Agnolo Gaddi or Lorenzo Monaco, Saints Julian, James, and Michael

James Jackson Jarves (1818–1888), Florence, by 1859

The panel support, of a vertical grain, retains its original thickness of 2.7 centimeters although it has been beveled around all its edges on the back. It is composed of four planks measuring, from left to right, 25.5, 21, 12.5, and 15 centimeters in width. Discoloration of the wood indicates the placement of a batten, approximately 8 centimeters wide, affixed across the panel 30 centimeters from the present bottom edge. Four nails securing this batten remain: none are driven into the third plank, and one in the second plank was driven in so close to its right edge that it provoked a long split in the wood. The second plank contains two large knots in the wood. The third plank is reused from an earlier structure: two horizontal grooves 38 and 57 centimeters from the bottom edge, possibly for the attachment of a handle (two nails and fragments of wood remain in the lower groove), are cropped at its left edge. A large hand-cut nail is driven through the fourth board and into the third along the depth of the panel near the top, presumably to resecure the join between them.

When the panel arrived at Yale, its upper profile had been extended, and it had been provided with a new gold ground, including elaborately punched and lettered haloes for the saints (fig. 1). The extensions and nineteenth-century gilding were removed by Andrew Petryn in 1958, exposing the wood support behind the saints and the gesso underlayer in several large local losses in the lower draperies at the left and center (fig. 2). The latter were inpainted by Andrea Rothe in a treatment of 1998, when it was also decided to regild the background, including the addition of an artificial, engraved craquelure, but to imitate the effect of the haloes only by a scribed perimeter line. In contrast to this distracting and unsightly intervention, the paint surface is beautifully preserved outside the discreet areas of total loss near the bottom, exhibiting only modest abrasion throughout.

Following initial generic attributions to Taddeo Gaddi¹ or, more vaguely and dismissively, to an unknown Giottesque painter,² the Yale Saints Julian, James, and Michael has been universally recognized as a work from the circle of Agnolo Gaddi. It was first labeled with the name Gherardo Starnina,³ when that artist was believed to have been a follower of Gaddi responsible for, among other things, the Castellani Chapel frescoes in Santa Croce and an altarpiece triptych in Berlin, the Gaddesque work most closely related in style to the Yale panel. Richard Offner more prudently and evasively called it simply a product of Agnolo Gaddi’s workshop,⁴ while Bernard Berenson, Hans Gronau, Charles Seymour, Jr., Federico Zeri, and Miklós Boskovits accepted an attribution to Agnolo Gaddi without qualification.⁵ Bruce Cole and Andrew Ladis returned to Offner’s model of identifying a small number of autograph creations by Agnolo Gaddi within a large orbit of workshop imitations; both Cole and Ladis included the Yale panel among the latter.⁶

Osvald Sirén, the first scholar to study the Yale panel in any detail, over a century ago, recognized it as the left wing—or a fragment of the left wing—of an altarpiece triptych,⁷ but only Gronau and Boskovits have advanced suggestions for identifying any other parts of its original complex. Gronau proposed a reconstruction based on a Virgin and Child Enthroned by Agnolo Gaddi (fig. 3) that had been discovered near Pisa, framed together with a predella panel believed also to be by Gaddi representing two miracles of Saint Michael the Archangel.⁸ Both panels passed into the Achillito Chiesa collection in Milan, where they were separated: the Virgin and Child is now in the Contini Bonacossi Collection at the Gallerie degli Uffizi, Florence, and the predella is now in the collection of the Yale University Art Gallery.⁹ Gronau assumed these two panels came from a single source—even though the predella would originally have stood beneath a lateral panel in its altarpiece, not the central panel—and he proposed joining Saints Julian, James, and Michael to the complex as the figure of Saint Michael within it would thus have stood in the position of honor, immediately to the right of the enthroned Virgin and above the predella panel now at Yale. He added to his reconstruction another predella panel showing two further scenes from the legend of Saint Michael, coincidentally also in the Yale collection and now recognized to be the work of Lippo d’Andrea (see Lippo d’Andrea, Two Scenes from the Legend of Saint Michael), and a third predella panel in the Vatican Pinacoteca,¹⁰ but he found no other fragments of the main tier of the altarpiece.

Gronau’s reconstruction was based exclusively on iconography—representations of scenes from the legend of Saint Michael are not common in early Florentine painting—and did not take into account the fact that the three predella panels had been painted by three different artists at widely separated dates. Additionally, both predella panels now at Yale feature a division of the picture field into two scenes separated by a gilded border, implying that each of them came from a pentaptych rather than a triptych, whereas the composition of Saints Julian, James, and Michael is more appropriate to a large altarpiece triptych. Boskovits rightly rejected Gronau’s proposal, although initially he accepted the relationship between the Chiesa Virgin and Child now at the Uffizi and the first of the two Yale predella panels.¹¹ He recently advanced an alternative proposal for Saints Julian, James, and Michael, linking it with a panel in the National Gallery of Art, Washington, D.C., representing the Coronation of the Virgin (fig. 4).¹² This association was defended in part on grounds of iconography, as images of the Coronation are frequently accompanied not by single saints in individual compartments but by the court of Heaven as symbolized by the presence of three (or more) saints within a single picture field, as in the Yale panel. The Washington and Yale panels, however, are not as closely linked in style as Boskovits argued—the Washington panel being a slightly later work—and batten nails in the two works do not align. The unusually low placement of the batten on the Yale panel, at the level of the figures’ hands, may imply that it was the center of three rather than the uppermost of two battens and that the panel once continued upward, possibly including clerestory figures that do not correspond to any other surviving altarpiece panels from the Gaddi circle. Evidence on the reverse of the Yale panel, furthermore, suggests either that it may possibly have been repurposed at a relatively early date for use as a door or that it was constructed from reused lumber, some of which had previously been part of a door. If the former were true, it may be that the complex of which the Yale panel formed part may have been dismantled much earlier than is usual with surviving trecento altarpiece fragments and, consequently, that no other members of the original structure survive.

Despite the preference of modern scholars to favor a more inclusive definition of the range of Agnolo Gaddi’s style and level of achievement—and their attendant rejection of overscrupulous efforts to mine his oeuvre in search of a cadre of minor masters working in his shadow—it is essential to consider the vast output, especially of fresco commissions, for which the artist was responsible over a relatively short span of time, from the mid-1370s to the mid-1390s. The efficiency and sustained level of this production, which far outstripped that of any other painter in late trecento Florence, would not have been possible without a stable, well-organized, and accomplished team of studio assistants, and it is not unreasonable to assume that some of these assistants might have gone on to pursue independent careers after Agnolo Gaddi’s death. The equivocal results of so many past attempts to assemble groups of paintings that might reveal distinct personalities is due not to a fallacy of concept but to a reliance on superficial criteria of discrimination. One group at least has been successfully isolated and broadly accepted as the work of a distinct artist, whom Gronau and Zeri have identified as the young Lorenzo Monaco.¹³ This group centers around the three predella panels and four pilaster bases of a dismembered altarpiece from the Nobili Chapel in Santa Maria degli Angeli, Florence, a documented work of 1387–88 presumed to have been commissioned from Agnolo Gaddi, with Lorenzo Monaco working as an assistant (see Lorenzo Monaco, Portrait of Bernardo di Cino Bartolini Benvenuti de’ Nobili and Four of His Sons: Bartolomeo, Carlo, Benedetto, Alamanno). While there is no reason to question the presence of Lorenzo Monaco in these seven panels, it is important to note that there is no hard evidence that Agnolo Gaddi was directly responsible for the upper registers of the altarpiece: none of the documentation relating to the commission actually names the artist or artists responsible for its creation. The eight pilaster saints,¹⁴ three pinnacles,¹⁵ and the main tier of this altarpiece¹⁶ may also be the work of Lorenzo Monaco. They are fundamentally different from all other panel paintings attributed to Agnolo Gaddi in their higher key palette that relies on a greater range of contrasts in bright, unmodulated colors and in their tendency to reinforce the outer contours of draperies by modeling them in deep, saturated tones rather than bordering them with a simple outline. Their drawing is more cursive and tends to exaggerate the plasticity of details, such as hands or of drapery folds enlivened with deep pockets of light and shadow, while simultaneously suppressing all but the most basic spatial devices. Agnolo generally inverts this process, preferring simplified planar forms of flat, clear color with light outlines, arranged in compositions that invoke a stronger notional sense of space.

Of all the surviving paintings presumed to have emerged from Agnolo Gaddi’s studio, the closest comparable works to the Yale Saints Julian, James, and Michael are the main panels of the Nobili altarpiece now in Berlin (see Lorenzo Monaco, Portrait of Bernardo di Cino Bartolini Benvenuti de’ Nobili and Four of His Sons, fig. 7). At the very least, this resemblance suggests a date for the Yale panel close to 1388. It also implies that it could be the work of Lorenzo Monaco, not Agnolo Gaddi. Consensus on this point may not be possible, but it is worth noting that the chief obstacle to integrating this painting (as well as the panels in Berlin) into the early career of Lorenzo Monaco is their incompatibility in style with the so-called Carmine Altarpiece, widely believed to be a documented work by Don Lorenzo from the late 1390s. First published by Zeri, the exceptional quality of each of the fragments of this altarpiece is regularly adduced to support their attribution to Lorenzo Monaco.¹⁷ Only Luciano Bellosi has seriously doubted this attribution, and although generally ignored, his objections deserve to be considered more seriously.¹⁸ The drawing in these panels is finer and less forceful than Lorenzo Monaco’s usual standard, while the palette is softer—more nuanced and less reliant on bright contrasts—and the range of colors more subdued. Spatial effects are both more ambitious and more lucid, and architectural detail, such as the finials of the Virgin’s throne, are articulated with a precision and structural clarity encountered nowhere else in Lorenzo Monaco’s oeuvre. Whether the presumed dating of the Carmine Altarpiece to 1398–99 is accurate or only approximate, it is irrelevant to a consideration of the chronology of Lorenzo Monaco’s development. A linear progression from the Yale and Berlin Saints, through the large figures of the San Gaggio altarpiece (in Florence and London¹⁹) and the scenes from its predella (in Berlin and Santa Barbara, California²⁰), to the Agony in the Garden in the Galleria dell’Accademia, Florence,²¹ is coherent and persuasive and argues on the whole for an attribution to Lorenzo Monaco for the Yale Saints Julian, James, and Michael. —LK

1. Sturgis 1868, 34, no. 21; Brown 1871, 14, no. 21; and Rankin 1905, 8, no. 21. ↩︎

2. Rankin 1895, 140, no. 21. ↩︎

3. Sirén 1916a, 51–53, no. 20, fig. 20; and van Marle 1924b, 573. ↩︎

4. Offner 1927a, 3, 20–21. ↩︎

5. Berenson 1932a, 214; Gronau 1950a, 44–47, pl. 26/fig. 6, pl. 27/fig. 7; Seymour 1970, 37–39, no. 22, fig. 22; Fredericksen and Zeri 1972, 76, 599; Boskovits 1975b, 298, 301; and Boskovits 2016, 133, fig. 2. ↩︎

6. Cole 1977, 73, 76; and Ladis 1992, 329, 333n20, fig. 12. ↩︎

7. Sirén 1916a, 51–53, no. 20, fig. 20. ↩︎

8. Gronau 1950a, 44–47, pl. 26/fig. 6, pl. 27/fig. 7. ↩︎

9. Inv. no. 1943.213; https://artgallery.yale.edu/collections/objects/45075. ↩︎

10. Inv. no. 528. ↩︎

11. Boskovits 1975b, 298, 301. The Chiesa Virgin and Child has subsequently been united in the Contini Bonacossi Collection at the Uffizi with four lateral panels by Agnolo Gaddi representing Saints Benedict (sometimes said to be John Gualbert), Peter, John the Baptist, and Minias; see inv. no. Contini Bonacossi no. 29, https://www.uffizi.it/en/artworks/triptych-gaddi-contini. See Cole 1977, 76–77; and Giovanna Ragionieri, in Caneva et al. 2018, 70. This reconstruction, which is at least plausible, has been rejected by Miklós Boskovits and others (such as Gaudenz Freuler), who instead suggest including a Virgin and Child Enthroned formerly in the Heinz Kisters collection, Kreuzlingen, Switzerland (sale, Sotheby’s, New York, February 1, 2018, lot 3), with the four Contini Bonacossi lateral panels. See Boskovits 1975b, 298, 300; and Freuler 1991, 203–5. These, however, seem incompatible in date, the ex-Kisters Virgin being an early work by Agnolo Gaddi, possibly from the 1370s, while the Contini Bonacossi Saints are mature works closer in date to 1390. The unusual shape and decoration of the gold ground in the ex-Kisters Virgin, furthermore, is entirely modern. Its similarity to the shapes of the Contini Bonacossi Saints is either fortuitous or imitative, as their frames, too, are modern. Giacomo Calogero has proposed identifying a panel in the Cathedral Museum at Mdina in Malta, representing Saints Julian and Benedict, as part of an alternative reconstruction for the Chiesa/Contini Bonacossi Virgin; see Calogero 2013, 34–39. Based on photographic evidence alone, this reconstruction, while not impossible, is not entirely persuasive. ↩︎

12. Boskovits 2016, 133. ↩︎

13. Gronau 1950b, 183–88, 217–22; Zeri 1964, 554–58; and Zeri 1965, 3–11. A more recent claim to recognize the work of Antonio Veneziano and Gherardo Starnina in passages of the frescoes in the Cappella Maggiore in Santa Croce, Florence, is extremely promising; see Monciatti 2014, 85–111, esp. 101–5. ↩︎

14. Indianapolis Museum of Art, inv. no. 2004.160A–D, https://collections.discovernewfields.org/artwork/75564?fromSearch=true; and Gemäldegalerie, Berlin, inv. no. 1138. ↩︎

15. Fondazione Francesco Federico Cerruti, Castello di Rivoli, Turin, inv. no. CC.17.P.GAD.1387.A26. ↩︎

16. Gemäldegalerie, Berlin, inv. no. 1039. ↩︎

17. Zeri 1965, 3–11. ↩︎

18. Bellosi 1965, 37–38. Marvin Eisenberg initially rejected the panels of this altarpiece from his list of accepted works by Lorenzo Monaco but retracted that position in an unpaginated addendum to his book just prior to publication; see Eisenberg 1989, 50–51, 187–88, and n.p. Angelo Tartuferi reproduced and catalogued the ten known panels from the altarpiece (plus four pilaster saints hypothetically added to it) in Tartuferi and Parenti 2006, 120–27. It should be noted that the gabled shape of the current frames on the main panels (today in Florence and Toledo, Ohio) of the structure is fanciful. It is possible that the pinnacle originally positioned above the center panel is to be identified with a Crucifixion formerly in the Samuel H. Kress Collection and now in the Mead Art Museum at Amherst College, Mass., inv. no. AC 1961.79, https://museums.fivecolleges.edu/detail.php?museum=all&t=objects&type=all&f=&s=1961.79&record=0. Conventionally attributed to Mariotto di Nardo, this painting, of exceptional quality, is fully consonant in style with the Toledo Virgin and the five predella panels from this structure. ↩︎

19. Galleria dell’Accademia, Florence, inv. nos. 1890 nn. 8604, 8605; and Courtauld Institute, London, inv. no. P.1966.GP.272, https://gallerycollections.courtauld.ac.uk/object-p-1966-gp-272. ↩︎

20. Gemäldegalerie, Berlin, inv. nos. 1063, 1108; and Santa Barbara Museum of Art, Calif., inv. no. 1967.15, https://collections.sbma.net/objects/2068/martyrdom-of-pope-caius?ctx=8b14e62b82dfdbd225b76dbadd351338f5678b0e&idx=0. ↩︎

21. Galleria dell’Accademia, Florence, inv. no. 1890 n. 438. ↩︎