As a boy in the 1490s, the future novelist Matteo Bandello was a novice monk at the Dominican monastery of Santa Maria delle Grazie in Milan, where his uncle Vincenzo was the prior. There he would pass the time watching Leonardo da Vinci at work on the north wall of the refectory, painting the great masterpiece of the Sforza era, the Cenacolo or Last Supper.
He would arrive early, climb up on to the scaffolding, and set to work. Sometimes he stayed there from dawn to sunset, never once laying down his brush, forgetting to eat and drink, painting without pause. At other times he would go for two, three or four days without touching his brush, but spending several hours a day in front of the work, his arms folded, examining and criticizing the figures to himself. I also saw him, driven by some sudden urge, at midday, when the sun was at its height, leaving the Corte Vecchia, where he was working on his marvellous clay horse, to come straight to Santa Maria delle Grazie, without seeking shade, and clamber up on to the scaffolding, pick up a brush, put in one or two strokes, and then go away again.83
Bandello was writing this decades after the event. There is some telescoping: Leonardo probably began painting the Last Supper in 1495, and was not therefore ‘working on’ the clay horse – exhibited in late 1493 – at the same time. But this is none the less an authentic glimpse of the maestro at work. It expresses his creative rhythm, the bursts of strenuousness interspersed by those puzzling spells of silent cogitation which others – particularly those who were paying him – tended to mistake for dreamy inactivity. And it gives us that marvellous image of Leonardo striding through the streets, under the noonday sun, not thinking of comfort or shelter or indeed anything else except a suddenly achieved solution to some tiny problem of compositional detail. Those ‘one or two strokes’ remind us of the painstakingly cumulative nature of his art. That massive sweep of visual narrative which one sees on the wall of the Grazie is made up of thousands of tiny brush-strokes, thousands of microscopic decisions. The familiarity of a world-famous painting makes it seem somehow inevitable – how could it be other than it is? – but every inch has been fought for.
The church of Santa Maria delle Grazie lay beyond the old Porta Vercellina to the west of the castle. For some years (you might have heard it muttered) it had been a virtual building-site. In 1492 the choir and apse were demolished to make way for a new tribune and cupola, designed by Bramante, and in tandem with this it was decided to enlarge the contiguous monastery buildings. The renovation of the refectory was certainly completed by 1495 – this is the date inscribed on Donato di Montorfano’s Crucifixion fresco which adorns the south wall. The Last Supper, on the opposite wall, was probably begun in this year. This whole renovation programme was commissioned and paid for by Ludovico, who intended the new setting to serve as a future Sforza mausoleum, a fitting monument to the ducal dynasty. This motivation became more pressing with the sudden deaths of his wife, Beatrice, and his daughter Bianca in 1497. These losses weighed heavily on the Moor, and introduced a period of sombre religiosity: his investment in the Grazie became emotional as well as financial, and he often took his meals at the refectory. Leonardo’s great mural – it is not technically a fresco, since it was painted in oils – was thus the pièce de résistance of a prestigious Sforza project: a thoroughly modern work to crown this elegant modernization.84
Studies for the Last Supper. Above: early compositional sketches on a sheet at Windsor. Below: studies for the heads of Judas (left) and St James the Elder.
To trace the making of ‘this restless masterpiece’, as Burckhardt called it, one must first turn to a sheet at Windsor which has an early compositional study in pen and ink.85 It is still rooted in traditional Last Supper iconography – Judas in the excluded position, seen from behind, on the near side of the table, and St John shown asleep next to Christ, a reference to his ‘leaning on Jesus’s bosom’ when the announcement of betrayal was made. Both these figures would be jettisoned in the final version.
There are two separate drawings on the sheet. The left-hand sketch has ten figures: the sheet has probably been trimmed, losing the remaining three. There are arches lightly sketched behind the group – first thoughts for the background of the painting, the ‘upper room’ in which the supper took place. The right-hand sketch shows four figures, but is essentially a study of Christ and Judas. Here Leonardo is focusing on the dramatic moment of identification: ‘He that dippeth his hand with me in the dish, the same shall betray me’ (Matthew 26:23). Judas has climbed off his stool and is reaching forward with his hand to the dish. Christ’s hand is tried out in two positions – raised as if about to reach forward; and having reached the dish and hovering in a moment of poignant contact with the hand of his betrayer. This smaller sketch has intensified the focus, found the dramatic fulcrum – that ghostly moment of contact. In focusing on this moment, Leonardo is already moving the story back a few frames from the more conventional Last Supper depiction showing the institution of the Eucharist.
The other constituent of this smaller sketch is the sleeping figure of St John with Jesus’s arm resting on his back: a note of tenderness which is scriptural – John is the disciple ‘whom Jesus loved’ – but in sceptical irreligious circles John ‘leaning on Jesus’s bosom’ was interpreted in terms of homosexuality. Among the blasphemies attributed to Christopher Marlowe a hundred years later was that Christ loved St John ‘with an extraordinary love’, and ‘used him as the sinners of Sodoma’. One remembers the Saltarelli episode, with its subtext of official disapproval of effeminate young models being used to depict angels and boyish Christs. Leonardo separated the figures in the final painting, though of all the disciples John remains the youngest and most beautiful.
A little later, perhaps, is the sketch at the Accademia in Venice, its red chalk gone over in ink by another hand.86 It seems cruder, largely because of the inking, but something of the compositional rhythm of the painting is emerging. The disciples are forming into groups; there is more emphasis on their individual features; they are identified in hurriedly written captions (Philip being mentioned twice). But Judas is still on our side of the table, and John is still slumped asleep.
These are glimpses into the early workings of Leonardo’s conception of the painting: miniature blueprints, made with rapidity and concentration and a questioning temper – this way or this? But, as so often with Leonardo, the roots of the painting run deeper, and, though these are the first actual studies for the Grazie Last Supper, we find in the reservoir of his sketchbooks a much earlier sheet, datable to c. 1480, with a trio of linked sketches: a group sitting at a table, a figure apart with his head in his hands, and a figure who is undeniably Christ, pointing with his finger at the fatal dish.87 These are not exactly studies for a Last Supper: the group is not the disciples, just five men at a table, passing the time in animated conversation – we might be at some village festivity, with men sitting round at trestle-tables. But something has triggered in Leonardo’s mind, which is realized in the swift poignant sketch of the Eucharistic Christ on the same sheet, and which grows to fruition in the great Milanese mural fifteen years later.
From the compositional studies in Windsor and Venice the focus moves in to the features of individual figures, and so we come to the famous series of heads in Windsor, mostly in red chalk, some highly finished. We see the characters emerging from a mist: Judas, Peter, St James the Elder, St Philip (the latter two almost certainly drawn from the same model, though in the painting they are individuated). There is a beautiful study for the hands of St John, and for the sleeve of St Peter.88 These studies are complemented by brief comments in the Forster notebooks – that a certain Alessandro from Parma provided the model for Christ’s hand; that ‘Cristofano da Castiglione, who lives at the Pietà, has a good head’. There is a note headed simply ‘crissto’, under which Leonardo writes, ‘Giovanni Conte, the one with the Cardinal of Mortaro’; this may tell us the name of the model for Christ. According to the well-informed Luigi of Aragon, who saw the painting in 1517, some of the disciples were ‘real portraits of Milanese courtiers and important citizens’.89
In a well-known passage, Leonardo lists some of the disciples’ reactions:
One who was drinking and has left the glass in its position and turned his head towards the speaker.
Another twisting the fingers of his hand together, turns with stern brow to his companion, and he with his hands spread shows the palms and shrugs up his shoulders to his ears, and makes a mouth of astonishment… Another who has turned, holding a knife in his hand, knocks over a glass on the table…
Another leans forward to see the speaker, shading his eyes with his hand.90
Some of these find their place in the finished painting – white-bearded St Andrew (third from left) shows his palms and shrugs up his shoulders. Others are transmuted, so that the man who turns with a knife in his hand (St Peter) is detached from the man who knocks over a glass, and the latter becomes a man (Judas) spilling a salt-cellar. At least one of these gestures is already enacted in that first compositional sketch in Windsor: in the smaller group, the figure between Christ and Judas ‘shades his eyes with his hand’.
In these emotional dynamics, as much as in compositional planning, lies the radically new conception of Leonardo’s Last Supper, breaking with the tradition inherited from the Middle Ages in which the disciples are a stiffly linear grouping ranged along the table. In Florence he would have seen versions by Taddeo Gaddi, Andrea del Castagno, Fra Angelico and Domenico Ghirlandaio.91 The latter’s elegant, almost languid fresco in the refectory of Ognissanti was completed shortly before Leonardo’s departure for Milan. In Leonardo’s version the line of diners is magically disrupted. We have instead a group whose outline is a kind of wave-formation, which Pietro Marani has compared to the optical diagrams of Paris MS C.92 The waves are formed of four sub-groups, each of three disciples: knots and huddles of men suddenly in crisis. Leonardo has also found his dramatic moment: not the institution of the Eucharist, nor the identification of Judas, but the first shell-shock of Christ’s announcement – ‘Verily I say until you, that one of you shall betray me. And they were exceeding sorrowful’ (Matthew 26:21–2). Thus the new fluency of the composition is in part the product of a narrative – almost a cinematic – decision: the moment that tells the story. This is well described by one of the first recorded commentators on the mural, Luca Pacioli. In the dedication of his Divina proportione, dated 14 December 1498, he writes:
One cannot imagine a keener attentiveness in the apostles at the sound of the voice of ineffable truth which says, ‘Unus vestrum me traditurus est.’ Through their deeds and gestures, they seem to be speaking among themselves, one man to another and he to yet another, afflicted with keen sense of wonder. Thus worthily our Leonardo created it with his delicate hand.93
Pacioli’s description is interesting because of his close connection with Leonardo at this time: it contains, perhaps, a refraction of Leonardo’s own statements – on the quality of ‘attentiveness’ and ‘wonder’ which racks the dramatic intensity of the focus on Christ, and the sense of interplay among the apostles. This is how the painting works: the figures not in a row, but intertwined, speaking ‘l’uno a l’altro e l’altro a l’uno’.
And then there is Judas: the villain of the piece, and yet in the preparatory profile study at Windsor (page 294) a man more ugly than evil – almost a grotesque, but with hints of remorse and self-disgust which touch the profile with tragedy, or indeed with Christian forgiveness. (The recent restoration of the painting has recovered subtleties in the faces lost beneath later retouchings; Judas is an example of a face now nearer to the preparatory drawing than it was before the restoration.) He recoils from the words of Christ even as his hand moves irrevocably towards the piece of bread he will dip in the dish.
Of the face of Judas in Leonardo’s Last Supper there is a well-known anecdote in Vasari: how the prior of the Grazie constantly badgered Leonardo ‘to hurry up and finish the work’, and complained of the artist’s dilatoriness to the Duke. In response Leonardo told Ludovico he was still searching for a face evil enough to represent Judas, but that if he did not succeed ‘he could always use the head of that tactless and impatient prior’ as a model. At this the Duke roared with laughter, and ‘the unfortunate prior retired in confusion to worry the labourers working in his garden.’ This is one of those Vasarian anecdotes that proves to have a kernel of truth, or at least of contemporary witness. The story is lifted from the Discorsi of Giambattista Giraldi Cintio, published in 1554, and Cintio in turn had it from his father, Cristoforo Giraldi, a Ferrarese diplomat who knew Leonardo personally in Milan. The Giraldi version of the story purports to be a record of Leonardo’s own words:
It remains for me to do the head of Judas, who was the great betrayer, as you all know, and so deserves to be painted with a face that expresses all his wickedness… And so for a year now, perhaps more, I have been going every day, morning and evening, down to the Borghetto, where all the base and ignoble characters live, most of them evil and wicked, in the hope that I will see a face which would be fit for this evil man. And to this day I have not found one… and if it turns out I cannot find one I will have to use the face of this reverend father, the prior.94
Whether or not the story is true, we are close here to an authentic recording of Leonardo. This is how Cristoforo Giraldi, who knew him, remembers or imagines him speaking: ‘Ogni giorno, sera e mattina, mi sono ridotto in Borghetto…’
The painting of the Last Supper began with the plastering of the refectory wall with an even layer of intonaco, which forms the structural base for the mural.95 The intonaco of the middle section – where the main action was to be painted – is coarser, to provide better adhesion for the paint-layers on top: the join between the two sections is perceptible as a faint horizontal line near the middle of the foreshortened ceiling. One of the discoveries to emerge from the latest restoration is the vestige of a sinopia, or outline drawing, done directly on to the plaster – ‘extremely concise red lines, executed freehand and with a fluid brush-stroke… to define the masses for his composition’. After this the gesso or ground was applied: modern analysis shows this to be ‘a slightly granular mixture, 100–200 microns thick, composed of calcium carbonate and magnesium with a proteinaceous binding agent’, and on top of that came a thin imprimatura of lead white. At this stage a number of incisions were made on the surface, mainly defining the form and perspective of the architectural setting, and – an eerie moment of precision – a small hole was punched in the centre of the pictorial area: the vanishing-point. This hole can be seen in a magnified photograph: it is at a point on the right temple of Christ.
All these preparations remind us that this was a collective studio work (a point missing from Bandello’s account, with its misleading sense of artistic solitude). Leonardo did not work alone on the painting, as Michelangelo reputedly did on the Sistine Chapel, but had a team of assistants – among them probably Marco d’Oggiono, who would produce one of the earliest copies of the mural; Salai, now about sixteen and working as a garzone; and Tommaso Masini, whose participation in a later large-scale mural (theAnghiarifresco in Florence) is documented. To these trusted assistants may be added the new intake of pupils and assistants whose names are found on two sheets of the Codex Atlanticus:96
Ioditti came on 8 September at 4 ducats a month.
Benedetto came on 17 October at 4 ducats a month.
The year referred to is either 1496 or 1497. Four ducats is the charge Leonardo exacts for their retta, or board and lodging; against this they can earn money for work they do as garzoni; thus by the end of the year Benedetto has earned nearly 39 lire – just under 10 ducats – which is approximately what he owed for his retta over that ten-week period. Benedetto’s name appears also on an undated sheet, partly cut off at the margin, which records the studio personnel at around the same time:
[… ] nco 4
[… ]iberdo 4
Gianmaria 4
Benedetto 4
Gianpetro 4
Salai 3
Bartolomeo 3
Girardo 4
The first name is probably ‘Franco’, and may be a reference to Francesco Galli, known as Il Napoletano; the fifth may refer to Giampietrino Rizzoli; and the penultimate, who pays the lower retta of 3 ducats, may be Bartolomeo Suardi, known as Il Bramantino, the pupil of Leonardo’s friend Bramante.
The painting itself probably began with the three heraldic lunettes above the depicted scene; they are now much damaged, but fragments of inscriptions and coats of arms, and a marvellous wreath of fruits and grasses, can still be seen. The painting of the central scene probably began on the left-hand side. We enter here the period described by Matteo Bandello – of intense work and arms-folded contemplation. Bandello’s description is borne out by technical data: ‘Leonardo’s slow progress is confirmed by a fewpentimenti[rethinks] and by the attentive refinement of important details… Each figure and each object on the table shows minor or significant revisions of outlines, which stray into the adjacent colours, testifying to the fact that Leonardo allowed himself great freedom in returning more than once to a given motif.’97 Among the pentimenti identified by the restorers is a modification of the position of Christ’s fingers, which were more extended in the original version.
In the summer of 1496, while he was at work on the Last Supper, Leonardo was also decorating certain rooms (camerini) – probably the apartments of Duchess Beatrice – in the Castello Sforzesco. This is the same job he mentions in the torn-up letter I quoted earlier: ‘Remember the commission to paint the rooms…’
On 8 June 1496 there was some kind of scene, a rare losing of his cool. It is noted by one of the Duke’s secretaries, who writes, ‘The painter who is decorating the camerini caused something of a scandal today, and for this reason he has left.’98 This tension is perhaps connected to another fragmentary draft letter to the Duke, in which Leonardo complains of financial strains: ‘It vexes me greatly that you should have found me in need, and… that my having to earn my living has forced me to interrupt the work and to attend to lesser matters instead of following up the work which your Lordship entrusted to me.’99 The important work is almost certainly the Last Supper, and the ‘lesser’ labours which have distracted him from it may be the decoration of the Duchess’s chambers.
The tone of this letter is very tetchy: consider the scarcely veiled sarcasm of the phrase ‘my having to earn a living’ – an inconvenience with which the Duke was unfamiliar. He continues: ‘Perhaps Your Excellency did not give further orders to Messer Gualtieri, believing that I had money enough… If your Lordship thought I had money, your Lordship was deceived.’ The reference is to Gualtiero Bascapé, elsewhere described as ducalis iudex dationum, the Duke’s judge of gifts, i.e. his dispenser of payments. It appears that some expected ‘gift’ had not been received: by ‘gifts’ are essentially meant payments provided in a form too irregular to be considered a salary. Accounts of Leonardo’s payment for the Last Supper vary wildly. According to Bandello he was on an annual salary of 2,000 ducats, but another well-informed source (Girolamo Bugati, a friar at the Grazie in the mid sixteenth century) says the Moor paid him only 500 ducats per annum.100 This would still compare well with the Virgin of the Rocks, for which Leonardo and Ambrogio de Predis had demanded 1,200 lire, or approximately 300 ducats.
In this griping letter, in this unexpected explosion of temper – ‘something of a scandal’ at the castle – we are privileged with a back-stage glimpse of Leonardo under the profound creative pressure of the Last Supper, which pressure is always exacerbated rather than relieved by diversions. It is the same Leonardo snapped by Bandello’s camera, striding grimly up the hot silent street towards the Grazie.
Bandello has another story, which shows Leonardo in more relaxed mood – chatting with a distinguished visitor to the Grazie, Cardinal Raymond Peraud, Bishop of Gurck. Documents confirm the Cardinal’s presence in Milan in late January 1497.101 Leonardo climbs down off the scaffolding to greet him. ‘They discoursed on many things,’ recalls Bandello, ‘and particularly on the excellence of painting, and some who were there said they wished they could see those paintings of antiquity that are so highly praised by great writers, so that they could judge whether the painters of our day could rank with the ancients.’ Leonardo also entertains the company with the picaresque story of the young Filippo Lippi being seized by ‘Saracens’ and held as a slave, and eventually winning his freedom because of his skill at drawing. This story is familiar from Vasari’s Life of Filippo Lippi.102 Two questions arise: Did Vasari take the story from Bandello? And did Bandello really first hear it from the lips of Leonardo da Vinci? The most one can say in each case is: possibly. Bandello’s Novelle were first published in Lucca in 1554 – four years after the first edition of the Lives – but they were certainly written earlier and were possibly available in manuscript. As for Leonardo, he could well have heard the story from Filippo’s son Filippino, whom he knew in Florence in the 1470s, and with whom he had cordial relations. It is equally possible that the opportunist Bandello was making a good story even better by spuriously ascribing it to Leonardo.
Leonardo was still at work on the Last Supper in summer 1497. There is an entry in the monastery’s ledger for that year, recording a payment of 37 lire to some workmen ‘for work done on a window in the refectory where Leonardo is painting the Apostles’. And on 29 June 1497 Ludovico writes a letter to his secretary, Marchesino Stanga, in the course of which he says that he hopes that ‘Leonardo the Florentine will soon finish the work he has begun in the refectory’, so that he can ‘attend to the other wall of the refectory’.103 A note of ducal impatience is perhaps discernible here.
As is well known, Leonardo’s great mural was innovative in less auspicious ways. He used a mix of oil and tempera for the painting, instead of the traditional buon fresco technique of painting on fresh plaster. This enabled him to work more slowly, and to repaint, but its disadvantages soon became apparent as the paint began to flake off. An inherent problem with damp exacerbated the situation. The deterioration of the paint surface was already visible during his lifetime. In 1517 the diarist Antonio de Beatis noted that the mural was ‘beginning to spoil’, and by the time Vasari saw it in the 1550s, there was ‘nothing visible except a muddle of blots’.104 This is doubtless a reason for the many early copies made of it, two of them Marco d’Oggiono’s and Giampietrino’s – by painters who were probably involved in the creation of the original. It is also the reason for the extensive and intrusive restoration projects; the earliest which is documented was done in the early eighteenth century, but was probably not the first. In the 1930s, comparing the current state of the mural with these early copies, and with the preparatory studies in Windsor, Kenneth Clark lamented the loss of subtle nuances of expression beneath the deadening hand of the restorer: he thought ‘the exaggerated grimacing types, with their flavour of Michelangelo’s Last Judgement’, suggested ‘a feeble mannerist of the sixteenth century’.105
The painting’s inherent, self-inflicted fragility seems now part of its magic. Reduced within a few decades to a ‘muddle of blots’, vandalized by Napoleonic soldiers in the early nineteenth century, and narrowly missed by Allied bombs in the summer of 1943, it is a miracle that it has survived at all.
The latest and most ambitious restoration, under Pinin Brambilla Barcelon, was unveiled in 1999, after more than twenty years’ work at an estimated cost of 20 billion lire (approximately £6 million). Much of it was targeted at removing the superimpositions of previous restorations: an encrustation of varnishes and over-paintings which were teased away, scab by microscopic scab, in the hope that some original pigment remained beneath. In the words of Brambilla Barcelon, the painting was treated ‘as if it were a great invalid’.106 The restoration had its critics, as always – it had ‘lost the soul’ of the original – but what we now see is much closer to what Leonardo and his assistants painted on that wall, watched by the wide-eyed boy Bandello, just over 500 years ago. Closer, but of course partial – only about 20 per cent of the original picture surface survives. The painting hovers like a ghost on the wall, vestigial yet ravishingly restocked with expressions and gestures, and with the simple yet compelling details of that last meal: the half-filled beakers of wine, the filigree weave on the tablecloth, the knife which in the emotion of the moment St Peter grips like a murder-weapon.