Madonna and Child, c. 1310/1315

The painting in the National Gallery of Art presents the Virgin and Child according to a variant of the compositional scheme known as the Hodegetria Virgin, in which Mary’s right hand, instead of indicating her Son, holds a rose that identifies her as the “rose of Sharon”[1]  [1]
Song of Solomon 2:1: “I am the rose of Sharon, and the lily of the valleys.” In medieval religious thought, the rose is often linked to the figure of Mary, who is praised as rosa speciosa or rosa gratiae divinae; cf. M. Schmidt and S. Egbers, “Rose,” in Marienlexikon, ed. Remigius Bäumer and Leo Scheffczyk, 6 vols. (St. Ottilien, 1993), 5:548–552. Eugenio Battisti’s conjecture of a connection between flower and rosary, on the other hand, is unfounded. See Eugenio Battisti, Piero della Francesca, 2 vols. (Milan, 1971), 1:389, 528 n. 548. It cannot be excluded that the iconography of the Virgin with a rose was influenced by transalpine models, for which see Julian Gardner, “Panel Paintings Attributed to Giotto in American Collections,” Center / National Gallery of Art, Center for Advanced Study in the Visual Arts 19 (1999): 71–72; Julian Gardner, “Giotto in America (and Elsewhere),” in Italian Panel Painting of the Due­cento and Trecento, ed. Victor M. Schmidt, Studies in the History of Art (Washington, DC, and New Haven, 2002), 171–176. However, the motif of the Madonna presenting the Christ child with a rose predates Giotto and is attested in Tuscan painting since the thirteenth century; cf. Dorothy C. Shorr, The Christ Child in Devotional Images in Italy during the XIV Century (New York, 1954), 114. It is also known that an oratory was built around an image of the “Madonna delle Rose” at Lucca in 1309; see Angela Protesti Faggi, “Un episodio di protogiottismo a Lucca: La ‘Madonna della Rosa,’” Antichità viva 27 (1988): 3–9. and hence the Church, mystic bride of Christ.[2]  [2]
Ever since the third century, the Canticles (the Song of Solomon) had been considered an allegorical description of God’s relation with the Church and hence in medieval exegesis an expression of Christ’s love for the Church. See Frank L. Cross, ed., The Oxford Dictionary of the Christian Church (London and New York, 1958), 1270. The gesture of the Christ child, his left hand extended to grasp his mother’s forefinger, also presumably has a symbolic as well as a playful or affectionate significance; in other versions of the composition the child even pulls on his mother’s hand with the forefinger pointed towards him, as if actively soliciting her designation of him as a lamb, sacrificial victim.[3]  [3]
“The isolated forefinger of the Virgin’s right hand would seem to possess a variety of meanings,” observed Dorothy C. Shorr in The Christ Child in Devotional Images in Italy during the XIV Century (New York, 1954), 168. In the case of the painting being discussed here, since the right hand of Mary is occupied in holding the symbolic rose, the artist used the other hand in a Christological sense: the Christ child grasps his mother’s forefinger as if to point it towards himself, thus designating him as victim. Shorr also observed, “This pointing finger has already been seen in French sculpture of the late thirteenth century, when the so-called Vierge dorée of Amiens Cathedral points to the Child seated on her arm.”

Both the profile and shape of our painting, known to scholars as the Goldman Madonna, suggest that it was the centerpiece of a dispersed polyptychPolyptych
 Type of object with several panels, usually an altarpiece, although it may also fulfil other functions. The polyptych normally consists of a central panel with an even number of side-panels, which are sometimes hinged to fold. Although in principle every object with two panels or more may be called a polyptych, the word is normally used as a general term for anything larger than a triptych. As with diptychs and triptychs, the size and material can vary. —Victor M. Schmidt, Grove Art © Oxford University Press [fig. 1]   [fig. 1] Reconstruction of a dispersed polyptych by Giotto: a. Saint John the Evangelist (fig. 2); b. Madonna and Child; c. Saint Lawrence (fig. 3) (see also Reconstruction). The appearance of the multipart altarpieceAltarpiece
 An image-bearing structure set on the rear part of the altar, abutting the back of the altarblock, or set behind the altar in such a way as to be visually joined with the altar when viewed from a distance. It is also sometimes called a retable, following the medieval term retrotabulum. The altarpiece was never officially prescribed by the Church, but it did perform a prescribed function alternatively carried out by a simple inscription on the altarblock: to declare to which saint or mystery the altar was dedicated. In fact, the altarpiece did more than merely identify the altar; its form and content evoked the mystery or personage whose cult was celebrated at the altar. This original and lasting function influenced the many forms taken by the altarpiece throughout its history. —Alexander Nagel, Grove Art © Oxford University Press to which our Madonna and Child originally belonged remains uncertain due primarily to the various restorations that have altered its appearance, especially by modifying its external profile and eliminating the original surface of the back of its wooden support. Despite these uncertainties, it seems probable that the polyptych in question consisted of a series of rectangular panels topped by equilateral triangular gables—an archaic type of altarpiece, but one represented in Florence not only in works by Giotto but also in paintings by the circle of Pacino di Bonaguida and Jacopo del Casentino,[4]  [4]
Even if it has perhaps been cropped below by c. 2.5 cm, as Monika Cämmerer-George suggested, the panel in the National Gallery of Art has the squat proportions of paintings dating to the first two decades of the fourteenth century, in which the panels of the main tier, if of rectangular format, often are or were surmounted by triangular gables. See Monika Cämmerer-George, Die Rahmung der toskanischen Altarbilder im Trecento (Strasbourg, 1966), 71. Examples of this type, apart from the laterals in Chaalis (cf. note 9 below), probably included Giotto’s panel of Saint Anthony of Padua in the Berenson Library at Villa I Tatti, The Harvard University Center for Italian Renaissance Studies, Florence; cf. Franco Russoli and Nicky Mariano, La raccolta Berenson (Milan, 1962), no. 1. Though now shorn of gables, the panels attributed to Pacino di Bonaguida from the polyptych now divided between the Galleria dell’Accademia in Florence (no. 6146) and the Barnes Foundation in Philadelphia, or nos. 1961.1.2 and 1961.1.3 in the El Paso Museum of Art attributed to Jacopo del Casentino, also exemplify the same type of altarpiece. See, respectively, Richard Offner and Miklós Boskovits, A Critical and Historical Corpus of Flor­entine Painting: The Fourteenth Century, sec. 3, vol. 2, Elder Contemp­­oraries of Bernardo Daddi, new ed. (Florence, 1987), 155–161; Richard Offner, A Critical and Historical Corpus of Florentine Painting: The Fourteenth Century, sec. 3, vol. 7, The Biadaiolo Illuminator, Master of the Dominican Effigies (New York, 1957), 100–101. and in those realized in the circle of Duccio di Buoninsegna (Sienese, c. 1250/1255 - 1318/1319) in Siena.[5]  [5]
For example, cf. the polyptych no. 28 in the Pinacoteca Nazionale in Siena attributed to Duccio di Buoninsegna (Sienese, c. 1250/1255 - 1318/1319) himself, and polyptych no. 33 in the same gallery, attributed to the Master of Città di Castello (Italian, active c. 1290 - 1320); the polyptych by the Goodhart Duccesque Master in the church of Monterongriffoli; and that by Ugolino da Siena in the Clark Art Institute in Williams­town, Massachusetts. See James H. Stubblebine, Duccio di Buoninsegna and His School, 2 vols. (Princeton, 1979), 2: figs. 131, 188, 262, 450. According to Monika Cämmerer-­George, Die Rahmung der toskanischen Altarbilder im Trecento (Strasbourg, 1966), 71, the type of altarpiece in question is of Sienese origin; but clearly altarpieces of this type, even if they have not survived intact, must have been widespread both in Florence and in Arezzo, as suggested inter alia by the example of the early fourteenth-century polyptych now divided between the Palazzo Vecchio in Florence (Loeser bequest) and the Yale University Art Gallery, New Haven (nos. 1943.203, 1946.13). The resemblance of our painting to the image of Saint Stephen in the Museo Horne in Florence has long been noted:[6]  [6]
Richard Offner, “A Remarkable Exhibition of Italian Paintings,” The Arts 5 (1924): 244, was the first to note that “a St. Stephen in the Horne collection in Florence​ . . . ​happens to have originally stood to our Virgin’s right in the same original polyptych.” The hypothesis was immediately accepted by Mather and by Curt H. Weigelt in 1925 and by practically all the subsequent literature, at least until Dillian Gordon pointed out that “the Washington panel​ . . . ​has been gilded with orange bole” while “the St. Stephen​ . . . ​was definitely gilded with green [earth].” From this observation she deduced not that the panels belong to different altarpieces but instead that they probably were executed by different artists. This hypothesis is rather improbable, because while the realization of an altarpiece could have been entrusted to more than one painter, it is unlikely they would each have been left the freedom to choose their own techniques, including the kind of preparation to be used before applying the gilding. Cf. Frank Jewett Mather, “Two Attributions to Giotto,” Art Studies 3 (1925): 25–27; Curt H. Weigelt, Giotto: Des Meisters Gemälde (Stuttgart, 1925), 242; Dillian Gordon, “A Dossal by Giotto and His Workshop: Some Problems of Attribution, Provenance and Patronage,” The Burlington Magazine 131 (1989): 526. [Editor’s note: see also Cecelia Frosinini, “St Stephen" in Dominique Thiebaut, ed., Giotto e compagni (Paris, 2013), 139–140, for a discussion of the recent technical examination of the painting. This examination revealed that St Stephen was also gilded with a red bole on top of a green earth layer. Miklós Boskovits did not survive to learn of these new findings.] affinities between them are evident in style, in dimension, and in details of the ornamental repertoire.[7]  [7]
Both paintings are surrounded by ornamental borders with decorative motifs incised freehand into the gold ground. The ornamental borders are framed in both panels by a double row of dots impressed into the gold ground. The halos also have incised decoration freehand, hence without the use of punched motifs, and are surrounded both inside and outside by concentric circles formed of dots. The border of the child’s halo is further punctuated by three rosettes. However, the gold groundGround
 The layer or layers used to prepare the support to hold the paint. in the Saint Stephen was laid over a preparation of green underpaintUnderpainting
 An initial layer of paint applied to a ground that begins to define shapes and values. instead of the usual red bole,[8]  [8]
See note 6 above. Gilding laid on a green underpaint foundation has also been observed elsewhere in Giotto’s oeuvre, in particular in his stories of Christ, now divided among the Isabella Stewart Gardner Museum in Boston (Presentation of Jesus in the Temple), the Berenson Library at Villa I Tatti, The Harvard University Center for Renaissance Studies in Florence (Entombment), the National Gallery in London (Descent of the Holy Spirit), the Alte Pinakothek in Munich (Last Supper, Crucifixion, Descent to Limbo), and the Metropolitan Museum of Art in New York (Adoration of the Magi). See also Dillian Gordon, “A Dossal by Giotto and His Workshop: Some Problems of Attribution, Provenance and Patronage,” The Burlington Magazine 131 (1989): 524–531. [Editor’s note: see also Dominique Thiébaut, ed., Giotto e compagni (Paris, 2013), 130–151, for a discussion of the results of the cleaning of this painting, which Miklós Boskovits did not survive to examine. The National Gallery of Art has preserved his manuscript as he submitted it. In the case of these Giotto panels, it is important to point out that Boskovits wrote about them before technical examination in advance of exhibitions at the J. Paul Getty Museum and at the Louvre revealed new information about their construction.] making their common origin unlikely. Many art historians have also proposed that two panels depicting Saint John the Evangelist and Saint Lawrence combined with busts of angels in the upper register [fig. 2]   [fig. 2] Giotto, Saint John the Evangelist, c. 1310/1315, tempera on panel, Musée Jacquemart-André, Abbaye royale de Chaalis-Institute de France, Fontaine-Chaalis[fig. 3]   [fig. 3] Giotto, Saint Lawrence, c. 1310/1315, tempera on panel, Musée Jacquemart-André, Abbaye royale de Chaalis-Institute de France, Fontaine-Chaalis, in the Jacquemart-­André collection, Paris, and now displayed in the former abbey of Chaalis, outside Paris, belong to the same polyptych as our Madonna. This, in spite of some structural discrepancies between the Chaalis saints and the Washington Madonna, is possible.[9]  [9]
Roberto Longhi, “Progressi nella reintegrazione d’un polittico di Giotto,” Dedalo 11 (1930–1931): 285–291, was the first to propose the reconstruction of a dispersed polyptych with the Washington Madonna at its center, flanked to the left by the Saint John the Evangelist in Chaalis and the Saint Stephen in the Museo Horne, and to the right by the Saint Lawrence in Chaalis and a still unidentified panel. The hypothesis met with a generally favorable reception among art historians, though Monika Cämmerer-George (1966) contested it, noting incongruities between the presumed laterals of the altarpiece in terms of both the decorative borders around the outer edge of the gold ground and the respective size of the panels and the proportions of the saints represented in them. See Monika Cämmerer-­George, Die Rahmung der toskanischen Altarbilder im Trecento (Strasbourg, 1966), 71–74, 208 n. 233. Curiously, Francesca Flores d’Arcais, though considering Longhi’s reconstruction valid, also adduced some stylistic divergences between the panels in Chaalis and Florence, in Francesca Flores d’Arcais, Giotto (Milan, 1995), 320, 324. Only in recent years, however, were Cämmerer-­George’s doubts accepted by Giovanna Ragionieri, in Giovanni Previtali and Giovanna Ragionieri, Giotto e la sua bottega, ed. Alessandro Conti, 3rd ed. (Milan, 1993), 147 n. 217, 329, 348, 391; and Alessandro Tomei, “Giotto,” in Enciclopedia dell’arte medievale, 12 vols.  (Rome, 1995), 6:671. Miklós Boskovits (2000) and Angelo Tartuferi (2000) firmly ruled out the common origin of the four panels in Miklós Boskovits, “Giotto: Un artista poco conosciuto?” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 182–186; Angelo Tartuferi, in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 118. On the panels in Chaalis, cf. Nicolas Sainte Fare Garnot, “Brève histoire d’une redécouverte,” in Primitifs italiens du Musée Jacquemart-­André (Paris, 2000), 104–105, who still seemed to accept Longhi’s reconstruction.

The reception of the painting, of unknown Florentine provenance, was rather cold when it suddenly appeared on the horizon of art historical studies in 1917, at a time when research on Giotto’s style had led to an extreme limitation of works held to be autograph. Bernard BerensonBerenson, Bernard
 (June 26, 1865–October 6, 1959) Art historian and connoisseur. Son of a Lithuanian timber merchant who emigrated to the United States with his family in 1875, he was educated at the Latin School, Boston, and at Harvard University, where he studied Greek, Latin, Sanskrit, Hebrew, and German. In an unsuccessful application for a traveling fellowship to Europe, he wrote, ‘Art prevails in this programme because it is there that I feel myself weakest. One can study literature here . . . but art not at all.’ On his subsequent visit to Europe in 1885, financed by friends, his rapid visual self-education led to the decision to settle in Italy and to devote his life to the study of Italian art. —William Mostyn-Owen, Grove Art © Oxford University Press, evidently the first art historian to have had the opportunity to examine it, attributed it at first to Bernardo Daddi (Florentine, active by 1320, died probably 1348), immediately followed by Edward Robinson (1920) and Wilhelm R. Valentiner (1922, 1927).[10]  [10]
Bernard Berenson, “Italian Paintings,” Bulletin of the Metropolitan Museum of Art 15 (1920): 160–161; Edward Robinson, “The Fiftieth Anniversary Celebration,” Bulletin of the Metropolitan Museum of Art 15 (1920): 75; Wilhelm R. Valentiner, The Henry Goldman Collection (New York, 1922), n.p., no. 1; Wilhelm R. Valentiner, “The Henry Goldman Collection,” Art News (1927): 9–10. A more exhaustive analysis of the painting soon led Berenson to change his mind and to admit its close kinship to Giotto; he later assigned it to “one of the painters of the frescoes designed by Giotto in the lower church of San Francesco in Assisi.”[11]  [11]
See Provenance, note 2. Commenting on the work many years later, Bernard Berenson observed: “si cita ancora la mia attribuzione​ . . . ​al Daddi, sebbene io rammenti d’aver scritto al Dr. Valentiner molto tempo prima che pubblicasse il catalogo di quella collezione [i.e., that of Henry Goldman] per informarlo che io ritenevo opera di uno dei pittori che eseguirono gli affreschi disegnati da Giotto nella chiesa inferiore di San Francesco ad Assisi” (My attribution to Daddi is still cited, even though I recall having written to Dr. Valentiner long before he published the catalog of the Goldman collection to inform him that I considered [it] the work of one of the painters who executed the frescoes designed by Giotto in the lower church of San Franceso in Assisi). See Bernard Berenson, “Quadri senza casa: Il Trecento fiorentino, 1,” Dedalo 11 (1930–1931): 988 n. 1. In fact, Wilhelm R. Valentiner, though maintaining the attribution to Daddi, also noted that the panel “is so closely related to the work of Giotto that one who is not professional would immediately think of the master himself.” Wilhelm R. Valentiner, The Henry Goldman Collection (New York, 1922), n.p., Introduction. Like Berenson, Raimond van Marle (1924) and Richard Offner (1924) also referred the panel to an “assistant of Giotto,” identifying the hand of the same anonymous artist in other paintings now generally recognized as the work of Giotto himself.[12]  [12]
Raimond van Marle, The Development of the Italian Schools of Painting, vol. 3, The Florentine School of the 14th Century (The Hague, 1924), 185–190, tried to reconstruct the catalog of an “Assistant of Giotto’s” whom he dubbed the “Master of the Six Scenes from the Life of Christ.” This name referred to the series of panels cited in note 8 above (without the scene of the Descent of the Holy Spirit, now in the National Gallery in London). The Dutch scholar attributed both this group of paintings and the small versions of the Crucifixion in the Gemäldegalerie in Berlin and in the Museum of Fine Arts in Strasbourg to the same hand, but they are now generally recognized as autograph works by Giotto. Van Marle also added to the catalog of the “Master of the Six Scenes” the painted crucifix in the Florentine church of San Felice in Piazza, a panel still disputed between those who regard it as an autograph work by Giotto and those who assign it to a studio assistant. Richard Offner, “A Remarkable Exhibition of Italian Paintings,” The Arts 5 (1924): 244, considered both the Goldman Madonna and the Saint Stephen in the Museo Horne in Florence the work of a “nameless master” of “considerable gifts who must have worked very close to Giotto.” Similar opinions were then expressed by Curt H. Weigelt (1925), Wolfgang Fritz Volbach (1926), Valentiner (1926), Pietro Toesca (1929, 1933), Berenson (1930–1931, 1932, 1936), and Robert Oertel (1953, 1968).[13]  [13]
Curt H. Weigelt, Giotto: Des Meisters Gemälde (Stuttgart, 1925), 242; Wolfgang Fritz Volbach, in Henry Thode, Giotto, ed. Wolfgang Fritz Volbach, 3rd ed., Künstler-­Monographien (Bielefeld and Leipzig, 1926), 159 n. 62; Wilhelm R. Valentiner, A Catalogue of Early Italian Paintings Exhibited at the Duveen Galleries, New York, April to May 1924 (New York, 1926), n.p., Introduction; Pietro Toesca, La pittura fiorentina del Trecento (Verona, 1929), 60 n. 9; Pietro Toesca, “Giotto di Bondone,” in Enciclopedia italiana di scienze, lettere ed arti, ed. Istituto Giovanni Treccani, 36 vols. 1929–1939 (Milan, 1933), 17:218; Bernard Berenson, “Quadri senza casa: Il Trecento fiorentino, 1,” Dedalo 11 (1930–1931): 974, 988 n. 1; Bernard Berenson, Italian Pictures of the Renaissance: A List of the Principal Artists and Their Works with an Index of Places (Oxford, 1932), 236; Bernard Berenson, Pitture italiane del rinascimento: Catalogo dei principali artisti e delle loro opere con un indice dei luoghi, trans. Emilio Cecchi (Milan, 1936), 203; Robert Oertel, Die Frühzeit der italienischen Malerei (Stuttgart, 1953), 91; Robert Oertel, Early Italian Painting to 1400 (London, 1968), 98. Frank Jewett Mather Jr. (1925), however, began to speak of the panel now in Washington as a fully autograph Giotto, and this recognition, reaffirmed by such reputable scholars as Carlo Gamba (1930), Roberto Longhi (1930–1931), Lionello Venturi (1931, 1933), and Mario Salmi (1937), was generally accepted—with a few exceptions [14]  [14]
Once they came into possession of the painting after Goldman’s death, Duveen Brothers, Inc., commissioned expert opinions from the most reputable connoisseurs of the day. Since some of these expertises are dated 1938 (copies in NGA curatorial files), it must be assumed they were collected in anticipation of the painting’s sale to the Samuel H. Kress Foundation. Among the scholars approached by Duveen, Giuseppe Fiocco, Roberto Longhi, Wilhelm Suida, and Adolfo Venturi all unhesitatingly confirmed the attribution to Giotto, whereas Berenson and F. Mason Perkins attributed it respectively to a “close follower of Giotto” and “one of the closest and most gifted of Giotto’s anonymous pupils.” Not invited to express his opinion on this occasion, Richard Offner entrusted his view of the Goldman Madonna to a highly critical article devoted to the Italian paintings in the new National Gallery of Art in Washington, intended for the Art News but never in fact published; all that remains of it are the page proofs, copies of which are in the NGA curatorial files. “Thus the peculiarities of the National Gallery panel”—observes inter alia the great American connoisseur—“would commit it, like the altarpiece to which it belongs, to Giotto’s studio, where the living presence of the master communicated to this work a considerable proportion of its admirable qualities, but in which the actual execution was due to an assistant.” Offner’s judgment of our Madonna was, after all, highly positive; he combined it with a proposed dating “toward the end of the first quarter of the century,” close to that supported by Longhi and various other proponents of Giotto’s authorship. It is worth adding that the newspaper article by Carlyle Burrows cited in Provenance note 5 triumphantly reported the “final judgment” of Bernard Berenson, as pronounced viva voce by the “nonagenarian authority on Italian Renaissance art” in 1958: Berenson had then decided, reported Burrows, to accept the panel as a genuine “late Giotto.” He would memorialize this oral opinion in his posthumously published volume, Bernard Berenson, Italian Pictures of the Renaissance: Florentine School, 2 vols. (London, 1963), 2:81. The fact that the recent Giotto monograph by Michael Schwarz (2009) did not cite the Washington Madonna presumably means that he did not accept the painting as an autograph work by Giotto. Cf. Michael Viktor Schwarz, Giotto (Munich, 2009). —following the painting’s entry into the Gallery.[15]  [15]
Frank Jewett Mather, “Two Attributions to Giotto,” Art Studies 3 (1925): 25–27; Carlo Gamba, “Problemi artistici all’esposizione di Londra,” Il Marzocco 35, no. 17 (1930): 2; Roberto Longhi, “Progressi nella reintegrazione d’un polittico di Giotto,” Dedalo 11 (1930–1931): 289–291; Lionello Venturi, Pitture italiane in America (Milan, 1931), no. 26; Lionello Venturi, Italian Paintings in America, trans. Countess Vanden Heuvel and Charles Marriott, 3 vols. (New York and Milan, 1933), 1: no. 32; Mario Salmi, “La mostra giottesca,” Emporium 86 (1937): 358.

Opinions have differed in evaluating the date of the painting. A substantial body of the art historical literature considers the panel the work of Giotto’s later maturity and dates it to the years 1325–1330 (Mather 1925; Brandi 1938–1939; Carli 1951, 1955; Battisti 1960; Walker et al. 1961; Berenson 1963; Bologna 1969; Bellosi 1974, 1994; Laclotte 1978; Shapley 1979; Brandi 1983; Cavazzini 1996),[16]  [16]
Frank Jewett Mather, “Two Attributions to Giotto,” Art Studies 3 (1925): 25–27; Cesare Brandi, “Giotto,” Le Arti 1 (1938–1939): 126; Enzo Carli, Giotto (Florence, 1951), n.p; Enzo Carli, La pittura senese (Milan, 1955), 96; Eugenio Battisti, Giotto (Geneva, 1960), 115–117; John Walker, Guy Emerson, and Charles Seymour, Art Treasures for America: An Anthology of Paintings and Sculpture in the Samuel H. Kress Collection (London, 1961), 4–6; Bernard Berenson, Italian Pictures of the Renaissance: Florentine School, 2 vols. (London, 1963), 2:81; Ferdinando Bologna, Novità su Giotto: Giotto al tempo della cappella Peruzzi (Turin, 1969), 98; Luciano Bellosi, Buffalmacco e il Trionfo della morte (Turin, 1974), 52; Luciano Bellosi, “Ancora sulla cronologia degli affreschi del Cappellone di San Nicola a Tolentino,” in Arte e spiritualità nell’ordine agostiniano e il Convento San Nicola a Tolentino: Atti della seconda sessione del convegno “Arte e Spiritualità negli Ordini Mendicanti”; Tolentino, 1–4 settembre 1992, ed. Graziano Campisano (Rome, 1994), 189; Michel Laclotte, “Reconstruction and Revaluation,” Apollo 108 (1978): 387; Fern Rusk Shapley, Catalogue of the Italian Paintings, 2 vols. (Washington, DC, 1979), 1:219–221; Cesare Brandi, Giotto (Milan, 1983), 147; Laura Cavazzini, Giotto (Florence, 1996), 33. or, more cautiously, within the third decade of the century (Toesca 1933; NGA 1985; Lunghi 1986; Bonsanti 1992, 2000; Tomei 1995).[17]  [17]
Pietro Toesca, “Giotto di Bondone,” in Enciclopedia italiana di scienze, lettere ed arti, ed. Istituto Giovanni Treccani, 36 vols. (Milan, 1933), 17:218; National Gallery of Art, European Paintings: An Illustrated Catalogue (Washington, DC, 1985), 177; Elvio Lunghi, “Giotto,” in La Pittura in Italia: Il Duecento e il Trecento, ed. Enrico Castelnuovo, 2 vols. (Milan, 1986), 2:576; Giorgio Bonsanti, “La bottega di Giotto e la Croce di San Felice,” in La croce giottesca di San Felice in Piazza: Storia e restauro, ed. Magnolia Scudieri (Venice, 1992), 79, 89 n. 50; Giorgio Bonsanti, “La bottega di Giotto,” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 66, 71; Alessandro Tomei, “Giotto,” in Enciclopedia dell’arte medie­vale, 12 vols. (Rome, 1995), 6:671. The conviction that it should be dated to c. 1320 or even earlier is equally widespread (Longhi 1930–1931; Salmi 1937; Cecchi 1937; Duveen Pictures 1941; Frankfurter 1944; Florisoone 1950; Gamba 1961; Walker 1963; Rossi 1966; De Benedictis 1967; Dal Poggetto 1967; Previtali 1967, 1990; Venturoli 1969; Tartuferi 1987, 2000, 2007; Bandera Bistoletti 1989; Flores d’Arcais 1995; Boskovits 2000; Ragionieri 2002, 2007).[18]  [18]
Roberto Longhi, “Progressi nella reintegrazione d’un polittico di Giotto,” Dedalo 11 (1930–1931): 289–291; Mario Salmi, “La mostra giottesca,” Emporium 86 (1937): 358; Emilio Cecchi, Giotto (Milan, 1937), 115 – ​116; Duveen Brothers, Duveen Pictures in Public Collections of America (New York, 1941), no. 11; Alfred M. Frankfurter, The Kress Collection in the National Gallery (New York, 1944), 14; Michel Florisoone, Giotto (Paris, 1950), 117; Carlo Gamba, Il Museo Horne a Firenze: Catalogo con 40 illustrazione (Florence, 1961), 35; John Walker, National Gallery of Art, Washington, DC, (New York, 1963), 68–69; Filippo Rossi, ed., Il Museo Horne a Firenze, Gallerie e musei minori di Firenze (Florence, 1966), 134; Cri­stina De Benedictis, “Firenze: Omaggio a Giotto ” Antichità viva 6, no. 3 (1967): 34; Paolo Dal Poggetto, ed., Omaggio a Giotto: Celebrazioni nazionali nel vii centenario della nascita di Giotto (Florence, 1967), 11; Giovanni Previtali, Giotto e la sua bottega (Milan, 1967), 112–113, 376–377; Giovanni Previtali, “Giotto,” in Dizionario della pittura e dei pittori, ed. Enrico Castelnuovo and Bruno Toscano, 6 vols. (Turin, 1990), 2:604; Paolo Venturoli, “Giotto,” Storia dell’arte 1, no. 2 (1969): 154, 156, 158; Angelo Tartuferi, “Dipinti del Due e Trecento alla mostra ‘Capolavori e Restauri,’” Paragone 38, no. 445 (1987): 53, 59 n. 30; Angelo Tartuferi, “Giotto: Una nuova immagine,” in Giotto: Guida alla mostra, ed. Angelo Tartuferi (Florence, 2000), 21; Angelo Tartuferi, Giotto (Florence, 2007), 125, 127; Sandrina Bandera Bistoletti, Giotto: Catalogo completo dei dipinti (Florence, 1989), 126; Francesca Flores d’Arcais, Giotto (Milan, 1995), 320, 324; Miklós Boskovits, “Giotto di Bondone,” in Dizionario biografico degli italiani, 82 vols. (Rome, 2000), 55:412, 413, 418; Giovanna Ragionieri, “Giotto di Bondone,” in La pittura in Europa: Il dizionario dei pittori, ed. Carlo Pirovano, 3 vols. (Milan, 2002), 2:360; and Giovanna Ragionieri, in Luciano Bellosi and Giovanna Ragionieri, Giotto e la sua eredità: Filippo Rusuti, Pietro Cavallini, Duccio, Giovanni da Rimini, Neri da Rimini, Pietro da Rimini, Simone Martini, Pietro Lorenzetti, Ambrogio Lorenzetti, Matteo Giovannetti, Maso di Banco, Puccio Capanna, Taddeo Gaddi, Giovanni da Milano, Giottino, Giusto de’Menabuoi, Altichiero, Jacopo Avanzi, Jean Pucelle, i fratelli Limbourg (Florence, 2007), 303, 305. Supporters of such a dating are in essence those who have placed it between the frescoes in the Peruzzi and Bardi Chapels in Santa Croce in Florence, such as Federico Zeri (1957), Cesare Gnudi (1958), Roberto Longhi (1968), Lamberto Busignani (1993), and Julian Gardner (2002), but also those who have not accepted Giotto’s direct authorship and have limited themselves to pointing out the painting’s kinship with the Christological cycle in the lower church of San Francesco in Assisi, such as Berenson (1930–1931) and Tantillo Mignosi (1975).[19]  [19]
Federico Zeri, “Due appunti su Giotto,” Paragone 8, no. 85 (1957): 78; Cesare Gnudi, “Giotto,” in Enciclopedia Universale dell’Arte, ed. Istituto per la collaborazione culturale, 15 vols. (Florence, 1962), 6:236; Roberto Longhi, Edizione delle opere complete di Roberto Longhi, 14 vols. (Florence, 1961–1984), 4:6; Alberto Busignani, Giotto (Florence, 1993), 206, 307; Julian Gardner, “Giotto in America (and Elsewhere),” in Italian Panel Painting of the Duecento and Trecento, ed. Victor M. Schmidt, Studies in the History of Art (Washington, DC, and New Haven, 2002), 176; Bernard Berenson, “Quadri senza casa: Il Trecento fiorentino, 1,” Dedalo 11 (1930–1931): 974, 988 n. 1; Almamaria Tantillo Mignosi, “Osserva­zioni sul transetto della Basilica Inferiore di Assisi,” Bollettino d’arte 60 (1975): 136, 141 n. 38.

Various hypotheses have also been advanced about the presumed destination of the polyptych of which the Washington Madonna formed part. The proposal, first formulated by Mather (1925), that it was one of the four polyptychs that Ghiberti mentioned in Santa Croce, met with wide support.[20]  [20]
Frank Jewett Mather, “Two Attributions to Giotto,” Art Studies 3 (1925): 25–27. See Julius von Schlosser, ed., Lorenzo Ghibertis Denkwürdigkeiten: I Commentarii, 2 vols. (Berlin, 1912), 1:36: “Nell’ordine de’ frati minori [in Florence] quattro cappelle et quattro tauole” (In the order of friars minor of Florence [i.e., Santa Croce] four chapels and four altarpieces). Vasari, ever since the first edition (1550) of his Vite, specified that the chapels in question were that with the stories of Saint Francis (the Bardi Chapel), that of the Peruzzi, that of the Giugni, and a fourth to the left of the sanctuary (the family chapel of the Tosinghi and Spinelli). Giorgio Vasari, Le vite de’ più eccellenti architetti, pittori, et scultori italiani, da Cimabue insino a’ tempi nostri nell’edizione per i tipi di Lorenzo Torrentino, Firenze 1550, ed. Luciano Bellosi and Aldo Rossi (Turin, 1986), 119. Later, observing that one of the panels in Chaalis represented Saint Lawrence and that in the Museo Horne Saint Stephen, Gnudi (1959), convinced like many others of the common origin of these panels with the Goldman Madonna, cautiously suggested as its original site the chapel dedicated to these saints in Santa Croce — ​​that is, the chapel of the Pulci and Berardi families.[21]  [21]
On this chapel, the penultimate to the left of the cappella maggiore, see Walter Paatz and Elisabeth Paatz, Die Kirchen von Florenz: Ein kunst­geschichtliches Handbuch, 6 vols. (Frankfurt am Main, 1940), 1:574; Richard Offner, Miklós Boskovits, and Enrica Neri Lusanna, A Critical and Historical Corpus of Florentine Painting: The Fourteenth Century, sec. 3, vol. 3, The Works of Bernardo Daddi, new ed. (Florence, 1989), 122–145. The frescoes by Bernardo Daddi, with stories of Saints Lawrence and Stephen, probably date to the mid-third decade or slightly before. However, since they are superimposed over earlier frescoes (or possibly merely geometrical decoration), the chapel could have been officiated several years before Daddi’s intervention and perhaps even furnished with an altarpiece. Although many found the proposal convincing, Gardner (1999, 2002)[22]  [22]
Julian Gardner, “Panel Paintings Attributed to Giotto in American Collections,” Center / National Gallery of Art, Center for Advanced Study in the Visual Arts 19 (1999): 71; Julian Gardner, “Giotto in America (and Elsewhere),” in Italian Panel Painting of the Duecento and Trecento, ed. Victor M. Schmidt, Studies in the History of Art (Washington, DC, and New Haven, 2002), 176. placed it in doubt: according to this scholar, the original altar blocks that survive in the family chapels in the east transept of Santa Croce were too small ever to have supported an altarpiece some three meters in length, as the polyptych in question must have been.[23]  [23]
Comparison with Giotto’s surviving polyptychs that include half figures of saints (the ones in the Badia, today in the Uffizi, Florence, and Santa Maria del Fiore [the Duomo]) shows a ratio between height and width of c. 2.5:1. If we assume that the panel in the National Gallery of Art originally had a size and shape similar to the panels of the two saints now in Chaalis, it must have been topped by an equilateral triangular gable and had an overall height of approximately 125 cm. We can then calculate the total width of the five-part polyptych as c. 312 cm. On the structure and proportions of Giotto’s polyptychs in general, see Monika Cämmerer-­George, Die Rahmung der toskanischen Altarbilder im Trecento (Strasbourg, 1966), 50–85.

An alternative hypothesis, formulated by Venturi (1931, 1933), identified the panel now in Washington and its presumed companions as components of the lost polyptych in the church of the Badia in Florence.[24]  [24]
Lionello Venturi, Pitture italiane in America (Milan, 1931), no. 26; Lionello Venturi, Italian Paintings in America, trans. Countess Vanden Heuvel and Charles Marriott, 3 vols. (New York and Milan, 1933), 1; no. 32. Ugo Procacci (1962), however, refuted the proposal and succeeded in identifying the former Badia altarpiece with the still intact polyptych that entered the Museo di Santa Croce in the nineteenth century.[25]  [25]
Ugo Procacci, “La tavola di Giotto dell’altar maggiore della Chiesa della Badia fiorentina,” in Scritti di storia dell’arte in onore di Mario Salmi, ed. Filippa Aliberti, 3 vols. (Rome, 1962), 2:9–45. The polyptych has been on display in the Uffizi, Florence, since 1957. Carlo Ludovico Ragghianti (1949) advanced yet another proposal, according to which the panels now divided among the museums of Chaalis, Florence, and Washington were components of the polyptych painted by Giotto for the church of San Francesco in Borgo Sansepolcro,[26]  [26]
Carlo Ludovico Ragghianti, in Giorgio Vasari, Le vite dei più eccellenti pittori, scultori ed architetti, ed. Carlo Ludovico Ragghianti (Milan, 1949), 4:282 n. 69. The “tavola di man di Giotto, di figure piccole, che poi se n’è ita in pezzi” (painting by the hand of Giotto, with small figures, which was then broken into pieces), putatively rediscovered in Borgo Sansepolcro, was reported by Vasari in the second edition (1568) of Giorgio Vasari, Le vite de’ più eccellenti pittori, scultori e architettori nelle redazioni del 1550 e 1568, ed. Rosanna Bettarini and Paola Barocchi, 6 vols. (Florence, 1967), 2:113. The tendency now is to accept the hypothesis of Martin Davies, who identified the fragments of this polyptych with the stories of Christ divided among the museums of Boston, London, Munich, New York, and the Berenson Library at Villa I Tatti, The Harvard University Center for Italian Renaissance Studies, Florence. See Martin Davies, The Earlier Italian Schools, National Gallery Catalogues (London, 1951), 181; and note 8 above. but this hypothesis has been found unconvincing. 

If the problem of the destination of the polyptych remains unresolved, by observing the characteristics of polyptychs by Giotto himself and by other Florentine painters (for example, that by Jacopo del Casentino, now in New Orleans)[27]  [27]
Isaac Delgado Museum of Art, New Orleans, no. 61.60. The polyptych given to Jacopo del Casentino is one of the few Florentine polyptychs of the period to have survived intact; see Richard Offner and Miklós Boskovits, A Critical and Historical Corpus of Florentine Painting: The Fourteenth Century, sec. 3, vol. 4, Bernardo Daddi, His Shop and Following, new ed. (Florence, 1991), 179–189. datable within the first quarter of the fourteenth century, we can conclude that the altarpiece of which the Goldman Madonna formed the center is unlikely to have been very different in appearance from that conjectured by Longhi (1930–1931)[fig. 4]   [fig. 4] Reconstruction of the Badia polyptych by Giotto as proposed by Roberto Longhi, from left to right: Saint Stephen, Museo Horne, Florence; Saint John the Evangelist (fig. 2); Madonna and Child, NGA; Saint Lawrence (fig. 3). By flanking the panel in the Gallery with panels of saints of slightly smaller size, we would obtain an ensemble similar in dimensions to those of the former Badia polyptych (which we know was intended for the high altar of that church). It would have considerably exceeded in width those polyptychs executed by Giotto for side chapels or for long-established altars in older churches.[28]  [28]
The Badia polyptych, painted by Giotto for the early medieval church that had been undergoing enlargement and reconstruction from 1284 onward, measures 91 × 340 cm. The polyptych painted by Simone Martini (Sienese, active from 1315; died 1344) in 1319 for the high altar of the Dominican church of Santa Caterina in Pisa measures 195 × 340 cm; that by Pietro Lorenzetti (Sienese, active 1306 - 1345), commissioned in 1320 for the main altar of the pieve of Santa Maria in Arezzo, 298 × 309 cm; and the panel by Meo da Siena, dated 1333 and destined for the high altar of San Pietro in Perugia, now in the Städelsches Kunstinstitut at Frankfurt, 59.5 × 304 cm. Smaller polyptychs were made for altars in side chapels or for the high altars in early churches, where the altar tables were generally of reduced size. We may cite as an example Giotto’s Stefan­eschi altarpiece in Old Saint Peter’s in Rome, whose original dimensions, despite its prestigious destination, were just over two and a half meters in width. On the problem, cf. Julian Gardner, “Fronts and Backs: Setting and Structure,” in La pittura nel XIV e XV secolo, il contributo dell’analisi tecnica alla storia dell’arte, ed. Hendrik W. van Os and J. R. J. van Asperen de Boer (Bologna, 1983), 297–322; Julian Gardner, “Giotto in America (and Elsewhere),” in Italian Panel Painting of the Duecento and Trecento, ed. Victor M. Schmidt, Studies in the History of Art (Washington, DC, and New Haven, 2002), 160–181. It would be futile, based on current knowledge, to go any further in the field of conjecture; it will suffice to point out that among the churches in which panels by Giotto are mentioned by the earlier sources, the most probable provenances are likely to be the Florentine churches of Santa Croce and Ognissanti.[29]  [29]
Peter Murray’s useful compilation (1959), complemented now by the work of Michael Viktor Schwarz and Pia Theis (2004), listed, apart from the Badia polyptych, four panels by Giotto in the church of Santa Croce, one in San Giorgio alla Costa (the fragmentary Maestà now in the Museo Diocesano of Florence), a crucifix and a now lost image of Saint Louis of Toulouse formerly in Santa Maria Novella, and a crucifix and four panels in Ognissanti. See Peter Murray, An Index of Attributions Made in Tuscan Sources before Vasari (Florence, 1959), 79–89; and Michael Viktor Schwarz and Pia Theis, Giottus pictor, vol. 1, Giottos Leben (Vienna, 2004), 285–303. The tabula altaris in the chapel of the Palazzo del Bargello is perhaps cited by error in a manuscript of Filippo Villani’s De origine. Other manuscripts of Villani’s book only report the presence of Giotto’s frescoes in this chapel; cf. Schwarz and Theis 2004, 290.  Of the four panels cited in Santa Croce, that of the Baroncelli Chapel is still in situ. There are no firm reasons for assuming that the polyptych now in the North Carolina Museum of Art in Raleigh has a provenance from the family chapels of the east side of the transept of this church, though its relatively small size (105.7 × 250 cm) and the presence of Francis of Assisi among the saints represented (Santa Croce being a Franciscan church) make this a possibility. The hypothesis of Wilhelm Suida (1931), who considered this polyptych as intended for the Peruzzi Chapel, is now generally accepted. See Wilhelm Suida, “A Giotto Altarpiece,” The Burlington Magazine for Connoisseurs 59 (1931): 188–193. The proposed identification of one of Giotto’s Santa Croce panels with that formerly in the Bromley Davenport collection at Capesthorne Hall (Macclesfield) is more uncertain. This altarpiece, perhaps a product of Giotto’s atelier and indeed intended for the church of Santa Croce, was certainly not painted by the hand of the master; the proposed attribution to the young Taddeo Gaddi seems very convincing. Its original destination might have been the Lupicini Chapel or, less probable, the Bardi Saint Francis Chapel, in either of which, given its relatively small size, it could have been easily accommodated on the altar. Cf. Mina Gregori, “Sul polittico Bromley Davenport di Taddeo Gaddi e sulla sua originaria collocazione,” Paragone 25, no. 297 (1974): 73–83; and Federico Zeri, “Italian Primitives at Mssrs. Wildenstein,” The Burlington Magazine 107 (1965): 252–256. We may wonder whether so acute an observer as Ghiberti could ever have classified it as an autograph work by the master himself. The panels of saints dispersed between the Horne and Jacquemart-­André museums remain to be considered. That they come from the chapels in the east transept of Santa Croce seems improbable due to their measurements; see Julian Gardner, “Giotto in America (and Elsewhere),” in Italian Panel Painting of the Duecento and Trecento, ed. Victor M. Schmidt, Studies in the History of Art (Washington, DC, and New Haven, 2002), 181 n. 80. But the existence of a Giotto polyptych would also be conceivable in the chapel of Saint Louis, on the north side of the transept: this chapel, dating to the years 1332–1335, could have accommodated an altarpiece of a width exceeding 3 meters; cf. Irene Hueck, “Stifter und Patronatsrecht: Dokumente zu zwei Kapellen der Bardi,” Mitteilungen des Kunsthistorischen Institutes in Florenz 20 (1976): 263–270. It cannot be excluded, lastly, that one of Giotto’s four polyptychs might have been executed for the high altar of the old church of Santa Croce, in use at least until 1314 and perhaps until 1320/1330 and doubtless not left without appropriate decoration. See Walter Paatz and Elisabeth Paatz, Die Kirchen von Florenz: Ein kunstgeschichtliches Handbuch, 6 vols. (Frankfurt am Main, 1940), 1:500. As for the church of Ognissanti, one of the panels with an undoubted provenance from the church is the Maestà now in the Uffizi, Florence, often—though without foundation—considered an altarpiece for the high altar or displayed on its rood-screen; cf. Irene Hueck, “Le opere di Giotto per la chiesa di Ognissanti,” in La ‘Madonna d’Ognissanti’ di Giotto restaurata (Florence, 1992), 37–49. Whatever the case, the fact remains that the Maestà comes from Ognissanti, as does the Dormitio Virginis now in the Gemäldegalerie in Berlin (no. 1884) and the great painted crucifix still in situ in the church that Ghiberti cited in addition to the four panels. On the crucifix, see Marco Ciatti, ed., L’officina di Giotto: Il restauro della Croce di Ognissanti, Problemi di conservazione e restauro (Florence, 2010). Two other panels therefore remain to be identified. One might have been the “mezza Nostra Donna col fanciullo in braccio” that Ghiberti saw above the side door of the church leading into the cloister. It is possible, however, that this Madonna and Child was not a panel painting but a frescoed lunette. Julius von Schlosser, ed., Lorenzo Ghibertis Denkwürdigkeiten, 2 vols. (Berlin, 1912), 1:36.

With regard to the chronological position of the panel being discussed here, some features, such as the exclusive use of decorations incised freehand (hence the absence of punched motifs), seem to offer firm clues.[30]  [30]
Erling Skaug’s conclusion that “in Florence extensive punch work was introduced only about 1333–1334” was perhaps excessively cautious, since it was based only on securely dated paintings. Even so, Skaug admitted the occasional use of punches by Giotto before his terminus post quem, for example in the Stefaneschi altarpiece now in the Pinacoteca Vaticana (on which see the following note) and in that of the Baroncelli Chapel in Santa Croce. See Erling S. Skaug, Punch Marks from Giotto to Fra Angelico: Attribution, Chronology, and Workshop Relationships in Tuscan Panel Painting with Particular Consideration to Florence, c. 1330–1430, 2 vols. (Oslo, 1994), 1:34–36. This latter chapel was founded, according to the inscription on its external wall, in 1328; see Andrew Ladis, Taddeo Gaddi: Critical Reappraisal and Catalogue Raisonné (Columbia, MO, 1982), 88. In any case, the fact that the gold ground of the Goldman Madonna is totally devoid of punched decoration might be an indication that it predates both the Roman and Florentine polyptychs. In the Stefaneschi altarpiece, which ought to date to the early 1320s,[31]  [31]
The Stefaneschi altarpiece has lost its original frame, which probably bore the signature of Giotto and the date of execution. The fourteenth-century Liber benefactorum in the Basilica Vaticana did not hesitate to identify Giotto as the artist of the polyptych, while the manuscript of the canon Giacomo Grimaldi, dating to the early seventeenth century, said it was “circa annum MCCCXX depicta.” Cf. Wolfgang Fritz Volbach, Catalogo della Pinacoteca Vaticana, vol. 1, I dipinti dal X secolo fino a Giotto (Vatican City, 1979), 45–51. There are no good reasons to doubt the reliability of this affirmation, probably based on an inscription partially legible on the original frame. In addition, a punch mark is used in the Ognissanti crucifix, which is probably earlier than the Stefaneschi altarpiece. See Miklós Boskovits, “Il Crocifisso di Giotto della chiesa di Ognissanti: Riflessioni dopo il restauro,” in L’Officina di Giotto: Il restauro della Croce di Ognissanti, ed. Marco Ciatti (Florence, 2010), 47–62. the artist used at least one punched motif, and this type of decoration is increasingly found in his later paintings. The characteristics of the motifs incised freehand in the halos and the pseudo-­Kufic lettering in the broad ornamental border that runs around the margins of the panel [fig. 5]   [fig. 5] Graphic tracing of the halos and pseudo-Kufic lettering, Giotto, Madonna and Child, c. 1310/1315, tempera on poplar, National Gallery of Art, Washington, Samuel H. Kress Collection. (Joanna Dunn, National Gallery of Art, Washington) seem recurrent in the works by Giotto since the end of the thirteenth century [fig. 6]   [fig. 6] Graphic tracing of the decoration of the halo of Christ, Giotto, Painted Crucifix, San Felice in Piazza, Florence, but perhaps it is not by chance that the practice of surrounding the halos with a double row of dots appears no earlier than the Maestà in the Uffizi, Florence.[32]  [32]
Giotto had already used Kufic-type lettering in the halos of the mourners in the Santa Maria Novella crucifix, a painting generally recognized as one of the master’s earliest works, datable to c. 1290. Motifs similar to those of Mary’s halo in the Washington panel—interlaced motifs that are repeated in equal form from right to left, with the exception of two different decorative elements incised in the gold ground above the Madonna’s right shoulder—are matched in Christ’s halo in the Giottesque crucifix in San Felice in Piazza, Florence (figs. 5 and 6). Tracing based on Magnolia Scudieri, ed., Lacroce giottesca di San Felice in Piazza: Storia e restauro (Venice, 1992), 146. Invariably discussed as a product of Giotto’s shop and thought to be painted mainly by the master’s assistants, this work was dated to the early years of the fourteenth century by Giovanni Previtali, Giotto e la sua bottega (Milan, 1967), 137–138; Carlo Volpe, “Sulla Croce di San Felice in Piazza e la cronologia dei crocefissi giotteschi,” in Giotto e il suo tempo: Atti del congresso internazionale per la celebrazione del vii centenario della nascità di Giotto, Assisi, Padova, Firenze (Rome, 1971), 253–263; and Giorgio Bonsanti, “La bottega di Giotto e la Croce di San Felice,” in La croce giottesca di San Felice in Piazza: Storia e restauro, ed. Magnolia Scudieri (Venice, 1992), 53–90; and toward 1310 or shortly after, i.e., the time of the Peruzzi Chapel, by Luciano Bellosi, Giotto (Florence, 1981), 69; and Miklós Boskovits, “Una vetrata e un frammento d’affresco di Giotto nel museo di Santa Croce,” in Scritti di storia dell’arte in onore di Federico Zeri, 2 vols. (Milan, 1984), 1:44, the latter of whom considered the crucifix an autograph painting by Giotto. Similar ornamental motifs also appear in the frescoes of the Saint Nicholas chapel in the lower church of San Francesco in Assisi, painted, according to the more recent studies, no later than c. 1300; cf. Brigitte Klesse, Seidenstoffe in der italienischen Malerei des 14. Jahrhunderts (Bern, 1967), 192, no. 53 of the repertoire. For the dating of the Uffizi Maestà to the end of the first decade of the fourteenth century, cf. Giorgio Bonsanti, “La bottega di Giotto,” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 66; and Miklós Boskovits, “Giotto: Un artista poco conosciuto?” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 85. Adriano Pero­ni, “La Maestà di Ognissanti rivisitata dopo il restauro,” in La Madonna d’Ognissanti di Giotto restaurata (Florence, 1992), 30, argued, however, that the painting should date to the early years of the century. To these elements, which seem to indicate a relatively early date in Giotto’s career—within the second decade of the fourteenth century—we can perhaps add an observation regarding the red coif that covers Mary’s head and can be glimpsed, on either side of her face, below the pseudo-­Kufic hem of the mantle that covers her head. This is an archaic, byzantinizing motif that would disappear from Giotto’s authenticated works in the course of the 1320s.[33]  [33]
In medieval Byzantine devotional images of the Madonna and Child—and hence also in Italian ones—Mary’s hair is in general hidden within a kind of coif. The color of this article of clothing may vary, but at least in central Italy within the last decades of the thirteenth and first decades of the fourteenth century, it is usually red, as in the Washington Madonna. The Virgin is represented in this way in Giotto’s pre-1300 Maestà now in the Museo Diocesano in Florence, in the fresco on the inner façade of the upper church of the basilica in Assisi, and also in the Uffizi Maestà and in the Stefaneschi altarpiece. In Giotto’s later works, such as the polyptych in the Pinacoteca Nazionale in Bologna; the Madonna of Santa Maria a Ricorboli, also in Florence; and the polyptych in the Baroncelli Chapel in Santa Croce (c. 1328), the mother of Christ instead wears a wimple and a transparent veil that covers her head but leaves her blond hair visible below it. A relatively early date might also be suggested by the neckline of the Virgin’s dress. Though this is admittedly an unreliable clue, it concurs with other features in suggesting a date for the Washington painting no later than the early 1320s.[34]  [34]
Bellosi, who was the first to use the width of the scooped neckline as an indication of dating, compared the neckline of the Washington Madonna with those painted by Pietro Lorenzetti in 1320 in the polyptych of the Pieve of Santa Maria in Arezzo, in Luciano Bellosi, “Moda e cronologia: B) per la pittura di primo Trecento,” Prospettiva 11 (1977): 12–14. Of course, that does not imply a dating ad annum; the fact remains that we find increasingly wider necklines than that of the Goldman Madonna in Giotto’s last decade of activity.

A stylistic reading of the painting seems to confirm the chronological position suggested by the abovementioned data. Its morphological features connect the Goldman Madonna with the central phase of Giotto’s career, what might be defined as his “Peruzzi phase.” Unfortunately, the frescoes of the Peruzzi Chapel (in Santa Croce), much admired by the sources and by artists in the past, are now reduced to almost total illegibility by the radical abrasionAbrasion
 A gradual loss of material on the surface. It can be caused by rubbing, wearing, or scraping against itself or another material. It may be a deteriorative process that occurs over time as a result of weathering or handling or it may be due to a deliberate attempt to smooth the material. to which they have been subjected. Other paintings have survived from the same phase, in which the artist appears no longer satisfied with the serene classicism of his Paduan paintings. Solemnity and monumentality were no longer enough: a more circumstantial, naturalistic description of the events, and a deeper participation of the protagonists in them, were needed. While further accentuating the physical stature and presence of his figures, Giotto now strove to underline their active involvement in the emotional climate of the scenes. These were also the years of the cycle of frescoes in the Magdalene chapel in the lower church of San Francesco at Assisi (for which there are good reasons for dating it to 1308).[35]  [35]
This is the date of the Cappella della Maddalena frescoes proposed by Giovanni Previtali, Giotto e la sua bottega (Milan, 1967), 90–94; Lucia­no Bellosi, Giotto (Florence, 1981), 59; Cesare Brandi, Giotto (Milan, 1983), 115–118; Filippo Todini, “Pittura del Duecento e del Trecento in Umbria e il cantiere di Assisi,” in La Pittura in Italia: Il Duecento e il Trecento, ed. Enrico Castelnuovo, 2 vols. (Milan, 1986), 2:395; Giorgio Bonsanti, “La bottega di Giotto,” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 69; Miklós Boskovits, “Giotto di Bondone,” in Dizionario biografico degli italiani, 82 vols. (Rome, 2000), 55:411. The Assisi document also seems to confirm this date, asserting that the painter Palmerino di Guido repaid a credit on Giotto’s behalf in January 1309; cf. Valentino Martinelli, “Un documento per Giotto ad Assisi,” Storia dell’arte 19 (1973): 193–208. Evidently, the painter had been present and was working in the Umbrian city until shortly before this date. Francesca Flores d’Arcais, Giotto (Milan, 1995), 272, proposed a later date (1315–1318) for the Magdalene chapel cycle. The polyptych in the North Carolina Museum of Art in Raleigh can be assumed to be only one or two years later: it is stylistically close to the mural cycle in Assisi.[36]  [36]
Cf. note 29 above. It cannot be excluded that the Raleigh polyptych belonged to the Peruzzi Chapel. However, the altarpiece is probably earlier than supposed in the past: Miklós Boskovits argued for a dating of 1310 in Miklós Boskovits, “Il Crocifisso di Giotto della chiesa di Ognissanti: Riflessioni dopo il restauro,” in L’Officina di Giotto: Il restauro della Croce di Ognissanti, ed. Marco Ciatti (Florence, 2010), 52, 61 n. 27. Apart from the semidestroyed frescoes of the Peruzzi Chapel,[37]  [37]
Most art historians of the period accepted the dating of the Peruzzi Chapel cycle to 1310–1315. Cf. Giovanni Previtali, Giotto e la sua bottega (Milan, 1967), 105–107; Luciano Bellosi, Giotto (Florence, 1981), 62; Cesare Brandi, Giotto (Milan, 1983), 185; Francesca Flores d’Arcais, Giotto (Milan, 1995), 252–261; Giorgio Bonsanti, “La bottega di Giotto,” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 66; Miklós Boskovits, “Giotto: Un artista poco conosciuto?” in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 86–87. Hayden B. J. Maginnis, Painting in the Age of Giotto: A Historical Reevaluation (University Park, PA, 1997), 131, seemed to prefer a slightly later dating. the surviving stained-glass windows of the last bay of Santa Croce before the transept (now replaced by copies: the original windows are housed in the Museo dell’Opera)[38]  [38]
See Miklós Boskovits, “Una vetrata e un frammento d’affresco di Giotto nel museo di Santa Croce,” in Scritti di storia dell’arte in onore di Federico Zeri, 2 vols. (Milan, 1984), 1:39–45; Miklós Boskovits, in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 138–140. must also date to the early 1310s. The painted crucifix in the Florentine church of the Ognissanti, together with the Dormitio Virginis painted for the same church (now in the Gemäldegalerie in Berlin), also should date to this phase.[39]  [39]
The Dormitio Virginis in the Gemäldegalerie in Berlin, no. 884, is generally dated to c. 1310. See Giovanni Previtali, Giotto e la sua bottega (Milan, 1967), 109–112; Luciano Bellosi, Giotto (Florence, 1981), 59; Miklós Boskovits, ed., Frühe italienische Malerei: Gemäldegalerie Berlin, Katalog der Gemälde, trans. Erich Schleier (Berlin, 1988), 56–61; Francesca Flores d’Arcais, Giotto (Milan, 1995), 237–242. The crucifix in Ognissanti, although commissioned for the same church, is generally thought to be slightly later: according to Previtali 1967, 137–138, in the third decade of the century; according to Bellosi 1981, 69, “di un’epoca assai posteriore” (of considerably later period) than the Peruzzi Chapel frescoes. Miklós Boskovits, “Il Crocifisso di Giotto della chiesa di Ognissanti: Riflessioni dopo il restauro,” in L’Officina di Giotto: Il restauro della Croce di Ognissanti, ed. Marco Ciatti (Florence, 2010), 58, considered the painting executed by the mid-1310s, like Flores d’Arcais (1995), 243–244, who proposed the period 1310–1315, and Giorgio Bonsanti, in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 147–150, who preferred the date c. 1315.

Between 1311 and 1315, Giotto seems to have been engaged in paintings especially for his hometown.[40]  [40]
Cf. Michael Viktor Schwarz and Pia Theis, Giottus pictor, vol. 1, Giottos Leben (Vienna, 2004), 103–112. In the years that followed, Giotto’s painting is dominated by a gothicizingGothic
 Term used to denote, since the 15th century, the architecture and, from the 19th century onward, all the visual arts of Europe during a period extending by convention from about 1120 to c. 1400 in central Italy, and until the late 15th century and even well into the 16th century in northern Europe and the Iberian Peninsula. The early gothic style overlapped chronologically with Romanesque and flourished after the onset of Renaissance art in Italy and elsewhere. The term gothic is applied to western European painting of the 13th century to the early 15th century. Unlike gothic architecture, it is distinguished more by developments in style and function than in technique, and even in these areas there is considerable national and regional diversity. The applicability of the term to Italian painting is debated, as is its usefulness in accounting for developments in Netherlandish painting from the early 15th century. Contact with Byzantine art was close in the early 13th century, but after c. 1250 survived principally in the Holy Roman Empire and Italy. —Peter Kidson, Grove Art © Oxford University Press tendency, for example in the murals in the transept of the lower church of Assisi, where we may note a greater pursuit of realistic effects and a greater attention to the elegance of pose and expressiveness of gesture of his figures. So novel are these developments that the paintings he executed during these years have often been denied the status of autograph works; instead, art historians have postulated the leading presence at the master’s side of an assistant, the so-called “Parente di Giotto.”[41]  [41]
On the much-debated problem of the so-called “Parente di [i.e., relative of] Giotto,” a painter to whom—according to some art historians—the great master delegated much of the actual execution of very important commissions during the second and third decade of the fourteenth century, cf. a summary of the contributions of Giorgio Bonsanti and Miklós Boskovits, in Giotto: Bilancio critico di sessant’anni di studi e ricerche, ed. Angelo Tartuferi (Florence, 2000), 55–73, 75–94. The features of the works belonging to this phase are in any case different from those expressed in the Washington Madonna. The painting being discussed here finds its most convincing parallels in passages of the frescoes in the Magdalene chapel, for example with the bust of the titular saint in the vault, and also in the image of the Virgin in the Raleigh polyptych.[42]  [42]
In their recent contribution, Maria Clelia Galassi and Elizabeth Walmsley (2009) observed analogies between the painting technique of the Washington Madonna and the panel in the Museo Horne in Florence and differences between the execution of these latter and the Ognissanti Madonna and the Raleigh polyptych. This may be explained, in my opinion, by their different dates of execution. See Maria Clelia Galassi and Elizabeth Walmsley, “Painting Technique in the Late Works of Giotto,” in The Quest for the Original: Underdrawing and Technology in Painting; Symposium 16, Bruges, 21 – ​23 September 2006 [Colloque pour l’Étude du Dessin Sous-­Jacent et de la Technologie dans la Peinture], ed. Hélène Verougstraete (Leuven, 2009), 116–122. In the Goldman Madonna, however, the more slender proportions of the protagonists, the more fluent calligraphy of the contours, and the greater complexity of the draperies, in comparison with the abovementioned painting, reveal a more advanced stage in Giotto’s development. The conduct of the protagonists of our painting seems to confirm this: the child is presented by the artist no longer as the infant pantocrator of the Maestà in the Uffizi but with a more human touch, with charm and infantile immediacy. The sternness is mitigated: the child’s little face is softened, its tenderness heightened by the smile that seems to play on his lips. His gestures—right hand stretched out to touch the rose in his mother’s hand, left hand to grasp her forefinger—are intimate and playful. As already in the great Maestà, the lips of the child and those of his mother are parted, not for any solemn declaration but to engage in an affectionate and intimate conversation. 

Whatever our painting’s destination, it ought to date to the years around 1310 to 1315, when the two panels of Chaalis (components, perhaps, of the same polyptych) also saw the light of day.

Miklós Boskovits (1935–2011)

March 21, 2016