Unreal Nature

July 31, 2009

Flying Colors

Filed under: Uncategorized — unrealnature @ 7:15 am

“Restoration” is a very poorly chosen term, and strictly speaking, it signifies something which cannot be done. The genuine restoring of a painting is obviously something which is possible only to its original creator.”
— Max Sourner (1921), The Materials of the Artist

… a picture is never finished. No artist has ever painted an image frozen in time; all painting is a perpetual process, with every scene destined to rearrange its tonal contrasts as time does its work on the pigments. When John Ruskin said, “Every hue throughout your work is altered by every touch that you add in other places,” he might have added, “and all that happens subsequently.” When the artist has been reduced to the dust of centuries, time — sometimes personified by an overzealous restorer — goes on remodeling the colors, bringing darkness here and bleaching there, mocking our attempts to pronounce authoritatively on the coloristic intentions of the image’s creator. Even the simple act of cleaning is, as on art restorer noted, “an act of critical interpretation.”

Today’s extracts are taken from the book, Bright Earth: Art and the Invention of Color by Philip Ball (2001).

… By the eighteenth century there was still not one of the spectral colors with which artist considered themselves adequately supplied in pigments. For blues, azurite and ultramarine were scarce, and smalt, difficult to work with in the first place, had a tendency to fade owing to leaching of cobalt from the pigment particles. (The smalt sky of Valázquez’s otherwise radiant Immaculate Conception, produced about 1618 has turned a depressing brown.) Indigo, an organic dye, was not lightfast. The report of the discovery of Prussian blue in 1710 began by saying, “Painters who mix oil with their colors have few that represent blue, and those such that, rightly, they wish for [some] more satisfactory.”

The richest reds were lakes — primarily cochineal and madder — but these might not last. In Joshua Reynolds’s Anne, Countess of Albemarle (c. 1759-1760), the countess’s pallid face was once a healthier pink; the complexion has become deathly as the color drained from the (probably cochineal) red lake.

Of pure and bright purples there were still none, and for orange, realgar had only ever been used sparingly. Greens were still a particular problem. The Italian Giovanni Angelo Canini commented in the mid-seventeenth century that mixed greens should always be made up “fresher” than they actually appeared in nature because they would darken with age. A century later Robrt Dossie’s Handmaid to the Arts proclaimed that “the greens we are forced at present to compound from blue and yellow are seldom secure from flying or changing.”

This was largely the result of using fugitive yellow lakes. Foliage would often turn a bizarre and sickly blue color as the yellow component (sometimes a glaze over blue underpaint) lost its hue under the influence of light. Pieter Lastman’s Juno Discovering Jupiter with Io (c. 1618) has acquired dark blue-green vegetation, and Jan van Huysum’s Flowers in a Terracotta Vase(1746) had patchy blue leaves owing to the fading of the lake pigment. In 1830 J.-F.-L. Mérimée commented that “in several Flemish paintings, leaves of trees have become blue, because the yellow lake, mixed with ultramarine, has disappeared.”

It was hoped that Prussian blue, discovered in the early eighteenth century, would at least satisfy the desire for a blue to replace ultramarine. But already by the middle of the century, its tendency to fade had become evident. Dossie warned that the lighter, brighter, and most attractive varieties of Prussian blue (which contain white alumina) are “extremely subject to fly, or to turn to a greyish green.” This propensity to fade is accentuated when the pigment is mixed with a high proportion of white, as it often was for eighteenth-century skies. The skies in several paintings by Gainsborough (such as Gainsborough’s forest, c. 1749), Watteau (Récréation italienne, c. 1715-1716), and Canaletto (Venice: Campo San Vidal and Santa Maria della Carita, 1726-1728) are all pearly and washed out where once they would have been a deeper blue.

… The Impressionists and their successors would surely have benefited from the services of someone like [George] Field. As it was they tended to throw caution to the wind — and pay the price. Some of the new pigments, such as cobalt blue and lemon yellow (strontium chromate), aged very well; others did not at all. Zinc yellow (zinc chromate) goes greenish in oil, which proved disastrous for Georges Seurat’s carefully judged pointillism. Discoloration of zinc yellow is apparent in Bathers at Asnières (1883-1884), and the lawn of La Grande Jatte (1884-1885) is peppered with brown spots owing to deterioration of a chrome yellow. There can be few more emphatic examples of how the limitations of materials can undermine an artist’s delicate coloristic intentions.

If painters have difficulty with color stability, digital photographers have it a thousand times worse. We don’t even have a (dis)colored original. We have only a digital file — data — that must be read and interpreted. By what machine? By whom? In what cultural/historical context? We have profiled monitors and profiled paper/printer combinations and software conversions from files to monitor then from computer to printer. We have papers that are, at best known (hoped) to be stable for one hundred years — if, if you have ever managed to get a satisfactory print out of the series of interpretations in the journey from file to paper. 




  1. Interesting: I’ve just been asked to have a go at Photoshopping a scan of a watercolour to paint out foxing, and ran into exactly that issue. For initial broad colour adjustment, I corrected for overall fading and yellowing (see here: left = before, right=after). But the client (not confidential BTW) wants it merely to be de-foxed, and the overall pallid tea-stained appearance kept. Go figure.

    Comment by Ray Girvan — July 31, 2009 @ 2:21 pm

  2. That’s a good example of how restorations get done — where the editing decisions come from. I like your what you did, though I’m sure you’re guessing at which colors to use.

    Comment by unrealnature — July 31, 2009 @ 7:25 pm

  3. I’m sure you’re guessing at which colors to use

    Quite. Mainly, adjusting the global colour balance away from yellow, and upping the saturation, to make the sunlit wall not-too-vivid white: see improved current draft. But it’s anyone’s guess what pigment-specific changes have happened.

    Comment by Ray Girvan — August 1, 2009 @ 10:21 am

  4. “it’s anyone’s guess what pigment-specific changes have happened.”

    That’s what I meant.

    The new version is interesting, but I think maybe I like the first one better. The first was more “watery.” That’s (one of the many) trouble(s) with Photoshop. So many choices; so little time…

    Comment by unrealnature — August 1, 2009 @ 2:50 pm

  5. I bow completely to the artistic aspect of this subject. It does seem sad, though, that we don’t have paintings in their original colors, even if all works are “works in progress.”

    Another aspect that is fascinating to me involves what makes these colors. (A related area is the colors in pottery glazes.) Some of the colors come from the crystal field splitting of electron states in the transition metals. This is obviously what is going on in Prussian Blue but it is so molecularly complex that I doubt anyone has sorted it out from a molecular orbital standpoint.

    On the other hand, the mader dyes that you refer to, typically alizarin, often use Aluminum as a fixer (particularly to cloth). What is going on here, I think, is that these dyes contain a carbonyl and a carboxy group (-OH) and that, by binding to the metal, there is a facilitated charge transfer band that floats down into the visible.

    Now, why is this interesting? Because, in some other universe evolution would provide its minions with an expanded visual spectrum and “people” will see not only into the ultraviolet, but also into the infrared and, maybe, polarized light (like my friend the mantis shrimp.)

    Imagine the aesthetics they might have. If only now, for the last several hundred years, we have had Prussian Blue. And before that many paintings settled into drabness and lack of color. These spectral beings would not have to worry. Their palate would be immense. They might look at a “View of Delft” and see things we would never begin to see.

    But, by then, you could Photoshop it.

    Comment by Dr. C. — August 9, 2009 @ 5:11 pm

  6. The guy that wrote this stuff is a chemist (or rather, he majored in chemistry and took his PhD in physics). He manages to make the chemistry interesting even to me. I’m sure Dr. C knows most of this already but anyway, here’s what Ball says about Prussion Blue:

    In many transition metal compounds, the rearrangement of electrons induced by light absorption is confined largely to the metal ions themselves. But in some cases the electrons are shifted more dramatically. Iron’s red signature is produced by the movement of an electron onto the metal ion from an adjacent oxygen ion — a so-called charge transfer process, which in this case diminishes the positive charge of the iron. The same thing, more elaborately staged, is played out in the pigment Prussian blue. Here the crystal lattice contains a mixture of iron ions in two different charged states, interspersed with cyanide ions. Absorption of red light can send an electron across a cyanide “bridge” between metal ions of different charge.

    Here is some cool stuff about how they analyze old paintings to find out what paint was used:

    Some pigments may be prepared in forms that are identical in chemical composition but slightly different in the way that the atoms are arranged in the crystal. Realgar, the orange sulfide of arsenic, is an example: in addition to the normal crystal form, there exists a different form called pararealgar, which is also orange. To tell them apart, one needs to know where the atoms are. This can be deduced from the pattern of X-rays scattered off the pigment particles. The regular stacks of atoms in the crystals reflect the X-rays more srongly at some angles than others, a phenomenon called diffraction. For a jumble of grains in different orientation, this gives rise to a series of concentric rings of reflected X-rays that can be recorded on photographic film. The position and brightness of the rings constitute a record of where the atoms are located in space. By such means, it has been possible to deduce that Paolo Veronese made use (doubtless unwittingly) of both realgar and pararealgar in the orange hues of his Allegories (c. 1570s).

    The organic dyes that give lake pigments their colors are harder to identify with certainty. It is no good knowing that the dyes contain carbon, hydrogen, and oxygen, since that is true of most natural products and will not help us distinguish madder lake (colored by alizarin and purpuriin) from kermes carmine lake or cochineal carmine lake (both colored by carminic and kermesic acids). There are, however, some wet chemical methods that can be used to distinguish them, even to the extent of differentiating kermes carmine from cochineal carmine. And infrared spectroscopy reveals the distincitve vibrationis of the different organic molecules, provided that the sample is big enough to offer measurable spectra.

    … [He gets into thin-layer chromatography.] By using so-called high-performance liquid chromatography, it is even possible to distinguish between cochineal lakes in which the dyes have come from different sources: Old World (Polish) or New World insect species, which produce different proportions of much the same coloring molecules.

    Comment by unrealnature — August 9, 2009 @ 7:36 pm

  7. I stand in awe.

    Comment by Dr. C. — August 12, 2009 @ 9:03 pm

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Blog at WordPress.com.

%d bloggers like this: