Support: The panel was constructed from five pine boards with the grain oriented vertically.[1] The width of the boards from left to right is as follows: 4.7 cm, 22.7, 20.3, 20.9, and 16.0 cm. The panel retains its original reverse and varies in thickness from 1.7-2.2cm. The reverse of the panel is coated with an even layer of a fibrous material[2] (see fig. 7 above). There are two original horizontal cross-pieces, the ends of which were cut away by about 7.5 cm on the left (when looking at the reverse) and about 5 cm on the right, likely during a past structural intervention on the panel. The battens are held with nails hammered in from the front (under the preparation layers) and clinched on the back. Most of the nails on the back have been trimmed, likely when the battens were cut, but the clinching can still be seen in some of the nails in x-ray (fig. 8). Nail fragments and empty nail channels can be seen to the left and right of the cut-down battens about 2 cm from the edges. The lower batten has been partially and roughly carved away, possibly in an attempt to make it more flexible without removing the original nails.
The greenish-blue painted borders on the front are not original and complicate assessment of the original dimensions of the composition and of the entire support. Furthermore, there are non-original wooden additions along the bottom and right edges, located within the area of the greenish-blue borders. The wood addition on the bottom edge was attached to the original panel by a half-lap join; the addition measures 2.8 cm high on the front and 4.8 cm on the reverse. The wood addition on the right was attached by a butt join and measures 3.5 cm wide. The painted borders are about 4cm wide on average; the panel itself is not perfectly square and so they vary in width from 3.5 to 4.5 cm. It can be assumed that the edges of the original panel were at least slightly planed down on the right and bottom to attach the later wood additions, but it is likely that all four edges were trimmed by a greater degree, as observations about the preparation suggest.
The original preparation on the panel appears to extend to the edges. Fibrous material—seemingly similar to that visible on the reverse—is present across the entire panel. It is visible in the x-radiograph as bunches of fibers that are located throughout, including at the extreme edges (fig. 8). It is most noticeable at the ends of the cross-battens, which have been cut away together with the fibers, leaving no confusion that the fibers seen in x-radiograph are on the surface. In fact, the fibers seem to be cut off at the edges: some bunches are truncated halfway, as opposed to how bunches would be expected to taper off, curve, or become slightly sparser, at the edge of the panel.[3] This appearance suggests that the panel was trimmed on all edges by some degree. The x-radiograph also reveals the presence of a lead white-containing priming layer, applied with broad brushstrokes, across the entire original panel. Presumably the fibrous material constituted the lowermost layer, embedded in, and perhaps followed by a white ground layer, with the additional white priming layer on top (see Preparation below).
On the other hand, incisions in the preparation suggest that the original dimensions of the painted and gilded composition are nearly preserved. Located near all four corners, yet not centering the final composition, are short incisions, measuring about 8–12 cm in length, that were made in the white ground before the gilding was applied (fig. 9). The incisions along the left and the top are located near the edges of the painted border. The incisions along the bottom line up with the bottom of the banderoles and along the right nearly line up with the right side of the niche for Saint Francis. However, there is original gilding beyond the incised lines on the right and bottom: 3 cm and 2.5 cm respectively. Comparing the width of the original gilding at these perimeters with the left and the top reveals that it is slightly narrower on right and bottom, that is to say, the composition is not exactly centered by the current bluish-green borders either. In fact, the end of the gilded surface would be expected to lie another .5 cm to the right of the current right edge of the gilding and another about .9 cm below the current bottom edge of the gilding, at just the point where the original panel was trimmed, and the addition adhered. Importantly, although the preparation is present up to the edges, there is no evidence of gilding beyond the incisions underneath the greenish-blue border. Thus, it is likely that similar corner incisions were once present where the wood is now trimmed, and that they represented the original dimensions of the composition.
The overall panel could have been somewhat larger on all sides, but it is difficult to estimate its appearance without knowledge of the rest of the altarpiece and its construction. Two final pieces of evidence support the conjecture that the overall panel is nearly preserved. A series of small, old nail holes near the top of the panel are apparent in the x-radiograph. The location of these, extending slightly into the top of the composition, suggests that they could be related to an applied tracery frame molding. The x-radiograph also reveals the presence of two plugs in square-shaped holes at the lower corners, possibly related to face-nailed applied cross-grain framing battens or to a larger polyptych construction.[4] The board on the left side is narrower in width relative to the others; it is likely that all were similar in their widths. Additionally, evidence of nails on the far right of both battens are unusually close to the panel edge also suggesting this board was wider.
Preparation and Underdrawing: As noted above, the panel was prepared with a whitish ground, which was not analyzed, as well as an additional lead white-containing priming layer. It appears that the front of the panel was prepared with a fibrous material, like that visible on the back. In the absence of cross-sectional paint samples, the exact stratigraphy cannot be determined with certainty. The gilded regions were further prepared with an ochre-colored mordant or bole (see Gilding below for more observations about the gilding and the stratigraphy). In addition to the corner incisions that seem related to the dimensions of the composition, the saints’ niches were also incised prior to gilding.
Examination with infrared reflectography revealed the presence of an extensive underdrawing, executed with a liquid medium (fig. 10).[5] The underdrawing described all of the important contours of the two saints as well as the dragon. The artist also used much hatching to create shading. These hatchmarks, consisting of broad, short, and curved lines, are quite similar in appearance to the wet-in-wet paintstrokes apparent on the surface (see Paint Layers for more detail). The artist made a few significant changes to the underdrawing in the painting stage, including Michael’s diadem, the angle of his wings, of his right arm, and the appearance of the dragon (see Catalogue Entry and Ainsworth 2008, pp. 113–15).
Paint Layers: The painting technique is arresting for integrating occasional instances of quite virtuosic brushstrokes, and bold strategies with a highly polished painting approach. In general, the fleshtones are well blended, with smooth transitions that lend the figures a sculptural quality. In the x-radiograph, it is clear that the artist mixed some lead white throughout the fleshtones, but not to the point of achieving much opacity. The thinly applied paint takes advantage of the white ground to yield an alabaster-like glow in the figures.
There are a few quirks in the technique, which may be related to the scale of the painting and of the entire altarpiece to which it belonged, and its intended viewing distance. Some elements seem quite exaggerated, for example, the blue at Francis’s stigmata and wound, which at close viewing is more than the slight bluish tinge of wounded, bruised flesh but an unearthly blue that reads even from a great distance. Likewise, the shape of Francis’s eyebrow and eyelid—which look like a series of stylized waves, with points and curves—appear harsh and unnatural at close viewing, but at a greater distance effectively convey the saint’s emotion. Saint Michael’s eyes are haunting at a distance for their great depth, and up close it is apparent that the artist achieved this by eschewing a contour for the lower lid in favor of a gradation of black, softly blending into grey, beneath his downcast upper lid.
In contrast to the highly blended paint noted above, the artist made effective use of bold, visible brushwork in select areas. In the wings of Saint Michael, a brush—perhaps a dry brush—was used to add short parallel hatchmarks into the wet paint, which partially exposed some of the bright white ground beneath, yielding a sense of light shimmering on the feathers. A similar effect is apparent along a highlight on Francis’s tonsured head. Perhaps this open, virtuosic brushwork was also adopted with the large-scale and the viewing distance in mind.
Overall, the painting is in relatively fair condition considering its age but has suffered from varying degrees of abrasion. In some areas only the upper glazes are slightly rubbed, while in perhaps the most compromised passage, in Michael’s shield, the abrasion partly interrupts forms. In another example, the alteration in Francis’s brown robe is only truly apparent at close inspection, where the broad strokes of the underlayer are quite apparent due to the thinness of the upper paint layers. The condition must be kept in mind when making comparisons with other works. At first inspection, the hands of Saint Francis seem quite different in technique from hands in other paintings by the artist, with far less detail and definition when compared with those in
Christ Appearing to his Mother (The Met,
22.60.58), for example, which are far smaller in scale. The rubbing of the upper layers of paint somewhat interrupts the continuity of the fleshtones, which were likely incredibly smooth originally, and diminishes the alabaster-like quality. In closer examination, there are some brown contours for creases in the fingers remaining, as also seen in the Christ panel, and the two are in fact comparable, even accounting for the great difference in scale.
In addition, paint that originally overlapped the gold background was especially vulnerable to damage, as is commonly encountered in paintings of this age. The trompe l’oeil banners for the two saints have also suffered significantly: the paint on the inscriptions is rubbed and the red and blue fictive nails that affixed the banners are damaged. Old restorations attempted to clarify the remaining paint. The blue line at the lower edge of Saint Michael’s banner is not original, and somewhat mars the trompe l’oeil effect.
Gilding: The gilding has elements indicative of mordant gilding including: the regularly spaced lines of loss in the gold, as would happen where squares of gold leaf met but did not overlap when applied over a mordant, as well as the general crack pattern that is characteristic of mordant gilding.[6] However, there are a few instances of “lay lines,” regions where adjacent gold leaves overlap, which can only occur in water gilding. These are most readily apparent near the top of the panel, and it is likely that there are other instances of these lay lines, but these are now hard to detect due to the general condition issues, including rubbing, later toning, and re-gilding. It is difficult to account for the evidence of two different systems of gilding unless the panel underwent two wholescale gilding campaigns, first with water gilding and then, perhaps due to damage, it was re-gilded using a mordant. Alternatively, the gilder could have used some sort of hybrid method. In addition to the earliest gilding campaign, there is also a significant amount of restoration, including local re-gilding.
The warm brown glaze on the gilding is partly original, but seemingly strengthened in later treatments. In some passages, the handling of the glazes is similar to the painting technique in the figures, including some discrete strokes and hatches, to the upper right of Francis and the scraping away of the glaze with parallel hatching—possibly a stiff, dry brush—to the left of Francis’s thumb. These areas of glazing must be original. There are a few instances of fabric tamping where the glazes are the darkest.
As noted above, no incisions were found for the figures, only the niches, thus the gilder must have followed the underdrawn design. The painter then finessed the contours slightly, by overlapping the gold background where necessary, and sometimes by a substantial amount, as in the tail of the dragon. The saint’s name banners were painted on top of the gilding, but the lack of reserve for these does not necessarily mean that they were not planned from the outset: the gilder may not have wanted to leave gaps in his gilding for such small forms, preferring continuity in his application.
Sophie Scully 2023
[1] It is not possible to date pine wood by dendrochronology, report on the wood type by Peter Klein, May 24, 2015, is in the files of the Department of Paintings Conservation.
[2] The fibers are currently being analyzed and the results of this investigation are forthcoming.
[3] Observations made by Alan Miller, Conservator, and Kristin Holder, Research Scholar, Department of Paintings Conservation.
[4] See n. 3.
[5] Infrared reflectography was acquired with a Merlin Indigo InGaAs near-infrared camera fitted with a StingRay macro lens customized for the wavelengths covered by the camera, 0.9 to 1.7 microns, 2007.
[6] Observations on the gilding technique were made by Cynthia Moyer, Associate Conservator, Department of Paintings Conservation.
