4 Responses »

Van Gogh shows his true colours…

Sometimes, my job is like an episode of CSI: I can solve mysteries about a work of art with the help of evidence I find in the lab. A few months ago, researchers from the Van Gogh Museum came to me and requested that I examine a small painting of a seated nude girl using X-ray fluorescence spectrometry (XRF). Van Gogh produced this painting in Paris in the spring of 1886, at the studio of his teacher Fernand Cormon. He painted it on a small canvas that he had previously used for a flower still life, as we know from examining an X-ray photograph of the piece. The photograph is difficult to interpret, but we can make out a bouquet of flowers in a tall vase. X-ray photographs are always black and white, however. I knew that the XRF technique would not only produce a sharper image, but also tell us the colours of the concealed flower still life.

Vincent van Gogh, Nude girl, seated, 1886, Van Gogh Museum, Amsterdam (Vincent van Gogh Stichting)

Vincent van Gogh, Nude girl, seated, 1886, Van Gogh Museum, Amsterdam (Vincent van Gogh Foundation)

X-ray photograph of the painting Nude girl, seated. The red lines were added.

X-ray photograph of the painting Nude girl, seated, showing the underlying flower still life. The red lines were added.

.

My research associate Teio thought it might be possible to date the work more accurately if we found out what kinds of flowers Van Gogh had painted. In the X-ray photo, the flowers appear to be bunched into racemes. They could be hyacinths, but lilacs are also a possibility. Lilacs bloom from late April to early June, while hyacinths flower from April to mid-May or even as early as January in a pot indoors. If these flowers are hyacinths, then Van Gogh could conceivably have made this painting in Antwerp, before moving to Paris. Yet that seems improbable, because the type of canvas that he used is associated with his Paris period.

.

For several weeks, the painting visited our laboratory at the Cultural Heritage Agency of the Netherlands. During that time, the XRF scanner was used to measure the elements present in the paint. We scanned not only the uppermost paint layer, point by point, but also the layers of the flower still life below. This involved irradiating the paint with X-rays. In response, each element produced characteristic radiation. This showed me what elements are present, allowing me to deduce what pigments must have been used in each paint layer.

As fascinating as this may sound, the actual scanning is quite boring! The scanner took a week to cover a few dozen square centimetres of the paint surface. At least twice a day, I would stop by to check whether it was stuck. I made a back-up every day, because if the computer crashed, I certainly didn’t want to lose the results of the scan.

The XRF scanner measuring the elements present in the paint.

The XRF scanner measuring the elements present in the paint.

Checking the XRF scanner

Checking the XRF scanner

.

The scan for the element mercury revealed that the red pigment vermilion was used for the flowers. With a microscope, we had previously found traces of violet on the paint surface. This colour probably consists of ultramarine blue (which cannot be detected by an XRF scan) mixed with red (specifically, vermillion).

The scan did not provide a definite answer as to the type of flower depicted, since both hyacinths and lilacs can be violet in colour. It did give us a much clearer image of the concealed painting, however, along with a wealth of new data that will help us to answer other research questions in future.

XRF scan of copper and arsenic in the underlying flower still life

XRF scan of copper and arsenic in the underlying flower still life

Based on the XRF scans I made this reconstruction of the colours of the flowers in the underlying still life

Reconstruction of the colours of the flowers in the underlying still life

.

But Teio suggested that if we took a closer look at the X-ray, we might be better able to date the concealed picture, and therefore the painting of the nude girl as well. The bunched flowers are clearly emerging from the top of a tall vase. This is a more natural place for lilacs than for hyacinths, which are normally attached to a bulb planted in a pot. If these are lilacs, then Van Gogh could have painted the flower still life in the second half of April and then painted the nude girl over it in mid or late May, fairly soon after the flower painting was reasonably dry and just before leaving the Atelier Cormon.

This is exactly the kind of thing that makes my job so exciting! My area of expertise is materials research and technical examination of art, but as a chemist I have a different way of thinking from Teio’s. Through discussion, we can refine our research questions and come up with ideas for further investigation.

.

You can find more information about this painting and other re-used Van Gogh canvases in the Van Gogh Museum, where the display Van Gogh’s studio practice: Canvases re-used will be on show until June 2012.

 

4 Responses to “Van Gogh shows his true colours…”

  1. Zofia Marcinkowska says:

    I am fascinated by the discovery. Please send me information. I wish you success.

  2. Douglas says:

    This is wonderful and very interesting. I wish you unlimited success in this research and am interested to see what it will uncover. Greetings from England!

  3. Rene Mendizabal says:

    It’s a fantastic research. I’m not sure if Van Gogh would be very happy to see again some paintings that he wanted to eliminate after all. All the best and keep your quest for the truth.

  4. The hidden flowers remind me of mimosa. Van Gogh might have liked their bright warm yellow colour.
    Fascinating studies.