Neroccio di Bartolomeo de’ Landi, The Annunciation

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

The panel support, of a horizontal wood grain, has been neither thinned nor cradled and is 3.8 centimeters thick. It comprises four planks measuring approximately 10, 11.5, 13.5, and 13 centimeters, from top to bottom. The seams between the top two planks and the second two planks have opened slightly; the bottom seam is well adhered except at the furthest-right edge, viewed from the front. Three tapering dovetail channels once holding battens, probably not original, are cut across the grain and were filled with sections of wood running with the grain during a treatment by Gianni Marussich in 1998. The outer edges of the panel show no sign of cutting in modern times, and scattered remnants of gesso preparation along them may be original. The lower 7 centimeters at either side of the panel are cut in a straight vertical, cropping the arc of the lunette short of its full extension; it is unclear whether this is original or a later intervention.

The paint surface overall is lightly abraded but reasonably well preserved. Small losses around the edges, along the center seam, in the foot of the Virgin’s bed, and beneath the Archangel’s right foot and above his left foot—the latter caused by a knot in the panel support—were inpainted by Patricia Garland in 1998–99. The Virgin’s blue robe, by contrast, is missing almost entirely and was completely reconstructed in this campaign. Mordant gilding in the haloes of the two main figures, surrounding the apparition of God the Father at the top center, and defining the vanes of the archangel Gabriel’s wings is largely missing and has not been reconstructed, but the delicate patterns it formed are visible in small remnants of leaf and in the glue residue where it was applied.

One of the most justly famous, frequently published, and imperfectly understood paintings in the collection of the Yale University Art Gallery, the Annunciation depicts the moment of the Incarnation of Christ as the Archangel Gabriel delivers the divine salutation to the Virgin Mary. Loosely based on the composition of Simone Martini’s authoritative model in the altarpiece he painted with Lippo Memmi for Siena Cathedral in 1333¹ but adapted to a lunette-shaped picture field, the Virgin is shown at the right, seated demurely on a carved marble throne with gilded griffon armrests and an aedicular backrest with gilded capitals. She is interrupted in her reading by the dynamic entrance of the Archangel at the left. Dressed in a blue tunic and billowing pink robe, he kneels in flight, his right hand raised in a gesture of blessing and his left holding a curling banderole inscribed “AVE GRATIA PLENA.” A gilded vase with a stalk of lilies stands on the marble pavement between the two figures. At the far right is a view into the Virgin’s bedchamber; a building behind the Archangel at the left is decorated with a frieze of classical figures in relief, as is a low stone wall closing off the courtyard at the rear. A door cut into this wall offers a glimpse of a river landscape; four trees are visible rising above the ledge of the wall, and a vision of God the Father—holding a globe and blessing the Virgin before and below Him—appears in the clouds at the top of the sky.

Although James Jackson Jarves believed the Annunciation to be Florentine and, specifically, by Piero del Pollaiuolo, this opinion did not outlast the various manuals of his collection published between 1860 and 1871.² William Rankin said of the painting in 1895 only that it has nothing in common with the work of the Pollaiuolo brothers.³ While he declined to offer an alternative attribution, he did introduce a critical observation that has dogged scholarship of the painting ever since: that “the principal figures are so much weaker” than the remarkable spatial sense of the architecture and the landscape “as to suggest a less vigorous hand.” Bernard Berenson first recognized the correct attribution of the painting to the Sienese master Neroccio de’ Landi, an identification confirmed by F. Mason Perkins and by Charles Fairfax Murray and repeated in every published reference to the painting since then.⁴ Luigi Dami, expanding upon Rankin’s observations, or arriving independently at the same conclusion, introduced the suggestion that while the figure of the Virgin is incontrovertibly by Neroccio, the limpid spatial structure and architectural background of the scene could not have been his invention: “In nessun’altra opera posteriore Neroccio saprà preparare per le sue figure un teatro così vasto e così splendido” (In no later work did Neroccio know how to prepare so vast and splendid a theater for his figures).⁵ Dami contended that few Sienese painters around 1475—the date he felt most suitable for the Annunciation—were capable of understanding this overtly Florentine approach to pictorial construction, Francesco di Giorgio being a notable exception.

The intimations of Dami’s critique were developed more fully by Frank Mather, who assigned the painting outright to Francesco di Giorgio, and by Arthur McComb, Helen Comstock, George Edgell, Selwyn Brinton, and Raimond van Marle, who viewed the painting as a collaboration between the two artists.⁶ This point of view became ensconced as a commonplace in subsequent studies and was repeated or amplified by Allen Weller, Gertrude Coor, Charles Seymour, Jr., Gustina Scaglia, Ralph Toledano, Max Seidel, and others.⁷ A contrary opinion was articulated by John Pope-Hennessy, who observed that “the fact that Neroccio shared a studio with Francesco di Giorgio [between 1469 and 1475] has given rise to the belief that the two artists were jointly responsible for executing certain paintings . . . but there is no valid reason to believe that any extant panel is the joint product of both artists.”⁸ Pope-Hennessy’s position remained a minority point of view. Seymour went so far as to claim that “from the recent X-rays of the Annunciation it is beyond question that two hands are present in the work; moreover, the familiar styles of both Neroccio and Francesco di Giorgio are clearly discernable in each.”⁹ The X-rays in question were published by F. Clancy, who added that “the archaically passive Virgin may be given to the conservative Neroccio. Francesco, however, must be associated with the more active and weightless figure of the angel. The figure of God the Father must also go to Francesco. The dynamic and foreshortened pose is unlike anything Neroccio produced. . . . Finally, the architectural setting must logically go to Francesco in view of his well known involvement in architecture.”¹⁰ The fallacious logic of these assertions was first called out by Everett Fahy, who correctly observed that an attribution of the Annunciation to two distinct artists is untenable: “The recent X-rays . . . are homogeneous in appearance, revealing no discrepancies in handling.”¹¹ Garland presented further technical evidence that the structure of the paint layers is uniform in distinct parts of the picture surface.¹² While her conclusion does not follow that this demonstrates authorship by one painter alone, visual evidence does argue that no part of the Annunciation was painted by any artist other than Neroccio.

The persistent belief that the Jarves Annunciation was a collaborative work led to two corollary assumptions repeated through most of the literature concerning it. The first of these was that the painting must date between 1469 and 1475, when Neroccio and Francesco di Giorgio are documented as having formed an artistic partnership. The second assumption followed from the first: that the altarpiece for which this painting originally functioned as the lunette cannot be identified, since no altarpieces by Neroccio from this period are known and none by Francesco di Giorgio is either of an appropriate size or missing a lunette. Even those authors who rejected the theory of collaboration within the work clung to an early date for it—Pope-Hennessy, for example, believed it to have been painted ca. 1470, while Luke Syson opted for the years between 1475 and 1480—and therefore continued to believe that it is the lone surviving fragment of an unidentified altarpiece.¹³

It has often been remarked that Neroccio’s style changed relatively little over much of the course of his career, becoming noticeably different only in the last few years of his life. This contention, however, is partly the result of random or inattentive dating of his many devotional images of the Virgin and Child. Relying on the fixed points of his dated altarpieces of 1476, 1492, and 1496, a progressive chronology of his development is not difficult to construct, leading from the decorative linearity and blonde tonalities of his early efforts through an incipient naturalism and increasingly atmospheric modeling to the solid forms and warm tones of his last works. Within such a chronology, the most telling points of comparison for the Jarves Annunciation are consistently to be found with works clustering around the 1492 altarpiece from Montepescini, now in the Pinacoteca Nazionale, Siena (fig. 1). Paintings such as the Stoclet Madonna formerly in Brussels¹⁴ or the Virgin and Child with Saints Bernardino and Catherine of Siena in the Pinacoteca Nazionale, Siena,¹⁵ share with these the studied fall of light across gently rounded forms that is at once more advanced than the artist’s effete drawing and palette in the 1476 Saint Michael altarpiece, also in the Pinacoteca,¹⁶ and less developed than his overt attempts to imitate Luca Signorelli’s oil-based modeling in the 1496 Montisi altarpiece.¹⁷ The type of the archangel Gabriel in the Jarves Annunciation finds an exact counterpart in the figure of Saint Sebastian standing in the Montepescini altarpiece to the left of the Virgin’s throne, while the elaborate carved decoration of this throne—including its griffon armrests—forms a suitable analogue to the decorative forms of the seat and low garden wall in the Annunciation. The Jarves Annunciation is exactly the same width as the Montepescini altarpiece (128.5 cm). The conclusion is inescapable that they originally formed part of a single liturgical ensemble.¹⁸

Restoring the lunette to its original context atop a conventional rectangular altarpiece with gold ground and punched margin, an isocephalic line of standing saints clustered closely together, and the Virgin seated on a throne, the architecture of which is projected in exaggerated foreshortening to enhance its decorative effect, forces a reevaluation of the supposedly innovative spatial structure of the courtyard setting of the Annunciation. Neroccio has borrowed the trappings and vocabulary of Albertian perspective to lay out the grid of his marble pavement but in an irrational manner that creates a fantasy setting, not a measurably coherent illusion of depth. The receding orthogonals do converge in a single vanishing point, but this is considerably and inappropriately higher than the horizon line. The intervals between parallels marking depth in space, furthermore, do not diminish according to any mathematical progression; some of the intervals actually lengthen as they move backward, creating an unresolved tension between the notional distance behind the picture plane the parallels are meant to indicate and the height along the picture plane that they actually displace. This tension is most acutely noticeable in the exaggerated distortions of geometric form visible at the extreme ends of the foreground, a distortion familiar to amateurs of Sienese painting from the freely contrived fantasy landscapes of Giovanni di Paolo. The figures, furthermore, are not situated logically within this spatial grid nor are they scaled in such a fashion that would allow the grid to function as a matrix lending empirical credibility to the setting. This is, in fact, the furthest thing from the “perfectly realised linear perspective” that it is often claimed to be.¹⁹

The contention, finally, that the grisaille figures adorning the building facade behind the Archangel at the left and the parapet wall closing off the courtyard in the background are too accomplished for Neroccio’s capacities and too accurately reflect familiarity with classical prototypes to reveal his essentially Gothic predisposition as an artist is baseless. Neroccio was a collector of classical statuettes, a number of which were catalogued in the inventory of his estate after his death.²⁰ He was also responsible, contrary to prevailing opinion, for the not dissimilar figures painted only a few years after this in the background of his Claudia Quinta in the National Gallery of Art, Washington, D.C.²¹ Whether the individual figures included in the frieze(s) in the Jarves Annunciation were intended to have specific and linear iconographic associations relevant to the painting’s theme or were included for their decorative value alone is debatable. It is certain, however, that the artist did not need to call upon a specialist collaborator for their realization. —LK

1. Gallerie degli Uffizi, Florence, 1890 nos. 451–53, https://www.uffizi.it/en/artworks/annunciation-with-st-margaret-and-st-ansanus. ↩︎

2. Jarves 1860, 53, no. 76; and Brown 1871, 21, no. 65. ↩︎

3. Rankin 1895, 148. ↩︎

4. Berenson 1897, 206; Berenson 1902, 156; Perkins 1905, 77; and Murray, as reported in Rankin 1905, 11. ↩︎

5. Dami 1913, 141–42. ↩︎

6. Mather 1914, 968; McComb 1924, 19; Comstock 1928b, 35; Comstock 1946, 51–52; Edgell 1932, 247; Brinton 1934, 109; and van Marle 1937, 262, 286, 292. ↩︎

7. Weller 1943, 50, 70–71, 74, 82, 398; Coor 1961, 43–44; Seymour 1970, 191–92; Scaglia 1980, 9, 23n26; Toledano 1987, 90–91; and Seidel 1991, 57–59. ↩︎

8. Pope-Hennessy 1947, 21. ↩︎

9. Seymour 1970, 192. ↩︎

10. F. Clancy, in Seymour et al. 1972, 28. ↩︎

11. Fahy 1974, 285. ↩︎

12. Garland 2003b, 66, 68. ↩︎

13. Pope-Hennessy 1947, 22; and Luke Syson, in Syson 2007, 138. ↩︎

14. Coor 1961, no. 6, fig. 43. ↩︎

15. Inv. no. 285. ↩︎

16. Inv. no. 282. ↩︎

17. Coor 1961, no. 26, fig. 79. ↩︎

18. Possibly but not necessarily incidental to this reconstruction is the fact that the blue of the Virgin’s robe in both altarpiece and lunette are deteriorated to the same degree and in the same fashion. ↩︎

19. Luke Syson, in Syson 2007, 137. ↩︎

20. The inventory, dated November 26, 1500, is published in full by Milanesi 1854–56, 3:7–9; and in Coor 1961, 152–59. Item 183 in the inventory is “una testa di tucto rilievo antica” (an antique head fully in the round); item 267 is “43 pezi di forme di rotture antiche di gesso” (gesso [copies] of antique fragments); and item 268 is “3 gessi d’Apollo” (3 gesso [copies] of Apollo). Numerous pieces of “marmo antico” (antique marble) might also refer to antiquities or antique fragments but are not described in sufficient detail to identify with confidence. ↩︎

21. Inv. no. 1937.1.12, https://www.nga.gov/collection/art-object-page.14.html. The most useful study of the eight figures of heroes and heroines that formed a suite with the Claudia Quinta is to be found in Syson 2007, 234–45, where the proposal is endorsed that the series was begun in 1493 and that the panels by Francesco di Giorgio (Scipio Africanus, Museo Nazionale del Bargello, Florence, inv. no. 2023 Carrand) and Neroccio were completed by the Griselda Master. A contrary opinion, that the paintings could be two or more years later and that the Claudia Quinta is entirely to be attributed to Neroccio, is expressed by the present author in Kanter 2000b, 147–55. ↩︎