AIC's 41st Annual Meeting

Art historical debate has percolated for more than half a century over the identity of the four Bacchanals (1635-6) commissioned from Nicolas Poussin by Cardinal Richelieu for his chateau in Poitou, France. Three of the four subjects have been identified as the Triumphs of Pan, Bacchus, and Silenus, respectively. The success of this commission led to demand for the production of copies close in date to the originals, some of which are represented in major collections. There has been general consensus that Triumph of Pan(National Gallery of London, NG6477) was an original part of that commission. However, the status of Triumph of Bacchus (Nelson-Atkins Museum of Art, 31-94) as the original commissioned version has at various times been doubted, in spite of sharing a common provenance with Pan until 1850. An incomplete early provenance and apparent stylistic discrePancies, particularly in relation to Pan, have caused serious doubts about the inclusion of Silenus (National Gallery of London, NG42) among Richelieu’s original commission. It is today classified by the National Gallery as a painting “after Poussin”.

In preparation for the Nelson-Atkins’ catalog of its French painting collection we sought to employ recent innovations in computerized “thread counting” for automated analysis of canvas weave variations to resolve the status of the Nelson-Atkins Bacchus. This approach has proven particularly powerful in demonstrating commonalities and distinctions between canvases of nominally identical average thread counts with art historical implications, as among the paintings of Vincent van Gogh ("The Burlington Magazine", February 2012), court portraits by Velazquez ("The Burlington Magazine", September 2012), and paintings by Johannes Vermeer ("Metropolitan Museum Journal", 2012). utomated weave comparison held the potential to settle the questions surrounding Bacchus in the event that the artist had employed part of the same bolt of canvas for Pan. Existing National Gallery radiographs of Pan were digitized and made available for our comparison with Bacchus through this process, resulting in a very close match of their warp thread-spacing variations (imposed primarily by the loom and resembling a bar code when color-coded and mapped). When subsequently performed on radiographs provided for the National Gallery Triumph of Silenus its warp spacing variations were also matched to a very high degree with those of Pan, providing compelling evidence that all three paintings were executed on sections of the same bolt of canvas. This form of evidence, connecting the three works to a single canvas as it does, relates them more closely than other forms of analysis might, such as a demonstration of the shared use of a common set of pigments, and is unaffected by variations in condition of the paintings. This outcome should lead to a significant reassessment of Silenus and, more generally, of the criteria upon which scholarly opinion has been formed in attempting to resolve questions of authorship among roughly contemporary Old Master paintings. A comparison of painting materials and working methods employed for Bacchus and Pan is now proceeding, which will be underpinned by the certainty that they were contemporary products of the same studio.

*author for correspondence
Robert Erdmann1*, C. Richard Johnson2, Mary Schafer3, John Twilley 4
1 Assistant Professor, University of Arizona, Materials Science and Engineering and Program in Applied Mathematics
2 Geoffrey S. M. Hedrick Senior Professor of Engineering, Cornell University, School of Electrical and Computer Engineering
3 Associate Conservator of paintings, Nelson-Atkins Museum of Art
4 Andrew W. Mellon Science Advisor, Nelson-Atkins Museum of Art