it. Neither he nor the course can tell you what the “slow curing process” is, and the course will not guess: a protein film left warm for half a year has several plausible routes to insolubility and no evidence here distinguishes them. What the fact does establish is the economics of the trade. A factory holding six months of stock in a loft was not being patient; it was avoiding a solvent step on an industrial scale, and the alcohol and steam routes are described by Reilly as the “instantaneous” alternatives for someone who cannot wait.

Function of every ingredient

Isopropyl alcohol, 70 per cent — the coagulant, and the only active ingredient. Propan-2-ol, C₃H₈O, a small water-miscible alcohol. It is here because albumen is insoluble in it and is coagulated by it, and because that coagulation does not reverse when the alcohol dries off. Reilly’s own words for what the step achieves are “to render the first coating insoluble”, and he names the property as the reason the trick works at all — alcohol’s coagulation of albumen is “a property that is useful to obtain multiple coatings of albumen on a single sheet”.

What more of it does. Raise the strength toward neat alcohol and the coagulation happens faster than the alcohol can penetrate, so the surface sets before the interior does. Reilly’s word for the result is “unevenly”, and unevenness in the first coat is not a cosmetic fault: it becomes a thickness variation in the finished paper, and Reilly says thickness variations show up immediately in the toning bath as areas that tone more quickly and deeply. A print blotched in the gold bath is the symptom of a bath mixed too strong.

What less of it does. Drop the strength and the water in the bath gets to work first. The layer partially dissolves before it is set, so the sheet loses albumen at the very step meant to preserve it, and the second float then dissolves what is left. The failure looks exactly like having skipped the hardening step, which is why the diagnosis under Troubleshooting has to work backwards from the bath strength rather than from the appearance.

What it interacts with. Two things, and neither is optional reading. It is the reason the chloride is hard to keep in solution, because ammonium chloride is far more soluble in water than in alcohol — see the arithmetic under Mixing. And it is incompatible with silver nitrate in a way that is not merely inconvenient: the course’s incompatibilities page records, from CAMEO, that silver nitrate with alcohols may form silver fulminate. That is a storage and mixing rule and it is set out under Incompatibilities below.

Ethanol instead. Wall’s 1912 bath is ethyl alcohol, so the substitution is published and not a guess. Ethanol coagulates protein by the same route; what differs is the strength at which it does it and the fact that Wall’s is applied by floating rather than immersion. The two baths are compared under Variants. Denatured or methylated spirit is a poorer choice than either for a reason Wall himself gives: methylated spirit sold retail after 1891 carried a fraction of a per cent of petroleum which separates on the addition of water and renders the liquid milky, “partially unfitting it for several uses to which it was formerly applied in connection with photography”.

Ammonium chloride, at the albumen’s own percentage — the ingredient whose job is to do nothing. NH₄Cl, the salt of a strong acid and a weak base, and the chloride the nineteenth-century trade used almost universally for salting albumen. In the albumen solution it is the reagent that becomes silver chloride. Here it is none of that. It is present at exactly the concentration at which it has no tendency to move, so that the chloride already in the first coat stays where it is.

What more of it does. Overshoot the albumen’s chloride and the gradient reverses: the bath is now richer in chloride than the coat, and chloride will tend to move into the layer. The consequence would be a first coat with more halide than the albumen was formulated to carry, which Reilly’s own account says buys nothing — more chloride than is necessary “only results in higher silver consumption without conferring any additional benefit”. That direction is the course’s inference from Reilly’s stated rule and his stated chloride figures, not something he says about the bath; what he says is that the bath should match. There is also a hard practical ceiling, since the salt is close to the limit of what the bath will hold at all.

What less of it does — including none at all. Then the coat is leached. Reilly’s figures give the consequence: below about 1.5 per cent the paper is less sensitive and gives a thinner, less rich image, gaining a little contrast from a thin negative in exchange. A leached first coat under an unleached second one gives a paper whose halide is stratified, and the practical signature is a print that looks thin and lacks depth despite an obviously thick, glossy coating — the one failure that a look at the surface cannot diagnose. Wall’s bath, it should be said, carries no chloride at all and Wall reports no trouble; whether that is because his sheets were coagulated by floating rather than immersion, and so wetted on one side only and for a shorter time, is a reasonable hypothesis and is untested. It is listed as an experiment.

Sodium chloride instead. Reilly allows either chloride in the albumen and says prints of similar colour and contrast may be expected from either, or from both in combination. The matching rule therefore reads across: a sheet salted with sodium chloride wants sodium chloride in its coagulating bath. Reilly prints no figure for that case, and the course does not supply one beyond the rule itself, which is that the bath carries what the albumen carries.

Water — not the diluent, but the reason the bath works at all. It has two jobs and both are essential. It is the solvent that gets the chloride into a liquid that is mostly alcohol, which is why it goes into the vessel first. And the 30 per cent of the bath that is water is what stops the alcohol setting the surface of the layer before it has reached the inside: a bath of pure alcohol is the one Reilly says coagulates unevenly. A reader who thinks of the water here as a way of using less alcohol has the formula upside down. The water is part of the mechanism; the strength is a compromise between two failure modes and not an economy.

Nothing else, and the absences are worth naming. There is no acid, although the albumen it serves is deliberately acidified — the acid’s work is done in the albumen and there is no reason to repeat it here. There is no hardening agent of the gelatin kind, no alum and no aldehyde: albumen does not need one, because alcohol alone renders it insoluble, and adding one would introduce a hazard for no gain. And there is no wetting agent, although both albumen floats take one. Reilly specifies the surfactant for the albumen tray and not for this one, and the reason is visible in the operation: alcohol wets paper on contact and needs no help doing it.

Interactions

With the albumen solution — a dependency between two formulary entries, which is unique here. No other bath in this formulary has to be reformulated because a different formula was changed. This one does: change the chloride in the albumen solution and the coagulating bath must change with it, in the same direction and to the same figure. Two batches of albumen at different chloride strengths need two coagulating baths, and a single bath used for both will leach one and enrich the other. Label the bottle with the chloride percentage, not just its name.

With the silver sensitiser — less than you would expect. Reilly does not raise the sensitising float for a double-coated sheet: 2.5 to 3 minutes on a 10 to 12 per cent bath “insures adequate sensitization of even doubly albumenized paper”. That is a specific claim about this specific combination and it should not be generalised to other strengths, and the modern kit’s alternative — a 15 per cent bath and a 6-minute float — shows how far a different procedure can sit from it. What does change is handling: a double-coated sheet is stiffer and curlier, it must not be bone dry when it meets the silver, and the ban on letting silver solution reach the back of the sheet is harder to obey on a sheet that will not lie flat.

With the toner, decisively. Albumen prints need a much stronger gold bath than salted papers and tone more slowly even so, because albumen is less permeable and the silver particles are more protected; a double-coated sheet is more of both. Use the gold baths at full strength — the gold-borax and gold-thiocyanate entries carry the working strengths — expect longer, and judge by inspection. And expect the toner to report on your coating: unevenly coated areas tone faster and deeper.

With the fixer, and this is where permanence is decided. Reilly’s albumen fixing regime is an alkaline 15 per cent sodium thiosulfate bath, two trays, four minutes in each, both made up fresh on the day — see the alkaline fixing bath. The reason for two baths is chemical rather than economical, and the reason the times matter more here than anywhere else is that coagulated albumen resists penetration by solutions. Double the albumen and you have doubled the barrier between the fixer and the silver it has to reach. Nothing in Reilly licenses extending his times for a double-coated sheet, and the course will not invent an extension; what it will say is that this is the step where a double-coated print is most likely to be under-processed, and that the residual-thiosulfate test exists for exactly this reason.

With ammonia fuming — a lever you should not reach for. Fuming raises the sensitivity of a salted sheet, and Reilly’s assessment is that it is unnecessary in modern practice: its value was greatest where a paper of low chloride content was sensitised on a weak bath, and much of that value can be had instead by raising the silver bath. Double coating is a procedure whose whole risk is a chloride-poor first coat, so a worker tempted to fume because the prints are thin is treating the symptom. Check the coagulating bath’s chloride first.

With the rawstock. Abney’s Rives and Saxe comparison is the period statement of a modern problem: a porous stock takes albumen into its body rather than onto its surface, and double coating is one way of getting a surface layer onto it. Wall’s warning bounds the same effect from the other side — do not float longer than eighty seconds “or the albumen will sink into the body of paper”. The interaction is that a porous sheet benefits more from a second coat than a smooth one does, and also curls more when it gets one.

With drying temperature. Reilly’s factories dried at 30 to 50 °C and the higher temperature bought gloss. That is an interaction because it is a partial substitute: a warmer drying room raises the gloss of a single coat, which is one of the things a second coat is for. It does not raise the amount of albumen on the sheet, so it does not shorten the density range the negative must have. If you are double coating for gloss alone, try the drying room first.

Variants

Wall’s bath, 1912: two volumes of alcohol to one of water, applied by floating. The other published method, in one sentence: “Double albumenised paper is made by coagulating the first layer of albumen by floating on a mixture of two volumes alcohol and one volume of water. The paper is now dried and again floated on the salted albumen.” Three differences from Reilly’s, and they are not equally important.

🧮What Wall's two-to-one actually is, in modern terms

Wall’s “alcohol” is defined in his own dictionary: used without qualification it means common or ethylic alcohol, that is, rectified spirit. He gives two standards for it — the old one containing 16 per cent of water at specific gravity 0.838, the new British Pharmacopoeia one containing 10 per cent of water at 0.834.

2 × 84 % ÷ 3 = 56 % alcohol

Old standard, two volumes in three

2 × 90 % ÷ 3 = 60 % alcohol

New B.P. standard, two volumes in three

So Wall’s bath is somewhere around 56 to 60 per cent ethanol against Reilly’s 70 per cent isopropanol — the same region, comfortably clear of both of Reilly’s failure boundaries. Two caveats, both the course’s. Wall does not say whether his percentages of water are by weight or by volume, so the figures above are approximate rather than exact. And ethanol and isopropanol are not interchangeable at equal percentages as protein denaturants, so the arithmetic shows the two baths are in the same neighbourhood and does not show that they are equivalent.

The first difference is the alcohol, and it is the least important: both are short-chain alcohols and both coagulate albumen. The second is the strength, and 56 to 60 against 70 is a real gap but sits well inside Reilly’s stated bounds. The third is the significant one: Wall floats and Reilly immerses. A floated sheet is coagulated from one side only, and only for as long as the alcohol takes to work through from the coated face. That is a genuinely different operation, and it is the most plausible explanation for the fourth difference — Wall’s bath carries no chloride, and he reports no trouble from leaching. A short one-sided contact has far less opportunity to wash the halide out than a pull-through does. The course records that as a hypothesis, because neither author addresses the other.

The heat routes. Two are published, at two temperatures, and both rest on the same fact that albumen coagulates above 65 °C.

• Steam, from Reilly: “subject the albumen to a current of steam, which in effect cooks the albumen and renders it insoluble”. No time, no apparatus, no further detail. It is the historical “instantaneous” alternative and it is recorded here as published rather than recommended: a sheet of paper held in a steam jet is also a sheet of paper being wetted, and nobody publishes how that is managed.
• A dry-mount press or an iron, from the modern kit: the press at 250 °F (about 121 °C) for 2 minutes with the sheet between two clean archival papers, or a clothes iron at its highest setting moved slowly back and forth for 2 to 3 minutes. That procedure hardens both coats, the second as well as the first, and its floats are 6 minutes rather than Reilly’s 1 to 1.5. It is the route to prefer where ventilation is the limiting factor, and it is a complete alternative rather than a substitution into Reilly’s method: the two sets of times belong to two different procedures and should not be mixed a line at a time.

Six months in a warm loft. The trade’s own answer, which Reilly calls the simplest and most widely used method of the nineteenth century, attributing it to Stiefel’s Sensitized Papers of 1894. It is not available to a modern worker in any practical sense, and it is here because it explains the industry: a factory with a loft did not need a solvent, and the two “instantaneous” methods exist for people who cannot wait half a year.

The non-variants: two ways of getting gloss that are not this procedure at all. Both are worth knowing because both are cheaper.

• Hot drying. Reilly: drying rooms at 30 to 50 °C, and the higher the temperature the glossier the paper, offered explicitly as a way to improve the gloss and depth of single-coated albumen.
• Rolling and burnishing. Abney: “Saxe paper can be rendered nearly as glossy as Rives by doubly albumenising and rolling” — note that he pairs the two. Wall rolls his albumenised paper between smooth rollers when thoroughly dry as a matter of course. The Getty names burnishing alongside double coating as the reason later prints are glossier, and adds a consequence the others do not: burnished and heat-treated prints show a reduced network of surface microcracks, or none.

Neither adds albumen. Both change the surface. If what you want is the negative behaviour — a shorter density range — only the second coat will give it to you.

And the rung below, which is a different formula rather than a variant of this one. Diluting the albumen instead of coating twice moves down the same axis in the opposite direction, and it has its own entry at the albumenised salting solution. Vogel gives the whole ladder in one sentence and names its top rung: thick albumen gives a glossy surface and brilliant prints, “therefore the thickly albumenized paper is called brilliant albumen paper”.

Safety

Level B, and the reason is entirely the alcohol tray.

The course’s classification rubric admits “volatile organic solvents beyond isopropanol in small quantities” to Level A and no further. A litre or more of 70 per cent isopropanol standing open in a tray, with the sheets then hung up to dry in the same room, is not a small quantity, and the criterion it meets at Level B is the plain one: a procedure that generates vapour.

⚠The flash point is below room temperature

The numbers, from the international chemical safety card for isopropyl alcohol: flash point 11.7 °C closed cup, explosive limits 2 to 12 per cent by volume in air, vapour pressure 4.4 kPa at 20 °C, auto-ignition 456 °C. CAMEO adds the figure that decides where the vapour goes: a vapour density of 2.07, heavier than air, so it sinks and pools. The aggregated GHS classification is Danger, with H225 highly flammable liquid and vapour, H319 causes serious eye irritation, and H336 may cause drowsiness or dizziness.

Read those figures carefully, because they are for the pure liquid and the tray holds a 70 per cent solution. Water raises a flash point, so the bath’s is somewhere above 11.7 °C; the course has found no published flash point for 70 per cent aqueous isopropanol and will not estimate one, because an estimated flash point is exactly the kind of number Rule 1 exists to keep out of a safety section. What can be said without estimating anything is that a flash point of 11.7 °C for the substance is below any temperature at which a darkroom is habitable, that dilution moves it up by an unpublished amount, and that the only safe working assumption is that the tray is giving off ignitable vapour. Treat it as one. The controls follow directly:

• No ignition source anywhere in the room. No naked flame, no hot plate, no dry-mount press running, no soldering iron, no smoking. The safety card’s own line is that the vapour mixes well with air and explosive mixtures are easily formed.
• Ventilation at bench level and below. The vapour is twice as dense as air; it collects low and will not clear itself from a sealed cupboard or a windowless bathroom.
• Cover the tray between sheets and mix only what the session needs. This is a safety control and a quality control at once, since an open tray also loses strength.
• The drying sheets are part of the hazard. Alcohol-wet paper hanging in a room is evaporating solvent into that room for several minutes. Dry them where the vapour goes somewhere.
• Goggles for pouring, nitrile gloves, and the safety card’s warning about the central nervous system in mind: drowsiness and dizziness are the early signs of too much vapour, and the correct response to noticing them is to leave the room, not to open a window and continue.

☠An alcohol-damp sheet must never go into a dry-mount press or under an iron

Both hardening routes on this page are published and both are sound. Combining them in that order is not. A press at 121 °C or an iron at its highest setting is an ignition-capable heat source, and a sheet still damp with 70 per cent isopropanol is carrying a flammable liquid onto it — whatever the exact flash point of that solution turns out to be, it is far below the temperature of the press. Nothing in either source suggests doing this and neither warns against it, because neither author is describing the other author’s method.

This instruction is the course’s own, drawn from the two published hazard classifications rather than from either procedure. Choose one route. If a sheet has been through the alcohol bath, it is dried in air and pressed only when it is dry and odourless — and if you are pressing at all, prefer the heat route from the start and keep the solvent out of the room entirely.

The other reagent. Ammonium chloride carries the signal word Warning with H302, harmful if swallowed, and H319, causes serious eye irritation. It is a solid, weighed once per batch, handled with ordinary care and a technique that does not raise dust. It contributes nothing to the level of this page.

What is not a hazard here, and why. There is no silver in this bath and none in either albumen tray, so nothing on this page stains, nothing has to be collected as silver-bearing waste, and the whole silver nitrate regime belongs to the sensitising bench and not here. Nothing on the sheet is light-sensitive at any point in this procedure, so there is no safelight requirement and no dim-light handling; every step is done in white light. There is no acid, no alkali, no oxidiser and no reducing agent in the bath, so there is no route from anything in the tray to a toxic gas — the only vapour in the room is the alcohol, and it is a flammability and narcosis problem rather than a toxicity one at these concentrations. There is no heating in the alcohol route at all, so there is no hot-work hazard on that path; the heat route trades the solvent hazard for a burn hazard and a scorched-paper hazard, which is a real trade and is why both routes are given.

What is downstream and is not Level B. Sensitising, printing under ultraviolet, gold toning and fixing all have their own hazards and their own pages. The fact that a coated sheet is inert should not be read as a statement about the process it belongs to.

Storage

The bath. A stoppered bottle, cool, out of sunlight, and stored as a flammable liquid — the safety card’s own wording is “fireproof, separated from strong oxidants, cool, well closed” — and, absolutely, away from silver nitrate. Nothing in the mixed bath decomposes: it is an alcohol, a chloride and water, with no oxidation to guard against and no biological load. What degrades it is evaporation, and evaporation only happens where you leave it open. Label it with the alcohol strength and the chloride percentage, because a bath matched to a 1.5 per cent albumen and a bath matched to a 2 per cent albumen are different baths and are indistinguishable by eye.

The alcohol bottle it is made from is a different question. CAMEO’s reactivity profile for isopropanol carries the hazard that catches people out: it “reacts with air or oxygen to form dangerously unstable peroxides”, so a half-empty bottle that has stood for years is not the same article as a fresh one. Buy the alcohol in a size the bench will use, and do not keep an ageing part-full bottle for a procedure that needs a litre or more at a time.

The single-coated sheets, waiting for the second float. Dry, piled, under weights. The flattening is functional: a sheet that is not flat cannot be pulled evenly through a shallow tray, and the marked edge has to survive, so mark it in pencil on the back where handling will not remove it.

The finished double-coated sheets. Reilly’s rule for albumenised paper generally, and it applies with more force here because the layer is thicker: cool, dry, and if the sheets must be rolled, rolled albumen side out, so that the layer is stretched rather than compressed and cracks less. Flat under weights is better than rolled. Do not let them get bone dry before sensitising — Reilly’s remedy is a night in a damp place or a few hours in a closed box with a dish of water, and a sheet that is too dry will not properly absorb the silver solution.

The dry chemical. Ammonium chloride takes up water from the air and cakes; keep it closed.

Incompatibilities

Silver nitrate and alcohol must never meet, and on this page they are in the same workflow. This is the incompatibility that matters most here and it is easy to overlook, because the alcohol tray and the silver tray are days apart in the procedure and inches apart on the shelf. The course’s incompatibilities reference records the pair from CAMEO: silver nitrate with alcohols may produce silver fulminate, which can explode when disturbed. The rule the course draws from it is stated there and is repeated here because this page is one of the few places where both substances are in play: do not mix concentrated silver nitrate solution with alcohol, and do not store the two together.

In practice, on this bench: separate shelves, separate trays that are never swapped, separate funnels and measuring cylinders, and the alcohol bottle nowhere near the silver bottle. There is no reason whatever for the two to be in the same cupboard, and one very good reason for them not to be.

Silver nitrate and albumen, by design. The two halves of the process must meet only inside the paper. A drop of silver solution in either albumen tray coagulates the protein where it lands and precipitates silver chloride at the same moment; the result is unrecoverable. This is stated on the albumen solution page and applies twice as often here, because there are two floats.

Alcohol and oxidisers generally. Nitric acid, permanganate, dichromate, persulfate: a flammable liquid and an oxidiser are a fire, and several of those appear elsewhere in this course. Nothing in this procedure calls for any of them; the incompatibility is a storage rule.

Alcohol and heat, which is the same rule as the flash point. Covered under Safety, and repeated in the Never callout there because it is the one combination this page can foresee a reader making.

Heat and albumen, which is not a safety matter but will ruin the work. Albumen coagulates above 65 °C. Do not warm the albumen solution to make it flow better, do not warm the coagulating bath to speed it up, and do not dry a coated sheet against a radiator hot enough to cook it. The 30 to 50 °C of the historical drying rooms is deliberately well below the threshold.

Waste

The spent coagulating bath is a flammable-liquid waste and not a drain stream. Kodak’s environmental guidance for amateur photographers lists what may be sewered — developers, stop baths, fixers after silver recovery, wash water — and then lists what may not, naming solvents and flammable materials explicitly. A litre of 70 per cent isopropanol carrying a little dissolved albumen and a little ammonium chloride is a solvent waste, and the alcohol is the reason.

What it does and does not contain. It contains alcohol, ammonium chloride, water, and whatever albumen came off the sheets. It contains no silver, because it is used before the sheet has ever met a silver bath. That is worth stating plainly: this is one of the few waste streams on the alternative-process bench that has nothing to do with silver recovery, and treating it as though it did would send it down the wrong route entirely.

The two routes, and both have limits. Traces on cloths and swabs evaporate, and evaporation of a small residue in a shallow dish left outdoors away from buildings is a route the wet-plate literature uses deliberately; it is not a route for a litre. Bulk liquid is bottled, labelled with its contents, and taken to whatever route your jurisdiction provides — a household waste and recycling centre’s chemical cupboard for a domestic user in the United Kingdom, a licensed waste contractor for a trade user.

And the caveat that governs all of it. The course cannot tell you what your jurisdiction permits, and flammable liquids are among the most tightly regulated wastes there are. The disposal policy sets out what the course can and cannot say. Local regulation governs.

The albumen side. Surplus albumen goes back into its bottle through a filter and is not waste until the batch is finished with; when it is, it is a food-grade protein waste and the albumen solution page covers it.

Troubleshooting

The sheet is no thicker after two coats than after one — no more gloss, no more depth, no change in the negative it wants. The hardening failed, and Reilly names the mechanism: the second float dissolved off what the first one left. Four causes, in the order to check them. The bath was too weak, either as mixed or after standing open through a session. The sheet was not dry when it went into the bath. The dip was too brief — Reilly says “briefly” and gives no number, so the shortest useful dip is something you have to find. Or the bath was skipped, which happens more often than it sounds when the step is invisible in the result.

Patchy, blotchy or streaky gloss on the finished sheet, which then tones unevenly. Two suspects. If the bath was mixed strong or with neat alcohol, this is Reilly’s stated failure of an over-strong bath: uneven coagulation. If the bath was right, it is a pull-through problem — a sheet dragged through at varying speed is in contact for varying times. Pull slowly and steadily, and remember that whatever unevenness you leave here the gold bath will find, because less heavily coated areas tone more quickly and deeply.

The bath clouds or throws crystals when the alcohol goes in. The chloride has exceeded what the mixed solvent will hold. Do not filter and carry on: filtering removes the chloride you put in for a reason. Remake it, dissolving the salt in the water first, and if it still will not hold, work at the concentration the bath will carry and record what that was. See the arithmetic under Mixing.

Prints are thin and lack depth despite an obviously thick, glossy coating. The signature of a leached first coat: the halide is concentrated in the upper part of the layer because the lower part lost its chloride to a chloride-free or under-salted bath. Check the bath’s chloride against the albumen’s before you touch anything else — before the silver bath strength, before the exposure and long before reaching for ammonia fuming, which treats the symptom.

The second float traps bubbles where the first did not. Expected: a coagulated surface is less readily wetted than a raw one. Sweep the tray, use the wetting agent in the albumen, lay the sheet in one continuous movement, and start the clock only when it lies flat with no bubbles. With a thin rawstock a trapped bubble shows through as a light circle on a dark ground; with a thicker one you must lift the sheet and inspect its underside.

A thick hard rind along one edge, and sheets that will not lie flat. The runoff bead was left to dry. Blot it with a cloth several times during each drying, at both coats. Once it has set it is extremely difficult to work with and the last droplet takes a very long time to dry.

The finished paper is brittle and the layer cracks when handled. Reilly’s own warning, and it is a sign of a coating that has gone past useful: two coats “may even be so heavily coated that they are brittle and hard to tone”. Consider a shorter float on the second coat, a slightly diluted albumen for the second coat only, or accepting that this particular albumen and this particular rawstock want one coat. Handle flat, store flat, and if you must roll, roll albumen side out.

The sheet curls badly at the silver bath and will not float without the solution reaching its back. Also expected — Reilly says double-coated papers curl more and are harder to manipulate in sensitisation and printing. Condition the sheet overnight in a damp place first, which he recommends anyway because a too-dry sheet will not absorb the silver properly, and flatten under weights before the session rather than during it.

Toning takes far longer than it did on single-coated sheets, or stalls. Expected, and the remedy is strength rather than patience alone: albumen needs a much stronger toning bath than salted paper because it is less permeable and the silver is more protected, and a double coat is more of both. Use the gold bath at full strength and judge by inspection.

Yellow highlight staining appears in the months after printing. This is albumen’s characteristic failure and it is not specific to double coating, but a double-coated print has more protein and a harder fixing problem. Look first at the fixing: two fresh baths, four minutes each, alkaline, both made up on the day. The AIC’s account of the causes names the bonding of image silver to sulfur-containing side groups of the protein among them, and a print left with residual thiosulfate has a second source of sulfur on top of that.

A batch that behaved yesterday behaves differently today, with nothing changed. Check the bath’s strength before anything else. It is the one component of this procedure that changes measurably while it sits, in the one direction that causes the failure at the top of this list.

Experiments

The gravimetric experiment, which is the only one that answers the question directly. Every other test of double coating is indirect. Weigh the sheets. Take six sheets cut to the same size, weigh each dry to 0.01 g, coat all six once and re-weigh when dry; coagulate three and leave three; coat all six a second time and re-weigh. The three that were coagulated should have gained roughly as much on the second coat as on the first; the three that were not should have gained little or nothing. That is Reilly’s central claim, stated as a number, and it takes one afternoon and a jeweller’s balance. Record the mass gain per square metre — it is the figure that makes every other result on this page comparable, and no source publishes it.

The alcohol strength series, which tests the mechanism rather than the recipe. The same albumen, the same rawstock, five coagulating baths at 50, 60, 70, 80 and 99 per cent isopropanol, chloride matched in every one. Keep the five baths small — 200 mL in a small tray is enough to draw a half-sheet through — because the fire controls under Safety get stricter as the strength rises and the 99 per cent arm is the most flammable liquid this course puts in an open dish. Weigh as above. Reilly’s account predicts a maximum somewhere in the middle: too weak and the coat is partly lost, too strong and it is set unevenly. Look at the 99 per cent sheets in raking light for the unevenness he describes, and tone one from each strength, because the toning bath is a more sensitive thickness map than your eye is. If the reading of Reilly’s two bounds under The mechanism is right, the optimum should shift downward for a thicker or drier first coat, which is a second run of the same experiment with the first float lengthened.

The chloride-matching experiment, which tests the one ingredient whose job is invisible. Three baths at the same alcohol strength: chloride-free, matched to the albumen, and at double the albumen’s chloride. Print all three from one negative on one silver bath to the same visual endpoint. The prediction from Reilly’s figures is that the chloride-free sheets give a thinner, less rich image and a little more contrast from a thin negative, and that the over-salted sheets give nothing extra for the extra silver they consume. Measure maximum density with a reflection densitometer if you have one and by careful visual comparison against a step tablet if you do not.

Wall against Reilly: floating against immersion. Wall coagulates by floating and uses no chloride; Reilly immerses and insists on it. The hypothesis under Variants is that these two differences are connected — that a one-sided, shorter contact leaches much less. Test it directly: four combinations — floated or immersed, crossed with a chloride-free bath or a matched one — everything else held, judged on maximum density and on the gravimetric gain. This is the experiment the course would most like to see, because it would resolve a disagreement between two sources neither of whom addresses the other.

Heat against alcohol. The modern kit’s press at 250 °F for 2 minutes against Reilly’s alcohol bath, on the same albumen and the same rawstock, judged by weight gain and then by print. Add a third arm at a lower press temperature — say 80 °C, comfortably above the 65 °C coagulation threshold and well below 121 °C — to find out whether the supplier’s temperature is a requirement or a convenience. Watch for scorching and for the sheet cockling, which the supplier’s own instructions warn about.

The permeability experiment, timed with a clock. Reilly’s claim that thicker albumen is less permeable is directly testable and rarely tested. Single-coated and double-coated prints from one negative, toned in the same bath to a matched colour: record the time each takes. Then fix them side by side and test both for residual thiosulfate at four minutes, at eight, and at twelve. The result is worth having for its own sake, because it tells you whether your fixing regime is adequate for the paper you are actually making.

The reversal test, for the instruction that costs nothing. Four double-coated sheets: two with the marked edge reversed for the second float and two hung the same way twice. Print all four from a negative with a large even area — a plain sky is ideal — and read the density from top to bottom of each. This measures how much of the gravity gradient the reversal actually cancels, which Reilly asserts and nobody quantifies.

The scale-length measurement, which turns all of the above into numbers. Sensitise a test sheet of each paper, print it under a 21-step tablet until the step-1 patch matches the density of the margin outside the tablet, process normally and count the distinguishable steps. That count is the paper’s scale length and it tells you what density range your negatives for that paper should have. Run it on a single-coated and a double-coated sheet from the same albumen and you have measured, in the only units that matter, exactly what the second coat bought you.