The Colour That Would Not Wash Out

ADyeing is among the oldest of the chemical industries, and for most of its history it was practised without any theory of why it worked. Fragments of madder-dyed linen from the Nile valley have been dated to the third millennium BC, and the plant sources of the classical palette, madder for red, weld for yellow and woad for blue, were in use across Europe long before anyone could say what a dye was. The difficulty facing early dyers was not obtaining colour, which a boiled plant will yield readily enough, but making it stay. A pigment that rinses away with the first wash is a stain rather than a dye, and the distinction between the two organised the entire trade. Guild statutes in fourteenth-century Florence and Bruges accordingly graded dyers not by the shades they could produce but by the materials they were licensed to fix colour with, and prosecuted those who passed off fugitive colours as permanent ones.

BThe most valuable of the ancient dyes came not from a plant but from a sea snail. Tyrian purple was drawn from a gland of several Murex species, and the yield was famously poor: reconstruction experiments in the 1980s suggested that something over ten thousand molluscs were consumed for a single gram of dyestuff. Because the compound responsible for the colour contains bromine, and no plant is known to make it, its presence in an archaeological textile has long been treated as a signature that cannot be forged. Anneke Rijs, a dye chemist at the Leiden Institute for Materials History, does not dispute the chemistry but questions the inference historians draw from it. Of the forty-one purple-toned fragments her laboratory examined between 2016 and 2023, only nine returned the brominated marker; the remainder had been coloured with madder laid over an iron-treated ground, a far cheaper route that ages into a similar hue.

CHer claim is not that ancient purple was cheap, since the written sources make its costliness plain, but that the frequency with which it is reported in the archaeological literature measures the sensitivity of the eye rather than of the instrument. Rijs is unusually explicit about what would refute her: a body of securely provenanced fragments in which attribution by eye and chromatographic analysis agree in more than four cases out of five. She also enters a caution that is often dropped when she is quoted. Her fragments came almost entirely from museum collections in temperate Europe, where burial conditions favour the survival of some compounds over others, and the bias this introduces has never been quantified. What she objects to is the habit of presenting a colour judged by eye as though it had been measured; of the twenty-six regional surveys her group reviewed, seventeen called a textile Tyrian without any chemical test at all.

DCormac Fenwick, an economic historian at Dunmore College, treats the argument about identification as secondary to a question of supply. Colour, in his account, was never limited by knowledge of dyestuffs but by access to mordants, the metallic salts that bind a dye to a fibre and without which most plant colours are fugitive. Salts of the necessary purity came from a handful of deposits, and when a substantial one was opened at Tolfa in the Papal States in 1461 the price in northern markets fell by roughly a third within a decade. Fenwick's maps of licensed dyehouses show them clustering along the routes by which those salts travelled rather than around the fields where dye plants grew. He accepts that his evidence is fiscal rather than technical, that customs registers record what was taxed rather than what was used, and that no series he has assembled runs unbroken for more than sixty years.

EThe synthetic turn is conventionally dated to 1856, when William Perkin, an eighteen-year-old student in London trying to synthesise quinine, precipitated instead a purple residue from coal-tar aniline and patented it as mauveine. What followed was not swift. The decisive commercial event came four decades later, when a German firm brought a synthetic indigo to market in 1897 after seventeen years of development and an outlay it afterwards put at eighteen million marks, more than its share capital at the outset. The effect on cultivation was abrupt: the area under indigo in Bengal, above 1.5 million acres in 1895, had fallen below one hundred thousand by 1913. Fenwick attributes the collapse to price and to the uniformity that mechanised printing demanded, since a vat of plant indigo varied from batch to batch. Rijs replies that the manufacturers' own laboratory books record the contest being won on fastness to light and to washing, tested and tabulated, before the synthetic had become the cheaper of the two.

FA more careful framing separates two questions that the debate tends to fuse. One concerns what a given textile was actually dyed with, which only analysis can settle; the other concerns why one dyestuff displaced another, which analysis cannot settle at all. Fenwick's registers and Rijs's chromatograms answer different questions, and neither converts into the other, which is why the current revival of natural dyeing is being argued about on terms that neither researcher supplies. What remains genuinely unresolved is whether reference collections are adequate, since identification depends on matching a sample against authentic dyed standards, and the standards now in circulation were themselves prepared in the twentieth century from cultivars that may differ from the historic ones. Both researchers agree that the next advance will come from the tedious business of building better reference material rather than from any new instrument, and both dislike the habit of announcing an identification before the comparison has been made.