A handheld or in-camera meter can fail in ways that are easy to miss: a flat battery, a backlit frame that fools a reflected reading, or a scene dominated by snow or dark foliage. The Sunny 16 rule provides an exposure estimate that depends on nothing but the weather, which makes it both a fallback when no meter is available and a way to confirm that a metered value is plausible.

Where the rule comes from

The rule predates the cheap, reliable light meter. Through the mid-twentieth century, Kodak, Ilford and Fuji printed exposure guides on the inside flaps and data slips of their film boxes so that an amateur without a meter could still come home with a printable negative. There is no single inventor; it is a manufacturer-disseminated rule of thumb that survived because the physics underneath it is dependable.

Direct midday sun is repeatable because the illuminance reaching the ground in clear conditions sits around 10,000 foot-candles, near enough constant from one sunny day to the next. In Exposure Value terms, that anchors the scene at EV 15 at ISO 100, which is exactly f/16 at 1/125 second. The classic Sunny 16 phrasing of 1/100 second at f/16 is therefore about EV 14.67, a third of a stop more generous than EV 15 because 1/100 is a little slower than 1/125. That extra third of a stop is harmless headroom rather than an error: it builds in a small allowance for the haze and atmosphere that take the real ground-level figure below the textbook 10,000 foot-candles.

That constancy only holds while the sun is high. Within roughly two to three hours either side of solar noon the figure is stable; beyond that the light reddens and weakens, the shadows lengthen, and the base setting no longer applies.

The base setting and its geometry

On a clear day a frontlit subject is correctly exposed at f/16 with a shutter speed equal to the reciprocal of the film’s ISO speed. Frontlit means the sun is behind you, within about 45 degrees of the line to your subject, throwing hard, distinct shadows you can see falling away from the camera. Those crisp shadows are the visual confirmation that the f/16 base applies.

Take Ilford FP4 Plus, rated ISO 125/22. The base sunny setting is 1/125 second at f/16, which is EV 15. Because exposure is the product of aperture area and time, every whole-stop trade holds the total light constant: f/11 at 1/250, f/8 at 1/500, f/5.6 at 1/1000. Open the aperture a stop, halve the time; the negative receives the same light each way. HP5 Plus at EI 400 starts one place along that ladder: 1/500 second at f/16, then f/11 at 1/1000, and so on.

The rule works on incident light, the light falling on the subject, not the reflected light an in-camera meter reads. Because it ignores subject tone, it is not pulled off course by a white wall or a black coat the way a reflected reading is.

The full condition ladder

The standard exposure guide opens the aperture one stop at a time as the light softens, judged by the shadows the subject casts, while shutter speed and ISO stay fixed:

• f/22 — snow or sand, dark and sharp-edged shadows
• f/16 — sunny, distinct shadows
• f/11 — slight overcast, soft-edged shadows
• f/8 — overcast, barely visible shadows
• f/5.6 — heavy overcast, no shadows
• f/4 — open shade or sunset, no shadows

The f/22 snow-and-sand row closes the first loop opened above. You stop down not because the sun is brighter but because the highly reflective field raises overall scene luminance. This is also where a reflected meter betrays you: pointed at snow, it tries to render it middle grey and underexposes by about two stops, so a reflected reading on snow needs +1 to +2 EV of compensation to match what the incident-based rule gives you directly.

Backlighting closes the second loop: add one stop. A backlit subject sits in its own shadow facing the camera, so opening up by a stop restores the detail a reflected reading would have crushed.

Where it breaks: shade, dusk and reciprocity

The bright end of the ladder is trustworthy partly because the exposures are short. Film stops behaving linearly once exposures grow long, a phenomenon called reciprocity failure: the emulsion needs proportionally more light than the metered time suggests. Ilford correct for this with Tc = Tm^p, where the metered time Tm in seconds is raised to a film-specific exponent. For HP5 Plus p = 1.31; for FP4 Plus p = 1.26. The correction only matters from about 1 second upward. A metered 10 seconds on HP5 Plus becomes roughly 10^1.31, around 20 seconds of real exposure.

In full sun your times are 1/125 or 1/500, far from that threshold, so the estimate stays honest. But the f/4 open-shade and sunset end of the table is exactly where times can stretch toward and past a second, and there the raw Sunny 16 figure can fall short. The rule does not warn you; you have to know to apply the reciprocity correction yourself.

The Looney 11 sibling

The same reciprocal-of-ISO logic generalises to any directly sunlit subject. A full Moon photographed from Earth is lit by that same sun, and it is correctly exposed at f/11 with the shutter set to the reciprocal of the ISO: f/11 at 1/100 second on ISO 100 film. This is the Looney 11 rule, and it is a useful reminder that what Sunny 16 really measures is sunlight on a surface, not the time of day on Earth.

Using it as a sanity check

Held against a meter, the rule flags gross errors with a concrete number. Suppose you are shooting FP4 Plus in open front-lit sun, where Sunny 16 expects f/16 at 1/125, but the meter reads f/4. That is a four-stop gap, far outside the rule’s tolerance, and it almost always means a fooled reflected reading off a dark subject, a wrong ISO on the dial, or a 2x or 4x ND filter still on the lens, rather than a genuine exposure.

As a back-of-envelope reference the rule lands within about a stop, the residual coming from atmosphere, sun angle, season and latitude. That tolerance is comfortable on a forgiving film. Ilford rate HP5 Plus at ISO 400/27 but state it can be metered anywhere from EI 400 to EI 3200 with appropriate development, and in practice it holds usable detail from roughly one stop under to two or more stops over box speed. A meterless estimate that is right to within a stop sits well inside that margin.