Bought En Bloc, and What the Record Loses

Buying a collection whole buys the objects and leaves their histories behind

A collection sold as one lot transfers ownership of every object in it and, almost always, none of the evidence about where those objects came from. The invoices, receipts, dealer correspondence and annotated catalogues that document individual acquisitions belong to the collector, not to the collection, and they stay with the estate, get dispersed, or are simply thrown away.

What the buyer receives instead is a single transaction: one date, one price, one counterparty, covering hundreds of objects. For the institution this is efficient. For the record it is a compression from which most of the information never recovers, because each object's own history has been replaced by the collection's history.

The effect shows up years later, when someone tries to write a provenance for one piece. The chain runs back to the collector and stops. Everything before that has to be reconstructed from outside the object: from sale catalogues, from other collectors' correspondence, from the annotations strangers made in their own copies of the catalogue. It is archaeology performed on the art market, and it is slow.

Seven objects in this archive entered public collections this way, through three transactions. One is the clearest case in the literature, and the Museum that made it has published the evidence against itself.

In April 1904 the Metropolitan bought a collection it had not seen

The Metropolitan Museum of Art acquired its founding collection of arms and armour in a single purchase in April 1904, and it did so without examining what it was buying. Stuart W. Pyhrr, then the Museum's curator in charge of arms and armour, set out the transaction in the Metropolitan Museum Journal in 2012 and used the phrase in his opening sentence: the Museum made its first important acquisition of arms and armour by purchasing en bloc the collection of Charles Maurice Camille de Talleyrand-Perigord, duc de Dino.

The collection comprised almost five hundred pieces, assembled in France over the last two decades of the nineteenth century, and was then considered the finest private collection of arms and armour in Europe. The Museum committed just over $250,000, the largest sum it had paid to that date for a single acquisition. The New York press announced it with considerable fanfare on 15 May 1904.

What the press did not report, and what Pyhrr states plainly, is that the Museum had purchased the collection sight unseen, based solely on the enthusiastic recommendation of one of its trustees, and without reference to Dino's privately published catalogue. The trustee was Rutherfurd Stuyvesant. The catalogue in question had been written by Baron de Cosson, a respected English authority who had been Dino's paid adviser on the collection since about 1896, and published in 1901. It existed, it was authoritative, and nobody at the Museum consulted it before committing the money.

The purchase worked, on its own terms. It gave the Metropolitan a collection of international repute, prompted the appointment of Bashford Dean as guest curator in 1904, and led to the establishment of the Department of Arms and Armor on 28 October 1912. The objects were the right objects. It is the record that came off worst.

The accession numbers are the surviving trace of the transaction

Every object from the Dino purchase carries an accession number beginning 04.3, and that prefix is the most reliable provenance evidence any of them has. The Metropolitan's own collection data gives all of them the same credit line, Rogers Fund, 1904, and the same accession year.

Read the middle two columns together and the difficulty is visible in the Museum's own cataloguing. Two of these four are described by the Museum as composite or comprehensively reworked, and the dates it gives are ranges spanning three centuries. Those descriptions are the product of a century of subsequent research, not of anything that arrived in 1904.

The prefix is doing work no prefix should have to do. It is standing in for the provenance of each object, and it says only that the object was in one Frenchman's house before April 1904.

Fewer than half the catalogue entries carried a provenance

The single most damaging fact about the Dino collection's paperwork is that there was almost none. Pyhrr states it directly: the duc de Dino appears to have left no written account of his activity as a collector, nor were any dealers' invoices or receipts preserved with the collection.

The printed catalogue, which the Museum did not read before buying, turns out to be only marginally helpful in this regard. Of its three hundred entries, fewer than half cite a provenance. That is the ceiling on what the Museum could have learned even if it had done the reading, and it means that for more than half of the collection's catalogued objects, the trail begins with Dino and there is nothing behind him.

What has since been recovered came from outside the collection entirely. Pyhrr reconstructs the sources from the surviving correspondence of Dino's friends and fellow collectors, notably de Cosson himself, Costantino Ressman, William Henry Riggs and the Parisian amateur Jean-Jacques Reubell, together with annotations that Riggs made in his own copy of a sale catalogue. The de Cosson papers are in the Library of the Royal Armouries at Leeds. A hundred years of scholarship has been spent rebuilding, from other people's letters, the information that would have fitted in a box of receipts.

Some of what has been rebuilt is uncomfortable. Dino's most important armour, a composite harness incorporating pieces from the Cloud Bands garniture made in 1554 for the future Philip II of Spain, came from the Fountaine collection, and many of the Fountaine arms had been among elements of armour stolen from the Real Armeria in Madrid and brought to London for sale at Christie's in 1839 and 1840. That is a fact about title, and it surfaced only because someone went looking in the sale catalogues.

The restorer who built the composites later joined the museum

Daniel Tachaux completed and mounted armours for the nineteenth-century Paris trade and then spent a decade on the Metropolitan Museum's staff. His work appears in the Dino collection before the sale: when Costantino Ressman commissioned the Paris armourer Alexandre Lebon in October 1891 to complete the missing parts of an armour and fabricate a helm of appropriate type, Lebon employed Tachaux for the job.

The economics of that transaction are worth stating exactly, because they explain the whole practice. The restored and completed armour cost Ressman a total of 7,950 francs. His price to Dino was 30,000 francs. Completion was not conservation; it was value creation, and everybody in the chain understood that.

Tachaux's work impressed Bashford Dean enough that Dean hired him as the Metropolitan Museum's armourer in 1909. He stayed until 1919, when he left to work for the collector Clarence Mackay. The Museum's photographs of the Armor for the Joust of Peace credit him, and the record in this archive names him as the maker of the armour's most visible element.

There is nothing furtive in this. Tachaux was a highly skilled craftsman doing what the market and then the museum asked of him, and the Metropolitan has published the connection itself. The consequence for the record is nonetheless real: some of the same hands that assembled the objects were later employed to look after them, and the boundary between an object's manufacture, its restoration and its curation is not where a reader assumes it is.

The wider verdict is the Museum's own. Modern scholarship, Pyhrr writes, has revealed that some of the better-known and most frequently published Dino objects are composites or outright fakes. Dino himself was not immune to it either: his relationship with the dealer Bachereau ended in 1899 in a dispute over a fake weapon sold to him as genuine.

A bequest can force a museum to take the fakes with the masterpieces

The Waddesdon Bequest reached the British Museum in 1898 as a single collection of nearly 300 objects, and the museum received the whole of it. Baron Ferdinand Rothschild left it on his death, and it consists, in the British Museum's own description, of exceptionally important medieval and Renaissance pieces as well as a number of nineteenth-century fakes.

That combination is not an embarrassment the museum hides. It presents the two together as a picture of the late nineteenth-century art market, which is the honest and the more interesting reading: what a great collector bought tells you what was on offer and what was believable, and the fakes are evidence of the same market that produced the masterpieces.

It is also the clearest illustration of what arriving whole means. A curator selecting object by object declines the doubtful ones. A collection arriving under a will does not offer that choice, and the receiving institution acquires the collector's judgment along with the collector's things, including the occasions on which that judgment failed.

The Holy Thorn Reliquary is in that bequest, and it is among the finest surviving pieces of Parisian court goldsmiths' work of about 1400. At the time it was accessioned its origins were, in the museum's phrase, completely lost, and it was catalogued as Spanish, sixteenth century. The object that is now the centrepiece of its own gallery entered the collection under a wrong attribution, inside a group that also contained forgeries, because the whole group came at once.

When the lot is the collection, the object has no price

An en bloc transaction publishes one figure for many objects, and no arithmetic recovers the price of any single piece. The Metropolitan paid just over $250,000 for almost five hundred pieces, which yields an average and nothing else; the Negroli helmet and a plain munition breastplate did not cost the same, and the transaction does not say what either cost.

The Hope Diamond shows the same effect in a purchase of two objects or two hundred. Harry Winston bought the McLean jewels in 1949 while the estate was being settled, and the reported figures for that purchase cover the whole collection rather than the stone alone. There is consequently no purchase price for the most famous diamond in the world at the most recent moment it was sold, only a price for a group of which it was part.

Henry Walters took it furthest. On his death on 30 November 1931 he bequeathed to the City of Baltimore an extensive art collection begun by his father William T. Walters, together with two buildings and an endowment, and the museum opened to the public in 1934. Some 36,000 objects transferred in one instrument. Every one of them has the same acquisition event in its provenance, and none of them has a price.

None of this is a defect in the institutions. It is a structural property of the transaction, and the honest response is to record it rather than to paper it over with an estimate. Where a record in this archive carries no figure for an object bought inside a larger lot, that absence is deliberate and it is the accurate statement of what is known.

How to read a collection name in a provenance

Treat the collection as one link, not as a proof of everything behind it. Duc de Dino collection establishes that the object was in France in the 1880s or 1890s and nothing further, and for more than half the collection the catalogue itself offers nothing further either.

Ask whether the archive travelled with the objects. This is the question that separates a rich collection provenance from a poor one, and the answer is usually recoverable: the de Cosson papers at the Royal Armouries survive and have been worked, while Dino's own receipts do not exist. Where the archive is elsewhere, say where.

Expect the attributions to have moved. An en bloc purchase imports a collector's catalogue wholesale, and the corrections arrive over the following century one object at a time. A date range in a museum record that spans three hundred years is not vagueness; it is the visible scar of that process.

And do not read an accession prefix as provenance. 04.3 tells you which transaction brought an object into the Metropolitan. It tells you nothing whatever about the four hundred years before it, which is precisely the information the transaction destroyed.