Like anything else, works of art inevitably change over time. Some of the pigments in Renaissance paintings are unstable, and the varnishes often used to protect the surface gradually become discoloured. Alterations in temperature and humidity, which happen anyway but are more acute when pictures are moved, cause panels to crack, canvases to sag, and paint to flake. Frescoes suffer from other problems. The Italian technique of buon fresco, involving the use of water-based mineral pigments on wet plaster, is remarkably durable, since the paint itself forms a chemical bond with the lime of the plaster, but the lime can react with various types of atmospheric pollution, often with unfortunate consequences. Further problems can be caused by damp, and still others by the intrinsically more fragile techniques of painting a secco, on top of dry plaster. Marble sculpture, which is chemically similar to the plaster used for frescoes, can also suffer from pollution and, as is inevitable, gradually acquires a darker surface, while bronzes too develop a natural patina (although, as with marble, this can be added artificially too).

Even if the physical changes were reversible, which most of them are not, it still would not follow that we could remove centuries of accumulated dirt to reveal the pristine work of art beneath. Many major works have been cleaned before, sometimes in destructive ways, or repainted, or both; and the process of cleaning can itself involve chemicals which may react with the materials below, sometimes with consequences that only gradually become apparent. Despite these dangers, restoration is seldom undertaken solely to counter physical threats to works of art, such as cracking or flaking; rather it aims to improve their appearance. Just what constitutes an improvement is of course conditioned by current taste. Few restorers would now contemplate covering Renaissance pictures with yellow varnish, but Dürer wrote in 1509 that all varnishes except the kind he made himself were yellow. We do not know whether this was true, but it seems likely that yellow varnish was common at that period and, presumably, that painters took account of it in their pictures.

The modern practice of covering old paintings with clear varnish therefore does not necessarily give a fair idea of how they originally looked, any more than does the habit of displaying them under strong artificial light. But the brilliant colors that we so often see today certainly meet the expectations of a public accustomed to slides and color reproductions. These are almost always brighter than the originals, even when the effect of colorplates is not artificially enhanced by thick black borders, as in the magazine FMR.

Most restoration used to be done by practicing painters, but now it has turned into a highly specialized activity, with a large and growing body of scientific literature. The process of professionalization and the introduction of new techniques for the physical examination of the works themselves often give an impression of objectivity and rigor to an activity that still calls for subjective judgments, not just about the intended final result, but even about the means employed. In the 1950s it was common to replace damaged areas in frescoes with bare patches of plaster. But this practice, which has seriously compromised the appearance of masterpieces by Giotto and Piero della Francesca, has now been generally abandoned. A little later there was a fashion for peeling frescoes from the wall, but this too is now avoided whenever possible. Again and again new techniques or new chemicals have been introduced as marvelous panaceas, only to be discredited a few years later.

Even today, there are profound differences in approach and attitude among restorers, as James Beck emphasizes in Art Restoration: The Culture, the Business and the Scandal. Indeed, it is hard to think of a group of people readier to criticize their professional colleagues, at least in private. Many of them are lax about keeping detailed records, and some of the museums that employ them are equally secretive, while dealers are no better. This is in many ways understandable, given the exceptional responsibility that restorers and curators have and the fact that, whatever they do, they are almost bound to lay themselves open to criticism from those who had come to love the paintings in their former state. A consequence of this is that scholarly monographs on artists and even museum catalogs are often strikingly uninformative about the physical condition of the works themselves, let alone their restoration history.

It might be supposed that the most obvious dangers of overenthusiastic restoration could be avoided by close collaboration between restorers and art historians. But while this is entirely welcome, it does not always prevent mistakes from being made. James Beck’s co-author, Michael Daley, describes some unhappy episodes at the National Gallery in London, most of which, it should be said, occurred some years ago. Thus the removal of modeling from Uccello’s much-damaged Rout of San Romano made the picture look far flatter than the companion pieces in Florence and Paris, while Sebastiano del Piombo’s Raising of Lazarus was given a white ground of a kind he never used, with the result that the colors were heightened in a way which is clearly misleading. Although there are a number of distinguished exceptions, restorers are not always particularly well informed about the history of art, nor especially interested in it. Likewise, art historians seldom have the scientific training to judge the full implications of the courses of action proposed to them. At the same time professional curiosity can encourage them to uncover as much as possible of the original paint surface. Thus we find restorations justified as much on the grounds of the historical knowledge that has been produced as of the aesthetic results. These two goals do not always require the same kind of intervention.


All these issues are discussed with clarity and passion by James Beck. In the past few years he has become well known, or notorious, for his criticism of two important examples of recent restoration, the cleaning of the ceiling of the Sistine Chapel and that of Jacopo della Quercia’s tomb of Ilaria del Carretto in Lucca. His comments about the latter led to a bizarre charge of defamation brought by the Italian restorer, and episode which earned Beck widespread sympathy even from those who do not share his views. With some reason, he sees the prosecution not just as the angry response of a man whose professional competence had been questioned, but as a manifestation of the strength of the vested interests involved in restoration, which in Italy has become a flourishing industry.

While it is not entirely clear that the cleaning of the tomb has caused irreversible damage, Beck is certainly entitled to the view that the removal of every scrap of patina from the marble has altered its appearance for the worse; and his comment that the impact of the work could have been greatly enhanced by the simple expedient of improving the natural light in the church is unanswerable. The issues involved with cleaning the Sistine ceiling, painted between 1508 and 1512, are more complex, and the implications more profound, not just because this is one of the half dozen most famous European paintings, but also because its appearance has been irrevocably and radically changed, and with it our image of Michelangelo as a painter.

The dramatic consequences of the cleaning were evident from the first, in the lunettes of the ancestors of Christ high on the side walls. These had previously been dim and almost monochromatic, with the figures immersed in deep shadows. The removal of a thick layer, or layers, of glue and dirt transformed them in an almost miraculous way. The brilliant colors of the buon fresco, laid on with great rapidity and precision, looked utterly unlike anything by Michelangelo that anyone had previously seen. Like many other art historians, I was able to view the first of the cleaned lunettes from the scaffold, and my initial reaction was entirely positive. At that time I accepted without question that Michelangelo had worked almost exclusively in buon fresco. Both Giorgio Vasari in 1550 and Giovan Battista Armenini in 1587 state that glue was used in secco painting, but the restorers were confident that all the glue on the lunettes had been added in earlier restorations, in order to brighten the colors beneath. On the strength of this visit and of photographs of further details published from time to time, I later defended the restoration in a review of a book about the cleaning of the lunettes.

By the time that I returned to the Sistine chapel last summer the whole ceiling had been cleaned, and I must now admit that my earlier enthusiasm was misplaced. Seen in their entirety from the ground—as they should be—the frescoes create a decidedly disagreeable impression: the colors are gaudy, so that the costumes tend to overwhelm the faces and limbs, the figures look crude and often flat, and the architecture seems insubstantial and pedantic. Most troubling of all, the earlier coherence of the decoration has been lost, so that the stylistic changes in Michelangelo’s approach as he moved down the Chapel, which were previously muted by a uniform tonality and a rich play of shadows, have now become all too apparent. Restrained grandeur has been replaced by garish confusion. Talking to friends, I find that my unease is widely shared; and it is certainly noticeable that the completion of the restoration has not attracted the kind of acclaim that greeted the unveiling of the lunettes.

Although it is often said that the cleaning has greatly increased our understanding of Michelangelo’s achievements and priorities as a painter, what we see today is difficult to reconcile with the comments of his contemporaries. If this were true only of the lunettes it would not be particularly troubling, since these are the least prominent part of the decoration. But the discrepancy applies to the entire ceiling. What had always struck observers in the past, even in Michelangelo’s own lifetime, was that he was not much of a colorist, but excelled in creating an impression of relief. This view seemed to fit very well with a famous comment he made in a letter written for publication in 1547, when asked which was better, painting or sculpture:


I say that it’s my impression that painting is more esteemed the more it tends towards relief, and relief [sculpture] is considered worse the more it tends towards painting; and therefore I used to suppose [i.e., before I read your letter] that sculpture was the lantern of painting, and that the difference between the two was like that between the sun and the moon.1

For Michelangelo, then, painting was at best a pale reflection of sculpture. Commenting on this passage, probably in 1564, Vasari’s great friend Vincenzo Borghini remarked that the artist’s critics,

partly in jest, but also partly with an element of justified scorn, say that his colors made him choleric, since he could never master them or control them as he wished; and it is certainly the case that in his paintings he was more concerned with relief than with coloring. 2

Equally striking is the testimony of Paolo Giovio, writing between 1524 and 1529. He said nothing about the colors of the ceiling, but praised Michelangelo primarily for his handling of light and shade, mentioning in particular the body of Holofernes, receding and concealed in gradually diminishing light, and the figure of Haman, on which the light was emphasised by a contrast with the shadows. After the cleaning, these observations no longer seem very apposite. But there are several other passages in sixteenth-century texts in which Michelangelo’s qualities as a painter were characterized in much the same way.

Remarks of this kind were easy enough to understand in the context of the ceiling before cleaning, but unless we believe that the tonality of the fresco was drastically altered within a few years of its completion, leaving no memory of its original appearance in the minds of Michelangelo’s contemporaries, their observations seem extraordinary misleading. The only significant piece of evidence for such a change is provided by a German visitor to Rome in 1536, Johann Fichard, who said that the colors of the paintings in the Sistine chapel seemed to be very much darkened, and that this was doubtless due to soot. That he was struck by their dark tonality is clear, but whether he was right about the cause is more questionable. Fichard had never been in Rome before, and the fact that he ascribed the frescoes to Raphael suggests that he had not investigated their history very closely.

Defenders of the restoration, however, have argued for the authenticity of the present coloring on the grounds that it was of decisive importance for younger Florentine painters such as Pontormo and Rosso. Yet the analogy proposed with Pontormo’s exceptionally damaged frescoes at Galluzzo is hardly convincing, since these have certainly lost a great deal of secco. The resemblance of the ceiling to the better preserved Florentine frescoes from this period is much less convincing, while the absence of anything comparable in frescoes painted in Rome is also striking. As a colorist, Michelangelo now seems almost closer to Disney than to his contemporaries.

The clearest evidence that the ceiling today does not conform to its original appearance is provided by early copies. An anonymous and incomplete copy of most of the vault and lunettes, at Windsor, and a detailed drawing by Giulio Clovio of the seated figure of Jonah, both showing part of the altar wall before Michelangelo painted the Last Judgment and therefore dating from no later than 1534, include a very prominent shadow behind Jonah’s left foot. (See illustrations on this page.) Most of this shadow, regularly shown in later copies, disappeared during the present cleaning. Likewise a substantial area of shading on the right knee of Ezekiel, clearly visible already in sixteenth-century engravings, has also vanished, and similar losses are evident throughout the ceiling. It is difficult to see how such shadows could all have been due to dirt or repainting, as supporters of the restoration would have us believe.

Quite soon after the cleaning began there were protests from prominent painters on both sides of the Atlantic, but unfortunately art historians and restorers do not normally give much weight to the views of such people. Another vocal critic was an Italian academic named Alessandro Conti, whose book on the restoration was published in 1986, 3 and whose assistance is warmly acknowledged by Beck. The basic thesis of both writers is that Michelangelo used secco extensively and that this was eliminated by the cleaning; the presence of the glue was not simply due, as the Vatican authorities claimed, to the intervention of previous restorers attempting to enliven the colors or repaint isolated areas of damage.

The present restorers, by contrast, believe that Michelangelo used secco only for corrections and for the limited application of pigments that could not be put on in any other way, notably gold and some blue. Their claim that virtually all the glue must have been added later is based on evidence of three kinds: first, the testimony of the artist’s earliest biographers, Vasari and Ascanio Condivi; second, information discovered about previous restorations; and third, physical data derived from the frescoes themselves.

Condivi in 1553, followed by Vasari in 1568, certainly said that Michelangelo painted the ceiling in fresco, and was prevented from putting “the final touches” (l’ultima mano) on the first half of the vault and some gold and blue (which would necessarily have involved secco) on the second half, because of the impatience of the pope. These comments do not necessarily indicate, as has often been argued, that Michelangelo used no secco at all, except for small areas of gold and blue; they could simply indicate that he did not use as much as he had intended.

However, supporters of the cleaning also drew attention to Vasari’s more general remarks, elsewhere in his book, about the merits of buon fresco and the shortcomings of secco, and argued that Michelangelo, as a good Florentine, would have used the less “manly” (virile) and durable technique of secco as sparingly as possible. But Conti pointed out that Vasari’s comments were made in the context of the debate about the relative merits of painting and sculpture, and that his emphasis on the difficulty of buon fresco, which calls for an assured technique and experience in anticipating how colors change when they dry, needs to be seen as a polemical point, but not necessarily as an account of normal pictorial practice. In fact, as has recently been established, Michelangelo certainly used secco extensively on the Last Judgment, but this was not mentioned by Vasari, who instead emphasized his insistence on painting it in fresco.

In fact, secco painting was widely practiced in the Renaissance, as we know both from Armenini and from the treatise of Cennino Cennini, written almost certainly before 1427. Neither writer shared Vasari’s view that secco tended to turn black, a comment which might reflect his personal experience and could also in part explain his own preference for a procedure involving some use of oil, rather than traditional methods of fresco. The written sources, then, certainly do not provide clear evidence that Michelangelo himself had any reservations about using secco on the ceiling. Indeed, Condivi’s comment about the final touches cannot easily be reconciled with the idea that he objected to this technique in principle. In this context it is significant that shortly after starting work on the ceiling he ordered some lake, a dark red pigment, from Florence. This can only be applied a secco, and Armenini mentions that, mixed with black, it was applied as a wash to retouch areas of shadow, adding that some artists combined the technique with the use of glue.

It is reported that discontinuous layers of glue were to be found all over the ceiling, but the idea that all of this was applied by past restorers is impossible to substantiate from the surviving records. Limited repairs were undertaken in the 1560s, and it has been argued that green repainting applied to some blue drapery in the Sacrifice of Noah implies that this area was already covered by glue that had turned yellow; but it has not shown whether this had been applied by Michelangelo himself, or later. Further restoration was done in the chapel in 1625, but there is no evidence that this extended to the ceiling or involved glue. In 1710–1712 the ceiling was cleaned by a minor painter named Annibale Mazzuoli, using sponges soaked in wine. Mazzuoli evidently did some repainting, which could have involved limited application of glue. It is possible that he covered large parts of the vault with this material, in order to protect the surface of the plaster and perhaps to brighten the colors, and equally possible that he removed some that had been put on earlier; but there is nothing in the documents to indicate that any of this happened, let alone any contemporary comment that the appearance of the ceiling was radically changed. Nor is there evidence of any drastic intervention at a later period. Mazzuoli, then, is the only credible suspect for the extensive addition of shadows, but his guilt is unproven, and in any case many features that have now disappeared can already be seen in copies made before his intervention, as has been mentioned above.

The technical evidence is likewise less conclusive than has sometimes been supposed. Cross-sections taken before cleaning reveal a thin layer of soot beneath the lowest layer of glue, more explicitly described in the case of the lunettes as “a thin, but very tenacious layer of grey particles, consisting of fatty dusts and soot.” The implication is that dirt had built up before the glue was applied, and that this cannot therefore be the work of Michelangelo. Before accepting this conclusion, one would like to know more about the very tenacious layer of gray particles; and it should be pointed out that the presence of soot could merely indicate that dirt settled on the plaster in the period between Michelangelo’s application of buon fresco and secco, or even that he used lamp black in his secco pigment.

In the opinion of the restorers, however, the strongest evidence that Michelangelo only painted in buon fresco comes from a discovery made in the later stages of the cleaning, namely a fragment of fresco from the torso of one of the seated nude youths, which was covered by plaster during the repairs of the 1560s. There is no glue on it at all. But this proves very little, since it is not suggested that Michelangelo covered the entire ceiling with secco, or with glue. Charles Heath Wilson, who examined the frescoes from a scaffold in the last century, noted the presence of extensive secco, which he confidently ascribed to Michelangelo himself, but he specifically indicated that he had failed to find traces of this on the seated nudes.

The whole question of later repainting may be settled when a full report of the restoration is published, because at the moment we have no more than isolated pieces of technical evidence released in a piecemeal way. Until such a report becomes available, I am inclined to believe that Conti and Beck are right in claiming that Michelangelo did use some kind of secco to modify the bright tones of his buon fresco, and in particular to enrich the shadows and enhance the modeling. This seems to me the only reasonable way of accounting for the comments of sixteenth-century observers, for the loss of shadows shown even in copies predating the Last Judgment, and for the depressing effect that the ceiling makes today. This is not to deny that some of the glue may have been added after Michelangelo’s death or that the ceiling had become darker over the centuries. But it is at least open to question whether this darkening justified the removal of almost everything except the buon fresco.

It has been claimed that this was done not just to reveal Michelangelo’s own work, but also because the glue was itself pulling away fragments of the underlying pigment, necessitating urgent intervention. The problem seems to have been noticed only after the cleaning of the lunettes, and one wonders whether it was quite so urgent as was feared and whether it called for such radical treatment. Beck and Daley go so far as to suggest that, far from protecting the frescoes, the restoration has itself exposed them to further damage. The choice of solvent, AB57, caused some disquiet from the first, especially since it was developed for cleaning marble statues rather than frescoes. In places it was used in conjunction with a solution of dimethylformamide; and Beck mentions a suggestion by an expert on restoration that the brightness of the cleaned paintings may be partly due to a reaction provoked in the plaster by this combination of chemicals. After cleaning, the lunettes were then covered with an acrylic resin called Paraloid B72, but doubts were also soon expressed about the wisdom of this treatment, which was subsequently abandoned. Most of the ceiling, which had survived for centuries under the glue, now has no protective covering at all, and the consequences, in the polluted atmosphere of the chapel, remain to be seen.

With the benefit of hindsight, it would surely have been better, if only to set everyone’s mind at rest, if a careful examination of the entire ceiling, the Last Judgment, the copies, and the available records of earlier restorations had been made before a decision was taken to proceed. A more cautious approach of this kind has recently been adopted for another Renaissance masterpiece. Over the past few years a detailed study has been made of Piero della Francesca’s cycle of frescoes in Arezzo, as a preliminary step toward preventing further deterioration. This has revealed an immense amount about the history of the chapel, its micro-climate, and the technique of the paintings themselves. In the present context, the most remarkable discovery is that Piero used secco far more extensively and in much less conventional ways than had previously been suspected, often employing radically different methods even within the same scene. Although most of his work remains, much has disappeared, for example the leaves that once covered a huge tree in the Death of Adam, of which only tiny traces of green pigment still survive. This particular finding has some bearing on another restoration project discussed by Beck, that of the frescoes by Masaccio and Masolino in the Brancacci Chapel in Florence.

Here, as in the Vatican, the frescoes were covered with a thick layer of dark material, whose removal has revealed the colors of the buon fresco, as well as exposing certain features, such as clouds, that were previously invisible. But the results have not been entirely positive. Now, for example, there are troubling differences in tone between adjacent giornate, the areas of wet plaster painted in a single session, of a kind that it is difficult to believe that the artists or their patrons would have considered acceptable. Various areas of secco have also been removed, particularly foliage; and although it has been claimed that they were all later additions, we cannot say whether they replaced earlier painting of the same kind. Particularly in the light of the discoveries at Arezzo, it seems odd that the trees in Masaccio’s Tribute Money should be almost bare. Yet in their treatment of secco the restorers have not been consistent: Adam and Eve have lost their fig leaves, but the haloes of the Apostles have been replaced and some steps entirely repainted on the basis of indications found in the underdrawing.

The cleaning of the Brancacci Chapel has produced less disturbing results than that of the Sistine Chapel, but both demonstrate that restoration can have negative as well as positive consequences. No one would deny that we can see more of the buon fresco of Masaccio and Masolino than before, and more of Michelangelo’s ceiling. But we cannot be confident that the present appearance of either work would have given much satisfaction to the artists themselves, or even that they would have regarded the recent changes as improvements. Recalling with some shame my early support for the cleaning of the ceiling, I now find it difficult to believe that the right procedure was adopted. If there are lessons to be learned from this unhappy episode, they are that we ought to be prepared to tolerate darker pictures, and that in restoration the wisest course is often the most conservative. By reminding us of these lessons, in the face of hostility from his professional colleagues and even a threat to his liberty, Beck has done a valuable service for everyone who cares about the art of the past.

This Issue

November 18, 1993