Paintings by famous early Renaissance artists are rare in Australian collections, so the realisation that the Saint George and the dragon, c.1430 (fig. 1),¹ The work was acquired by the Felton Bequest in 1949 with the name Saint George slaying the dragon (A. J. L. McDonnell, letter to the Secretary, Felton Bequests’ Committee, 26 April 1949. p. 2, including ‘Report on work of art’, Felton Bequest Correspondence, 2/20, National Gallery of Victoria Library, Melbourne). However, as the saint has discarded his sword, has only a dagger in his hand, and as the princess bears a chain with what seems to be a collar, it is likely that, in this version of the subject, the dragon will be tethered rather than slain. Thus, the more general title Saint George and the dragon is more fitting. in the National Gallery of Victoria is by Paolo Uccello, one of the legendary artists of the Italian Renaissance, is a notable development in the study of European paintings in this country. Having belonged to a number of discerning British collectors in the nineteenth century under various attributions, the work’s true authorship was recognised in 1968 by the eminent Italian art historian and connoisseur Roberto Longhi. Since Carlo Volpe’s persuasive endorsement of 1980, the majority of authors have accepted the attribution. A close look at Uccello’s career helps to locate the work in his artistic development with some precision. The recent technical examination of this work and the Annunciation, c. 1430 (fig. 2), in the Ashmolean Museum, Oxford, with which it has long been associated, and a comparison of these with Uccello’s better-known works, provides further evidence for their attribution to him. As well, it reveals the sophistication of Uccello’s technique, even when working on a small scale.

The provenances of the Saint George and the Annunciation are indicative of the growing taste for Italian early Renaissance paintings in Britain from the first half of the nineteenth century. The Saint George was most probably among the more than seventy paintings offered for sale by the London art dealer Samuel Woodburn (1786–1853) to the National Gallery, London, in 1846, at a time when there had been calls for it to represent the early Italian school in its collection. The Saint George appears in the undated List of Woodburn’s Collection of Early Italian Pictures, housed in the Gallery’s archives, as item number five, attributed to Andrea Orcagna.² National Gallery, London, archives. Also among the works offered was the magnificent central panel of the Virgin and Child with Angels from Masaccio’s Pisa altarpiece, then thought to be by Gentile da Fabriano. The Gallery rejected the offer on Woodburn’s terms;³ For a summary of the circumstances of Woodburn’s offer to the Gallery, see: S. Avery-Quash, ‘The growth of interest in early Italian painting in Britain with particular reference to pictures in the National Gallery’, in D. Gordon (ed.), The Fifteenth Century Italian Paintings, vol. I, National Gallery Catalogues, London, 2003, pp. xxvi–xxviii. however, the Masaccio was eventually acquired by the Gallery in 1916, and it purchased the now celebrated Saint George and the dragon by Uccello from the Lanckoronski Collection in 1959.⁴ Gordon, 2003, pp. 219, 403. Following Woodburn’s death, his Saint George was sold from his estate in 1860 to a buyer simply recorded as ‘Campanari’. This may well have been the Italian Marquis Domenico Campanari, a client of Christie’s in the mid nineteenth century. The painting was Lot 59, still attributed to Orcagna, and was described as ‘an exquisite work of the highest rarity and interest’.⁵ ‘Catalogue of the very celebrated & valuable series of capital pictures, by the greatest early Italian Masters, formed under singular advantages, by that distinguished Connoisseur, the late Samuel Woodburn, Esq. […]’ Lot 59 (9–11 June 1860), Christie’s archives, London. In 1848 Domenico Campanari purchased Bronzino’s Portrait of a woman (Maria Salviati?), now in the M. H. De Young Memorial Museum, San Francisco, which he took for a lost portrait of Vittoria Colonna by Michelangelo, and published it as such (L. Østermark-Johansen, ‘The matchless beauty of widowhood: Vittoria Colonna’s reputation in nineteenth-century England’, Art History, vol. XXII, no. 2, June 1999, p. 270). He also seems to have acted as the London agent for his brother Carlo, a leading excavator and collector of Greek, Roman, and Etruscan artefacts, with an estate at Toscanello. Many hundreds of objects from Carlo’s collection were bought for the British Museum in the nineteenth century. Another brother, Secondino, and their father were also involved in researching and publishing early Tuscan art (G. Dennis, Cities and Cemeteries of Etruria, London, 1848, pp. 441–7).

The painting appeared shortly thereafter as Lot 135 in the 1863 sale of the Reverend Walter Davenport Bromley’s collection. Davenport Bromley (1787–1863) was among the first significant collectors of Italian early Renaissance paintings in Britain; he owned such important works as Giovanni Bellini’s Agony in the Garden, now in the National Gallery, London, and Giotto’s Dormition of the Virgin, now in the Stäatliche Gemälde Sammlung, Berlin.⁶ D. Sutton, ‘From Ottley to Eastlake’, Apollo, August 1985, pp. 86, 88. The Saint George was bought by ‘Burton’,⁷ ‘Catalogue of the celebrated collection of pictures, of the Rev. Walter Davenport Bromley, deceased: which will be sold by auction, by Messrs. Christie, Manson & Woods, at their Great Rooms, 8, King Street, St. James’s Square, on Friday, June 12, 1863, and following day. […]’ Lot 135 (12 June 1863), Christie’s archives, London. possibly the Mrs F. W. Burton who sold it in 1867 to James Carnegie, the ninth Earl of Southesk (1827–1905), according to an inventory of 1904 from Kinnaird Castle.⁸ Hugh Brigstocke, Assistant Keeper, National Gallery of Scotland, letter to Christopher Lloyd (with extract from Catalogue of Pictures at Kinnaird Castle, 1904), 23 January 1969, Ashmolean Museum Curatorial File. From 1922 the painting was loaned to the National Gallery of Scotland, following a fire at the castle.⁹ Catalogue of Pictures at Kinnaird Castle, 1904 (including later MS add:), National Gallery of Scotland archive. It was then sold, presumably by Sir Charles Alexander Carnegie (1893–1992), the eleventh Earl of Southesk, to the London dealers Ellis and Smith who, in turn, sold it to Agnew’s on 9 February 1949.¹⁰ Stock no. 11, in stockbook no. 12, Agnew’s archive, London. On 11 July it was acquired by the Felton Bequest as a work of Domenico di Bartolo, on the enthusiastic recommendation of the Felton Bequest Advisers A. J. L. McDonnell and Kenneth Clark, who felt it would be ‘an ornament to the Gallery and a constant source of delight to its visitors’.¹¹ A. J. L. McDonnell, letter to the Secretary, Felton Bequests’ Committee, 26 April 1949, p. 3 (copy). It was accessioned in the same year. The Annunciation was donated to the Ashmolean Museum in 1850 together with Uccello’s Hunt in a forest and a substantial number of Italian early Renaissance works by The Hon. W. T. H. Fox-Strangways (1795–1865), the fourth Earl of Ilchester, another pioneering British collector of these works.¹² C. Lloyd, A Catalogue of the Earlier Italian Paintings in the Ashmolean Museum, Oxford, 1977, pp. xv, 61–2, 172–5.

The Saint George was published in 1927 in volume nine of Raimond van Marle’s monumental survey The Development of the Italian Schools of Painting, where it appeared with an attribution to the Sienese artist Domenico di Bartolo.¹³ R. van Marle, The Development of the Italian Schools of Painting, The Hague, 1927, vol. IX, p. 544. The following year Roberto Longhi re-attributed it to an anonymous direct follower of Uccello in an article on the subject on the artist’s late-Gothic followers, principally Giovanni di Francesco.¹⁴ R. Longhi, ‘Ricerche su Giovanni di Francesco’, Pinacotheca, vol. VI, no. 1, anno 1, July–August 1928, p. 38. Subsequently, the work’s attribution continued to oscillate between minor or anonymous Sienese or Florentine artists. In his 1935 discussion of Uccello’s collaborators and followers who worked on the mural paintings in the Chiostro Verde of Santa Maria Novella, Georg Pudelko observed that the Saint George is by the same hand as the Annunciation in the Ashmolean Museum, showing influences of Fra Angelico and Uccello.¹⁵ G. Pudelko, ‘The minor masters of the Chiostro Verde’, Art Bulletin, vol. XVII, no. 1, March 1935, p. 72 n. 8. In 1939 John Pope-Hennessy described the artist as a Florentine student of Uccello in a footnote to his discussion of the Sienese artist Sassetta,¹⁶ J. Pope-Hennessy, Sassetta, London & Toronto, 1939, pp. 202, 203 n. 133. and no author has since doubted that the two paintings are by the same hand. In the 1968 re-publication of his above-mentioned article, Longhi revised his attribution of the Saint George without explanation, giving it to Uccello himself.¹⁷ Longhi, ‘Me pinxit’ e quesiti Caravaggeschi 1928–1934, Florence, 1968, p. 25.

Carlo Volpe, in his 1980 article in the Italian journal Paragone, initiated an important development in the critical reception of the Saint George and the Annunciation, and for the understanding of Uccello’s early career. He wrote it following the discovery in the church of San Martino Maggiore in Bologna of Uccello’s beautiful though, sadly, damaged Adoration mural painting, hidden for centuries under a layer of whitewash. Volpe accepted Longhi’s expansive re-assessment of Uccello’s oeuvre, which included a number of works whose attributions to Uccello had been disputed, but not Longhi’s argument that Uccello reached his artistic maturity late in life. Uccello incised the date of the Adoration into an area of drapery in the foreground of the work. The first three digits clearly read 143, the last digit, however, can be read variously as 1, 2, or 7.¹⁸ That the date is original is suggested by the nature of its incisions, in which a fine instrument must have been used, creating slight ridges along the length of some strokes perhaps where the still wet material of the intonaco (fresco layer) or arriccio was pushed to one side. This appears different to the graffiti elsewhere on the fresco, in which the incisions are much coarser. The white material in parts of the incisions must be the remains of the whitewash that covered the fresco until 1977. In any case, with a date in the 1430s, the discovery of the charming, nocturnal scene of the Adoration re-affirmed Uccello’s masterful synthesis of perspective and the sculptural modelling of forms early in the Renaissance, still evident despite the disturbing losses it has suffered. Volpe attributed the Adoration to Uccello and dated it to a relatively early part of his career as well as the Virgin and Child detached mural painting (Museo di San Marco, Florence), the Quarate predella (Museo Diocesano, Florence), the Stories of the Virgin and Saint Stephen mural paintings in the Mircovaldi Chapel of Prato Cathedral, the Oxford Annunciation and the Melbourne Saint George.¹⁹ C. Volpe, ‘Paolo Uccello a Bologna’, Paragone, no. 365, anno XXXI, July 1980, pp. 3–28. Since 1980 Uccello’s authorship of the Saint George and the Annunciation has been widely accepted.²⁰ See A. Angelini, ‘Paolo Uccello’, in L. Bellosi (ed.) Giovanni di Francesco e l’arte fiorentina di metà Quattrocento, exh. cat., Casa Buonarroti, Florence, 16 May–20 August 1990, p. 73; M. Boskovits, The Martello Collection: Further Paintings, Drawings and Miniatures 13th–18th Century, Florence, 1990, p. 140; F. & S. Borsi, Paolo Uccello, trans. E. Powell, New York, 1994. (orig. Italian ed. Milan, 1992), pp. 346–7; A. Padoa Rizzo, La Capella dell’Assunta nel Duomo di Prato, Prato, 1997, p. 17; L. B. Kanter, ‘The ‘cose piccole’ of Paolo Uccello’, Apollo, August 2000, p. 17; T. Gott, ‘Attributed to Raolo Uccello …’ in Painting and Sculpture before 1800 in the International Collections of the National Gallery of Victoria, 2003, p. 23.

There is ample stylistic evidence for the attribution of the two works to Uccello. As seen in the Karlsruhe Adoration²¹ The attribution of the Karlsruhe Adoration to Uccello has been contested, but was accepted by Franco and Stefano Borsi, 1994, pp. 290–2. (Staatliche Kunsthalle, Karlsruhe), the landscape of the Saint George is divided into distinct horizontal bands and, like the Crucifixion with the Virgin and Saints (Museo Thyssen-Bornemisza, Madrid), a kind of aerial perspective is created in the Saint George and the Annunciation through the use of contrasting bands of colour in the landscapes as they recede to the horizon. Volpe noted the similarity of the horse in the Saint George to the one in the Quarate predella²² Volpe, 1980, p. 17. and the dragon’s wings in the Saint George bear the same dark circles between their webbing as the crest of the knight just behind Tolentino in the Battle of San Romano (National Gallery, London), as well as the dragon’s wings in the Saint George and the dragon in the Musée Jacquemart-André in Paris. The dark circles on the wings of the Paris dragon have honey-coloured crescents inside them, comparable to the gold crescents on the wings of the Melbourne dragon.²³ The moon and other crescents appear in many of Uccello’s works, though in the context of the Saint George, they may serve a particular iconographic purpose. Given the strong allusion to God the Father as the sun and the fact that the name of the city, Silena, is similar to the name of the Greek moon goddess, Selene, the iconography of the work seems to be based on a contrast between Christianity and paganism expressed through the motifs of the sun and the moon. Golden rays of ‘sunlight’ and the moon appear in the top left and right corners, respectively, of Uccello’s Saint George and the dragon in the Musée Jacquemart-André. Dragons do not often have horns; however, the Melbourne dragon has one painted with a scalloped base as though constructed like a piece of metalwork,²⁴ The design of the dragon is related to a drawing previously in a private collection (current location unknown) published by Bernhard Degenhart and Annegrit Schmitt (‘Uccello: Wiederherstellung einer Zeichnung’, Albertina-Studien, 1963, vol. I, p. 114 and fig. 17). The drawn and painted dragons share a long neck curved in an S shape, the horizontal bands of scales on the front of the neck, the two rows of circular scales on the back of the neck, and the large head encircled by shaggy hair and a horn (or horn-like) projection from the forehead. The drawing was attributed by Degenhart and Schmitt to an anonymous fifteenth-century Florentine artist copying the lost Fight between dragons and lions painting by Uccello in the Palazzo Medici. It may, however, be derived from a model-book drawing used by Uccello and his contemporaries. as is the unicorn’s horn on the standard of the Battle painting in the Louvre. The decorative punte di diamante (pyramid-shaped points) on the ceiling of the portico in the Annunciation are comparable to those on the underside of the sarcophagus of the Equestrian monument for Sir John Hawkwood in Florence Cathedral, and the arabesques of Gabriel’s windswept drapery are similar to those of the embodiment of Faith on the ceiling of the Microvaldi Chapel in Prato.

Payments made to Uccello for his work as a young assistant in Lorenzo Ghiberti’s workshop,²⁵ J. Beck, ‘Uccello’s apprenticeship with Ghiberti’, Burlington Magazine, vol. CXXII, no. 933, December 1980, p. 837. and Ghiberti’s influence, especially of his most famous work, the Doors of Paradise, as Michelangelo is said to have named them, has been seen in Uccello’s works. The Saint George seems to rely on compositional formulae from the David and Goliath panel of the Doors of Paradise. In both compositions fighting takes place in the foreground, with a rocky outcrop forming a backdrop and a walled city filling almost the entire width of the background. Similarly, the Annunciation seems to rely on compositional formulae from the Essau and Jacob panel of the Doors of Paradise. In each, the portico dominates the scene around which the narrative unfolds, with God the Father in an upper corner. Uccello, like Ghiberti, included a curtain draped over a beam to soften the hard lines of the architecture. Technical examination of the Annunciation discussed below shows that the opening of the portico facing the viewer was first drawn and incised as a round arch and the doorway leading inside was drawn and incised as an arch. The arch closest to the viewer was ultimately painted with a pointed top and the doorway was made rectangular. Thus, the initial design of the portico was initially closer to the design of Ghiberti’s portico in the Essau and Jacob panel of the Doors of Paradise, with its round arches. The exact dates of production for the panels of the Doors of Paradise are unknown but must fall between 1425, when Ghiberti received the commission, and April 1437, when all the panels were cast.²⁶ R. Krautheimer & T. Krautheimer-Hess, Lorenzo Ghiberti, Princeton, 1956, pp. 191–92. Since Uccello was in Venice from some time after 5 August 1425 to at least 12 July 1427,²⁷ The precise dates of Uccello’s departure from and subsequent return to Florence are unknown. Uccello’s will (cited in G. Gaye, Carteggio inedito d’artisti dei secoli XIV, XV, XVI, Florence, 1839, vol. I, pp. 147–8) was written in Florence on 5 August 1425. His 1427 tax return was lodged by his relative Deo Beccuti, who stated that Uccello left for Venice more than two years before (Archivio di Stato di Firenze, Catasto, 55, San Giovanni, Drago, fols 707–707v, dated 12 July, among tax returns from 1427). Uccello’s tax-return of January 1431 seems not to be autograph, since it is written in the third person (‘sua incharichi’). Furthermore, the handwriting is similar to that of Deo Beccuti, and so he may have submitted Uccello’s 1431 tax return as he had the previous one. (Archivio di Stato di Firenze, Catasto, 381, San Giovanni, Drago, fol. 779. Though undated, the entry is between others dated 30 January 1431, for example, those beginning fols 764, 788, in the series of returns compiled in chronological order). This does not necessarily imply that Uccello had not already returned from Venice, since he may have led a peripatetic life at the time. his knowledge of Ghiberti’s designs for the Doors of Paradise most likely dates from after his return to Florence. Most recent authors have dated the Saint George to the early 1430s, although Volpe and others have dated the Annunciation to about 1420.²⁸ Volpe, 1980, p. 18; F. & S. Borsi, 1994, p. 345.

Further indirect evidence that the works date from the early 1430s is provided by Uccello’s stylistically similar Virgin and Child detached mural painting (fig. 3), formerly in a house of the Del Beccuto family in Florence and now in the Museo di San Marco in the same city. Anna Padoa Rizzo has plausibly associated the work with debts owed to Uccello by Deo di Deo del Beccuto, a wealthy and powerful relative from his mother’s family, recorded in Uccello’s tax statements of 1431 and 1433. Although the cause for the debts is not specified, none was declared in Uccello’s 1427 tax statement, and so the debts may relate to work done by Uccello for Deo Beccuti following his return to Florence.²⁹ A. Padoa Rizzo, ‘L’Età di Masaccio, Firenze, Palazzo Vecchio, giugno 1990’, Antichità viva, nos 2–3, anno XXIX, p. 58. Despite the differences in media, all three works feature gold grounds, large amounts of blue (identified in the Oxford and Florence paintings as the expensive pigment lapis lazuli,³⁰ For the Annunciation, see M. Spring, M975 Uccello, The Annunciation (Ashmolean Museum, Oxford) Report on analysis of paint samples, June 2004, Conservation Department, National Gallery, London. The micro-sample was taken on 27 November 2002; for the Virgin and Child, see Volpe, 1980, p. 18. and it may be in the Melbourne painting as well) and a reliance on Ghibertian compositional formulae. The Virgin and Child is similar in form and colouring to the Virgin and Child polychrome stucco sculpture in the Victoria and Albert Museum, London, attributed to Ghiberti’s workshop. The precious materials and small scale of the Saint George and the Annunciation suggest they were commissioned by a wealthy patron or patrons for a domestic context, as appears to have been the case for the Del Beccuto Virgin and Child.

Technical investigations of the Saint George and the Annunciation using scientific analyses were undertaken for the first time between 2002 and 2004.³¹ In September of 2002 the Ashmolean Museum and the National Gallery, London, kindly agreed to undertake a series of technical analyses of the Annunciation. The work travelled to London in November; four X-radiography prints, together covering the whole surface of the painting, were made on 14 November. The irr prints of details of the painting arrived in March, made from mosaics assembled from data recorded on 21 November 2002. The painting was examined in infrared by Rachel Billinge, Rausing Research Associate in the Conservation Department of the National Gallery, London. Infrared reflectography was carried out using a Hamamatsu C2400 camera with an N2606 series infrared vidicon tube. The camera is fitted with a 36mm lens to which a Kodak 87A Wratten filter has been attached to exclude visible light. The infrared reflectography mosaics were assembled on a computer using Vips-ip software. For further information about the software, see . Scientific analyses of the Saint George were undertaken at the National Gallery of Victoria. The X-radiography images were made by John Payne, Senior Conservator of Paintings, and Carl Villis, Conservator of Paintings before 1800, using 4 mA, 35kV, 25-second exposure, Agfa D7 X-radiographic film and an Andrex CP 490 x-ray unit; the mosaic was assembled by Villis. Infrared images of the obverse and reverse were made in March 2004 by Garry Sommerfeld, Senior Photographer, NGV, using PowerPhase FX scanback on Sinar 5 × 4 camera, Kodak Wratten 87c filter and tungsten lights. For each work the support is a single piece of wood with vertical grain, with no trace of original frames or substantial remains of barbes (ridges of paint and/or ground) around the outside of the paint surfaces, which have been damaged and restored in places in both works. However, there is no reason to think that the painted surfaces have been reduced significantly. The round top of the Saint George is probably the original shape of the paint surface,³² The cloth interlayer visible in the X-radiography of the Saint George, discussed later in the body of this article, stops at a point below the arch, presumably because it was easier not to cut the cloth to follow the round part of the panel. as is the case with the Karlsruhe Adoration where traces of the original arch-shaped frame remain. The sunburst behind God the Father in the Saint George harmonises perfectly with the round arch format. Given its smaller dimensions and different shape, the Saint George was probably not originally part of an integral ensemble with the Annunciation, as has been intimated by some authors.³³ Borsi, 1994, p. 346. The Saint George may have been set within a tabernacle-style frame to hang on a wall or to stand alone on a piece of furniture, possibly with wings. Two small areas of restoration on the vertical edges of the painting at the springing of the arch suggest that the original frame had pilasters whose capitals impinged slightly onto the surface of the painting. The condition of the paint surface on both works is quite good overall despite old worm infestations in the panels, visible in the X-radiography of the Annunciation and in the exposed sides of the Saint George. The photograph of the latter, published by Van Marle, shows losses to the dragon’s wings that have since been restored, possibly when the work was with Agnew’s, since the restorer Horace Buttery was paid £18 in relation to the work at this time, although no explanation for the payment was recorded.³⁴ Buttery’s record books (MS. 1112–1993, Fitzwilliam Museum, Cambridge, housed at the Hamilton Kerr Institute) do not contain any reference to the Saint George.

Close inspection of the X-radiography of the Annunciation also reveals the presence of a single piece of fine-weave cloth that extends to, or close to, the four edges of the panel. Cloth is also visible in places at the edges of the Saint George and its X-radiography shows that a single piece of fine-weave cloth covers the panel up to a point just below the top of God the Father’s papal tiara (fig. 4). Cloth interlayers were commonly used by Italian Renaissance artists to provide a smooth surface for the application of the ground and subsequent paint layers on a panel; however, there were variations in the quality of cloth employed and the extent of its use. Uccello applied cloth strips to cover knots and joins between panels of the Hunt in a forest and covered almost the entire panel of the Battle in the Musée du Louvre with pieces of cloth.³⁵ M. Kemp et al., ‘Paolo Uccello’s Hunt in the forest, Burlington Magazine, vol. CXXXIII, no. 1056, March 1991, p. 176 n. 2. In the Karlsruhe Adoration, however, he applied roughly torn patches of coarse-weave cloth.³⁶ The X-radiography was kindly provided by Dr Dietmar Lüdke, Senior Curator for Old Masters, and the Conservation Department at the Stäatliche Kunsthalle, Karlsruhe. The extensive use of fine-weave cloth in the Saint George and the Annunciation is indicative of their technical similarity and their high value.

By comparing the quality of the lines that appear in infrared reflectography (IRR), the X-radiography and on the paint surface, it is possible to distinguish underdrawing from incisions and painted lines and to establish Uccello’s working procedure in the Annunciation. The underdrawing is most easily identified when lines in the irr do not correspond to incisions or painted lines on the surface. The lowest depiction of Gabriel was drawn with his left hand holding a lily stem, appearing from behind his right sleeve (fig. 5), but this detail was ultimately not incised or painted. The Holy Spirit was drawn next to the top right of the capital of the free-standing pillar, but was painted a fraction lower. An S-shaped curve for a contour of the drapery of the Virgin’s robe that appears in the IRR a little way below the book was neither incised nor painted (fig. 6). The medium of the underdrawing is difficult to determine, but may be a combination of drawing with the point of a brush and metalpoint for the straight lines.

The IRR shows ruled lines for the architecture and freehand lines for the figures and contours of drapery, with little or no hatching for shadows and no obvious monochrome wash drawing. It seems that the Uccello first drew the outlines of the portico as a box-like, wire-frame structure with a ruler and then drew the Virgin and the lowest Gabriel within the architectural space. The principal lines of construction for the portico pass through the Virgin. Similarly, the vertical edges of the far pillar extend through the lowest Gabriel’s sleeve and an incision passes through his halo (fig. 5). There are other ruled lines in the area occupied by the Virgin’s body, whose significance is unclear, but which appear to be improvised construction lines. The IRR shows numerous adjustments to the drawing for the architecture, suggesting that there was no detailed, auxiliary preparatory drawing for the portico, with the details being largely worked out on the panel, nor does the construction appear to have been measured. For example, the decorative frieze along the top of the building facing the viewer was divided into approximately, but not exactly, equal-sized rectangles in the underdrawing as a guide for the repeated arabesque motif.

Varying amounts of underdrawing and incisions are present in all the figures and the landscape in the Annunciation. The incisions in the drapery follow the underdrawing somewhat loosely. The freely executed arabesques in the incisions for the drapery and figures in both works, such as the corkscrew curls of God the Father’s hair in the Saint George (fig. 1), show an artist completely at ease with his technique. Where geometric precision was required, tools were used; compasses were used to draw the haloes for the seraphim in the Annunciation, as indicated by the points visible in the centres of the unpainted ones.

Perhaps the most interesting aspect of the design stage in the Annunciation is the geometric construction of perspective in the architecture, an aspect of painting with which Uccello is synonymous. It seems that the artist drew an approximate foreshortened square in perspective for the ceiling of the portico, drew diagonals between the corners of the ceiling to find the middle of the square and then, after many attempts, divided the square into a grid to provide the basis of the decoration of the pyramid-shaped points. The final grid is based on eight rows of squares by eight rows, while only those visible through the pointed arch were actually drawn and incised. In the underdrawing for the cornice around the freestanding pillar, separated from the capital by a block, the artist first determined the position of the four corners of the pillar, including the corner which is not visible (approximately), he then drew two diagonals between these corners to establish the correct angle for the corners of the cornice as they extend beyond the pillar. Having worked out the final version of the cornice he then extended the principal lines to the left so that the cornice of the far pillar would be correctly aligned. He also extended construction lines to align their capitals (figs 5 & 8). Of this fairly extensively underdrawn perspective construction, many lines do not appear in the X-radiography or as incisions on the paint surface. Thus, the artist generally incised, or incised more strongly, those lines that he intended to be visible in the final composition.

A similar geometric approach to the planning of perspective is visible in the Adoration in San Martino Maggiore. The lines of construction incised into the arriccio (preparatory layer) for a pyramid-shaped point on the inside edge of the right side of the architectonic frame show that Uccello drew two diagonals between the corners of a foreshortened square to find its centre, as he did in the Annunciation, but then extended a perpendicular line from this point to arrive at the correct position for the top of the punta in relation to the base. Piero della Francesca later followed a similar procedure when he constructed the pyramid-shaped roof of a house in perspective in his treatise On Painted Perspective.³⁷ See De Prospectiva Pingendi, Book 2, Proposition 9, fol. 25v., c.1470–1480, Biblioteca Palatina, Parma.

In contrast, the perspective construction of the city in the background of the Saint George is minimal, showing that Uccello’s approach to perspective was not dogmatic. The infrared image reveals a little freehand underdrawing in the battlements and buildings (fig. 7), a number of buildings were incised freehand and there is a long, ruled incision as a guide for the battlements along the front of the city wall. The infrared image and X-radiography show that numerous changes were made to the design of the city during its execution. The representation of architecture in the Saint George serves as a background to the narrative, rather than a mis en scène, as it does in the Annunciation, explaining why Uccello gave less attention to this aspect of the work.

Striking features of the irr of the Annunciation are the very dark areas that are light-to-mid-grey in the painting, such as the fur lining of God the Father’s robes, his hair and beard, the fur lining of the Virgin’s robes and the ground in front of the portico (figs 5 & 6). This is not the result of a dark preparatory layer under the surface, as is sometimes the case with Uccello’s areas of green paint,³⁸ For an example, see Kemp et al., p. 169. but the effect of black pigment in the grey paint. A micro-sample taken by the Scientific Department of the National Gallery, London, from the ground just to the left of the portico was found to contain a single layer of grey paint, composed of lead white and a relatively coarse black pigment, identified as probably a carbon black, possibly a mineral black such as graphite.³⁹ Spring. A comparable phenomenon occurs in the Hunt in a forest, where the fur cuffs and collar of a hunter’s costume, which are a mid-brown colour on the paint surface, appear black in the IRR, probably because of an admixture of black pigment.⁴⁰ See Kemp et al., fig. 13. A micro-sample taken from a brown horse revealed finely ground yellow and red ochre, black and lead white (p. 178). Although no micro-sample was published for the cuffs or collars, it seems reasonable to assume that they too contain an admixture of black pigment.

An aspect of the technique of the Saint George visible to the naked eye is the use of coloured glazes over gold and silver leaf. A similar technique is recorded since at least the twelfth century,⁴¹ M. Doerner, The Materials of the Artist and Their Use in Painting with Notes on the Techniques of the Old Masters, trans. E. Neuhas, rev. edn, 2nd impression, London, 1969, p. 321, noted that the eighth-century Lucca Manuscript described the ‘pictura translucida’ technique of applying resin and oil-based colours over tin foil. was used across Europe, and was particularly popular in Florence in the first half of the fifteenth century. In the Saint George a large part of the dragon’s wings and body are painted with semi-transparent green glazes over gold leaf, reinforced with black hatching in the shadows. The same technique of using green glazes over gold leaf was used by Uccello in the surcoat of Tolentino’s page in the London Battle. The adjacent brocade is executed with red glazes over gold leaf and the sallet (helmet) in his hand is executed with red glazes over silver leaf. Like Saint George’s armour, the page’s armour is executed with opaque blackish glazes over silver leaf.⁴² A. Roy & D. Gordon, ‘Uccello’s Battle of San Romano’, National Gallery Technical Bulletin, vol. XXII, 2001, p. 10. The gold bands of God the Father’s papal tiara in the Annunciation, and the corresponding feature in the Saint George (fig. 1), as well as the princess’s gold girdle, show traces of red glazing. The media for the two works would seem to be basically tempera with some oil glazing. These works would have made a sumptuous impression when first painted, with large areas of exposed silver and gold leaf as well as areas covered in jewel-like coloured glazes.

The gold ground around the splendid figure of God the Father in the Saint George is incised with ruled, radiating lines and hexa-prong punchwork is used to create the alternating areas of stippled texture (fig. 1). Uccello used a similar, perhaps identical, punch in the pomegranate designs on Tolentino’s headdress in the London Battle.⁴³ For a macro-photograph showing the punch-work, see P. Roccasecca, Paolo Uccello: Le Battaglie, Milan, 1997, p. 47. In the Annunciation similar punchwork appears in the seraphim, although not with sufficient clarity to determine the type of punch used. Round punches were used to embellish the musical Angels’ haloes, while the Virgin’s and the lowest Gabriel’s haloes were incised by hand with meandering motifs, in a manner distinct from the Saint George.

There are other minor differences between the works. In the Annunciation the facial features are predominantly modelled softly, in red-brown tones, while in the Saint George graphic strokes of black define the upper eyelashes, the irises and pupils of the eyes and the lines between the lips, closer to Fra Angelico’s style in such works as the Fiesole predella in the National Gallery, London. Conversely, the pastel colours of some of the Angels’ robes in the Annunciation are close to Fra Angelico’s delicate palette, while the strongly contrasting colours of the Saint George are characteristic of Uccello’s sometimes wonderfully unorthodox palette.

An intriguing aspect of the Saint George is the presence of fingerprints in the blue paint of the walled city’s tower and the horse’s saddle. Like Van Eyck and Leonardo, Uccello used his fingers to manipulate the glazes of his paintings while they were still fresh. In the London Battle he used his thumb and fingers to modulate the blackish glazes over a layer of silver leaf.⁴⁴ Roy & Gordon, p. 10. The fingerprints in the Saint George do not appear in areas with glazing; however, this does not exclude the possibility that the artist worked the paint intentionally. The Karlsruhe Adoration also shows extensive use of the artist’s fingertips to work blackish glazes over gold leaf in the drapery, particularly in the angels’ robes. But the fingerprints in the Saint George are too smudged and partial to provide a match with those in the Battle or the Adoration.

The Melbourne Saint George and the Oxford Annunciation exhibit the valuable materials, sophisticated and refined technique and strong Ghibertian influence of Uccello’s works from about the early 1430s. They date from a critical point in his career, following his return to Florence from Venice, in which he affirmed his position as one of the leading mural and panel painters of the early Renaissance. Not long after he would paint such masterpieces as the Equestrian monument for Sir John Hawkwood in Florence Cathedral and the three Battle paintings now in the National Gallery, London, the Musée du Louvre, Paris, and the Galleria degli Uffizi, Florence. In the sixteenth century Giorgio Vasari named Uccello, with Brunelleschi, Donatello, Ghiberti and Masaccio, as one of the remarkable generation that renewed the art of Florence in the Renaissance.⁴⁵ G. Vasari, Le vite de’piú eccellenti architetti, pittori, et scultori italiani, da Cimabue insino a’ tempi nostri, originally published Florence 1550, L. Bellosi & A. Rossi (eds), Turin, 1991, vol. I, p. 266. Following the peregrinations of the Saint George and the Annunciation in the nineteenth and twentieth centuries, the reasons for his fame can now be appreciated as far afield as Melbourne and Oxford.

Hugh Hudson, Fellow in Art History, University of Melbourne (in 2006).

Notes

 

Acknowledgements

Thanks are due to Dr Catherine Whistler at the Ashmolean Museum for approving the scientific analyses of the Annunciation; Rachel Billinge at the National Gallery, London, for making the irr of the work and discussing the interpretation of its scientific analyses; Dr Gerard Vaughan at the National Gallery of Victoria for approving the scientific analyses of the Saint George; and Carl Villis, John Payne and Garry Sommerfeld for making the scientific images; Dr Ted Gott and Michael Watson for assisting my research in the Gallery’s archives. Many thanks also go to Julian Agnew and Jane E. H. Hamilton, Agnew’s, London; Aidan Weston-Lewis, National Gallery of Scotland; Dr Alexandra Villing, British Museum; Isobel Siddons, National Gallery, London; Lynda McLeod, Christie’s, London; Dr Dietmar Lüdke and staff of the conservation department, Staatliche Kunsthalle, Karlsruhe; and the Research and Graduate Studies Committee at the School of Art History, Cinema, Classics and Archaeology, the University of Melbourne, for awarding a grant to this project.

This article is a development of the author’s paper ‘The materials and technique of two panel paintings attributed to Paolo Uccello: The Oxford Annunciation and the Melbourne Saint George’, in H. Verougstraete & R. van Schoute (eds), Le Dessin sous-jacent et la technologie dans la peinture: copies, répliques, pastiches, colloque XV, Paris, (forthcoming, 2006), including analysis of the scientific images of the Saint George that were not available when the paper was given. This research was undertaken towards a PhD at the University of Melbourne, ‘Paolo Uccello: The life and work of an Italian Renaissance artist’, completed in September 2005.

1     The work was acquired by the Felton Bequest in 1949 with the name Saint George slaying the dragon (A. J. L. McDonnell, letter to the Secretary, Felton Bequests’ Committee, 26 April 1949. p. 2, including ‘Report on work of art’, Felton Bequest Correspondence, 2/20, National Gallery of Victoria Library, Melbourne). However, as the saint has discarded his sword, has only a dagger in his hand, and as the princess bears a chain with what seems to be a collar, it is likely that, in this version of the subject, the dragon will be tethered rather than slain. Thus, the more general title Saint George and the dragon is more fitting.

2     National Gallery, London, archives.

3     For a summary of the circumstances of Woodburn’s offer to the Gallery, see: S. Avery-Quash, ‘The growth of interest in early Italian painting in Britain with particular reference to pictures in the National Gallery’, in D. Gordon (ed.), The Fifteenth Century Italian Paintings, vol. I, National Gallery Catalogues, London, 2003, pp. xxvi–xxviii.

4     Gordon, 2003, pp. 219, 403.

5     ‘Catalogue of the very celebrated & valuable series of capital pictures, by the greatest early Italian Masters, formed under singular advantages, by that distinguished Connoisseur, the late Samuel Woodburn, Esq. […]’ Lot 59 (9–11 June 1860), Christie’s archives, London. In 1848 Domenico Campanari purchased Bronzino’s Portrait of a woman (Maria Salviati?), now in the M. H. De Young Memorial Museum, San Francisco, which he took for a lost portrait of Vittoria Colonna by Michelangelo, and published it as such (L. Østermark-Johansen, ‘The matchless beauty of widowhood: Vittoria Colonna’s reputation in nineteenth-century England’, Art History, vol. XXII, no. 2, June 1999, p. 270).
He also seems to have acted as the London agent for his brother Carlo, a leading excavator and collector of Greek, Roman, and Etruscan artefacts, with an estate at Toscanello. Many hundreds of objects from Carlo’s collection were bought for the British Museum in the nineteenth century. Another brother, Secondino, and their father were also involved in researching and publishing early Tuscan art (G. Dennis, Cities and Cemeteries of Etruria, London, 1848, pp. 441–7).

6     D. Sutton, ‘From Ottley to Eastlake’, Apollo, August 1985, pp. 86, 88.

7     ‘Catalogue of the celebrated collection of pictures, of the Rev. Walter Davenport Bromley, deceased: which will be sold by auction, by Messrs. Christie, Manson & Woods, at their Great Rooms, 8, King Street, St. James’s Square, on Friday, June 12, 1863, and following day. […]’ Lot 135 (12 June 1863), Christie’s archives, London.

8     Hugh Brigstocke, Assistant Keeper, National Gallery of Scotland, letter to Christopher Lloyd (with extract from Catalogue of Pictures at Kinnaird Castle, 1904), 23 January 1969, Ashmolean Museum Curatorial File.

9     Catalogue of Pictures at Kinnaird Castle, 1904 (including later MS add:), National Gallery of Scotland archive.

10     Stock no. 11, in stockbook no. 12, Agnew’s archive, London.

11     A. J. L. McDonnell, letter to the Secretary, Felton Bequests’ Committee, 26 April 1949, p. 3 (copy).

12     C. Lloyd, A Catalogue of the Earlier Italian Paintings in the Ashmolean Museum, Oxford, 1977, pp. xv, 61–2, 172–5.

13     R. van Marle, The Development of the Italian Schools of Painting, The Hague, 1927, vol. IX, p. 544.

14     R. Longhi, ‘Ricerche su Giovanni di Francesco’, Pinacotheca, vol. VI, no. 1, anno 1, July–August 1928, p. 38.

15     G. Pudelko, ‘The minor masters of the Chiostro Verde’,  Art Bulletin, vol. XVII, no. 1, March 1935, p. 72 n. 8.

16     J. Pope-Hennessy, Sassetta, London & Toronto, 1939, pp. 202, 203 n. 133.

17     Longhi, ‘Me pinxit’ e quesiti Caravaggeschi 1928–1934, Florence, 1968, p. 25.

18     That the date is original is suggested by the nature of its incisions, in which a fine instrument must have been used, creating slight ridges along the length of some strokes perhaps where the still wet material of the intonaco (fresco layer) or arriccio was pushed to one side. This appears different to the graffiti elsewhere on the fresco, in which the incisions are much coarser. The white material in parts of the incisions must be the remains of the whitewash that covered the fresco until 1977.

19     C. Volpe, ‘Paolo Uccello a Bologna’, Paragone, no. 365, anno XXXI, July 1980, pp. 3–28.

20     See A. Angelini, ‘Paolo Uccello’, in L. Bellosi (ed.) Giovanni di Francesco e l’arte fiorentina di metà Quattrocento, exh. cat., Casa Buonarroti, Florence, 16 May–20 August 1990, p. 73; M. Boskovits, The Martello Collection: Further Paintings, Drawings and Miniatures 13th–18th Century, Florence, 1990, p. 140; F. & S. Borsi, Paolo Uccello, trans. E. Powell, New York, 1994. (orig. Italian ed. Milan, 1992), pp. 346–7; A. Padoa Rizzo, La Capella dell’Assunta nel Duomo di Prato, Prato, 1997, p. 17; L. B. Kanter, ‘The ‘cose piccole’ of Paolo Uccello’, Apollo, August 2000, p. 17; T. Gott, ‘Attributed to Paolo Uccello …’ in Painting and Sculpture before 1800 in the International Collections of the National Gallery of Victoria, 2003, p. 23.

21     The attribution of the Karlsruhe Adoration to Uccello has been contested, but was accepted by Franco and Stefano Borsi, 1994, pp. 290–2.

22     Volpe, 1980, p. 17.

23     The moon and other crescents appear in many of Uccello’s works, though in the context of the Saint George, they may serve a particular iconographic purpose. Given the strong allusion to God the Father as the sun and the fact that the name of the city, Silena, is similar to the name of the Greek moon goddess, Selene, the iconography of the work seems to be based on a contrast between Christianity and paganism expressed through the motifs of the sun and the moon. Golden rays of ‘sunlight’ and the moon appear in the top left and right corners, respectively, of Uccello’s Saint George and the dragon in the Musée Jacquemart-André.

24     The design of the dragon is related to a drawing previously in a private collection (current location unknown) published by Bernhard Degenhart and Annegrit Schmitt (‘Uccello: Wiederherstellung einer Zeichnung’, Albertina-Studien, 1963, vol. I, p. 114 and fig. 17). The drawn and painted dragons share a long neck curved in an S shape, the horizontal bands of scales on the front of the neck, the two rows of circular scales on the back of the neck, and the large head encircled by shaggy hair and a horn (or horn-like) projection from the forehead. The drawing was attributed by Degenhart and Schmitt to an anonymous fifteenth-century Florentine artist copying the lost Fight between dragons and lions painting by Uccello in the Palazzo Medici. It may, however, be derived from a model-book drawing used by Uccello and his contemporaries.

25     J. Beck, ‘Uccello’s apprenticeship with Ghiberti’, Burlington Magazine, vol. CXXII, no. 933, December 1980, p. 837.

26     R. Krautheimer & T. Krautheimer-Hess, Lorenzo Ghiberti, Princeton, 1956, pp. 191–92.

27     The precise dates of Uccello’s departure from and subsequent return to Florence are unknown. Uccello’s will (cited in G. Gaye, Carteggio inedito d’artisti dei secoli XIV, XV, XVI, Florence, 1839, vol. I, pp. 147–8) was written in Florence on 5 August 1425. His 1427 tax return was lodged by his relative Deo Beccuti, who stated that Uccello left for Venice more than two years before (Archivio di Stato di Firenze, Catasto, 55, San Giovanni, Drago, fols 707–707v, dated 12 July, among tax returns from 1427). Uccello’s tax-return of January 1431 seems not to be autograph, since it is written in the third person (‘sua incharichi’). Furthermore, the handwriting is similar to that of Deo Beccuti, and so he may have submitted Uccello’s 1431 tax return as he had the previous one. (Archivio di Stato di Firenze, Catasto, 381, San Giovanni, Drago, fol. 779. Though undated, the entry is between others dated 30 January 1431, for example, those beginning fols 764, 788, in the series of returns compiled in chronological order). This does not necessarily imply that Uccello had not already returned from Venice, since he may have led a peripatetic life at the time.

28     Volpe, 1980, p. 18; F. & S. Borsi, 1994, p. 345.

29     A. Padoa Rizzo, ‘L’Età di Masaccio, Firenze, Palazzo Vecchio, giugno 1990’, Antichità viva, nos 2–3, anno XXIX, p. 58.

30     For the Annunciation, see M. Spring, M975 Uccello, The Annunciation (Ashmolean Museum, Oxford) Report on analysis of paint samples, June 2004, Conservation Department, National Gallery, London. The micro-sample was taken on 27 November 2002; for the Virgin and Child, see Volpe, 1980, p. 18.

31     In September of 2002 the Ashmolean Museum and the National Gallery, London, kindly agreed to undertake a series of technical analyses of the Annunciation. The work travelled to London in November; four X-radiography prints, together covering the whole surface of the painting, were made on 14 November. The irr prints of details of the painting arrived in March, made from mosaics assembled from data recorded on 21 November 2002. The painting was examined in infrared by Rachel Billinge, Rausing Research Associate in the Conservation Department of the National Gallery, London. Infrared reflectography was carried out using a Hamamatsu C2400 camera with an N2606 series infrared vidicon tube. The camera is fitted with a 36mm lens to which a Kodak 87A Wratten filter has been attached to exclude visible light. The infrared reflectography mosaics were assembled on a computer using Vips-ip software. For further information about the software, see . Scientific analyses of the Saint George were undertaken at the National Gallery of Victoria. The X-radiography images were made by John Payne, Senior Conservator of Paintings, and Carl Villis, Conservator of Paintings before 1800, using 4 mA, 35kV, 25-second exposure, Agfa D7 X-radiographic film and an Andrex CP 490 x-ray unit; the mosaic was assembled by Villis. Infrared images of the obverse and reverse were made in March 2004 by Garry Sommerfeld, Senior Photographer, NGV, using PowerPhase FX scanback on Sinar 5 × 4 camera, Kodak Wratten 87c filter and tungsten lights.

32     The cloth interlayer visible in the X-radiography of the Saint George, discussed later in the body of this article, stops at a point below the arch, presumably because it was easier not to cut the cloth to follow the round part of the panel.

33     Borsi, 1994, p. 346.

34     Buttery’s record books (MS. 1112–1993, Fitzwilliam Museum, Cambridge, housed at the Hamilton Kerr Institute) do not contain any reference to the Saint George.

35     M. Kemp et al., ‘Paolo Uccello’s Hunt in the forest, Burlington Magazine, vol. CXXXIII, no. 1056, March 1991, p. 176 n. 2.

36     The X-radiography was kindly provided by Dr Dietmar Lüdke, Senior Curator for Old Masters, and the Conservation Department at the Stäatliche Kunsthalle, Karlsruhe.

37     See De Prospectiva Pingendi, Book 2, Proposition 9, fol. 25v., c.1470–1480, Biblioteca Palatina, Parma.

38     For an example, see Kemp et al., p. 169.

39     Spring.

40     See Kemp et al., fig. 13. A micro-sample taken from a brown horse revealed finely ground yellow and red ochre, black and lead white (p. 178). Although no micro-sample was published for the cuffs or collars, it seems reasonable to assume that they too contain an admixture of black pigment.

41     M. Doerner, The Materials of the Artist and Their Use in Painting with Notes on the Techniques of the Old Masters, trans. E. Neuhas, rev. edn, 2nd impression, London, 1969, p. 321, noted that the eighth-century Lucca Manuscript described the ‘pictura translucida’ technique of applying resin and oil-based colours over tin foil.

42     A. Roy & D. Gordon, ‘Uccello’s Battle of San Romano’, National Gallery Technical Bulletin, vol. XXII, 2001, p. 10.

43     For a macro-photograph showing the punch-work, see P. Roccasecca, Paolo Uccello: Le Battaglie, Milan, 1997, p. 47.

44     Roy & Gordon, p. 10.

45     G. Vasari, Le vite de’piú eccellenti architetti, pittori, et scultori italiani, da Cimabue insino a’ tempi nostri, originally published Florence 1550, L. Bellosi & A. Rossi (eds), Turin, 1991, vol. I, p. 266.