The Scientific Examination of the Sanders Portrait of William Shakespeare
by Marie-Claude Corbeil, Senior Conservation Scientist, Analytical Research Laboratory, Canadian Conservation Institute, Department of Canadian Heritage
The publication of several articles about what could be an authentic head and shoulders portrait of William Shakespeare (“Is this the face of genius” by Stefanie Nolen, The Globe and Mail, Friday May 11, 2001; “It’s time to reveal Shakespeare to the world” by Stefanie Nolen, The Globe and Mail, Saturday May 12, 2001; “Careful analysis finds nothing to dispute authenticity” by Anne McIlroy, The Globe and Mail, Saturday May 12, 2001) has generated a lot of interest about the scientific examination of the portrait. This brief article will attempt to describe the examination procedures that were carried out at the Canadian Conservation Institute (CCI).
Let’s begin with a few words about the painting and some of the hypotheses that have been made about it. The portrait is presumed to have been painted by an unknown artist named John Sanders. It measures 42 cm high by 33 cm wide and is executed on panel, as are many old paintings (contrary to popular belief, canvas is not the only support used by painters; other types include wood, ivory, copper, and, more recently, particleboard). In the case of this portrait, two planks have been glued together to form a panel of the appropriate size. A date (“AN° 1603”) appears in red in the upper right corner. The location of this date (very close to the right edge of the panel with part of the last number missing) suggests that the right side of the panel may have been ground down or chipped off by accident. A paper label is glued on the back of the panel but it is badly damaged and the writing it bears is no longer legible.
The painting was the subject of an article in The Connoisseur in 1909. The author, a man named Spielmann, had transcribed the inscription on the label that identified the man in the portrait as William Shakespeare at the age of 39 years. But Spielmann did not accept that the painting was authentic; he concluded it was a “relatively modern” copy or fake. Among other things, he declared that the date had been added long after the portrait had been painted, that the costume had been extensively retouched or overpainted, and that the paper of the label was not that old.
When the portrait arrived at CCI, careful consideration was given to the procedures that would be used to date it. Examination of a painting by scientific methods should always be a cautious process, with only the necessary tests being performed. The first test suggested to the owner of the painting was tree-ring dating of the wood panel. By measuring the distance between growth rings and comparing this data to reference curves, it is possible to determine when a tree was cut, which will in turn provide an earliest possible date for the painting on that wood panel. For example, the portrait could not have been painted in 1603 if the wood on which it was executed was not cut until 1850. The tree-ring dating was done by an expert in the fieldPeter Klein from Hamburg University. His analysis showed that the wood was oak from the Baltic region, that the earliest possible date for the execution of the painting was 1597, and that a date of execution from 1603 onward was plausible.
However, this fact alone could not verify the age of the painting. A 19th-century forger could easily have painted the current portrait on top of an old painting. To check out this possibility, the painting was radiographed (a radiograph of a painting reveals what is beneath the surface paint layer in the same manner that a medical x-ray reveals the bones under skin). In this case the radiograph did not show the presence of another painting underneath the portrait we could see.
The painting was then examined with various photographic techniques using infrared and ultraviolet radiation. Ultraviolet fluorescence photography highlights any retouching done to a painting after it is finished. The results revealed nothing, thereby calling into question Spielmann’s allegation that the costume had been extensively retouched or modified.
The next step was to analyse the materials used to paint the portrait. Paintings generally consist of several superimposed layers, and the materials found in the layers and the way in which they are superimposed are characteristic of the era and geographical location in which the painting was done. However, because many painting materials and techniques have been used over several centuries and in broad geographical locations, this type of analysis cannot date a painting definitively. On the other hand, should the analysis reveal the presence of a painting material that was not introduced until the 19th or 20th century, it could prove that a painting is a copy or a fake.
Microscopic paint samples were taken with the tip of a scalpel blade and were analysed using several instrumental techniques. The results indicated that the nature of the materials in the portrait and the way in which they were used were consistent with those of paintings of the Northern School (which includes England) from the early 17th century. First a calcium carbonate and glue layer had been applied to the wooden support, followed by a second layer (an “imprimatura” layer) consisting of a mixture of lead white and calcium carbonate in oil. All of the pigments identified in the paint layers (cinnabar, or vermilion obtained by sublimation, red lake, charcoal black, orpiment, and lead white) were traditional pigments; no anachronisms were found. In addition, close examination of the date revealed nothing in the way the red paint was applied to indicate that the date had been applied at some point in time after the portrait was finished.
Finally, the paper label on the back of the painting was examined to determine whether or not it was contemporaneous with the painting. Analysis showed that it was rag paper made from linen fibres (rag paper was traditionally used in the 17th century; pulp paper is much more modern). Analysis also showed that the label was glued to the wood with starch paste. The last step was to date the paper. To obtain a sample that was large enough for radiocarbon dating, paper was cut along the sides of the label. This sample was submitted to Roelf Beukens at IsoTrace Laboratory, a laboratory affiliated with the University of Toronto that specializes in radiocarbon dating. He concluded that the paper could date anywhere between 1475 and 1640.
The owner of the painting considered having other materials (such as the ink or the glue from the label or the paint itself) dated, but this proved to be technically impossible. Radiocarbon dating is not a simple thing; often the sample needs to be pretreated to extract the carbon that will be analysed. The dating of complex mixtures such as ink and paint is further complicated by the risk of contamination which would give false dates. More importantly, even though modern techniques have greatly reduced the size of sample required for such an analysis, it is still necessary to remove several tens of milligrams of material to obtain enough carbon after pretreating the sample. The removal of such large samples would have greatly compromised the integrity of the painting as a work of art and testimony of the past.
However, the results of the tests that were done were conclusive: the painting was executed on wood that dated from the correct period; the materials and the way in which they were used were consistent with a painting done in England in 1603; no anachronistic material was found; and the label identifying the subject of the portrait was made of rag paper dating from 1640 at the latest. All these elements indicate that the painting is indeed an old painting and not a relatively modern copy or fake (as stated by Spielmann).
But is the Sanders painting a portrait of William Shakespeare? It is not within CCI’s area of expertise to answer this question. However, by establishing that the painting is from the appropriate time period, CCI has contributed to furthering the debate.