Most developers reduce exposed silver halide to metallic silver and leave the gelatin otherwise untouched; the negative’s density is purely a function of how much silver was deposited. Staining pyro developers behave differently. As they reduce silver they also deposit a coloured, image-wise dye in the gelatin, so the finished negative is part silver and part stain. The consequence is not cosmetic. Because the stain forms in proportion to the silver, it adds the most density exactly where a negative carries the most silver — in the highlights — and so functions as a contrast-reducing mask that is built into the image rather than applied as a separate step.

The Developing Agents and How Stain Forms

The two staining agents are pyrogallol (1,2,3-trihydroxybenzene) and pyrocatechin, more commonly pyrocatechol (1,2-dihydroxybenzene). Both are polyhydroxybenzenes that reduce exposed silver halide to silver in alkaline solution, and in doing so are themselves oxidised. It is the oxidation products that stain.

Neither agent works alone, which is the first thing the names hide. PMK stands for Pyro-Metol-Kodalk: the bulk of the reducing is done by metol, a fast superadditive primary developer, and the pyrogallol regenerates the metol while contributing the stain. In Sandy King’s Pyrocat-HD the same role falls to phenidone (or, in the substitute formula, metol), paired with pyrocatechin. The polyphenol is the staining partner, not the engine.

The oxidation chemistry is well characterised outside photography. In alkaline solution pyrogallol autoxidises through quinone intermediates to purpurogallin and related coloured products; kinetic studies, such as Abrash’s 1989 work in the International Journal of Chemical Kinetics, describe these end-products and their absorption in the visible. Crucially, the coloured product accumulates at the development sites, in register with the silver, rather than fogging the whole frame — but only because almost no sulphite is present to wash it away. In a conventional developer such as D-76, sodium sulphite dissolves the oxidation products before they can settle as image stain, and the same sulphite acts as a mild silver solvent that erodes grain edges. Stripping the sulphite is the single decision that preserves both the stain and the grain.

A second effect rides along. The oxidation products also harden, or tan, the gelatin locally — the same crosslinking exploited industrially to gel gelatin with polyphenols. Because hardening is heaviest where development is heaviest, it stiffens the boundary between a dense area and an adjacent thin one, restricting the lateral diffusion of fresh developer across that edge. This is an adjacency, or Eberhard-type, edge effect: development is held back just inside the dense side of a boundary and pushed just outside it, sharpening the transition. The tanning the formula is named for and the acutance it is prized for are the same mechanism seen from two ends.

Mixing and Working Dilutions

Both developers are mixed as two stock solutions and combined only at use. Stock A is kept acidic so it keeps on the shelf; the alkali lives in Stock B, and development — with its oxidation and staining — only begins once the two meet water together.

PMK Stock A carries metol 5 g, sodium bisulphite 10 g and pyrogallol 50 g made to 500 ml; the bisulphite is there as an acidifier to stabilise the stock, not as a working preservative. Stock B is sodium metaborate (Kodalk) 300 g per litre — the K in the acronym, the alkali. The standard working dilution is 1:2:100, one part A to two parts B to a hundred of water, used at 20 C / 70 F.

Pyrocat-HD, published by Sandy King on unblinkingeye.com in 2000 as a phenidone-pyrocatechin substitute for PMK, follows the same two-part logic. Stock A: sodium metabisulphite 10 g, pyrocatechin 50 g, phenidone 2 g (or 25 g metol in the substitute) and potassium bromide 1 g per litre. Stock B: potassium carbonate 750 g per litre. It runs at 1:1:100 for normal work, 2:2:100 for expansion, and as dilute as 1:1:200 to 1:1:400 for semi-stand development. Pyrocatechin is far less prone to aerial oxidation than pyrogallol, which is why Pyrocat-HD is the more stable and forgiving of the two to mix and store.

A Worked Example in Zone-System Terms

Take Ilford FP4+ rated at EI 80. In PMK at 1:2:100 and 20 C, ten minutes gives a normal contrast index — an N development. To fit a contrasty subject onto a normal grade you contract to N-1 by cutting the time to eight minutes; to expand a flat subject toward N+1 you extend to about thirteen. The highlight-masking claim becomes measurable here: developed normally, the same negative that would block its Zone VIII and IX values in a non-staining developer instead carries proportionally more stain in those zones, holding their separation without driving the silver density up.

HP5+ at EI 320 follows the same pattern in PMK — roughly thirteen minutes for N, ten for N-1, and as long as twenty-six minutes for a full N+2 expansion. The same FP4+ in Pyrocat-HD at 1:1:100, 70 F, reaches a comparable normal contrast in about eight minutes (HP5+ around thirteen, T-Max 400 around twelve on a rotary processor). Pan F+ at EI 32 needs about nine minutes in PMK, Delta 100 at EI 80 about eleven. These are starting points to be calibrated against your own metering and densitometry, not constants.

Reading and Printing the Stain

The stain is where most pyro problems begin, because it does not read the way silver does. A white-light visual densitometer effectively ignores it; you must read pyro negatives on the blue channel of a colour densitometer for silver papers, or on a UV densitometer for alternative processes such as platinum or kallitype. The wavelength you measure changes the answer dramatically. Because Pyrocat-HD’s brown stain filters UV more strongly than it does blue, the same negative reads as a meaningfully higher printing density range under UV than in blue light — and higher still relative to a white-light reading that all but ignores the stain. The stain that is nearly invisible to your eye is a substantial slice of contrast to a UV-sensitive process, which is exactly why a Pyrocat-HD negative can serve both silver and alternative-process printing.

Stain colour decides how cheaply that density prints. PMK lays down a yellow-green stain; Pyrocat-HD a brown one. Variable-contrast and graded silver papers are most sensitive to blue light, and a brown stain absorbs less blue than a yellow-green stain does, so a Pyrocat-HD negative costs less printing exposure for the same masking effect. On variable-contrast paper the yellow-green PMK stain has a second action: sitting heaviest in the highlights, it selectively holds back the paper’s blue-sensitive high-contrast layer there, softening those tones by grade as well as by density.

One step undoes all of it. Pyro negatives must be fixed in an alkaline fixer — Hutchings recommended Photographers’ Formulary TF-4 — because an acid fixer strips much of the stain on its way out. Follow a pyro process with an ordinary acid rapid fixer and you will lose the very effect you developed for, and wonder why your negatives print thin.

Image Stain, General Stain, and a Dated Correction

The quality metric practitioners actually judge by is the ratio of image stain to general stain. Image stain is proportional: it tracks the silver, sits in the highlights, and does the masking. General stain is overall base-plus-fog colouring the whole frame, including unexposed borders — it adds nothing useful and only costs printing exposure. A good staining developer maximises the first and minimises the second.

That distinction dates a piece of received PMK lore. Hutchings formulated PMK in 1979 and popularised it through View Camera in 1991 and The Book of Pyro the same year; his original procedure had you return the fixed film to the used, oxidised developer for about two minutes to boost the stain. Current consensus is that this after-bath adds mostly general stain — overall fog — rather than proportional image stain, and it is no longer recommended. The correction is itself an illustration of the metric: more colour on the negative is not the goal; more proportional colour is.