In this series, 'Conservation in focus', Simon Gillespie, Director of Simon Gillespie Studio, explores the art of conservation. In each article, Simon looks in depth at a particular technique or tool that is crucial to conservators and explores the challenges – and triumphs – associated with their everyday work.
It is important to keep in mind that a conservator can never return an artwork to the exact same condition as was when it left the artist's studio. Conservation treatment, then, involves a balance of undoing the ravages of time while accepting that the artwork is a physical object subject to the laws of nature. The goal of a conservator is to stabilise the artwork so that it can be preserved and to integrate any repairs in such a way that is sympathetic to the original.
When describing a painting, conservators will tend to work from the support forward: is it canvas, wood or something else? Then they focus on the ground layer, then the paint layer, then any varnishes, later overpaint... although the order of treatment will depend on what is most urgent to stabilise the painting. For these stories, I am going to work from the back to the front of the artwork, beginning with painting supports.
Most paintings from before the sixteenth century are tempera on wooden panel. The adoption of canvas as a painting support happened in different places at different times, and one notable example is Venice; when its inhabitants became increasingly wealthy and started commissioning artworks, they found that paintings on wood degraded at an alarming rate in their city, due to the humid and salty air. So, Venice led the way in the widespread practice of painting in oil paint on canvas, because those materials were much better suited to the saline, humid Venetian air, and because canvas was easily transported and available in plentiful quantity from Venice's sailcloth industry.
Even if a panel painting is not stored in a palazzo built into the seabed, it is still vulnerable to changes in humidity or structural weaknesses. Panels shrink and expand depending on humidity levels and can change shape as they age – glues used to join two or more wood planks to make bigger panels will fail, or the wood may split.
The panel painting Autumn is an obvious example: when it first arrived at our studio, from Birmingham Museum, it was in two separate pieces. The join between the two horizontal planks of wood had failed entirely and the picture had not been deemed worth repairing until Bendor spotted its hidden quality. The picture had been kept out of view in storage because it was in such poor condition, whereas now that the panel has been painstakingly repaired at the studio, the artwork can be enjoyed as a masterpiece.
The painting of Meleager and Atalanta by Jacob Jordaens from Swansea Museum is a good case study of the kinds of issues panels can exhibit.
The panel is made up of five horizontal butt-joined planks of wood, and the adhesive in all four joins had failed over the years. At some point, possibly during the nineteenth century, someone had made an attempt to rejoin the planks with 'butterfly joints' – a carpentry method where a piece of wood shaped like a bow tie is inlaid across the join between two planks, serving as a bridge or staple between them. Unfortunately, these butterfly joints had also failed. When the painting arrived at the studio, the planks were held together inside the frame by a complicated support frame on the back. A long-suffering picture!
To resolve the structural issues, the support frame was first dismantled from the panel, releasing the five planks of the painting. The panel joins were cleaned to remove the old glues and the planks were then securely rejoined with a suitable adhesive. Once rejoined, the natural warp inherent in each plank due to ageing resulted in the panel having a naturally forming curve, with the middle protruding by about an inch compared to the top and bottom.
The really exciting find was the discovery of panel marks on the reverse of the panel. Just as gold and silver jewellery have hallmarks to identify who made them and where, panel makers in the past would stamp their panels. The stamps found on this artwork were those of the city of Antwerp and of the panel maker Gabron, which together indicated a date of 1619–1621.
Another interesting find was that only the four top planks were joined with dowels. The texture of the bottom plank is notably different, suggesting a different application of ground layer and preparation. I believe this plank may have been added by the artist to enable him to enlarge the composition as he worked out where he wanted to place his figures.
All of this was hugely important for Bendor to make his case to the experts to have the painting recognised as a work in which Jordaens tests his ideas for the composition of his most famous work, Meleager and Atalanta, a large painting on canvas which hangs in the Museo del Prado in Spain. Bendor wrote about the painting for Art UK in 2016.
Much art conservation involves undoing or mitigating the damage caused by past restorers' interventions. (You'll see this will become a theme over the course of my next articles). This was exemplified in the Pollok House portrait of the Duke of Buckingham by Peter Paul Rubens, one of the world's most famous artists.
The portrait was made of two vertical planks joined together, with the join running down the middle of the sitter's left eye. Unfortunately, the join had failed in the past and, as part of an attempt at restoration, the edges of the panel had been shaved down, losing a strip 3–4mm wide. This may sound very small, but as it went down the middle of the face it cut down part of the eye of the sitter, which is a crucial part of any portrait. The planks had also been unevenly rejoined, creating a step at the join instead of a smooth plane. In an attempt to disguise the damage, overpaint had been applied along it, extending over the original paint. The result? The masterpiece had been made to look like a later copy of inferior quality. Our treatment therefore focused on undoing the deleterious effects of these interventions to return to the artist's intention.
In the past, to provide support to weakened panels, it was common practice among restorers to apply a 'cradle' to the reverse: a grid of wooden slats. The slats running in one direction will be fixed to the panel and immovable whereas the slats running perpendicular to the first will be able to slide slightly, to allow for natural movements in the panel. The problem with cradles is that often the slats that are supposed to be mobile end up becoming stuck over time if they are not regularly monitored, and the resulting pressure between the stuck slats and the panel trying to move will cause damage.
Another issue is that to apply the cradle in the first place, the panel is thinned down, a practice which irreversibly alters the original piece and therefore would now raise ethical questions. The panel Winter by Peter Breughel the elder had had a cradle attached to it, to stop it from warping. The cradle's mobility needed improvement.
Its pair, Spring, had been treated in the mid-twentieth century. Instead of being cradled, the panel was thinned down from the back and adhered to a new secondary support: a plank of balsa wood. Who could forget Bendor's cry of disbelief – 'Is that MDF?!' – as the picture was removed from its frame? The intervention appears to have been an attempt to strengthen the original oak panel, which had suffered two severe splits.
Nowadays, specialist conservators have developed improved methods for cradling. To make a panel tray to reinforce the Virgin and Child from Cardiff's National Museum of Wales, we had the pleasure of working with Britta New, a conservator at The National Gallery in London who is an expert in making supports for panel paintings. The new structure Britta designed gently supports the panel whilst allowing small movements in response to humidity changes.
One exciting thing about panels is the information that can be gathered by analysing the tree rings (called 'dendrochronology'). A tree's growth is visible in the rings of its trunk: each ring marks a complete cycle of the seasons. The thickness of the ring depends on rainfall, so trees from the same region tend to develop the same pattern of ring widths in a given period of time. This means we can use data about trees in Europe to analyse a panel's ring patterns and draw conclusions as to where and when the tree grew and when it was cut down. Based on information about the maturing time for planks before they were used for painting on (a process called 'seasoning'), we can now determine the earliest date an artwork could have been started.
It's an astonishing science. For example, with Rembrandt's Self-Portrait at Knightshayes, Peter Klein's dendrochronology told us that the tree was felled in 1621, which means it would have been ready for use in painting by the late 1620s, when Rembrandt was working in Leiden. Another fascinating discovery was that the panel of the Knightshayes portrait came from the very same tree as that of the Kassel version, which strongly supports the argument that the panels were painted in Rembrandt's workshop, whether by Rembrandt himself or by students under his supervision.
In my next article, I will look at another common support for artworks: canvas.
Simon Gillespie, Director at Simon Gillespie Studio
Previous: Paint analysis for art conservation