Low light and fast subjects often demand a shutter speed the metered film speed cannot supply. The common response is to rate Kodak Tri-X 400 (400TX) at a higher exposure index and extend development to compensate. “Pushing to EI 1600” is the two-stop version: you meter the film as though it were four times more sensitive than its nominal ISO 400, then develop it longer. Understanding what this trade buys, and what it costs, means separating two things the camera collapses into one act — exposure, which is fixed at the moment the shutter opens, and development, which you control afterwards.

What ISO Speed Actually Measures

A film’s true speed is set by its emulsion and defined by a sensitometric measurement, not by the dial on your meter. Under ISO 6:1993, the speed point m is the place on the characteristic curve where density first rises 0.10 above base-plus-fog. Development is then adjusted so that a second point, 1.30 in log-exposure further along, sits 0.80 in density above m — an average gradient of about 0.80 ÷ 1.30 = 0.615. The arithmetic speed is S = 0.8 ÷ Hₘ, where Hₘ is the exposure in lux-seconds at the speed point.

Rating Tri-X at EI 1600 changes none of this. It simply delivers two stops less light to every part of the frame. The exposure you intended for the speed point now lands two stops below it, well down on the toe, near base-plus-fog. No development scheme moves the speed point — the published push times don’t recover the rated speed at the shadow; they only build density higher up the curve. The “speed” you gain at EI 1600 is a practical evaluation, not foot speed. Kodak and Ilford both say so plainly in their datasheets.

Expose for the Shadows: A Worked Example

Ansel Adams’ rule in The Negative (1981) — expose for the shadows, develop for the highlights — becomes concrete in the Zone System. Place an important textured shadow on Zone III, an important bright detail on Zone VII or VIII, and use development to set where the highlights land.

Take a dim interior. You meter a shadowed wall you want to read as a textured Zone III, and a bright window you want to hold near Zone VII. At box speed EI 400 you place that wall on Zone III and the negative records it cleanly on the straight line. Rate the film at EI 1600 and you give that same wall two stops less exposure: Zone III drops to Zone I, onto the toe near base-plus-fog, where adjacent tones no longer separate. The texture is gone before the developer ever touches the film.

Meanwhile the window is being asked to climb. Push development is, in Zone System terms, forced N+2 expansion applied to a scene of already normal contrast — the kind of expansion you would reserve for a flat subject, now imposed on everything. Expansion raises high values: a window placed on Zone VII is pushed up toward Zone VIII and IX, against the shoulder. That is the trap of the push. The shadows fall off the bottom while the highlights are driven over the top, and detail survives only in a shrinking middle band.

What the Datasheet Actually Says

Kodak’s numbers come from Tri-X 320 and 400 Films, Technical Data F-4017, May 2007 edition. Normal EI 400 development in D-76 stock is 6¾ minutes at 68°F (20°C) in a small tank — not the “roughly seven minutes” of folklore. The full small-tank push table, at 68°F with agitation at 30-second intervals:

• EI 1600 (2-stop): D-76 stock 9½ min; D-76 1:1 13¼ min; HC-110 dilution B 6 min; T-MAX 8¾ min; T-MAX RS 7¾ min; XTOL 9¾ min; XTOL 1:1 13¼ min.
• EI 3200 (3-stop): D-76 stock 11 min; D-76 1:1 16 min; T-MAX RS 9½ min; XTOL 11½ min. Kodak explicitly states 400TX can be underexposed by three stops and pushed.

The contrast cost is quantified, not just asserted. F-4017 gives the contrast index target for normal Tri-X development as 0.56. The push times above all develop past that figure — that is the precise meaning of “increased contrast.” Extending development steepens the curve because it acts preferentially on the heavily exposed regions: highlight densities climb while the toe, already starved of exposure, barely moves.

Choosing a Developer

D-76 is not the only option, and for a push it may not be the best one. The choice splits into general-purpose developers (D-76, HC-110, XTOL) and speed-increasing developers built around phenidone (Microphen, Acufine, and Ilford’s DD-X). A phenidone–hydroquinone superadditive developer can buy a genuine fraction of a stop of shadow speed by rendering developable grains that a metol developer would leave untouched; D-76 in stock dilution mostly just adds contrast.

Ilford’s own HP5 Plus datasheet leans the same way: for maximum film speed at EI 3200 it names Microphen as the powder developer of choice (and DD-X as the liquid). If shadow separation is what you are trying to protect, that points you at a speed-increasing developer rather than a contrast-building one. Reach for HC-110 dilution B (6 minutes at EI 1600) only when you want the convenience and the steep result, not the most detail.

Grain, and Why Push Magnifies It

Tri-X’s diffuse RMS granularity is 17, rated “fine” — measured at a net diffuse density of 1.0 through a 48-micrometre aperture at 12× magnification, in HC-110 dilution B at 68°F. For context, Tri-X 320 reads 16, and modern tabular-grain 400 films sit lower still; the number is meaningful only against a comparison. A push drives it upward.

The mechanism sits at the grain. A silver-halide crystal becomes developable only once a stable latent-image speck forms at a sensitivity site — by the Gurney–Mott model, roughly four or more silver atoms clustered together. Grains that caught too few photons never reach that threshold, which is exactly why push processing cannot recover shadow detail: the latent image was never written. Extended development then works on the grains that did form a speck, reducing them more completely, growing filamentary silver, and overlapping adjacent developed clumps into larger structures. More development, more complete reduction, coarser grain.

Practical Controls and the Highlight Problem

Agitation is a contrast lever, not a ritual. Kodak’s F-4017 scheme is 5 to 7 inversion cycles in the first 5 seconds, then a repeat every 30 seconds; the datasheet warns that tank times under 5 minutes risk uneven development. More vigorous agitation brings fresh developer to the highlights more often, raising both contrast and grain — restraint here partly counters the push’s worst tendencies.

Low light forces long exposures, and Tri-X reciprocity failure then compounds the shadow loss already described. F-4017’s corrections: at an indicated 1 second, add 1 stop and cut development 10%; at 10 seconds, add 2 stops and cut 20%; at 100 seconds, add 3 stops and cut 30%. The development cuts matter on a push, where you are extending time in the first place.

The practical ceiling is set by the highlights. As bright values pile against the shoulder they “block up” into a single featureless maximum density. The cure is at the metering stage and in the development: keep speculars and light sources off the shoulder by metering for them deliberately, and consider reduced or compensating development — dilute, less-agitated, or two-bath schemes — to tame the highlight pile-up at the cost of some contrast you would otherwise add.

Tri-X in Context

Tri-X is not the only fast film that pushes. Ilford HP5 Plus carries an official ceiling of EI 3200 in DD-X, Ilfotec HC, Microphen, or Ilfotec RT Rapid, with Microphen named alongside DD-X as Ilford’s pick for maximum film speed. At EI 1600, HP5 Plus runs 13 minutes in DD-X (1+4) at 20°C, or 14 minutes in ID-11 stock (Ilford’s D-76 equivalent); at EI 3200, 20 minutes in DD-X. The figures land near Tri-X’s, but the films do not look alike under the strain, and the developer that protects HP5’s shadows is named for a reason. Push processing extends the conditions under which a frame can be made at all — but on either film it trades the ends of the tonal scale for a workable middle.