Oddly Enough / Episode 08

The Face Under the Grass

A Van Gogh meadow had a woman’s face underneath it. How could anyone see her without taking the grass away?

A faint X-ray clue led researchers to the chemistry of the paint, and to a portrait that stayed on the same canvas.

Follow the hidden portrait ↓
LATER LANDSCAPEEARLIER PORTRAIT
Original illustration of two paint layers, separated for clarity. The generic face and grass are not copies of the painting or its recovered image.
01 / The picture on the wall

Look closely. It is still grass.

Van Gogh’s Patch of Grass is a close-up of a small patch of ground. Blades and flowers fill the canvas. Painted in Paris in 1887, it belongs to the Kröller-Müller Museum in the Netherlands.

Under that bright surface sits an earlier painting: the head of a woman, made during his time in Nuenen.

Van Gogh had used the canvas again. The grass covered the portrait, but the older paint remained underneath.

See the actual painting in the museum’s collection.

02 / A clue, not yet a face

The X-ray left a question.

Earlier X-ray examinations had already shown a faint head beneath the landscape. By the time of the 2008 investigation, the hidden composition was a known clue.

The harder question was what that face looked like. Different layers of paint contributed to the old X-ray image, leaving the features difficult to read.

A team including researchers from Delft and Antwerp, museum specialists and synchrotron scientists tried a more selective approach. They would ask where particular elements were present in the paint.

03 / The paint answers back

A red cheek leaves a chemical clue.

At DESY in Hamburg, an intense, narrow X-ray beam scanned a region of the painting. A detector collected the characteristic X-ray signals emitted by elements in the paint.

Working across the area built a map: where was mercury concentrated? Where was antimony?

Those elements offered clues to the pigments. Mercury helped trace reddish parts of the hidden face; antimony helped reveal its lighter tones. Reading the distributions together made features visible that the earlier image had left indistinct.

X-RAY BEAMDETECT SIGNALSMERCURYANTIMONYREAD TOGETHER
A schematic of the method. The coloured symbols are invented explanatory marks, not the team’s maps or a reconstruction of the real woman.
04 / The image beneath the image

The grass stayed on the canvas.

The team produced an approximate colour reconstruction of the woman’s head. It came from interpreting the paint’s elemental signals, rather than photographing a newly exposed surface.

The landscape had not been stripped away. The older composition could be studied while it remained underneath.

The museum points to the contrast between the darker Nuenen portrait and Van Gogh’s brighter Paris work. Two stages of an artist’s development occupy the same canvas.

You can still look at the grass. Now you know there is someone underneath.

A little further into the story

How can an element tell you a colour?

An element is a clue to a pigment, not a colour label by itself. The study associated mercury with a red pigment and used additional spectroscopy to identify antimony-bearing Naples yellow in lighter parts of the face.

That chemical interpretation is why the result is described as an approximate reconstruction. It does not recover every original colour with certainty.

Was the whole painting scanned?

The reported investigation scanned a 17.5 × 17.5 cm region covering the head, over two days. A map was assembled from measurements across that area.

The scan duration describes data collection. Interpreting the signals also required chemical analysis and art-historical comparisons.

Can this cleanly separate every hidden painting?

Elemental maps can combine signals from paint at the surface and below it. If different compositions use similar materials, their contributions can be difficult to distinguish.

The layers therefore need interpretation alongside other evidence. This successful case does not make an X-ray scan a universal undo button for paintings.

Follow the evidence

Original explanatory illustrations. The drawings are schematics, not experimental images or data.