Monday, February 20, 2012

Hamerton's Ultimate Limited Palette (1876)

The limited palette of Anders Zorn - comprised of yellow ochre, vermilion, white, and black -
had ancient roots;  the legendary Greek painter Apelles would have been using a similarly
restricted range of colors around 330 BC.

There have long been two factions amongst painters concerning the approach to outfitting a palette:  those who believe in using a restricted palette, and those who prefer using an open palette.  The first group believes that color harmony in a painting can best be achieved through the restricted use of colors, and that those who use many pigments are liable to over-rely on the convenience of available colors and become lackadaisical, using pure colors where they are likely to damage the aesthetics of a painting.  The latter group believes that the beauty of the natural world is best represented through the use of a full selection of chromatic colors, and that artists who hold fast to a limited palette are reduced to painting by formula and are capable of producing only gray or brown works of art.  Each group tries to overcome the difficulties of mastering color in its own way; each side finds their own method the easier, and each side has certainly produced works of unquestionable beauty and significance.

Sir William Etty
"With three colours and white - anything approaching to  a yellow, a red, and a blue-
(Etty) could produce a sweetly coloured picture."

In the 1870s, artist, author, art critic, and publisher, Philip Gilbert Hamerton, decided to tackle the differences between the two camps of palette users, and reconcile the groups.  Though Hamerton was obviously biased in favor of the limited palette, he felt there were too many artists during his time who were rigidly employing a far-too-restricted selection of colors, and that those artists, when claiming they could reproduce the colors of nature with 5 or fewer pigments, were sadly mistaken.  These artists, whom Hamerton referred to as the oligochromatists, could only work in low-key,  and if they used the common three-color palette - Yellow Ochre, Light Red, Cobalt Blue, plus black and white (Ivory Black and Flake White) - they would never be able to depict accurately an object such as a rose or an orange.  A limited palette was fine, felt Hamerton, and certainly preferable for the student or less-experienced artist, but any palette could not be chromatically complete without a bright yellow, a bright red, and a bright blue.

Hamerton spent two years of experimentation into the fewest colors necessary to sufficiently imitate "every hue in the world," and in the 1876 edition of his magazine, The Portfolio, published his results.  Here are his conclusions:

It is easy to imitate dull colours by mixture, but not so easy to imitate bright ones;  hence in a restricted palette it is necessary to have bright pigments.  An intensely bright chromatic yellow is essential, and also a red and a blue.  Again, there are certain practical necessities which are due to the nature of the materials at our command.  In theory one red ought to suffice, but in practice it so happens that htere is no red pigment which, by mixture with another primary, will yield at pleasure either the red of the rose or that of the fox-hunter's coat.  Vermilion must be one of our reds, but we need a rose-red besides.  For the same reason a bright chromatic yellow, though necessary, is not enough;  we need another quality of yellow for work in mixtures.  Thus, if we take pale cadmium for our bright yellow, we shall require yellow ochre for many offices which cadmium could not possibly fulfil.  As to the blues, when ultramarine is not too purple one blue is enough.  White and black we presume to be necessary from the beginning.  Our list is already extensive enough to imitate a great variety of greens, but not all greens.  You could not quite accurately imitate either malachite or an emerald with it.  Let us add a good useful green, then, the Emerald Oxide of Chromium.  We want a brown, too, for the darkening of yellows and the neutralising of greens;  a brown that will go well with yellow ochre, and form agreeable tints with white.  We should like Asphaltum best for this, but it is too dangerous.  The next best brown for our present purpose is Vandyke Brown. 
Our restricted palette is now complete.  It is composed of 

White - Pale Cadmium Yellow - Yellow Ochre - Vermilion - Rose Madder - 
Ultramarine - Emerald Oxide of Chromium - Vandyke Brown - Black. 

Seven colours, with the addition of white and black - nine pigments in all.  With these every hue in the world may be imitated closely enough for the purposes of art.  The tints given by this palette may not always be so bright as the originals, but by the help of opposition they may be made sufficiently bright for the purposes of a colourist.  Even this deficiency of brightness is a safeguard against glare. 
It may be a question whether pale Cadmium is the best pigment we could select for our bright yellow.  If not adulterated, we should give it preference.  Aureolin might, perhaps, be used instead of it, but as Aureolin is the same price as Cadmium there may still be a risk of adulteration.  A very pure and true yellow is Strontian Yellow.  It has not quite so much body as Cadmium, but it bears a high character for stability...  Our experiments have, however, been conducted throughout with pale Cadmium. 
There can be no question about Yellow Ochre.  It is one of the most valuable pigments we possess, combining so perfectly with white at one end of the scale and brown at the other. 
Vermilion is especially useful for its power of reddening ochre, making purples with blue, and neutralising greens.  In a restricted palette it is of constant use.  In a palette not restricted the utility of Vermilion is not so much felt, because the red earths in a great measure supply its place. 
Rose Madder should have the true decided rosy tint, as that is necessary for the completion of our chromatic scale.  It is believed to be permanent, but we do not really know yet whether it will last for centuries.  We certainly know, however, that it is very much safer than insect lakes, and there can be no complete colouring without a lake of some sort. 
It is, unfortunately, impossible to use genuine Ultramarine in the ordinary work of painting, on account of its great cost.  The French Ultramarines commonly sold are very useful pigments, but not ideally perfect, because they incline too much to violet.  When the only accessible Ultramarine has a violet tinge, it may be well to admit cobalt upon the palette, in addition. 
The Emerald Oxide of Chromium is one of the most useful pigments we have, and when not adulterated it is one of the safest.  In a restricted palette it becomes of the utmost importance, and is the sheet-anchor of the greens, as vermilion is of the reds.  A distinguished living landscape-painter says that in his art there are two principles, the red principle and the green principle, and that it is chiefly the opposition and interchange of these two which constitute the colour of landscape.  It is therefore well for a landscape-painter to see that his palette is strong on these two points. 
Vandyke Brown is said to be got from peat, and to be of a bituminous state.  It is semi-transparent, and both rich and deep in colour, but, unfortunately, it becomes dull and opaque in drying.  It has a reputation of being a 'bad' drier;  but the truth is that it dries well, though tardily.  One of the Cappagh browns might be substituted for Vandyke Brown if slow drying were considered a great objection. 
There are plenty of Blacks to chose from;  but when you only take one it must be really black, and not a very dark brown or grey.  Ivory Black, when of the best quality, is on the whole the most eligible.  Black Lead makes better grey tints in combination with white, because they are more nearly neutral.  It is quite a delusion to imagine that the intense blacks ever give really neutral shades in their combination with white:  their greys are usually anything but neutral. 
The following is a candid account of one of the tests to which our restricted palette has been subjected.  The reader will please remember that these are merely tests, and that in actual practice a painter does not trouble himself to imitate consciously the pigments which are absent from his palette.  This is not necessary, but such tests as these prove that the palette has the same chromatic resources which it would have if the pigments which it can imitate were actually present.  To make the test more severe, the original pigments and the mixed tints were placed side by side on a mill-board, left to dry for two years, and then varnished. 
Naples Yellow. - Imitated with White, Cadmium, and Yellow Ochre.  This was very like the best Naples Yellow. 
Jaune Pinart. - White, Cadmium, Yellow Ochre, in other proportions.  Very near. 
Lemon Yellow. - White, Cadmium, and a very little Oxide of Chromium.  A close imitation. 
Orange Cadmium. - Cadmium and Vermilion.  Accurate in hue, but rather less bright than the original. 
Light Red. - Vermilion, Yellow Ochre, Vandyke Brown.  A very close imitation. 
Venetian Red. - Vermilion, Yellow Ochre, Rose Madder, and a little Vandyke Brown.  As close as possible. 
Indian Red. - Vermilion Red, Rose Madder, Black, and a little Yellow Ochre.  In hue quite indistinguishable from the original. 
Cobalt. - Ultramarine, White, a little Yellow Ochre, and a very little Cadmium.  Very near in hue, but less transparent than Cobalt. 
Prussian Blue. - Ultramarine, Black, an atom of Chromium Green.  Deceptively close in hue, but the imitation cannot have all the transparency, or the wonderful colouring power of the original. 
Ultramarine Ashes. - Ultramarine, Black, and White.  Very near in hue, but rather more opaque. 
Indigo. - Ultramarine and Black.  A very close imitation indeed. 
Raw Sienna. - Yellow Ochre, Vandyke Brown, and a very little Rose Madder.  Near to the original, but not identical in hue.  The mixture cannot have the fine colouring properties of Raw Sienna, but the beauty of Raw Sienna may be got, when required, by opposition. 
Burnt Sienna. - This pigment is easier to imitate closely than the preceding one.  An accurate imitation may be produced with Yellow Ochre, Vermilion, much Rose Madder, and some Vandyke Brown. 
Emerald Green (Vert Véronèse). - Cadmium, White, Oxide of Chromium, Ultramarine.  An extremely close approximation in hue, but the imitation is not so bright or intense as the original.  This is not much of a misfortune, as the original is one of the crudest pigments we have. 
Malachite Green. - Oxide of Chromium, Ultramarine, Yellow Ochre, Cadmium, and White. Scarcely distinguishable from the original. 
Terre Verte. - Oxide of Chromium, Yellow Ochre, Black, White.  A close imitation in hue, but not in body. 

Besides these tests we made an extensive series of compound tints, going round the chromatic compass, and always found that the palette was able to give the tint required.  It is not necessary to trouble the reader with a minute description of these.  The conclusion was, that our restricted palette had in itself all the resources necessary for full colouring, and we believe that it is not possible to generate the whole progeny of intermediate tints from fewer originals.  We believe, also, that whenever a colourist composes a palette of nine pigments, which shall have all the chromatic possibilities, that palette must of necessity be arranged exactly on the same principles as the one we have given.  The necessities of the case leave us no choice.  It is not a matter of taste at all, but of compulsion by the nature of things.  We do not say that the artist must of necessity take the very pigments we have named, but he must take their equivalents.

(Sir Joshua Reynolds) seldom, if ever, used so many as nine pigments on one picture.
His own portrait was painted with five only.  On the 22nd of June, 1770, he says
what his fixed practice was, and this only includes five pigments, namely,
White, Yellow Ochre, Lake, Ultramarine, and Black.  His conviction was that four
are sufficient.  'Another circumstance,' he says, 'that tends to prejudice me in favour
of their colouring'
(he is speaking of the ancients) 'is the account we have of some
of their principal painters using but four colours only.  I am convinced the fewer
the colours the cleaner will be the effect of those colours, and that four are sufficient
to make every combination required.'

The fullest palette ever used by reynolds, so far as we are able to ascertain, was as follows:
White - Orpiment - Yellow Ochre - Lake - Carmine - Ultramarine - Blue Black - Black

When full colour is not an object, a valuable restricted palette may easily be arranged with five pigments;  three to represent (in some measure) the primaries, and white and black.  The palette of five pigments may often amply suffice for the dead-colouring of a picture, and it may even happen that the subject of the picture allows of its being entirely painted in five pigments.  There are works by Corot, Daubigny, and others, which might have been easily painted with so restricted a palette as that.  We have found the following a very good simple palette for dead-colouring: -  

Flake White - Yellow Ochre - Cobalt Blue - Light Red - Ivory Black 

In publishing our experiments in few pigments we are far from desiring to imply any condemnation of the many excellent artists who willingly use all safe pigments which the colour-maker can prepare for them.  We say simply, that it is sometimes a convenience to have complete chromatic possibilities in nine pigments, and we have thought that it might be useful to point out those necessities which decide which pigments must be used when only nine are chosen.  After some practice the artist plays upon his materials as a musician plays upon his strings or keys, and it is always a comfort to him to know that he has the full chromatic range in the substances on his palette.  When, on the other hand, the chromatic range is not complete, it is most desirable that the artist should be clearly aware of it, and not waste efforts in trying for results which are not attainable with the means at his disposal.  We admit that the knowledge of materials cannot make a colourist, but we believe that it may shorten the road to excellence when a student has naturally an eye for colour.  There are certain facts about pigments which are undeniable, for they can be demonstrated scientifically.  The chromatic completeness of incompleteness of a palette admits of this positive proof. 
We quite believe that... young artists and all amateurs should use as few pigments as possible.  (The) reason for this (is) that the more pigments an artist has on his palette the less is he likely to become fully acquainted with the colouring powers of each.  It may even be well for them (young artists and amateurs) to paint consciously with a palette so limited as to be incomplete, if only harmonious in its own key.  There does not seem, however, to be any serious reason why a mature artist should not use as many pigments as he likes, if all he admits are permanent.  Painters are always likely to be divided into the two classes which we have called polychromatists and oligochromatists:  the former will include those who care most for the brightness and vividness of particular hues;  the latter those who look for harmony of general effect.  There is also another side to this question which we have been compelled to omit altogether from this paper.  An artist may like a pigment for other qualities than its colour:  he may like it for its body, or its transparence, or its fluidity:  he may like it in oil, because it dries well;   or in water, because it washes well.  Artists often become attached to pigments for reasons of this kind, and they are certainly good reasons, from the practical point of view.

Hamerton, Philip Gilbert, Technical Notes:  Experiments on a Restricted Palette, "The Portfolio", Seeley, Jackson, and Halliday, London 1876 pp. 130-136


Chloe Cumming said...

So much great stuff! Always mulling over these kinds of decisions when planning paintings I never have time to execute.

Frank Gardner said...

Good info here Matt. I always love talk of limited palettes.

MCGuilmet said...

Really enjoyed this one, thanks!

oilsketch said...

Very interesting post!

john pototschnik said...

Really excellent, Matthew. Thanks.

Jim Gibbons said...

Wonderful information! I know a little about limited palettes but now know a whole lot more. cheers

Christopher Adam Lessley said...

This is a fantastic bit of research. I especially enjoyed the bit about the painting being left to dry for two years to test color steadfastness. Thank you so much for all the work you put into this blog! I'm rabid for every new post...

Phil Kapitan said...

This is wicked cool. Since this was done in 1876 though, I'm wondering if there are better colors to be used now perhaps than some of the ones recommended on that palette? Like Vandyke Brown for there a more modern pigment that would have almost the exact effects but not be as fugitive?


innisart said...
This comment has been removed by the author.
innisart said...

@Phil As Hammerton said, "We do not say that the artist must of necessity take the very pigments we have named, but he must take their equivalents."

VanDyke Brown is almost a warm black which is somewhat transparent, and when white is added to it, produces a warm gray. I suppose there are other possibilities to replace it, but usually, the other earth colors tend to be more chalky.

Winsor & Newton makes a VanDyke Brown which they consider to be of high quality. They give it an "A" rating for permanency (one step down from the most lightfast). It is made from bituminous earth like that Hammerton describes.

Gamblin makes a VanDyke Brown which "matches the 18c. original in masstone and tint but... unlike the peat moss-based original, is a permanent color." Gamblin's is made from Bone Black and Iron Oxide (PBk 9 and PR 101).

Gamblin also makes a permanent Asphaltum, the color Hammerton thought was better than VanDyke Brown except for the factor that Asphaltum was "too dangerous."

Phil Kapitan said...

Nice! Thanks for the info man. I definitely will be experimenting with this palette. Seems very solid.