Load the same roll of Kodak T-MAX 100 into two tanks and you can make it look like two different films. Develop one in Kodak D-76 diluted 1+1 and the grain stays tight and the edges soften; develop the other in Rodinal 1+50 or in Geoffrey Crawley’s FX-1 and the grain coarsens visibly while the fine detail snaps into crisper relief. The emulsion is identical. What changes is the chemistry working on it, and that single comparison is the whole subject in miniature: the same lever that suppresses grain tends to blunt edges, and the lever that sharpens edges tends to bring grain forward. This holds even for a film as fine as T-MAX 100, which Kodak rates at an RMS granularity of 8 and classes as extremely fine.

What Grain Physically Is

A black-and-white emulsion is a suspension of silver halide crystals in gelatin. In a conventional cubic or octahedral emulsion these crystals run from a few tenths of a micrometre up to a couple of micrometres across. Exposure renders some of them developable, and the developer reduces each entire crystal to metallic silver. A single silver particle is far too small to see in any normal print. What you read as grain is not one crystal but the clumping of many developed particles into irregular aggregates, separated by clear gelatin where no crystal was reduced, the eye reading that random distribution of opaque clusters against the transparent base as texture.

Tabular-grain emulsions rearrange this geometry. Kodak introduced T-GRAIN with the T-MAX films in 1986: thin, flat crystals roughly 0.2 to 1 micrometre across with a high aspect ratio, lying flat in the coating. Because they scatter less light sideways through the emulsion, they resolve more detail for a given speed, which is why T-MAX 100 reaches 63 lines/mm at a test-object contrast of 1.6:1 and 200 lines/mm at 1000:1 when developed in D-76 at 20C. Ilford’s Delta 100 uses a comparable core-shell tabular crystal for the same reason. Tabular crystals also respond differently to solvent developers than thick conventional ones, since there is far less crystal volume for a solvent to etch away.

Granularity Is Not Graininess

It is worth keeping two words apart. Granularity is a measured property of the film: the root-mean-square fluctuation in density read by a microdensitometer through a 48-micrometre circular aperture, on an area developed to a net diffuse density of 1.0, at 12X magnification. Graininess is the subjective texture an observer actually sees at a given enlargement. The two are linked by Selwyn’s law, which holds that for a not-too-small aperture, RMS granularity multiplied by the square root of the aperture area is roughly constant. That is precisely why the 48-micrometre aperture must be quoted alongside any figure: change the aperture and you change the number, so a granularity value with no stated conditions means nothing.

Kodak’s published numbers give the scale a shape. T-MAX 100 reads 8, extremely fine; Tri-X 400 reads 17, which Kodak still classes as fine. Both are diffuse RMS values measured at density 1.0 through the 48-micrometre aperture at 12X, so they are directly comparable to each other. They are not comparable across brands: Ilford does not publish RMS figures for FP4 Plus, HP5 Plus or Delta 100, describing their grain only qualitatively. For consumer films Kodak itself moved to the Print Grain Index, a perceptual metric read under diffuse enlarger illumination on a uniform scale where a change of 4 units is a just-noticeable difference for ninety per cent of observers and a value around 25 marks the visual threshold of graininess. PGI numbers cannot be set against RMS granularity directly.

The Silver-Solvent Lever

The most direct chemical control over graininess is the silver-solvent action of sulphite. D-76, the Kodak standard since 1927, carries 100 grams of anhydrous sodium sulphite per litre, alongside 2 grams of metol, 5 grams of hydroquinone and 2 grams of borax. At that concentration the sulphite dissolves the outermost layers of the halide crystals and of the developing silver, etching the clumps smaller and smoothing their edges. That is the fine, slightly soft grain D-76 is known for. Dilute it 1+1 and, in Kodak’s own words, you get a sharper negative with slightly more grain, because weaker sulphite cannot etch the grain edges as hard and so preserves the contrast between clumps.

High-acutance developers take the same idea to its limit. Crawley specified that FX-1 hold its sulphite below 6 grams per litre, just 5 grams in the working solution against D-76’s 100, with metol at 0.5 grams, sodium carbonate at 2.5 grams, and a trace of potassium iodide; he warned that more sulphite would regenerate the developing agent and erase the definition he was after, though dropping below about 4 grams hurts keeping properties. Willi Beutler’s original formula, which FX-1 descends from, works on the same low-sulphite principle. Rodinal, the old Agfa p-aminophenol developer now made by Adox as Adonal to the 2005 recipe, carries its sulphite mostly as a preservative rather than as a solvent, and at the high dilutions it is used at the silver-solvent action all but disappears. The crystals develop with little etching, and the grain reads as distinct, hard-edged clusters, more so the more you dilute it.

Edge Effects and the Mackie Line

Perceived sharpness is largely a matter of how abruptly density changes across an edge in the negative, and that is governed by adjacency effects rather than resolving power alone. Where a densely exposed area meets a lightly exposed one, the developer in the dense area exhausts locally and accumulates restraining bromide. That bromide diffuses sideways into the adjacent thin area and suppresses development there, while fresh developer migrates the other way and pushes the dense side darker still. The result is a dark-then-light border straddling the edge, the Mackie line, which the eye reads as extra crispness.

Solvent sulphite blunts these boundaries at the same time as it dissolves grain, which is the structural reason a high-solvent developer measures finer yet looks softer. A low-solvent developer with sparse agitation does the opposite, and you can exploit it deliberately: Rodinal at 1+100, semi-stand or stand developed with only a token agitation at the start, lets the bromide accumulate and broadens the edge band into pronounced halos. Standard agitation runs the other way: an initial burst, then five to seven inversions every 30 seconds, drives fuller development and larger, more contrasty clumps. Restrained or stand agitation lowers graininess and strengthens edge effects, at a real cost in effective film speed.

Choosing Across the Whole Chain

The trade-off is decided as much by format as by chemistry, because enlargement multiplies grain. A 35mm frame at 24x36mm needs roughly 7 to 8 times linear magnification to fill a 10x8 inch print; a 6x6cm frame needs about 3.5 times; a 4x5 sheet only about 2 times. Grain that is intrusive from 35mm developed in Rodinal can be invisible from 4x5, which means the larger format lets you reach for the sharper, coarser-grained developer with no penalty. On 35mm you weigh the same negative in D-76 1+1, finer and softer, against FX-1 or Rodinal, grainier and crisper, and the right answer follows from how far you intend to enlarge.

The numbers here come from Kodak’s own datasheets, F-4016 for T-MAX 100 and F-4017 for Tri-X 400, and from Kodak technical publication E-58 on the Print Grain Index; the high-acutance formulae from Geoffrey Crawley’s FX-1 and Willi Beutler’s original developer. For the sensitometry behind granularity, acutance and edge effects, Anchell and Troop’s The Film Developing Cookbook and Ansel Adams’ The Negative remain the standard references.