The Cost of Colour

Open any design tool and the colour picker offers you every colour at the same price, which is nothing. Move the cursor a centimetre and the most expensive pigment in the history of European painting arrives, free, instantly, in unlimited quantity.

That condition is about two hundred years old. For most of the history of picture-making, colour had a price, the price varied enormously between colours, and because it did, the palette worked as a hierarchy before anyone decided what the picture was about. I have been thinking about what we lost when that stopped being true, and about the costs that did not disappear so much as move somewhere harder to see.

When blue was a budget line

Lapis lazuli. The source of natural ultramarine, and for centuries the reason blue cost what it cost. Photo by James St. John, CC BY 2.0, via Wikimedia Commons.

Ultramarine means, literally, beyond the sea. It was named for where it came from rather than what it looked like, which tells you which fact mattered to the people buying it.

It was ground from lapis lazuli, and the lapis came from the Sar-e-Sang deposit in Badakhshan, in what is now Afghanistan. That material had been moving west into Mesopotamia and Egypt since the third millennium BCE. By the Renaissance the finished pigment was, at times, valued alongside gold, and the reason is not only distance. Extracting usable blue from the stone required a long process of grinding, kneading and washing that discarded most of the material, so the pigment ran to roughly ten times the price of the rock it started as.

Titian, Bacchus and Ariadne, 1520–1523. National Gallery, London. The sky is natural ultramarine, and it is the most expensive decision in the painting. Public domain, via Wikimedia Commons.

This is why the Virgin's robe is blue. Not primarily because blue signified heaven, although the iconography was available and got used. Because contracts specified it. A patron paying for a panel would name the pigment and the quantity, and the most costly material on the account went onto the most important figure in the picture.

Read that as an interface and it is remarkable. The visual hierarchy of the painting was underwritten by a procurement decision. Importance was expressed in a currency the viewer could not see but could reliably infer, because everybody knew what blue cost. Devotion and the invoice pointed at the same figure.

1824: a prize that ends a thousand-year constraint

In 1824 the Société d'Encouragement pour l'Industrie Nationale offered six thousand francs to anyone who could make a synthetic ultramarine that cost no more than three hundred francs per kilogram.

Jean-Baptiste Guimet arrived at a process in 1826 and kept it secret. The prize was awarded to him on 4 February 1828. That same year Christian Gottlob Gmelin, professor of chemistry at Tübingen, reached a method independently and, crucially, published it. A priority dispute followed and has never entirely settled: France upheld Guimet's claim to the money, while Gmelin is generally credited with starting the industry, because publishing is what starts industries.

Inside a decade the most expensive colour in European painting was a manufactured commodity. What had constrained artists for six hundred years stopped constraining them, and it stopped very quickly.

Table 1. The end of expensive blue, compressed into four years.

I do not think this was a loss for painting. It was obviously a gain. But something did go, and it went without being noticed, because nobody misses a constraint while it is being removed.

When every colour costs the same, colour can no longer signal importance by itself. Hierarchy stops being inherited and has to be authored.

That is the condition we work in now, and it is why colour systems need rules rather than taste. A palette in which every swatch is equally available carries no ranking of its own. Whatever ranking it appears to have is one somebody wrote down, and if nobody wrote it down, there is not one.

The costs that did not go away

Cheap colour was not free. It was priced somewhere else, and for a long stretch of the nineteenth century it was priced in the health of the people who applied it.

Carl Wilhelm Scheele produced a copper arsenite green in 1775 that was brighter and more saturated than anything then available, and it sold enormously: wallpapers, dress fabric, artificial flowers, toys, sweet wrappers. It also contained arsenic.

Matilda Scheurer was nineteen years old when she died in 1861. Her job was dusting artificial foliage with green powder. Her death was reported widely, and the reporting is a large part of why the danger became public knowledge at all.

A specimen page from Robert M. Kedzie, Shadows from the Walls of Death, 1874. The green is the actual arsenical wallpaper, bound into the book. Public domain, National Library of Medicine, via Wikimedia Commons.

The object above deserves its own paragraph, because it is one of the strangest pieces of information design I know of. In 1874 Robert Kedzie, a physician in Michigan, wanted to warn the public about arsenical wallpaper. Description was not working. So he mounted actual samples of the poisonous papers onto pages, bound them into a volume, and sent copies to public libraries around the state.

The warning and the hazard are the same object. To make the danger legible he had to hand people the danger. Libraries were instructed to keep it away from children, several destroyed their copies, and the surviving volumes are handled today with gloves and, in some collections, sealed. It is the most literal possible answer to the question of how you make an abstract risk feel real, and the answer turns out to have a cost of its own.

The last wallpaper brands using the green stopped in the 1890s. The compound continued as an insecticide into the 1930s.

One correction, because this one travels further than the truth

You will often read that Napoleon was killed by the arsenic in his green wallpaper at Longwood. He did live in rooms with arsenical green, and arsenic has been detected in his hair. The theory nonetheless does not survive later examination, and the cause of death is stomach cancer. I believed the wallpaper story for years before checking it, which is exactly why I am writing it down.

Mummy brown, and a tube buried in a garden

Some of the cost was not in the application. It was in the sourcing.

Mummy brown was a brown pigment made from Egyptian mummified remains, ground and combined with white pitch and myrrh. Human and animal mummies were both used, and so, at times, were the Guanche mummies of the Canary Islands. The earliest recorded seller is a Paris shop of 1712 trading under the name À la momie. Demand fell away by about 1915.

In 1881, at lunch at Edward Burne-Jones's house, the pigment came up in conversation. Burne-Jones took it for a colour name. Lawrence Alma-Tadema, who was present, told him he had personally seen a mummy being ground at a colourman's. Burne-Jones went upstairs, came back down with his tube, and said it should be given a decent burial. Rudyard Kipling, then about sixteen, was there and later wrote that they all went out and helped, and that to this day he could drive a spade within a foot of where that tube lies. Georgiana Burne-Jones recorded the episode in her 1906 Memorials.

Edward Burne-Jones, The Last Sleep of Arthur in Avalon. Museo de Arte de Ponce. Among the works in which mummy brown has been suspected, on the basis of colour behaviour rather than analysis. Public domain, via Wikimedia Commons.

The trade's own ending is documented, and the wording is better than anything I could write. In October 1964 TIME ran a short piece on the pigment's disappearance, quoting Geoffrey Roberson-Park, managing director of the London colourmen C. Roberson & Co.:

We might have a few odd limbs lying around somewhere, but not enough to make any more paint. We sold our last complete mummy some years ago for, I think, £3. Perhaps we shouldn't have. We certainly can't get any more.

Three pounds, and a note of regret about the inventory.

Delacroix's Liberty Leading the People appears on lists of works in which the pigment is suspected, as does the Burne-Jones above. I want to be precise about the status of that claim: it rests on the behaviour of the colour, not on analysis of those specific canvases. Suspected is the correct word, and shortening it to "Delacroix painted with mummies" is how a hedge becomes a falsehood in one step.

Eugène Delacroix, Liberty Leading the People, 1830. Musée du Louvre. Public domain, via Wikimedia Commons.

Indian yellow, or how to be wrong correctly

This one is here because of its method, not its colour.

In 1883 T. N. Mukharji, working at the request of the botanist Joseph Hooker, reported from Monghyr in Bihar that a transparent yellow pigment was produced from the urine of cows fed exclusively on mango leaves, collected, and reduced over heat. The account became standard and was repeated for over a century.

In 2002 Victoria Finlay examined the story properly and found the evidence thinner than its reputation. The printed source was a single letter. She found no local oral testimony for the practice. And a frequently cited 1908 ban on the pigment had no legislative record behind it. Her doubt was not contrarian, it was correct practice: a claim resting on one letter should be held loosely.

Then in 2018 a team re-analysed the surviving samples that Mukharji had sent to Kew. They found euxanthic acid and euxanthone as expected, and in some samples hippuric acid, a ruminant metabolite excreted in urine. The paper describes this as the first substantial link between animal urine and the pigment. Mukharji was right.

Table 2. Nobody in this sequence behaved badly except the century that repeated a single letter without saying so.

The person who doubted was not defeated by the person who confirmed. The doubt was right given what was knowable in 2002, and the confirmation was possible because someone in 1883 kept the physical sample. That is the whole argument for preserving evidence rather than only conclusions, and it is the reason I keep raw material around long after I have stopped needing it.

Video. "A History of Color: An Audio Tour of the Forbes Pigment Collection," Harvard Art Museums. The collection exists so that the material evidence survives the explanation.

2016: scarcity is reintroduced by contract

Geology stopped making colour scarce. Contracts started.

Surrey NanoSystems announced Vantablack in 2014, a vertically aligned carbon nanotube array that traps incident light rather than reflecting it. The company's own figure is worth stating exactly as the company states it: Vantablack absorbs 99.965% of light at 750 nm. It is usually paraphrased as 99.965% of visible light, which is a different and larger claim than the measurement supports.

In 2016 Anish Kapoor acquired exclusive rights to the material for artistic use. The precise wording of that arrangement is worth care, because the sentence everyone quotes is a journalist's summary of a manufacturer's FAQ rather than the manufacturer's own words, and the FAQ is no longer online. What is not disputed is the effect: one artist, and only one, could use it.

Stuart Semple's response was to manufacture the opposite. He produced an extremely saturated pink, priced it at a few pounds, and attached a condition to the purchase which is still live on his shop today:

By adding this product to your cart you confirm that you are not Anish Kapoor, you are in no way affiliated to Anish Kapoor, you are not purchasing this item on behalf of Anish Kapoor or an associate of Anish Kapoor. To the best of your knowledge, information and belief this paint will not make its way into that hands of Anish Kapoor.

The typo is in the original, and has survived years of editing, which I find quietly perfect. Kapoor replied by posting a photograph of his middle finger dipped in the pink. Semple then produced progressively blacker paints available to anyone: Black 2.0, then 3.0, and now Black 4.0, which is sold openly at consumer prices.

It is easy to read this as a stunt, and it partly is. It is also the oldest argument in the history of pigment, restated with lawyers instead of mountains. For most of history the answer to "who may use this colour" was decided by distance and labour. For a few decades in the twentieth century the answer was nobody in particular. Then somebody noticed that the question could be answered by contract, and it was.

Back to the picker

Two things follow, and I hold them at different confidence levels.

The first I am sure of. Zero marginal cost means hierarchy is entirely authored. No colour in a modern palette arrives carrying importance, so if a system's most saturated colour is not reserved for something, it is reserved for nothing, and users will correctly learn to ignore it. The rule has to exist in writing or it does not exist.

The second is softer, and closer to a suspicion than a claim. The cost of materials did not vanish from our work, it moved out of view. Compute, extraction, the labour behind the datasets, the attention of the person receiving the thing. I am not going to pretend I have quantified any of that. But the history of pigment is fairly consistent on one point: the cost was usually visible to somebody, and usually not to the person looking at the picture.

Sources

Ultramarine prize, dates and the Guimet–Gmelin priority question: WebExhibits, Pigments through the Ages. Mummy brown: "Techniques: The Passing of Mummy Brown," TIME, 2 October 1964; Georgiana Burne-Jones, Memorials of Edward Burne-Jones, 1906. Indian yellow: R. Ploeger, A. Shugar, G. D. Smith and V. J. Chen, "Late 19th century accounts of Indian yellow: The analysis of samples from the Royal Botanic Gardens, Kew," Dyes and Pigments 160 (2019), 418–431, doi:10.1016/j.dyepig.2018.08.014; Victoria Finlay, Colour: Travels Through the Paintbox, 2002. Vantablack absorption figure: Surrey NanoSystems. Purchase condition and current prices: Culture Hustle. Kedzie's volume is digitised by the National Library of Medicine.

I write about art in Korean on my Naver blog, 베이니월드, where the same subjects get a different treatment.

Related reading

• The Caption Did the Work

• Eighty-three, or Ninety-one

• What the Bauhaus Preliminary Course Actually Taught

• Ulm, and the Idea That Design Could Have a Method

More from the Moc Mallen Studio journal