Many museum collections that have been built over a long time – collections such as that of the National Gallery of Victoria – contain works that have not been studied in any great depth and that have seldom been on display. The value of collecting can sometimes be measured by another generation’s ‘rediscovery’ of these works, of which the Descent from the Cross, c.1511–20, an early-sixteenth century Antwerp panel, is an example.1 The dating of the panel is based on the results of the dendrochronological analysis carried out by Dr Peter Klein, of the University of Hamburg, who also provided valuable comments about the construction and condition of the panel (see P. Klein, ‘A Note on Dendrochronological Analyses of Panel Paintings at the National Gallery of Victoria’, in this issue of the Art Bulletin, pp. 44–5). For further discussion of the Descent from the Cross, see U. Hoff, European Paintings before 1800 in the National Gallery of Victoria, 4th edn, Melbourne, 1995, p. 117.
The Felton Bequest acquired the Descent from the Cross from the estate of Miss Lucy Smith in 1931. Lucy Smith, who had died in 1904, had inherited the painting, along with the Brighton property known as Castlefield, from her father, John Matthew Smith (d. 1898). John Smith had been a prominent and successful businessman and solicitor, with interests in railways, hotels and real estate. He is believed to have brought the Descent from the Cross with him when he came to Melbourne from England in 1846.2 Details of the Smith family history have been provided by the Brighton Historical Society. I am also indebted to Paul de Serville for providing a detailed synopsis of the life of John Matthew Smith. Smith is recorded as having been born 13 June 1815/16 at Broomsgrove, Worcester. Though his descendants traced the family history in England, the relevant documentation has subsequently been destroyed.
Though on display when it was first acquired by the Gallery, the painting subsequently spent many years in storage. The condition of the work in 1994, when it was examined by the author, was probably much as it had been for the past hundred years or more. The central plank of the panel was cracked, the varnish was severely discoloured and the painting showed extensive areas of drying cracking associated with deteriorated passages of restoration (fig. 1). The work was framed in an ornate nineteenth-century gilded composition frame.3 For composition frames, see J. Simon, The Art of the Picture Frame: Artists, Patrons and the Framing of Portraits in Britain (exh. cat.), National Portrait Gallery, London, 1996, pp. 41–3.
An X-radiograph taken of the painting in 1994 (fig. 2) revealed the losses in the paint layer that had been the underlying motivation for the previous restorations; particularly notable were the losses associated with the realigning of the two vertical joins in the panel (see below). The areas of overpainting would, however, prove to be far more extensive than the losses. It also became apparent that any new restoration of the painting would involve some challenging reconstruction.
Prior to the treatment proper, sections of the painting were cleaned to make it possible to assess the state of preservation of the paint layer and the quality of the painting (fig. 3). The results of the test cleaning were dramatic and encouraging, and a proposal to clean the painting was approved. Though most of the heavily discoloured varnish could be removed with conventional solvent mixtures, the areas of overpainting were resistant to acceptable solvents and required removal with a scalpel, aided by magnification (fig. 4). The cleaning worked through several campaigns of restoration, some very old, with the most recent likely to have been undertaken in the early nineteenth century, prior to the work’s coming to Australia. As expected, the cleaned painting (fig. 5) presented the areas of loss that had been visible in the X-radiograph, but also revealed passages of beautifully preserved paint layer, most notably the faces of all the figures other than the central kneeling woman (figs 6 & 7).
As with many sixteenth-century panel paintings, the Descent from the Cross presented structural problems to deal with. The oak panel had at some stage been taken apart along its original joins and its three planks had been subsequently reglued out of plane. This displacement had caused much of the paint loss along the joins (the losses being the result of the planing, or sanding, of the misaligned edge in order to create a smooth transition). At some point the panel had also been thinned to 4 mm and cradled. The discontinuous cracking in the central plank was a consequence of the contractions of the wood being countered by the cradling: its structure enclosed and constrained its crosspieces and effectively prohibited the movement of the pane1.4 Closed cradles such as this have been found on a number of panels in the Gallery’s collection. Although in the 1920s and 1930s panels were cradled, on recommendation, after their acquisition and prior to journeying to Australia, the Descent from the Cross appears to have come to Australia some eighty-five years before entering the Melbourne collection. The cradling is believed to date to before 1846. The cradling was removed and the cracks glued. The process of removing the cradle, and the associated extensions to the corners of the panel, returned the painting to its original ogee shape at the top. No attempt was made to open the joins in the panel and to realign the planks. With the cradle removed, the panel presented a very thin structure, which was given support in the form of a lightweight, rigid panel tray.
With the structural work carried out, attention turned to the restoration of the painting. Old filling material was removed from the losses in the sky, and new fills, carved to match the texture of the surrounding paint, were put in place. Elsewhere, the surfaces of existing fills were adjusted. As it had not been possible, during cleaning, to completely remove the overpainting in the lower left corner (fig. 8) – without damage to the underlying paint – this area required reworking of the existing overpainting, as well as reconstruction in areas where losses in the paint layer had been revealed by the cleaning (fig. 9). The 34 re-establishment of recession within the picture plane was the aim of the restoration of the foreground area.
Particularly complex was the reconstruction of the face and hand of the central female figure. In the previous restorations this figure had been awkwardly overpainted (fig. 10). After cleaning, elements of the features that had been hidden by the overpainting were visible, and the original brushstrokes that had formed the eye and mouth were more clearly defined (fig. 11). On the basis of this information and of that provided by other faces and hands in the painting, a reconstruction evolved (fig. 12). The existence of a painting that closely resembles the Melbourne work, and that shows the face of the kneeling woman similarly tilted towards the observer, further sustained this reconstruction process.
The collection of the Museum Onze-Lieve-Vrouw ter Potterie in Brugge contains a complete Descent from the Cross triptych, a work that is dated 1520 on the reverse of its left wing.5 For a full description of the triptych and its restoration, see D. Fallon, ‘Une Intervention d’Adriaen Ysenbrant ou de son atelier dans un triptyque maniériste de 1520, conservé à Bruges’, in Bulletin de l’Institut Royal du Patrimoine Artistigue, vol. XIX, 1982–83, pp. 133–41. Though this discussion centres around an argument that Adriaen Isenbrandt was responsible for the reworking of the head of the fainting Virgin, Isenbrandt cannot be seen as the primary hand behind either the Brugge or the Melbourne painting. The central panel of the Brugge triptych (fig. 13) is a dose transcription of the Melbourne panel. The similarities are particularly evident in a number of technical observations, most notably details of an X-radiograph of the Brugge panel. Though the fainting Virgin in the Brugge picture is represented with a hooded mantle, the X-radiograph (fig. 14) reveals the figure to have originally been painted wearing the turban headwear of the figure of the Virgin in the Melbourne panel. More critically, the X-radiograph reveals that the horizon line and the central mountainous rocky outcrop in the Brugge panel were originally painted lower (fig. 15), conforming to the handling of these elements in the Melbourne painting. Similarly, though the background architectural elements occur on the right in the Brugge panel, the underdrawing reveals the intention to place these elements on the left-hand side, as in the Melbourne picture.
Despite these similarities, and although the two works are directly linked by subject and by the structure of the primary figure group, as well as by the intentions revealed beneath the surface, the Brugge and Melbourne panels are not by the same hand. The Melbourne painter is more precise and complex in his rendering of faces, fleshtones and general detailing. The fabrics are richer in texture, and the colour more strongly orchestrated throughout.
As with any major treatment of a painting, the work on the Descent from the Cross has revealed aspects of both the technique of the painting itself and the construction of the support. As already noted, the panel comprises three vertical oak planks (fig. 16). The planks are quartersawn, and dendrochronological dating reveals that the two outer planks have come from the same tree, which was around 270 years old when felled.6 See note 1 above. The oak is from the Baltic (Polish) region, and is therefore characteristic of Flemish panels of the fifteenth and sixteenth centuries. According to the dendrochronological analysis, the earliest date the painting is likely to have been executed is 1511. The three planks of the panel were originally joined with dowel pins between their abutting flat surfaces. The dowel holes would have been revealed when the panel was thinned for its cradle, probably in the early nineteenth century, and again became apparent when the cradle was removed (fig. 17).
The panel was found to have been prepared with a ground of chalk and glue. The underdrawing, which was executed with a brush, revealed that there were a number of changes between intention and execution. The architecture of the town in the left background has been simplified from its representation in the underdrawing (figs 18a and 18b), and a figure has been left out. There have been slight shifts in the alignment of Christ’s feet (fig. 19). The underdrawing is handled differently in different areas of the composition, with single hatching being used to represent shadowed areas; the rendering of the figures is tight, while the drawing is loose and less prescriptive in the background. A curious arc of broken lines follows the basic form of the face of the woman supporting Christ’s body (fig. 20).7 Reflectograms used to examine the underdrawing were made by Michael Varcoe-Cocks, with equipment funded through a joint Research Infrastructure and Equipment grant made available to a number of institutional conservation studios, and coordinated through the University of Melbourne.
What is notable about the palette in this painting is the remarkable juxtaposing of red and green in the shot fabric of the garment worn by the central kneeling woman. Also worthy of comment is the construction of the brocade on the cloak of Joseph of Arimathea, the man seen standing at the right of the composition. Despite appearing to follow the folds of the fabric, the brocade pattern (derived from the pomegranate) is in fact flat and may have been stencilled or traced across the folds. This shorthand method of creating the illusion of complex painting is sufficiently convincing to be overlooked at first glance (fig. 21).
The Brugge panel retains its original frame, and this was used as a model for reproduction in the framing, the final stage of the treatment of the Melbourne pane1.8 For a detailed description of the frame of the Brugge panel, see H. Verougstraete-Marcq & R. van Schoute, Cadres et supports dans la peinture flamande aux 15e et We siècles, Heure-Le-Romain, 1989, p. 211. The reproduction frame was made from oak; the joinery techniques used were those of the early sixteenth century.
The restoration of the Descent from the Cross took over two years of work but returns to the Gallery’s collection a painting that, because of its condition, had rarely been displayed. The painting revealed by the restoration process deserves our attention not simply because of its quality but because it presents a rich field for research and interpretation (fig. 22).
The Gallery has a small but representative group of early Flemish works in its collection. They reflect various themes and provide a balanced representation of a complex period. Within this group, perhaps the nearest neighbour to the Descent from the Cross is not a painting but the Carved Retable of the Passion of Christ, executed in Antwerp after 1511.9 Hoff, pp. 118–19, repr. p. 118. Both works might be described as stylistically transitional, and both were made in the same city – perhaps no more than ten years apart and most likely in studios that were within walking distance of one another. It is one of the great delights of the Melbourne collection that two such works have found their way to the other side of the world, more than four hundred years after they were made.
John Payne, Senior Conservator of Painting, National Gallery of Victoria (in 1999).
Acknowledgements
The proposal to investigate the treatment of the Descent from the Cross came from James Mollison, former Director of the National Gallery of Victoria, who brought the panel out of storage. The treatment could not have proceeded without the support of my colleagues in the painting conservation studio – Michael Varcoe-Cocks, Carl Villis and Linda Waters – whose discussions and encouragement were invaluable.
Notes
1 The dating of the panel is based on the results of the dendrochronological analysis carried out by Dr Peter Klein, of the University of Hamburg, who also provided valuable comments about the construction and condition of the panel (see P. Klein, ‘A Note on Dendrochronological Analyses of Panel Paintings at the National Gallery of Victoria’, in this issue of the Art Bulletin, pp. 44–5). For further discussion of the Descent from the Cross, see U. Hoff, European Paintings before 1800 in the National Gallery of Victoria, 4th edn, Melbourne, 1995, p. 117.
2 Details of the Smith family history have been provided by the Brighton Historical Society. I am also indebted to Paul de Serville for providing a detailed synopsis of the life of John Matthew Smith. Smith is recorded as having been born 13 June 1815/16 at Broomsgrove, Worcester. Though his descendants traced the family history in England, the relevant documentation has subsequently been destroyed.
3 For composition frames, see J. Simon, The Art of the Picture Frame: Artists, Patrons and the Framing of Portraits in Britain (exh. cat.), National Portrait Gallery, London, 1996, pp. 41–3.
4 Closed cradles such as this have been found on a number of panels in the Gallery’s collection. Although in the 1920s and 1930s panels were cradled, on recommendation, after their acquisition and prior to journeying to Australia, the Descent from the Cross appears to have come to Australia some eighty-five years before entering the Melbourne collection. The cradling is believed to date to before 1846.
5 For a full description of the triptych and its restoration, see D. Fallon, ‘Une Intervention d’Adriaen Ysenbrant ou de son atelier dans un triptyque maniériste de 1520, conservé à Bruges’, in Bulletin de l’Institut Royal du Patrimoine Artistigue, vol. XIX, 1982–83, pp. 133–41. Though this discussion centres around an argument that Adriaen Isenbrandt was responsible for the reworking of the head of the fainting Virgin, Isenbrandt cannot be seen as the primary hand behind either the Brugge or the Melbourne painting.
6 See note 1 above.
7 Reflectograms used to examine the underdrawing were made by Michael Varcoe-Cocks, with equipment funded through a joint Research Infrastructure and Equipment grant made available to a number of institutional conservation studios, and coordinated through the University of Melbourne.
8 For a detailed description of the frame of the Brugge panel, see H. Verougstraete-Marcq & R. van Schoute, Cadres et supports dans la peinture flamande aux 15e et We siècles, Heure-Le-Romain, 1989, p. 211.
9 Hoff, pp. 118–19, repr. p. 118.