The Witt Library, housed in a fine old Robert Adam building, is the largest collection of photographs of paintings, drawings, and engravings in the world-a million and a half black-and-white copies, all annotated, in thick green file boxes stacked ceiling-high in every available inch of space. In the basement is the Dutch section, and it was there I went first, where "J. Vermeer of Delft" is given a four-foot shelf along the wall of a long, narrow corridor. This may seem rather a lot for a painter with forty works at most, but the Witt, as its director once told me with admirable British nonchalance, "is uncritical as to attribution." What that means is that the files include many copies-a great many copies-of paintings of doubtful authenticity or outright fakery.
And that is what makes it so useful. "We provide," Dr. Rowlande had further explained, "not so much a catalogue raisonne of an artist's work as a quarry of information to be discriminatingly mined, so to speak."
At the Witt, and nowhere else in the world, it is possible to look at most of the dubious paintings that have been successfully put forward as Vermeers at one time or another, and to compare them with the entire small body of unquestioned Vermeers. It is also possible to mine the quarry for factual information-provenances, cuttings, sales records-that is impossible to get anywhere else.
I probably ought to explain that the three paintings from the cache required a different approach than the one I'd used on the others. That is, my initial premise for the pictures from Florence had been that the forgery, if there was a forgery, was an old copy of an already existing painting; that there were two identical Durer self-portraits for example, one fake and one real, and that the fake was masquerading as the original. That avenue had turned out to be a blind alley.
But for the cache there were other possibilities. They had been purchased by Bolzano's father between 1930 and 1939-when only the most primitive scientific techniques for assuring the authenticity of paintings were available, and they had been out of sight since 1944. If one of them was a fake, it was not going to be a copy of an actual Vermeer, Rubens, or Titian, but a centuries-old painting in the style of one of them. Possibly it hadn't been intended as a forgery (what would be the point of forging the unknown Vermeer in the seventeenth century?) but had been altered later on. In any case, it had gotten by the experts of the 1930s.
If one of them was a fake. And if I couldn't find that out in the Witt, I wasn't going to be able to find out anywhere.
It always takes me a while to get used to the filing system. The materials under an artist's name are not arranged chronologically, or artistically, or by "period," but in the way that's most helpful to the people who use the place. Most of those who come are on errands like mine: They have an old painting whose authorship they doubt, so they want to look at every picture they can find with a similar composition, to see what it might be-other than what it's purported to be.
So Vermeer's pictures, for example, are organized under headings such as "Single Figures, Male, Full Length, Turned to Left" and "Males, Less Than Full Length, Without Hands." I found Young Woman at the Clavichord under "Single Figures and Portraits, Women, Less Than Full Length, With Hands, Turned to Right."
There was only one version, and it was identified as Bolzano's, and it matched the one in the exhibition perfectly. I turned over the large gray card to which the photograph was attached, hoping to find a provenance, and I did:
Young Woman at the Clavichord was perhaps in the collection of Diego Duarte, Antwerp, 1682, or in an anonymous sale (Jacobus Abrahamsz, Dissius of Delft?), Amsterdam, 16 May 1696, or in an anonymous sale, Amsterdam, 11 July 1714 (Lot 12). It was apparently in the collection of Graf von Schonborn at Pommerfelden near Bamberg, allegedly by 1746, and passed with the greater part of his collection to the Lacroix family, Paris. After several anonymous sales, apparently purchased by Charles Sedelmeyer in 1892 and sold by him to Lawrie and Co. in London in February 1893. Lent by T. Humphrey Ward to the Royal Academy, 1896. Acquired by the Bolzano family, Florence, in 1933. Expropriated by German government in 1944, present whereabouts unknown.
Also attributed to G. ter Borch, q.v. Also attributed to J. van Cost, q.v. Also attributed to K. Dujardin, q.v. Also attributed to P. de Hooch, q.v.
I'm not sure what I'd expected to find, but I'd hoped for something more useful. But this is a pretty typical provenance for an Old Master, and as you can see, it raises more questions than it answers, with lengthy gaps, and "apparently," "allegedly," "anonymous," and "perhaps" sprinkled throughout. And then there was that mess of "Also attributed's," all of which except the de Hooch were news to me.
I looked in the files of the other three painters as instructed by the q.v.'s, but came away with no reason to think that any of them had anything to do with it. No surprise.
Then I went rapidly through all the Vermeer files, glancing at each picture. I was searching for parts of Young Woman at the Clavichord that might show up in other pictures. There is a common kind of fake, of a Vermeer, say, in which the forger borrows a pair of hands from The Music Lesson, a mouth from The Geographer, eyes from The Love Letter, and so on, and weaves them into a single picture that thus has many Vermeer touches, even if it lacks the unity of a Vermeer whole. The great artists, on the other hand, while they repeated themes or entire paintings, rarely cannibalized little pieces of their own work.
As expected, I found nothing to suggest that Young Woman at the Clavichord was anything but an original and thoughtfully integrated composition. And I was more than ever convinced-almost certain now-that whatever Peter meant by "Down your alley," he didn't mean that anyone but J. Vermeer of Delft had painted this one.
That left Titian's Venus and the Lute Player and Rubens's Rape of the Sabines, and although the Rubens files were only a few yards from Vermeer's, I was freezing down there in the stony cellar, so I climbed up two flights to where the Italian collection is, and where the temperature is kept almost livable. (Years ago I complained to Dr. Rowlande about the Witt's heating, and was given a lecture on the pitiful American dependence on central heating instead of sensible underwear.)
The bright oval room in which Titian's files are stored was once the house's dining room, and in its center is an elegant twenty-foot-long Adam table, probably the original one, now just a comfortably worn worktable. On it I spread the contents of a folder labeled Venus With Musicians. Bolzano's Venus was there, along with a lengthy provenance and an envelope full of cuttings. The provenance told me nothing, but one of the cuttings quickly solved the riddle I'd come with: Why was this relatively early Titian painted in a style not associated with the artist for another forty years?
The cutting was from a 1951 paper by a Yale professor:
The Firenze Venus has long been ascribed to the year 1583. Clever biographical extrapolations by Sabrioli, however, now suggest that the correct date may be 1538, with the earlier ascription being attributable to an accidental transposition of digits in the seventeenth century. The current observer, though no art historian, finds Sabrioli's ingenious deductions thoroughly convincing.
Well, this observer didn't. With no disrespect intended to Sabrioli's biographical extrapolations, they were wrong. And the fact that all the post-1951 cuttings used the 1538 date merely meant that they were wrong, too. On stylistic grounds, Venus and the Lute Player was 1583, not 1538, and that was that. Sabrioli made a mistake. Case closed.
But the Titian had other problems, the main one being that there were six different versions: the one from Bolzano's collection, the well-known one in the Fitzwilliam Museum, one in Dresden with a doubtful provenance, one in Berlin with a black organist substituting for the lute player, and two in the Prado, also with organists, but white instead of black, with one of the two lacking the customary Cupid smirking nearby.