Saint Andrew and Saint Benedict with the Archangel Gabriel [left panel], shortly before 1387

This panel is part of a triptychTriptych
 A picture consisting of three parts. The term denotes both the object itself and its compositional form. As an object, the triptych may vary in size and material, but usually consists of a central panel flanked by wings (or shutters), which may be hinged; as a compositional form it is a tripartite structure, often with an emphasized central element. Although its imagery was, until the 19th century at least, predominantly religious, the object as such was not tied to a specific function. —Victor M. Schmidt, Grove Art © Oxford University Press that consists of two laterals with paired saints (this panel and Saint Bernard and Saint Catherine of Alexandria with the Virgin of the Annunciation [right panel]) and a central panel with the Madonna and Child (Madonna and Child Enthroned with Twelve Angels, and with the Blessing Christ [middle panel]). All three panels are topped with similar triangular gables with a painted medallion in the center. The reduction of a five-part 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 into a simplified format with the external profile of a triptych may have been suggested to Florentine masters as a consequence of trends that appeared towards the end of the fourteenth century: a greater simplification in composition and a revival of elements of painting from the first half of the Trecento.[1]  [1]
A similar solution—a five-part polyptych reduced to triptych format—in fact appears on the main side of Giotto’s Stefaneschi altarpiece (Pinacoteca Vaticana, no. 40.120) and later (probably at a date close to 1340) in the fragment of a triptych by Jacopo del Casentino now in a private collection; see 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), pl. xlix. About 1365, Matteo di Pacino revived this scheme in an altarpiece of similar structure now in the Gal­leria dell’Accademia in Florence (no. 8463); see Michela Palmeri, in Di­pinti, vol. 1, Dal Duecento a Giovanni da Milano, Cataloghi della Galleria dell’Accademia di Firenze, ed. Miklós Boskovits and Angelo Tartuferi (Florence, 2003), 174–181. Perhaps a few years later in date is another triptych of similar format, attributed to the Master of San Lu­cchese near Poggibonsi, destroyed in 1944; reproduced in Bernard Berenson, “Quadri senza casa: Il Trecento fiorentino, 2,” Dedalo 11 (1930 – ​1931): 1050, as a work by Jacopo di Cione. Another altarpiece of the same format is the triptych painted by the Master of the Misericordia and Niccolò Gerini in collaboration, now in the church of Sant’Andrea at Montespertoli near Florence; reproduced in Richard Offner and Klara Steinweg, A Critical and Historical Corpus of Florentine Painting: The Fourteenth Century, sec. 4, vol. 5, Giovanni del Biondo, pt. 2 (New York, 1969), pl. xliv, as workshop of Giovanni del Biondo, probably executed around 1380.  Agnolo Gaddi followed this trend in several of his works. He demonstrates this in the three panels being discussed here by his deliberate revival of motifs that had been abandoned by most Florentine painters since the mid-fourteenth century. To present the Madonna seated on a throne of Giottesque type,[2]  [2]
The throne is similar to the one that appears in the polyptych signed by Giotto now in the Museo Civico Nazionale in Bologna (no. 284). Thrones of this type, with a high, triangular-topped backrest but of simple structure and convincingly drawn in perspective, appear in the early 1360s in Giovanni da Milano’s polyptych now in the Museo Civico at Prato; in the fragmentary polyptych by Cenni di Francesco, dated 1370, in the church of San Cristofano a Perticaia near Florence; in the polyptych by Pietro Nelli and Niccolò Gerini in the pieve at Impruneta, dated 1375; and thereafter ever more frequently in the last quarter of the century. Cf. Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370 – ​1400 (Florence, 1975), pls. 86 and 61.  instead of concealing the structure of the throne with a gold-embroidered cloth of honor as in most paintings realized by masters in the circle of OrcagnaAndrea di Cione
 (Orcagna; Orgagna; Arcagnuolo) (born 1315–20; died Florence, 1368) Painter, sculptor, and architect, thought to have also been active as a poet. He was trained as a painter and referred to himself as “pictor” on the tabernacle in Orsanmichele. Details of his training are not known, but his first surviving works reveal various influences, especially of Maso di Banco and Taddeo Gaddi. —G. Kreytenberg, Grove Art © Oxford University Press, was a sort of archaism at this time. Agnolo scrupulously describes this seat and at the same time exploits its form to create three-dimensional effects. Yet these archaizing motifs are combined with more forward-looking features. Gaddi’s progressive adjustment to the innovative late-­GothicGothic
 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 taste of his time is thus attested by various aspects of the triptych, such as the pastiglia decoration in the gables, the now lost decoration of the frame,[3]  [3]
On the original decoration of the frame, cf. Mary B. Bustin, “Recalling the Past: Evidence for the Original Construction of Madonna Enthroned with Saints and Angels by Agnolo Gaddi,” Studies in the History of Art: Conservation Research 57 (1996–1997): 50–64. The fashion of covering the gable zones of altarpieces with pastiglia ornamental motifs in relief before they were gilded began to spread from Orca­gna’s shop in the second half of the Trecento.  the rich orientalizing carpet that covers the floor,[4]  [4]
The sumptuous decoration, with pairs of facing animals, of the brocaded fabrics used to cover the throne of the Madonna or the floor on which the saints stand is a phenomenon characteristic of Florentine painting in the second half of the fourteenth century, especially in paintings of the circle of Orcagna, but also in panels produced in the bottega of Agnolo Gaddi. Cf. Brigitte Klesse, Seidenstoffe in der italienischen Malerei des 14. Jahrhunderts (Bern, 1967), 283, 316, 341.  and even the crowded composition of the central panel.

Lionello Venturi published this triptych in 1931 under the name of Gherardo Starnina.[5]  [5]
Lionello Venturi, Pitture italiane in America (Milan, 1931), no. 52.  This attribution was based on the now discarded theory of scholars who had tried in the first three decades of the century to reconstruct the oeuvre of a putative disciple of Agnolo Gaddi, to whom the conventional name “Madonnenmeister” or “Compagno d’ Agnolo” was given and who was later identified with Gherardo Starnina.[6]  [6]
Oskar Wulff (1907) coined this unfortunate conventional name. Osvald Sirén (1914–1915) later altered it to “Compagno d’Agnolo,” and shortly thereafter he formulated a hypothesis identifying this artist with Gherardo Starnina (1916, 51–53). See Oskar Wulff, “Der Madonnenmeister: Ein sienesisch – ​florentinische Trecentist,” Zeitschrift für Kunstgeschichte 20 (1907): 195–210, 227–236; Osvald Sirén, “Early Italian Pictures, the University Museum, Göttingen (Conclusion),” The Burlington Magazine for Connoisseurs 26 (1914): 107–114; Osvald Sirén, ed., A Descriptive Catalogue of the Pictures in the Jarves Collection Belonging to Yale University (New Haven and London, 1916), 51–53. The proposal was accepted by Raimond van Marle, The Development of the Italian Schools of Painting, vol. 3, The Florentine School of the 14th Century (The Hague, 1924), 565–573, and some other authors in the following decade.  However, 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 recognized that a good part of the work given to the “Compagno d’Agnolo” belongs to Agnolo himself. The studies of Ugo Procacci [7]  [7]
Referring to the group of paintings usually cited under the name “Compagno d’Agnolo,” Bernard Berenson wrote: “Una delle più singolari aberrazioni della critica recente è stata quella di attribuire tutte queste Madonne allo Starnina; ma non è necessario perdere il tempo a dissipare errori che il tempo stesso disperderà” (One of the most singular aberrations of modern criticism is that of attributing all these Madonnas to Starnina; but it is not necessary to waste time dissipating errors that time itself will dissipate). See Bernard Berenson, “Quadri senza casa: Il Trecento fiorentino, 3,” Dedalo 11 (1930–1931): 1303. After the discovery of the remains of the cycle of Starnina’s documented frescoes in the church of the Carmine in Florence, the hypothesis of the anonymous master’s identification with Starnina was gradually abandoned. See Ugo Procacci, “Gherardo Starnina,” Rivista d’arte 15 (1933): 151–190; Ugo Procacci, “Gherardo Starnina,” Rivista d’arte 17 (1935): 331–384.  on Starnina finally put the proposed identification to rest, though it continued to enjoy residual credit for some time to come.[8]  [8]
The triptych now in the National Gallery of Art evidently came to Duveen Brothers accompanied by expertises from Robert Langton Douglas and Osvald Sirén, both supporting the attribution to Starnina; see Duveen Brothers, Duveen Pictures in Public Collections of America (New York, 1941), nos. 24–25. Lionello Venturi (1931, 1933), Wilhelm R. Valentiner (1933), and Ugo Galetti and Ettore Camesasca (1951) also accepted that attribution. Roberto Salvini (1935–1936), not accepting the identification Gherardo Starnina = Compagno d’Agnolo, attributed the painting to the latter. Cf. Lionello Venturi, Pitture italiane in America (Milan, 1931), no. 52; Lionello Venturi, Italian Paintings in America, trans. Countess Vanden Heuvel and Charles Marriott, 3 vols. (New York and Milan, 1933), 1: no. 63; Wilhelm R. Valentiner, ed., The Sixteenth Loan Exhibition of Old Masters: Italian Paintings of the XIV to XVI Century (Detroit, 1933), no. 11; Ugo Galetti and Ettore Camesasca, Enciclopedia della pittura italiana, 3 vols. (Milan, 1951), 3:2351; and Roberto Salvini, “Per la cronologia e per il catalogo di un discepolo di Agnolo Gaddi,” Bollettino d’arte 29 (1935–1936): 287.  In the National Gallery of Art, the altarpiece was cataloged as a work by Agnolo Gaddi, and since the 1960s, art historians have unanimously accepted this attribution. 

The present writer proposed that Agnolo Gaddi’s altarpiece might have been executed for the sacristy of the church of San Miniato al Monte (Florence),[9]  [9]
Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370–1400 (Florence, 1975), 118–121.  of which the Alberti were patrons and for whose decoration Benedetto di Nerozzo Alberti left funds in his will of 1387.[10]  [10]
On Benedetto Alberti’s will, see Luigi Passerini, ed., Gli Alberti di Firenze: Genealogia, storia e documenti, 2 vols. (Florence, 1869), 2:187. The codicil in question was attached to the will drawn up in 1377 (now lost and known only from a seventeenth-century abstract). In it the testator already instructed that “si facessi dipignere la Sagrestia di San Miniato al Monte con gli Armadi et Finestra et af[freschi]” (that the sacristy of San Miniato al Monte should be painted [and provided] with cupboards, [stained glass] window, and frescoes); Giovanni Felice Berti, Cenni storico-artistici per servire di guida ed illustrazione alla insigne Basilica di S. Minato al Monte e di alcuni dintorni presso Firenze (Florence, 1850), 156. So, when Alberti ten years later made testamentary provision that the “sacrestia ecclesiae sancti Miniatis ad Montem de prope Florentiam compleatur et compleri et perfici debeant picturis, armariis, coro, fenestra vitrea, altari et aliis necessariis” (the sacristy of San Miniato al Monte of Florence should be completed and perfected with paintings, cupboards, a choir stall, stained glass window, an altar, and all other necessary things), this decoration might already have been planned and perhaps even in part realized. Ada Labriola and Federica Baldini accepted the provenance of the Gallery's panels from the sacristy of San Miniato in Ada Labriola, “La decorazione pittorica,” in L’Oratorio di Santa Caterina: Osservazioni storico-critiche in occasione del restauro, ed. Maurizio De Vita (Florence, 1998), 52; Federica Baldini, in L’Oratorio di Santa Caterina all’Antella e i suoi pittori, ed. Angelo Tartuferi (Florence, 2009), 159.  The reasons adduced at that time in support of such a hypothesis were, it must be admitted, not quite convincing: referring to the inscription in Saint Benedict’s book to the “admonition” (an administrative sanction by the Florentine government) against Alberti in 1387 and his subsequent exile is open to question. Furthermore, I erroneously asserted that Saint Giovanni Gualberto was represented in the painting. The saint to the right of the Virgin is, in fact, Bernard of Clairvaux, but the presence of this saint in the altarpiece is actually a further argument in support of a provenance from the sacristy of San Miniato. Saint Bernard was the patron Saint of Benedetto Alberti’s son Bernardo, who in his will dated 1389 left money for masses to be celebrated annually pro anima dicti testatoris (for the soul of the said testator) in the family chapel in San Miniato, which had evidently already by that date been consecrated.[11]  [11]
See Stefan Weppelmann, Spinello Aretino und die toskanische Malerei des 14. Jahrhunderts (Florence, 2003), 381.  The representation of Saint Andrew, who was the patron saint of a predeceased son of Benedetto Alberti, also links the altarpiece to the sacristy of San Miniato. As to the fourth saint, Catherine of Alexandria (standing on a broken wheel), she was evidently much venerated in Benedetto’s family. This is proved by the fact that, in his will of 1387, he bequeathed money for the decoration of an oratory near Florence (Santa Caterina dell’Antella) dedicated to the martyr saint of Alexandria (and decorated by a cycle of frescoes illustrating scenes from her life by Spinello Aretino); additionally, his son Bernardo wished to build a monastery and a church in her honor.[12]  [12]
See Stefan Weppelmann, Spinello Aretino und die toskanische Malerei des 14. Jahrhunderts (Florence, 2003), 381.  The Alberti family’s veneration of Saint Catherine may have been based on the popular etymology of her name (catherine = catenula) diffused by Jacopo da Varazze in Legenda aurea,[13]  [13]
Jacopo da Varazze, Legenda aurea, ed. Giovanni Paolo Maggioni, 2 vols. (Florence, 1998), 2:1205.  with reference to the chain represented in the Alberti coats of arms.

Although the provenance from San Miniato remains a hypothesis, it still seems to me a quite plausible one that, if correct, would give us the certainty that by 1830 the triptych was still on the altar of the sacristy. An altarpiece can apparently be seen still in situ in a sketch of the sacristy’s altar wall [fig. 1]   [fig. 1] C. A. R. Roller, ""Design of the east wall of the sacristy of San Miniato in Florence,"" from Tagebuch einer italienischen Reise (Journal of My Trip to Italy in the Years 1829 and 1830), 1:7, June 1830, Rittersaalverein Castle Museum, Burgdorf, Switzerland made in that year by architect Christoph Robert August Roller (1805–1858), in his Tagebuch einer italienischen Reise (Castle Museum Burgdorf, Burgdorf, Switzerland). Unfortunately, the sketch, to which Stefan Weppelmann kindly drew my attention, is very small and certainly not sufficient for the identification of the triptych in the Gallery. What may be said for certain is only that an altarpiece composed of five panels stood on the altar of the sacristy of San Miniato in 1830, but by 1836 this altarpiece was no longer there, as Stefan Weppelmann rightly observed. It was removed and sold presumably by the Pia Opera degli Esercizi Spirituali, which had owned the furniture and decorations of the church since 1820.[14]  [14]
See “Regesto dell’Abbazia fiorentina di San Miniato,” La Graticola 4 (1976): 117–135. See Stefan Weppelmann, Spinello Aretino und die toskanische Malerei des 14. Jahrhunderts (Florence, 2003), 381.

As for its date, the Gallery’s first catalog (1941) cautiously suggested “the last quarter of the XIVth century,” while the volume devoted to the Duveen Pictures (1941) proposed an approximate date of c. 1380.[15]  [15]
National Gallery of Art, Preliminary Catalogue of Paintings and Sculpture (Washington, DC, 1941), 69; Duveen Brothers, Duveen Pictures in Public Collections of America (New York, 1941), nos. 24–25.  More recent publications in general support a time frame within the 1380s, although without explaining the reasons for this proposal. Arguing for a provenance from the sacristy of the Florentine church of San Miniato al Monte, Miklós Boskovits (1975) attempted a more precise dating shortly after the codicil dated 1387 was appended to the testament of Benedetto di Nerozzo degli Alberti, its putative patron.[16]  [16]
Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370–1400 (Florence, 1975), 118.  For his part, Bruce Cole (1979) stylistically linked the Gallery triptych with the cycle of frescoes in the choir of Santa Croce in Florence, for which he proposed a date of execution in the years c. 1388–1393.[17]  [17]
Bruce Cole, Agnolo Gaddi (Oxford, 1977), 21–26.  Given the lack of securely datable panels by Gaddi, with the exception of the composite altarpiece of the Cappella del Crocifisso, still in the church of San Miniato,[18]  [18]
Even the dating of this work, for the most part identified with the altarpiece for which Agnolo was paid between 1394 and 1396, has been questioned. What is certain is that payments were made to Agnolo during those years for “la tavola di San Miniato a Monte.” In 1396, however, because the artist had died in the meantime, his brother Zanobi received the balance; see Bruce Cole, Agnolo Gaddi (Oxford, 1977), 65, 67. No doubt correctly, art historians generally have assumed that the documents refer to the altarpiece placed on the altar of the crucifix of Saint Giovanni Gualberto in San Miniato, which has come down to us with its components rearranged. Originally, the panels of the polyptych representing the stories of Christ and full-length figures of Saint Giovanni Gualberto and Saint Minias formed an ensemble that must have contained at the center the much-venerated crucifix, which, according to legend, had spoken to Saint Giovanni Gualberto. When the crucifix was transferred to the Vallombrosans of Santa Trinita in Florence in 1671, the painted panels were rearranged in such a way as to fill the gap created by its removal. Though Cole (1977, 51–56) contested the identification of the existing altarpiece of the crucifix of Saint Giovanni Gualberto with that cited in the documents, his opinion met with little support. Cole argued instead that the documented painting should be identified with the triptych of the Contini-­Bonacossi bequest now in the Uffizi, Florence. But this latter painting has existed in its present form only since the 1930s, when the dispersed panels of two different altarpieces were arbitrarily cobbled together during an unscrupulous restoration; cf. Bruce Cole, Agnolo Gaddi (Oxford, 1977), 51–56; Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370 – ​1400 (Florence, 1975), 66, 216 n. 85, 298; Gaudenz Freuler, ed., Manifestatori delle cose miracolose: Arte italiana del ’300 e ’400 da collezioni in Svizzera e nel Liechtenstein (Einsiedeln, 1991), 204; Christoph Merzenich, Vom Schreinerwerk zum Gemälde: Florentiner Altarwerke der ersten Hälfte des Quattrocento (Berlin, 2001), 228. Freuler, however, believed that at least the two laterals of the Contini-­Bonacossi polyptych formed part of the altarpiece documented in 1394–1396.  various scholars have attempted to construct a chronology for the artist based on an analysis of the punched decoration of his work; however, this effort has failed to yield any precise indication for the Washington altarpiece other than a vague association with a relatively late phase in the painter’s activity.[19]  [19]
Among the punch marks used by Agnolo, Erling Skaug (1994) especially observed two that can be identified in Taddeo Gaddi’s triptych dated 1334, now in the Staatlichen Museen of Berlin. Skaug inserted the National Gallery of Art painting in the late phase of Agnolo Gaddi. Both the investigations of Mojmir S. Frinta (1998) and the later analysis of Skaug himself (2004) largely concurred, however, in suggesting that the very same punches were used throughout the artist’s career. 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), 1:260–264; 2: punch chart 8.2; Mojmir Svatopluk Frinta, Punched Decoration on Late Medieval Panel and Miniature Painting (Prague, 1998), 100, 244, 283, 290, 322, 481; Erling S. Skaug, “Towards a Reconstruction of the Santa Maria degli Angeli Altarpiece of 1388: Agnolo Gaddi and Lorenzo Monaco?” Mitteilungen des Kunsthistorischen Institutes in Florenz 48 (2004): 245, 254–255. Some of the punch marks found in the Washington triptych also appear in the later San Miniato altarpiece, documented in 1394–1396 (on which see note 18 above), while others had already been used in Coronation of the Virgin in the London National Gallery, generally considered an early work by the artist.

In the course of his career, especially between the 1380s and the early 1390s, Agnolo Gaddi produced a number of polyptychs, now in part dismantled and dispersed, of which at least the surviving central panels propose a composition close to that of the triptych discussed here. I refer in particular to Madonna and Child with Eight Angels (now united with laterals that did not originally belong to it) in the Contini-­Bonacossi bequest to the Uffizi, Florence;[20]  [20]
Osvald Sirén, “Addenda und Errata in meinem Giottino-­Buch,” Monatshefte für Kunstwissenschaft 1 (1908): 1122–1123, published the central panel with an attribution to Agnolo Gaddi and stated its present whereabouts as the Masi collection in Capannoli (Pisa). It is very probable, therefore, that the painting has a provenance either from Pisa or from Capannoli itself, a spa (at present a resort) in the district of Pisa that used to belong to Piero Gambacorti, the governor of Pisa, who owned a castle there; see Emanuele Repetti, Dizionario geografico, fisico, storico della Toscana, contenente la descrizione di tutti i luoghi del Granducato, Ducato di Lucca, Gar­fagnana e Lunigiana, 11 vols. (Florence, 1833–1849), 1:452. The provenance of the laterals now attached to this painting is unknown; all we know is that in 1952 they were the property of Alessandro Contini-­Bonacossi, who, according to Cole, had purchased them in Rome prior to 1931. Cf. George Kaftal, Saints in Italian Art, vol. 1, Iconography of the Saints in Tuscan Painting (Florence, 1952), 747–748; Bruce Cole, Agnolo Gaddi (Oxford, 1977), 76. Contini then commissioned special frames to be made for them, similar to the original frame of the Madonna from Capannoli. Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370–1400 (Florence, 1975), 298, dated the latter to c. 1380–1385, while Cole, as we have seen, considered the recomposed triptych to be genuine and associated it with the two documents of 1394–1396. Caterina Caneva, in Gli Uffizi: Cata­logo generale, 2nd ed. (Florence, 1980), 277, dated the execution of the whole altarpiece to 1375–1380; Ada Labriola, “Gaddi, Agnolo,” in Dizionario biografico degli italiani, 82 vols. (Rome, 1998), 51:146, to the 1380s; while Sonia Chiodo, “Gaddi, Agnolo,” in Allgemeines Künstlerlexikon: Die bildenden Künstler aller Zeiten und Völker, ed. Günter Meißner, 87 vols. (Munich, 2005), 47:113, mentioned it as a “late work.”  Madonna and Child Surrounded by Eight Angels in the church of San Lorenzo at Borgo San Lorenzo;[21]  [21]
Boskovits’s proposal (1975, 117, 296) to attribute the Borgo San Lorenzo Madonna to Agnolo and to date it to 1380–1385 has in substance been accepted by the more recent literature; cf. Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370 – ​1400 (Florence, 1975), 117, 296; 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), 1:261; Maria Matilde Simari, “Il Mugello,” in Il Mugello, la Valdisieve e la Romagna fiorentina, ed. Cristina Acidini and Anna Benvenuti Papi (Milan, 2000), 56. A modern (nineteenth- or twentieth-century) copy of this painting, attributed to Agnolo himself, was with Wildenstein & Co. in New York in 1954; see Dorothy C. Shorr, The Christ Child in Devotional Images in Italy during the XIV Century (New York, 1954), 114.  the triptych in the Staatliche Museen of Berlin, in which six angels are placed around the throne and a further pair are in the gable;[22]  [22]
The predella of the painting, now in the Louvre, Paris, has been recognized as that formerly in the Nobili chapel at Santa Maria degli Angeli and dated to 1387–1388 on the basis of information derived from the sources by Hans Dietrich Gronau, “The Earliest Works of Lorenzo Monaco, 2,” The Burlington Magazine 92, no. 569 (1950): 217–222. Fe­de­rico Zeri cautiously conjectured that the predella belonged to the triptych now in the Gemäldegalerie in Berlin, and Bruce Cole more firmly accepted this proposal. See Federico Zeri, “Investigations into the Early Period of Lorenzo Monaco, 1,” The Burlington Magazine 106 (1964): 554–558; Bruce Cole, Agnolo Gaddi (Oxford, 1977), 75, 84–87. For a recent résumé of the problems relating to a reconstruction of the altarpiece, see Erling S. Skaug, “Note sulle decorazione a punzone e i dipinti su tavola di Lorenzo Monaco,” in Lorenzo Monaco: Dalla tradizione giottesca al Rinascimento, ed. Angelo Tartuferi and Daniela Parenti (Florence, 2006), 106–110; see also note 23 below.  and Madonna and Child Flanked by Twelve Angels, now in a private collection in Milan.[23]  [23]
The painting, which measures 150 × 87 cm, was illustrated in Servizio per le ricerche delle opere rubate, Bollettino 17 (1994): 60, published by the special police unit of the Carabinieri devoted to the recovery of stolen works of art, formerly with an attribution to Agnolo Gaddi. Later, the reported theft of the painting was shown to be mistaken, and the work was republished by Gaudenz Freuler, “Gli inizi di Lorenzo Monaco miniatore,” in Lorenzo Monaco: Dalla tradizione giottesca al Rinascimento, ed. Angelo Tartuferi and Daniela Parenti (Florence, 2006), 80, with the same attribution. The variety of the angels’ poses, the softness of the modeling, and the motif of angels supporting the crown over Mary’s head (which recurs in paintings of Gaddi’s late phase) in any case suggest a late dating, probably in the last decade of the century.  None of these is securely dated, but the Berlin triptych can in all probability be identified with that formerly on the altar of the Nobili chapel in Santa Maria degli Angeli, which bore the inscription “An D 1387 Bernardus Cini de Nobilibus fecit fieri hanc cappellam.”[24]  [24]
Gemäldegalerie, Berlin, no. 1039; see Bruce Cole, Agnolo Gaddi (Oxford, 1977), 76, 84–87. Attributed to the school of Agnolo Gaddi by Osvald Sirén, Don Lorenzo Monaco (Strasbourg, 1905), 41, it was reinstated as an autograph work by Gaddi himself by Bernard Berenson, Italian Pictures of the Renaissance: A List of the Principal Artists and their Works with an Index of Places (Oxford, 1932), 213. For the lost inscriptions, see Dillian Gordon, The Fifteenth Century Italian Paintings, National Gallery Catalogues (London, 2003), 197 n. 3.  This gives us a useful point of reference not only for defining a chronological sequence of the paintings in question but also, as we shall see, for the dating of Gaddi’s great cycles of Florentine frescoes. Another chronological point of reference, albeit an approximate one, is 1383, the date of the testament of Michele di Vanni Castellani, in which he made bequests for the construction and decoration of a family chapel in Santa Croce, the chapel that would later be frescoed by Agnolo Gaddi.[25]  [25]
See Bruce Cole, Agnolo Gaddi (Oxford, 1977), 78–79, and, concerning the will of Michele Castellani, 61, doc. 7. Attributed by Vasari and by much of the later art-historical literature to Gherardo Starnina, the mural paintings in the chapel were attributed by Berenson to Agnolo Gaddi, at least “in great part.” See Bernard Berenson, Italian Pictures of the Renaissance: A List of the Principal Artists and Their Works with an Index of Places (Oxford, 1932), 213. More recent studies have in general accepted this view, though some art historians believe that they detect the hand of various assistants of Agnolo in the cycle. Cole (1977, 14–15) thought that “three masters” worked in the chapel with some degree of autonomy: the anonymous master of the stories of Saints Nicholas and Anthony, the equally anonymous master of the scenes from the life of the Baptist, and Agnolo Gaddi himself, who, Cole argued, executed part of the lunette of Zacharias and the stories of Saint John the Evangelist.  The style of this decoration suggests a period of execution not much later than the will; indeed, most art historians tend to place the execution of the cycle in the years immediately following 1383.

Although successive restorations have now made it difficult to assess, the Borgo San Lorenzo panel [26]  [26]
The Borgo San Lorenzo panel was extensively retouched in 1864 and then restored c. 1920; see Francesco Niccolai, Mugello e Val di Sieve: Guida topografica storico-artistica illustrata (Borgo S. Lorenzo, 1914), 430. Since then it has been subjected to at least two other restorations. The alterations in the appearance of the painting following these treatments are documented in photographs nos. 886, 68589, and 93950 of the Soprintendenza in Florence.  seems the earliest of the group. It was perhaps painted even before the frescoes in the Castellani chapel, with which it has affinities in its use of dense shadows in modeling, in the rigid profiles of the angels, and in the deeply channeled and brittle-looking folds of their garments. Between that work and the Nobili triptych now in Berlin can be placed both the Madonna of the Contini-Bonacossi bequest (it too now altered by retouches) and the triptych in the Gallery. In contrast to these latter two, the animated composition of the panel destined for the Nobili chapel seems to represent a further step forward, in the direction of the more dynamic compositions and the more delicate modeling that characterize the painter’s final phase, to which the above-mentioned Madonna surrounded by twelve angels now in a private collection can, I believe, be ascribed.[27]  [27]
The close resemblance between the passage with the blessing Christ child seated on his mother’s lap and the fresco by Agnolo Gaddi now in the Museo di Pittura Murale in Prato suggests that the two paintings are close in date, presumably contemporary with the painter’s documented activity in Prato in the years 1391–1394. For the fresco in Prato, see Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370–1400 (Florence, 1975), fig. 260; for the documents in question, see Giuseppe Poggi, “Appunti d’archivio: La Cappella del Sacro Cingolo nel Duomo di Prato e gli affreschi di Agnolo Gaddi,” Rivista d’arte 14 (1932): 355–376; Bruce Cole, Agnolo Gaddi (Oxford, 1977), 63–65.

If such a chronological sequence of the altarpieces executed by Agnolo in the 1380s is plausible, the Gallery triptych ought to date to a period slightly preceding 1387—that is, slightly preceding the execution of the other and more important enterprise promoted by Benedetto di Nerozzo Alberti, the frescoing of the choir in Santa Croce.[28]  [28]
We have no secure documentary evidence for the dating of the frescoes in question. Roberto Salvini, L’arte di Agnolo Gaddi (Florence, 1936), 31–85, considered them to predate the decoration of the Castellani chapel, but the more recent art historical literature in general indicates 1387 as the terminus ante quem for the execution of the cycle; Sonia Chiodo, “Gaddi, Agnolo,” in Allgemeines Künstlerlexikon: Die bildenden Künstler aller Zeiten und Völker, ed. Günter Meißner, 87 vols. (Munich, 2005), 47:113. The fact is that while Benedetto degli Alberti in his testament dictated in 1377 indicated Santa Croce (where his family had the patronage of the cappella maggiore) as the place where he wished to be buried, he made no mention of the realization of the frescoes in this chapel in the codicil added to his will ten years later, when he made testamentary provision for the funding of other artistic enterprises; for the text of the codicil, which also cites the relevant passage from the testament of 1377, see Luigi Passerini, ed., Gli Alberti di Firenze: Genealogia, storia e documenti, 2 vols. (Florence, 1869), 2:186–194. It seems logical to infer from this evidence that at the time the codicil was added the mural decoration of the chapel had already been finished and that everything was ready for Benedetto’s burial. In my view, however, the stylistic evidence suggests a later, or more protracted, date for the very demanding enterprise of frescoing the chapel, which could have begun c. 1385 but could well have been prolonged for years due to the political setbacks that struck the family. Accordingly, the present writer suggests the date 1385–1390; Miklós Boskovits, Pittura fiorentina alla vigilia del Rinascimento, 1370–1400 (Florence, 1975), 297; and Bruce Cole, Agnolo Gaddi (Oxford, 1977), 21–27, c. 1388–1393. In dating the frescoes in the choir, various stylistic features should be borne in mind: the more complex and crowded compositions; the more elaborate language of gesture; the numerous genre details; and the tendency towards a softening of the forms, modeled with light tonal passages of chiaroscuro. These features all suggest that the cycle of the cappella maggiore in Santa Croce is stylistically more advanced than the decoration of the Castellani chapel (executed, as we have seen, sometime after 1383), but no doubt antecedent to the frescoes in the Cappella della Cintola in Prato Cathedral, for which the painter received payments in 1392–1394. See Giuseppe Poggi, “Appunti d’archivio: La Cappella del Sacro Cingolo nel Duomo di Prato e gli affreschi di Agnolo Gaddi,” Rivista d’arte 14 (1932): 363–369.  Various similarities can be identified between passages of that cycle and the Washington triptych, in confirmation of the chronological proximity of the two works: the bust of Saint Andrew [fig. 2]   [fig. 2] Detail of Saint Andrew (left), Agnolo Gaddi, Madonna and Child with Saints Andrew, Benedict, Bernard, and Catherine of Alexandria with Angels, shortly before 1387, tempera on poplar, National Gallery of Art, Washington, Samuel H. Kress Collection recurs, in similar form, in the scene of the Making of the Cross [fig. 3]   [fig. 3] Detail of spectators, Agnolo Gaddi, The Making of the Cross, 1385–1390, fresco, Cappella Maggiore, Santa Croce, Florence. Image: Scala/Art Resource, NY, in the group of spectators to the extreme right of the fresco, while analogies can also be identified between the other saints of the triptych and the busts of the prophets inserted in the ornamental friezes that articulate the chapel’s decoration. Close similarities have also been observed between the lateral saints of the Gallery triptych and the fragments of an altarpiece now in Indianapolis.[29]  [29]
Clowes collection, Indianapolis Museum of Art: Ian Fraser et al., A Catalogue of the Clowes Collection, Indianapolis Museum of Art Bulletin (Indianapolis, 1973), 6. Bruce Cole, Agnolo Gaddi (Oxford, 1977), 27, rightly observed the “very close stylistic and positional relationship with the Washington saints,” deducing from this affinity that “both works were in the artist’s shop at the same time.” Perhaps it would be permissible to speak of the reuse of the same model employed for the Washington triptych in another similar and slightly later triptych, of which the panels now in Indianapolis formed part. I suspect, however, that the Indianapolis panels belonged to the triptych of the Nobili chapel in Santa Maria degli Angeli, the Florentine church of the Camaldolese order (see notes 21 and 23 above). In fact, the lost framing pilasters of the triptych now in Berlin could have contained, to judge from their measurements, two of the Indianapolis panels superimposed on each side. It is also worth pointing out that the paintings in the American museum come — ​​like the Berlin triptych — ​​from the Solly collection, and that the white habit with which Saint Benedict and the other monk immersed in reading are portrayed would have been very suitable to adorn a chapel in a church belonging to the Camaldolese order (a reformed branch of the Benedictines). Laurence Kanter informs me that he is about to publish further evidence for identifying the Indianapolis saints and their counterparts at the University of Gottingen as the front and lateral faces of the framing pilasters of the Nobili altarpiece, together with the missing pilaster base from the altarpiece predella.  We have no secure evidence to help us date these fragments, probably the remains of the decoration of the lateral pilasters of a polyptych roughly contemporary with, or perhaps slightly later than, the triptych being discussed here. In conclusion, therefore, the Washington altarpiece exemplifies a stage in the artist’s career in which he embarked on the gradual discovery of the innovative features of late-­Gothic art. This led him to develop greater elegance in poses, more delicate and harmonious arrangement of draperies, and more spontaneous vitality in the conduct of the angels thronged around the sides of the throne as if drawn magnetically to the child. A clear sign of the innovations of the phase in which Agnolo painted the frescoes in the choir of Santa Croce is also the artist’s polychromy: abandoning the somber palette of previous works, he now prefers or utilizes combinations of delicate pastel colors.

Miklós Boskovits (1935–2011)

March 21, 2016