The record behind the sheet in your tube: the file we worked from, what we measured, what we chose and why, what we are allowed to sell, and the framing dimensions for this sheet.
The sheet you have
- Work
- Piet Mondrian (Dutch, 1872–1944), Lozenge Composition with Yellow, Black, Blue, Red, and Gray, 1921
- Object
- Oil on canvas, 60 × 60 cm, hung on its point — 84.85 cm from side tip to side tip
- Held by
- The Art Institute of Chicago, Gift of Edgar Kaufmann, Jr.
- Accession
- 1957.307 (AIC object 109819)
- File
- The museum's own IIIF endpoint, 6578 × 6578 px, taken directly from the institution and not from an aggregator
- Crop
- [193, 207, 6386, 6480] → 6193 × 6273 px, the tightest rectangle that holds the canvas
- Printed at
- 50 × 50 cm sheet: image 41.46 × 42.00 cm, 0.49× — 379 dpi available
70 × 70 cm sheet: image 52.43 × 53.11 cm, 0.62× — exactly 300 dpi, printed pixel for pixel
Both files built at a flat 300 dpi, the most matte fine art paper resolves. Nothing enlarged in software. - Made in
- The certified house nearest your address, to one specification: 200 gsm FSC-certified fine art paper, twelve pigment inks, same colour profile, same border geometry.
Why a diamond
It is a square canvas rotated forty-five degrees and hung on its corner, and the lines inside it are not rotated with it: they stay level with the floor. Mondrian painted only about seventeen of these. The format is his argument made visible. On an upright canvas you can still read the frame as containing the composition; on a diamond you cannot, because the bands leave the picture at four diagonals and obviously continue outside it. 1921 is also the year the language itself is finished — after two years in Paris clearing out the last regular grid, he arrives at unequal rectangles, bands of unequal weight, three primaries and nothing else.
What we measured
The four tips were found in the museum's file at native resolution, in 400-pixel windows rather than on a reduced copy: left at x 193, right at x 6386, top at y 207, bottom at y 6480. The horizontal diagonal is therefore 6193 px, and since a 60 cm square on its point measures 84.85 cm tip to tip, the scale is 72.99 pixels per centimetre. That is the anchor.
The vertical extent is deliberately not used as an anchor. It comes to 6273 px, which is 85.95 cm — 1.10 cm more than the diagonal. That extra centimetre is the depth of the stretcher, visible along the two lower edges because the canvas was photographed from very slightly above. For the same reason the horizontal diagonal sits 38 pixels above the middle of the frame and tilts by a quarter of a degree. None of it was straightened: our rule is that a picture is only rotated when a second copy from another collection proves the tilt belongs to the photograph, and here it plainly belongs to the camera angle rather than to the canvas.
The grey in the corners
A diamond cannot fill a rectangle. The museum photographed the canvas against a plain light grey, so any rectangular crop keeps grey in the four corners — here about half the area of the crop. We could only take it out by cutting the picture away from its background, and that is retouching. The same decision was made for Talbot's Wrack, where the crop keeps cream album paper in the corners.
So the grey stays. Look at the lower tip and you can see the soft shadow the canvas throws on it, and at the vertex itself a small dark loss where the two edges meet, with a short crack running back along one of them. All of that prints. You are looking at a photograph of a painting, not a diagram of one.
Why the large sheet stops at 0.62×
The file carries 6193 pixels across the diamond. At our floor of 300 dpi that is 52.43 cm of image and no more. Our usual border would have wanted 58.8 cm on a 70 cm sheet, which would mean inventing pixels, so instead the picture is set at exactly 300 dpi and the side margins run wider than anywhere else except the Blechen. On that sheet the crop is printed pixel for pixel; nothing is resampled at all.
What we are allowed to sell
The work — and this one is worth reading. Mondrian died in 1944, so in Switzerland, the EU and the UK the seventy-year term ran out on 1 January 2015. The United States runs on a different rule. The painting was, so far as anyone can show, never published in the copyright sense before 1978 — it was with Léonce Rosenberg on consignment in 1921–22, first exhibited in 1935 and first published in 1951, according to the Art Institute's own record — and an unpublished work of an author who died in 1944 falls under 17 U.S.C. 303(a): life plus seventy, with a floor at the end of 2002. It has therefore been free in the United States since 1 January 2003. That is also the Art Institute's own determination, and it is the determination we rely on.
Correction, 3 September 2026. The first version of this page said the work was "painted and shown in 1921" and that a restored American term of 95 years from publication expired on 1 January 2017. We could not support that, and we have withdrawn it. If the painting were held to have been published before 1978 — at the 1951 exhibition, say — the restored term would run much longer, so the catalogue does not stand on that branch at all. It stands on the never-published reading and on the museum's declaration, and we would rather write that down than leave a tidier sentence we cannot prove. The house rule that no Mondrian painted after 1930 enters this catalogue stays in force as a conservative filter, with the added condition that the holding museum must itself declare the work public domain.
The photograph. The Art Institute of Chicago publishes images of public-domain works under Creative Commons Zero, for any purpose including commercial use; a caption is requested, not required, and it is on the product page. One honest qualification: because the photograph shows the depth of the stretcher, it is strictly a photograph of a three-dimensional object, and that is the one case where Swiss law would give a photographer a right of his own. It does not bite here, because the museum has waived its layer — but our clearance at that level rests on the waiver rather than on the usual argument that the original is flat. We would rather say so than not.
What we do not claim. This is a giclée reproduction printed with pigment inks. It is not a painting and is not sold as one.
A note on colour
Measured by our own method this picture qualifies for the Blue collection — 80 % of its coloured area falls in the blue band. It is not in the Blue collection. The blue is Mondrian's white, which is a cool grey-blue and covers a third of the canvas; raise the saturation threshold slightly and the blue share collapses to 32 % while the yellow and red rise to two thirds. Somebody arriving from a search for blue pictures would be sent a white one. The measurement stands as measured and the decision is recorded against it.
Framing this sheet
- 50 × 50 cm sheet
- A 50 × 50 cm frame, no mat. Nearest standard 20 × 20 in. Margins 4.27 cm each side, 3.72 above, 4.28 below.
- 70 × 70 cm sheet
- A 70 × 70 cm frame. Nearest standard 28 × 28 in. Margins 8.79 cm each side, 7.86 above, 9.03 below.
- If you mat it
- Go one frame size up and cut the window three millimetres inside the printed image on every side: 40.86 × 41.40 cm for the small sheet, 51.83 × 52.51 cm for the large one. Off-white or warm white board, acid-free. Do not trim the sheet.
Please measure before you buy or cut. These figures describe the file we print from. Sheets are cut by machine to about a millimetre and frames vary between makers. We replace a print that arrives damaged or wrong; we cannot replace a frame or mat that turns out not to fit.
All records: museartis.com/pages/sheet-records · Questions: info@museartis.com