Break/Fix: The Film That Came Back Wrong

The situation

You have read the seven mechanism pages of this part. This session tests whether you can use them, which is a different thing.

Five sets of symptoms follow, each with the session record that produced it and without the answer. Work each one before reading its hints, write down what each observation eliminates, and only then look at the diagnosis. A sixth case is given at the end with no hints at all.

After the cases, you will produce two faults deliberately — safelight fog on paper and a deliberately under-fixed print — so that two of the most consequential signatures in this part are learned from specimens you made on your own paper, under your own safelight, rather than from a description.

The five cases:

1. A thin, flat 35 mm film with clear rebates.
2. A print with a grey border and a base that will not stay white.
3. A negative with a regular pattern of marks at the film edges.
4. A paper that will not reach black in a freshly mixed developer.
5. A strip of film with a violet cast, and a yellow stain that appeared weeks later.

Before you start

Level A. The practical half of this session uses working-strength print developer, stop bath, fixer and wash at the dilutions the manufacturers publish, in ordinary trays, at room temperature. Nothing is heated, nothing concentrated is handled, no mains equipment is opened and no bath is mixed from solids. The controls are the ones the whole printing sequence carries: nitrile gloves, eye protection, ordinary room ventilation, tongs, and one dedicated labelled vessel per bath. Kodak’s own handling publication for photographic processing chemicals asks for no more than that at these concentrations — gloves and eye protection, adequate ventilation, chemicals kept off the skin, and no eating or drinking in the work area.

What is not a hazard here, and why. There is no acid concentrate on this bench: the stop bath is a citric-acid concentrate already diluted 1+19, and the fixer is an ammonium thiosulfate concentrate at 1+4. Neither evolves a vapour at these strengths in a room with ordinary air movement, which is why ventilation appears as a control and not as an engineering requirement. There is no potassium ferricyanide anywhere in this session, so the acid incompatibility that sets Level B on the reducer pages does not apply. And the fixer’s own hazard is to the drain rather than to you: spent fixer carries dissolved silver.

Three points specific to this page.

The deliberately under-fixed print is a specimen that will keep changing. It carries residual thiosulfate and residual silver by design. Label it with its processing details and its date, store it in its own sleeve, and keep it away from finished work, because residual thiosulfate can migrate to prints stored in contact with it. It is a reference, not a photograph, and it belongs in a labelled envelope rather than in a box with your prints.

The safelight specimen is run with the enlarger off and the room otherwise dark, so that the only variable is the safelight. If the enlarger’s lamphouse leaks, you will measure the leak and call it the safelight.

Waste. Spent fixer is collected separately for silver recovery; spent developer separately again. Local regulation governs both, and the course cannot tell you what your authority accepts. Part XII owns the silver-recovery route.

Estimated cost

Band ££. No capital purchase is needed beyond a working darkroom: the session uses the enlarger, trays, tongs, thermometer and safelight you already have, and the diagnostic half needs only negatives, a light box or a window, a loupe and a rule. See the planner for the standing cost of the darkroom itself.

Estimated consumables cost

Consumed	This session	Sourced price	Cost this session
Photographic paper, variable-contrast RC, 5 × 7 in	6 sheets — one range-finding, one safelight specimen, one safelight control, three for the fixing series	£16.06–£44.71 per 25 to 100 sheets, dated 5 September 2026, which is £0.45–£0.64 a sheet	£2.68–£3.85
Paper developer, working strength	500 mL of 1+9, that is 50 mL of concentrate	£10.52–£20.03 per 500 mL to 1 L of concentrate	£0.53–£2.00
Stop bath, working strength	500 mL of 1+19, that is 25 mL of concentrate	£10.66–£12.18 per 500 mL of concentrate	£0.53–£0.61
Fixer, working strength	500 mL of 1+4, that is 100 mL of concentrate	£21.05–£25.98 per 1 L of concentrate	£2.11–£2.60
Fibre-based paper, one sheet, optional	1 sheet, if you want the fixing series on the material where under-fixing does its worst	The planner’s only fibre record is 8 × 10 in, £47.04 for 25 sheets to £150.28 for 100, or £1.50 to £1.88 a sheet; this session works at 5 × 7 and no fibre listing at that size is recorded	—
Sodium sulfide solution for the residual-silver test	a few drops, from Part XII’s session	Not priced by the planner, which names sodium sulfide as a gap and this session as one of the three reasons for it	—
Wash water	a few tens of litres	Not priced by the planner, which names it as a gap of a different kind: a water tariff rather than a retail listing	—
Sleeves, labels, chinagraph	one sleeve and two labels	None. The sleeving and the bench stationery are two separate gaps the planner names	—

The priced rows give £5.85 to £9.06 for one run. That is a floor and not a total: four rows carry no dated price, and the concentrate figures assume the smallest useful volume is mixed and discarded after the session, which is what a controlled specimen requires and is more wasteful than a printing evening would be. The five paper cases and the sixth blind case cost nothing at all — they are worked from the page.

Equipment is excluded on purpose: the enlarger, safelight, trays, tongs, thermometer, timer, loupe and light box are not consumed by this session, and a page that quietly counted them would distort the figure the consumables calculator needs.

Alternative route

The five cases, the sixth blind case, the evidence list and the atlas-entry exercise need no darkroom. They are the larger half of this page by time, they are the half that teaches the method, and they are worked with a pencil.

The two specimens need a darkroom and there is no substitute that makes a specimen. What can be substituted is narrower and the page says so rather than pretending otherwise.

• For the safelight specimen, substitute a room test. Take a sheet of paper out of the box inside a changing bag, cut it in two, keep one half in the bag, and lay the other on a surface in whatever space you load film in, under whatever light that space has, for a stated time. Develop both later, or have them developed together. You will have measured your loading space rather than your safelight — a smaller question, honestly answered, and one that matters if you load film at all.
• For the under-fixed specimen, there is no substitute that produces one, and inventing a “simulation” would be exactly the fabrication this course forbids. What survives is the reading half: find a print that has yellowed — a family album, a junk shop, a charity-shop frame — and work the fixing page’s location-based diagnosis on it. Somebody else made you the specimen.

Without a densitometer, everything on this page still works. Every discriminator here is a comparison — this step against that step, this half of the sheet against the other half, this rebate against a known-good rebate — and comparisons are read by eye on a light box. What you lose is the ability to say how much, which matters for the record and not for the diagnosis.

The symptoms

Case 1 — the thin, flat film

What you have. A 36-exposure roll of HP5 Plus. Every frame is thin and flat: the shadows show no detail at all, the highlights are weak, and printed on grade 2 the whole roll is grey. The rebates are clear and look exactly like the rebates on the last roll, and the edge numbers along the bottom are crisp and easy to read.

Session record.

Film	HP5 Plus, 135-36, in date, from a box kept in a cupboard
Camera	35 mm SLR with through-the-lens metering; the roll before this one was DELTA 3200
Developer	ID-11 stock, freshly mixed from a sealed packet, first film through it
Time and temperature	7½ minutes at 20.0 °C, measured with a spirit thermometer checked against a second one
Agitation	Four inversions in the first 10 seconds, four again in the first 10 seconds of each minute
Stop, fix, wash	ILFOSTOP 1+19 for 10 s; RAPID FIXER 1+4 for 3 minutes; the fill-and-invert wash

Case 2 — the grey border

What you have. Every print from one evening has a grey border where the easel blades masked the paper, and a base white that is not white anywhere. Prints look flat at every grade; going harder makes the light tones worse rather than better. The effect is not equal across the evening: the prints you fussed over are worse than the ones that went straight through.

Session record.

Paper	MULTIGRADE RC DELUXE, opened six months ago, kept in its bag inside its box
Safelight	ILFORD 902 filter, installed when the darkroom was built in 2019; 25 W bulb fitted last winter when the old one failed; lamp 0.7 m above the developer dish
Developer	MULTIGRADE 1+9, freshly mixed, 20 °C, 1 minute
Working habit	Prints inspected face-up in the developer under the safelight; a difficult print may sit there two or three minutes while you decide
Enlarger	Off during processing; room door has a folded towel along the bottom

Case 3 — the regular pattern at the film edges

What you have. A 35 mm negative with a regular series of small density differences along both long edges, encroaching a few millimetres inwards. They repeat at a fixed interval. The middle of every frame is clean. It is on the whole roll.

Session record.

Film	FP4 Plus, 135-36
Developer	ID-11 1+1, 20 °C, 13 minutes
Agitation	“Vigorous inversion for the full 10 seconds at the start of every minute, and again at 30 seconds, because the last roll was mottled”
Tank	Two-reel plastic tank with one film in it and 300 mL of solution
Everything else	Standard, and the same as the previous six rolls

Case 4 — the paper that will not go black

What you have. A test strip in which the steps stop getting darker after the fourth step, and the tone they stop at is a dark grey rather than a black. The paper’s white is clean. More exposure adds nothing below that step. It happens on every sheet.

Session record.

Paper	MULTIGRADE FB CLASSIC, glossy, new box, in date
Developer	MULTIGRADE 1+9, mixed twenty minutes ago, 20 °C, 2 minutes with continuous rocking, four sheets through it
Stop	ILFOSTOP 1+19, 10 seconds
Fixer	RAPID FIXER 1+4, fresh, 4 minutes — “the film time, to be safe”
Wash	45 minutes running water
Safelight	Tested last month and labelled: safe working time 3 minutes at the dish

Case 5 — the violet cast and the late yellow stain

What you have. A strip of 35 mm negatives that came out of the fixer looking slightly milky and violet in the rebates. You decided it was the film’s dye and filed it. Six weeks later the strip has a yellow-brown stain, strongest in the clear areas, and the strip next to it in the same sleeve has begun to look faintly yellow at the edge where the two touch.

Session record.

Film	FP4 Plus
Fixer	RAPID FIXER 1+4, mixed eleven weeks ago, “about twenty films” through it, no running total kept
Fixing time	3 minutes, by habit; no clearing-time test was ever run on this bath, fresh or used
Wash	The fill-and-invert method, but “two changes rather than three, because it was late”
Storage	Cut into strips and sleeved the same evening, in a polyester sleeve, in a folder in a warm cupboard

Evidence to collect

The same list every time, before any hypothesis. Collecting it takes ten minutes and it is the difference between a diagnosis and a preference.

The evidence list, and why each item is on it

1. 1The rebate, and base plus fogAgainst clear fixed base or a known-good strip. It is the only control you have that the lens never touched, and it either eliminates a whole family of causes or redirects the entire diagnosis.
2. 2Both sides of the material under raking lightA lamp at a grazing angle across the surface, and a loupe focused on one surface and then the other. This is what separates a scratch from a mark, and an emulsion-side event from a base-side one.
3. 3The leader and the unexposed endThe leader was exposed to full room light when you loaded the camera and should be at maximum density. A grey leader convicts the developer or the fixer without any reference to the pictures.
4. 4Frame-to-frame consistencyEvery frame the same points at the process, the film or the camera body. One frame different points at that exposure. Do this before anything else that costs money.
5. 5The session recordTimes, temperatures, dilutions, agitation, and the running total on every bath. If there is no record, that is itself a finding and it is the first thing to change.
6. 6The age and history of every solution usedWhen mixed, how stored, how full the bottle was, and how much has been through it. Half the faults in this part were created on a shelf.

Write it down before you have an opinion. The value of the list is that it is collected in the same order every time and independently of what you suspect, which is what stops you from gathering only the evidence that supports the interesting hypothesis.

The questions in order, and the discipline

Ask them in this sequence, and after each one write down what the answer eliminated — not what it suggested.

1. Is it on every frame or on one?
2. Is it inside the image area, or does it reach the rebate and the inter-frame gap?
3. Does it follow a feature of the equipment, or a feature of the picture?
4. Is it periodic, and does the period match something you can measure?
5. Did it exist when the material was wet, or did it arrive later?
6. Does it appear on other sessions’ material, or only this one?

✕The discipline is the writing, not the questions

Anyone can ask six questions. What separates diagnosis from guessing is the record of what each answer ruled out, and the reason is not tidiness.

Without the record, a hypothesis that has already been eliminated comes back — because it is interesting, or because somebody suggests it, or because the next observation is ambiguous and it fits. With the record, you can look at the line that says rebate normal: development and film history eliminated and stop.

Kodak’s Z-133 builds the same discipline into a laboratory procedure from the other end: before acting on an out-of-control process at all, verify it — check the densitometer, and run a second control strip. And after every corrective action in its troubleshooting table, without exception: process another control strip to confirm that the change returned the process to control before resuming normal work. Verify before, verify after. Neither is optional and both are usually skipped.

Hints, in increasing order of disruption

Take only the ones you need. Each case has three, and the third gives the answer away.

Case 1. (a) What in the strip is evidence about the developer alone, independent of anything the camera did? (b) The edge numbers were exposed at the factory; what does it mean that they are crisp? (c) What was the previous roll, and what does a camera’s ISO setting do between rolls?

Case 2. (a) The border was masked by the easel. What light could have reached it? (b) The prints you fussed over are worse. What accumulates with time? (c) Two numbers in the session record are outside the manufacturer’s published recommendation, and one of them is not the one you first thought of.

Case 3. (a) Measure the interval. What else on the film has that interval? (b) Which end of the agitation ladder produces marks with a direction and a period? (c) Read the reason given in the record for the agitation scheme, and ask what fault it was fixing.

Case 4. (a) The developer is fresh, the paper is new and the base is clean; which steps of the process remain? (b) Compare the fixing time in the record against the maker’s published time for paper, not for film. (c) What is the manufacturer’s own word for what a long fixation does to a print?

Case 5. (a) Two faults are described, six weeks apart. Are they one event or two? (b) What test would have distinguished them at the time, and can it still? (c) Notice what has happened to the neighbouring strip in the sleeve.

The diagnosis

Case 1 — underexposure by two stops, with normal development

The diagnosis. The film was exposed at EI 1600 and developed for EI 400.

The evidence that decides it. Two observations, and they are independent.

The rebates are normal and the edge numbers are crisp. Edge numbering is a latent image put on the film at the factory; nothing but development brings it up, and nothing the camera does can affect it. Crisp edge numbers and a normal rebate density say the developer did its job, at the right strength, for long enough. That eliminates every development cause in one look.

The shadows show no detail at all. That is the exposure witness from the negative page: if the darkest important tone carries no density above base plus fog, it was never recorded.

Why it is flat as well as thin, which is the part that misleads people. Underexposure does not only move the negative down; it moves the subject’s whole range left onto the toe of the curve, where the gradient is shallow. A subject range that would have been rendered across a healthy stretch of the straight line is now compressed into the toe, so the negative is low in contrast as well as low in density — from one cause, not two. A reader who reads “thin and flat” as “underexposed and under-developed” has invented a second fault.

The confirming test. Look at the camera’s film-speed setting. The previous roll was DELTA 3200, and a dial set for it and not reset is two stops of underexposure on a 400 film. If the camera has no manual dial, look for a DX override. Then shoot a short roll of the same stock at EI 400 through the same camera and develop it identically; if it is normal, the camera and the process are both exonerated and the setting was the fault.

The fix. Nothing for these negatives — no development change, no intensifier, no reducer. The information is not there. Print what can be printed on a hard grade, accept that the shadows are empty, and reshoot anything that matters.

The prevention, as a procedure rather than a resolution. Add “check the film-speed setting” to the loading routine, in the same breath as checking the frame counter. A resolution to be more careful fails the next time you are in a hurry; a step in a routine does not.

Case 2 — safelight fog, from a wrong bulb at a wrong distance through an aged filter

The diagnosis. The border was masked from the enlarger and is grey, so it was fogged by light that was not the enlarger’s. Everything in the record points at the safelight, and three separate things about it are wrong at once.

The evidence.

• The border is grey. The easel blades masked it from the enlarger, so the exposure that greyed it did not come through the lens.
• The severity tracks time under the lamp. The prints you fussed over sat longest, and they are the worst. That is a dose accumulating with time, which is what a safelight delivers and what a constant leak does not.
• The base will not stay white and the print looks flat at every grade. ILFORD describe exactly this as the appearance of low-level safelight fogging: “not seen as safelight fog, but only as a general loss of photographic quality, particularly reduced contrast and lack of clear highlights.”

Three published numbers, three violations.

In the record	Published recommendation	Source
25 W bulb	15 W	ILFORD MULTIGRADE FB CLASSIC; Kodak K-4
Lamp 0.7 m from the dish	not less than 1.2 m (4 ft)	ILFORD safelight sheet; Kodak K-4
Filter installed 2019, never changed	change each year and record the installation date	ILFORD safelight sheet

The filter is the one people miss, and it is the one that fails invisibly: a faded dye passes shorter wavelengths than it did when new, and the lamp looks the same colour to you.

The confirming test. The packet test, on one sheet. Lay an opaque object on a sheet of paper at the place where the safelight is brightest — usually over the developer dish — leave it for the time a print actually spends there, and develop it. The shape of the object on a grey ground is the answer. Then run Part XVI’s full test to establish a safe working time and label the lamp with it.

The honest note. The packet test convicts the room, not specifically the safelight. Run it once with the safelight on and once with everything off, and the difference between the two is the safelight’s contribution. The record’s folded towel along the door is not a light seal, and it is worth eliminating before you buy a filter.

The fix. A 15 W bulb, the lamp moved to at least 1.2 m, a new filter with the installation date written on it, and a working time established by measurement rather than by habit.

The prevention. The lamp carries a label giving the paper, the distance, the bulb and the safe working time, and the filter has a replacement date on it. The label is on the lamp rather than in a notebook because the next person to wonder, including you in eighteen months, will look at the lamp.

Case 3 — surge marks, from over-agitation applied to fix the wrong fault

The diagnosis. Solution driven through the perforations by excessive agitation, producing extra development beside each hole.

The evidence. The marks are at both long edges, they are periodic, and the period matches the perforation pitch on the same strip. Nothing else in the processing chain repeats at an interval set by the film’s own geometry. Measure it rather than assume it.

The story in the record is the useful part. The agitation was doubled deliberately, to cure mottle on the previous roll. Mottle is the under-agitation failure; surge is the over-agitation failure; they are two ends of one lever. The printer moved from one end to the other in a single step and arrived at the opposite fault. The correct move was to the middle: the published scheme.

The confirming test. One roll at the manufacturer’s published scheme — ILFORD’s four inversions in the first ten seconds of each minute, or Kodak’s five to seven inversion cycles in five seconds every thirty — with everything else held. If the marks go and the mottle does not return, the scheme was the fault.

⁇What the course can and cannot cite for this identification

Kodak’s troubleshooting table documents the category: “streaks of non-uniform density” against excessive or uneven developer agitation, and “mottle — areas of non-uniform density” against inadequate agitation. That is a manufacturer’s statement and it is what licenses the two-ended-lever argument.

The course has not found a manufacturer publication naming the sprocket-hole surge mark specifically. The flow-through-the-perforations account is presented as mechanism — it follows from the transport argument and from the geometry — rather than as a sourced identification, and the atlas entry records the same limitation. If you find a Tier 1 source that names it, that entry is where the correction belongs.

The fix and the prevention. Adopt one published scheme, write it into the processing table as inversions and intervals, and change one thing at a time when a fault appears. Note that changing the agitation also changes the effective development, so the contrast index you had established is no longer yours until you re-establish it.

Case 4 — over-fixing, at the film time on paper

The diagnosis. The print’s deepest tones are being etched by four minutes in a fixer whose published time for fibre paper at that dilution is one minute.

The evidence, by elimination. The developer is fresh, at temperature, at the published dilution, and was given two minutes — inside the 1½ to 3 minutes ILFORD publish, and comfortably inside the six minutes they say makes no noticeable difference. The paper is new and in date, so ageing is out. The base is clean and the safelight is tested and labelled, so fog is out. Four sheets through a litre is nowhere near the 50 fibre sheets ILFORD’s capacity table allows, so exhaustion is out. That leaves the steps after development, and the record has a number in it that is four times the published one.

The manufacturer’s own word for the mechanism is in their sheet: there is no benefit in extending fixation beyond the recommended time, and “some loss of print quality might be seen when long fixing times are given due to image etching”. The deepest tones lose density because a fixer, given enough time, dissolves image silver as well as halide — and the print’s smallest deposits go first.

The confirming test, and it is one sheet. Make two identical prints. Fix one for the published minute and one for four minutes, wash and dry both identically, and compare the deepest step side by side. This is the cleanest experiment on the page: one variable, one control, one comparison.

The fix. Use the paper time. ILFORD RAPID FIXER at 1+4: 30 seconds for RC, 1 minute for fibre, against 2 to 5 minutes for film.

The prevention. Write the fixing time on the fixer bottle’s label, per material, and note that the film time and the paper time are different numbers rather than a range. The error here came from caution, which is why “be careful” is not a prevention: the printer was being careful, in the wrong direction.

Case 5 — under-fixing, with a washing failure that cannot now be separated from it

The diagnosis, as far as the evidence carries. The material was under-fixed. The violet, milky cast in the rebates at the time was retained halide and retained sensitising dye; the yellow-brown stain six weeks later is residual chemistry converting to silver sulfide.

The evidence.

• A milky, violet cast in the rebates is the classic retained-halide appearance, and Kodak’s own table pairs “milkiness, most apparent in low-density areas” with “retained silver halide due to inadequate fixing”.
• The bath is eleven weeks old with about twenty films through it and no clearing-time test ever run. ILFORD’s capacity for RAPID FIXER at 1+4 is 24 films of 135-36 per litre. Twenty is at the edge of it, and “about twenty” without a running total is not a measurement.
• The stain arrived weeks later and is worst in the clear areas — where there is least image silver and most of whatever the fixer left.
• The neighbouring strip in the sleeve has begun to yellow at the contact edge. That is migration, and it is evidence about residues rather than about the image.

!Where the evidence stops, and why the page says so

Two causes remain live and this session cannot separate them.

The stain is consistent with residual silver — complexes the tired bath never carried far enough to be washable, which is a fixing failure — and with residual thiosulfate — fixer that the shortened wash left behind, which is a washing failure. The record contains a defect at both steps: a bath at the edge of its capacity with no test, and a wash cut from three changes to two.

Part XII’s two tests would have separated them at the time, and their two-by-two says which. They cannot do it cleanly now: the residues have been converting to silver sulfide for six weeks, so a negative result no longer means the residue was absent, only that it is no longer in the form the test detects.

So the honest result is two live hypotheses and one instruction: run both tests on the next session’s material, before filing anything, and find out which of the two defects in the record is actually operating in your darkroom. That is a real result, not a failure. A diagnosis that names one cause here would be a preference dressed as a finding, and it would send you to fix the wrong step.

Note also which way the ambiguity leans in practice: an overworked fixer produces a washing fault, by loading the material with complexes a wash struggles to shift. So when both are in the record, the fixer is fixed first and the wash re-measured afterwards.

The fix. Nothing for these negatives. Silver sulfide is not silver, and the stain is the image now.

The prevention, as three procedural changes. Measure the clearing time of every fixer bath when it is fresh and write it on the bottle, because that number cannot be recovered later. Keep a running total on the bath and discard at the published capacity or at twice the fresh clearing time, whichever comes first. And do not truncate a wash because it is late — a wash cut short costs a negative that a reshoot cannot replace.

The fix, as a table

Every fix on this page is a change to a procedure. None of them is a resolution.

Case	Fault	The procedural change
1	Two stops of underexposure from an unreset film-speed setting	“Check the speed setting” enters the loading routine, beside checking the frame counter
2	Safelight fog from bulb, distance and filter age together	Lamp labelled with paper, distance, bulb and safe working time; filter carries its installation date
3	Surge marks from over-agitation applied to cure mottle	One published agitation scheme written into the processing table; one change at a time thereafter
4	Over-fixing at the film time on paper	Fixing times per material written on the fixer bottle
5	Under-fixing, with a washing failure not separable after the event	Fresh clearing time measured and recorded per bath; running total kept; the wash is not shortened

Reproducing two faults deliberately

Now make two of these faults on purpose, under control, so that you have a physical reference.

Specimen A — safelight fog, at a stated distance and time

The geometry is Part XVI’s test, and the series is its series extended past the safe end. The point is not to establish a safe time — that is Part XVI’s job — but to produce a sheet on which fog is unmistakably visible, so that you know what your own paper looks like when it is fogged by your own lamp.

1. Set the published condition. Safelight filter at the maker’s recommended type, 15 W bulb, lamp 1.2 m from the bench, everything else in the room off, enlarger off, door shut. Give your eyes fifteen minutes to adapt and look for leaks while you wait.

2. Find the mid-grey exposure. In total darkness, with the enlarger and no negative, make a short test strip and find the exposure that gives a pale grey of roughly 0.2 to 0.3 density. This is Part XVI’s stage 1 and the number carries over.

3. Expose the “after” patch. On a fresh sheet, in total darkness, give that mid-grey exposure to a patch at one end, and notch the sheet’s corner so you know which way round it is. Paper is more sensitive to safelight fogging after it carries a latent image, so this is the critical half.

4. Run the safelight series past the safe end. Carry the sheet to the brightest place in the room, lay it emulsion up, and uncover successive bands with a masking card so that they receive cumulative times of 0, 4, 8 and 16 minutes.

   The four-minute step is ILFORD’s own stated limit — an SL1 or 902 at 15 W and 1.2 m “should be safe for up to 4 minutes” — so the series starts at the published boundary and runs to four times it. If even the 16-minute band shows nothing, double again (0, 16, 32, 64) on a second sheet and record what it took. The specimen is made when a band shows, and the time it took is the datum.

5. Process it with a control. A second sheet, given the identical mid-grey patch and kept in a light-tight envelope in the same room throughout, developed in the same dish at the same time. Every difference between the control patch and the test sheet’s zero-minute band is a processing difference; every difference along the test sheet is safelight.

6. Label it and keep it. Paper, safelight filter and its installation date, bulb wattage, distance, the four cumulative times, the date, and the mid-grey exposure used.

Specimen B — an under-fixed print, kept

The parameters. ILFORD RAPID FIXER at 1+4 gives fibre paper 1 minute and RC paper 30 seconds. Make a series at a quarter, a half and the full time, so that the specimen has its own control built in.

1. Make three identical prints from one negative — a print with a clean white, a real black and a large light area, because that is where the fault will show.
2. Develop, stop and fix all three identically except the fixing time: 15 seconds, 30 seconds and 60 seconds on fibre (or 8, 15 and 30 seconds on RC).
3. Wash all three fully and identically. This matters: the specimen is meant to isolate a fixing failure, and a shortened wash would add a second variable and reproduce Case 5’s ambiguity on purpose.
4. Confirm with Part XII’s residual-silver test. One drop of the diluted sodium sulfide on a white area of each, excess blotted at once, compared against the reference spot. The 60-second print is the reference the other two are read against.
5. Dry all three and keep them together, labelled.

⚠This specimen keeps changing, and it must be stored as one

The 15-second print carries residual silver and residual thiosulfate by design. It is not stable and it is not meant to be: watching it change over a year is half its value as a reference.

Store it in its own sleeve, in its own envelope, away from finished work. Residual thiosulfate can migrate to prints stored in contact with it — Case 5’s neighbouring strip is exactly that happening by accident — so a deliberately under-fixed print filed with your prints is a slow accident waiting.

Label the envelope with what it is, why it is under-fixed, the fixing time, the date, and a line saying that it is a reference specimen and not a print. Somebody else will open that box one day.

Photograph it or scan it at intervals if you want a record of the change, since the print itself will not stay comparable to its own earlier state.

Why the specimens matter more than photographs of faults

The course holds no plates of these defects, and the atlas says so on its front page. This session is the answer to that, and it is a better answer rather than a substitute for a worse one.

A photograph of somebody else’s fault answers the wrong question. What you want to know at the bench is not what does safelight fog look like but does my paper, under my safelight, look like this? A plate cannot tell you, because it was made on a different paper under a different lamp and reproduced through a printing process that changed it again.

A specimen you made is calibrated. You know the distance, the wattage, the filter’s age and the exposure time, because you set them. So the specimen does not just show you a fault; it tells you what conditions produced it, on your material, which is what you need in order to recognise a smaller version of it later.

And it lives in the right place. A reference belongs in the darkroom, in a labelled envelope beside the enlarger, not in a book on a shelf in another room. The test of a reference is whether you actually pick it up when you are wondering, and a sheet in a drawer six inches from the dish passes that test in a way that a page does not.

Writing the atlas entry

The reference contribution of this session is one troubleshooting atlas entry, written to the course’s format, about a defect you produced or diagnosed yourself.

The format is fixed and it is the same six sections every published entry carries:

Section	What goes in it
What you see	The appearance, in words precise enough that somebody can match it against their own material. No photograph, and no describing one.
Likely causes	In the order they turn out to be the answer, not the order they occur in the process.
The chemistry and physics	The mechanism, from a source, with the source named. If you cannot source it, say so in the entry rather than inventing it.
Diagnostic questions	The observations that separate this defect from the ones it resembles. An entry with no discriminator here is a description.
Corrective action	What to do with the material in front of you — including, often, that nothing can be done, and why.
Prevention	A change to a procedure. Not a resolution.

Three rules that make the difference between an entry and a note.

Name your discriminator. The best entries in the atlas are built round one observation that settles it: a refix that fails, a pitch that matches the perforations, a rebate that is normal, a texture that is present in the rebate. Find yours and put it in a callout.

Cite your own specimen as evidence, with its parameters. “On a 15-second fixation of MULTIGRADE FB at 1+4, washed fully, the residual-silver test spot read distinctly yellow against a 60-second control” is evidence. “Under-fixed prints go yellow” is a rumour.

Say what you could not establish. An entry that names the limit of its own evidence is worth more than one that papers over it — Case 5 above is the model, and so is the surge-mark callout.

What to record

Permanently, in a form you will find again:

1. The case — the symptom, in the words you would use to search for it.
2. The evidence — the full list, including the items that showed nothing, because a negative observation is what eliminates a cause.
3. What each observation eliminated, in order.
4. The diagnosis, and where two causes remained live, both of them.
5. The confirming test, and its result.
6. The procedural change, written as a change to a routine rather than an intention.

That record is what makes the second occurrence a ten-minute job instead of an evening. It is also what makes a pattern visible: three separate faults over a year that all trace to the same unmeasured bath is a finding that no single session can produce.

And the closing rule of this part:

A defect that has been explained is a lesson. A defect that has been explained and prevented is a process.

The sixth case, given blind

No hints, no diagnosis, and no answer on this page. Work it in writing, show your reasoning, and record what each observation eliminates.

What you have. A 120 roll of FP4 Plus, six frames. Four of the six show a broad, soft-edged pale streak running away from the large bright sky area at the top of each frame, down across the darker landscape below it. The streaks have no sharp boundary. On every affected frame the streak runs the same way relative to the film — from the perforated edge inward — even though two of those frames were taken with the camera turned on its side, so that the sky is at the left of the picture rather than the top. The two unaffected frames are the two with no large bright area in them at all. The rebates are normal and the edge numbers are crisp.

Session record.

Film	FP4 Plus, 120, in date
Developer	ID-11 1+3, 20 °C, 20 minutes
Agitation	Four inversions at the start, then one gentle inversion every five minutes — “a semi-stand scheme from a video”
Tank	Two-reel plastic tank, one film, 500 mL
Stop, fix, wash	Standard, all baths at 20 °C
Previous rolls	Five rolls in the same tank at the standard scheme, all clean

Three questions to answer in writing before you look anything up.

1. Which of the six questions in the questions in order section does the sentence about the camera being turned on its side answer, and what does it eliminate?
2. Why do the two frames without a large bright area show nothing?
3. What single change to the record would you make first, and what would you expect it to cost you?

Recap

Five cases, worked from evidence rather than from a hunch, each with a confirming test. A thin flat film with crisp edge numbers is underexposed and not under-developed, and it is flat because the subject slid onto the toe. A grey border was fogged by light that did not come through the lens, and the severity tracking time under the lamp identifies which light. A period that matches the perforations is a surge mark, and it appeared because a fault at one end of the agitation lever was cured by jumping to the other end. A paper that stops climbing below black, with everything else eliminated, was fixed for four times the paper’s published time. And a violet cast with a late yellow stain leaves two causes live that this session cannot separate, which is a real result and comes with an instruction rather than an answer. Then two specimens made on purpose, labelled and kept, because a calibrated fault on your own material answers a question no photograph can. And one atlas entry, in the course’s own six sections, with your discriminator named and your evidence cited.

Check your understanding

Question 1. In Case 1, why do crisp edge numbers and a normal rebate eliminate the developer, when the negatives are visibly thin?

• Because the rebate is thicker film and develops faster
• Because edge numbering is a latent image put on the film at the factory, so nothing but development can bring it up and nothing the camera does can affect it; crisp numbers at normal density therefore say the developer worked at the right strength for long enough, whatever the pictures look like
• Because the rebate is fixed rather than developed
• Because edge numbers are printed in ink after processing

Show the answer and why

Answer: Because edge numbering is a latent image put on the film at the factory, so nothing but development can bring it up and nothing the camera does can affect it; crisp numbers at normal density therefore say the developer worked at the right strength for long enough, whatever the pictures look like

This is the most useful piece of free evidence on a roll of 35 mm film, and it is routinely ignored. Two things on every film were exposed by somebody other than you — the factory's edge printing, and the leader, which met full room light when you loaded the camera. Both should be at full density after correct development, and both are entirely independent of the exposure you gave. A grey leader or faint edge numbers convict the developer or the fixer in one look; crisp ones exonerate them just as firmly.

Question 2. Case 3's printer doubled the agitation to cure mottle on the previous roll and produced surge marks. What is the general lesson, in terms of the diagnostic method rather than the chemistry?

• That mottle should be ignored
• That agitation should always be minimal
• That when one fault is the failure mode at one end of a control and another is the failure mode at the other end, a correction that jumps from one end to the other will land on the opposite fault; the move is to the published middle, and then one change at a time with a controlled comparison
• That the tank was the wrong type

Show the answer and why

Answer: That when one fault is the failure mode at one end of a control and another is the failure mode at the other end, a correction that jumps from one end to the other will land on the opposite fault; the move is to the published middle, and then one change at a time with a controlled comparison

The printer's reasoning was not stupid — mottle is under-agitation and more agitation is the right direction — but the size of the step was uncontrolled, and there was no published reference to move to. Kodak's own table names both ends explicitly, which is what makes the lever visible. The method lesson generalises well beyond agitation: whenever a fault has an opposite, ask what the published setting is before deciding how far to move, and change nothing else at the same time so that the result is readable.

Question 3. Case 5 ends with two live causes rather than a diagnosis. Why is that reported as a result rather than as a failure of the session?

• Because the negatives are lost either way, so the cause does not matter
• Because the evidence genuinely cannot separate a fixing failure from a washing failure six weeks after the residues began converting to silver sulfide — the record contains a defect at both steps, and naming one would be a preference dressed as a finding that would send the reader to fix the wrong step; the useful output is an instruction to run both tests on the next session's material, before filing anything
• Because the tests are unreliable
• Because two causes are always possible in photography

Show the answer and why

Answer: Because the evidence genuinely cannot separate a fixing failure from a washing failure six weeks after the residues began converting to silver sulfide — the record contains a defect at both steps, and naming one would be a preference dressed as a finding that would send the reader to fix the wrong step; the useful output is an instruction to run both tests on the next session's material, before filing anything

The distinction matters because the preventions differ: a fixing failure is fixed by a clearing-time test and a capacity log, and a washing failure by a longer or better-designed wash. Guessing sends you to one of those and leaves the other running. Note also the asymmetry the callout draws out — an overworked fixer produces a washing fault by loading the material with complexes a wash cannot shift — which is why, when both defects are in the record, the fixer is corrected first and the wash re-measured afterwards.

Question 4. Why does the safelight specimen use a series of 0, 4, 8 and 16 minutes rather than Part XVI's 0, 1, 2 and 4?

• Because longer exposures are more accurate
• Because the two exercises have opposite aims. Part XVI's series brackets the threshold in order to establish a safe working time, so it sits around the published four-minute limit. This specimen is meant to produce fog that is unmistakably visible, so it starts at that limit and runs to four times it — and if nothing shows, the instruction is to double again and record what it took, which makes the specimen a measurement rather than a recipe.
• Because paper is less sensitive when it is fresh
• Because the enlarger is off for this test

Show the answer and why

Answer: Because the two exercises have opposite aims. Part XVI's series brackets the threshold in order to establish a safe working time, so it sits around the published four-minute limit. This specimen is meant to produce fog that is unmistakably visible, so it starts at that limit and runs to four times it — and if nothing shows, the instruction is to double again and record what it took, which makes the specimen a measurement rather than a recipe.

It is worth noticing that the series is derived from the published condition rather than invented: ILFORD state that an SL1 or 902 at 15 W and 1.2 m should be safe for up to four minutes, so four minutes is the boundary the series starts at, and every other number is a doubling of it. That also means the specimen carries information even when it "fails" — a lamp that shows nothing at 16 minutes is a better lamp than the recommendation assumes, and the time it eventually takes is a number worth writing on the label.

Question 5. What does Kodak's Z-133 require before acting on an out-of-control process, and after every corrective action?

• A written incident report before, and a management review after
• Before: verify the out-of-control condition — check that the densitometer is working and the calculations were right, and process a second control strip to confirm it. After: process another control strip to confirm that the change has returned the process to control, before resuming normal work.
• Nothing before; a full re-calibration after
• A supplier notification in both cases

Show the answer and why

Answer: Before: verify the out-of-control condition — check that the densitometer is working and the calculations were right, and process a second control strip to confirm it. After: process another control strip to confirm that the change has returned the process to control, before resuming normal work.

Verify before, verify after — and both halves are routinely skipped by people who are otherwise careful. Acting on an unverified fault means you may be correcting an instrument error or an arithmetic slip; resuming work on an unverified correction means you find out whether it worked on the material you cared about. The home equivalents are cheap: a second strip through the suspect bath before you discard it, and a clip test or a control strip after you change anything.

Check my answersClear