WHEN the UK's National Gallery in London was bequeathed an exquisite painting of the 'Virgin and Child with an Angel' in 1924 officials must have been delighted an early 16th-century masterpiece by Francesco Francia, the artist from Bologna, was to grace the museum's collection.

Until, that is, an almost exactly similar work turned up for auction in London in 1954. Problem which was the original and which a copy? In 1998 it looked like the London painting had been accepted as genuine. But recent research has been carried out, and the picture examined using infrared reflectography. That technique revealed what lies beneath the paint.

According to Rachel Billinge, a researcher at the gallery “We could see little dots, indicating that the image had been 'pounced' from a cartoon, which is a perfectly good Renaissance technique. But then I looked at the hair of the angel, and saw what looked like graphite pencil marks.”

And that was the worst possible news. Graphite was available in only one place in the early 16th century Cumbria. The lovely pencilled curls could not have been drawn by an Italian in the 16th century and the work could not be an original. In fact, it was probably made in the 19th century.

This and other fakes and mistakes from the gallery's collection are to go on show in a major exhibition called 'Close Examination' from this June. The gallery will be dusting off some of its most embarrassing acquisitions.

In 1874, for instance, the gallery successfully bid for two Botticellis. One of them, 'Venus and Mars', is now one of the museum's most famous and recognisable works. The other is not. Soon after its purchase it was quietly removed from the walls of the gallery and attributed to an anonymous follower of the master.

The exhibition will focus attention on the role of the gallery's scientific department, which has pioneered the latest techniques in infrared imaging as well as x-ray techniques, pigment analysis and dendrochronology, a technique whereby wood can be dated by examining its rings.

The show will explore how such techniques allowed scholars to reveal how a Giovanni Bellini portrait of a friar was over-painted with the emblems of St Peter Martyr and how one portrait, acquired in 1990, had been tampered with — given a bright-blue background using a pigment not available until the 18th century — to make it resemble a more valuable Holbein, perhaps by an art dealer on the make.

All this leaves Billinge with a healthy respect for the copyist and faker. “Sometimes the faker has gone to such lengths you can respect their techniques — much more so than the originals, churned out in a workshop by some bored apprentice,” she said.

— The Guardian, London