Giotto di Bondone and Workshop
NG 5360 
Pentecost

Dillian Gordon, 2011

Extracted from:
Dillian Gordon, The Italian Paintings Before 1400 (London: National Gallery Company and Yale University Press, 2011).

c. 1310–18(?)

Egg tempera on poplar, 45.5 × 44.0 cm

The Apostles are shown gathered together in a loggia which has a coffered ceiling and is surmounted by three gables with traceried rose windows; the back wall is pierced by three windows with closed wooden shutters, and at the front is a balustrade with cosmati panelling and a wooden gate. The Apostles have tongues of fire on their heads, and the Holy Spirit descends in the form of a white dove. Outside, the multitude of different nations, symbolised by an old man and two youths, are gathered, each man marvelling at hearing the Apostles speak in his own language (Acts of the Apostles 2:1ff.).

Technical Notes

Panel structure and condition

Poplar board,¹ with a horizontal grain. Painted surface 45.5 × 44.0 cm; painted up to the edges all round; thickness 3.0–3.3 cm. The panel has been cut on all sides, but a comparison with the others in the series (see below) indicates that the painted area has not been significantly reduced in size. The gesso at the right edge rises slightly to form a lip or barb left by the removal of an engaged frame.

The reverse (fig. 1) shows that the board tapers slightly and a wedge‐shaped original insert completes the base at the lower right‐hand corner (left corner seen from the back). The back has traces of old gesso bearing the imprint of canvas and a minute fragment of red paint imitating porphyry. There is no gesso where a vertical batten c. 6.5 cm wide was originally placed slightly off centre (seen from the back 16.7 cm from the left edge at the top and 16.0 cm at the bottom, 21.0 cm from the right edge at the top and 21.3 cm at the bottom). Removal of the batten has revealed the holes of the nails that attached it to the board. The number 99 has been painted over the batten mark.²

On the back is a paper label on which is written in ink: This picture is the property of Mrs Henry Coningham, Plinlimmon, Shooters Hill, SE 18. 1.9.39.

Painting condition and technique³

Cleaned and restored in 1983.

A vigorous and bold underdrawing of quite broad lines done with a liquid medium is clearly visible under a microscope, but only faintly visible with infrared reflectography, which suggests that the composition was drawn with a non‐carboniferous material such as iron‐gall ink.

Some, but not all, of the architectural lines have been incised.

A green bole has been used as a base for the water gilding (also found in the other panels of the same series; see below) rather than the usual red bole.⁴ Small superficial scratches cover the entire gilded background.

Each halo has been incised with several concentric circles, variously, even randomly, spaced. The gold used for the haloes extends considerably further than necessary, so that, for example, the pink robe of the Apostle on the right and a large part of each Apostle’s face are painted over gold, with the result that many of the outlines have flaked and have had to be retouched. The outline of the profile of the Apostle on the extreme right has been almost entirely reconstructed, as has the beard of the Apostle third from the right in the front row. The Apostle just visible at the extreme left, whose face is painted entirely over gold, seems initially to have been omitted from the composition, but inserted after the completion of the gilding.⁵ In order to ensure the adhesion of the paint to the gold, the gold has been scratched in some places to provide a key for the paint.⁶

Mordant gilding, using a very thin mordant which is now impossible to detect even where the gold has flaked, has been used for the cosmati wall inlay, the coffered ceiling and the borders of the draperies, and is now somewhat worn.

The white dove (fig. 2) seems to have been added at a late stage, since the white paint goes over the mordant‐gilded [page 231][page 232] patterns of the coffered ceiling (mordant gilding was usually done in the final stages). The dove has been repainted, but traces of original white paint show that it was part of the original composition.

The repainting of the dove is connected with the application of the thick mordant‐gilded rays emanating from it, symbolising the descent of the Holy Spirit to the Apostles. These rays are not original; they have been applied over cracks in the paint surface. Instead, the Holy Spirit was originally represented by lines of gilded tin,⁷ which radiated from the dove to fall on the heads of the Apostles alongside the tongues of flame (fig. 3). These lines have all but disappeared; some of the incised lines used as guidance are just visible in raking light and in the X‐radiograph, passing across haloes and in the green background; a line of decayed tin, which now looks black, can be seen in the hair of the Apostle added at the extreme left in the back row, and tiny sporadic traces of tin are still to be found in the green background and elsewhere. A speck of tin and an incised line show that the ray directed towards the Apostle second from the left in the front row was originally at an angle, which crossed the face of the Apostle in red and was therefore moved to a different position. The gilded tin would have been considerably raised, standing out against the painted background and burnished gold haloes.

The overall condition of the painted surface is reasonably good, except for the damage caused by a horizontal split in the wood across the middle. The flesh is very thinly painted using a pale pink mixture of white and vermilion, given a slight greenish tinge in the half‐tones by the use of black. Detailing on top is outlined with a reddish brown. There is no green underpaint (fig. 4).

The distribution of colours in the composition is relatively simple – pinks alternating with greens, offset with the vermilion of one of the Apostles’ robes and the mustard‐coloured garment of the foreground figure on the left, achieved with orange and yellow earth mixed with an unidentified yellow lake. The pink of the garment of the figure at the extreme left is painted with red lake mixed with white, with a lake glaze and touches of ultramarine for the lining. The greenish‐blue draperies have an underlayer of azurite and white with glazes of pure azurite and a little ultramarine on top. In the case of the figure in the mustard‐coloured garment at the front, azurite has been used for his hose but the paint has darkened to appear almost black.

The tongues of flame are often double, to convey the description given in the Bible of ‘cloven tongues’, painted with a translucent orange‐coloured paint over vermilion mixed with white.

As well as being very thinly painted, the green background is somewhat abraded. The greenish effect has been achieved with an upper greenish‐yellow layer (lead‐tin yellow ‘type II’ mixed with a little black) painted over a warm pink made up of red earth and white, which is now showing through where the upper layer has flaked. Some pink particles, visible in cross‐section in the lower part of the yellow layer,⁸ suggest that the yellow layer was applied very soon after the pink layer, thus catching up some of the pink particles, and may therefore have been an intended tonal effect rather than a change of mind.

Azurite has been used for the cosmati inlay pattern of the architecture, contrasting with the red and gold.

Iconography

It is unusual to omit the Virgin Mary from the scene (compare Barnaba da Modena, NG 1437, p. 29), but Giotto also omits the Virgin in his fresco of the subject in the Arena Chapel, Padua. The closed doors may be intended to draw parallels with the Apparition behind Closed Doors, when the resurrected Christ appeared to the Apostles and they received the Holy Spirit (John 20:19–23).

The setting is unusual and appears to combine both secular and religious architectural features: the room with [page 233] its coffered ceiling and wooden shutters comes from domestic architecture, while the cosmati inlaid marble and the gables with traceried rose windows come from church decoration, and the enclosure with marble panels on either side of a wooden gate is reminiscent of the entrance to choir enclosures found in Roman basilicas, for example in San Clemente and Santa Maria in Cosmedin.⁹

Companion Panels

NG 5360 is one of a series of seven scenes: the others are the Nativity with the Epiphany (New York, Metropolitan Museum; fig. 5); the Presentation (Boston, Isabella Stewart Gardner Museum; fig. 6); the Last Supper and the Crucifixion (both Munich, Alte Pinakothek; figs 7 and 8); the Entombment (Settignano, Villa I Tatti, Berenson Collection; fig. 9); the Descent into Limbo (Munich, Alte Pinakothek; fig. 10). The seven paintings are linked together by their subject matter, style, dimensions¹⁰ and technique: all have the rare and distinctive green bole beneath the gold‐leaf background rather than the usual red bole. Like NG 5360, the painting of the Nativity in New York also includes gilded tin, which was used for the star behind the hill.¹¹

X‐radiographs (figs 12 a–g) revealing the pattern of the wood grain confirm that the seven panels were originally part of a single board of wood approximately 308 cm long, with the narrative scenes in chronological order.¹² A barb or lip of gesso where the integral frame has been removed at the left edge of the Nativity and at the right edge of the Pentecost confirms that these were the outermost panels, at the left and right respectively. Both these scenes have been made up at the base with wedges of wood. The intervening scenes seem to have lost little at the sides in the sawing up, and may originally have been divided by plain or decorated bands of gold leaf, rather than raised mouldings, since no barbs remain on the other panels. Alternate panels – the Nativity, the Last Supper, the Entombment and Pentecost – have traces of a vertical batten, not centred but irregularly placed, which may have been there to prevent the panel from bowing.¹³ Once the battens had been nailed on, the back was covered with a layer of fine canvas and gesso and then finished with a deep red fictive porphyry, of which traces remain on the Nativity, the Crucifixion, the Entombment, the Descent into Limbo and Pentecost.¹⁴ This finishing of the back suggests that the panels were intended to be visible from the back.

Function

Whether the series was autonomous or formed part of an altar piece is a matter of debate. Giovanni Previtali called the series a ‘paliotto francescano’, but also implausibly suggested that the panels could have formed shutters around a statue of the Virgin and Child or a reliquary.¹⁵ Everett Fahy suggested that they could have been sacristy cupboard doors.¹⁶ However, a vertical reconstruction is ruled out by the technical evidence (see above).

Cesare Gnudi thought that the panels would have formed a predella and suggested that they could have gone beneath the Washington/Horne/Musée Jacquemart André panels (for which see below).¹⁷ Keith Christiansen also reconstructed the seven panels as a predella.¹⁸ Giorgio Bonsanti took up the information reported by Cornelia Syre that, according to an agent in Rome, the series originally consisted of twelve scenes, and suggested that scenes from the Life of the Virgin could have gone above in a two‐tier structure.¹⁹ However, because the seven scenes were painted on a continuous board of wood there would need to have been originally fourteen scenes in two tiers.

The present author considers that the panels almost certainly constituted an autonomous altar piece. Long low dossals with narrative scenes were common in Umbria, particularly in Perugia, during the late thirteenth and early fourteenth centuries.²⁰

Attribution and Date

The attribution and the date of the series are problematic. The extent of the involvement of Giotto has been the subject of disagreement, and almost every combination of circumstances has been proposed: the panels have been seen as autograph paintings, as designed by Giotto and executed by his workshop, and as entirely by his workshop with no evidence of Giotto’s participation.²¹ The dating of the panels has ranged from soon after 1305 to around 1330.

Several critics do not accept the autograph status of the series. Bernard Berenson considered them to be purely a workshop product.²² Francesca Flores d’Arcais attributed them to Giotto’s workshop/‘Parente di Giotto’ with no evidence of the participation of Giotto himself, painted not long after 1305: she explored the close relationship of the composition of each of the seven panels to the frescoes in the Arena (Scrovegni) Chapel in Padua, completed by 1306, and to the frescoes in the right transept of the Lower Church of San Francesco, Assisi.²³

However, most critics accept the involvement of Giotto himself to a greater or lesser degree. For example, Roberto Salvini and Luisa Vertova saw the seven panels as having been designed by Giotto and executed by his assistants.²⁴ Roberto Longhi saw them as being by Giotto and suggested (given the presence of Saint Francis in the Crucifixion) that they formed [page 234] the altar piece in the Peruzzi Chapel in Santa Croce, Florence, which was frescoed by Giotto.²⁵ However, it is now generally agreed that the altar piece formerly in the Peruzzi Chapel, dedicated to Saints John the Baptist and John the Evangelist, is the polyptych now in the North Carolina Museum of Art, Raleigh, as first suggested by Wilhelm Suida.²⁶

Giovanni Previtali, on the basis of an analysis of the Boston Presentation, which he saw as falling between the Presentation of the Arena Chapel and the Presentation of the right transept of the Lower Church in San Francesco, Assisi, dated the seven panels after 1305 and before 1319, probably around 1310–15. He suggested that they could have formed one of the four altar pieces that Ghiberti said Giotto had painted for Santa Croce in Florence.²⁷ He drew attention to their stylistic similarities to the polyptych/polyptychs composed of the Virgin and Child (Washington, National Gallery of Art), Saint Stephen (Florence, Horne Museum), and Saint John the Evangelist and Saint Lawrence (Chaalis, Musée Jacquemart André), and the similar decoration on the headdress worn by the Virgin in the Washington panel and in the Nativity in New York.²⁸

Ferdinando Bologna, who dated the altar piece around 1327, agreed with the autograph status of the seven panels, and explored their stylistic links with the Washington/Horne/ Musée Jacquemart André panels, which he dated to soon after 1328, that is, contemporary with the frescoes by Bernardo Daddi in the Pulci Chapel in Santa Croce, dedicated to Saints Stephen and Lawrence. Bologna (following Gnudi) proposed the Pulci Chapel as the original site of the Washington/Horne/Musée Jacquemart André panels. He compared the two youths in the Pentecost to the two figures on either side of the scene of [page 235]Saint Francis renouncing his earthly goods in the uppermost tier of the Bardi Chapel, a scene which he saw as having been designed by Giotto and executed by Maso di Banco.²⁹

Most recently, the series has been described by Giorgio Bonsanti as a general diffused collaboration, close to the frescoes of the Bardi Chapel and painted around 1320–5, before Giotto’s departure for Naples in 1328,³⁰ with Miklós Boskovits discussing them as a mature work of the late 1320s³¹ and Angelo Tartuferi dating them around 1325.³²

The powerful draughtsmanship, the subtle painting of features and hair, and the sophisticated experimentation with colour in the Pentecost suggest that much of the panel was the work of Giotto himself, with the weaker passages, such as the spindly hands, possibly painted by assistants.³³

The design across the seven scenes of the altar piece is noticeably uneven: dramatically powerful compositions with three‐dimensional figures, such as the Nativity with the Epiphany and the Crucifixion, contrast with the unconvincing perspective of the ciborium in relation to the tiled floor in the Presentation and the clumsy perspective of the tomb in the Entombment, where the left arm of the Mary supporting the Virgin is curiously truncated. There is a notable lack of any attempt to arrive at a centralised viewpoint for the seven scenes: each is a self‐contained entity. Giotto’s involvement appears to have been piecemeal, perhaps indicating that at the time he himself was engaged on a more important project, leaving much of the execution to his workshop.

The composition in the Pentecost is similar to that of the equivalent scene in the Arena Chapel, while the architecture of the Last Supper is similar to that of the Marriage at Cana in the chapel, giving a terminus post quem of 1306. Stylistically the panels would appear to be after the Magdalene Chapel of around 1307–8 in the Lower Church of San Francesco, Assisi, and before the polyptych/polyptychs composed of the Washington/Horne/Musée Jacquemart André panels, themselves close to the Bardi Chapel. Julian Gardner has convincingly argued that the Bardi Chapel was painted between 1317 and 1321, and preceded the Peruzzi Chapel.³⁴ Some details of the altar piece that included NG 5360 can be compared with the Magdalene Chapel, such as the architectural idiom, and the way of painting trees; the gesture of the standing bearded man at the left of NG 5360 is similar to that of one of the onlookers seated at the table in the scene of Mary Magdalene washing Christ’s feet.³⁵ A date in the second decade of the fourteenth century for the altar piece that included NG 5360 is further suggested by the iconographic and stylistic similarities with the frescoes of the Infancy of Christ by Giotto’s workshop in the right transept of the Lower Church of San Francesco, which themselves are datable around 1311–13.³⁶ Similarities of the Presentation with that in Assisi were discussed by Previtali (see above). Isolated details of the altar piece are to be found in the Assisi frescoes, such as the thatched stable in the Nativity, which is very similar to that in the equivalent scene in the Lower Church.

Patron

The altar piece was presumably commissioned by the man and woman kneeling at the foot of the cross in the Crucifixion. Recent examination under a microscope confirms that the male donor is tonsured.³⁷ He is therefore a cleric. He is wearing [page 236]a pale green alb with an embroidered apparel and embroidered cuffs; his collar is an amice; hanging from his left wrist is a gold textile decorated with jewels, which matches his collar, and which seems to be a maniple.³⁸ The woman wears a white kerchief, and a blue dress painted with azurite mixed with some lead white and yellow lake.³⁹ The man looks up at Christ, while the woman looks across at the swooning Virgin.

Original Location

Keith Christiansen reiterated Previtali’s suggestion that Santa Croce may have been the original location of the altar piece, on the basis of Ghiberti’s statement, and suggested that it could have come from the Bardi Chapel itself, noting that the scenes from the Life of Christ would complement the frescoes there, which show scenes from the life of Saint Francis, with the Stigmatisation above the entrance to the chapel: Saint Francis was seen as the alter Christus.⁴⁰ However, according to Vasari the altar piece in the Bardi Chapel was a Crucifixion by Ugolino; this has been identified by Hayden Maginnis as possibly the panel in the Thyssen‐Bornemisza Collection.⁴¹ Furthermore, it should be noted that the width of the Bardi Chapel is only around 432 cm (the width of the altar piece that included NG 5360 was at least 308 cm), and it would seem unlikely that Giotto could have designed a long low altar piece for a tall thin chapel.

Another possibility regarding the original location of the altar piece to which NG 5360 belonged was proposed by Bologna, who took up Martin Davies’s idea that the seven panels are identifiable with ‘una tavola di man di Giotto di figure piccole, che poi se n’è ita in pezzi’ (‘a painting by the hand of Giotto of small figures, which subsequently was broken into pieces’). The painting was said by Vasari to have been taken from Borgo San Sepolcro to Arezzo in 1327 by Pietro Saccone di Mala, a brother of Guido Tarlati, Bishop of Arezzo, who died in 1327 and for whose tomb, according to Vasari, Giotto made drawings; the pieces were retrieved and taken to Florence in the sixteenth century by an acquaintance of Vasari, Baccio Gondi.⁴² Bologna further considered that the presence of Saint Francis in the Crucifixion indicated that the altar piece came from the church of San Francesco in Borgo San Sepolcro, whose altar table is inscribed 1304.⁴³ The status of the church is reflected in the fact that in 1305 Napoleone Orsini, Cardinal Legate in Umbria, granted 100 days’ indulgence to whoever visited the main altar.⁴⁴ The fact that the altar piece was finished with fictive porphyry on the reverse could support the case for its having come from Borgo San Sepolcro, where the Franciscans later, in 1426, commissioned a double‐sided altar piece, eventually painted by Sassetta in 1437–44.⁴⁵ Moreover, the altar table, which measures approximately 342.5 cm,⁴⁶ could well accommodate the seven narrative panels plus the hypothetical intervening gold bands and a frame. In such a setting the seven scenes would mirror the seven arches of the altar below.

Vasari does not give the original location of the painting by Giotto, and the date of the removal would have been comparatively soon after the panels had been painted (whichever of the dates discussed above one considers correct).

The information Vasari supplies regarding the removal of the altar piece in 1327 from Borgo San Sepolcro is quite specific, and he gives the correct date for the death of Guido Tarlati.⁴⁷ Saccone, Guido Tarlati’s brother, a military man, is not described as having removed the paintings by force but merely as having taken them (‘condusse’), and this could suggest that he had a legal right to the altar piece, which could in turn suggest that it might have been commissioned by Tarlati.⁴⁸ Alternatively, Saccone might have felt free to remove the painting once his brother had died and could no longer prevent him from doing so. If the kneeling donor were to be identifiable as Guido Tarlati, then this would possibly be before he became Bishop of Arezzo in 1312,⁴⁹ while he was still ‘arciprete’, since no mitre is visible.⁵⁰ However, were the kneeling cleric to be Tarlati, one might question why it was commissioned for Borgo San Sepolcro and not Arezzo. It could be that Tarlati’s relationship with the Franciscans in Arezzo was poor: he excommunicated a number of them in 1314.⁵¹ It could also be that the commission was instigated by the woman, possibly the mother or a sister, kneeling prominently in the foreground of the Crucifixion, who may have had some connection with Borgo San Sepolcro and a particular devotion to the Franciscan Order.

Carl Strehlke has made the interesting suggestions that the woman may be no relation to the cleric but may represent a donor who had made a substantial contribution to the cost of the altar piece, and that the presence of Saint Francis could refer to the name of the donor.⁵² It could equally refer to the name of the female donor.

The somewhat unusual selection of scenes could suggest that the two Infancy scenes were chosen by the female donor, while the choice of the Last Supper, Crucifixion and Pentecost, all with Eucharistic connotations, could have been the preferred choice of a cleric. Strehlke remarks on the absence of important scenes such as the Resurrection.⁵³ Instead, the relatively rare scene of the Descent into Limbo, which is common in Riminese dossals, was included.⁵⁴

It was previously suggested by the present author that the altar piece could have come from San Francesco, Rimini, where Giotto painted a large Crucifix, and which was in effect the burial church of the Malatesta family.⁵⁵ The possibility that the male donor is a layman is now ruled out by the fact that he has a tonsure. However, the selection of scenes indicates a possible location of the altar piece in Rimini and the surrounding area.⁵⁶ Moreover, the iconography of the panel in the Metropolitan Museum is unusual in being a hybrid of the Nativity and the Adoration of the Magi. These scenes are traditionally shown separately, as in the Arena Chapel and in the Infancy scenes of the Lower Church of San Francesco, Assisi. It may be significant that these two scenes are combined in subsequent Riminese painting (see, for example, figs 15 and 16 on pp. 212–13).⁵⁷ The reason for their combination here may be due to the wishes of the female donor and the need to include as many episodes as possible from the Infancy of Christ, but not at the expense of the Passion and post‐mortem scenes desired by the cleric.

It seems unlikely that the altar piece was commissioned for a church dedicated to Saint Francis. The rare examples of [page 237]members of the secular clergy commissioning paintings for a Franciscan church mostly concern San Francesco, Assisi, the burial church of the saint.

The simple format would have meant that the altar piece could have been painted wherever Giotto and his workshop happened to be and sent to its final destination.⁵⁸

Exhibited

London 1989–90, National Gallery, Art in the Making. Italian Painting before 1400, 29 November–28 February (2).

Provenance

In the collection of Prince Stanislas Poniatowski in 1822;⁵⁹ together with the Nativity, the Presentation and the Entombment in the Poniatowski sale, Christie’s, London, 9 February 1839, lots 101–4;⁶⁰ NG 5360 was lot 104, bought by Hall. Bequeathed by William Coningham (1815–1884) to his son W.J.C. Coningham, who bequeathed it to his widow,⁶¹ who gave it in 1922 to Major Henry Coningham. Bequeathed to the National Gallery by his widow, Geraldine Emily Coningham, in memory of her husband and of Mrs Coningham of Brighton, 1942.

Notes