Experiment: Comparing Developing Agents

D-76 contains two grams of metol and five of hydroquinone instead of seven grams of either. Part VIII gave the reason the trade names give — the pair does more than the parts — and then refused to teach the mechanism, because no source the course holds states it.

That refusal leaves something a bench can settle. Whether the pair exceeds the parts is not a question about mechanism at all; superadditivity is a question about how much silver appears, and eight strips can answer it without naming a single intermediate.

Purpose

To develop one exposure batch in six baths differing only in which agents are dissolved in them — metol alone, hydroquinone alone, Phenidone alone, ascorbate alone, metol with hydroquinone, Phenidone with hydroquinone — each agent at exactly the concentration the pair uses it at, and to test whether the pairs do more than their singles.

Hypothesis, part one. Hydroquinone alone at this bath’s pH is nearly inactive: its pK_a is 9.88, so at a measured pH near 9.3 only about a fifth of it is ionised, and Kodak’s 1928 primer says its potential is so low that it is rarely used alone and is generally used with Elon.

Hypothesis, part two. The two pairs each develop more than the sum of their singles. Densities add: if the agents worked independently, the silver laid down by the pair at any exposure would be the silver laid down by one plus that laid down by the other. That is the additive prediction, and this session tries to break it.

The control. The shared reference strip — D-76 1+1 for 11 minutes at 20 °C, the standard condition of the test-negative lab — ties this session to the others. Within the session the single-agent cells are the controls: M is the control for MQ, and P and Q together are the control for PQ.

The one variable that changes is the agent load. The design work below is almost entirely about making that sentence true, because six agents dissolved in one base do not naturally leave its pH where they found it.

!What this experiment can prove and what it can only bound

A proper additivity test needs a densitometer, because it compares the density of the pair against the sum of the densities of the singles, step by step, and this part has none until Part XXVII. A step reading settles it in the one case where the arithmetic collapses: if the hydroquinone-alone strip is nearly blank, the additive prediction for the pair is simply “the same as the metol strip”, and no densities are needed to test that. So this is a falsification test, made sharp on purpose by choosing a partner that contributes almost nothing alone, with the strips archived so that Part XXVII can turn the falsification into a number.

Learning objectives

By the end of this session you will be able to:

• Build one buffered base that six different agent loads cannot move, and show by arithmetic why the published borax quantity would not have done it.
• Prepare and use a 1 % w/v stock solution for a quantity too small to weigh, holding the added volume constant across every cell so dilution does not travel with the agent.
• Choose between an acid and its sodium salt as a formulation decision rather than a weighing one.
• State the additive prediction for a pair, say what would falsify it, and say what a step reading cannot establish even when it does.
• Rank activity, fog and the colour of the used bath, and describe the difference in image character between two developers of equal activity.

Prerequisites

• Lab: preparing standard test negatives, for the batch, the standard condition and your resolution limit.
• Superadditivity, which owns the phenomenon, Kendall’s ratios and the unverified mechanism; alternative developing agents, which owns ascorbate; classical developing agents and alkalis, buffers and pH.
• Concentration and dilution and measurement and uncertainty.
• The SOPs for opening the laboratory, mixing from a stock, calibrating a pH meter, checking the balance and labelling a container.

Safety classification

Level A: a standard home darkroom, with the weighing worked at Level B controls.

• The solutions in use. Alkaline developers by the hundred millilitres in open cylinders, with gloves and eye protection — the handling of Part VIII’s mixing lab.
• The weighing, which is the whole of the elevated risk. A hundred grams of sulfite, twenty of borax — a Level B substance on its reproductive classification — and four agents down to one hundredth of a gram. Phenidone is weighed at a scale where a puff of dust is a large fraction of the dose, exactly the case Princeton’s arts-safety guidance has in mind when it asks for an enclosure, local exhaust or an approved dust respirator. Tray, still air, one jar open at a time, gloves, eye protection, particulate mask, no film open in the room. A reader who is pregnant, breastfeeding or trying to conceive should read the borax page before opening the tub.
• Catechol and pyrogallol are deliberately absent. Both would fit and both are Level B; excluding them is what keeps the session Level A.
• Waste. Alkaline developer in one labelled container; fixer and its first rinse, silver-bearing, in another. Never combined.

What is not a hazard here, and why. There is no sulfur dioxide risk from a bath carrying 0.79 mol/L of sulfite: sulfite releases the gas on meeting an acid, and this session has no acid step, the stop being a plain water rinse. Sodium ascorbate is the mildest substance on the bench — 98.9 per cent of 361 ECHA reports say it meets no GHS criterion, a well-populated finding rather than a thin one — and it changes nothing about the protective equipment, which is set by the metol and hydroquinone on the same tray.

Hazards

Metol: skin sensitisation, H317 under a harmonised CLP entry, cumulative and not reversible — gloves throughout. Hydroquinone: serious eye damage and aquatic toxicity, in four of the eight baths and in the reference — eye protection from the first jar to the last cap.

Phenidone: harmful if swallowed, toxic to aquatic life. H302 in 98.4 per cent of reports and H411 in all of them, with the same statements in the harmonised entry, so the classification is agreed at law. What is thin is the investigation: PubChem holds no reactivity datasheet, no International Chemical Safety Card and no NIOSH Pocket Guide entry. Absence of a finding is not a finding of absence, and the safety data sheet for your jar is the authority for it.

Borax: reproductive toxicity, and dust. H360 at 93.2 per cent of reports, with a workplace exposure limit for borates in HSE’s EH40. Sodium ascorbate meets no GHS criterion in 98.9 per cent of reports and is still a powder at the balance. The waste stream — about 1.0 litre — carries agents notified as very toxic to aquatic life with long-lasting effects.

Required PPE

• Single-use nitrile gloves, 0.2 mm, per HSE’s COSHH essentials sheet P1, changed when contaminated.
• Eye protection throughout.
• A particulate mask for the weighing, which is what allows a Level B weighing inside a Level A session and matters more here than anywhere else in the part, because of the hundredth-gram weighings.
• An apron or overall, clean dry hands or lint-free gloves for the film, and dedicated labelled utensils with a separate set for fixer.

Ventilation

Nothing here evaporates in a way that puts a substance into the air, so ventilation is not removing a vapour: it is diluting during the ten minutes when powder jars are open. HSE’s COSHH essentials sheet P1 sets more than five air changes an hour with a through draught for manual film development, and that is the standard. Over five hours it also carries away the ammonia smell of a rapid fixer, which is amenity rather than control.

Materials

Item	Quantity	Note
Wedge strips from the Part IX exposure batch	8	Datum-notched; identity nips at allocation
Negative sleeves and card label slips	8	One slip per strip
Small clear jars for the used cells	8	For the colour ranking
Random reading cards	8	Shuffled, for blind reading
White card	1	The background every colour reading is made against

Chemicals

Chemical	Quantity	Form
Sodium sulfite, anhydrous	100 g	Solid, for 1 L of base at 100 g/L
Borax	20 g	Solid, for the same litre at 20 g/L, weighed at Level B controls
Metol	0.50 g	Solid: 0.25 g into each of two cells
Hydroquinone	2.52 g	Solid: 0.63 g into each of four cells
Phenidone	1.00 g	Solid, for 100 ml of a 1 % w/v stock; about 0.09 g is consumed
Sodium ascorbate	1.13 g	Solid, one cell
D-76 stock, from Part VIII’s lab	40 ml	Solution, for the reference strip at 1+1
Rapid fixer (ammonium thiosulfate type)	500 ml at 1+4	Twice the clearing time
Wetting agent	500 ml at 1+200	Final rinse
Water, at 20 °C	about 4 L	Base, rinse and wash

Equipment

Nine 100 ml measuring cylinders in a rack, standing in a tray deep enough to serve as a water bath, and eight small beakers for mixing and pH readings. A balance reading to 0.01 g, checked against a reference; a pH meter with pH 7 and pH 10 buffers; a thermometer to 0.1 °C; a 10 ml syringe and a 100 ml cylinder for the stock. Three shallow trays, tongs one pair per solution with a separate set for fixer, a drying line, a lightbox and a flatbed scanner.

Estimated cost

££, and unusually for this part the cost is in the agents rather than the film: phenidone is bought in the smallest jar sold and about 0.09 g is consumed here. The planner carries the numbers.

Estimated consumables cost

Unusually for Part IX the cost is in the agents rather than the film — and the most expensive substance on the page, phenidone, is bought in the smallest jar sold and 0.09 g of it is consumed here.

Consumed	This session	Sourced price	Cost this session
Sodium sulfite, anhydrous	100 g, for 1 L of base	£13.68–£19.98 per 1 kg, anhydrous	£1.37–£2.00
Borax	20 g, in the same litre	£9.98 per 200 g, decahydrate	£1.00
Metol	0.50 g, two cells	£16.20 per 50 g (£0.32 a g)	£0.16
Hydroquinone	2.52 g, four cells	£11.89 per 50 g (£0.24 a g)	£0.60
Phenidone	1.00 g weighed, about 0.09 g consumed	£9.79 per 10 g, checked 7 September 2026 (£0.98 a g)	£0.98 weighed, about £0.09 consumed
Sodium ascorbate	1.13 g, one cell	None. The file prices ascorbic acid at £11.90 for 50 g and says on its own face that sodium ascorbate is a different substance it does not price	—
D-76 stock, from Part VIII	40 mL for the reference strip	Costed in Part VIII’s mixing lab	—
Wedge strips from the Part IX exposure batch	8	Costed in preparing-standard-test-negatives	—
Rapid fixer concentrate	100 mL, to make 500 mL at 1+4	£21.05–£25.98 per 1 L of ammonium thiosulfate concentrate, diluted 1+4 for film	£2.10–£2.60
Wetting agent	2.5 mL, for 500 mL at 1+200	£28.70 per 1 L of concentrate, diluted 1+200	£0.07
Negative sleeves and card label slips	8 pockets, 8 slips	None. A named price gap: sleeving that passes the Photographic Activity Test	—

The priced rows come to £6.28 to £7.41 for one run of this session, at the retail ranges read on 5 and 7 September 2026 and recorded in the laboratory planner. That is a floor, not a total: 2 of the 11 rows carry no dated price, so they are counted as nothing here and are certainly not free. A priced entry is a dated range to plan against, never a quotation.

Phenidone now has a dated UK figure and sodium ascorbate does not. The price file carries ascorbic acid and states on its own face that the sodium salt is a different substance sold under a different name — which is precisely the distinction an ascorbate developer is written around, and precisely why one substitution here is not the other. The subtotal above therefore carries one of the two substances the session is actually about, and not the other.

Waste streams

1. Alkaline developer, about 1.0 L, carrying all four agents, the sulfite and the borate. One labelled container, per the general waste SOP.
2. Spent fixer and its first rinse, silver-bearing, to the silver stream.
3. Rinse and wash water: the rinse after the developer goes with stream 1, the first change of wash water after fixing with stream 2, the rest to the general stream.

Never combine streams 1 and 2: one is a concentrated sulfite solution and the other is acidic.

Alternative route

Darkness is needed for one operation: giving eight strips their identity nips, about fifteen minutes with the light off, or under a deep red safelight on ILFORD ORTHO Plus.

The route without phenidone. A reader who cannot get it can run four cells — M, Q, MQ and the reference — and still test additivity completely, because the falsification argument needs one pair and one nearly-inactive partner. What is lost is the comparison between the two pairs.

Preparation

About 60 minutes, the evening before or the same morning.

The base, and why the published borax quantity will not do

Every cell is 120 ml of one base plus 5.0 ml of added liquid, made up to 125 ml. The base is D-76 with the agents taken out and the borax multiplied by ten: sodium sulfite 100 g/L and borax 20 g/L. The multiplication is the whole design, because each agent brings its own acid load and the loads are unequal. Metol is a hemisulfate: at 2.0 g/L its sulfuric acid releases 0.0116 mol/L of protons the moment it dissolves. Hydroquinone at 5.0 g/L and pH 9.3 is about a fifth ionised, releasing about 0.009 mol/L. Sodium ascorbate releases essentially none, because ascorbic acid’s pK_a of 4.04 means the salt arrives already ionised. Phenidone releases an amount the course cannot calculate, because no dissociation constant for it appears in the IUPAC dataset used for the other agents.

🧮Two grams of borax against twenty

Work the proton bookkeeping across the sulfite and borate couples — pK_a2 of sulfurous acid at 7.19, pK_a of boric acid at 9.27 — and the six cells come out like this.

Borax in the base	M	Q	P and ascorbate	MQ	PQ	Spread
2 g/L, as in D-76	9.10	9.19	9.56	8.97	9.19	0.59 units
20 g/L, as in D-76R	9.24	9.26	9.31	9.18	9.26	0.13 units

At the published quantity the six cells span more than half a pH unit, and the activity series has just measured what half a unit does to a developer. Worse, the pH would move with the agent, so the two could never be separated. At twenty grams the spread is 0.13 units, inside the ±0.2 specification of a teaching electrode — a weaker statement than it looks, since you cannot show the cells are at one pH, only fail to show they are not.

The arithmetic assumes ideal behaviour at 0.8 mol/L, and you can see how poor that is: the same method on real D-76 gives 8.97 against ILFORD’s measured 8.60 to 8.70 for ID-11 stock. Read it as three tenths high, and as directions. The phenidone cells it cannot reach at all, which is why every cell’s pH is measured on the day.

The rest of the preparation

1. Open the laboratory to the SOP; check the balance against a reference.
2. Mix 1 litre of base: 100 g of sulfite into about 700 ml of water at 40 to 45 °C, then 20 g of borax, which is slow cold, then up to 1 L and cool to 20 °C. Give it a version code per the scheme.
3. Make 100 ml of a 1 % w/v phenidone stock: 1.00 g up to 100 ml in water at about 40 °C, the temperature ILFORD specify for the part A of their own powder developers. The course has no verified solubility figure for phenidone in water — PubChem’s heading returns no values — so this stock is a practical arrangement, not a sourced one. If a gram will not dissolve, halve the stock strength and double every volume below rather than reaching for a solvent, which would add an ingredient to two cells and not the others. Make it on the day; there is no keeping figure either.

📓Why 0.044 g of phenidone is measured with a syringe and not a balance

Phenidone goes in at 0.35 g/L, which is 0.044 g in 125 ml — known to about ±23 per cent on a balance reading to 0.01 g. A 1 % w/v stock moves the measurement to a volume: 4.4 ml delivers the same 0.044 g, a 10 ml syringe reads to 0.1 ml, and the error in the gram that made the stock is shared identically by both phenidone cells rather than made twice.

Every cell then receives exactly 5.0 ml of added liquid, and in four of them all of it is water: otherwise the phenidone cells would be four per cent more dilute than the rest, which is far below what a step reading can see and therefore exactly the kind of confound that never announces itself.

4. Weigh the agents into eight labelled beakers, at Level B controls, recording every mass actually weighed. At 0.25 g on a 0.01 g balance the metol is known to ±4 per cent — the largest single uncertainty in these compositions, and it belongs in the report.

   Cell	Metol	Hydroquinone	Sodium ascorbate	1 % phenidone stock	Water to make 5.0 ml
   M	0.25 g	—	—	—	5.0 ml
   Q	—	0.63 g	—	—	5.0 ml
   P	—	—	—	4.4 ml	0.6 ml
   A	—	—	1.13 g	—	5.0 ml
   MQ	0.25 g	0.63 g	—	—	5.0 ml
   PQ	—	0.63 g	—	4.4 ml	0.6 ml
   Q-long	—	0.63 g	—	—	5.0 ml

   Each beaker then takes 120 ml of base, giving 125 ml at 2.0 g/L metol, 5.0 g/L hydroquinone, 9.0 g/L sodium ascorbate and 0.35 g/L phenidone.

⚗Three quantities that were chosen, and the reason for each

Metol 2.0 g/L and hydroquinone 5.0 g/L are D-76’s own, because the pairs have to be pairs a formulary actually prints and the singles have to be at the same concentrations as in the pairs.

Phenidone 0.35 g/L is Kendall’s substitution ratio applied to that metol: his 1941 patent claims that replacing metol with about one fifth to one sixth of its weight of 1-phenyl-3-pyrazolidone gives similar development characteristics, and on 2 g that is 0.33 to 0.40 g.

Sodium ascorbate 9.0 g/L is a molar equivalent, not a formula — hydroquinone at 5.0 g/L is 0.045 mol/L and the salt at 198.11 g/mol matches it. No published ascorbate quantity exists for a base like this, because XTOL’s composition has never been published: Kodak’s sheet gives a working pH of 8.20 ± 0.05, states that it contains no hydroquinone, and stops. The salt is chosen over the acid deliberately, because ascorbic acid at the same molarity would dump 0.045 mol/L of protons into the bath — four times metol’s load — and the cell would be a pH experiment wearing an agent’s name. That is a formulation decision, not a weighing decision.

5. Calibrate the pH meter at 7 and 10 per the SOP, bracketing the samples, and measure every cell at 20 °C once its agent has dissolved, recording all seven with the electrode’s stated accuracy beside each. Do not adjust anything: the readings are the evidence that the base did its job, and if two cells differ by more than the electrode’s accuracy, the comparison has a pH difference inside it and the report must say so.
6. Mix 80 ml of D-76 at 1+1 for the reference strip, freshly, as Kodak’s sheet directs.

Procedure

Eight strips, one equal time, two batches. The 300 minutes divides as 40 to bring the bench to temperature and re-measure the cells, 15 in darkness allocating strips, 90 for the two wet batches and the wash, 10 for the colour ranking while the strips hang, 60 while they dry, and 85 for the blind reading and the scanner pass.

Why the equal time is 12 minutes

Every cell except two gets 12 minutes at 20 °C — half again ILFORD’s published 8 minutes for D-76 at stock with this film. The extra time is spent on the weak cells: a single-agent bath registering nothing at 8 minutes tells you only that it is slower than D-76, which you knew. The MQ cell is over-developed as a result, which is a predicted cost rather than a spoiled strip. The exceptions are the reference, which keeps the standard D-76 1+1 for 11 minutes, and Q-long, a second hydroquinone cell at 24 minutes.

🧪What the second hydroquinone strip is for

“Nearly inactive” and “slow” are different claims and one strip cannot tell them apart: a dead bath gives no more density at 24 minutes than at 12, a slow one gives visibly more. That is Hurter and Driffield’s logic about bromide, borrowed — a restrainer retards rather than stops — and one extra strip turns an impression into a two-point series.

Stage 1 — the bench, about 40 minutes

1. Bring the bench, the cylinders and the cells to 20 °C and hold within 0.3 °C, the limit Kodak’s process-control publication sets before temperature affects the result. Check between batches.
2. Re-measure and record the pH of all seven cells on the day. The ascorbate cell is the one to watch: CAMEO records that ascorbate solutions oxidise in air at a rate rising with pH, accelerated by iron and copper, so mix it last.
3. Pour 80 ml of each cell into its labelled cylinder and keep the rest in the beaker for the colour comparison.

Stage 2 — allocate the strips, in darkness, about 15 minutes

4. With the light off, take eight strips from the tin. Working from the datum notch, give each its identity nips: one to four on the top edge for M, Q, P and A; one to four on the lower edge for MQ, PQ, Q-long and the reference.
5. Close the tin, put it away, turn on the light, and lay each strip beside its cylinder.

Stage 3 — the wet run, about 90 minutes

The agitation script, identical to every other strip in the part: lower the strip in and move it gently up and down for the first 30 seconds, then lift, drain for two seconds and re-immerse once at the start of each subsequent minute.

Two batches of four, and one strip that stays in twice as long

M — metol 2.0 g/L

develop 12 minrinsefix

Q — hydroquinone 5.0 g/L

develop 12 minrinsefix

P — Phenidone 0.35 g/L

develop 12 minrinsefix

A — ascorbate 9.0 g/L

develop 12 minrinsefix

MQ — metol + hydroquinone

develop 12 minrinsefix

PQ — Phenidone + hydroquinone

develop 12 minrinsefix

Shared reference, D-76 1+1

develop 11 minrinsefix

Q-long — hydroquinone, 24 min

develop 24 minrinsefix

0 min15304560

The singles go first so that the pairs are developed by hands that already know the script, and Q-long starts with the second batch and finishes alone. Four strips a minute apart is the most one pair of hands can agitate on a one-minute cycle.

6. Batch one: M, Q, P and A, started one minute apart, twelve minutes each.
7. Batch two: MQ, PQ, the reference and Q-long, one minute apart, as soon as batch one is out of the developer. The reference gets 11 minutes and Q-long 24.
8. After each strip: rinse one minute in two changes of plain water, then fix for twice the clearing time measured on a scrap of the same film. Every strip in Part IX gets a water rinse rather than an acid stop, for the reason the test-negative lab sets out; the cost ILFORD name is an increased risk of processing marks and stains.
9. Wash all eight together, 5 to 10 minutes within 5 °C of the developer, final-rinse in wetting agent at 1+200, hang in still dust-free air. Do not squeegee.

Stage 4 — the colours, immediately, about 10 minutes

10. Pour each used cell into its clear jar, stand the eight against the white card in one light, and rank them before anything else. Write a colour and an intensity for each — water clear, faint straw, clear yellow, amber, brown — never “darker than the last one”. A used developer’s colour goes on changing for hours and this reading does not survive the evening.

Expected observations

M, metol alone. A real image: Kodak’s 1928 primer describes the Elon image as coming up quickly and gaining density slowly, so expect a strip that finds its threshold and then does not go far. Q, hydroquinone alone, should be nearly blank — the prediction the whole design leans on — and Q-long then decides between slow and dead.

P, Phenidone alone. Weak, and possibly weak for two reasons at once: Kendall’s patent uses the agent alone at 6 g/L and as a metol substitute at 0.3 g/L, so a cell at 0.35 g/L holds by far the fewest reducing molecules on the bench. A strip uniformly thin, clearest steps included, suggests the agent ran out; one thin at the toe and closer to normal at the top, that it was simply slow. A, ascorbate alone, is the genuine unknown: no published concentration exists for a base like this.

MQ and PQ. Both denser and faster than any single. If they are not, either the additive hypothesis has survived — worth reporting loudly — or the mixing went wrong, and the pH readings are the first place to look.

The colours and the fog. Expect the hydroquinone cells to yellow and the metol cell to be uninterpretable, since the course cannot name metol’s oxidation product. Read every masked patch against clear fixed base.

What is happening chemically

Ionisation: why the same agent is a different developer at a different pH

An agent hands electrons to silver from its ionised form. Hydroquinone’s pK_a of 9.88 puts the ionised fraction at about 6 per cent at ID-11’s measured 8.65, 21 per cent at this session’s 9.3 and about three quarters at a paper developer’s 10.4 — which is why ILFORD measure Bromophen stock at 10.30 to 10.50 and why the hydroquinone-alone developers of the formularies are caustic. Metol’s amine is already neutral above about pH 8, so its ionisation barely moves. That difference is why this session pins the pH instead of letting each agent find its own.

C₆H₆O₂ + 2 OH⁻ → C₆H₄O₂ + 2 H₂O + 2 e⁻

Hydroquinone giving up two electrons in alkaline solution and becoming quinone; the electrons go to silver ions at the crystal surface

Regeneration: the account this course draws and does not assert

The standard explanation is that the primary agent works at the crystal surface and the secondary, in solution, reduces the oxidised primary back to its working form, so a small quantity of the primary is used many times over. Part VIII marks that account unverified and this page does not upgrade it: no Tier-1 source the course holds states it, and the course cannot name metol’s oxidation product, so the step cannot be written as an equation without an unnamed species on both sides. What can be written is the analogous reaction Kodak’s primer states, in which sulfite restores hydroquinone from quinone:

C₆H₄O₂ + SO₃²⁻ + H₂O → C₆H₆O₂ + SO₄²⁻

A developing agent restored by a cheaper reductant — the only regeneration in this chemistry that the course can source

That shows the idea is not exotic. It does not show that hydroquinone does the same for oxidised metol — which is why this experiment is worth running. It measures the effect without touching the mechanism, and a pair that beats its singles is a fact about density whatever explains it.

⁇A measurement this session could contribute to, and one it cannot

Could contribute to. Whether the excess is the same for both pairs: similar margins make the effect look like a property of pairing, while a much larger margin for the pyrazolidone pair is a quantitative fact any mechanism must account for, and Kendall’s twentyfold observation says to expect it.

Cannot contribute to. The identity of any oxidation product, the semiquinone, or the redox potential of any agent at developer pH. Part VIII marks all three unsourced, none is visible in a wedge step, and a colour in a jar is not a compound.

Data to record

Session header and the exposure batch block, copied by name from the test-negative lab. The cell table: name, notch code, random reading number, every mass actually weighed, the volumes added, the base’s and the cell’s version codes, and the measured pH with the electrode’s stated accuracy.

The reading table: threshold step, scale length, a description of the masked patch against clear fixed base, scanner code values with the scan settings, and the date. The colour table, eight rows, in one sitting against a white card, with the time. The deviations, including the ones that feel too small to matter.

Analysis

1. Read blind. Shuffle the cards, assign a random number to each strip, read all eight in one sitting in random order on one lightbox, and open the key only when the table is full.
2. Apply the resolution limit first: no difference smaller than it is reported as a result.
3. Check the reference against the control set. If it has moved by more than that limit, every conclusion becomes a statement about differences between cells rather than about absolutes.
4. Rank the six cells by activity, threshold step first, scale length second; rank fog separately.
5. Test additivity, explicitly. Before you look at the pairs, write down the additive prediction for each — MQ should equal M plus Q, and PQ should equal P plus Q, in density at every step.

The additivity test, and why a nearly inactive partner makes it decidable by eye

1. 1M — the single that works — metol at 2.0 g/L, the concentration the pair uses it at
2. 2Q — the single that nearly does not — hydroquinone at 5.0 g/L and pH 9.3, about a fifth ionised
3. 3The additive prediction — M plus Q — which, when Q is nearly blank, is just M
4. 4MQ — the pair — everything above the dashed line is the result this session exists to find

Drawn as the logic, not as data: the heights are the shape of the prediction and no strip has been measured. The point of the drawing is that the shaded gap is visible without a densitometer only because the second column is nearly nothing.

6. State the case you are in, in one of three sentences. “Q was blank, so the additive prediction for MQ was M, and MQ exceeded M by n steps: additivity is refuted.” Or: “Q produced measurable density, so the prediction needs densities and this session bounds rather than settles it.” Or: “MQ did not exceed M by more than the resolution limit: a null result.” Choosing honestly between them is the analysis.
7. Compare the two pairs. MQ against PQ, at one time and one pH with the primary agent at Kendall’s substitution ratio, is as close as a home darkroom gets to the question the 1941 patent asked.
8. Read the colours as evidence about the bath, not the strip. Kodak’s primer warns that hydroquinone’s sulfonates are colourless, so a bath that stayed clear is not thereby a bath that stayed strong: colour is a positive tell whose absence proves nothing.
9. Scan everything in one pass with the wedge included and every automatic function off, then repeat steps 4 and 5 on the numbers. Where eye and scanner disagree about an order, that pair is inside the resolution of both and neither has separated it.

Art track: what the ranking leaves out

Two developers can reach the same contrast and give different negatives, and the best short account of the difference is a hundred years old: Kodak’s 1928 primer says an Elon image comes up quickly and gains density slowly, while a hydroquinone image comes up slowly and gains density steadily and rapidly. A pair is a blend of those behaviours rather than an average.

That is a claim about the shape of the negative, worth looking for even though the reading method cannot measure it. Compare M and Q at whatever steps they both managed — metol’s strip separating the faint steps and then flattening, the hydroquinone strip doing the opposite — then ask which parent MQ resembles at the toe and which at the top. A negative whose toe is well separated prints shadows that hold detail without burning; one whose top end keeps climbing prints highlights that resist blocking.

Write one paragraph naming which of your six baths you would use for a subject you actually photograph, and what specifically about the strip made you choose it. That is the art-track deliverable, and the assignment asks for it again.

Troubleshooting

What you see	Likely cause	What to do
The phenidone will not dissolve in 100 ml	Unknown solubility; the course has no figure	Warm to 40 °C as ILFORD do for their part A; failing that, halve the stock strength and double every volume, recording the change
The P strip is uniformly thin, clearest steps included	The cell may be agent-limited rather than slow	Re-run one strip at the same concentration in twice the volume: more agent at the same concentration separates “ran out” from “too slow”
The A strip is much weaker than expected	Ascorbate oxidation, which CAMEO records is accelerated by alkaline conditions and by iron and copper	Mix that cell last and use it first, and note your water source: the course cannot say whether a published ascorbate developer carries a sequestrant, because XTOL’s composition is unpublished
Drying marks read as density	Squeegeed, or the wetting agent was wrong	Re-wash and re-dry. Never squeegee a strip Part XXVII will measure

Clean-up

Once the colour reading is written down, pour each jar into the alkaline developer container, rinse each vessel twice into the same container and only then into the sink; fixer and its first rinse go to the silver container. Keep developer and fixer utensils separate — Kodak names uncleaned mixing equipment among the leading causes of solution contamination. Wipe the balance and the tray, throw the gloves away, close the laboratory to the SOP.

Storage

Discard all seven cells once their colours are read, and the phenidone stock with them: there is no keeping figure for it, and a stock of unknown age is a variable pretending to be a reagent. Keep the base only if a repeat run is days away, capped full and dated. Date the phenidone jar on opening: a developer uses tenths of a gram, so the smallest jar sold will be old when it is used, and the course has no published shelf-life figure for the dry solid.

Archive every strip, fixed for twice its clearing time, washed to the published standard, dried untouched, sleeved with the eight-field card slip the test-negative lab sets out — with this session’s specifics: the agents and their concentrations, the base’s and the cell’s version codes, the measured pH with its temperature, the equal time and the temperature range. A strip whose agent load is not on its slip is a strip that was never made.

Disposal considerations

Two containers, labelled with contents and date, never combined.

The alkaline developer waste, about 1.0 L, carries all four agents. Metol and hydroquinone are notified as very toxic to aquatic life with long-lasting effects and phenidone one category below — about 0.09 g in the session, so the waste is dominated by the hydroquinone and the alkali. The ascorbate cell is bottled with the rest, although sodium ascorbate carries no GHS classification in 98.9 per cent of 361 reports, because the cell also holds 100 g/L of sulfite and 20 g/L of borate and has been through a film. The fixer and its first rinse are silver-bearing and go to the silver stream.

ILFORD’s advice to domestic users in the United Kingdom is to bottle each waste chemical separately, label it and take it to a Household Waste and Recycling Centre. That is one country’s answer; the disposal page sets out why the question is jurisdictional. Check your local regulations; they govern, and they differ between authorities within one country.

Questions

1. Your hydroquinone strip is not blank: it carries three faint steps. What does that do to the additivity test, and which of the Analysis’s three report sentences must you now write?
2. Every cell receives 5.0 ml of added liquid and in four of them all of it is water. Name the error that guards against, and why it would have been invisible in the strips.
3. A reader proposes ascorbic acid instead of sodium ascorbate. Give the arithmetic, then the reason it is a formulation change rather than a weighing change.
4. Q-long comes back identical to Q. Write the conclusion in one sentence, then name the one additional cell you would run to test it.
5. The MQ cell measures pH 9.18 and the P cell 9.44. What has happened, and what does it do to a comparison between the MQ and PQ strips? And the metol cell’s used developer is a deep amber: what does this course allow you to write about it, and what does it not?

Further experiments

Phenidone as a developer in its own right. Kendall’s Example I uses the agent alone at 6 g/L — twenty times his Example II. One cell at 6 g/L turns “Phenidone alone is weak” into “Phenidone alone at the substitution concentration is weak”, which is a far more useful sentence.

The same six cells at carbonate pH. Add 10 g/L of sodium carbonate and repeat. The prediction is sharp: hydroquinone should gain far more than metol, because at pH 10.4 about three quarters of it is ionised against a fifth here, while metol’s amine is neutral in both.

Check your understanding

Question 1. Why must the single-agent cells use exactly the concentrations the pairs use, rather than a concentration at which each agent works well on its own?

• To save chemicals
• Because "the sum of the singles" is only defined if each single is at the same concentration it has in the pair; otherwise the comparison is between different baths and the additive prediction means nothing
• Because the manufacturers specify it
• Because the pH would otherwise differ between cells

Show the answer and why

Answer: Because "the sum of the singles" is only defined if each single is at the same concentration it has in the pair; otherwise the comparison is between different baths and the additive prediction means nothing

Superadditivity is the claim that a bath containing A at concentration a and B at concentration b does more than a bath of A at a plus a bath of B at b. Change either concentration in the singles and you are no longer testing that claim. It is also why the Phenidone cell is run at 0.35 g/L, which is Kendall’s substitution ratio applied to D-76’s metol, even though his own patent uses the agent alone at twenty times that: the higher concentration answers a different and easier question, and it belongs in Further experiments.

Question 2. The base carries 20 g/L of borax rather than D-76’s 2 g/L. What does the extra borax buy?

• A higher pH, which makes every agent more active
• Buffer capacity: at 2 g/L the six agent loads spread the cells over about 0.59 pH units, so pH would move with the agent and the two could never be separated; at 20 g/L the spread falls to about 0.13
• Better solubility for the phenidone
• Longer keeping for the base

Show the answer and why

Answer: Buffer capacity: at 2 g/L the six agent loads spread the cells over about 0.59 pH units, so pH would move with the agent and the two could never be separated; at 20 g/L the spread falls to about 0.13

Each agent brings its own acid or base load. Metol is a hemisulfate and releases about 0.0116 mol/L of protons; hydroquinone releases about 0.009 mol/L by ionising; sodium ascorbate releases essentially none because it arrives already ionised; and phenidone releases an amount the course cannot compute at all, since no dissociation constant for it appears in the IUPAC dataset. A borate pool ten times larger absorbs all of that. The pH does rise a little as well, but the buffering is the point and the rise is a side effect that applies equally to every cell.

Question 3. What makes the additivity test decidable by eye in this design, when a proper test needs a densitometer?

• The step wedge converts density directly into a number
• The prediction collapses: if one single is nearly blank, the additive prediction for the pair is just the other single, so a pair that is obviously denser and faster than that single has falsified additivity without any density being measured
• Blind reading removes the need for densities
• The scanner supplies calibrated densities

Show the answer and why

Answer: The prediction collapses: if one single is nearly blank, the additive prediction for the pair is just the other single, so a pair that is obviously denser and faster than that single has falsified additivity without any density being measured

Densities add; wedge steps are positions and do not. So in general you cannot test additivity by counting steps. The design gets round that by choosing a partner that is nearly inactive alone, which makes the sum of the two singles indistinguishable from one of them. Everything the pair does beyond that is excess, and it is visible. What the method still cannot do is say how large the excess is, which is why every strip is archived for the densitometer in Part XXVII.

Question 4. A student reports that the hydroquinone cell "stayed clear, so it was still fresh". What is wrong with the inference?

• Nothing; colour is a reliable indicator of developer condition
• Kodak’s own primer records that quinone reacts with sulfite to give colourless hydroquinone mono- and disodium sulfonates, and warns that a colourless developer is no indication of undiminished power — so colour is a positive tell whose absence proves nothing
• Hydroquinone developers never change colour
• The colour should have been read after the strips dried

Show the answer and why

Answer: Kodak’s own primer records that quinone reacts with sulfite to give colourless hydroquinone mono- and disodium sulfonates, and warns that a colourless developer is no indication of undiminished power — so colour is a positive tell whose absence proves nothing

The same book gives two fates for the quinone produced by development, and reconciles neither: one returns hydroquinone at the sulfite’s expense, and the other converts it to colourless sulfonates that are not hydroquinone. On the first route the reservoir is topped up; on the second it is drained while the bottle stays clear. That is why the analysis treats a yellowed bath as evidence and a clear one as no evidence, and why the fogged-strip activity test used in the break/fix page exists at all.

Question 5. Sodium ascorbate is used rather than ascorbic acid at the same molarity. What would the acid have done to the cell?

• Nothing, since the base is buffered
• Released about 0.045 mol/L of protons — roughly four times metol’s load — which even a 0.105 mol/L borate pool would not have absorbed without moving, so the cell would have been a pH experiment wearing an agent’s name
• Precipitated the sulfite
• Reacted with the borax

Show the answer and why

Answer: Released about 0.045 mol/L of protons — roughly four times metol’s load — which even a 0.105 mol/L borate pool would not have absorbed without moving, so the cell would have been a pH experiment wearing an agent’s name

Ascorbic acid’s pKa1 is 4.04, so at any developer pH the acid form has long since given up its first proton — into the bath. The salt arrives with that proton already gone. The mass conversion is straightforward, 198.11 over 176.12 or a factor of 1.125, but the conversion answers the question about ascorbate ions and no part of the question about pH, which is why Part VIII calls the swap a formulation change rather than a weighing change.

Question 6. What does the Q-long strip, developed for 24 minutes instead of 12, actually test?

• Whether hydroquinone develops more slowly than metol
• Whether hydroquinone at this pH is slow or effectively inactive: a bath that is merely retarded gives visibly more density in twice the time, while one that is inactive gives no more
• Whether the equal time of 12 minutes was long enough for every cell
• Whether fog rises with development time

Show the answer and why

Answer: Whether hydroquinone at this pH is slow or effectively inactive: a bath that is merely retarded gives visibly more density in twice the time, while one that is inactive gives no more

It is Hurter and Driffield’s argument about bromide, borrowed. They showed a restrainer retards rather than stops, because the image appears in full force if sufficient time is allowed, so a single strip at a single time cannot distinguish a suppressed reaction from an absent one. One extra strip turns an impression into a two-point series for the cost of eight minutes in a session that is already running — and it matters here because the whole additivity argument leans on the hydroquinone cell being nearly blank for a reason you can state.

Check my answersClear