Assignment: Time in the Frame, Moving Water, Clouds, People and Traffic

The intention

Make three negatives in which the duration of the exposure, and not the framing, is the decision that produced the picture — and predict, in writing and in numbers, what each moving thing will do before you open the shutter.

Every previous assignment treated exposure time as something you compute in order to get a correct density. Here it is a compositional control with its own arithmetic. A square empties, a waterfall turns to a single mass, a cloud streaks or washes out, a friend appears as a transparent presence you placed on purpose. All four of those are the same calculation run with different numbers, and the calculation is short enough to do on the back of the log sheet.

Learning objectives

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

• Compute the local presence fraction of a moving subject — which is not its time in frame — and convert it into a predicted change in exposure at that patch, in stops.
• Predict the smear length of a moving subject from its speed, its distance and the focal distance, and say at what exposure it stops reading as an object.
• Bracket duration at constant total exposure by changing the pinhole rather than the time.
• State the two reasons the linear presence rule describes the light and not the density, and design a bracket that works despite them.
• Work beside water and traffic without either becoming an incident, and photograph people who will not appear in the picture without that becoming an ethical problem.

Prerequisites

Architecture, Low Viewpoint and the Flat Plane for public-space working and the session log, and Still Life and Controlled Light for the exposure arithmetic and the three-stop scale. Part VI’s exposure and reciprocity lesson and Part IV’s latent-image behaviour own reciprocity failure, which is the limit on everything here.

Safety classification

Level A. No new chemistry. What is new is where you stand: beside moving water, beside a road, or in a place full of people who are moving and you are not.

• Substances. The same three baths indoors, at the same dilutions.
• Energies. The sun for a long session; moving water, which carries more energy than it looks like.
• Procedures. A tripod in a public place; exposures during which you are not looking at the camera.
• Waste. Used fixer is silver-bearing and is collected.

What is not a hazard here, and why. Long exposures are not a hazard in themselves: the camera is inert for four minutes exactly as it is for four seconds, it emits nothing and stores nothing. The paper is no more sensitive to handling than it was in the darkroom. There is no chemical difference whatever between this assignment and the last one. The hazards are all locational, and they arise because the method requires you to stay in one place for a long time and to attend to a timer rather than to your surroundings.

Hazards

Water. These are the course’s own rules for this assignment, stated as conditions rather than as guidance from an authority. Set the tripod on a dry, level surface well back from any edge, not on wet rock, weed-covered stone, a mud bank or a lip that has been undercut. Never wade, and never step onto a rock in a current, for a photograph. For coastal work check the tide before you go and know where the water will be in two hours, because the exposure that starts on dry sand can finish somewhere else. Do not put yourself between water and a wall, or in a gully, where a route back can close. If you are working alone, tell somebody the location and the time you expect to be back.

Roads. The Highway Code’s rule 1 is that pavements and footways should be used where provided, and rule 2 that where there is no pavement you keep to the right-hand side so that you can see oncoming traffic. A camera photographing traffic goes on the pavement side of any barrier, never on a carriageway, a central reservation, a cycle lane or a bridge parapet. Rule 7 is the one that matters when a timer goes off across the road: “When it is safe, go straight across the road — do not run.”

Obstruction. As on the architecture page, section 137(1) of the Highways Act 1980 makes it an offence to wilfully obstruct the free passage along a highway in England and Wales. A four-minute exposure means four minutes of tripod, so leave a clear passing width and stay with it.

Sun. A session built around long exposures is a session spent standing still outdoors. ICNIRP’s shadow rule applies: shadow shorter than you means the ultraviolet is at its strongest, which under clear skies in late spring and summer is the four hours around midday.

Required PPE

• Sun protection — hat, sleeves, sun cream, and shade between sheets — is the protective measure for the exposure half of the session.
• Footwear with grip, for any work near water or on a bank. This is protective equipment on this page in the way gloves are protective equipment in the darkroom.
• Back indoors: nitrile gloves, eye protection and one pair of tongs per tray.

Ventilation

Not among the outdoor controls, for the obvious reason. Indoors, an openable window or an extractor and not a sealed cupboard, ILFORD’s general instruction; none of the three baths generates a vapour that needs extraction.

Materials

Item	Quantity	Notes
Variable-contrast RC paper	12 sheets	Nine planned, three spare.
Three subjects, chosen in advance	3	One water or cloud, one crowd or traffic, one planned figure.
A willing person	1	For the partial-presence figure. See the consent note.
A watch with a second hand, and a separate timer	1 each	The watch times the subject, the timer times the exposure. They are different measurements.
A tape measure or a pacing distance you have calibrated	1	Subject distance is an input to every calculation here.
Prediction sheet, folded	3	Predictions on the left, observations on the right.

Chemicals

Chemical	Quantity	Form
Paper developer concentrate	100 mL	Diluted 1+9 to make 1 L; 1 minute at 20 °C for RC paper.
Stop bath concentrate	50 mL	Diluted 1+19 to make 1 L; about 10 seconds. A citric acid stop.
Rapid fixer concentrate	200 mL	Diluted 1+4 to make 1 L; 30 seconds for RC paper. An ammonium thiosulfate fixer.
Water	about 12 L	Dilution and washing.

Equipment

The camera with its flat back; at least three pinhole plates of different diameters, because the duration bracket on this page is done by changing the hole; the pinhole register; a tripod; a lux meter or a phone app; a timer; and the processing kit from the previous sessions.

Estimated cost

Cost band £. Twelve sheets and a session’s chemistry. Dated prices live in the laboratory planner.

Waste streams

Used fixer is silver-bearing and is collected, with the first change of wash water, into the labelled container Part II established. Local regulations govern, and they differ between authorities even within the United Kingdom.

Preparation

1. The arithmetic of presence

A patch of the negative receives light from whatever was in front of it, for as long as it was there. If a moving subject occupies that patch for a fraction φ of the exposure and the background occupies it for the rest, the total light arriving is simply the sum:

H = φ E_subject T + (1 − φ) E_background T

Exposure at a patch crossed by a moving subject

Divide through by what the bare background would have delivered, and write r = E_subject/E_background for the ratio of their luminances:

H / H_background = φ r + (1 − φ), in stops: log₂(φ r + 1 − φ)

Change at that patch, relative to the bare background

That is the whole of it, and it is exact for the light. Two consequences fall straight out, and both are counter-intuitive enough to be worth exposing a sheet to confirm.

What a moving subject does to the patch it crosses, against the fraction of the exposure it is there

• very dark subject, r = 0.1
• dark subject, r = 0.2
• mid subject, r = 0.5
• bright subject, r = 2
• very bright subject, r = 4

Show the numbers behind this plot

Five curves of exposure change in stops against presence fraction from zero to one, one curve for each ratio of subject luminance to background luminance. All five start at zero change when the fraction is zero, because a subject that is never there changes nothing, and all five are curved rather than straight, because stops are logarithmic while the light adds linearly. For a very dark subject at one tenth the background luminance, the change is only minus nought point one four of a stop at a tenth presence, minus nought point three seven at a quarter, minus nought point eight six at a half, minus one point six two at three quarters, minus two point four at nine tenths, and minus three point three two only when the subject is present for the entire exposure. For a dark subject at one fifth the background, the same fractions give minus nought point one two, minus nought point three two, minus nought point seven four, minus one point three two, minus one point eight four and minus two point three two. For a mid subject at half the background, minus nought point zero seven, minus nought point one nine, minus nought point four two, minus nought point six eight, minus nought point eight six and minus one. For a bright subject at twice the background the curve rises instead, reaching plus nought point one four, plus nought point three two, plus nought point five eight, plus nought point eight one, plus nought point nine three and plus one. For a very bright subject at four times the background it rises to plus nought point three eight, plus nought point eight one, plus one point three two, plus one point seven, plus one point eight nine and plus two. The shape is the lesson. Every curve is flat near the left-hand end and steep near the right, so half the exposure gives far less than half the effect, and a solid rendering of a dark subject needs almost all of the exposure rather than most of it. A shaded band marks the region below about a quarter of a stop, inside which the change is too small to find on a paper negative at all.

Series	Presence fraction φ — share of the exposure the subject occupies that patch	Change in exposure at that patch, stops
very dark subject, r = 0.1	0.00	0.00
very dark subject, r = 0.1	0.05	-0.07
very dark subject, r = 0.1	0.10	-0.14
very dark subject, r = 0.1	0.25	-0.37
very dark subject, r = 0.1	0.50	-0.86
very dark subject, r = 0.1	0.75	-1.62
very dark subject, r = 0.1	0.90	-2.40
very dark subject, r = 0.1	1.00	-3.32
dark subject, r = 0.2	0.00	0.00
dark subject, r = 0.2	0.05	-0.06
dark subject, r = 0.2	0.10	-0.12
dark subject, r = 0.2	0.25	-0.32
dark subject, r = 0.2	0.50	-0.74
dark subject, r = 0.2	0.75	-1.32
dark subject, r = 0.2	0.90	-1.84
dark subject, r = 0.2	1.00	-2.32
mid subject, r = 0.5	0.00	0.00
mid subject, r = 0.5	0.05	-0.04
mid subject, r = 0.5	0.10	-0.07
mid subject, r = 0.5	0.25	-0.19
mid subject, r = 0.5	0.50	-0.42
mid subject, r = 0.5	0.75	-0.68
mid subject, r = 0.5	0.90	-0.86
mid subject, r = 0.5	1.00	-1.00
bright subject, r = 2	0.00	0.00
bright subject, r = 2	0.05	0.07
bright subject, r = 2	0.10	0.14
bright subject, r = 2	0.25	0.32
bright subject, r = 2	0.50	0.58
bright subject, r = 2	0.75	0.81
bright subject, r = 2	0.90	0.93
bright subject, r = 2	1.00	1.00
very bright subject, r = 4	0.00	0.00
very bright subject, r = 4	0.05	0.20
very bright subject, r = 4	0.10	0.38
very bright subject, r = 4	0.25	0.81
very bright subject, r = 4	0.50	1.32
very bright subject, r = 4	0.75	1.70
very bright subject, r = 4	0.90	1.89
very bright subject, r = 4	1.00	2.00

Computed from log₂(φr + 1 − φ); not measured. It describes the light arriving, not the density you finally see — the material's characteristic curve and its reciprocity failure both intervene, which is what the bracket is for. The curve is drawn to show the shape, not measured from a real material. Your own materials will differ, and measuring them is what the sensitometry part of the course is for.

First: half the exposure is nowhere near half the effect. A dark figure at a fifth of the background’s luminance, standing for exactly half a four-minute exposure, takes 0.74 of a stop off that patch. That is a distinct but transparent ghost, not a half-solid person. To make them read as solid you need nine-tenths of the exposure, and even then you get 1.84 stops rather than the 2.32 of standing there the whole time. Ghosts are easy and solid figures are hard, which is the opposite of most people’s expectation.

Second: the fraction that matters is local, not global. A person walking through the frame is in the frame for as long as it takes them to cross it, but they are in front of any given patch only for as long as it takes them to walk their own width.

🧮DO THE MATHS: why a busy square empties

The scene. A camera at f = 40 mm, 3 m from the nearest walkers, on a 100 mm sheet. A person 0.5 m wide walks at 1.2 m/s. The exposure is 60 seconds.

Time in frame. At 3 m the 100 mm sheet covers 100 × 3000/40 = 7.5 m of the scene. Crossing 7.5 m at 1.2 m/s takes 6.25 s, which is 10.4 per cent of the exposure. That is the number people compute, and it is the wrong one.

Local presence. At any single point in the frame, that walker blocks the background only while their own body passes it: 0.5 m at 1.2 m/s = 0.42 s. As a fraction of 60 seconds, φ = 0.0069.

The change at that patch. With r = 0.2, that is log₂(0.0069 × 0.2 + 0.9931) = −0.008 stop. Eight thousandths of a stop, against a material whose whole scale is about three stops. It is not faint. It is absent.

A car is worse. A 4.5 m car at 13.4 m/s (30 mph) passes a point in 0.34 s, so on the same 60 s exposure φ = 0.0056 and the change is again a hundredth of a stop. Moving traffic does not record in daylight, no matter how much of it there is. At night it will, and for a reason that has nothing to do with this arithmetic — its lamps are enormously brighter than the background, so r is not 0.2 but hundreds, and the night assignment picks that up.

And the person who stops. Someone who pauses for 20 s of a 4-minute exposure has φ = 20/240 = 0.083, giving −0.10 stop with r = 0.2. Still invisible. To be seen at all you have to stand still for a serious fraction of the exposure, which is why the ghost has to be arranged rather than hoped for.

2. Smear, and when a thing stops being a thing

Smear geometry: from subject speed to a length on the paper

1. 1Subject travel, v × T — measured in the world, at distance u
2. 2Magnification m = f / u — the same similar triangles that give image size
3. 3Smear on the paper = m v T — a length you can predict, and then measure with a rule
4. 4The threshold — smear matters once m v T exceeds the blur circle b ≈ d

Everything on this page reduces to one multiplication. Predict the smear before the exposure, write it in millimetres, and measure it on the dry negative.

Δ = m v T = (f / u) v T

Smear length on the paper

Subject	u	v	f	Image speed	Smear in 60 s
Walker	3 m	1.2 m/s	40 mm	16 mm/s	960 mm
Car at 30 mph	10 m	13.4 m/s	50 mm	67 mm/s	4,020 mm
Wave or river surface	20 m	1.5 m/s	50 mm	3.75 mm/s	225 mm
Cloud	1,000 m	10 m/s	50 mm	0.5 mm/s	30 mm

Smear starts to matter once Δ exceeds the blur circle, since below that it is hidden inside the softness the hole already imposes. Setting Δ = d = 0.25 mm and solving for T gives the exposure at which each subject stops being a shape: 0.016 s for the walker, 0.004 s for the car, 0.067 s for the water and 0.5 s for the cloud. Every one of those is far shorter than any exposure you can make at f/200 on paper. Nothing that moves is ever rendered as itself in this part. The only question is how far it has smeared, and therefore whether it reads as a streak, as a smooth mass, or as nothing at all.

That is what sets the three subjects apart.

Water at a few tens of metres smears a couple of hundred millimetres in a minute — several sheet widths — so at exposures over about ten seconds a river or a fall becomes a single smooth mass whose tone is the average luminance of the surface over the exposure. What you lose is the texture: the individual crest, which is present at one point for perhaps half a second, has φ = 0.017 in a 30 s exposure and shifts that patch by about a twentieth of a stop. It is gone. What you gain is a tone you can place deliberately against the rocks, and an edge between water and stone that is completely clean because the stone did not move at all.

Cloud is the interesting one because it sits at the boundary. At 0.5 mm/s of image speed a one-minute exposure streaks a cloud 30 mm across a 100 mm sheet, which reads unmistakably as a streak with direction and length. Ten minutes spreads it 300 mm — three sheet widths — and the sky washes to an even tone with no structure at all. The streak-to-wash transition is somewhere around two to four minutes for typical cloud, and it is under your control to within a factor of two, which is about as much control as anything on this page offers.

People and traffic, by the arithmetic above, simply do not appear. The picture of a busy square is a picture of an empty square with the pavement slightly greyed, and that emptiness is the subject.

📓Where the linear rule stops describing what you will see

The presence relation is exact for the light arriving. It is not exact for the density you get, for two reasons this course has already established and has not solved.

The characteristic curve is not a straight line. A change of 0.74 stop near the middle of the material’s response produces a visible density difference; the same 0.74 stop down in the toe produces almost nothing. So a ghost placed against a bright wall and the same ghost placed against a dark doorway are not the same picture, even at the same φ. Part XIII builds the curve that would let you predict this properly.

Nobody publishes a reciprocity correction for paper used as a camera negative. Part VI searched and said so. Worse for this page: reciprocity failure is intensity dependent, so the patch that had a dark figure in front of it for half the exposure was at a lower illuminance for that half, and suffered the failure more. Whether that makes the presence relation measurably non-linear on paper at exposures of minutes, this course does not know, and it will not invent an answer.

So the assignment is designed to find out. Stage 3 below is a controlled test of exactly this question, with a control, and its result is a real measurement of a real quantity that is not in the literature. ILFORD’s related observation about films is the hint to work from: they state that “the contrast is increased with long exposures”, because shadows sit at a lower illuminance than highlights and suffer the failure more. If that holds for paper, the dark-figure ghost should come out fainter than the linear rule predicts, and the bright-figure ghost should come out close to it.

3. Bracketing duration instead of exposure

You cannot bracket duration by changing the time, because changing the time changes the density as well, and then you cannot tell which effect you are looking at. Change the hole instead. Total exposure goes as T/N², so to double the duration while holding the total constant you need N² to double, which means the diameter falls by a factor of √2.

Plate	N at f = 50 mm	Exposure relative to the 0.25 mm plate	In stops	Duration for the same density
0.50 mm	f/100	0.25 ×	−2.00	quarter
0.35 mm	f/143	0.51 ×	−0.97	half
0.25 mm	f/200	1.00 ×	0.00	reference
0.20 mm	f/250	1.56 ×	+0.64	half again
0.15 mm	f/333	2.78 ×	+1.47	nearly triple

So with the four plates you pierced in Part VI you can span roughly eleven to one in duration at constant density — from a quarter to nearly triple the reference. That is the instrument this assignment needs, and it is the reason the camera was built modular.

The cost is a change in sharpness, since the plates are not all at the optimum diameter. Record which plate produced which duration, because a comparison across plates carries a confounded variable, and the honest way to handle a confound is to name it rather than to hope it is small.

4. Write the predictions, then fold the sheet

For each of the three planned exposures, before leaving: the intention; the subject distance u; the speed v, either measured or estimated with a stated method; the intended duration T; the computed smear Δ in millimetres; for anything you want to record, the local presence fraction φ and the predicted change in stops; the computed exposure and the plate that gives the duration you want; and one sentence naming what you expect not to appear.

The making

Stage 1 — The smooth mass (45 minutes)

Water or cloud. Two sheets at durations a factor of four apart, at constant total exposure via the plates — for example the 0.35 mm plate for the short one and the 0.15 mm for the long. Predict the smear length for each in millimetres, and predict which one will read as streaks and which as a wash.

Stage 2 — The square that empties (40 minutes)

The busiest place you can reasonably work in. One sheet, the longest duration your light allows. Predict, in writing, that the people will not appear, and predict the change in stops at a patch they cross. This sheet is a negative prediction, and confirming a negative prediction is a real result.

Stand where you can see the camera, and out of the frame. Count the people who walk through and write the number in the log; it makes the emptiness of the negative far more striking six months later.

Stage 3 — The placed figure, and the linearity test (50 minutes)

Planning a partial-presence figure, and turning it into a test with a control

1. 1Choose the tone you want, in stopsDecide first: a barely-there presence at about −0.3 stop, a clear ghost at about −0.75, or a nearly solid figure at −1.8 or more. That is the target, and it goes in the log before anything else.
2. 2Measure r, the luminance ratioRead the lux reflected from the person and from the background they will stand against, or estimate from their tones against your printed grey scale. r = subject over background. Dark clothes against a pale wall is around 0.2; a white shirt against a dark doorway can be 3 or more.
3. 3Solve for φFrom stops = log₂(φr + 1 − φ), rearranged: φ = (2^stops − 1)/(r − 1). For −0.75 stop with r = 0.2 that is φ = (0.595 − 1)/(0.2 − 1) = 0.51, so the sitter holds for just over half the exposure.
4. 4Convert to seconds and rehearse itφ × T is the hold. On a 4-minute exposure at φ = 0.51 that is 122 seconds. Rehearse the entry and the exit: walk in briskly, settle, hold, walk out briskly, because the walking is itself a presence fraction and you want it small.
5. 5Expose the control on the same sheetHave the sitter stand for the WHOLE of a second exposure at the same total, in the same place. That gives you the φ = 1 point, where the linear rule predicts log₂(r) exactly, and the two sheets together are the test.
6. 6Reconcile, and say which way it missedMeasure both patches against the grey scale. If the partial figure is fainter than the prediction and the full figure is on it, that is evidence that reciprocity failure is bending the relation — the finding this course could not look up.

The control is what turns a portrait trick into a measurement. Without the φ = 1 sheet you cannot tell a non-linear presence relation from a wrong estimate of r.

Consent, and where the rule actually comes from. Article 2(2)(a) of the UK GDPR provides that the Regulation does not apply to “the processing of personal data by an individual in the course of a purely personal or household activity”, so a student’s own photographs made for their own study fall outside it. Data protection law therefore does not supply the rule here, and the course supplies its own, which is not a legal claim and is deliberately stricter than the law:

• The person who holds for the frame agrees in advance, having been told how long the exposure is, how long they must stand still, and that they will appear as a partial presence rather than as a portrait.
• They can stop at any point, and a spoiled sheet is a cheap price.
• Do not photograph identifiable children for these assignments without a parent or carer present and agreeing.
• Passers-by who will not appear in the picture have not consented to anything and do not need to have — but if somebody asks what you are doing, tell them, and if somebody asks not to be included, the answer is yes, even though the arithmetic says they were never going to be.

This is the point at which the ghost picture stops being purely a calculation. A picture of a person made out of their absence is a strong thing, and it is worth being explicit with your sitter about what it will look like before rather than after.

Stage 4 — Process everything together (30 minutes)

Developer 1 minute at 20 °C, stop about 10 seconds, fixer 30 seconds, wash. Process the whole session as one batch. Every conclusion on this page is a comparison between sheets, and a drifting developer turns a comparison into noise.

Stage 5 — Measure, then unfold the predictions (15 minutes)

Measure the smear lengths with a rule on the dry negatives. Judge the presence patches against your printed grey scale. Only then unfold the prediction sheets.

Expected observations

• The water will be smoother than you expected and its tone lighter on the negative than the rocks, because it has been averaging a bright sky reflection for the whole exposure.
• The clouds will either streak or vanish, with very little in between, and the changeover will be somewhere in the low minutes.
• The square will be empty, and the pavement will look slightly grey and slightly odd, which is the sum of thousands of hundredth-of-a-stop contributions.
• The ghost will be fainter than you wanted. Almost everybody’s first placed figure is, and there are three candidate reasons — φ smaller than planned because the hold was shorter than the plan, r misjudged, or the non-linearity this stage is testing for. The log is what separates them.
• The smear predictions will be close, usually within ten or twenty per cent, because Δ = m v T contains nothing chemical. When a smear measurement is badly out, the wrong input is almost always the speed.

Evaluation and reconciliation

This is the assessed part, and it is a table rather than a paragraph.

Element in the frame	Predicted φ	Predicted change, stops	Observed against the grey scale	Difference	Most likely cause
The placed figure
The control figure, φ = 1		1.00 stop is log₂ r
A named passer-by, if you timed one
The water or cloud, as smear length	Δ predicted, mm	—	Δ measured, mm

Fill it in for every moving element you predicted. Then write one paragraph answering the single question the stage-3 control was designed to answer: did the partial figure land where the linear presence rule said it would, relative to the full-presence control? If it came out fainter, say so and say by how much in stops. That number is not in any manufacturer’s literature, and a season of them is worth more than any of the tidy predictions on this page.

The critique

One paragraph per selected negative, following the critique sequence:

1. Intention. What did the duration decision say it would produce?
2. Optics. Blur circle as expected; corners as the cos⁴ figure predicts.
3. Exposure. Which plate, which duration, and what density resulted.
4. Tone. Where did the three-stop scale run out? A smoothed waterfall is a large area of one tone, and it is very easy to put that tone in the wrong place.
5. Time. Predicted against observed presence, in stops, for every moving element.
6. Composition. Did the smear point where you wanted it to? Direction of smear is a compositional line, and it is one you chose when you chose the viewpoint.
7. The one change. An action with a number in it.

Troubleshooting

What you see	Likely cause	What to do
The placed figure is invisible	φ much smaller than planned, or r close to 1	Time the hold with a watch rather than counting; put the sitter against a background of clearly different tone
The figure has a hard outline on one side	They arrived or left slowly, or shifted their feet	Rehearse the entry and exit; a brisk walk in and out keeps the transit’s own φ small
The water is a flat, featureless white	It averaged a bright sky and blocked up	Shorter duration with a bigger hole, or a viewpoint where the water reflects something darker
Clouds gone entirely, sky an even tone	Duration well past the wash threshold	That is a result, not a failure — but record the duration, and take the streak version with a plate two sizes larger
Smear measured much longer than predicted	Speed estimate too low, or the subject was nearer than you paced	Measure u with a tape once per session; pacing drifts
Everything much thinner than the arithmetic says	Reciprocity failure, uncorrected because nobody publishes a paper correction	Bracket two stops either side and log the sheet that worked; that row is a data point in your own curve
The whole sheet fogged after a long exposure	A light leak that is invisible at short exposures	Repeat Part VI’s leak test at the duration you are actually using, not at 30 seconds

Clean-up, storage and disposal considerations

Trays emptied developer, stop, fixer, with the fixer into the labelled silver-bearing container. Negatives dried face-up and sleeved flat, labelled to the log row. Prediction sheets and the reconciliation table are stored with the negative, because on this assignment they are the result.

Disposal. Used developer and stop carry no silver; used fixer does and is collected rather than discharged, because of the silver-thiosulfate complexes it contains. Local regulations govern, and they differ between authorities even within the United Kingdom.

What to submit

• One selected negative and its working positive — whichever of the three best answers its own intention.
• The completed reconciliation table for every moving element you predicted.
• The paragraph on the linearity control, with a number in stops.
• The three folded prediction sheets, unedited.
• One critique paragraph.

Further work

• The presence curve, measured. Five sheets, same sitter, same background, same total exposure, with φ set to 0.1, 0.25, 0.5, 0.75 and 1.0. Judge each against the grey scale and plot the result over the computed curve for your measured r. That is the experiment this page could only sketch, and it takes one afternoon.
• A duration series on one subject. The same waterfall at 5 s, 20 s, 80 s and 320 s, at constant total exposure through four plates. The point at which texture disappears is a number about that water, that distance and that focal distance, and it will transfer to nothing else — which is why it is worth having for your own sites.
• The development question. Kodak publish, for TRI-X, development reductions of 10, 20 and 30 per cent at metered exposures of 1, 10 and 100 seconds, to hold back the contrast rise that long exposures bring. Nobody publishes the equivalent for paper. Try 20 per cent less development on one of a matched pair of long-exposure sheets and see whether the highlight that blocked comes back.