Short answer
Cleaning removes material that lies over the paint — discoloured varnish, soot, grime, earlier restorers' coatings and retouchings — and the colour beneath is usually cooler, brighter and higher in contrast than anyone living has seen. It is contested because it cannot be undone, because deciding which layers are later additions is a judgement, and because the result replaces an appearance people knew. The cases on record include disputes that inquiries settled in the institution's favour and one where a treatment visibly created a new problem. This page explains the evidence; it gives no method, and the decision belongs to trained conservators.
Three kinds of layer are involved. Natural-resin varnish oxidises to yellow and then brown, filtering every colour under it. Soot and dust settle on any surface near candles, lamps or city air. And restorers add things: glue or oil rubbed on to 'revive' dull colour, toned varnishes, retouching that later discolours. On the Sistine ceiling the chief restorer's account names all three — smoke, dust and, above all, animal glue with oil, gum arabic and ox gall applied in past campaigns, partly to hide salt efflorescence. A discoloured layer is not purely an optical nuisance, though. The National Gallery's vermilion study observes that a thick yellow varnish filters light and keeps dirt, and with it chloride, away from the paint; thin, clear varnishes let more light through. Removing a layer changes what the paint is exposed to as well as how it looks.
A yellow-brown film subtracts blue from everything beneath it, so skies read greenish, whites cream and shadows merge. The eye adapts to a uniform cast and stops registering it, which is why a half-cleaned picture looks more startling than either state alone. When the cast goes, hue relationships the painter set up between cool and warm passages return, and so does contrast in the darks. Viewers who learned the picture in its darkened state can experience the result as loss — of unity, of 'tone', of age. That reaction is not simply ignorance. A cleaned surface may reveal pigments that have themselves altered, so that the newly exposed colour is neither the darkened one people knew nor the one originally painted. Smalt that has gone grey, faded red lakes and browned copper greens all look more obviously wrong once the unifying yellow is gone.
In London, the National Gallery's 1947 exhibition of cleaned pictures 'unleashed a barrage of criticism', in the words of its own archive: works by Rubens, Koninck, Velázquez and Rembrandt were alleged to have been spoiled. A government inquiry that year and an internal investigation in 1951 both found for the Gallery. In Rome, the cleaning of Michelangelo's Sistine frescoes between 1980 and 1994 exposed colour so bright that it changed how his painting is described; critics led by the art historian James Beck argued that the restorers had removed shadows Michelangelo himself added a secco. And in Los Angeles in 2012–14 the Getty removed from Pollock's Mural a varnish applied during a 1973 treatment, which had dulled the surface: there the layer taken off was an earlier conservator's well-intended work.
A varnish is meant to be removable, and the American Institute for Conservation's guidelines ask that materials be chosen with their effect on future examination and treatment in mind. But removal itself goes one way. If a layer held original material — a glaze, a final a secco touch — it is gone, and the argument over whether it was original can no longer be tested on the object. Paint also becomes more vulnerable as it ages. The National Gallery notes that the soft lead-soap inclusions in old oil paint are easily damaged during cleaning, and that solvents and humidity changes could further alter the fragile, water-attracting crust on degraded smalt. The Canadian Conservation Institute states that varnish removal is not always advisable or necessary and must only be done by someone with conservation training, because solvents can irreversibly damage paint.
Several records in the dataset are sequences rather than single events. At Quaracchi a consolidation treatment turned green-altered azurite dark blue, and the blue changed again within two years; analysis showed the treatment had produced unstable copper hydroxide. On the Sistine ceiling, glue applied to mask salt bloom became the dark coating a later generation removed. Pollock's Mural was stabilised in 1973 at the cost of a dulled surface and a fixed sag. None of these was careless by the standards of its day. They are the reason professional codes ask for every intervention to be documented and for each method to be weighed against its effect on future treatment — and the reason nothing on this page should be read as guidance on doing any of it.
| Object | What changed in appearance | Method | Rests on | What is disputed or uncertain |
|---|---|---|---|---|
| Michelangelo, Sistine Chapel ceiling and Last Judgment (Vatican Museums) | Frescoes seen 'as if through a smoked glass' emerged with bright, strongly contrasted colour. | Cleaning of soot, dust and earlier restorers' glue coatings by the Vatican's own restorers. | The restorers' technical examination, which attributed the dark layers to smoke and to glue applied in past restorations. | Critics, notably James Beck, argued that Michelangelo finished shadows a secco in carbon black and glue, and that the cleaning removed his work along with the dirt. The restorers held that the layers were later. The removal cannot be undone, so the dispute cannot be settled by looking again. |
| Paintings shown in the 1947 exhibition of cleaned pictures, including works by Rubens, Koninck, Velázquez and Rembrandt (The National Gallery, London) | Paintings returned from wartime storage were exhibited after cleaning, looking unlike the darkened pictures the public knew. | Cleaning in the Gallery's newly founded Conservation Department. | The Gallery's own records of the treatments and the two inquiries that followed. | It was alleged that four named pictures had been spoiled. A government inquiry in 1947 and an internal investigation in 1951 found for the Gallery; the Gallery's director remained in the argument over picture cleaning for the rest of his career. |
| Jackson Pollock, Mural (1943) (University of Iowa; treated at the J. Paul Getty Museum and Getty Conservation Institute) | A surface dulled by a varnish applied in 1973 was changed visibly in depth and surface character once that varnish was removed, by the Getty's account. | Removal of the aged varnish by the Getty's conservators, with a new shaped stretcher. | Technical study of Pollock's materials and of the 1973 treatment. | Not disputed in the source. It is recorded here because the thing removed was itself a conservation treatment: the 1973 work stabilised the paint but dulled the surface and fixed a sag in place. |
| Wall painting of San Antonio Abate, church of San Pietro at Quaracchi (Near Florence) | A green-altered area of azurite turned dark blue after consolidation, then changed colour again within two years. | A standard two-stage consolidation treatment used on wall paintings. | Infrared spectroscopy and X-ray diffraction of the altered and treated areas. | The study's own conclusion is that the procedure is questionable where copper chlorides are present: the 'recovered' blue was unstable copper hydroxide, not azurite. |
| Painted figures of the Terracotta Army (Lintong, Shaanxi) | Polychromy that would otherwise be lost with its ground is kept in place. | Consolidation of the paint layers by polymerising a monomer mixture within them. | Chemical study of the deteriorated organic ground layer. | Introducing a polymer into the paint is not something that can be taken out again, which sits uneasily with the reversibility that conservation codes ask for; the authors present it as the answer to an otherwise unsolved problem. |
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactStrong evidence
The National Gallery's 1947 exhibition of cleaned pictures drew allegations that works by Rubens, Koninck, Velázquez and Rembrandt had been spoiled; a government inquiry in 1947 and an internal investigation in 1951 both found in the Gallery's favour.
Caveat: The Gallery's own archive summary of a dispute to which it was party.
Source: Cleaning and conservation (1946–1967), archive series NGA3/2/2
FactModerate evidence
According to the chief restorer, the Sistine Chapel frescoes were obscured by dust and candle smoke and above all by animal glue with oil, gum arabic and ox gall applied by earlier restorers, partly to hide white salt efflorescence from rainwater infiltration.
Caveat: A participant's account of the 1980–1994 restoration.
Colourwise interpretationModerate evidence
Critics of the Sistine Chapel restoration, including James Beck, argued that Michelangelo had applied shadows a secco in carbon black and glue and that the cleaning removed them; the restorers attributed the dark layers to later accretion. The question remains disputed.
Based on: Colourwise's summary of two opposed published positions: the restorer's account in a journal article and the critics' case as reported in a reference work. Colourwise takes no side.
Caveat: Disputed; the critics' position is cited second-hand.
Source: Restoration of the Sistine Chapel frescoes; Michelangelo Buonarroti: Restoration of the frescoes on the vaulted ceiling and the Last Judgment in the Sistine Chapel (Conservation Science in Cultural Heritage, 2017)
FactStrong evidence
A varnish applied to Jackson Pollock's Mural during a 1973 treatment had dulled its surface; conservators at the Getty removed the aged varnish during treatment begun in 2012.
Source: Getty Completes Conservation of Milestone Painting by Jackson Pollock
ConventionStrong evidence
Removal of discoloured varnish is not always advisable or necessary and must only be carried out by someone with conservation training, because solvents can irreversibly damage paint.
FactModerate evidence
A thick, discoloured varnish can act as a light filter and a barrier to dirt for vermilion paint beneath it, so removing it changes the paint's exposure as well as its appearance.
Caveat: An observation in the study's discussion; the authors list the protective role of varnish as an open question.
Reviewed 6 October 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.