Earliest copy of Mona Lisa found in Prado

Experts say the painting was completed at the same time as Leonardo’s original


The restoration of a copy of the Mona Lisa in the Prado in Madrid has revealed the work to be a contemporary version of the masterpiece, created in the artist’s studio at the same time that Leonardo was working on the original. Scholars believe the two versions were completed side by side.

The discovery was announced at a conference at the National Gallery in London last month, coinciding with its Leonardo exhibition (“Leonardo da Vinci: Painter at the Court of Milan”, until 5 February). The replica promises to tell us more about the master’s Mona Lisa, since there are areas of the painting that are better preserved than in the original at the Louvre in Paris.

Until recently, curators at the Prado had no idea of the significance of their copy. Although the portrait is finely painted, it lacks Leonardo’s delicate landscape and had a dull, black background. Visually, this had a deadening effect on the image of the female figure, believed to be Lisa Gherardini.

Last year, conservators at the Prado discovered that the black background was overpainted, and beneath it lay a landscape. During the past few months this covering has been painstakingly stripped away at the Madrid conservation studio, with the final area of overpaint about to be removed as we went to press. This has revealed what Prado technical specialist Ana González Mozo describes as “a high quality work”.

A photograph was presented at the London conference, revealing the picture’s condition after 90% of the black paint had been removed, leaving just a small section in the upper right corner.

The restored painting is expected to be unveiled in Madrid later this month. It is then due to be loaned to the Louvre, as a late addition to its exhibition, “Leonardo da Vinci’s Saint Anne” (29 March-25 June).

There are dozens of replicas of the Mona Lisa from the 16th and 17th centuries. Although the Prado version was considered by some specialists as early, the absence of the landscape meant that it has, until now, aroused little interest (there is no substantive entry on it in the Prado’s collection catalogue). The Prado and the Louvre now both accept the work to be the earliest copy of the masterpiece.

The Prado picture was believed to be painted on oak, and it was assumed to be by a northern European artist. José Ruiz Manero, the author of a study of Italian art in Spanish collections, concluded that it was Flemish.

The support was examined in 2011 and found to be walnut, which was used in Italy (as was poplar, used for the original Mona Lisa). The works are also close in size: the Louvre’s painting is 77cm x 53cm and the Prado’s copy is 76cm x 57cm.

As well as the hidden landscape, there was an even greater surprise to come: infrared reflectography results were compared with those obtained in 2004 when the original Mona Lisa was examined.

The underdrawing of the replica was similar to that of the Mona Lisa before it was finished. This suggests that the original and the replica were created next to each other at the same time, as the composition evolved. Leonardo worked on the Mona Lisa in Florence, at some point between 1503 and 1506.

Bruno Mottin, the head conservator at the Centre de Recherche et de Restauration des Musées de France, says that the most likely painter of the Prado copy was one of Leonardo’s two favourite pupils: Andrea Salai (who originally joined his studio in 1490 and may have become his lover) or Francesco Melzi (who joined around 1506).

Prado curator Miguel Falomir believes the replica can probably be identified with a portrait listed in a 1666 inventory of Madrid’s Alcázar Palace, with the black overpaint probably added in the mid-18th century. The reason for this addition remains obscure, since the landscape is in good condition. Possibly the overpaint was to integrate the picture into an interior decoration.

The replica offers important insights into Leonardo’s original. In the copy there are areas that are better preserved than in the Louvre painting, giving more detail of the spindles of the chair, the frill on the edge of the fabric on Gherardini’s chest, and the veil around her left arm. It also reveals more about Leonardo’s studio practice and how a key pupil would work on a second version alongside a master.

Examining Leonardo’s paintings

Technical investigations reveal fresh evidence about some of the artist’s best known works

london. The conference on Leonardo’s technical practice, held at London’s National Gallery on 13-14 January, probed beneath the surface of some of the master’s works.

The Louvre’s Virgin of the Rocks (left) had been assumed to be in poor condition, but the recent infrared investigation is encouraging. New images show “less damage than had been feared”, according to the conservator Bruno Mottin. Infrared also reveals pentimenti in the figure of the angel. Louvre curator Vincent Delieuvin concluded that “restoration of the painting is possible.” Cleaning is not yet on the agenda, although Leonardo specialist Martin Kemp did speculate that the French specialists “might be softening us up for an announcement”.

The recent cleaning of the Louvre’s Virgin and Child with Saint Anne has proved highly controversial (see p8). The key questions are whether Leonardo’s glazes may have been partially removed and how much of the varnish was removed over the ultramarine. The painting is due to be unveiled at an exhibition at the Louvre on 29 March.

There was a debate over the two versions of The Madonna of the Yarnwinder: the Duke of Buccleuch’s picture (centre, stolen in 2003, recovered in 2007, and currently in the National Gallery exhibition) and the version formerly owned by the Lansdowne family (sold by Wildenstein in 1999 to a New York collector and exhibited last spring at the Art Institute of Chicago). An infrared examination was recently undertaken on the Lansdowne Madonna by Florence’s Opificio delle Pietre Dure. Superintendent Cristina Acidini accepted the work as authentic and reported evidence of a basket, which the artist had originally depicted at the foot of the yarnwinder.

The National Gallery accepts the Buccleuch version as the primary version and the Lansdowne work was not requested for the exhibition. Martin Kemp and Thereza Wells, in their new book, Madonna of the Yarnwinder, argue that Leonardo was involved with both versions—a finding that seems to be confirmed by the Florence examination.

A New York conservator, Dianne Dwyer Modestini, made a presentation about the newly discovered Salvator Mundi (right). We can therefore publish for the first time an image of the painting when it was sold in 2005 in what is believed to have been a minor auction in the US. The picture presumably looked very similar when it sold at Sotheby’s in 1958 as a work “after Boltraffio” for £45. Among the interesting points is that Christ had a beard (not present in the original painting, according to Modestini)—presumably to make it more similar to a Hollar engraving of 1650. It is assumed that Hollar misinterpreted the dark area under Christ’s chin as a beard. The painting is likely to go into store when the National Gallery show closes on 5 February and it may soon come onto the market.

In a report on drawings, Carmen Bambach, a specialist at New York’s Metropolitan Museum, said that the disputed sketch of Madonna and Child with St Anne and a Lamb is on paper with a watermark of a head in profile, which she believes was only used from 1579. She also argues that the composition is based on an engraving by Antonio da Trento, who died in 1550.

If correct, the drawing, still presumably owned by dealer Simon Dickinson, cannot be a Leonardo. It was previously valued at more than $7m (The Art Newspaper, February 2011, pp57, 59). The Louvre is holding a technical symposium on Leonardo on 20 June.