The Pure Silver Portfolio: The Specification

Nobody is going to mark this. That single fact governs everything below, because a criterion nobody can apply is not a standard — it is a wish with a number in it.

So the capstone specification is written to one rule: every requirement is something you can check against your own materials, or it is not a requirement. A density you can read on your own instrument. A wash test you can run at a sink. A clearing time you can measure with a stopwatch. A second print you can lay beside the first. Where a criterion cannot be made checkable — and four of them cannot, for reasons this page states rather than hides — the specification says so, and gives the judgement you have to make in its place.

How this page is used

Read it once now, at planning time, and again at the end when you fill in the rubric. In between it is a reference rather than a lesson: each requirement below states what satisfies it, what evidence proves it, and which capstone page produces that evidence.

The eleven requirements, the evidence each needs, and where it is produced

1. 11 — Build or substantially modifyEvidence: a written performance goal with a number, a before measurement, an after measurement, and an instrument certificate. Produced by Stage 1.
2. 22 — Calculate pinhole characteristicsEvidence: the complete calculation sheet, checked against a hole you measured. Produced by Stage 1.
3. 33 — Expose original negativesEvidence: the exposure log, one entry per sheet or frame, written at the moment of exposure. Produced by Stage 3.
4. 44 — Formulate chemistryEvidence: the mixing record with nominal and actual masses, and a version identifier. Produced by Stage 2.
5. 55 — Process in that chemistryEvidence: the run record naming the version identifier that processed each negative. Produced by Stage 3.
6. 66 — Document the formulationEvidence: a formulary-standard page for your solution, including the function of every ingredient. Produced by Stage 2.
7. 77 — Measure the processEvidence: a control-strip curve, contrast index, base-plus-fog, exposure index, clearing time with capacity used, residual-thiosulfate result — each with an uncertainty. Produced by Stage 3.
8. 88 — PrintEvidence: the finished prints, their printing maps, and one print remade from its map alone. Produced by Stage 4.
9. 99 — Tone one imageEvidence: the toned print, its matched untoned control, and the matched-patch densities before and after. Produced by Stage 5.
10. 1010 — Cross-print in an alternative processEvidence: the print, its coating and exposure record, and the negative route declared. Produced by Stage 5.
11. 1111 — Evaluate permanenceEvidence: residual-chemistry results on the actual portfolio materials, the storage plan, and a permanence statement whose four parts stay separate. Produced by Stage 6.

Eleven requirements, eleven pieces of evidence, six stages. Nothing here is satisfied by an assertion; every row names an object or a document a reader could pick up.

Requirement 1 — Build or substantially modify equipment

What satisfies it. A new instrument, or a substantial modification to one you own. “Substantially” is a test rather than a feeling, and it has three parts:

1. A written performance goal with a number in it, dated before the work starts.
2. A measurement of the starting state, made with an instrument whose uncertainty you can quote.
3. A measurement of the finished state, by the same method, and a stated verdict on whether the goal was met.

The test that makes it real is the fourth clause: the result must have been capable of failing. Buying an enlarger and levelling it is not a modification, because there was no goal it could have missed. Converting the same enlarger to an LED head and demonstrating that opposite corners of the field now agree within 0.04 in density where they did not before is a modification, because the measurement could have come out the other way.

!Three worked judgements on the boundary

“I recalibrated my bought densitometer against a certified step tablet.” This qualifies if it was a calibration with a stated before and after — residuals across all twenty-one steps, the scale constant that changed, and the certificate line updated. It does not qualify as a bare re-zeroing, because nothing could have failed.

“I replaced my enlarger’s lamp with a brighter one.” This does not qualify on its own. It qualifies if the goal was stated as a number — an exposure time brought under a stated ceiling, or a uniformity figure held while the output rose — and both states were measured.

“I built a drying box out of a plastic storage crate.” This qualifies if it had a goal you could miss: a drying time, or a measured reduction in airborne dust settling on a test sheet. It does not qualify as “it dries plates now”, because that was never in doubt.

The permitted menu is on Stage 1 and everything on it is an instrument this course has already taught. You may not invent a new class of apparatus for the capstone; the requirement is about engineering discipline, not novelty.

Requirement 2 — Calculate pinhole characteristics

Mandatory for every capstone, whatever instrument you built. A student whose Stage 1 project is an LED safelight still does this calculation, on a pinhole they make and measure for the purpose if they own no pinhole camera. The reason is that it is the course’s shortest complete example of optics predicting a picture: seven quantities, all derived from three measurements, all checkable against a photograph.

The required set, each shown with its working rather than only its answer:

Quantity	What it is checked against
The chosen focal distance	Measured, not nominal — the plate-to-hole distance you actually have
The optimum diameter for that focal distance	Under the constant Part VI adopts, with the wavelength stated
The diameter actually made, and how it was measured	Your pinhole register from Part VI, with the measurement’s own precision
The effective f-number that follows	Consistency with the exposures the camera actually needs
Geometric blur and the diffraction limit at that diameter, and the point where they cross	Whether your hole is on the sharp side or the diffracted side of the crossing
Angle of view and coverage on the format in use	The negative: does the image circle reach the corners
Illumination falloff at the corner	Your own corner density, and the excess over the cosine-fourth prediction

📓One number the course refuses to give you

Part VI states that real pinhole falloff is generally worse than cosine-fourth because of the tunnel effect of a plate with thickness, and that it could not find a published measurement of the excess to put a figure on it. So the specification asks you to compute the cosine-fourth prediction, measure your own corner, and record the difference in stops with your plate thickness beside it. It does not ask you to meet a falloff limit, because no limit exists that this course could stand behind. That distinction — a measurement required, a threshold refused — recurs throughout this page and is the honest shape of a specification built on a corpus with gaps in it.

Requirements 3 to 5 — Expose, formulate, process

These three are one chain and the specification treats them as one.

Requirement 3, an original negative, means all four of: exposed by you during the capstone; on material you have characterised, in the sense of Part XV’s assignment; with a completed exposure-log entry written at the moment of exposure; and processed in the chemistry of requirement 4. Archive negatives from earlier parts may be printed and may appear among the finished works. They do not satisfy requirement 3, because they were made before the instrument and the chemistry existed.

Requirement 4, formulated chemistry, means at least one working solution mixed from individually weighed raw chemicals. A kit does not satisfy it. A concentrate diluted does not satisfy it, however carefully. A formula copied without understanding does not satisfy it either, and the evidence for understanding is requirement 6.

The reason for the strictness is worth stating, because it looks like purism and is not. The capstone is testing whether you can predict what a change will do. That prediction is only meaningful about a composition you know, and you only know a composition you weighed. It is also why one solution is enough: the discipline is in the knowing, not in the number of bottles.

Requirement 5 is then almost automatic: the negatives of requirement 3 are processed in the solution of requirement 4, and the run record names the formula version identifier that did it. If you remix the developer mid-project, the identifier changes and every subsequent negative traces to the new one. A project in which some negatives cannot be traced to a version has failed requirement 5 even though every bath was mixed correctly.

✕The commonest way this chain breaks

Formulating the developer after the shooting. It seems efficient — expose while the weather holds, mix later — and it is the single most expensive mistake available in this project. If the formulated developer turns out to give a contrast index you cannot print, the negatives are already made and the only remedy is to shoot them again. The plan’s sequencing rule exists for this: the chemistry is formulated and verified before the first exposure of the campaign.

Requirements 6 and 7 — Document and measure

Requirement 6, the documentation standard, is the formulary page template applied to your own solution. Identity and provenance; purpose; the ingredient table with quantity, form and function; mixing order with the reason the order is what it is; working dilutions; behaviour and the conditions it was measured under; image characteristics as observed; storage and the shelf life you can evidence; incompatibilities; waste route. And the section that carries the weight: the function of every ingredient, written for every one of them rather than for the interesting ones. An ingredient present for a reason that has nothing to do with reducing silver is exactly the one whose paragraph proves you understood the formula.

Requirement 7, the measurement standard, names six numbers. Each is quoted with its uncertainty and each traces to an instrument with a certificate.

Number	The construction it is quoted under	Where the uncertainty comes from
A characteristic curve from a control strip	Plotted from your own densitometer readings, one point per step	The densitometer certificate, by density band — it is not one figure for the instrument
Contrast index	The course’s convention, never a bare “CI”	Propagated from the density and log-exposure terms, which are never added together
Base plus fog	The unexposed strip processed alongside, not the film’s clear edge	The densitometer’s repeatability in the low band
Effective exposure index	The course’s speed criterion, modelled on ISO 6	The sensitometer’s log-exposure uncertainty, plus the density term divided by the local gradient
Clearing time, with the capacity used	The Part XI fixing criterion	A stopwatch and an eye, and the page says what that resolution is
A residual-thiosulfate result	The Part XII washing criterion	Visual comparison against a control half of the same sheet

!How these numbers are written, and how they are not

Three forms of words are fixed by earlier parts and this specification requires them.

“EI 320, measured under the criterion defined in Part XIII, which is modelled on ISO 6, on emulsion batch NNN in [solution version] at 20 °C for N minutes, read on instrument serial NNN under its certificate of [date].” Never “the ISO speed is 320”. The course does not hold ISO 6’s normative text and makes no conformance claim.

“Contrast index 0.62, measured under the criterion defined in Part XIII.” Never a bare “CI 0.62”, because three constructions in common use give three different numbers on the same curve.

“Dmax 2.08 relative to tile serial NNNN.” Every reflection density the course’s own densitometer produces is relative to your white reference tile, and the day somebody compares it with a published figure is the day the omission costs something.

On reflection density, which three of the capstone’s activities need. The Part XV densitometer does have a reflection capability, and Part XV fixes the procedure for using it — the paper-anchor convention: one sheet with a base-white area and a maximum-black area, processed to the maker’s own recommendation, dried flat in air, read 24 hours after it is dry to the touch, at five places on each area, mean and spread recorded, against the tile and the trap. The specification requires that convention wherever a print density is quoted, and requires the tile serial with the number. A reader who has not built the reflection attachment records the print densities as outstanding rather than estimating them, and Stage 4 gives the visual-comparison route that stands in for them, with what it cannot do stated.

Requirements 8 to 10 — Print, tone, cross-print

Six to ten finished works. Among them, at least one contact print, at least one enlarged silver print, at least one toned print, and at least one made in an alternative process.

The overlap rule. The four categories may overlap: a toned enlargement satisfies two of them at once, and a toned Van Dyke would satisfy two more. What the specification forbids is satisfying all four with fewer than four separate works. A portfolio in which one much-treated print carries every category is not a body of work; it is a demonstration piece with a set around it.

A minimum-compliant six-work portfolio, and what each work carries

1. Work 1 — contact printfrom an original negative, in the frame or under glass and a weightcategory: contact
2. Work 2 — enlargementwith a printing map a stranger could followcategory: enlargement
3. Work 3 — enlargement, tonedwith its matched untoned control retained as evidence, not exhibitedcategories: enlargement + toned
4. Work 4 — alternative processwith its coating and exposure record and the negative route declaredcategory: alternative
5. Work 5 — enlargementthe one remade from its map alone in a later session
6. Work 6 — contact printthe second of a pair that argues something the singles could not

Four separate works carry the four categories; works 5 and 6 exist because the set needs them. Note that work 3's untoned control is documentation rather than a seventh work — it goes in the documentation set, not on the wall.

What “finished” means, stated as five tests you apply to each object:

• Dry, and dry long enough to be judged: a fibre print read the day after it is dry to the touch, not the evening it left the wash.
• Flat, by the method Part XIX specifies, and flat enough to sit under a mat without buckling.
• Spotted and trimmed, with the spotting medium recorded.
• Mounted or sleeved to a stated standard, and the standard is stated by naming what the material’s own packet claims rather than by the word “archival”.
• Repeatable from its printing map — negative, paper and batch, developer and its age, filtration, aperture, magnification, base exposure, every local move with its area and its increment in stops, and a version number.

That last one is the requirement with teeth, and Stage 4 makes you prove it on at least one work by printing it again from the map alone in a later session and writing down where the two copies differ.

Requirement 11 — Evaluate permanence

The evidence rule here is the strictest on the page, and it is a rule about sentences rather than about procedures.

Every permanence claim is classified as measured, published or inferred, and the three are not allowed to blur. Part XX’s four-part statement is the template and the capstone requires it unchanged: what was done; what published procedure it follows, named by publisher and document rather than by the word “archival”; what was tested and against what criterion; and what is inference, labelled as such.

A claim with no measurement behind it is written as an inference and never as a fact. That is the whole of requirement 11 in one line, and it is checkable by reading your own paragraph with a pencil.

What you can honestly measure is bounded, and the boundaries are the course’s rather than your apparatus’s:

• The residual-thiosulfate test tells you the wash worked, by the Part XII criterion: no colour difference between a strip of the control sheet immersed about three minutes in 1 per cent silver nitrate and the untreated part of the same sheet, compared wet in subdued light. It gives a pass or a fail against a control, not a figure in micrograms per square centimetre.
• The residual-silver test tells you the fixing worked, by ILFORD’s sulfide test read against a reference spot on a print known to be well fixed. Same shape: a comparison, not a quantity.
• Both are named by the reagent you actually mixed. ILFORD’s sulfide test is the composition Kodak publishes as ST-1, so naming ST-1 for it is accurate. A plain 1 per cent silver nitrate immersion is not HT-2, which is acidified and read against a card the course does not hold. Stage 6 sets out the rule; the failure it prevents is the first of the four below, an unrecorded number wearing a borrowed name.
• Neither certifies permanence. They detect failure, not degree, and both are visual comparisons whose resolution is your eye’s.
• Protective toning is a manufacturer’s recommendation, not a measured outcome for your print. Kodak state that toning converts the silver image to an inert compound and that all their toners protect it; the standard written to measure the effectiveness of such conversion, ISO 18915, explicitly does not recommend treatments and puts times and temperatures outside its scope. So the honest sentence names the manufacturer’s claim as theirs, states that the degree of conversion in your print was not measured, and stops.
• An enclosure or mount claim is a claim about the packet. IPI’s own guidance is that a board meeting ISO 18902 has a pH between 7.0 and 9.5, at least 2 per cent calcium carbonate buffering and no lignin — and that none of those requirements on its own implies sufficient preservation quality. Write what the packet says; do not upgrade it.

The documentation set

Thirteen artefacts. The format column is what a reader would need to be handed, and where the course already publishes a sheet for it, that sheet is what you use rather than a form of your own.

Artefact	Format	Produced on
Project plan	One page: subject, instrument, chemistry, materials, schedule, budget, contingency	Planning
Whole-project hazard assessment	One row per substance: GHS classification from the current sheet with its version and date, quantity handled, control, incompatibilities, storage	Planning
Instrument certificate and design record	Part XV’s certificate format, plus the dimensioned sketch, the bill of materials and the changes made during the build	Stage 1
Formula record with a version identifier	The formula version record sheets, coded by the versioning SOP	Stage 2
Formulary-standard page for your solution	The section set of a formulary entry, including the function of every ingredient	Stage 2
Exposure logs	The exposure log sheets, or the browser exposure log exported	Stage 3
Processing records	The lab notebook sheets, one per run, naming the solution version	Stage 3
Control-strip data and curves	The curve-plotting sheets, plus the raw readings as run, strip, step, log_H, density, sd, n	Stage 3
Printing maps	One per finished print, versioned	Stage 4
Toning and alternative-process session records	The toning session record and the coating and exposure record from Parts XX to XXV	Stage 5
Residual-chemistry results	One row per sample, by the tests SOP	Stages 3 and 6
Permanence and portfolio statements	Prose, to the templates named above	Stage 6
The chemistry report	Its own page’s structure, thirteen sections	The report

The rubric

Four criteria, four bands each. A band is claimed against a named piece of evidence in the documentation set — an artefact and a page number, not a feeling. A band claimed with no named evidence scores as the band below it, and that rule is what makes self-assessment worth doing.

Accuracy of the chemistry and the numbers

Band	What it looks like
4	Every number carries an uncertainty traceable to a certificate; the two axes are never added together; the constructions are named; a disagreement with a manufacturer’s figure is discussed rather than averaged away
3	Numbers are traceable and constructions named, but the uncertainty treatment is partial — a single instrument figure quoted where the certificate gives bands
2	The measurements were made and are correct, but are quoted bare: no uncertainty, or a construction not named
1	Numbers appear whose origin cannot be established from the documentation

Craft of the objects

Band	What it looks like
4	The set reads as one hand on one day; every work meets the five finishing tests; the remade-from-map print differs from its original only in ways you can explain
3	All works finished to the standard, with one or two visibly from a different session
2	Works are finished but inconsistent in surface, tone or presentation, and the inconsistency is not accounted for
1	Works are unfinished by the specification’s tests — damp-judged, unspotted, or not repeatable from their maps

Coherence of the body of work

Band	What it looks like
4	The set argues something the individual pictures do not; the sequence has a reason; at least one good print was removed because it did not serve the set, and the documentation says which and why
3	The set is coherent and sequenced, but nothing was removed — everything printed was included
2	The works share a subject but not an argument; the order is arbitrary
1	A collection of separate successes

Honesty of the documentation

Band	What it looks like
4	Every claim is classified; gaps are named at the point they bite; substitutions are declared with what each cost; at least one sentence states something you do not know
3	Claims are classified and substitutions declared, but the gaps are collected in a closing paragraph rather than named where they matter
2	The record is complete and accurate but makes no distinction between what was measured and what was inferred
1	A claim is presented as measured that was not measured

The failure conditions

These are not bands. Any one of them means the requirement it touches is not met, whatever the rest of the project looks like, and all four are checkable by you in an afternoon.

• An unrecorded number. A figure appears in the report or on a label that no notebook entry, log line or reading file supports.
• A formula altered without saying so. A published formula reproduced with a quantity changed and no variant declaration. Historical formulas are never silently altered; that is a rule of the course and not a preference.
• A permanence claim presented as measured when it was inferred.
• A work that cannot be reprinted from its own map. Demonstrated by trying, not by inspection.

The safety boundary

Nothing in the capstone is above Level B. This is a rule of the specification, and three consequences follow that a project plan has to respect.

No procedure requiring a fume cupboard. If the verified classification of a reagent makes local exhaust the recognised control, that route is Level C and is outside the capstone. Stage 5 names which toners and which alternative processes survive that test against their own safety data sheets rather than by reputation.

No licensed waste route beyond ordinary photographic and silver-bearing collection. Spent developer, stop and fixer are collected by stream, the fixer retained for silver recovery, and selenium and sulfide toner waste collected in its own labelled container. A process whose waste needs a route beyond that is outside the specification. Local regulation governs disposal in every jurisdiction, and the disposal page explains why the course can describe chemistry and general practice and cannot give you a permission.

No work on the mains side of anything. The course’s line is the one Part XVII states: everything you build stays at extra-low voltage, and anything at mains potential is a bought, certified, sealed appliance, plugged in, used as its maker directs, and never opened.

!A correction to an assumption this specification was drafted on

An earlier draft of this specification said that a certified enclosed switching module was the permitted route to controlling a mains lamp from an instrument you built. That product could not be found. Part XVII’s market survey of 5 September 2026 set five criteria before looking, searched the categories that could plausibly satisfy them, and reported that the devices which take a low-voltage trigger are not sealed appliances and the sealed appliances will not take one.

So the specification states the three routes that do exist and no fourth. Convert the light source to a low-voltage LED head, in which case your own timer switches it and nothing is at mains potential. Or leave the mains enlarger on its own switch and count the exposure by ear, which costs resolution and nothing else. Or buy a certified enlarger timer with its own switched socket. A capstone build that cannot be completed within those three is not permitted, and the survey’s own limit is stated with it: a survey by direct supplier reading cannot prove a universal negative, and a product that reopens the question would be welcome rather than embarrassing.

The Level D processes of Part XXVI may be discussed in the chemistry report and are never performed. Mercury development, cyanide fixing and reduction, wet-plate collodion, the uranium and mercury treatments, the dichromated colloids: understanding them is part of understanding what you do, and that part exists so that the understanding is available without the procedure. A portfolio containing one of them is not a portfolio that scored badly. It is outside the specification.

Alternative route

Five obstacles, one route each, and what each route costs. Every substitution used is declared in the documentation, as a declared substitution naming the obstacle, the route taken and what was not tested as a result. That condition applies to all of them and is not negotiable, because an undeclared substitution is a hole in the record exactly where a reader looks for the explanation.

No darkroom. A changing bag and a daylight tank cover every negative, exactly as they have since Part II. Requirement 8’s contact print is made by controlled room light on a bench, or in a UV box for a printing-out material. Requirement 10 is unaffected — most alternative processes are handled in subdued room light and exposed to daylight or UV. What is lost is the enlargement, and the specification says so rather than pretending a contact print supplies it: magnification, the enlarger’s own contribution and local control at a distance are what requirement 8’s second half tests, and there is no substitute for them. A reader on this route declares the gap, and the rubric’s honesty criterion can still reach band 4 with it, because a named limitation is documentation working correctly.

No extraction. The permitted list narrows to reagents whose own classification does not make local exhaust the recognised control. Stage 5 carries that list, verified against safety data sheets rather than assumed, and the practical answer for most readers is a toning session run outdoors or at an open window with a fan, and an alternative process chosen from the iron-based end. The cost is a narrower choice of image colour, and it is a real cost rather than a nominal one.

No woodworking tools, no bench, no space to cut. Requirement 1’s test is the performance goal, not the fabrication. A measured modification to bought equipment satisfies it in full, and so does an instrument assembled from a bought enclosure and off-the-shelf parts. The cost here is close to nil, which is why the “substantially” test was written as a measurement rather than as a description of effort.

Limited mobility, or a shared kitchen. The reduced-volume, single-tray route: smaller formats, one tray used in sequence with a rinse between, chemistry made up in the quantity the session needs rather than by the litre, everything cleared away between sessions, and nothing left standing. The cost is throughput — this route makes six works over more sessions than four trays would — and one technical consequence worth planning for: a single tray makes two-bath fixing awkward, so plan the fixer capacity log carefully and retire baths early rather than stretching them.

Unable to handle a class of chemical. Some substitutions cost nothing and some remove a requirement’s teaching, and the difference has to be stated honestly rather than smoothed over.

Class you cannot handle	Route	What it costs
Sulfide toners	Selenium, gold, or an iron-blue toner, each with its own hazard profile — see Stage 5	Nothing structural: requirement 9 asks for a toned image with a stated intention, not for a specific chemistry
Selenium	A sulfide or gold route, or an iron-blue route	Nothing structural, but the protective-toning discussion in the report changes, because the manufacturers’ claims differ by toner
Silver nitrate	This one bites. The residual-thiosulfate test uses 1 per cent silver nitrate	Requirement 7’s residual result and requirement 11’s measured sentence both become unavailable. Record them as not testable, with the reason as a property of your circumstances, and write the permanence statement with the measurement section stating what was not done. That is a genuine loss of evidence and the rubric’s accuracy criterion caps at band 3
Developing agents that are skin sensitisers	Formulate a solution that contains none — a fixer, a stop bath or a plain-hypo bath satisfies requirement 4, provided the negatives are still processed in something whose composition is known	Requirement 4 is met; requirement 6’s function-of-every-ingredient section becomes shorter, and the report’s developer sections rest on a bought developer’s disclosed components rather than on your own composition
Fine powders of any kind	A bought concentrate for the working baths, with requirement 4 satisfied by a solution made from a liquid raw material of stated composition	This is the substitution with the largest cost, because requirement 4’s whole purpose is the composition you know. Declare it plainly; the accuracy criterion caps at band 2 for the chemistry, and the rest of the rubric is unaffected

What this specification cannot make checkable

Four things, named here rather than dressed as criteria.

How permanent your prints actually are. No test in this course, and no test available at a sink, answers it. The residual tests detect failure rather than degree, and the standards that carry thresholds are ones the course names by number and does not hold. The judgement you make instead is the one Part XII asks for: how permanent does this print need to be, and what did you do about it.

Whether your falloff, your reading spread or your illumination uniformity is acceptable. The course has established no acceptance limits for these — the curve-plotting worksheet prints none for exactly this reason, and the calibration records leave the limits to the student against a cited method. So the specification asks for the measurement and the comparison, and asks you to state your own limit and where it came from.

Whether the body of work is any good. The rubric’s coherence criterion can check that a set argues something and that a print was removed for a reason. It cannot check whether the argument was worth making. That judgement is yours and other people’s, and the portfolio statement exists so that you have to make it explicitly enough to be disagreed with.

Whether the degree of toning conversion you achieved is enough for anything. The standard written to measure conversion effectiveness prescribes no treatment and no criterion a print can meet, so there is nothing to be enough for. The requirement is therefore to measure what changed — hue, matched-patch densities before and after — and to say what was not measured.

Recap

Eleven requirements, six stages, thirteen documented artefacts, four rubric criteria and four failure conditions. The one rule holding it together is that a requirement is a thing you can check on your own bench: a number with a construction and an uncertainty, an object with a map that reproduces it, a test with a control carried through the same session, a sentence whose class is labelled.

Where a check is unavailable the specification says which and why, and asks for the judgement instead. Where a facility is unavailable it gives a route and states the route’s cost. And the condition on every route is the same one that governs the whole document: declare it. The capstone is not assessed on whether you had a full darkroom. It is assessed on whether a reader could tell, from what you wrote down, exactly what you did and exactly what you did not.

Check your understanding

Question 1. Three readers describe their intended Stage 1 work. Which one is a substantial modification under this specification?

• Buying a second-hand enlarger, cleaning the condensers and levelling the baseboard until a grid negative prints square
• Replacing an enlarger lamphouse with an LED head, having first measured that opposite corners of the field differed by 0.11 in density, and measuring afterwards that they agree within 0.03
• Buying a new lens for the enlarger because the old one was soft in the corners
• Reorganising the darkroom so the wet and dry sides no longer overlap

Show the answer and why

Answer: Replacing an enlarger lamphouse with an LED head, having first measured that opposite corners of the field differed by 0.11 in density, and measuring afterwards that they agree within 0.03

The test has four parts and the fourth is the discriminating one: a written goal with a number, a before measurement, an after measurement, and a result that could have failed. Only the LED conversion has all four — the corner disagreement was measured first, so the conversion could have made it worse or left it unchanged, and the outcome is a fact rather than an impression. The other three are all worth doing and none of them could have failed: a levelled baseboard, a sharper lens and a better layout are improvements whose success was not in question, so there was nothing for a measurement to establish.

Question 2. Your capstone report contains the sentence: "The prints were selenium toned for archival permanence and will last well over a century." How many failure conditions does that sentence trip?

• None, provided the toning was actually carried out
• One: the word "archival"
• Two: an inference presented as fact, and an unrecorded number
• It trips no failure condition but scores band 2 on honesty

Show the answer and why

Answer: Two: an inference presented as fact, and an unrecorded number

The century figure is the unrecorded number — no source in this course's corpus attaches a number of years to a toned silver-gelatin print on the strength of its toning, so it comes from nowhere the documentation can name. And "for archival permanence" states an outcome where the evidence supports a mechanism at most: the manufacturers recommend toning as protection against oxidising gases, the degree of conversion in your print was not measured, and the standard written to measure conversion effectiveness prescribes no criterion a print can meet. The honest version names the manufacturer's claim as theirs, states that conversion was not measured, and makes no claim about duration.

Question 3. A reader cannot handle silver nitrate at all. Which of the eleven requirements does that actually affect, and how does the specification handle it?

• Requirement 11 only, and the permanence statement is simply omitted
• Requirements 7 and 11: the residual-thiosulfate result and the measured sentence in the permanence statement both become unavailable, and both are recorded as not testable with the reason, capping the accuracy criterion at band 3
• None: the residual tests are optional throughout the course
• Requirements 3, 5 and 7, because silver nitrate is used in processing

Show the answer and why

Answer: Requirements 7 and 11: the residual-thiosulfate result and the measured sentence in the permanence statement both become unavailable, and both are recorded as not testable with the reason, capping the accuracy criterion at band 3

The residual-thiosulfate test is the one place in the capstone that needs silver nitrate, and it feeds two requirements: it is one of requirement 7's six numbers and it is the measured sentence in requirement 11's four-part statement. Silver nitrate is not used in processing, so requirements 3 and 5 are untouched. The specification's handling is the same everywhere a route is closed: complete what can be completed, record the missing item as not testable with the reason, and declare the substitution. What it does not do is pretend the loss is costless — this is the one substitution in the table that removes a measurement rather than exchanging one route for another, so the accuracy band is capped and the reason is stated.

Question 4. Why does the specification require reflection densities to be quoted as "relative to tile serial NNNN" rather than as a bare number?

• Because the densitometer is homemade and its readings are therefore approximate
• Because the course-built reflection scale is anchored to an uncertified white tile, so the numbers are internally consistent but not on any external scale — and the omission only costs something on the day somebody compares one with a published figure
• Because reflection density changes with humidity and the tile records the conditions
• Because ISO 5-3 requires a serial number on every density measurement

Show the answer and why

Answer: Because the course-built reflection scale is anchored to an uncertified white tile, so the numbers are internally consistent but not on any external scale — and the omission only costs something on the day somebody compares one with a published figure

The qualifier is about the reference, not about the instrument's quality. A reflection scale needs a top anchor, and the anchor here is a white tile the course did not buy a certificate for, so every number on the scale is stated relative to that specific tile. Within your own work that is entirely sufficient — comparisons across your own sessions are valid, which is what set consistency and dry-down need. What is not valid is comparing the number with somebody else's, and writing the serial every time is what stops that comparison being made by accident three years later.

Check my answersClear