Eight impressions, one plate

There are no photographs of the sheet this project is about. There are, however, photographs of other impressions of the same plate, released by their holding institutions under licences that permit reuse, and eight of them are on this site.

Three from the Rijksmuseum, catalogued as states one, two and four. One from the National Gallery of Art. Two from the Metropolitan Museum of Art. One from the Art Institute of Chicago. One from the Cleveland Museum of Art. All public domain or CC0, all credited, all linked back to the record they came from.

The British Museum's three impressions are not among them, and their absence is instructive: the British Museum licenses its images CC BY-NC-SA, which is not a public-domain licence, so this project links to them rather than copying them — even though they are the best-documented records of the lot, and the only ones that state what physically distinguishes the states.

What the comparison shows

Put the Rijksmuseum's state one next to its state four and the difference is not subtle. These are the same lines from the same copper, separated by roughly a hundred and seventy years and by the hand of a Paris publisher who re-cut the signature to make the plate saleable again.

Between them sits everything that happens to a piece of metal that keeps being useful.

What it cannot show

A screen is not a print room. Comparing digital images tells you about composition and about gross differences in state and wear. It does not reliably tell you about the two things that most distinguish impressions of the same state — the inking, and the paper.

Inking is a decision made at the press, and it survives digitisation poorly: a photograph normalised for a collection database has already had some of that information flattened out of it. Paper is worse. The tone, the surface, the weight, the way light sits in it — none of that comes through, and the watermark is invisible unless the institution has separately photographed the sheet in transmitted light. Almost none of them have.

Which is why every one of these images is captioned with what it is: a photograph of a different object, made by someone else, for their purposes, not ours.

The thing the exercise is actually for

Comparators are how you build a scale. Before you can say anything about wear in one impression, you need to know what early and late look like on this specific plate. Eight images from five institutions is the beginning of that scale, and it is more than this project could have assembled on its own.

It exists because five museums decided their photographs of public-domain works should be public too. That decision, made by other people for other reasons, is what makes a small privately-held research project possible at all.

Addendum, 12 August 2026. There are nine now, and the ninth is the one that mattered.

When this note was written the comparator set covered states one, two and four, and the gap at state three was described here as a gap in what museums had digitised. It was not. RISD Museum in Providence holds an impression catalogued as New Hollstein 170.iii — the same state as this sheet — and publishes it under CC0. It had been sitting in this project's own census since the first pass, with its state field blank, because the museum's record puts state, inscriptions and dimensions behind disclosure panels that render empty until they are opened, and the pass that read the page recorded the emptiness as the museum's answer. A search summary that gave the correct state was checked against that reading and rejected.

So the sentence in this note about what museums have not photographed was, for state three, a sentence about what this project had not clicked. The comparison it argued for — that you cannot judge wear in one impression without knowing what early and late look like on this specific plate — is unchanged, and is now possible for the state that actually matters here. The correction is logged as CH-002.