Why Kodak Film Sensitivity Depended on Cows and Mustard

The working dose Samuel Sheppard wrote into his patent was 2 to 3 grains of thiosinamine per 100 pounds of dry gelatin, which works out to roughly three or four parts per million. Added to a gelatin stripped of its natural sensitizer, that trace could raise the emulsion's sensitivity by a factor of ten or twenty. Nudge it upward and the film fogged.

One Man's Meat Is Another Man's Poison

Gelatin was not supposed to be an ingredient. It was the carrier, the clear jelly holding silver halide crystals in place on a glass plate or a strip of celluloid, and the properties emulsion makers argued about were mechanical: setting point, melting point, viscosity, jelly strength. Factories built their own apparatus to measure those. One standard rig used a metal disk pressed into a beaker of jelly while mercury was run into a graduated cylinder above it until the surface broke, and a good hard emulsion gelatin was expected to hold about 1,500 grams before it gave way.

None of those numbers predicted speed. F. F. Renwick, who spent years as the chemist at Ilford in London, wrote in 1921 that gelatin, paper, and collodion cotton were the three raw materials demanding the greatest care in selection, and that there were "no generally recognized methods of testing any of these products for photographic purposes except the obvious one of trying them on a small scale and noting the results before proceeding to large-scale use." He added that there seemed little prospect of ever devising a reliable chemical test.

Eight years later, E. J. Wall said the same thing in flatter language in his 1929 survey of emulsion practice: the one reliable test for a gelatin was to make a batch of emulsion with it under working conditions and look at what came out. A delivery one factory rejected as useless could run beautifully in another factory's process. Speed traveled with a supplier and a batch number, not with a formula, and an emulsion maker who found a delivery that worked had no way of describing what made it work.

The Hunt for Gelatin X

The break came from an experiment that reframed the question. R. F. Punnett, working in Eastman's emulsion department, showed that an extract could be prepared from an active gelatin, one that yielded fast emulsions, and that stirring that extract into a relatively inert gelatin made the inert gelatin fast. Whatever was doing the work was a substance carried in the gelatin, not a property of the protein itself. That is a solvable problem.

Sheppard's laboratory then spent years running the raw materials down, stage by stage through the manufacturing chain. What Wall called the mysterious "Gelatine X" was run down to allyl isothiocyanate, allyl mustard oil, present in the active gelatins of the day in amounts somewhere between one part in a million and one part in 300,000. Sheppard published the work in the Photographic Journal in 1925, and again in the British Journal of Photography the same year. Wall, an independent author with no stake in the Kodak laboratory, called the paper the most important contribution to the science and practice of emulsion making yet published.

Part of the reason the answer took so long is that the search needed a blank to work against. Sheppard's laboratory had one: an ash-free gelatin prepared by a purification route he and his colleagues Sweet and Benedict had published in the Journal of the American Chemical Society in September 1922, in a paper nominally about the elasticity of purified gelatin jellies. An emulsion made with that gelatin was close to useless, producing a developable image only after impractically long exposures if at all. Useless emulsion is an excellent instrument. Against a floor that low, a candidate compound either moves the needle or it does not, and the effect of a part per million becomes something you can see on a strip of developed film.

The mechanism Sheppard proposed is still recognizable in modern emulsion chemistry. In an ammoniacal emulsion, the isothiocyanate converts to allyl thiocarbamide, which binds the silver halide and then breaks down in alkali, depositing silver sulfide on the crystal. Those deposits have to be small and discrete. Formed correctly, they are sensitivity centers, sites where a handful of photons is enough to build a developable latent image speck. Formed too generously, they make the grain developable with no exposure at all, which is fog, and Sheppard noted that heavy doses could also drive the emulsion into reversal.

What Sheppard Actually Filed

The paper was the announcement; the patent was the claim, and it came first. Sheppard's US patent 1,574,944 was filed on June 6, 1924, and granted on March 2, 1926, which means the identification was in hand more than a year before the Photographic Journal saw it. It does not claim one compound. It claims a family defined by structure, any compound carrying a divalent atom of the sulfur group double-bonded to a single metalloid atom: thiocarbamides, selenocarbamides, tellurocarbamides. Thiosinamine, the allyl-substituted thiourea, is singled out as especially useful. So is allyl isothiocyanate. So, cheerfully, is telluro-mustard-oil.

The patent reads like a document written by someone who has just turned a mystery into a purchase order. The maximum usable dose, Sheppard writes, is found in any specific case "as readily determined by routine test as, for example, the proportion of sulfur permissible in steel for a particular purpose." Gelatin could now be bought on its physical merits alone, its setting behavior and viscosity and hardness, with any shortfall in sensitizing power made up from a bottle.

The patent also runs the process backward, and that half matters as much as the additive. Sheppard describes treating active photographic gelatins with oxidizing agents to "oxidize the sulfur in the sensitizing substituted (allyl) thiocarbamide or corresponding carbimide, which occurs naturally in photographic gelatins," converting it to something inert. On a factory scale, he recommends folding the oxidation into the liming or deliming steps already part of gelatin manufacture, using calcium hypochlorite or sodium peroxide. Strip the accident out during production, then add back a weighed quantity. A second patent, 1,623,499, filed with Punnett in June 1925 and issued in April 1927, covered the sulfur compounds in combination with reducing agents.

Emulsion makers had spent 40 years relying on ripening, the extended digestion at temperature that grows the silver halide crystals and raises sensitivity by increments nobody could quite predict. Sheppard states in the patent that his compounds act on their own, whether ripening steps are used or not, and can wholly or partially replace them. He preferred to keep them, because the two effects stack. A slow, uncertain, hours-long physical process had acquired a chemical shortcut that could be weighed out on a balance.

Publication flushed out the competition. The Lumière brothers responded that they had already examined roughly 900 candidate substances and were using several commercially, including thiocarbamide and guanidine sulphocyanate, and had simply never published any of it.

Where the Mustard Came From

Allyl isothiocyanate is the compound that makes mustard hot. It forms when plants in the mustard family are crushed and an enzyme meets a stored glucosinolate, and it is the pungent principle in mustard oil, horseradish, and wasabi. Photographic gelatin was rendered from slaughterhouse and tannery stock, cattle hides and bones above all. A sulfur compound characteristic of mustard plants, showing up in a protein rendered from cows, points in one obvious direction.

The trade drew that conclusion and made a joke of it. When Sheppard retired from the Kodak Research Laboratories on January 1, 1948, after 35 years, a note in Nature summarized his best-known work this way: he and his co-workers found that the sensitizing action of gelatin is influenced by certain sulphur compounds, "among the most important of which are derivatives of mustard oil." The note then reported the line that had grown up around the finding, "that if the cow had not a taste for mustard we should have no fast photographic emulsions."

Nature is reporting a saying there, not a feeding trial, and Sheppard's own patent language goes no further than stating that the compound occurs naturally in photographic gelatins. The forage explanation is the standing account of how a mustard-plant chemical ended up in a cow, and it was carried in Nature by people who knew the chemistry, but it reached print as the industry's explanation rather than as a controlled result. Tracing a particular shipment of hides back to a particular pasture is a different order of claim, and the published record does not carry it.

Somebody had blundered into the effect long before anyone could account for it. Wall records that H. J. Newton, working in the dry collodion era, recommended an infusion of tea and mustard seeds as a plate preservative and reported that it gave more sensitivity than tea alone. Mustard seed was also among the raw materials Sheppard worked with in the laboratory.

Speed Stopped Being a Property of a Delivery

The version most photographers carry around is that a film's speed is a physical fact about the emulsion, the way its spectral response or its grain structure is a physical fact. Load Tri-X, meter at 400, and the 400 feels like something inherent, discovered rather than decided. Measuring it was not the hard part, and had not been for a long time: Leon Warnerke had a working sensitometer by 1880, and Hurter and Driffield published their method for determining the sensitiveness of plates in 1890. Specifying it in advance was the hard part. For roughly the first 45 years of gelatin dry plates, speed was a property of a delivery. It arrived on a wagon, it varied between suppliers and between lots from the same supplier, and the chemistry that set it had no name.

What Sheppard changed was the direction of control. Once the active substance was named, a manufacturer could purify the carrier down to a known inert baseline and then dose it, which converts an agricultural variable into a manufacturing one. Kodak took the logic all the way up the supply chain, and by 1930 Eastman Kodak had taken over the American Glue Company's plant in Peabody, Massachusetts, and turned it to making photographic gelatine.

The molecule itself did not stay the answer. As G. Russell reported in Nature in 1956, allyl isothiocyanate was by then not usually present in gelatins made by current methods, and the sensitizers doing the work had been traced by Steigmann and Wood to thiosulfate and polythionates generated during processing. Sheppard had found the natural sensitizer in the gelatins of his own decade. Changed manufacturing left it out of the gelatins that came after, and the sensitizing was being done instead by labile sulfur species thrown off by the processing itself.

What the Number on the Box Is Doing

A modern film speed is a measurement made against a written procedure. Under the ISO method for black-and-white negative film, the sample is developed to a specified contrast, and the speed point is the exposure that produces a density 0.10 above base plus fog. The number printed on the box is the result of that test, not a description of the silver. Sulfur sensitization is still there underneath it, now normally paired with gold, dosed and held within limits that a quality lab checks batch by batch.

That is what Kodak published in 1925. Emulsion makers had been pushing speed up deliberately since the late 1870s, when extended ripening turned gelatin plates into fast plates. What they could not do was account for the share the gelatin itself contributed, which moved with the supplier and with the lot. Sheppard's compounds, as the patent puts it, offered "a novel, dependable and independent way for controlling the properties of photographic emulsions in addition to the customary expedients hitherto available." The largest uncontrolled variable in the process now had a name, a formula weight, and a supplier. Before that, as the trade liked to put it, the fastest plates in the world were fast because a cow somewhere had wandered into the mustard.

Lead image by Detroit Publishing Company, public domain. Source.