As a restorer who makes his living cleaning and restoring old master paintings, it is my intention in these notes to explain what this work entails – what can be done to prolong the life of a work of art whilst ensuring that its appearance remains as close as possible to its original condition – and to show the limitations of this craft.

Art restoration falls into several categories which can become confused, especially where a client asks for the repair of the irreparable or is disturbed by the revelation of brilliant colours which have been obscured by centuries of dirt, for example copal soot, candle smuts, tobacco smoke etc.

Conservation

Conservation is an aspect of the restorer’s work which is often taken for granted and is – as it should be – largely invisible. Its essence is the prevention of further decay if possible. A work of art which has been gravely damaged, for example by fire or water, may not be capable of conservation, and in this case the remedy is sympathetic reconstruction – always maintaining the integrity of whatever original material remains.

The old adage of not patching old garments with new cloth is true of works of art, and the introduction of any new material, such as relining a canvas, has to be made with the greatest caution. Some years ago, for example, it was thought wise to use wax as an adhesive to stick new canvas to the back of the old. This was believed a reliable method until it became evident over time that the wax tended to saturate the old canvas, producing a curiously flattened effect from the front. Thus what was believed a ‘safe’ method in conservation terms had a bad effect on the painting’s appearance.

Panels

Panels cause far more problems for the conservator than canvasses, since the latter’s weakening can normally be resolved by relining. Panels age in various ways – the way they have been treated in the past, the type of wood from which they are made and, above all, the way in which the planks are cut from the parent trunk.

In northern Europe until the end of the seventeenth century oak was the preferred and generally durable wood. This was gradually replaced by the more expensive (because imported) but equally durable mahogany. Most really serious problems with panels come from the softwoods used by the Italian artists. So many early Italian paintings have been subjected to somewhat violent treatment – sawing down the middle when the panel was painted on both sides, panels cut down to fit new frames or decor etc. Often, they have been subjected to centuries of injudicious cleaning to remove candle smoke (particularly insoluble in all reagents and solvents) – a serious problem for altarpieces.

Panels are also vulnerable if subjected to extreme changes in humidity, stable in either slightly damp or very dry environments (England versus Italy), their structure can become weakened if moved from one environment to the other and they can, quite literally, split apart suddenly, or the paint can flake and fall off the surface.

Canvas

Whilst canvas is a less vulnerable material, relining is normally required every few generations as the fibres weaken. An exception of a seventeenth-century unlined canvas is Vermeer’s Guitar player (London, Kenwood House, Iveagh Bequest). This gives the picture a freshness and vibrancy often lost when canvasses have been relined by the application of heavy irons. The unique state of this painting has sometimes led art historians to find the picture disconcerting. Canvasses can also be subject to physical damage, for example piercing by a sharp object, or damage by falling from a wall.

The remedies

Modern methods for relining canvasses, which require less heat and pressure, make the process much safer. Panels have often suffered in the past where attempts to halt decay have actually accelerated the process. The old practice of cradling – the application of sliding bars across the grain supported by bars glued along the grain – is now considered too great an intervention and is, of course, irreversible. In this case bulk rigidity is lost and the panel shrinks further across the grain.

Most vulnerable are large format panel paintings where the whole is made up of separate planks. Rubens favoured this method, using panels of the finest oak available at the time, but tension between the panels has caused the most serious problems. The most famous example of this was the splitting of the panel Chateau de Steen (London, National Gallery), caused by the failure of the central heating on a particularly cold night.

The restorer’s dilemma – paint loss

Most Old Masters have suffered some form of paint loss, generally minor. This is usually caused by scratches or abrasions or localised flaking. Great skill is required to correct this loss but it poses no ethical or intellectual problems. The dilemma begins where a significant area is lost, it may be quite small – a grimacing face in a Dutch genre piece, a few delicate petals in a flower piece – it is rarely a whole face or flower. In my experience skill, intuition and careful observation of other of the artist’s works, make it possible to reconstruct the missing part.

Although this may seem a sensible approach, there are strongly held alternative points of view. For example, the Yale University Art Gallery at New Haven, Connecticut, insists on leaving all paint loss plainly visible. This is just tolerable where loss is minor, but major damage becomes disfiguring and interferes with the appreciation of what remains. A compromise is possible, as with the major loss on a Duccio panel (London, National Gallery), where no evidence remains of what has been lost. The missing areas can be outlined so that the eye is not automatically drawn to the gap. This is not the end of the restorer’s dilemmas. The task of cleaning off both surface dirt and discoloured varnishes to reveal the painting as nearly as possible as the artist left it. With Dutch seventeenth-century works this is largely achievable owing to sound techniques and the care which has been taken with small pictures which have always been valuable. French pictures present more problems as they are often unevenly preserved, thus forcing the restorer to replace a glaze here and there, or, sometimes controversially, to tone down a brilliantly preserved detail in order to harmonise with the rest of the picture. This is a very difficult area and many would argue that it is an unethical practice. It is, however, the finished effect that is all-important and once a picture is severely out of balance, it is the restorer’s duty to take decisions about such matters as harmonious effect.

Specific artists create very special problems. For example, almost all the greens in the trees in Claude Lorraine’s landscapes have turned brownish through chemical deterioration. If a small piece of a tree is missing, the purist would insist on using the green known to have been used originally to replace it. Visually this would be a disaster for the harmony of the picture and the restorer must therefore take account of the passage of time when making his adjustments.

Glazes form the most difficult of the restorer’s decisions based on judgement. They are technically easy to apply as they consist of a more or less evenly spread layer of transparent colour over underpainting. The problem comes with knowing how much and what colour to apply. Tragically, few old master paintings retain their original glazes and almost all sixteenth- and seventeenth-century French and Italian paintings depend on them. The Dutch used glazes much less frequently and it is usually possible to gauge what to do by looking at similar works.

For other schools there is much less certainty even when there has been a policy of minimalist intervention. This is best seen in the early Flemish section of the Musees Royaux des Beaux-Arts in Brussels. It is likely that many of their early paintings retain their original glazes, but also retain later, and largely yellowed, varnishes. Some people claim that removing a yellowed varnish risks removing the glaze as well, and this was certainly true in many misguided eighteenth- and nineteenth-century restorations. Nevertheless, the effect on the Brussels collection is that the paintings look quite unlike similar works in the National Gallery, London, where there is no yellowed varnish on the early Flemish pictures.

Old restorations

The intervention of previous restorers is one of the most delicate problems for the restorer. It may not always be wise to remove the work of a previous restorer where it was well done; the old restoration may be in need of restoration itself (in one notorious case a restorer removed the retouching of Ruskin himself from an Italian panel and was left with a ruin which Ruskin had skilfully reconstructed).

The most unfortunate example of restorers’ misguided intervention tend to be in the private sector, where a whole collection suffered from the ministrations of the same person. Thus the collection of Old Masters given by General Guise to Christ Church at Oxford in 1765 suffered from cleaning by a restorer called ‘Old Bonus’. There were complaints at the time but there is no doubt that the pictures suffered, although modern restoration has minimised these effects.

These early methods often employed abrasives to remove varnishes, with the result that many Old Masters have acquired a rubbed effect. Sometimes early restorers who caused damage by abrasion covered up their efforts with well-handled overpainting. This can, on occasion, deceive even a well-trained eye, especially if the disguising overpaint is more than two hundred years old.

Dirt

As mentioned earlier, dirt comes in many forms. On cleaning George Stubbs’s Whistlejacket I had to contend with coal dust on its surface – its stately environment at Wentworth Woodhouse was in the heart of the South Yorkshire coal fields. The deleterious effects of the sulphur in coal smoke affect many of the paintings kept in the art galleries of Victorian industrial cities, and white lead pigment in Old Master paintings can change chemically with a tendency to blacken. This problem is now, fortunately, rare.

It was the custom in country houses to revarnish paintings every few years in a kind of ‘spring clean’ (sometimes frames were confused in this process which later caused problems with identification of artists and sitters). Paintings gradually acquired what became known as ‘gallery tone’ which meant that they tended to resemble each other, acquiring an even, golden glow. This expectation of ancient glow has coloured the attitude of those opposed to modern methods of cleaning and restoration. Microscopic inspection before the cleaning process starts, can often identify the different layers and can reveal tiny fragments of a long lost glaze.

The restorer is sometimes confronted with a past intervention which was intended to deceive. This may take the form of a fake signature which disappears in varnish removal, when it is revealed that it is sitting on a layer of overpaint – sometimes in such cases the remains of the real signature may appear elsewhere! The old repaint may be misleading in the form of a changed expression – from a grimace to a smile, or may involve a question of propriety – nudes given draperies. On occasions, lowering skies have been given a powder blue coat of paint to ‘cheer up’ a melancholy landscape (very common in the monochrome estuary and river views of Jan van Goyen). Our Victorian forefathers developed endless remedies of this type in order to ‘beautify’ or render less aggressive the endless variety of the Old Masters.

Pigment analysis can of enormous help when the restorer is presented with a difficult decision as to what may or may not be later additions, particularly when these are close in date to that of the painting. For example, the use of cheap, eighteenth-century ‘Prussian’ blue instead of the expensive Azurite or Lapis Lazuli becomes very obvious under analysis. It was van Meegren’s downfall that his pigment supplier diluted his Lapis Lazuli with the early nineteenth-century invention, Cobalt Blue, thus exposing his fraud.

In the final analysis, all the in-painting I have to do is readily removable as it is bound in an easily soluble medium. Certainly I spend a good deal of time researching the state of the painting I am about to restore before commencing work. I also seek the advice of art historians if I have any doubts which may be solved by academic discussion. My decisions are never automatic as each painting on which I work has its own identity, and of course, its own restoration history.

Popularity: 14% [?]