Francesco Francia and workshop
NG 179, 180 
The Buonvisi Altarpiece: Saint Anne with the Virgin and Four Saints (main panel) and the Pietà (lunette)

Giorgia Mancini, Nicholas Penny, 2016

Extracted from:
Giorgia Mancini and Nicholas Penny, The Sixteenth Century Italian Paintings, Volume III: Bologna and Ferrara (London: National Gallery Company and Yale University Press, 2018).

1510–12

NG 179 oil on canvas (transferred from panel),
201.3 × 185.4 cm, painted surface 198.7 (max) × 182.7 (max) cm;
NG 180 oil on wood, 98.5 × 191 cm (including additions)

NG179 signed on the throne: ‘FRANCIA.AVRIFEX. BONONIE̅SIS.P.’ (‘Francia. Goldsmith. of Bologna. p[ainted this]’)

Support

The main panel (NG 179) has been transferred from wood to canvas. The canvas has been marouflaged onto a synthetic support. From the pattern of losses visible in the photographs made after cleaning, the panel must have had six vertical joins and consisted of seven boards of wood, the outer two being relatively narrow. The joins run through Saint Sebastian’s right foot and just to the right of his head, through the Virgin’s right eye, Saint Anne’s left hand, Saint Lawrence’s left shoulder and behind Saint Benedict’s head.

The lunette (NG 180) was painted on a poplar panel (about 2.6 cm thick) composed of three boards joined horizontally and shaped as a semicircle. The spandrels are a later addition and increase the overall height of the panel by 4.5 cm and the width by 6 cm (3 cm each side). These were attached either in the late eighteenth or early nineteenth century, when this painting was converted from a crowning lunette to a separate gallery painting in a rectangular frame.

The back of NG 180 exhibits saw marks, but there is also evidence that the wood was split into boards and there are some traces of hand finishing, so the panel probably retains its original thickness. Dowels are visible in the X‐radiographs. Two small butterfly keys (one at the centre of each join) were inserted from the back to reinforce the joins, which were later also reinforced with fibres. Two vertical battens, apparently of pine, are inserted into channels with a dovetail profile in the back of the panel. They have been sawn at the lower edge of the panel, which is very uneven and has probably been cut – but by how much is not clear.

The panel has some woodworm damage and slight splits where it is restricted by the added spandrels but is otherwise in good condition. On the back, on the original part of the panel, are inscriptions and three wax seals.

Preparatory layers and preliminary drawing

The preparatory layer was largely removed from the main panel during the transfer carried out in 1891, but fragmentary remains in several cross‐sections of paint indicate that both panels were prepared with a layer of gesso.¹ A thin unpigmented layer was applied on top, and in one sample from the lunette it was possible to determine that this layer is composed of a drying oil.²

Lines that can be identified for certain as underdrawing are not obvious in the infrared images of the main panel. This could be because some of the underdrawing was lost in the transfer process.

In the infrared images of the lunette some drawing can be seen, consisting of simple outlines, very finely drawn. These are only identifiable as underdrawing where the painted outline does not follow the drawing exactly. This is the case in the left arm of the angel in red, which was painted lower than it was drawn (fig. 1). Smaller changes include enlargement of the eyes of several of the figures and adjustment of the positions of chins and hairlines. In the lower portion of the panel, where there is now a straight ledge of grey stone, are lines which appear to be continuations of the draperies and of the heel of Christ’s left foot and, by the foot of the angel in green, what looks like the top of a spray of lilies. This is in keeping with the evidence from panel examination that the lower edge has been cut. There seems therefore to have been a first idea for the composition that extended further down the panel. It must have been abandoned quite early in the painting process, since a sample from the ledge just below the angel in green shows that only the underpaint for his drapery runs beneath the grey stone paint.

In the infrared images of the main panel it is possible to see a few fine lines similar to the underdrawing on the lunette, and similar small changes to eyes and the contours of faces. The most significant change is to the position of the Christ Child’s head, which seems to have been drawn slightly lower and to the right of its painted position.

Medium and pigments

Linseed oil was identified as the paint binder in samples from Saint Benedict’s white robe and the blue sky in the main panel, and in green and red paint from the lunette.³

The sky is painted with ultramarine mixed with lead white over a layer of azurite and lead white. The Virgin’s cloak in the main panel – the brightest of all the blue draperies – employs these same blue pigments. The first layer consists of azurite mixed with varying amounts of white, but in some areas further layers were applied in pink and purple shades, the deepest purple‐red shadows being mixtures containing mainly red lake with a little ultramarine or azurite and colourless soda‐ash glass.⁴ The whole cloak was then glazed with natural ultramarine, applied with varying thickness to further refine the folds, and sometimes mixed with white for the very brightest lights. This distinctive layer structure is also to be found in the darker and more purple blue of Saint Anne’s cloak and in the robe of Saint Paul, with more reliance on the purple‐red shades in the undermodelling.

The deep blue of the Virgin’s cloak in the lunette is painted in the same way as Saint Anne’s. Surprisingly, the blue runs beneath the areas showing the green lining and beneath the whole of her red robe, which is only glazed rather thinly on top with red lake. The far brighter red of Saint Paul’s cloak on the main panel is instead achieved with multiple layers of undermodelling consisting of various mixtures of vermilion, red lake, lead white and bone black, finally glazed with red lake. All the layers include colourless soda‐ash glass as an additive.

The green lining of Saint Paul’s robe was painted first with a thin layer of verdigris, lead white and black; thick layers, consisting chiefly of verdigris, were applied over this. The angel in green in the lunette was also painted with verdigris. A sample from the knee shows that the lower half was painted first in a blue paint made from azurite and white.⁵ The upper part appears different and may not have the same structure, or it may be that the upper green paint is more thickly applied, obscuring more effectively what is beneath. The spotty final layer of verdigris has been blotted with a cloth over the whole drapery, presumably to make it uniform.

Beneath the thin black modern overpaint in the background of the lunette is the mid‐grey original paint consisting of lead white and lamp black. This has been applied over a light beige layer consisting of lead white with some black, yellow earth and colourless soda‐ash glass.

Green pigment was seen when examining the surface under low magnification in the flesh of every figure in the main panel, giving the greenish‐brown tone that is especially evident in the darkest shadows. This pigment was identified as green earth in paint samples, and found to have been mixed with lead white, vermilion, black, yellow ochre, red lake, a very small amount of ultramarine in places, and with colourless soda‐ash glass as an additive. The modelling of the flesh was built up in multiple layers – four or five are visible in most of the cross‐sections – working from relatively pale beige hues that probably constitute the mid‐tones, with darker shadows and the lightest highlights applied on top.

Gilding

Much of the gold decoration is reinforced. It could only be examined under low magnification, and observations concerning its original character in the main panel are advanced tentatively. The gold of the signature seems to have been applied on an insubstantial, perhaps colourless, mordant. The haloes seem to be drawn with shell gold (reinforced with later gold paint), perhaps with mordant gilding for the stronger line around the perimeter. The flecks of gold in the Christ Child’s three‐pointed radiance seem to be the work of a restorer, but they follow traces of an original. There are remnants of what seems to be an original gold pattern on Saint Sebastian’s loincloth, and also of a gilded fringe where it billows out below. There is a touch of gold on one of the arrows, in a line of light along the metal tip. The pattern on Saint Paul’s sword also appears to be rendered with shell gold, and there is a simple double line of gold around the hem of his red cloak, together with touches of gold around the cuff and neck of his blue robe, which suggest a gold‐patterned border. Saint Anne’s blue cloak has a solid line of gold around its edge, as do the cuffs of her red robe. Saint Lawrence’s alb has gold lines at the neck and small gold dots on the dark green half‐circles of the orphreys on his shoulders. Around the neck and cuffs of the Virgin’s red dress and the hem of her blue cloak is a slightly more elaborate gold‐patterned border, now very faint. Very little of this is visible in ordinary viewing conditions. In the lunette all the haloes seem to have been reinforced, probably because of the overpaint in the background. They were toned down in the most recent treatment. There are traces of original gold decoration in the form of a single line around the cuffs and hem of the robe of the angel in green, and around the neck, cuffs and hem of the robe of the angel in red. There are also touches of gold on the end of the feathers of both angels.

Condition

The main panel has suffered from being transferred on to canvas. The paint surface is uneven and has sunk in places. Its condition must, however, have been a serious problem before the transfer took place as it exhibits many of the same problems suffered by the lunette, which has not been transferred.

Both paintings have widespread flake losses. In the main panel there are small scattered losses from many areas but especially from the sky. There are also areas where losses are larger, including the outer edges and the robes of Saint Anne and Saint Paul. Both paintings exhibit areas where wide cracks have developed in the paint as it was drying and paint has subsequently been lost. This is especially bad in areas of orange or red drapery, such as the robes of the Virgin and Saint Anne, the cloak of Saint Paul in the main panel, and the robe of the angel in red in the lunette.

In both the main panel and the lunette the paint containing azurite has darkened and the paint containing ultramarine has blanched. Blanching can also be seen in areas of brown shadow such as the hair of Saint Sebastian in the main panel and of Christ in the lunette.

To disguise all these problems, both paintings had been extensively repainted before the most recent treatment; much of this old repaint was not removed and is now becoming more visible due to discoloration. The gilding is worn and has been reinforced, as has been mentioned in the previous section.

Conservation history

Both paintings are likely to have been cleaned when they were re‐framed for display in the Palazzo Buonvisi in the late eighteenth century (see ‘Provenance’ below). After the transfer of both paintings to the ducal palace in Lucca it was agreed, on 27 August 1827, that they should be restored. The work was probably undertaken soon afterwards by Michele Ridolfi or at least under his supervision.⁶ The spandrels were probably added at that date, although this could have been done in previous decades. The style of their decoration would be possible in 1800 but more likely in the 1820s.

The main panel was transferred from wood to canvas in 1891. Flaking paint on both panels was secured in 1940 and 1955. In 1971–3 the transfer canvas of the main panel was marouflaged onto a modern synthetic support. At the same time the painting was cleaned and restored, as was the lunette.

Saint Anne and the Virgin in the main panel

The altarpiece was commissioned for the Buonvisi chapel dedicated to Saint Anne, the mother of the Virgin Mary, in the church of S. Frediano, Lucca. An account of the church and chapel is provided in a separate section below. In the main panel the Virgin Mary and Saint Anne are seated close together on a throne, and Francia’s composition reflects their dual prominence here with an arch springing above each figure. They are united by the infant Christ, who is seated on his mother’s lap but leans towards his grandmother to take the fruit proffered by her (probably intended to be an apple but more like an apricot in size and colour) . The departures from, and debts to, the prevailing conventions for depicting Saint Anne and the Virgin are described in Appendix 1 below, but it is worth observing here that, among altarpieces with this subject, the painting is highly unusual in representing Saint Anne as smaller in size than her daughter. Francia took great care in painting the blues, as is explained in the section on pigments above. He was especially concerned to ensure that the Virgin’s blue is the brightest in the painting.

The Virgin is less directly engaged with her child because we are to understand that she is thinking of his destiny, which is foreshadowed by the Sacrifice of Isaac: the Old Testament episode represented by Francia as a relief carving in the capital beside her head. It was not a new idea to incorporate an Old Testament episode into the Virgin Mary’s throne or the surrounding architecture but it was ingenious to do so in a manner that suggests the Virgin’s thoughts, and tactful to give this an emphasis that no attentive viewer can miss but that does not disrupt the figural composition.⁷

The saints beside the throne in the main panel

Saint Sebastian and Saint Paul stand on the left side of the throne and Saint Lawrence and Saint Benedict on the right. Three of these are the name saints of the members of the Buonvisi family, who commissioned the altarpiece and for whom the chapel was built: the patron, Benedetto (1450–1513); his brother, Paolo (died 1484); and their father, Lorenzo (died 1451).⁸ Saint Sebastian, who was often invoked against the plague, may have been included on account of the plague that broke out in Lucca in 1510, and other such reasons have been given for – but are not necessary to explain – the presence of Saint Anne and Saint John the Baptist.⁹

Sebastian is depicted nude but for a loincloth, tied to a column and pierced by three arrows. It is typical of Francia at this stage in his development that the arrows are inconspicuous. Paul is, as usual, depicted with the sword with which he was executed. His pose repeats that of his neighbour and indeed their feet are almost identical. The rich colours he wears connect him with the Virgin above, although the blue of his tunic is greener in hue than that of the Virgin’s cloak; these colours also help link the principal panel with the lunette.

Lawrence, as always, is attired as a deacon, wearing a dalmatic (the long‐sleeved outer tunic worn by a deacon over his alb, or inner tunic of white linen). In this case the dalmatic is of a gold damask with a pattern outlined in scarlet, a textile that the patrons, involved in the silk trade, would have especially valued.¹⁰ The damask is divided by green orphreys into nine visible panels, most of which are decorated with bold foliate patterns but two of which feature unusual open knot designs. Lawrence was martyred in Rome in AD 258 – roasted, according to legend, on a grill, the handle of which he holds here with his right hand, together with a martyr’s palm in the other.

Saint Benedict was normally represented in a black habit but the Camaldolite Order, which followed his rule, wore white and the Buonvisi family was closely connected with the Camaldolese abbey of Cantignano.¹¹ Carmichael’s suggestion that Benedict may be a portrait¹² should not be dismissed, although, if it is one, the subject has surely been idealised. Benedict’s features differ from those of the other saints – in particular, his eyeballs seem safely lodged within their sockets – and a profile view was often employed for portraits incorporated in altarpieces. It is noteworthy that Benedict looks at Christ, whereas Paul and Lawrence look at Benedict. Given that these saints stand for family members, this is appropriate, since Benedetto was the true patron of the chapel. Sebastian, who was not chosen as a family saint, looks up to heaven.

In front of the pedestal of the throne the infant John the Baptist (holding a scroll inscribed ‘ECCE AGNVS DEI’ on one side, with the ‘ECCE’ legible in reverse on the other) raises his hand and lifts his head to acknowledge not only the infant Christ above him, but also Christ sacrificed, depicted in the lunette. His gesture also draws attention to Francia’s signature. The nimble posture of his seemingly weightless body is supplemented by the delightful melodic lines of the scroll in front of him and the two flying tails of the belt for his camelskin behind him.

The Pietà in the lunette

The lunette (NG 180) depicts the Virgin Mary with the dead Christ, flanked by two angels, against a dark background. The scale of the figures is larger than those in the main [page 173] panel, to compensate for the greater height at which they were placed.¹³ Christ’s body rests on his mother’s lap, his back supported by her right hand and his feet touching the ground. She looks down with eyes red from weeping (fig. 3). With great subtlety Francia makes her resemble her mother in the panel below, thus showing how grief has aged her. The colour of her blue cloak is also the same as that of Saint Anne (as noted above in the section on pigments). These two panels appear to have been the only parts of the altarpiece (no predella is recorded) and this is only one respect in which they are connected.

The angel on the left looks at the viewer, while supporting Christ’s head with his right hand, partially concealed by hair, his fingertips gently touching Christ’s cheek. A few shining corkscrew curls, of the kind that Francia delighted in painting a decade or so earlier, are blown back from this angel’s cheek (fig. 4).

The angel on the right, his hands joined in prayer, looks sorrowfully downward (perhaps at the wounds in Christ’s feet). This angel’s drapery, rising like flames behind the pale feet, has a movement and complexity that are faintly echoed in the flutter of Sebastian’s loincloth below. Both angels have large, dark wings, the outer ones sweeping in lines parallel with the edge of the lunette.

Only a fraction of the wound in Christ’s side is visible to us; no other marks of suffering can be detected. His expression is one of sorrow and repose, and the stiffness of his body is softened by gracious curves. The representation [page 174] of controlled and silent grief, all the more intense because it is so restrained, was greatly and justly admired in the past.¹⁴

From the mid‐fifteenth century the dead Christ had become a common subject for the upper register of altarpieces, especially in north Italy. It supplemented, and may perhaps have served as a substitute for, the Crucifixion, which had long been obligatory for chapels (generally placed upon, but sometimes hung above, the altar) and was obviously appropriate for the place where the eucharistic sacrament was consecrated. In many cases the body of Christ was shown on the edge of the tomb, supported by angels, as in the cimasa of Francia’s Felicini Altarpiece (Pinacoteca Nazionale, Bologna) and in numerous paintings by Giovanni Bellini and his followers. But some version of the Lamentation, an episode (not described in the Gospels) occurring between the Deposition and the Entombment, was also employed: a striking example of a lunette with this subject is the one installed in 1487 in S. Giorgio, Ferrara, by Cosimo Tura for the Roverella Altarpiece (now in the Louvre; the main panel is NG 772), which Francia would surely have known. Both Tura and Francia show Christ’s body on the Virgin’s lap, a subject known as the Pietà, which should be understood as a condensed and intimate version of the Lamentation. The Pietà was also popular in sculpture and, especially around 1500, for the main panels of altarpieces and for smaller devotional paintings (such as NG 2671).¹⁵ There is no consistency in the way that subjects of this sort were identified, either in the sixteenth century or later: an inventory of 1834 describes NG 180 as a ‘deposizione di Croce’.¹⁶

The architecture

During the 1490s Francia painted four monumental altarpieces featuring the Virgin and Child enthroned with saints, each of which included an elaborate architectural setting and a landscape beyond: the Felicini Altarpiece, the Manzuoli Altarpiece and the Scappi Altarpiece, all today in the Pinacoteca Nazionale, Bologna, and the Bentivoglio Altarpiece in S. Giacomo Maggiore, Bologna (p. 151, fig. 1).¹⁷ The setting Francia has supplied for the main panel here retains features found in these earlier works but is far simpler, clearer and cleaner. The landscape is now shrunk to a detail beside Sebastian’s knee, although the clouds have the same fleecy shapes with thin horizontal plinths that Francia had previously employed. The foliate ornament on the underside of the arch that springs forward from the pillar behind the throne is like that of the frieze in the Manzuoli Altarpiece but far more discreet. The idea of decorating the uppermost portion of a pier with a relief was earlier used in the Bentivoglio Altarpiece, albeit competing there with the decoration of a capital below. The concave form of the pedestal of the throne here (apparently hexagonal in plan and without steps) is anticipated by the lowest portion of the pedestal in the Felicini Altarpiece and is a motif that Francia, as a goldsmith, would have noted in Roman altars and the lower registers of candelabra.

The polychromy in Francia’s earlier architectural settings, generally involving combinations of marble with bronze that was gilded or patinated green, is now replaced with a less distracting grey and grey‐green. The very minimal articulation of the walls, with the frieze indicated by a simple moulding, is also notable.¹⁸ There are two monolithic columns of marble to either side of the foreground. They have no relation to the rest of the architecture and were perhaps inserted only to provide something to which Sebastian can be tied. The irregular, tawny horizontal banding is more like the bark of a cherry tree than any familiar pattern in stone.

Because the architecture has been reduced to quiet ‘background music’, its implausible character – especially the absurd meeting of the arches with the side walls – and the oddity of a throne with no back are seldom remarked upon.

Attribution and dating

Although the altarpiece by Francia is signed, and although it was mentioned by Vasari in the 1550 and 1568 editions of his Lives,¹⁹ Francia’s work was little known in Tuscany even by experts during the seventeenth and eighteenth centuries and the signature was mistaken as an abbreviation for the Florentine painter Franciabigio. The altarpiece was published as such by Vincenzo Marchiò in his guidebook to the city of 1721 and was recorded as such by several earlier and later manuscript sources.²⁰ By the close of the eighteenth century Francia was better known as an artist and in a list of 1778 this ‘most beautiful and precious’ work was correctly attributed to ‘Francesco Francia Bolognese’.²¹ The removal of the altarpiece from the Buonvisi chapel to the family picture gallery followed soon after. In the modern literature concerning Francia almost all scholars agree in considering this obviously large and prestigious commission as executed by the artist as well as signed by him.²² But there has been less agreement about its date.

It is clear from the wills dictated by the patron Benedetto Buonvisi (see Appendix 3) that Francia had received the commission by August 1510 and that by June 1511 the altarpiece had still not been completed but was expected. It is possible that Francia travelled to Lucca to oversee the installation of the altarpiece and that, while there, he received another major commission for the same church: the Immaculate Conception for the Stiatta chapel, which seems to have been settled upon between May 1511 and August 1512.²³ Even if he never travelled to Lucca, it seems unlikely that, had the first commission not been completed promptly, the second would have been so readily given. Moreover, delays would have prompted notarial activity and of this there is no trace. Thus there seems no reason to doubt that NG 179–180 was painted between 1510 and 1512.²⁴ This hypothesis is supported by the very similar treatment of the infant Christ and the identical facial type given to the Virgin in a painting by Francia and his workshop which is dated 1513.²⁵

The choice of an artist from Bologna

It is unlikely to be a coincidence that another artist from Bologna, Amico Aspertini, had been painting in fresco in the chapel of Saint Augustine in the same church of S. Frediano in 1508 and 1509, invited there by Pasquino Cenami, prior of [page 175]the community from 1505 to 1510, who had many contacts in Bologna. Either Aspertini or the prior or others in the circle of the prior (see Appendix 2) may also have suggested Francia. His paintings have a purified monumentality and austere decorum like that of Fra Bartolommeo, who had worked with great success for the cathedral of Lucca in 1509 but was, after 1510, completely absorbed by his work for the Florentine Republic.²⁶ It is indeed not impossible that Fra Bartolommeo himself had commended Francia’s work, which he is likely to have encountered in his travels (he was in Venice in 1508 and would probably have travelled there via Bologna). And if the Saint Benedict in this painting is indeed a portrait, then it is likely that Francia was supplied with a drawing – in which case that drawing may have been made by Fra Bartolommeo. Although encumbered with commissions elsewhere, Fra Bartolommeo retained close links with Lucca, presenting to the convent of S. Romano there in 1513 the altarpiece he had intended for S. Pietro Martire in Murano (dated 1508).²⁷

In his will of 1510 Benedetto Buonvisi refers to the artist as ‘a certain painter called Francia’ (‘cuidam pictori cui vocatur Francia’), a formula which strongly suggests that he had no personal acquaintance with him.²⁸ It may therefore support the supposition that the painting was not made in Lucca. As already mentioned, Francia was also commissioned to paint an altarpiece of the Immaculate Conception for the Stiatta chapel in S. Frediano.²⁹ This work follows a pattern established by the altarpiece of the same subject executed by Vincenzo Frediani (active 1481–1505) in 1502 for the church of S. Francesco (now in the Museo Nazionale di Villa Guinigi, Lucca). The fact that Francia’s altarpiece for the Stiatta chapel is based on a local model is not evidence that he travelled to Lucca, as he could have been provided with a sketch of this picture by his patrons,³⁰ but he may have done so – if only to supervise the installation of the Buonvisi Altarpiece.³¹

The patron

Benedetto Buonvisi was born in Lucca around the mid‐fifteenth century. Together with his brothers, Antonio, Ludovico and Paolo, he managed the family’s international banking business and its trading interests, especially in silk.³² The family’s pre‐eminence in local politics is demonstrated by the fact that Benedetto and Paolo were elected as senators (anziani) or counsellors (gonfalonieri) more than forty times over a period of 45 years, from 1468 to 1513. Furthermore, two of his brothers were members of the Consiglio Generale della Repubblica and of the Collegio dei Mercanti, and as such represented the city on diplomatic missions. Benedetto had strong ties with the Medici and when he sent a letter of condolence to Piero de’ Medici on the death of his father, Lorenzo il Magnifico (1449 –1492), he spoke of the latter as a ‘master and father’, on whom he had always relied. Dignitaries and ambassadors visiting Lucca were received by Benedetto in his house in Borgo S. Frediano. The Buonvisi had offices in Naples, Venice, Genoa, Lyons, Tours, Louvain, Antwerp and London. Antonio Buonvisi became the banker to Henry VIII and was an intimate friend of Thomas More.

The Buonvisi chapel in S. Frediano

NG 179 and 180 were commissioned by Benedetto Buonvisi for his family chapel in S. Frediano in Lucca. According to tradition this church had been founded in the late sixth century by the Irish priest Finnian (also called Fridianus, died 588), son of Ultach, King of Ulster. From the mid‐eleventh century it was governed by the Canons Regular of Saint Augustine. In the second half of the twelfth century it became one of the most important centres for the diffusion of the reformed Gregorian liturgy. Between the end of the fifteenth century and the early sixteenth the church underwent a major campaign of refurbishment: aisles were vaulted, new chapels were built, and a new chancel and choir stalls created.³³ These works were carried out mainly by Matteo Civitali (1436 –1501) and his nephew Masseo (died 1517/18) and their assistants, at the instigation of the prior, Pasquino Cenami, with whom the Buonvisi were closely associated: indeed, Benedetto was married to Filippa di Martino Cenami, and he was among the powerful laymen present at Pasquino’s election.³⁴

In wills dictated in 1488 and 1497, Benedetto Buonvisi expressed his desire to be buried in S. Frediano, ‘in the burial vault of his ancestors’.³⁵ In the second of these wills he made arrangements for a chapel to be built in the church, charging his brother Paolo, as executor, with responsibility for the construction of the chapel and its fittings, as well as for arranging for masses to be celebrated there.³⁶ Furthermore, Benedetto wanted his own mortal remains and those of his deceased brother Ludovico to be transferred to the chapel’s vault.³⁷ The same stipulations appear in the will dictated by Benedetto on 19 August 1506.³⁸ From yet another will, of 16 August 1510, it is clear that Benedetto himself had overseen the erection of the chapel and commissioned appropriate decoration and fittings.³⁹ On 11 June 1511, in the presence of the prior and canons, Benedetto endowed the chapel with properties in Matraia and Lucca.⁴⁰ This document states that the chapel, dedicated to Saint Anne, had been erected in the southern aisle, by the facade (‘in calce dicte ecclesie versus meridiem’), and that the altarpiece and other ornaments would be installed when they were complete.⁴¹ The patronage of the chapel was to belong to Benedetto and Paolo Buonvisi and then pass to their heirs.

The names of the craftsmen commissioned to undertake the decoration of the chapel are mentioned in the will of 16 August 1510.⁴² The patron required that the completed works should be assessed by experts.⁴³ The altarpiece (‘anconam’) by Francia and the ‘wooden ornament of the image’ (‘lignamine ornamenta ymaginis’) – that is, the frame carved and gilded by Ambrogio di Biagio – were being made at the same time (the painting, presumably, in Bologna and the frame in Lucca).

When the Buonvisi chapel was renovated at the end of the eighteenth century the original furnishings were lost, with the exceptions of NG 179–180, which were taken to the Buonvisi palace, and the two ledger stones that indicate the entrances to the vault. Rather touchingly, a framed photograph of the National Gallery’s painting donated by the theologian and art historian Montgomery Carmichael was placed on the chapel [page 176] around 1900 and remains there to this day.⁴⁴ Carmichael (who was British consul in Leghorn) must have hoped by this means to encourage visitors, and especially British travellers, to reflect on the original purpose and placement of this great altarpiece, now removed from a religious context, for we know that he was dismayed by the way art lovers, and indeed expert connoisseurs, ignored the religious content and function of paintings they studied in galleries.

Previous owners: from the Buonvisi chapel to the Buonvisi palace

The Buonvisi family continued to prosper throughout the seventeenth century, when there were three Buonvisi cardinals.⁴⁵ By 1768 the family palace near S. Frediano housed a collection of 185 pictures.⁴⁶ Later in the eighteenth century Francia’s altarpiece was moved there and shown as two paintings, doubtless therefore re‐framed and probably cleaned.⁴⁷ The move presumably anticipated the arrival of a new altarpiece, Saint Anne adoring the Infant Christ (fig. 2), sent to Lucca from Rome by the leading Lucchese painter of the day, Stefano Tofanelli (1752–1812).⁴⁸ By 1796 this very beautiful painting, worthy of a follower of Pompeo Batoni (1708–1787) and Anton Raphael Mengs (1728–1779), and deeply indebted to Raphael, was on the new altar in the superb neoclassical chapel designed by the leading local architect Giovanni Lazzarini (1769–1834). It is hard to be sure whether the prevailing motive was to refurbish the chapel in a modern style, employing the finest local talent, or to remove to a picture gallery a work which had become more valuable as an old master painting than effective as a devotional aid, but in any case the move reflected the increasing esteem for Francia. The altarpiece is flanked by two other fine modern works by local artists: one by Bernardino Nocchi, of the ‘Transition’ of Saint Anne, and the other by Francesco Cecchi (1717–1812), of the Birth of the Virgin.⁴⁹

In 1800, shortly after this outstanding act of artistic patronage, Conte Francesco Gerolamo, the last remaining male member of the Buonvisi family, died. There must always have been a strong element of dynastic retrospection in his creation of this chapel. It preserved the fine seventeenth‐century marble tomb of Cardinal Buonviso Buonvisi and included busts of the other two family cardinals placed in roundels in the lunettes above the two side paintings. The family property, and indeed responsibility for the chapel, passed to the nephew of Count Francesco Gerolamo, Francesco Montecatini and his heirs, who, for a while, added the name Buonvisi to their own.⁵⁰ Nocchi’s painting was not completed before 1804;⁵¹ Cecchi’s may also have been completed after 1800. A stucco frame identical to those around these works, and perhaps intended for a fourth painting, surrounds a monument to Montecatini and his wife, Mariana Santini, on the side wall opposite that of the cardinal.⁵²

An account of the formation and dispersal of the Buonvisi collection is given in an appendix here (pp. 460–2). It was the most important in the city of Lucca and had been greatly enhanced by Conte Francesco Gerolamo. In February 1819 we know that the collection consisted of [page 177]216 paintings.⁵³ After prolonged negotiations a group of the Buonvisi pictures, including NG 179–180, were sold, in the winter of 1823 or soon afterwards, to the Duchess of Lucca, who was afflicted with cancer and died in March of the following year.⁵⁴ Her son inherited and enlarged the collection, housed in the royal palace (acquiring more paintings from the Buonvisi), but then felt obliged to sell. (For a full account of this collection see pp. 473–8.)

Sale to the National Gallery: Galvani, Buchanon Buchanan and Flight

In 1840 the Duke of Lucca’s Venetian agent, Carlo Galvani (1805–1866), tried to sell the Francias to the National Gallery, together with a painting by Domenichino (1581–1641; Appendix 3 below, documents 7, 8). In July 1840 the Francias were included in the exhibition of the duke’s collection at the Society of Painters in Water Colours, together with 13 other paintings formerly in the Buonvisi collection.⁵⁵ In the same year both Francias were bought by E.G. Flight, from whom they were purchased by the Trustees of the National Gallery through William Buchanan (1777–1864) in June 1841 for £3,500.⁵⁶ The Trustees assumed that Flight was an agent for the Duke of Lucca (the paintings were described as ‘at present under your charge’) and the price was one which they had offered for the paintings through Buchanan in the previous year. The purchase was therefore recorded in the Gallery’s catalogue as from the duke.

Many years later, early in 1860, Flight wrote indignantly to the Director, Sir Charles Eastlake, asserting that the Francias had been ‘bought by myself and sold by me and delivered by me’. Eastlake must have explained that it had been assumed that he had been an agent because another letter from Flight followed, objecting that he had never been anyone’s agent but had bought the paintings for £3,200 ‘at my own risk and as my own speculation’ nearly a year before the sale to the National Gallery.⁵⁷ It is hard to explain this episode other than as evidence of Buchanan’s dishonesty. He cannot have communicated the offer made by the Trustees in 1840 of £3,500. Knowing that Galvani would accept £300 less, he must then have arranged for Flight to buy them at that price, speculating that he could, after a brief interval, sell to the Trustees at the price they had previously offered. Whether Buchanan took a cut or was doing a favour for one of his backers will probably never be known.

This acquisition, together with that of a Perugino (NG 181) acquired soon afterwards, doubtless reflects the dawning recognition by the Trustees that it would be valuable for the national collection to include work by the antecedents of the mature Raphael.

Provenance

See above. On the altar of the Buonvisi chapel in S. Frediano, Lucca (recorded there until 1721). Moved to the Buonvisi palace, Lucca, after 1768, probably in the 1790s. Sold by Francesco Montecatini (son of Nicolao Montecatini and Maria Caterina Buonvisi) to Maria Luisa di Borbone, Duchess of Lucca, in late1823 or early 1824. On her death in 1824 the paintings passed to her son, whose collection was sold in London between 1840 and 1841. Purchased in 1840 by E.G. Flight, from whom purchased in June 1841 by the Trustees of the National Gallery.

Copies

In Room 36 of the piano nobile of the Palazzo Ducale, Lucca, according to an inventory compiled on 15 November 1834, there was ‘Un quadro in tela con cornice dorata del Martini dall’originale del Francia rappresentante la Madonna con vari Santi’. Given that there were elsewhere in the palace copies of other great altarpieces in the duke’s collection, this was almost certainly a copy of the main panel of the altarpiece catalogued here.⁵⁸

Nineteenth‐century catalogues of the collection cite an ‘old copy’ of Francia’s altarpiece in the ‘Gallery of Berlin’.⁵⁹ This must refer to another version of the Pietà made by Francia or his workshop and acquired from the sacristy of the Badia in Bologna for the Gemäldegalerie in Berlin.⁶⁰ It is modelled on NG 180 but with the addition of another Mary and with the Virgin’s legs visible.

A reduced copy of NG 179 (canvas on panel, 74.9 × 64.8 cm) was sold as lot 28 at Christie’s on 13 November 1959 for 95 guineas. Another, also canvas on panel, was lot 165 in Sotheby’s, Sussex, on 22 November 1991 (51.5 × 47 cm). Many more copies are recorded of the lunette, NG 180, and some of these were made to furnish modern Anglican churches, which reveals the intense admiration felt for the piety and decorum of Francia’s art – and also reflects the fact that the lunette was not displayed as it is now, which would have made it impossible to copy exactly without a scaffold.

The copy of the Pietà in the church of the Holy Trinity at Coleman’s Hatch, East Sussex, built in 1912–13 and paid for by shipping magnate John McAndrew (1838–1925), had been made by the donor’s sister and installed in the church shortly after its dedication.⁶¹ Another copy was in the church of St Barnabas in Tunbridge Wells. This is probably the picture (canvas, 98 × 184 cm) that appeared at auction in London in 1992.⁶² Both this copy and the one in Coleman’s Hatch reproduce the painted spandrels, which are nowadays covered by the frame of the altarpiece. A further copy of the Buonvisi lunette, but reduced in size (canvas, 63.5 × 121.9 cm), featured in the Phillips sale in London on 16 February 1993 (lot 25).

Framing, display and reputation

It is possible that the original frame was far more ornamental than the paintings within it: in 1822 Tommaso Trenta described the Buonvisi altar as having had ‘many carvings’ (‘con diversi intagli’).⁶³ He also believed it to be the work of both Masseo Civitali and Ambrogio Pucci. According to a memorandum by Civitali’s son, Giuseppe (1511–1574), his father had died while working on the Buonvisi altar, which was then completed by Ambrogio Pucci.⁶⁴ However, it is possible that Giuseppe was referring to another altarpiece frame because Masseo died sometime between 4 November 1517 and 6 November 1518, and it seems unlikely that by [page 178] that time he would not have completed a commission he had received in 1510 from one of the most powerful patrons in Lucca. Furthermore, it is clear from Benedetto Buonvisi’s will that Ambrogio had received the commission to carve the frame and that it had not merely been transferred to him. At the same time some co‐ordination of design would be expected and some degree of collaboration is probable.

Although no frame known to be by Ambrogio Pucci has been identified, a rare example of surviving carving by Masseo and his collaborators, the altarpiece in the church of SS. Sisto e Margherita in Villa Collemandina (Lucca), painted between 1514 and 1519, may give an idea of the appearance of the original frame of the Buonvisi Altarpiece.⁶⁵

We know that the two paintings were framed separately in the gallery of the Buonvisi palace and then in the Palazzo Ducale of Lucca – and in the latter palace they also occupied different rooms both in 1829 and in 1834.⁶⁶ The paintings were also for many years separately framed in the National Gallery: the lunette with its added painted spandrels was displayed beside the main panel rather than above it.⁶⁷ This may surprise us but Eastlake, when considering the acquisition of the great Costa Altarpiece, thought it best not to acquire the original lunette because it ‘could not again be placed above the picture’.⁶⁸ When acquired, it seems likely that the paintings retained their Lucchese frames. The Trustees resolved to have the paintings ‘placed under glass’ in 1854, which was undertaken for NG 179 in January and for NG 180 in October by Critchfield. The documentation indicates cases placed over the frames.⁶⁹

It is not clear exactly when the two paintings were reassembled in the altarpiece frame that encloses them today (fig. 8). It must have been after 1887 when the French curator, connoisseur and collector Frédéric Reiset recommended that it be done.⁷⁰ A probable date would be 1891, when the main panel was transferred to canvas. It seems to be one of a group of altarpiece frames made for the National Gallery around that date by the firm of Dolman, probably at the instigation, perhaps to the specification, and certainly at the suggestion of Sir Frederic Burton, [page 179]Director from 1874 to 1894. This initiative attracted much attention. It was an acknowledgment that altarpieces came from a different context: a way of recalling their ecclesiastical origin without disrupting the gallery setting with an attempt at reconstruction.

The capitals are similar in design to those found on the frame for the great altarpiece by Costa (NG 1119; see p. 83) and the frieze is identical. The elongated and fluted pilasters are, however, unusual among the frames of this type and must be a response to the austerity of Francia’s style. The frame was clearly designed for glass and retains evidence of how the glass was fitted. In this case the artist’s name and the painting’s title and number are painted in gold letters in the plinth between the pedestals of the pilasters. The name of Francia, with ‘Raibolini’ in parentheses, reflects Burton’s eccentric attempt to depart from conventional usage in his new catalogue of the collection (Francia was found there under ‘Raibolini’).⁷¹ The designation of the regional school (Bolognese), also given on the frame, was a notable feature of the revised display arrangements made by Burton after completion of the Central Hall in 1887.⁷² Early photographs do not show the wooden spandrels which now conceal the neoclassical painted ones and these may have only been added after the paintings were cleaned in 1973. The toning of the gilding has greatly darkened.

By 1920 the paintings, reunited in Burton’s frame, occupied the north‐west wall of the Octagon in the centre of the ‘Barry Rooms’ (Room 39; fig. 7) as a sort of pendant with the Costa on the north‐east wall. From the south these altarpieces framed a view of Raphael’s Ansidei Madonna placed in Room 37 (then VIII) on a screen between the columns, which originally flanked (and again today flank) the entrance to Room 32 (then VII). This was part of Charles Holmes’s rearrangement of the collection after the First World War: the ‘sober illumination’, as well as the somewhat darkened splendour, recalled old ecclesiastical interiors and would, he hoped, help visitors ‘to understand more clearly the early religious pictures which at first sight may seem too brilliantly coloured, or too lavishly gilded’.⁷³

A curious paradox may be observed here: both Burton’s initiative and Holmes’s, by restablishing something of the painting’s original character as an altarpiece, made it harder to appreciate – and, above all, far harder to to admire (or copy) the lunette, which many ‘English lovers of paintings’ considered to have ‘no equal in the whole range of Italian art’.⁷⁴ Neither shared the very high general esteem for Francia. And in Holmes’s faint praise we may recognise the dawn of the indifference and distaste now commonplace among art lovers. Of the altarpiece he asks: ‘does it leave any definite impression upon us, other than that of something well designed, well coloured and generally imposing?’ He concedes that the lunette is ‘a masterpiece of formal arrangement’, ‘the colour … as rich and broad as could well be desired’, the ‘expression of sorrow’ such as has affected ‘generation after generation’, but, all the same, it somehow lacks ‘profundity of art’.⁷⁵ What had seemed a refined ideal is now found to be bland.

Appendix 1
Saint Anne and the Virgin during the Renaissance

Several ways of representing Saint Anne and her relationship with her daughter, the Virgin Mary, coexisted in the early sixteenth century. The oldest convention in Western Europe seems to have been for Saint Anne to be represented seated, with the Virgin, smaller in scale, on her lap, and the infant Christ on the Virgin’s lap. This hieratic and frontal arrangement must have been inspired by the convention for representing the Trinity with God the Father seated behind the Crucified Christ, smaller in scale, and the Holy Spirit. The image was known as ‘Saint Anne Trinitaire’, ‘Anna Selbdritt’ and ‘Sant’Anna metterza’.⁷⁶

A move towards naturalism may be observed in the altarpiece of about 1424 by Masaccio (1401–1428) and Masolino (1383–1447). Here the Virgin is the same size as her mother and on a separate, lower throne, so she is seated between the knees of Saint Anne rather than on her lap.⁷⁷ It is, however, not the case that the discrepancy of scale was always avoided thereafter. A convention whereby Saint Anne was depicted standing holding a miniature Virgin and Child even enjoyed some popularity around 1500 and, in Florence, where Masaccio and Masolino’s solution must have been well known, Fra Bartolommeo in his huge altarpiece (the Pala del Gran Consiglio) for Palazzo Vecchio (begun in 1510) borrowed the idea of the double throne but made Saint Anne distinctly larger than her daughter.⁷⁸ The same formula was used thirty years later in SS. Annunziata by Antonio di Donnino (1497–1547).⁷⁹ Francia was highly original in making his Virgin Mary larger than her mother, thereby indicating the effect of age as well as Saint Anne’s lesser theological eminence.

A more popular variant retained the idea of the Virgin seated on her mother’s lap but reduced or discarded the difference in scale and the frontality, allowing much more physical and psychological interaction between the three figures. Leonardo’s painting of Saint Anne, now in the Louvre, and his previous designs for the subject, of which the cartoon in the National Gallery (NG 6337) is the most famous example, belong to this category.⁸⁰ The altarpiece painted in Verona by Girolamo dai Libri (1474 –1555), of 1510–18 (NG 748), is another example.⁸¹

A third variant was for the Virgin Mary and Saint Anne to be seated side by side, as here in Francia’s altarpiece. This convention was the preferred one for sculpture and paintings that included other members of the Virgin’s family (and all three of Saint Anne’s husbands). Many such were made, especially north of the Alps, in the later fifteenth and early sixteenth centuries.⁸² In these side‐by‐side groups, whether or not the ‘holy kinship’ is included, Saint Anne commonly touches the infant Christ, offers him a fruit and exchanges looks with him – interaction of a kind that tended to be somewhat contorted when the Virgin was perched on her mother’s lap. The most celebrated and most venerated image in this category was certainly the marble group by Andrea Sansovino (1467–1529) completed by 1512 for the church of S. Agostino in Rome.⁸³ We might suppose that this [page 180] sculpture was an influence on Francia, but the date makes this unlikely and Sansovino still represents Saint Anne at a larger scale than the Virgin. If Francia had a model in mind, or was given one by his patron, it may, rather, have been a smaller sculptural group, many of which, carved in Germany, made no distinction of this kind.⁸⁴

It seems likely that local tradition and theological positions played some part in the choice of convention. As far as local tradition is concerned, we might suppose that patrons in Lucca would be inclined to follow Florentine examples established by Leonardo or Fra Bartolommeo, but Francia did not do so.⁸⁵ As for theology, the doctrine of the Immaculate Conception, whereby the Virgin was held to be born free of sin, was much debated in this period, but there was no question of the Dominican opponents of this doctrine seeking to diminish the sanctity of the Virgin’s mother, and they may well have been keen to assert their veneration for her (Fra Bartolommeo was of course a Dominican friar).

In representations of Saint Anne in the eighteenth century great ingenuity could be shown in preserving the superior stature of Saint Anne and at the same time not making the incongruity too disturbing. Tofanelli’s replacement altarpiece now in the Buonvisi chapel, in which the Virgin stands before her seated mother, is a good example of this (figs 5, 6).

Appendix 2
Cultural exchanges between Bologna and Lucca

So widely had the fame and the works of this craftsman spread throughout Lombardy, that even from Tuscany men sent for something by his hand, as they did from Lucca, whither there went a panel containing a S. Anne and a Madonna, with many other figures, and a Dead Christ above in the lap of His Mother; which work is set up in the Church of S. Fridiano, and is held in great price by the people of Lucca.⁸⁶This passage in Vasari’s Lives has been much glossed by modern scholars. As mentioned above, it seems likely that Francia was recommended through the same channels that only a few years earlier had brought Aspertini to Lucca. The relationships between humanists in Lucca and Bologna and the connections between their Augustianian foundations have been explored by Maurizia Tazartes in several publications.⁸⁷ Among the most active individuals in the cultural exchanges between the two towns was the Lucchese doctor and mathematician Leonardo Richi, who taught at Bologna University in 1497–8 and was a friend of the Bolognese humanist Giovanni Filoteo Achillini (1466–1538; known to both Francia and Aspertini). In his Viridiario, a collection of eulogies, Achillini mentions Richi as well as a member of the Trenta family and a ‘Medico’, most likely the internationally famous jurist and diplomat Giovan Marco de’ Medici, a close friend and political [page 181]associate of Benedetto Buonvisi.⁸⁸ It is clear that these three members of the Lucchese elite, all of them closely related to Pasquino Cenami and S. Frediano, had connections with Bolognese scholars and artists. Another intermediary might have been Girolamo di Paolo Buonvisi, nephew of Benedetto, employed by Pope Julius II as ambassador and banker, who was in Bologna from 1508 and probably knew Francia. But the key figure in the relationship between the two cities was surely Pasquino Cenami, who was appointed prior of the convent of S. Frediano in 1505.⁸⁹ He was born in Lunigiana (a small hilly region in northern Tuscany, near the border with Liguria), took a degree in theology at Bologna, and was prior of the convent of S. Frediano dei Sacchi in Bologna from 1482 to 1489.⁹⁰ After his return to Lucca he kept in touch with his compatriots in Bologna and with the Augustinians of S. Giovanni in Monte and S. Giacomo. It has been argued that Aspertini was in Lucca sometime before starting his work in the chapel of S. Agostino (1508–9), perhaps even before 1506.⁹¹

Appendix 3
Documents

1 Will of Benedetto Buonvisi, 2 September 1488

2 Will of Benedetto Buonvisi, 17 September 1497

3 Codicil to the will of Benedetto Buonvisi, 27 May 1500

4 Will of Benedetto Buonvisi, 19 August 1506

5 Will of Benedetto Buonvisi, 16 August 1510

6 Document pertaining to the Buonvisi chapel in S. Frediano, 11 June 1511

7 Letter from Carlo Galvani in response to an official enquiry from the National Gallery, dated 21 July 1840

8 Letter from Carlo Galvani, undated, but July 1840, in response to the Duke of Sutherland (Chairman of the Trustees)

Notes