Iona: version 2

2001. Unfinished study, oil on board, 18x14 inches.


Painted in two sessions of two and a half hours each including breaks. This one would have required another sitting really to bring it to life. It is darker and greener in colour than it would have been when finished. I used an experimental limited palette of titanium white, lamp black, pale yellow ochre and light red (red iron oxide PR 102). This is similar to the limited palette of Iona 1 but with lamp black replacing the cerulean blue, and the omission of burnt umber.

Experiment with Greek/Egyptian colours

There were several reasons for the experiment. First, I was amazed when I saw the 1st century AD mummy portraits (done by Greek painters in northern Egypt of Roman people). The flesh tones are in many cases extremely subtle, yet I read that the Greek painters used only four colours.

Fayum portraits book Some of these mummy portraits are owned by the British Museum, which did a big display of them some time back (the exhibition book by Euphrosyne Doxiadis is available from Amazon: click the image on the left for purchase information). The Greek painters used only four colours for the flesh tones: white, black, yellow and red, but until I had the book I didn't know which black, which red etc. I reasoned as follows: the greenish tones (which normally occur in the lighter shadows in the caucasian skin type) could only be got from a mixture of black and yellow if the black was bluish. Therefore it was probably lamp black. I am guessing that the red would have been red iron oxide and the yellow would have been yellow ochre: the yellow dug out of the ground and extracted from coloured clays, and the red obtained by heating the yellow. Presumably the white was white lead, or maybe chalk. I experimented along these lines and when I read Doxiadis's book I found there the same theory (pp98-99). Doxiadis says that: In modern times it has been asserted that the Greek palette of primaries must also have included blue. She goes on to say that experiments by someone called Tsarouchis demostrated otherwise. (I wonder why anyone would theorise incorrectly about this when the experiments required are so simple to do?)

Doxiadis also quotes the following from Pliny, writing in the first century AD:

...four colours only were used by the illustrious painters Apelles, Aetion, Melanthius and Nichomachus to execute their immortal works...

Pliny is referring to 4th century BC Greek artists, and referring to the much larger array of coloured pigments available in his day, he goes on to lament: is values of materials and not of genius that people are now on the lookout for.

Other reasons for experimenting with four colours are that limited palettes tend to produce a more harmonious impression, and also that in some ways it is easier. There are fewer problems to solve in terms of tone and colour, and fewer confusing mixtures scattered all over the palette.

More recently I saw the exhibition Rembrandt's women at the Royal Academy in London, and I wondered what Rembrandt used for green. The greens in old paintings are seldom very vibrant, yellows were always difficult (since the cadmiums and other yellows hadn't been invented, and orpiment did not mix well), and if you mix ultramarine with yellow ochre you don't get a very pleasant colour. Rembrandt in any case used very little blue, and to use ultramarine to mix green would have been a bit extravagant. Modern greens can be rather brash, and some (such as phthalo green) look very chemical and unnatural. I found that lamp black mixed with pale yellow ochre yields a quite pleasant but very muted green.

I was impressed by one portrait by Rembrandt in which a muted green background and the more intense green coat of the woman set off the flesh tones beautifully.

In the present case, although I believe the four colour Greek palette I have used here can be made to work, it would have saved a lot of labour had I allowed myself the use of burnt umber, instead of mixing three of my four colours to approximate it. The overall greenish effect would have been removed by overpainting the lit areas of the face with a warmer yellow red and white mixture. The Greek painters, as also Michaelangelo and other Renaissance painters, would often use a greenish underpainting for flesh tones. I have experimented with terre vert for this, but terre vert nowadays is usually an unholy mixture of other things, and the brand I bought is not a colour I like much.

Eric Hebborn in The Art Forger's Handbook claims that Frans Hals used only these four colours, and that Rembrandt used these four plus the earth colours: the umbers and siennas.

I pass on these observations to anyone interested in making or looking at paintings.


what's new home portraits

Martin Dace portraits
All images copyright 2001 Martin Dace