On Pigments

Getting to the bottom of how pigments (and indeed different preparations thereof) behave ultimately can become a bit of a rabbit hole. I will aim in this post to wander, perhaps circuitously, through some of my thoughts and considerations when exploring and using different pigments in different ways.

Classifying pigments

On a very basic level, pigments can be classified along two dimensions: natural or synthetic, and organic or inorganic. This gives us four broad categories. These are natural organic, natural inorganic, synthetic organic, and synthetic inorganic.

Natural Organic Pigments

are those made from plant or animal origin – though instantly the waters can be muddied to an extent as many of these are dyes and are laked (chemically bonded) to an inorganic precipitant (often aluminium hydroxide)¹. Examples of these could be madder root, cochineal, indigo, Tyrian purple, stil-de-grain (buckthorn berries). Other pigments may simply be ground plant or animal material – dragon’s blood resin, gamboge, charcoal, burned bones, peat, and (the slightly gross) asphaltum from mummies…

Natural Inorganic Pigments

are naturally occurring inorganic minerals, earths and rocks used as pigments – often metal oxides, oxide hydroxides, sulfides, sulfates, silicates and carbonates. This includes red, yellow and brown earths (iron oxides), green earths (clays containing large amounts of silica and the green minerals), slate, chalk, and coloured minerals such as lapis lazuli, cinnabar, malachite and so on. Virtually any coloured rock can become a pigment if it is ground finely enough.

Synthetic Inorganic Pigments

cover a wide range of compounds, including historically manufactured copper pigments such as bice, verditer, and verdigris, synthetic iron oxides made under controlled conditions, ranging through cobalts, cadmiums, Prussian blue, french ultramarine (chemically related to the blue mineral lazurite in lapis lazuli) right through to complex modern pigments like the Siccopal pigments and YInMn Blue.

Synthetic Organic Pigments

are those based on organic compounds in chemical terms (though a small subset may have other atoms like metals in their structure). These are often derived from petroleum products and tend to have wonderfully complex sounding names – arylide, diarylide, disazo condensation, perylene, isoindolinone, phthalocyanine, quinacridone, diketo-pyrrolo pyrrole, dioxazine and so on.

These categories can go some way to help us to understand the artistic properties and uses of various pigments, but first, before we explore these further it is useful to think about pigments as categorised by some of these properties, which can again be done in several different ways.

Transparent/Opaque/Staining/Granulating…

Perhaps the most basic way of describing a paint is as transparent/opaque and staining/granulating. But these are properties we observe in the paint rather than necessarily intrinsic properties of the pigment itself, since preparation and formulation can radically alter how a pigment behaves; this is indeed often given as basic information by manufacturers². For basic understanding of how a paint will behave, these are a good start, but instantly run into issues, for example some colours may be very opaque if you could apply them in sufficient mass-tone, but act to all extents as transparent in use due to how they wash out and often granulate. Smalt springs to mind: the individual particles can be relatively opaque, yet in a dilute wash the paint can behave almost entirely transparently, partly because the particles separate and settle. Others may stain but still granulate – or at least are flocculating and sedimentary (some cadmiums). Others are classically seen as staining, and stain they do, but flocculate strongly due to hydrophobic properties in certain preparations (Phthalocyanine Blue). To move further on from this there needs to be more nuance.

Kosvanec – The Transparent Watercolour Wheel

These definitions of pigments as “transparent”, “opaque”, staining, granulating and so on have basic meanings which as we have discussed above, and in this sense do not take too much to understand, but they have a specific meaning in Jim Kosvanec’s Book “The Transparent Watercolour Wheel”³ where he expands these to a set of specific qualities with general advice on how they react together in mixes and glazes, focussing on the technique of painting transparently to allow reflected light from the paper to dictate the colours in the work.

These are as follows:

Transparent Non-Staining

As they suggest, these are colours with some level of transparency, which do not stain the paper and can be lifted to leave limited trace.

The key colours Kosvanec lists here are: Rose Madder Genuine (NR9)⁴, Permanent Rose (PR60⁵), Cobalt Blue (PB28, also likely PB72, PB73 and PB74 which are other Cobalt Blue Pigments), Viridian (PG17), Hookers Green⁶, Aureolin (PY40).

In general they share similar handling characteristics – they will mix well with one another and you can generally mix more than two without getting too muddy. Mixing with opaque colours can clean up the opaques. Mixing with stains can be very effective, but can dye the transparent non staining colours and totally change the effect. Generally, they have a weak tinting strength and are light valued, making dark mixes difficult.

Semi-Transparent Non Staining

Reading on, you will note that there’s quite some cross over between the granulating pigments (discussed later) and the transparent non-staining and the transparency is to an extent created by the tendency to granulate rather than cover thickly and smoothly. The Semi-Transparent Non-Staining colours are deeper, and a little more intense.

Kosvanec Lists the following: New Gamboge⁷, Raw Sienna (PBr7), Quinacridone Gold (PO49)⁸, Quinacridone Burnt Orange (PO48), Burnt Sienna (PBr7), Vermillion⁹, Quinacridone Burnt Scarlet (PR206), Permanent Magenta¹⁰, Opera (Often PV19, with PR81/fluoro), Permanent Mauve (Manganese Violet, PV15), Prussian Blue/Antwerp Blue (PB27), Cyanine Blue¹¹, Sap Green¹², Green Gold (PY129).

Can be glazed without worrying about mud, but care needed to avoid lifting. Mixing with opaque colours can clean up the opaques. Mixing with stains can be very effective, but can dye the transparent non staining colours and totally change the effect. These tend to have darker values and higher tinting strength than the transparent non-staining pigments.

Transparent Staining

These are colours that stain the fibres of the paper, that are brash, intense, saturated. Kosvanec¹³ himself points out that the pigments used in these may well be very opaque in mass tone, but in use they are colours that would tend to be diluted and used in a wash. These are the “Winsors” from Winsor and Newton¹⁴, and similar from other brands, essentially most modern organic pigments such as arylides, phthalocyanines, naphthols, Dioxazine Violet, Indanthrone Blue and so on.¹⁵ Kosvanec includes Hookers Green light – perhaps meaning the Nitroso green, Carmine – uncertain if he means natural carmine – and Alizarin in this section. I would add some organo-metal compounds (nickel and copper azomethines), pyrroles, pyrazolones, and perylenes etc.

These can easily take over mixes, work well as glazes and are resistant to lifting. They will often flow into heavier colours – but this is to an extent affected by preparation.

Interestingly, Kosvanec does not mention the Quinacridones here in any depth. We’ll discuss them later, in any case.

Semi-opaque and Opaque

Kosvanec lumps together a number of colours here – though what is important is that he notes that the usage of these is often transparent. He includes here: Cadmiums, Hansa Yellow (something which which I would not agree), some Quinacridones, Cobalts and Ceruleans, Manganese Blue, Bice, Ultramarine. I would add Nickel Titanate, Nickel Dioxine and some of the more modern compounds – YInMn Blue, Siccopal Pigments, Bismuth Vanadates and so on.

Kosvanec notes that these will often mix ok with one other of their own type, and with transparent non-staining, but very quickly become muddy.

I will leave this section here, as I am going to expand somewhat on these shortly.

Whitened and Blackened Colours

Kosvanec¹⁶ refers to these as “semi-opaque and opaque colours which have some black or white in their composition”, which reads to me to sound like he is seeing many of these as mixes – however many of these are obviously single pigments. It is perhaps better to think of earths, blacks and whites. Kosvanec lists Naples Yellow (most likely thinking of a blend), Ochres, Umbers, Sepia, Indigo, Terre Vert, Mars Violet, Blacks, Whites and Tints, but also Hooker’s Green Dark (mentioned elsewhere) and Perylene Maroon (which I would class firmly a as a transparent stain.

In use, he suggests these are broadly the same as Semi-Opaques and Opaques, which for now I will leave as it is.

What is clear here is many of these are granulating or sedimentary, yet granulation is not restricted to these, some can stain, and not all follow a neat definition.

We need to go further.

Keene Wilson

Keene Wilson¹⁷ rehearses and expands a little on the work of Kosvanec, widening the discussion on the (semi)opaques.

He breaks down these into a few different categorisations, looking at “intense” opaques (in which he lists Cadmiums, Ultramarine, Mineral (Manganese) Violet, Cerulean, Manganese Blue, Hansa Yellow and Quinacridone Violet with the notes that two will mix, but must not be messed with, and echoes Kosvic that more than two will make mud, they mix nicely with transparent non-staining, but will be very dull when mixed with stains.,

He makes a special note on the Cadmiums that while these are some of the most intense colours (from opaques) they can be chalky and work in many ways better when pulled out, and do not like mixing with anything other than cadmiums.

He makes specific notes about some of the others, for example avoiding re-stating Cerulean, that Ultramarine does not mix well with various other opaques and so on. He also notes that Tom Hill (as so I) considers Hansa Yellow a transparent.

This breaks some of the outliers slightly out of Kosvanec’s assessment.

Where this is important is that it begins to make the distinction between granulation and sedimentary colours – while similar they are distinct characteristics. Moving on one also needs to consider flocculation¹⁸. For me, at least, Cadmiums, Mars Violet (Caput Mortuum), Oxide of Chromium (to name a few) are better described as sedimentary than granulating. Quinacridones and some other organics are granulating, but the particle sizes are small enough this is very different than what we would consider true granulation. Cobalts, which Kosvanec lists as transparent, are very granular. And Phthalocyanines can flocculate very strongly, producing an appearance which can resemble granulation, but the mechanism is different – this is due to their hydrophobic properties. The preparations they are released in often have specific additives to prevent this.

Enter the Quinacridones!

I hope that Caroline Buchanan will forgive the minor plagairism of a subheading in her great article for Daniel Smith, Making Sense of Staining, Sedimentary and Transparent Pigments – it was too good to ignore.

Buchanan suggests some different ways of looking at these characteristics, breaking the definitions down to Staining, Sedimentary, Luminous and then discussing the Quinacridones as a separate set.

This article goes quite far into technique and layering the different forms of pigment and is very much based around Daniel Smith’s range, but there are some useful insights to gain here.

Staining Pigments

Buchanan lists here Phthalos, Alizarin, “most of the Cadmiums”, Permanent Magenta, Hooker’s Green¹⁹, Indigo, Payne’s Grey and the Quinacridones. She notes these flow in washes and do not lift – this interests me as while Cadmiums do stain, they do not like painting over in my experience.

Sedimentary Pigments

Buchanan’s list here is broadly coterminous with that of Wilson, though she only includes orange from the Cadmiums, and brings the cobalts under this group.

Luminous Pigments

Buchanan renames the Transparents as Luminous, noting that all watercolours are effectively transparent, and this resonates with Wilson’s advice that Opaques are used transparently. It’s interesting to note almost all of these as she lists them (Viridian, Cobalt Blue, Aureolin) are granulating or textured. She discusses using these as glazes, and does note their sedimentary nature.

Quinacridones

Buchanan makes special note of the quinacridones as having luminosity with no negatives, and that they act as stains when used in a first wash. I agree with these assessments, and they are indeed wonderful pigments.

The question is now, if we are looking into this detail, how far do we go?

In Summary

The answer to the question above is plainly “how far do you want to go?”

In my mind a basic understanding of organic/inorganic and natural/synthetic gives a very good start in understanding both how some basic pigment properties will be evident.

Often granulating and staining is seen almost as a binary, which to me is unhelpful. I think at the very least we need to look at granulating and non granulating, staining and non-staining. Transparency needs to be considered, but this is more complex as indeed is the exact nature of granulation.

Then we have issues such as activity in wet, and tendency to backwash, and in all these we need to also consider the exact way the paint is made which can in some cases have as much as an effect.

Towards a set of properties for describing pigments

When we are discussing pigments and their uses – particularly in watercolour, although many of these properties matter in other media too – there are several things we want to know. The origin is perhaps less important in use, but some people more than others may wish to avoid animal products, or those with specific environmental concerns.

Certainly we want to know about things like lightfastness, and how they perform in use.

Alongside staining, granulation and opacity, we also want to describe the colour itself – hue, saturation and so on. Even here, however, the same pigment can behave differently according to particle size, impurities and the way it has been prepared. In general, I don’t feel a need to specify an exact position in a colour space; a visual swatch is more useful to me, although it can still be helpful to identify the general colour family

MacEvoy also looks at two further characteristics – that of activity in wet – how much the paint will diffuse out when dropped into a wash, and blossoming – the propensity to backwash if water meets the paint bead. These will of course be quite affected by the preparation (for example Schmincke watercolours are often quite active as I believe they use more oxgall than some, but particularly in some of the organic pigments there are trends.

So I propose, and will use in my descriptions the following:

Origin

• Natural Inorganic
• Synthetic Inorganic
• Natural Organic
• Synthetic Organic

This is pretty much agreed by all commentators subject to the discussion on laked pigments, above. For this schema will treat a natural organic dyestuff, laked to a simple alumina or clay base to be a natural organic pigment as the origin of the colour itself is predominantly from the natural compounds. I will also treat natural pigments which have been artificially calcined²⁰ as natural, for the same reason that their origin is natural, and the processing does not significantly alter the chemical makeup.

Type

This is where the matter becomes a little complex.

With synthetic organic pigments it’s relatively easy to break these down into chemical families and for the most part make reasonable assumptions about their properties therefrom. With natural organic pigments it perhaps makes less sense to categorise these in that way as many could be mixtures. For example, Tyrian Purple is predominantly 6,6′-dibromoindigo, which could be classified as an indole (a six-membered benzene ring fused to a five-membered nitrogen-containing pyrrole ring) yet the dye itself has other constituents, and as a pigment would traditionally be laked to stone powder. Indigo, is used as a pigment simply as the ground organic compound. Some natural pigments are very complex and contain various substances. Vine black, which is simply burned vegetable matter is mostly carbon.

Natural inorganic (mineral) pigments may be based on one chemical, but may naturally contain impurities. Synthetic inorganic pigments may simply be lab based synthesised versions of the main chemical (such as Ultramarine, which is synthetic Lazurite) or they may be something entirely different, though some of the simpler ones may occur as minerals, even if not used as pigments historically in that form.

Natural inorganic pigments

By this I will include naturally gathered earths, and minerals, including were these have been artificially calcined. The key definer here would be that they are gathered naturally

MacEvoy²¹ includes:

• Red earths
• Yellow earths
• Green earths
• Lapis lazuli
• Azurite
• Malachite

The list is likely to need expansion, particularly as we consider historical and less commonly used mineral pigments, and I will expand on this in a future post – for example I would most likely categorise Lapis Lazuli as “Silicate Minerals” (at least in mineralogical terms) and include Azurite and Malachite as a Copper Mineral – though if I was defining copper minerals I have with Lapis, this would be more accurately grouped as “Carbonate Minerals”.

Synthetic Inorganic

MacEvoy²² breaks these down into the following:

• Cadmium
• Chromium
• Cobalt
• Copper
• Iron
• Lead
• Magnesium
• Manganese
• Mercury
• Sulphur
• Titanium
• Zinc

This is obviously a useful (though perhaps incomplete, nowadays) list, as one may need to include cerium, bismuth, praseodymium and so on. Practically, there are considerations – For example Oxide of Chromium and Viridian are both (very similar) chromium compounds yet have very different characteristics in use, and Ultramarine is listed here as a Sulphur compound, and while, asi understand it, it is the sulphur predominantly responsible for the colour, the structure and properties are those of a silicate compound.

Natural Organic

As discussed above, this is quite complex. Many of these contain a key constituent that could be categorised with a chemical family – but this may not be the best way of considering properties. It’s likely any definitions here would need to be wider and perhaps more narrative.

Synthetic Organic

In most cases one can get an idea of properties from the chemical families of these pigments. A list, again taken mostly from MacEvoy²³ could be:

• azine²⁴
• monoazo (arylide)
• disazo (diarylide)
• disazo condensation
• benzimidazolone
• beta naphthol
• naphthol AS
• isoindolinone
• phthalocyanine
• quinacridone
• perinone
• perylene
• anthraquinone
• diketo-pyrrolo pyrrole
• dioxazine
• triarylcarbonium

Some of these catagories (e.g. phthalocyanines) include metal ions in the structure, and attached to this list could be metal complexes of other catagories.

Colour

This is based on MacEvoy²⁵ again, but is the most comprehensive and useful list I have considered:

• Green Yellow
• Yellow
• Orange Yellow
• Yellow Orange
• Orange
• Red Orange
• Orange Red
• Red
• Deep Red
• Violet Red
• Red Violet
• Violet
• Blue Violet
• Violet Blue
• Blue
• Green Blue
• Blue Green
• Green
• Yellow Green
• Earth Yellow
• Earth Orange
• Earth Red
• Brown
• White, Gray & Black

Warm/Cool

This is, as I have touched on and will again, a contentious matter. Where appropriate, I find this most useful in the context of a 4 colour (primaries plus green) wheel. This is simply due to usage, and the difficulty of mixing green from blue and yellow pigments.

• Cool Yellow
• Warm Yellow
• Warm Red
• Cool Red
• Warm Blue
• Cool Blue
• Cool Green
• Warm Green

Granulation and Sedimentation

• Granulating
• Slightly Granulating (by which I consider subtle granulations such as displayed by some quinacridones)
• Sedimentary — pigments which tend to settle rapidly and produce a sedimentary appearance – note that some of these may “granulate” in certain preparations.
• Flocculating
• Non Granulating

Staining

• Staining
• Moderately Staining
• Non-Staining

Transparency

• Opaque
• Semi-Opaque
• Semi-Transparent
• Transparent

This is a complex one; for example, as I have noted Smalt may appear opaque in its individual particles, but cannot be seen as anything but transparent in use. In general I mean opaque in the practical painting sense – can be overpainted light on dark. Semi transparent I would class predominantly as heavy transparent stains (Phthalos, Dioxazine).

Lightfastness

• Excellent
• Good
• Fair
• Fugitive

And I will include a qualifier where needed (for example, tendency to brown).

Activity in Wet

• Active in Wet
• Moderately Active in Wet
• Inactive in Wet

Additional information may be needed, as this can be significantly effected by the vehicle.

Blooming

• Susceptable to Blooming
• Moderately Susceptible to Blooming
• Inert

Afterword

This is a work in progress, and I am developing this as I explore the subject. I will update this post if appropriate and indicate. Posts discussing the four main categories will follow.

1. MacEvoy – https://www.handprint.com/HP/WCL/pigmt1.html#pigmenttypes – would classify this as the pure natural product “only modified by grinding, washing, filtering or heating”; I would include laking as acceptable as it is effectively still a natural pigment, simply fixed to a substrate and many of these have N (Natural) pigment codes in the CI. Lucy Mayes, in her recent book (https://londonpigment.com/products/the-natural-pigment-handbook-presale) conisders laked pigments as synthetic, though cites cochineal to be natural although this is often laked. ↩︎
2. See here for example. ↩︎
3. Kosvanec, J. (1994) The Transparent Watercolour Wheel – A Logical and Easy to Use System for Taking the Guesswork out of Mixing Watercolours. New York, Watson-Guptill Publications ↩︎
4. n.b. Kosvanec does not list CI numbers, these are my addition. ↩︎
5. Possibly – see note 2 ↩︎
6. Kosvanec states later in his book that he refers to Holbein’s Hooker’s Green (PG7, PY110, PY150), but there are various formations which may differ. The original Hooker’s Green was reputed to be Gamboge and Prussian Blue, followed by Naphthol green Dye – see https://www.oldholland.com/academy/prof-theo-de-beer-about-hookers-green/ – Nitroso Green (PG8) is also listed, this lattter being quite staining. ↩︎
7. Various pigments and mixes can carry this name ↩︎
8. Now obsolete as not manufactured ↩︎
9. Uncertain pigment code – it’s unlikely he is referring to real vermillion ↩︎
10. Most likely Quinacridone Magenta (PR122) ↩︎
11. Very uncertain to what her refers here as Phthalocyanines are very staining. ↩︎
12. Most commonly a mix ↩︎
13. Op. Cit. p.18 ↩︎
14. Winsor and Newton call their set of transparent stains made from various modern pigments “Winsor” colours – several other manufacturers do similar – they are generally modern synthetic organics. ↩︎
15. While many others plainly existed when Kosovanec wrote his book in 1992, his examples are limited by modern standards ↩︎
16. Op. Cit. P.22 ↩︎
17. https://www.keenewilson.com/page/4287/knowing-watercolor-pigments ↩︎
18. Clumping, without specifically granulating ↩︎
19. a mixed paint in their range ↩︎
20. I’ll explain this more in the future parts of this set of posts ↩︎
21. https://www.handprint.com/HP/WCL/pigmt1a.html ↩︎
22. https://www.handprint.com/HP/WCL/pigmt1b.html ↩︎
23. https://www.handprint.com/HP/WCL/pigmt1d.html ↩︎
24. Added September 2026 after further development ↩︎
25. https://www.handprint.com/HP/WCL/palette1.html ↩︎