Education · For lecturers and supervisors
Dossier v0.9Updated 2026-08-12Text CC BY 4.0Metadata CC0 1.0
Five things, all of them already public and none of them behind a request form. Every figure below is read from the data at build time, so this page cannot overstate what is actually there.
40 minutes of reading · general to graduate
A sequence, not a set: what an original print is, how states are defined and why catalogues disagree about them, how paper is used to date a sheet, what happens to a plate after the artist dies, how ownership history is read off the verso, and why a conservator argues against removing a stain. Each module carries its own seminar questions, which are unanswered here on purpose.
27 records · 20 institutions · CSV and JSON
Every row is transcribed from the holding institution’s own published record and carries the source URL against it, the evidence tier it was drawn from, and a review state. Institutions disagree with each other about state division and dimensions; those disagreements are preserved rather than normalised away, which makes the file useful for a data-methods class as well as a print-room one. It is downloadable as CSV and as JSON under CC0 1.0 — set it as a problem sheet without asking us.
4 states · 5 catalogue divisions compared
New Hollstein (Hinterding & Rutgers, Rembrandt) divides this plate one way; older catalogues divide it differently or not at all. The atlas sets the divisions side by side with the source for each, marks which have been verified against the catalogue itself and which are cited at second hand, and leaves the conflict standing. It is the clearest teaching example on the site of why a bare roman numeral in a collections database is a data-modelling failure.
12 questions · 5 marked suitable for a student
Each question states why it matters, what evidence exists now, what would resolve it, and what has already been tried and failed. That last part matters for supervision: a student can start from a documented dead end instead of rediscovering it. Accepted findings are credited by name on the record and in the changelog, permanently.
28 terms · 5 with an external source
Terms of art defined once, in plain language, each with a stable anchor so a reading list or a slide can link straight to the definition. Where a term has an authoritative external definition — the Getty AAT, the Canadian Conservation Institute — the entry cites it rather than paraphrasing it silently.
Cite the version, not the site. The record is versioned and the changelog is public, so a citation that names a version can be checked years later even after the underlying claim has been corrected — which several of them will be.
Suggested form
The Rembrandt Open Impression Project. Teaching pack, NP-RMB-001, version 0.9 (2026-08-12). https://rembrandt.aidispatch.xyz/education/teaching-pack — accessed [date of access].
Machine-readable forms — full dataset, census CSV, bibliography BibTeX, IIIF manifest — are listed on publications and data, with 13 bibliography entries exportable directly.
These are the open questions marked as suitable for a student: contained, resolvable with methods a taught postgraduate already has, and genuinely unanswered. Difficulty is our estimate of the work involved, not of the intelligence required. A resolved question is logged in the changelog with the contributor named.
| Ref | Question | Difficulty | Status | Why it is worth doing |
|---|---|---|---|---|
| RQ-01 | Does this impression carry the signature “Rembrandt f. 1638”, and if not, why not? | low | open | The owner's working record describes the sheet as “state III of IV, without signature.” The Rijksmuseum's record for its impression of this composition states the inscription “Rembrandt f. |
| RQ-02 | Which of the four New Hollstein states is this impression, and what distinguishes state III from state IV? | medium | open | State is the single field that most affects how a print is understood: it locates the sheet in the life of the plate. |
| RQ-05 | What is the chain-line spacing, and does it match any impression already recorded in the census? | medium | open | Chain lines are the widely spaced lines left in laid paper by the mould's supporting wires. |
| RQ-07 | Does the verso carry collector marks, dealer codes or inscriptions? | low | open | The back of a print is where its ownership history is usually written: collectors' stamps, dealers' stock numbers, pencil annotations, mount and hinge remnants. |
| RQ-12 | Which institutional impressions are the right comparators for assessing wear in this one? | medium | under investigation | Every wear judgement is a comparison, and the comparison is only as good as the comparators. |
If a student takes one on, tell us before they start — contribute — so two people do not do the same work in parallel, and so we can hand over what has already been tried.
The sheet can be examined in person, supervised, by researchers, conservators, curators and taught postgraduates with a supervisor. It is held in a private collection in Dubai, and it does not travel for a viewing — a session happens where the object is, or it happens as part of a loan to your institution.
What you want to look at, why this impression rather than one in a nearby print room, and what you would need in the room — magnification, transmitted light, a conservator present. A question we cannot answer from the record is the strongest possible case.
Handling is supervised throughout, the sheet stays in its mount unless a conservator removes it, and light exposure is logged against the object’s cumulative exposure ledger. Photography is by agreement in advance.
Anything you observe that changes a claim is published in the changelog with your name against it, and you keep the right to publish it yourself first. The project does not embargo a researcher’s own findings.
A seminar of twelve is not a study session. Institutional loan terms — including the lead time, the environmental conditions and the fact that there is no lender fee for scholarly projects — are set out in full on the loans page.
Stated plainly so you can plan around it rather than discover it the night before a seminar.
The sheet has not been photographed. Imaging follows conservator review, not the other way round, and no stand-in picture is published in the meantime. Every image on the site is of a different impression in a public collection, and each one says so.
The catalogue is in copyright and this project does not reproduce it, paraphrase it at length, or pretend to have consulted a copy it has not seen. Where a state division is cited at second hand, the record says so. Your library’s copy remains necessary.
Nothing here teaches what a sheet feels like, how a plate mark catches raking light, or how burr reads in the hand. Take the students to a print room. The census exists partly to help you find the nearest impression to do that with.
The project makes no attribution of its own and issues no opinion on authenticity. It records what published sources say and how good those sources are. Treat a confidence chip as a description of evidence, never as a verdict.
The record publishes no acquisition figure and no valuation, and a build gate blocks them mechanically rather than relying on editorial restraint. If your teaching needs that dimension, this is the wrong object.
When this sheet was printed, which state it is, who owned it before 2023 — all open. The pack is honest about that, but if you need a worked example with a known answer at the end, use a museum object with a full catalogue entry.
If something here would be more useful in a different shape — a different export format, a worked exercise, a version of the census filtered to your region’s collections — ask. Requests from people teaching this are the single best guide to what gets built next.