# The Last Link Is the Strongest One > Documentation in a provenance chain rises from 63 per cent at the earliest links to 96 per cent at the most recent. Seven of 266 records here end on a link that is not documented, and only two of the seven are title disputes. > The Study - 2026-09-12 > Canonical HTML: https://iconotheca.com/study/the-last-link-is-the-strongest-one/ ## Key points - Across 266 published records this archive grades 1,868 ownership links. Splitting every chain into five equal parts by position, the share graded documented rises from 63 per cent in the earliest fifth to 96 per cent in the most recent. - The first link in a chain is documented in 187 of 266 records. The last link is documented in 259 of 266. Doubt in a provenance sits at the beginning almost by definition, and current title is the part the evidence supports best. - Only seven records end on a link that is not documented, and they do not have a common cause. Two are contested title, three are institutions that have not published their own acquisition, one is a private owner nobody has observed since 2017, and one is an object that has been missing since 1946. - The three institutional cases matter more than they look. The Rijksmuseum's own catalogue entry for Durer's Rhinoceros does not state when or from whom it acquired the sheet, which means the museum with the clearest title in the group has the least documented last link. ## Documentation in a chain of ownership improves as it approaches the present ## Where the doubt falls, measured across 1,868 links ## Seven objects here end on a link that is not documented ## The two that are actually contested ## The three where the institution has not published its own acquisition ## The one nobody has looked at, and the one nobody can find ## What this means for anyone reading a provenance for a decision ## Why this archive is free to read Read a provenance from the bottom up and you will usually find that the part you need is the part that holds. The doubt in these chains is old, structural and concentrated in the first two links, and current title is the best-evidenced thing in the record. When the last link does fail, find out which kind of failure it is before you treat it as a claim: two of the seven here are contested, three are museums that have not published their own paperwork, one is a collector nobody has seen since 2017, and one is a sword that walked out of a police station eighty years ago. ## Questions **Where is a provenance chain usually weakest?** At the beginning. Measured across 1,868 graded ownership links in this archive, the earliest fifth of a chain is 63 per cent documented and the most recent fifth is 96 per cent. The first link in a record is documented in 187 of 266 cases and the last link in 259 of 266. **Does a gap in a provenance mean the title is bad?** Usually not by itself. In this archive the gaps cluster at the old end of the chain, where documents were never created or have not survived, while present ownership is documented in 97 per cent of records. A weak link close to the present is the one that should stop a reader, because it is rare. **What is the current status of the Cassirer Pissarro case?** The Supreme Court acted on 10 March 2025, and on 30 April 2025 the Ninth Circuit granted the plaintiffs' motion for remand and sent the case back to the district court for proceedings consistent with that opinion. Every United States court to reach the merits had previously upheld the Thyssen-Bornemisza Collection Foundation, but no such judgment currently stands. **Why would a museum's own acquisition be undocumented?** Because the acquisition record and the public catalogue entry are different documents, and many online entries carry the object number, the date of the work and the credit line without the date or source of acquisition. The Rijksmuseum's entry for its impression of Durer's Rhinoceros is an example: the object number is given and the acquisition is not. **How should a record describe an object that has disappeared?** By recording the last documented event and stopping there. The Honjo Masamune was surrendered at a Tokyo police station on 18 January 1946 and has not been seen since, so the final link in its chain is that surrender and its owner field reads unknown. Naming a speculative present holder would be an invention. ## Sources - [1] United States Court of Appeals for the Ninth Circuit, Cassirer v. Thyssen-Bornemisza Collection Foundation, No. 19-55616, Order filed 30 April 2025, on remand from the Supreme Court. Read in the court's own published PDF: 'Plaintiffs-Appellants' Motion for Remand to the District Court (Dkt. 182) is GRANTED.' - [2] Rijksmuseum, Amsterdam, catalogue entry for Rhinocerus (Neushoorn), Albrecht Durer, object number RP-P-OB-1473. Consulted for the acquisition information the entry does and does not carry. - [3] Worrell, S., 'The Crosby Garrett Helmet', Papers from the Institute of Archaeology, University College London, on the 2010 find, the Treasure Act position and the sale. - [4] Musee Guimet, collections page for the Begram material, 'Two female figures under a portico (Begram treasure)', for the hoard and its travelling history. - [5] The Art Newspaper, 'National Museum of Afghanistan reopens - but staff are unpaid and Taliban control the building', 10 December 2021, for the museum's condition and the return of travelling material.