A load of old Pollocks? Thirty-two 'works' found in an attic now suspected to be fakes

They may have looked like the demented drips and flicks of a child or a lunatic, but scientific analysis has shown the art of Jackson Pollock is anything but.

Now the techniques that revealed distinctive patterns in the American abstract expressionist artist's work have been applied to previously unrecorded works hailed as new finds last year. And the verdict is the 32 paintings which so excited the art world when they were unveiled last May could be fake.

In the wake of the analysis, the Pollock-Krasner Foundation, founded by Pollock's widow, Lee Krasner, issued a statement yesterday which said it was withholding final opinion as to the attribution until further research was completed and scholars reached a consensus.

It is commissioning further tests on half a dozen of the works whose existence was made known last year by Alex Matter, son of the photographer Herbert Matter and abstract painter Mercedes Matter, who were Pollock's close friends on Long Island, New York.

Mr Matter found the works on Long Island in 2002 amid his parents' belongings after his mother's death. He described them as "painted by Pollock in his signature dripped and poured style".

But Professor Taylor, of the Department of Physics at the University of Oregon, who has spent the last decade analysing the patterns in Pollock's work and compiled a database of analysis, disputes this description.

In a new article in the scientific journal Nature, the professor said fractal analysis - a mathematical way of analysing pattern formation - of six of the paintings showed they differed considerably from the patterns in works known to be by the artist.

"These differences indicate that Pollock's specific fractal signature has not been found in the submitted paintings. The analysis has also revealed that the patterns vary between the paintings, indicating they may have been painted by different hands," he said.

Pollock's known works show he was in control. "The large-scale fractals are a fingerprint of the artist's body motion [as he walked around a canvas]. But the small-scale fractals are also to do with his choices - his height over the canvas, the fluidity of his paint, angle and force behind the trajectory, and so on."

Francis O'Connor, co-editor of the Jackson Pollock Catalogue Raisonné, the definitive five-volume inventory of the artist's work, said: "Professor Taylor's fractal test results reinforce my own scepticism and reservations concerning the paintings.

"The historical documentation to date provides no conclusive proof that the new works can be attributed to Pollock. Further, a careful stylistic inspection of Pollock's poured works from the period in which the paintings are supposed to have been painted - roughly 1943 to 1950 - reveals no relation to Pollock's known stylistic development."

The possibility that the pieces were genuine gained currency because Pollock was an alcoholic and bartered some paintings for groceries before his death in a car crash in 1956 at the age of 44.

There were also notes in the handwriting of Herbert Matter identifying them as paintings carried out by Pollock in the 1940s which he had acquired as "gift + purchase".

But it is also known that Pollock's innovative work inspired many to try to emulate him. It is also now exceptionally valuable.

Given the large number of works in contention, the Pollock-Krasner Foundation decided it should be involved, although it disbanded its board for authenticating his works a decade ago.

Mark Borghi, the art dealer to whom Alex Matter showed the works, showed them in turn to Ellen Landau, a Pollock expert who believes they are genuine. The paintings are due to be exhibited this year. Nature magazine reports that Mr Borghi does not believe authentication should rely on fractal analysis alone because painters often produce things out of style.

The authentication of works of art normally relies heavily on expert visual assessment, supported by the analysis of materials and knowledge of their history. But tests on materials do little to help identify Pollocks as he often used off-the-shelf paints.

Professor Taylor's investigations, which involved watching films of Pollock at work, original paintings and dozens of imitations, did, however, suggest that despite varying working methods - such as flicking paint or letting it fall or run down a brush - genuine Pollocks all shared "a fractal composition that was systematic".