The whitest white I’ve been able to find is “radiant white” by Gamblin. It’s titanium white in poppy oil. Most of the time I prefer paints ground in linseed or walnut, but for this purpose it makes sense to use the whitest possible pigment and the most colorless binder available. I’m still painting out test strips on a neutral gray background, but I’d guess it’s a quarter Munsell value step than the next brightest titanium white I’ve played with. I’ll use it only when I need a very light highlight.

The darkest black I have is Williamsburg intense black. The pigment is listed as “carbon from gas flame.” The back label says: “warning: very slow drying.” It is just noticeably darker than bone (“ivory”) black. The slow drying can be compensated for somewhat with a drier such as lead napthenate. I will use it only for dark accents at the very last stage of painting, so drying time for this particular paint is not that important for me.

Update

2 May 2009:_ There’s a small highlight that I had previously painted in Old Holland titanium white. It’s light reflected from the shiny metal part of a clothes hangar. In real life this highlight is very noticeable, but on the painting, surrounded by relatively light tones, it did not stand out at all. I recently painted it in using pure Gamblin radiant white. It is noticeably brighter than before—giving an effect that is much more like what I was trying to depict.

Saturation runs on the same scale (often 1–100) regardless of the hue or the value. So the most intense blue, no matter what, is a saturation of 100. The most intense yellow, similarly, has a saturation of 100. That’s very convenient for people who design software interfaces. And if you don’t really understand color, it makes perfect sense. What it doesn’t do is model color accurately.

In reality, the most intense yellow is far more intense than the most intense blue. And the highest possible intensity varies depending on how light or dark the color is. The most intense yellows are found at a relatively light value. The most intense blues occur at somewhat darker values. At very high and very low values, maximum chroma goes way down. But you can always set the saturation to 100 to get the most intense color at the hue and value you have currently selected.

Chroma runs on a different scale. In Munsell, the most intense chroma for yellow runs up to about 18. The most intense blue is more like 12. (I’m doing this from memory, so these numbers might be a little off.) Those numbers change with hue and value, so that the Munsell color space is an irregular, lumpy cylinder. That is a much more accurate depiction of human color vision than the saturation model. It helps to think about color in this way when you’re trying to understand actual color relationships.

Update and Correction

28 February 2008:

David Briggs, in comments to this post, writes:

To clarify, David, my objection is that you seem to be judging saturation as if it was meant to be a measure of colour intensity. It isn’t, it’s a measure of colour purity. The term for intensity of colour of light is “colorfulness”, the product of the saturation of a light stimulus and its brightness.

Many colours reflect light of high saturation, but only those that reflect light of high colorfulness (high saturation AND BRIGHTNESS) have high chroma. RGB colours that emit maximum saturation red light range from very low chroma (nearly black) through moderate chroma (dark ruby reds) to very high chroma (bright red or R 255). The importance of this concept of saturation for painters comes from the fact that when a coloured surface turns from shade into light, the colours of the light reflected from it tend to follow a uniform saturation series such as this.

And darn it, he’s right. Thanks, David. As this Wikipedia entry on colorfulness puts it:

colorfulness is the perceived difference between the color of some stimulus and gray, chroma is the colorfulness of a stimulus relative to the brightness of a stimulus that appears white under similar viewing conditions, and saturation is the colorfulness of a stimulus relative to its own brightness.

I, with my obviously very small brain, still find it most useful to think about color using the characteristics of value, hue, and chroma. But saturation does not mean what I said it does. I apologize for any confusion.

Update

23 March 2008: Please note that the Rational Color blog has been transitioned to a forum. Membership is limited: you must apply and be approved. If you are a serious student of art you may want to do that. It is definitely a place for serious discussion of painting.