The eighty-five-year-old magistrate is shown without the stiff formality of many dogal portraits, almost as though he was ready to converse with the viewer.1 Ε. Pillsbury & W. Jordan, ‘Recent Painting Acquisitions – III: The Kimbell Art Museum’, Burlington Magazine, vol. CXXIX, no. 1016, November 1987, p. 773.
This description of Jacopo Tintoretto’s Portrait of Doge Pietro Loredan in the Kimbell Art Museum, Fort Worth, Texas (fig. 1), applies equally to the version of the painting in the National Gallery of Victoria (fig. 2). There has for some time been discussion about which of these paintings is the primary version.2 Discussing the Kimbell version, Pillsbury and Jordan note: ‘The painting has been accepted as Tintoretto’s primary version of the subject by Berenson, Tietze, Pallucchini, De Vecchi and all modern writers except VON DER BERCKEN who … maintained that the J. Ross Delafield picture [the Kimbell version] was a replica of the Melbourne version’ (p. 771 n. 11). Bercken’s position is reiterated by U. Hoff in European Painting and Sculpture before 1800, 3rd edn, National Gallery of Victoria, Melbourne, 1973, p. 154. An X-radiograph made of the Melbourne painting in 1989 (fig. 3) offers an opportunity to add technical information to the debate.
Examination of this radiograph reveals a number of pentimenti which are not evident in a radiograph of the Kimbell picture (fig. 4). These pentimenti point to important changes that have had considerable impact upon the way the Doge appears in the final image, suggesting that the image as it now exists in both versions was reconsidered and developed during the execution of the Melbourne work. The radiograph of the Melbourne painting reveals, for example, the original drawing of the neck and collar in an upright and stiff position, and the belt and waistline lower on the figure. The existence of these pentimenti reflects an intervention of a masterly nature in the development of the image, with the subsequent changes giving a greater sense of age to the Doge by relaxing his posture.
The following notes are based on a detailed reading of the Melbourne painting by means of an examination of both painted surface and radiograph. Also included is a discussion of the painting’s canvas support and its stretcher.
The stretcher on the Melbourne Doge cannot be considered original. Although the tacking or stretching edges of the original canvas have been removed, the canvas extends almost to the dimensions of the current stretcher. It can therefore be assumed that the original stretching system was larger than the current one.
The existence of yet another (smaller) stretcher is indicated by cracking in the paint and holes in the support, both of which are evident in the radiograph. Although the cracking and losses appear at the right, the left, the top and, to some extent, along the bottom edge, they are not, however, indicative of the original fold line and stretching system, since the image extends beyond the cracking and there is little evidence of distortion in the weave of the canvas.
The current stretcher measures 109.5 x 93.5 cm. The smaller stretcher would have measured around 105.5 x 88.5 cm. The dimensions of the original stretcher remain a matter of speculation.
The original support is a relatively fine, closely woven canvas of plain weave, with a thread count of around 12 x 15 per cm. The canvas finishes with an irregularly cut edge and an exposed margin of overpainted lining canvas on all sides. Though the painting has been lined, no major damage to the support is apparent. It is reasonable, however, to suggest that the original canvas has been cropped on all sides: when the dimensions of the Melbourne and Kimbell paintings are compared we find that the former picture is smaller in both height and width. At the same time, it is important to note that the image in the Kimbell portrait corresponds very accurately in scale to that in the Melbourne painting.
The Kimbell painting has along its bottom edge an original addition to the canvas, attached at the point at which the Melbourne image ends. There is no reason to believe that the Melbourne painting has not lost the section of image corresponding to this addition, and that the two works may not once have been of similar dimensions. It is also possible that the Melbourne canvas might initially have been only temporarily stretched for use as a working canvas.
The paint layer
The pentimenti in the paint layer, as revealed in the radiograph, are numerous, but are nevertheless important to highlight. The profile of the hat has been altered by flattening the crown and by changing what we see as the hat’s rear edge to a smooth curve. The section of the white cap that appears over the Doge’s left ear has been altered in position, along with the cord that dangles from it. The decorative balls on the left edge of the ermine cape have also been through a number of changes. It is apparent that initially only two large circular forms were drawn, corresponding to the second and third balls from the top. The last two balls in the second (current) set have also been overpainted, with the fur of the cuff and the thumb of the Doge’s right hand. Changes are also apparent in the drawing of the folds of the fabric of the sleeve above this hand. The lower edge of the ermine and the section of exposed left hand are also of interest. It would appear from the radiograph that the fur of the cape originally extended further towards the base of the picture and that the fur on the cuff consequently left a smaller area of the hand exposed. Also, the radiograph image indicates changes to the location and number of tails on the lower edge of the cape.
In the radiograph the curtain in the background shows a strong vertical highlight, which has subsequently been reworked to form the more complex angular fold we now see – a fold that serves to bring the eye of the viewer back into the portrait.
As noted earlier, it is the alterations to the neck, shoulder and waistline which carry the greatest weight in the development of the image. The radiograph shows the collar of the garment as having initially been drawn strongly vertical, giving significantly more stiffness to the posture. In the final image we see that the Doge’s left shoulderline intersects the collar at a higher point and therefore at a steeper angle than in the initial conception. The change to the shoulderline combines with the alteration to the profile of the cap to create a gentle sweeping curve that reduces the formality of the portrait (this more fluid curve is accentuated in the final image by means of a dark line). Related changes occur in the drawing of the Doge’s right shoulder. The radiograph suggests possibly four variations before the painter decided to work with the shoulder intersecting the neck at a low point.
The development of the neckline and shoulderlines is reflected in the alterations to the position of the waistline and belt. Originally, the waist and belt were drawn lower on the figure, thereby contributing to the overall formality of the portrait. But the waist was then shifted higher, with consequent changes being made to the folds in the garment, giving an enhanced sense of age and, with the changes to the neckline, reducing the stiffness of the image.
It should be noted that the waist and belt in the Kimbell version are placed almost half-way between the original and the revised position of the waist and belt in the Melbourne picture. Other slight changes that occur in the translation from the Melbourne to the Kimbell version include the following: in the latter painting the curve of the edge of the cape is flatter, the Doge’s right hand is higher, the folds in the background fabric are more complex. If it were not for these shifts, the translation would be exact.
Though the Melbourne painting is now smaller in dimensions than the Kimbell picture, the correspondence between the two images suggests a very direct relationship between the paintings. This direct correspondence of the images, together with the clear sense that the final conception of the portrait was worked out during the execution of the Melbourne painting, resolves the issue of which painting came first.
The informality that is such a notable feature of both portraits is a result of the changes that have taken place during the development of the Melbourne painting. It is in the Melbourne work that we see the shifts in posture, and a refinement of detail, taking place.
The radiograph of the Melbourne version offers an opportunity to reassess not only an individual work but also the working method of the artist’s studio. The similarities in the images under discussion, and the apparent similarity in the brushwork as revealed by the radiographs, support our contention of the primacy of the Melbourne version and point to the way in which the creative process evolved in the development of this portrait image.
John Payne, Painting Conservator, National Gallery of Victoria (in 1993).
This article would not have been possible had not the Kimbell Art Museum generously made available the radiograph of its painting for comparison. Both Dr Ursula Hoff and Dr Emma Devapriam have supported the Melbourne painting as the primary version and this article acknowledges and supports their view.
1 Ε. Pillsbury & W. Jordan, ‘Recent Painting Acquisitions – III: The Kimbell Art Museum’, Burlington Magazine, vol. CXXIX, no. 1016, November 1987, p. 773.
2 Discussing the Kimbell version, Pillsbury and Jordan note: ‘The painting has been accepted as Tintoretto’s primary version of the subject by Berenson, Tietze, Pallucchini, De Vecchi and all modern writers except VON DER BERCKEN who … maintained that the J. Ross Delafield picture [the Kimbell version] was a replica of the Melbourne version’ (p. 771 n. 11). Bercken’s position is reiterated by U. Hoff in European Painting and Sculpture before 1800, 3rd edn, National Gallery of Victoria, Melbourne, 1973, p. 154.