This blog post is part of our behind-the-scenes series focusing on Lady Dalrymple (c.1851–1853) by George Frederic Watts. Painting Conservator, Alexandra Lawson, talks us through the analysis of the painting.

After Lady Dalrymple was requested for the exhibition Whistler’s Woman in White (organised by the Royal Academy of Arts, London and the National Gallery of Art, Washington). This provided a brilliant opportunity to get the painting out of its frame and into the conservation studio, and use different technical analysis methods to further understand Watt’s painting techniques, and evolution of this painting.

One such technique is Infrared Reflectography, which captures an image of the infrared wavelength normally too long for the naked eye to see. Infrared can penetrate the paint layers to reveal the artist’s initial sketch, as it absorbs carbon-based black particles. In this detail of Sophia’s face we can see faint pencil lines articulating the nose and hollow of the eye, not visible in normal light. (Figure 1)

The earring and some areas of lace also appear dark even though they are red in normal light. This photo (Figure 2), taken through microscope shows the edge of the right earring. You can see a thin black paint layer corresponds with the darkest areas in the infrared image. This also tells us that this black layer also extends underneath the red in some places.

For the red headdress, Watts has used a paint mixture of pigments that include vermillion and iron oxide red, natural ultramarine blue, and a carbon black pigment. Natural ultramarine was an unusual choice at the time, as the much cheaper synthetic French ultramarine was widely used. These pigments were identified by taking a tiny sample using a scalpel, and using an optical microscopy technique to examine each particles properties in normal and polarised light.

Another way of looking at the paint composition is to take a cross-section of paint that captures the whole paint layer structure, like cutting a slice of cake (Figure 3). This sample from the same location as the dispersion above shows that the red layer was painted over the blue of the sky. Other samples from the image show pentimenti, for example the lower left tile was changed from red, to white to red again.

Watts often changed his mind as the painting process progressed, and the portrait of Sophia is a particularly good example. The painting has an intriguing label on the frame, which records that the picture is one half of a portrait of two sisters. The canvas has been cut down the right hand side of figure, following the form of the figure’s hands and folds of the dress. The image below (Figure 4) shows five different sections of canvas that have been attached with a glue paste to a lining canvas.

Using an x-radiograph (Figure 5) enables us to understand this altered area further. We found that the thread count (the weight of the linen canvas) was identical in both the canvas with the portrait, and the additional canvas sections. The preparatory layers are also very similar, consisting of a thin white layer that has been sanded to reveal the tops of the canvas weave. In x-ray the additions appear very dark, as in general the off-white priming layer has been left unpainted, leaving the off-white to form a column. The x-ray also revealed a white compositional element hidden under the terracotta tile. Could this be from the other sister’s dress?

While we haven’t yet discovered the other half of this painting, looking closely at the portrait has given us more information about Watts’ painting technique, and shown the many small and large changes that happened during the painting process.

This research was made possible by support of York Consortium for Conservation & Craftsmanship, who supported Alexandra Lawson carrying out the project.

Thanks also to Dr. Tracey Chaplin who provided training in polarised light microscopy, and De Lazlo Paintings Conservation, Sally Marriott, for supervising the project.