 |
| Jacques-Louis David Madame Pierre Sériziat with her son Émile 1795 oil on panel Louvre |
In 1985 the portrait above – painted on a support of thin wooden planks – suffered submersion in water during flood-conditions at the Louvre. The body of the work was successfully dried after this emergency, reportedly without harm to the wooden support or the paint layers, but the surface of the picture – at the level of the varnish-layer – remained marked with a pattern of textured vertical white tide-lines and the entire surface was dulled. Throughout 1986 the varnish-layer was slowly thinned across the entire painting and delicately manipulated until the offending marks could be persuaded to fade below the level of everyday observation, though they could not be absolutely eradicated.
 |
| Nicolas-André Monsiau Le Lion de Florence 1801 oil on canvas Louvre |
Monsiau's painting featured in the Paris Salon of 1801. The canvas was furnished with a topical narrative, claiming to illustrate a 17th-century episode that had actually occurred in Florence. One of the lions of the Grand Duke's menagerie was said to have escaped. A fleeing mother dropped her baby son and the lion seized the infant. When the desperate mother begged the lion for her son's life, the lion (activated by the nobility that Nature had conferred on the species) quietly set the boy safely on the ground. This edifying painting did not arrive at the Louvre until 1983, and had already been relined before the Museum bought it. Restoration in 1986-87 consisted of cleaning, removing old repaints, then filling gaps with new pigment. The relining had created a "pavement" pattern in relief when the paint was pressed with heat against the supporting canvas and its new adhesives. This flaw of texture was reproduced in the new repaints, so as to blend them "invisibly."
 |
| Anne-Louis Girodet de Roussy-Trioson Apollo, or, The Dawn 1814 oil on canvas Musée national du Château de Compiègne |
Napoleon began to renovate the Château de Compiègne in 1806, a project that continued until his reign ended. Of course the scheme included a particularly lush bedchamber for the Emperor. The Apollo above was commissioned for that room. Because of technical problems created inadvertently by the artist, the large blue sky quickly developed an overall pattern of "premature" cracks, revealing the white ground beneath. An early 19th-century restorer solved that problem simply enough by repainting the whole sky. In 1984 the problem resurfaced after the painting had been relined and cleaned. The old repainting of the sky, faded and unsatisfactory, was removed. The fresh ultramarine of the original sky was then revealed, with its intricate network of wide white cracks. This netting of white gaps was then painstakingly filled in with multiple layers of substitute material which was not designed exactly to match the original blue, but instead to permit the eye to register the modulations of the original blue without perceiving a second, interfering pattern. At close range this restoration work was clearly visible (as intended), but from a distance it largely disappeared.
 |
| François Gérard Daphnis et Chloé 1824 oil on canvas Louvre |
Daphnis and Chloe by François Gérard also suffered from technical difficulties created by the artist. Substances known as "drying agents" or "siccatives" became especially popular in the early 19th century. These encouraged quick drying of surface paint, permitting artists to add further layers sooner. But the drying of the lower paint layers was often by these same means unintentionally inhibited. This could eventually cause the different paint layers to pull apart, creating the wide "premature" crackling that afflicted the Apollo above. The dark background of Daphnis was similarly afflicted, with the two central figures surrounded by a vast yellowish spider-web. In neither painting were the figures touched by the cracks. They had been built up using more traditional (and safer) methods. In these cases, it was the backgrounds where the technological breakthrough of the day was disastrously adopted.
 |
| Édouard Dubufe Portrait of Madame F ca. 1865 oil on canvas Musée d'Orsay |
Édouard Dubufe was the most fashionable French portrait painter of the Second Empire. His Portrait of Madame F did not suffer afflictions as radical as those of its neighbors here. It was merely dirty, but in a complex way. Two separate layers of soiled and yellowed varnishes existed, one on top of the other. If this painting had been restored in London at the National Gallery in 1986, every scrap of varnish from both layers would have been scrubbed and scraped off. At the Louvre, the standard was more restrained. Restorers there were concerned to brighten the picture by cleaning it, but equally concerned not to allow the range of rose-pinks in the dress to display themselves too brightly. The long-time policy at the Louvre has been to "thin" a discolored varnish judiciously, when possible, leaving a fine layer remaining as "patina" to balance out the color values across the entire canvas. Nowadays, the National Gallery policy of total varnish removal has reportedly been scrapped, both because of its riskiness and because of increasing doubts that a "bare" canvas is more representative of an artist's intentions than a canvas with a "patina" layer left in place.
 |
| Édouard Manet Le Fifre 1866 oil on canvas Musée d'Orsay |
Manet's Fife Player was one of the earliest and most scandalous test-cases of an Impressionist masterpiece formally and publicly rejected by the Paris Salon. This happened in 1866. Only fifty years later an enlightened collector bequeathed the painting to the Louvre, where it was gratefully accepted. Eventually consigned to the Musée d'Orsay, it has become one of the most famous of all French paintings. A large Manet retrospective exhibition took place in 1983. Restorers and curators used that opportunity to compare the conditions and color-balances of many Manets brought together in physical proximity. They concluded that the Fife Player had yellowed significantly and should be cleaned. Yet the painting was too well known to be cleaned without controversy. A decision-making commission and much weighing of diverse options eventually resulted in the "moderation" of the varnish layer without thorough removal.
 |
| Claude Monet Portrait of Madame Gaudibert 1868 oil on canvas Musée d'Orsay |
Monet's early portrait of Madame Gaudibert shows his style much nearer to Manet's, before Monet began to fragment his forms and colors. The blue-gray curtain hanging in the background had been repeatedly retouched in the period after World War II with a then-popular mixture of tempera and oil thought to be color-stable. In fact the retouch-paint had darkened into strong contrasts against the original paint, which thus proved to be far more stable. With the retouches removed, the varnish layer was reportedly thinned and manipulated according to the standard Louvre method of revealing and balancing existing values across the painting as well as possible.
 |
| Gustave Courbet La mer orageuse 1870 oil on canvas Musée d'Orsay |
Courbet's stormy sea entered the Louvre in 1878, only a few years after its creation. The painting's many physical problems have thus been recorded and treated by the same institution for its entire lifespan, up to the present. Over a white ground Courbet placed brown-black layers of underpainting and a layer of varnish before he began to lay on the colors of his composition. The sky was painted thinly, while the sea received heavy layers of impasto. From early days the paint showed a reluctance to adhere properly, and began to flake. The canvas was relined in 1937, and the saturation with adhesives at that time provided about forty years of relative stability. By 1977 another major intervention became necessary to address areas of actual paint loss from some of the depicted waves. Fresh adhesives were applied and repaints were created to fill the losses. These repaints painstakingly duplicated the many layers Courbet had applied and are said to be completely invisible. There is, however, no reason to suppose the picture can ever escape from repeated cycles of loosening paint, temporary reattachments, and complex concealments.
 |
| Paul Cézanne L'Estaque 1878-79 oil on canvas Musée d'Orsay |
In the early 1980s there was a major effort at the Louvre to clean and restore a range of Impressionist paintings in preparation for the opening of the new Impressionist branch, to be known as the Musée d'Orsay. The Cézanne seascape above had been relined and cleaned in the 1950s. The old varnish deliberately left behind at that time had darkened further and cracked, especially in the foreground where paint texture was thicker and more irregular. Remnants of dark varnish embedded in troughs and depressions of Cézanne's paintwork were attacked with a scalpel under a microscope, in combination with solvents. The picture definitely looked brighter and fresher when it went on display in its new location in 1986. Modern synthetic varnishes will not discolor like old varnishes, it is sometimes claimed. But the claim of harmlessness has been made so often for various treatments and substances in the past, and has so consistently proved false, that new claims surely deserve to be met with suspicion.
 |
| Gustave Moreau Hélène glorifiée ca. 1890 oil on canvas Musée Gustave Moreau, Paris |
Moreau's Hélène of 1890 was never varnished. Like many advanced painters at the turn of the century, Gustave Moreau preferred the matte finish that helped to distinguish "modern art" from old-fashioned, glossy, 19th-century "Salon art." When Moreau's painting was examined by restorers at the Louvre in 1987, its only problem was a century's worth of sticky dust on the picture-surface. Safely dusting this surface was a two-part step – first with a soft brush to lift particles that could abrade the paint if they were pressed and dragged – and then a second careful pass using a dry eraser. With the painting now maintained in an air-filtered environment, it is possible that this satisfyingly simple procedure may not be needed again until another century has passed.
Descriptions of these restoration projects are based on the manual published by the Louvre in 1990 – Science et patience, ou, La restauration des peintures by Ségolène Bergeon.