Why do pigment names like ultramarine and ochre not match one exact colour?

The material came first, the colour word later

The table of dates shows the pattern. Ochre is recorded as a clayey earth in the late thirteenth century and as a colour name, 'brownish-yellow', from the mid-fifteenth. Vermilion was cinnabar and the red made from it for three hundred years before the adjective appears in the 1580s. Indigo was a blue powder from plants in the 1550s and a colour by the 1620s. Ultramarine, literally 'beyond the sea', named the pigment made from imported lapis lazuli in the 1590s and a colour resembling it by 1781. Carmine was a dyestuff in 1712 and a colour name in 1799. Cobalt was a metal in the 1680s and a blue in 1835. In every case painters and dyers used the word for the stuff, and the wider public later borrowed it for the look.

A material has a range; a swatch has one value

A pigment's colour depends on where it came from and on how it is prepared and applied. Ochre is the clearest case in the lexicon. The catalogue's ochre is an orange ochre at 60° of hue; the crowd survey's is a yellow ochre at 85°; they are ΔE00 14.4 apart, and the region has to span 55°–90° to hold both. Neither is wrong, because the earth itself runs from yellow to red. Umber is sold raw and burnt, the burnt kind redder, a distinction recorded since the 1640s. Cobalt's name covers more than one pigment; the catalogue has a cobalt green as well as a cobalt blue. A name that covers a family of materials cannot be reduced to one HEX code without choosing one member and dropping the rest.

What happens when the name moves to a screen

The second table takes the catalogue swatch for each pigment name and asks which lexicon regions contain it. Some land where a painter would expect: vermilion inside red and scarlet, carmine inside red and crimson, cobalt blue inside blue. Others show the convention parting from the material. The swatch for verdigris, a copper green, falls inside turquoise. Umber's falls among the greys and taupes, not the browns. The swatches for cerulean and prussian blue fall outside every region as drawn: they sit in the gaps between named blues that the lexicon's terms do not reach.

Indigo: one name, a dye and a band of the spectrum

Indigo shows how far a pigment name can travel. As a dye on cloth it gives a dark blue. In 1704 the name was also given to a band of the spectrum between blue and violet, and that second sense is the one the web inherited: CSS indigo (#4b0082) has a hue of 302° in OKLCH, well into purple, and falls inside the lexicon's indigo and purple regions. The crowd survey agrees with the web, not with the dyer. Someone who asks for indigo expecting the dyer's blue and receives the CSS colour has run into two senses of the name that have been diverging for three centuries. The lexicon's region follows the violet-leaning published values and says so.

Which to use: the pigment, the colour word or a value

Each does a different job. For paint that has to behave in a known way, lightfast, opaque or transparent, mixing predictably, what matters is the pigment actually in the tube, and the name on the label is a poor guide, since a traditional name does not guarantee the traditional pigment. For describing how something looks, the pigment name is evocative and imprecise, and a modified basic term or a measured value says more. For screens and print, only a value will do. The pigment names remain worth knowing because they carry history a number cannot, and because they explain why a paint called ultramarine and a button coloured 'ultramarine' need not match.

The data

How to check it yourself

1. Look the name up in an etymological dictionary and note whether the material sense or the colour sense came first, and how far apart they are.
2. Find two or three published screen values for the name and compare them; a wide spread means the name covers a family.
3. For paint, read the pigment listed on the tube or data sheet instead of the name on the front.
4. For a historical object, look for a conservation or technical report identifying the pigment before assuming the modern colour word describes it.

What the evidence says

Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.

• FactModerate evidence

  Ochre is recorded as the name of a clayey earth from the late 13th century and as a colour name from the mid-15th; ultramarine as a pigment from the 1590s and a colour by 1781; carmine as a dyestuff in 1712 and a colour name in 1799; cobalt as a metal in the 1680s and a blue in 1835.

  Caveat: First-attestation dates in a compiled reference move when earlier uses are found.

  Source: Online Etymology Dictionary

• FactModerate evidence

  Indigo is recorded as a plant dye from the 1550s, as the colour of that dye from the 1620s, and as the name of the violet-blue band of the spectrum from 1704.

  Source: Online Etymology Dictionary

• Colourwise analysisModerate evidence

  The Colourwise catalogue's ochre and the xkcd survey's ochre are 14.4 ΔE00 apart, at OKLCH hues of about 60° and 85°.

  Based on: CIEDE2000 and OKLCH hue computed by Colourwise from the catalogue swatch and the survey's published value.

  Caveat: Two screen values chosen to represent a pigment family; neither is a measurement of a pigment sample.

  Source: Color survey results, and the rgb.txt list of the 954 most common colour names

• StandardStrong evidence

  CSS Color Level 4 defines indigo as #4b0082, a dark violet, and sienna as #a0522d.

  Source: CSS Color Module Level 4

• Colourwise interpretationModerate evidence

  A single screen value under a pigment's name is a convention standing for a family of materials, and should not be read as the colour of the pigment.

  Based on: Published values for ochre differ by a visible amount; the etymology source records raw and burnt grades of umber; several catalogue swatches for pigment names fall in lexicon regions a painter would not associate with the pigment, or in none.

Related questions

Is ultramarine a colour or a pigment?

Both, in that order. It was the name of a blue pigment made from lapis lazuli from the 1590s, and is recorded for a colour resembling that pigment by 1781. Today the colour word is used far more widely than the mineral pigment.

Why is indigo purple on screens when the dye is blue?

The name has two senses. The dye gives a dark blue on cloth. Since 1704 the word has also named the violet-blue band of the spectrum, and the web's indigo follows that sense, landing well into purple.

What is the difference between umber and sienna?

Both are earth pigments with Italian names. The etymology source describes umber as a brown earth darker and browner than ochre, named either after Umbria or from the Latin for shade, and sienna as a brownish-ochre colour, from the Italian for 'earth of Siena', recorded by 1760. Burnt umber, a redder grade, is recorded from the 1640s.

Related topics

Go further

Sources

• Online Etymology Dictionary — Douglas Harper (Reference work)
• Color survey results, and the rgb.txt list of the 954 most common colour names — xkcd (Randall Munroe) (Reference work)
• CSS Color Module Level 4 — W3C (World Wide Web Consortium) (Official standard or government)
• Colourwise editorial synthesis — Colourwise (Colourwise editorial)

Reviewed 1 October 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.