Inpainting a George Romney Portrait
By Katie Rovito, Third-Year Graduate Intern in Paintings Conservation
Walking through the British art galleries, you might notice a cracking pattern or even an alligator skin–like texture on the surface of some of the paintings. These cracks formed when layers of paint dried unevenly. British painters of the late 1700s were notoriously experimental, trying out different materials to mimic the translucent glazes seen on old master paintings. Unfortunately, additives like natural resins, megilp, or bitumen caused severe cracking even within the artists’ lifetimes and continued to worsen with time. This was the case for a portrait by well-known 18th-century British artist George Romney (1734–1802) that I have been treating during my third-year graduate internship in paintings conservation at the Cleveland Museum of Art (CMA).
It is likely that Romney’s use of these types of materials contributed to the compromised condition of Portrait of Jane Hoskyns. Some 18th-century manuscripts recommended bitumen as a glazing color for its transparency and warm tone. Bitumen, also known as asphaltum, partially dissolves in oil and significantly slows the drying time. Recipes with bitumen often included various resins, waxes, and driers to counteract the poor drying and aging properties. Another possible culprit is megilp. Popularized by one of Romney’s contemporaries, Sir Joshua Reynolds (1723–1792), megilp is a thixotropic gel that forms when mixing linseed oil, mastic varnish, and a lead drier. The gel medium creates a transparent paint that can still hold impasto, or thick brushstrokes.
In addition to the drying cracks, these paint modifiers greatly impact the solubility of the paint film. Artists often coated their finished paintings with a natural resin varnish, such as mastic. It is common practice for conservators to clean and replace aged and yellowed coatings. When resins are added to oil paint, the solubility can become close to that of the varnish, making it nearly impossible to remove the varnish without also removing some of the original paint or glazes. This was especially true during earlier restoration campaigns when conservators used more aggressive solvent mixtures than we have available today.
There were signs that Romney’s Portrait of Jane Hoskyns had been overcleaned in a restoration prior to its acquisition by the CMA. Abrasion throughout the painting was apparent where original glazes had been removed from the peaks of the canvas weave. Transparent, thinly applied original tones that made up the shadows in the sitter’s dress ended abruptly where overcleaned. Similarly, a distinct line was visible along the lower right edge of the oval where the darker original glazes ended. This edge was likely once covered by a frame. When paint and varnish layers are hidden from light, such as under a frame, it slows polymerization, making the paint film more soluble and vulnerable to cleaning solutions.
When I first examined the painting, the varnish was only slightly discolored and the matte surface lacked proper saturation. Although some prior retouching addressed areas of overcleaning, the abrasion was still extensive. This was especially noticeable around the face, disrupting the soft gradations characteristic of Romney’s style. After performing solubility tests under a microscope, I removed the synthetic coating with a solvent established as safe.
The majority of my time on this treatment, though, has been spent in the inpainting phase. Gamblin Conservation Colors, which are stable and reversible, were ideal for reconstructing the abraded glazes. I met frequently with my supervisor, Dean Yoder (senior conservator of paintings), and Cory Korkow (curator of European paintings and sculpture, 1500–1800) throughout the treatment to consult on how far to take the inpainting. My strategy with the scattered areas of abrasion was to assess the painting in a holistic manner, slowly pulling together overcleaned passages. I worked under magnification using a tiny brush to carefully compensate for missing dots of paint.
There was more to consider when it came time to inpaint the stripped glazes in the dress and the sharp cleaning edge near the shoulder. Here, it was clear that a previous restoration had removed soft gradations of shadow intended by Romney. Fortunately, there were traces of the original glazes embedded in the impasto. Before inpainting the dress, I studied these glaze remnants and looked at drapery folds in other Romney paintings. Then, I first tested different reconstruction strategies with watercolor so they could be easily adjusted. The goal was to soften the abrupt cleaning edge by bringing the shadow up toward the top of the shoulder. Once Dean, Cory, and I were satisfied, I took off the watercolor and completed the inpainting with Gamblin Conservation Paint.
While Portrait of Jane Hoskyns likely had more glazes and darker shadows that we are not able to recover, I hope that my inpainting makes it easier to admire Romney’s work now that the viewer is less distracted by abrasion and missing glazes.
