Break/Fix: The Iron-Silver Print That Would Not Clear

Six prints are on the line and none of them is right. Two of the six are the same fault wearing different clothes, one of them is not a printing problem at all, and one of them was decided three weeks ago when a bottle was left uncapped.

The temptation is to change everything — fresh chemistry, a different paper, distilled water, a shorter fix — and print again. It would probably work, and it would tell you nothing. In this part in particular, that matters more than usual, because two of the six faults below do not appear until the print has been on a wall for a year, and a habit you did not identify is a habit you will still have when that year is up.

The situation

You have run the Van Dyke lab and the kallitype lab and you are back for a second weekend, printing both processes in one session on a new box of paper.

The bench: a bottle of Van Dyke sensitiser mixed six weeks ago, and a bottle of 20 per cent ferric oxalate opened in the spring and used four times since. You coated in the kitchen with the blind down and the ceiling light on, which is a fluorescent tube. You did not mask any of the sheets, so the coated area and the exposed area are the same rectangle. The kallitype developer is the sodium citrate bath from last time, topped up but not replenished by measure. The clearing bath is 3 per cent citric acid, one tray, used all afternoon. Your tap water is hard and reads pH 8.0. The hake brush is the one you also use for cyanotype.

The six results:

1. A Van Dyke whose highlights are a distinct buff-yellow, and which will not go white however long it washes.
2. A kallitype that looked fine when it dried in September and, taken out of the drawer in March, has visibly lost its shadows and gone yellow in the whites.
3. A Van Dyke that bleached away in the fixer, in front of you, over about four minutes, leaving a pale ghost.
4. A grey veil over a whole sheet, highlights and borders alike, on a print that is otherwise correctly exposed.
5. A kallitype with hard-edged pale patches and a diagonal light streak across one corner.
6. The ferric oxalate bottle, which was clear yellow-green in the spring and is now cloudy with something dark in the bottom.

Six symptoms, and — as in the cyanotype clinic — fewer than six faults. Two share a cause. One of them is the tap.

The symptoms

Describe what you see before you name it. Four distinctions do most of the work in these processes, and each separates two faults with different cures. Do it with the prints dry, in even daylight, beside a print you were happy with.

Say this	Not this	Because
The tone is grey	The tone is yellow	Ware’s distinction, and the most useful single sentence in siderotype diagnosis. Fog is unwanted image substance — silver made without light — and reads grey. Stain is unwanted residual chemistry, especially iron salts, and reads yellow. Different causes, different cures, and both can be present
The highlights lost density	The whole print lost density	A fixer that is too strong or too long empties the light values first, because the thinnest silver deposits lose the largest fraction of themselves. A print that was never dense enough is thin everywhere
It was like that when it dried	It changed in the drawer	The first is a printing fault you can reproduce. The second is a permanence fault, and the record is the only evidence you will ever have about it
The patch has a hard edge	The patch is a soft gradient	Hard edges come from liquid: an uneven pour, a dry spot, a splash. Soft gradients come from light, from coating and from drying

And one distinction specific to this part, which the cyanotype clinic did not need. Where in the sequence did it happen? These processes have four or five wet steps rather than two, and each of them can produce a light print. The clearing bath is supposed to lighten the print — King warns about it in terms, “the image will lighten considerably during clearing, but don’t worry because all the lost density will return during toning and fixing” — so a print that looks weak between the clearing bath and the fixer is not yet evidence of anything.

The three regions every sheet should carry

Almost everything below depends on comparing parts of one sheet with each other, and that is a decision you make before you coat, not after the print fails.

One sheet, three reference regions and a blocking patch

1. 1Uncoated margin — never sensitised. Any tone here is contamination in a bath or a tray, not a coating fault
2. 2Coated but masked border — sensitised, then covered during exposure. Any tone here happened without light — and this is also the clearing end point every kallitype source uses
3. 3Blocking patch beside the negative — a scrap of film opaque to the near ultraviolet. If it stays clean and the negative shadow does not, the negative is the fault
4. 4Picture area — the only region in which exposure is a candidate explanation at all

Read outward-in: margin, border, patch, picture. Drawn to show the arrangement, not to scale. The masked border does double duty in this part, because it is both the fog reference and the clearing end point.

Wynn White, who prints Van Dykes, cuts his mask from red construction paper and gives the reason in one clause: it “makes it possible to see whether or not the print has completely cleared or if there is any fogging”. Neither manufacturer’s sheet mentions masking at all. It is the single cheapest habit in this part, and today’s session did not have it.

Before you start

Level B, and the level is set by the bench rather than by any one substance. Diagnosis here means remaking sensitiser, coating, exposing, developing, clearing and fixing, so every hazard of the two process labs applies again: silver nitrate at H314 and H272, the oxalates at H302 and H312 with the kidney as the target organ, ultraviolet, and silver-bearing waste at every step. Gloves throughout, splash goggles for anything concentrated, one labelled vessel per solution, and nothing goes back into a stock bottle.

Three things about a diagnostic session in particular.

You will be making deliberately bad prints, so the trays sit fuller and longer than usual and there is more waste. Collect it under the silver-bearing waste SOP.

The residual-iron spot test uses a hexacyanoferrate reagent and the clearing baths are acid. Keep them physically apart, permanently and by layout: the test has its own dish, its own pipette and its own waste container, and nothing acidic comes to that bench. The safety card states that the solid decomposes on contact with acids producing toxic gases including hydrogen cyanide; Ware’s qualification is that the heating has to be very strong and the acid very concentrated; the course keeps the rule absolute anyway, because a part-used waste bottle is where the two actually meet.

And deteriorated sensitisers are waste, not drain. A condemned bottle of Van Dyke sensitiser is several grams of silver nitrate. It goes into the silver-bearing waste, labelled, and Part XII explains why that is an asset rather than a chore. Nothing oxalate-bearing shares a bottle with anything silver-bearing, on the safety card’s own warning about silver oxalate.

Alternative route

Without an ultraviolet unit the whole page works, and sunlight is the honest substitute for the Van Dyke, which is judged by inspection. For the kallitype it is worse than a substitute and the lab says why: with almost no print-out image the exposure has to be repeatable, so pair every differential and expose both halves in one session under an unchanging sky, recording conditions rather than a dose under the daylight exposure SOP.

Without a densitometer, everything here still works. Every discriminator on this page is a pattern — where the tone is, whether the border matches the picture, whether the fault was there wet or arrived on drying, whether a second sheet from a different box behaves differently. Patterns are read by laying two sheets side by side in even daylight. You lose the ability to say how much, which matters for the record and not for the diagnosis.

Without a spot-test reagent, the masked-border reading is the primary test in any case, and it is free. Record that the chemical test was not run rather than leaving a blank.

Without a second paper you cannot run the paper differential, which is among the most productive tests on the page. Any genuinely different sheet will serve, provided you write down what it was.

Without a wet bench at all there is no route to making the prints, and this page says so rather than pretending. What survives is the reading: the symptom vocabulary, the three reference regions, Ware’s algorithm and the case-by-case reasoning are all followable from a photograph of somebody else’s failed print, and are better read before the afternoon they matter than during it.

Evidence to collect

Cheapest first, and the cheapest tests here cost minutes.

The order to collect evidence in

1. 1Read the sheets you already haveDry, in daylight. Grey or yellow? Highlights or everywhere? Hard edge or soft? Where is the tone — margin, border, picture? Costs nothing and rules out whole branches
2. 2Test the waterpH by indicator paper, plus your supplier hardness figure. Two minutes, and it explains a large fraction of iron-silver failures on its own
3. 3Look at the two bottlesIs the Van Dyke sensitiser still clear? Is the ferric oxalate still yellow-green? A glance, and a bad bottle makes every other test meaningless
4. 4Run the ferricyanide test on the ferric oxalateTwo crystals and 2 mL of water. The suppliers publish it and it settles the commonest kallitype fault in about a minute
5. 5Spot-test a sacrificial corner of the failed printBlue means iron is still in the sheet. At its own station, with its own waste bottle
6. 6Coat a test sheet with all three regions and cut it into fourUncoated margin, masked border, blocking patch. This is the sheet Ware's algorithm is designed to read, and everything after it depends on having one
7. 7Run one differential at a timeWater, paper, sensitiser, bath. One variable per pair of half-sheets, both halves from the same coating
8. 8Only then change the exposureExposure is the variable everyone reaches for first and it explains fewest of these faults

The order is not arbitrary: the first four steps cost minutes and no materials, and between them they account for most of what goes wrong in this part.

The water test, in detail, because it is the one most readers skip. Fill a glass from the tap you wash in and read its pH; then find your supplier’s published hardness figure for your postcode. Photographers’ Formulary state the consequence for a Van Dyke in two sentences: “If your wash water is slightly alkaline, the iron salts will not be removed. Hard water is not satisfactory; it usually contains dissolved iron salts, which will contaminate the print.” Ware supplies the mechanism — above pH 4 iron(III) hydrolyses, and calcium promotes it — and adds the deadline: if the print dries with the hydroxide still in it, the hydroxide becomes goethite and dilute acid will no longer touch it.

The two spot tests, and what they do together. A drop of dilute potassium ferricyanide on a sacrificial corner goes blue if iron is present; a drop of the sulfide reagent from Part XII’s SOP goes brown or black if silver is present. No single published test separates residual iron from residual silver as the cause of a yellow-brown stain, and running both on the same strip is the course’s own combination rather than a published protocol — but it does separate them, and neither test alone does. The limits of the iron test are set out in full on the clearing lab and they matter here: no published sensitivity, responds to iron from any source, destroys what it touches.

Hints, in increasing order of disruption

Work down only as far as you need to. Each hint costs more than the one above it.

1. Look at the edge of the sheet before you look at the picture. Three of the six cases are decided by a region most printers trim off.
2. Ask what colour the fault is. Grey is image substance; yellow is residual chemistry. That one question splits the six into two groups of three.
3. Ask when the tone first appeared — on coating, during drying, after exposure, after wet processing, after the wash. That question is the whole of Ware’s algorithm.
4. Test the water and look at the bottles before you touch anything else. Both are free and both are more often the answer than the exposure is.
5. Run the ferricyanide test on the ferric oxalate. In a kallitype session it is the single highest yield per minute of anything on this list.
6. Split a sheet. One coating, cut in two after exposure, processed two ways. One variable, one coating, one exposure: the only comparison worth making.
7. Coat a sheet and expose it to nothing at all. A sheet that fogs in the dark has settled the question of whether light is involved.
8. Change the paper. The most productive single substitution in this cluster, and the one most readers try last.

The diagnosis

Ware’s algorithm, and why it belongs to this part more than to Part XXI

Mike Ware published a systematic separation of the causes of fogging — as Table 7.1 of Cyanomicon and as a standalone one-page handout — and Part XXI teaches it for the cyanotype. It applies here, and two features of it are written for these processes rather than for that one.

Cause six names residual iron. The handout’s sixth cause is “The clearing procedure is inadequate, leaving residual iron, etc.” A classic cyanotype has no clearing bath at all; a kallitype has two or three. That branch of the algorithm is a siderotype branch, and this part is where it is used in anger.

And the branches are marked POP and DEV. Ware’s own note is that “some ‘tests’ depend on whether the process is print-out (POP) or development (DEV). e.g. Fogging due to a faulty safelight may not be visible until wet processing is complete, especially for development processes.” This part is the only one in the cluster that contains both kinds, which is what makes those markings do work: the same fault announces itself at a different stage in a Van Dyke than in a kallitype.

The six causes, in Ware’s numbering, and where each one turns up in this part:

#	Ware’s cause	In an iron-silver session
1	The sensitiser is already decomposed	A ferric oxalate bottle past two or three months, which the ferricyanide test catches in a minute
2	A hostile chemical in the paper is reducing the sensitiser	The new box of paper, tested by drying a coated sheet in total darkness
3	The safelighting is inadequate under the conditions of working	The kitchen’s fluorescent tube
4	The mask or negative is not dense enough in its maximum value	Read against the blocking patch, and a print-out branch, so a Van Dyke test
5	The wet chemistry is faulty and is chemically reducing the sensitiser	The shared brush, and the developer nobody replenished
6	The clearing procedure is inadequate, leaving residual iron	The one tray of citric acid that did the whole afternoon

Ask, in this order	If yes
Fog apparent on coating?	Sensitiser decomposed (1), or a very hostile chemical in the paper (2)
Fog appears during drying? Then: does it still fog if dried in total darkness?	Yes — hostile paper chemical (2). No — faulty safelight (3)
Fog apparent after a print-out exposure?	Mask or negative not dense enough in its maximum value (4) — a POP branch, so a Van Dyke test
Fog after wet processing?	Sensitiser decomposed (1), faulty safelight (3), thin mask or negative (4), or faulty wet chemistry (5) — the DEV branch, and where a kallitype fault surfaces
Stain of sensitiser after the wash?	The clearing procedure is inadequate (6)
Stain in uncoated areas of the paper?	The wet chemistry is contaminated (5)

=How to run the algorithm when your failed print has no border

Today’s six prints were exposed without a mask, so the algorithm’s second column has nothing to read. You cannot run it backwards on a finished print; you run it forwards on the next one.

Coat one sheet with a margin left bare and a border that will be masked. Cut it into four before drying. Dry one quarter in room light, one in total darkness, expose the third under the negative with a blocking patch beside it, and process the fourth without exposing it at all. Four quarters, one coating, four stages — and between them they answer every row of the table in a single afternoon.

Case one: buff-yellow highlights that will not clear

This is stain, not fog, and the colour is the diagnosis — the fault the atlas files as highlights that will not clear. Yellow is residual chemistry, and in this family it is overwhelmingly iron. Ware’s identification notes for a platinum print name “pale yellow or straw-coloured staining — probably due to residual iron(III)” and add that it “may also occur in other siderotype processes, such as kallitype”.

Four candidates, and the differentials that separate them.

The clearing bath is exhausted. Cheapest to test: mix fresh and re-clear a sacrificial strip. The capacity is small — Photographers’ Formulary give two or three 8 × 10 prints per bath of 3 per cent citric acid, and King’s instruction is to “renew citric acid bath frequently, as this chemical is very inexpensive and proper clearing is absolutely vital to print stability”. Today’s session used one tray all afternoon.

The rinse before it was alkaline. Your tap reads pH 8.0. King: an alkaline first rinse “may form ferrous hydroxide compounds in the paper, making complete clearing difficult or impossible.” Photographers’ Formulary say the same from their side: water with a base pH “will make the print difficult to clear”. This is the tap, and it is the fault in today’s list that is not a printing problem.

The paper is unsuitable. Run the unexposed control sheet through development and clearing. King’s criterion is four to five minutes to a completely clear border; Photographers’ Formulary disqualify a paper needing more than ten. Today’s paper is a new box.

Or the developer was exhausted, so the sheet arrived overloaded. King’s warning is specific: an unreplenished developer accumulates iron and makes clearing progressively harder, producing “an unpleasant stain in the masked areas of the print”. Today’s developer was topped up but not replenished by measure.

And the test that settles which. The spot test tells you the iron is there. Which of the four put it there is answered by when: a first-print failure is the paper or the water, a failure that got worse through the session is the bath or the developer. Record the clearing time of every print and the answer arrives without an experiment.

!Two yellows, and no published test separates them

A yellow-brown highlight stain has two possible causes in these processes and they need different cures. Residual iron, which is what this case is about. And residual silver — unreduced silver salt that the fixer did not remove, which darkens over time and is a documented defect in its own right.

No source read for this course publishes a test that distinguishes them, which was one of the questions this page was set. What the course offers instead is the pair of tests above, run on the same sacrificial strip and marked as its own combination rather than a published protocol: ferricyanide goes blue for iron and the sulfide reagent goes brown for silver. A strip that gives both has both, which is not an unusual outcome, since a short fix and a short clear tend to travel together.

And there is a cheaper discriminator that costs nothing: the two faults have different histories. An iron stain is there when the print dries and gets worse slowly. A residual-silver stain often is not visible at all when the print dries and appears over weeks — which is case two.

Case two: the kallitype that faded in the drawer

This is the fault the whole part is arranged around, and it is the one you cannot diagnose today — the atlas case, seen from the printer’s end. The print was made in September, looked right, and has lost its shadows by March.

The mechanism. King states it plainly: residual iron(II), even in very small quantities, “will eventually oxidize the silver, and the image will fade”. Ware’s version for the family is that residual iron(III) will oxidise the image silver “with consequent degradation of the image”, and his structural point is why an iron-silver print is so vulnerable to it — the image is colloidal silver of about 20 nm with no binder over it, presenting a large surface area to anything that oxidises silver.

Fe³⁺ + Ag → Fe²⁺ + Ag⁺

The failure, in one line: the iron that made the picture takes it back

Three candidates, and the record is the only evidence.

Incomplete clearing, which is the majority answer and which case one has already put in front of you.

Incomplete washing, leaving thiosulfate and silver-thiosulfate complexes that decompose over years to silver sulphide. Photographers’ Formulary put the requirement in one sentence: for the print to be stable “the iron salt, the excess silver salt, and all the thiosulfate … must be removed”.

Or it was never fixed at all — or was fixed in an exhausted bath — leaving unreduced silver salt in the sheet, which is still light-sensitive and darkens on display while the image itself fades.

What the record shows in each case. If your log has the clearing time and it was long, suspect the clearing. If it has the wash time and it was short, suspect the wash. If it has the fixer’s age and how many prints had been through it, suspect the fixer. If it has none of those, you have learned nothing except that you need a better record, which is the real lesson of this case and the reason the clearing lab asks for an archival record separate from the session log.

⁇What a badly cleared kallitype looks like after a hundred years

The conservation evidence for this fault is thinner than it should be, and the reason is worth recording.

The obvious source is empty. The American Institute for Conservation’s Photographic Materials Group has a process entry whose title carries all seven names of this family — Van Dyke Brown, kallitype, brown print, sepia print, ferro-gallic, argentotype, argyrotype. Read on 7 September 2026, it carries its section headings, including Identification Characteristics, Conservation and Housing and Storage Considerations, and no text under any of them.

What the course has instead is Ware’s survival record. Dick Stevens, surveying many major American collections, found one historic kallitype. Ware’s own enquiry to curators and conservators turned up “only a handful”. And the few century-old specimens that have been identified “do show pronounced image fading and seriously yellowed highlights, suggesting the presence of residual iron(III)”. In early twentieth-century amateur archives he has himself identified probable kallitypes “mostly very deteriorated, showing severe iron stains, fogging, and fading”.

That is a description of case two at a hundred years, from a chemist who has looked at the objects, and it is the best evidence in this course’s corpus. It is not a conservation study, and the course says so.

Case three: the print that bleached away in the fixer

Watch the highlights, because they go first — the atlas files this as a printed-out image bleached in the fixer — and that is the discriminator between this fault and an under-exposed print.

Four candidates, in the order to test them.

The fixer was too strong — a rapid fixer or a paper fixer at working strength, which is four to five times the 5 per cent every source in this part specifies. Photographers’ Formulary’s warning is in their own emphasis: do not use a standard photographic fixing bath, or “the very finely divided silver metal of the Van Dyke print will be etched off the paper.”

It was too long. Both Van Dyke sheets give five minutes and both give a reason for the limit. Ware gives the mechanism from the argyrotype side: overlong immersion transforms the silver nanoparticles completely into silver sulphide, “causing the image to become badly faded”. The same bath that gives you the rich brown at two minutes takes the picture away at twenty.

It was too acid, or the water was. Ware’s objection to nitrate is that it “tends to dissolve the colloidal image silver during wet processing, especially under acidic conditions”, and Photographers’ Formulary’s New Kallitype sheet says the same from the practical end: “Acidic water accelerates the tendency of thiosulfate to bleach highlights.”

Or the print was under-exposed, and there was little silver there to lose.

The differential is cheap and it is the same one in both processes. Expose two identical strips from one coating. Fix one for one minute and one for five. If the short one is much stronger, the bath is the problem. If they match and both are thin, the exposure is.

And one thing that is not a fault. A kallitype is expected to bleach in the fixer. Bostick & Sullivan state it and give the only two answers they consider effective: tone before fixing, or overprint by a stop or two. A kallitype printer who has not toned and has not overprinted is seeing the process work as documented.

Case four: a grey veil over the whole sheet

Grey, not yellow, so this is fog: image substance where no light fell. In these processes that means silver reduced by iron(II) that was already present before the exposure, or made by something other than light.

Five candidates. Run Ware’s algorithm and they fall out in order.

The sensitiser is already decomposed — cause 1, and the first branch of the algorithm because it is free to check. For a kallitype this is the leading candidate and the test takes a minute: the ferricyanide test on the ferric oxalate. King’s warning is that a ferric oxalate solution degrades in water “with a resulting increase in print fog”, and his limit is two to three months. Today’s bottle was opened in the spring.

A hostile chemical in the paper — cause 2, and the test is Ware’s: dry a coated sheet in total darkness. If it still fogs, the paper is reducing the sensitiser. New box, new candidate.

The working light — cause 3, and today it has a name. The kitchen ceiling light is a fluorescent tube. Bostick & Sullivan warn about exactly this: fluorescent lighting may be used “but it is recommended that you limit the exposure of the chemicals and coated paper to them. This will limit the chances of fogging the image.” Wynn White found slight fogging on Van Dykes until he moved to a 7 W red bulb a metre above the coating area. Both Formulary kallitype sheets ask for a red safelight or very subdued incandescent light.

Heat during drying — a cause that is specific to this part and is not on Ware’s list. King: “A fan may be used to accelerate drying, but DO NOT force dry with heat, which may cause fogging.” Photographers’ Formulary cap it: do not heat the ferric oxalate solution, or the sensitised paper while drying it, above 50 °C, adding the honest qualification that tri-hydrogen ferric oxalate “is probably heat-sensitive, but the exact extent is not known”.

Or a contaminated brush — cause 5 in a different guise, and the subject of its own section below.

⚗Fog and stain are chemically different, and this is what each one is

The question this page was set is whether fog in these processes has a documented chemical cause distinct from stain, or whether the distinction is practitioner convention. It is chemical, and the sources define both sides of it.

Ware’s definition is the general one for siderotypes: fog is “unwanted residual image substance” and reads grey; stain is “unwanted other residual chemicals, especially ferric salts” and reads yellow.

In an iron-silver process that resolves to two named substances. Fog is metallic silver, reduced by iron(II) that was in the coating before the exposure — from a decomposed sensitiser, from a reducing agent in the paper, or from stray ultraviolet — and it is grey to grey-brown because that is what colloidal silver looks like. Stain is iron, mostly iron(III), hydrolysed in the fibres or bound to the cellulose, and it is yellow to straw-brown because that is what a hydrated iron(III) species on paper looks like; Ware’s mechanism for the deepening of that colour with time is the formation of oxo-bridged binuclear iron centres whose metal-to-metal charge transfer absorbs strongly in the visible.

The cures are as different as the causes. Fog cannot be removed — the silver is image substance in the wrong place — so the answer is always to prevent it. Stain can be removed, up to the point at which the print dries and the iron hydroxide becomes goethite. A reader who confuses them washes a fogged print for an hour and wonders why nothing happened.

Case five: hard-edged pale patches and a diagonal streak

Hard edges mean liquid — the atlas case is blotchy kallitype development — and in a kallitype the liquid is almost always the developer.

The pour. Every source warns about it because the image appears in seconds. Bostick & Sullivan: development “must be done quickly, or else watermarks may appear”. King pours the developer over a face-up print “as quickly as possible”. Photographers’ Formulary describe the tilted-tray method explicitly, holding the sheet clear of the pooled developer and lowering the tray so the solution runs across in one movement. Any part of the sheet that starts a second late shows it as a hard-edged patch.

Or the developer was cold, where a warm one was specified. The Formulary’s black-tone and brown-tone developers “work best if [they are] warm, (around 38 °C/100 °F or higher)”; the sepia one is a room-temperature bath given twice the time. A cold black-tone developer produces the same appearance more slowly and less obviously.

Or the developer is exhausted, in which case the clearing times through the session will already have told you.

The diagonal streak is a coating fault, not a development one, and the difference is the shape. Soft, directional, following the stroke: an atlas case. Wynn White adds a candidate specific to these processes that has nothing to do with technique: with some papers “the chemistry of the sensitizer doesn’t match the chemistry of the paper, resulting in blotchy or grainy images”, which is a paper substitution rather than a coating correction.

And one that looks like a coating fault and is a wet-processing one. Ware describes “bleeding” — a red-brown stain running off the dense areas during the first wash — as colloidal silver failing to be retained by the fibres, with remedies that are the sensitiser’s surfactant and the paper rather than the coating stroke.

Case six: the bottle that turned cloudy or dark

Four different clouds, and only two of them condemn a bottle. The atlas entry covers the general case; here is what each one is in this part.

A light precipitate in a fresh Van Dyke sensitiser is expected. Photographers’ Formulary say so: “A precipitate may or may not form. If a precipitate should form, disregard it” — with the one condition that it must not be transferred to the paper. On Ware’s reasoning for the analogous argyrotype it is most likely silver citrate, formed where the silver arrived somewhere locally rich before the acid had done its work. Decant and use.

A precipitate in a mixed kallitype sensitiser is expected too, and it has a name. Photographers’ Formulary: “It is very common for a precipitate of silver oxalate to form. This precipitate does no harm. If your sensitizer does form a precipitate, just be sure you do not transfer it to the paper you will be coating.” The free oxalate in a concentrated ferrioxalate solution is where it comes from.

A ferric oxalate solution that has gone from yellow-green to something else is decomposing, and this one does condemn it. The test is the ferricyanide test in the kallitype lab: in room light, yellow-brown to orange is good, green means a trace of iron(II), and “a blue cast — the deeper the blue, the poorer the quality”. King’s keeping figure is two to three months in solution; Ware’s from the literature is six to nine months in the dark, with the note that at least one respected platinum printer makes a fresh solution the night before every session. Today’s bottle was opened in the spring.

And a Van Dyke sensitiser that has gone dark is a different judgement. All three sources give it a long life: Photographers’ Formulary say it “will remain active for months”, Bostick & Sullivan give their premixed solution at least a year, and Wynn White reports a bottle “sitting around for a year or so” that was fine. No source read for this course states what a decomposed Van Dyke sensitiser looks like, and the course will not invent an appearance for it. What is documented is the direction: the solid citrate is light-sensitive and deliquescent, the mixed solution is light-sensitive, and both suppliers store it in a brown bottle in the dark.

The storage rule that follows from all of that. Brown bottles, dark, cool, capped — the Formulary’s New Kallitype sheet adds “don’t allow them to set for long periods with the cap off” — dated on the label, and tested against a fresh strip before any session that matters. For ferric oxalate, add the ferricyanide test to that list; it is a minute and it is the highest-yield minute in a kallitype session.

✕When to stop diagnosing and remake the bottle

There is a point at which further diagnosis costs more than the chemistry, and it is worth naming because printers pass it regularly.

Remake rather than diagnose when all three of these are true. The bottle is past its published keeping time. A cheap test — the ferricyanide test, or a coated strip beside a fresh one — points at it. And the fault is fog or a general weakness rather than something with a shape.

The arithmetic is on the side of remaking. A litre of kallitype developer is a few pounds of salts; a bottle of ferric oxalate is a few millilitres. Against that, an afternoon of prints costs paper, silver and the afternoon. The exception is the Van Dyke sensitiser, which carries several grams of silver nitrate and is the one bottle in this part worth arguing about — and which is also the one with the longest documented life.

And whatever you do, do not filter a decomposed sensitiser and carry on. Filtering removes the precipitate; it does not restore the iron(III) that has become iron(II), and it does not give back the silver that has already been reduced in the bottle.

Cross-contamination, which by now is the commonest inexplicable result

You are five processes into this cluster and the brushes have started to accumulate history. This is the fault that produces results nobody can explain, and it has a cheap and absolute answer.

Bostick & Sullivan give the rule on both their sheets. A brush used for one process is used for that process only: “Avoid using a brush that has been used to coat other alternative process chemistry to limit contamination. A glass coating rod can be used for multiple handcrafted processes if you wash it properly between uses.”

The chemistry behind it is specific and it is worth knowing which pairs are dangerous.

The residue	Meets	And produces
Cyanotype ferricyanide in a brush	An iron-silver sensitiser	Prussian blue where the iron(II) is, which is the ferric oxalate quality test performed accidentally on your print
An iron-silver sensitiser in a brush	A cyanotype sensitiser	Silver in a cyanotype, and a blue print with brown specks
Silver nitrate anywhere	An oxalate solution, dried	Silver oxalate, which the oxalic acid card names as explosive
An acid clearing bath	A hexacyanoferrate reagent or its waste	The one genuine gas hazard in this part, kept apart by layout
A metal implement	Any silver solution	Reduced silver, and a permanent mark

The absolute rules, and there are only three. A brush belongs to one process. A rod may be shared because it can be got genuinely clean, and is rinsed the moment it is put down. And nothing metal touches anything. See the atlas entry for the full cause list.

Today’s session used the cyanotype hake brush. That is worth a differential of its own: coat one sheet with the suspect brush and one with a rod from the same cup, and compare the borders.

How these faults relate to Part XXII, and what is new because of the iron

Three of today’s six would have been recognisable in the salted-paper clinic, and three would not.

Shared with every printed-out silver process. Bleaching in the fixer, because the image is unprotected colloidal silver in both. Fading from residual thiosulfate and from unreduced silver salt. And sulphiding and mirroring on a print that has been over-fixed or badly stored — Ware’s account of the fixer’s colour shift is the benign version of the same chemistry.

New here, and all three are the iron. The yellow stain, which a salt print does not have because there is no iron in it. The clearing bath, which exists only because of that stain and which is the step neither Van Dyke supplier sheet includes. And the fog from a decomposed sensitiser, which in a salted-paper process means silver reduced by something, and here means iron(II) that got there before the light did.

One difference of emphasis is worth stating. In Part XXII the characteristic long-term failure is the silver itself — sulphiding, mirroring, oxidation of colloidal metal. Here the silver is the same and there is a second reagent in the paper actively attacking it. That is why this part has a clearing lab and a toning lab where Part XXII has a toning lesson, and it is the whole of what iron adds to the family.

The fix

Case one, buff highlights. Mix a fresh clearing bath and re-clear a sacrificial strip; if that fixes it, the bath was exhausted and the cure is to renew it every two or three prints and to record clearing times. Acidify the rinse water to just below pH 7, or use distilled — this is today’s tap. Run the unexposed control sheet through development and clearing before committing more of the new paper. And replenish the developer by measure rather than by topping up.

Case two, the print that faded. There is no fix for that print. The fix is the record and the regime: clear to a tested end point, wash to the published time, tone. The clearing and toning lab is the whole answer, and King’s position — that all untoned kallitypes will eventually fade — is the reason it is not optional if the print matters.

Case three, the bleached print. Mix the fixer fresh at 5 per cent, check the pH of the water it is made with, hold it at about 20 °C because warmer water accelerates the bleaching, and time it. For kallitypes, tone before fixing or overprint by a stop or two, which are the only two answers Bostick & Sullivan consider effective.

Case four, the grey veil. Test the ferric oxalate; if it is blue-cast, discard it into the silver-bearing waste and buy fresh. Switch the fluorescent tube off and work under tungsten or warm LED with the window shaded. Dry with a fan and not with heat. Then re-run the four-quarter test sheet to confirm which one it was, because you have just changed three things at once and the point of a diagnosis is to know which.

Case five, the patches. Adopt the tilted-tray pour, put a thermometer in the developer, and use a tray at least one size larger than the sheet. If the streak survives all of that, change the brush, then the paper.

Case six, the bottle. Discard the ferric oxalate. Keep the Van Dyke sensitiser and test it against a fresh strip. Label and date everything, and add the ferricyanide test to your pre-session routine.

And the change that fixes three of the six at once: mask every sheet. It costs a piece of red film and it gives you the border that Ware’s algorithm, King’s clearing end point and the fog-or-stain distinction all need.

What to record

Two records, and the second is what makes a failure into evidence.

The diagnostic log, one line per test. What you changed, what you held constant, what happened, and what it eliminated. A test whose result you did not write down is a test you will run again in March.

And the archival record, for any print you intend to keep, which is the one case two proves you need:

Field	Why it is there
Date, process, paper and batch	The paper is the commonest single fault in this part
Sensitiser batch and its age in days	Both bottles have keeping limits and one of them is short
Working light, and drying method and temperature	Two of Ware’s six fogging causes live here
Developer, its temperature, and how many prints it had done	Exhaustion shows up as a clearing failure, not a thin print
Rinse water pH	The single most under-recorded variable in this part
Clearing bath, number of baths, and the time to a white border	This is the measurement that predicts case two
Spot-test result, before and after	The only direct evidence about residual iron you will ever have
Toner, strength and time	The other half of the permanence argument
Fixer strength, temperature, time, and its age	Case three, and half of case two
Wash time	The other half of case two

Three habits worth more than the table. Mask every sheet. Time the clearing rather than watching it. And keep the failures — labelled, dated, with their records — because a print that fades in a drawer is the only experiment in this part that runs at the timescale the part is actually about.

Recap

Colour first. Grey is fog and is image substance made without light; yellow is stain and is residual chemistry, mostly iron. The two have different causes, different cures, and one of them cannot be washed out at all.

Then ask when it appeared, which is the whole of Ware’s algorithm — an algorithm written for siderotypes, whose sixth cause names residual iron and whose branches are marked for print-out and development processes because this part contains both.

Almost every fault here is the paper, the clearing or the fixer, and the tests for all three are cheap: a pH strip, an unexposed control sheet through the clearing bath, and two identical strips fixed for different times.

The two tests that catch four of the six are the ferricyanide test on the ferric oxalate bottle, before the session, and the residual-iron spot test on a sacrificial corner, after it.

Cross-contamination is now the commonest inexplicable result. A brush belongs to one process, a rod is shared only because it can be got genuinely clean, and nothing metal touches anything.

And what iron adds to the salted-paper fault list is exactly three things: the yellow stain, the clearing bath that exists because of it, and a fog whose cause is iron(II) that arrived before the light did.

Check your understanding

Question 1. A print has a tone in the highlights. What single question splits the diagnosis fastest, and what are the two answers?

• How long was the exposure — too long or too short
• What colour is it — grey means fog, which is image substance made without light; yellow means stain, which is residual chemistry and in this family means iron
• Was it a Van Dyke or a kallitype
• Was the paper buffered or unbuffered

Show the answer and why

Answer: What colour is it — grey means fog, which is image substance made without light; yellow means stain, which is residual chemistry and in this family means iron

Ware's distinction is the most useful single sentence in siderotype diagnosis, and in an iron-silver process it resolves to two named substances: fog is metallic silver reduced by iron(II) that was there before the exposure, and stain is iron, mostly iron(III), hydrolysed in the fibres or bound to the cellulose. The cures are as different as the causes - fog cannot be removed and must be prevented, while stain can be cleared right up until the print dries and the iron hydroxide becomes an oxyhydroxide that dilute acid will not touch.

Question 2. Which two tests, run on every session, would catch four of the six faults on this page before the print was dry?

• A pH test on the wash water and a densitometer reading of the maximum density
• The ferricyanide test on the ferric oxalate bottle before the session, and the residual-iron spot test on a sacrificial corner after it
• A safelight fog test and a clearing-time measurement
• A step wedge in every print and a control strip through the fixer

Show the answer and why

Answer: The ferricyanide test on the ferric oxalate bottle before the session, and the residual-iron spot test on a sacrificial corner after it

The ferricyanide test on the bottle catches the decomposed sensitiser, which is the leading cause of fog in a kallitype and the first branch of Ware's algorithm, and it takes a minute with two crystals. The spot test on the print catches residual iron, which is the cause of case one and the mechanism of case two. Between them they cover the buff highlights, the fade in the drawer, the grey veil and the cloudy bottle. The water pH test is genuinely valuable and belongs in the same routine; it is one test rather than two, and it catches a subset of the same faults.

Question 3. You are told Ware's fogging algorithm was written for cyanotype. Which features of it show that it is a siderotype algorithm covering this part as well?

• Its sixth cause is "the clearing procedure is inadequate, leaving residual iron", and a classic cyanotype has no clearing bath
• Its branches are marked POP and DEV, and this part is the only one in the cluster that contains both kinds of process
• It asks the reader to compare a coated-but-masked border with an uncoated margin
• It names ferric oxalate as the sensitiser throughout

Show the answer and why

Answer: Its sixth cause is "the clearing procedure is inadequate, leaving residual iron", and a classic cyanotype has no clearing bath, Its branches are marked POP and DEV, and this part is the only one in the cluster that contains both kinds of process

Cause six names residual iron, which is a fault of the processes that need a clearing bath rather than of the one that does not, and the print-out and development markings only do work where both kinds of process exist - Ware's own note is that fogging from a faulty safelight may not be visible until wet processing is complete, especially for development processes. The border-and-margin comparison is general good practice and is not specific to any family. The handout names no sensitiser at all; it is written at the level of the class.

Question 4. Three prints from one afternoon: the first cleared in four minutes, the third in seven and the sixth in twelve. What has happened?

• The paper is too absorbent, which would have shown on the unexposed control sheet
• Something is accumulating through the session - the clearing bath is exhausted, or the developer is loaded with iron and each sheet arrives carrying more of it - and the fix is to renew the bath and replenish the developer by measure
• The rinse water became more alkaline as the session went on
• The fixer is exhausted, which shows up as a clearing problem

Show the answer and why

Answer: Something is accumulating through the session - the clearing bath is exhausted, or the developer is loaded with iron and each sheet arrives carrying more of it - and the fix is to renew the bath and replenish the developer by measure

The pattern is the diagnosis: a fault that is present on the first print is the paper or the water, and one that grows through the session is a bath filling up. Photographers' Formulary give a clearing bath a capacity of two or three 8 × 10 prints, and King's warning about an unreplenished developer is that accumulated iron makes the print progressively harder to clear and leaves a stain in the masked areas. The fixer comes after the clearing bath and cannot cause it; and a paper fault would have shown on print one.

Question 5. A yellow-brown stain in the highlights could be residual iron or residual silver. How do you tell?

• A published spot test distinguishes them by the colour it produces
• No published test in this course's sources separates them; the course runs both spot tests on one sacrificial strip - ferricyanide goes blue for iron, the sulfide reagent goes brown for silver - and marks that pairing as its own combination
• Residual silver is always darker than residual iron
• Only a densitometer reading in the blue channel can separate them

Show the answer and why

Answer: No published test in this course's sources separates them; the course runs both spot tests on one sacrificial strip - ferricyanide goes blue for iron, the sulfide reagent goes brown for silver - and marks that pairing as its own combination

This was one of the questions the page was set, and the honest answer is that the corpus contains no single published test for it. Running both tests on the same strip does separate them, and it is common for a strip to give both, since a short fix and a short clear tend to travel together. There is also a free discriminator in the history: an iron stain is present when the print dries and worsens slowly, while a residual-silver stain is often invisible at first and appears over weeks.

Question 6. When is the right answer to stop diagnosing and remake the sensitiser?

• As soon as two prints in a row go wrong
• When the bottle is past its published keeping time, a cheap test points at it, and the fault is fog or general weakness rather than something with a shape - and the exception is the Van Dyke sensitiser, which carries several grams of silver and has the longest documented life
• Never; a sensitiser can always be restored by filtering
• Whenever a precipitate appears in the bottle

Show the answer and why

Answer: When the bottle is past its published keeping time, a cheap test points at it, and the fault is fog or general weakness rather than something with a shape - and the exception is the Van Dyke sensitiser, which carries several grams of silver and has the longest documented life

The economics favour remaking: a bottle of ferric oxalate is a few millilitres and an afternoon of prints costs paper, silver and the afternoon. Filtering is the trap - it removes the precipitate and does not restore iron(III) that has become iron(II) or give back silver already reduced in the bottle. And a precipitate is not by itself a condemnation: both Photographers' Formulary sheets describe an expected precipitate, silver oxalate in the kallitype sensitiser and an unnamed one in the Van Dyke, and both say to disregard it provided it is not carried onto the paper.

Check my answersClear