Fern’s Photo

🧪 Techniques · ☀️ Exposure · 🤿 deep end · 4 min · updated 19 Oct 2026

Reciprocity failure: the arithmetic of long exposures

Past about a second, many films stop obeying the exposure rule they taught you. Ilford publishes an exponent per film, Kodak a table, Fujifilm a threshold. Here is how to do the sums, with worked examples and a field card.

What film gives you here. Film's response bends at low light, and for many films the bend is published, with testing where the guidance ends. A strange and useful kind of character.

The reciprocity law says that half the light for twice the time gives the same exposure. Film obeys it across the normal range of shutter speeds and then, as the light gets dim and the exposures long, it quietly stops. Ilford's technical sheet describes the effect in one sentence: give the same total exposure over a longer time and the density you get is lower, because the film is less efficient at forming stable development centres at low light levels 1. The fix is more exposure than the meter says, and the makers tell you how much.

Two ways of publishing it

Ilford gives each film a single factor, P, and a formula: corrected time in seconds equals the metered time in seconds raised to the power P. Exposures of one second or less need nothing 1. Their own example is HP5 Plus at a metered 10 seconds: 10 to the power 1.31 is 20.4, so you give 20 seconds 1. The factors in the current sheet are HP5+ 1.31, FP4+ 1.26, Delta 100 1.26, Delta 400 1.41, Delta 3200 1.33, Pan F+ 1.33, SFX 1.43, XP2 1.31, Ortho+ 1.25, Kentmere 100 1.26 and Kentmere 400 1.30 1. The sheet says the film fact sheets carry graphs based on a single factor for all films, and that the per-film factors come from newer work 1, which is why an old datasheet in a drawer may disagree with it.

Kodak prefers a table. For Tri-X, a metered 1 second gets 2 seconds (or one stop wider) with development cut by 10 per cent; a metered 10 seconds gets 50 seconds (or two stops) with development cut 20 per cent; a metered 100 seconds gets 1200 seconds (or three stops) with development cut 30 per cent 2. Foma's table for Fomapan 100 is blunter: multiply 1 second by 2, 10 seconds by 8 and 100 seconds by 16, or open the aperture 1, 3 or 4 stops 4.

Worked examples

Our arithmetic from Ilford's formula, rounded as Ilford suggests. A metered 30 seconds on HP5 Plus: 30 to the 1.31 is about 86 seconds. On FP4+: 30 to the 1.26 is about 73 seconds. On Delta 400: 30 to the 1.41 is about 121 seconds, four times what the meter said. At a metered 100 seconds, HP5 Plus wants around 417 seconds and FP4+ around 331 1. A scientific calculator, or a phone turned sideways, does the x to the y; Ilford's sheet even says so.

The field card

Metered HP5 Plus 1 FP4 Plus 1 Tri-X 2 Fomapan 100 4 Acros II 3
1 s 1 s 1 s 2 s 2 s 1 s
10 s 20 s 18 s 50 s 80 s 10 s
30 s 86 s 73 s between the rows between the rows 30 s
100 s 417 s 331 s 1200 s 1600 s 100 s

The Ilford columns are computed from the published factors; the Kodak and Foma columns are the makers' own rows. Kodak says the table is hard to interpolate and provides graphs for the times in between 2.

Why contrast climbs

Here is the part most guides skip. In a night scene the shadows are receiving far less light than the highlights, so they can lose more to reciprocity failure than the highlights do, and the gap between them grows. Ilford says it directly: contrast may increase with long exposures because of the difference in light levels within the image, and if it does, pulling the development may be required 1. Kodak builds the same idea into its table as the 10, 20 and 30 per cent development cuts 2. So a long-exposure roll may want less development than a daytime roll, depending on the scene's range, and it is worth keeping the two apart so you can decide separately. Any change in development is made in the tank with the usual care: gloves, eye protection, ventilation, and the fixer kept for collection.

The short end

Reciprocity has a short-exposure end too, which matters with electronic flash. Kodak's Tri-X table says that at 1/10,000 of a second you should open up half a stop and develop 15 per cent longer, and that even at 1/1,000 there is no exposure change but development should be 10 per cent longer 2.

The films that barely fail

Fujifilm's Acros II sheet says no compensation is needed for exposures shorter than 120 seconds, and only half a stop from 120 to 1000 seconds 3; that is our reason for calling it the film to load for night and pinhole work. Kodak says Portra 400 needs no filter or exposure correction from 1/10,000 second to 1 second, and asks you to test for anything longer 5. For colour the threshold is the film's own: Kodak asks for tests on Portra 400 beyond one second 5, and when a colour film does fail, its layers can drift apart as well as lose speed.

Doing it in the field

Meter, work out the corrected time, and if it runs past your camera's longest timed speed, switch to bulb with a cable release and count, or use a phone timer. Note the frame, the metered time and the time you gave; when you develop, decide whether the roll is a pull. Then look at the negatives against a window and see whether the shadows came through. If they did not, next time give more; the film's character includes being honest about its limits, and so should your notebook.

Keep going