The other two paintings seen by Cardinal d’Aragona and his secretary, Antonio de Beatis, are difficult to date. One, the painting of Saint Anne with the Virgin and Christ Child, had been discussed by Isabella d’Este’s agent, Pietro da Novellara, in 1501, and Vasari would also mention it, although without having seen the painting: he only described the cartoon, thereby misleading generations of Leonardo scholars, who would regard the marvellous painting in the Louvre as the work of Leonardo’s pupils.19 Novellara also added his views on the meaning of the painting and there is little to add to his interpretation in theological terms – although this was the main and most communicable interpretation: behind it lie many other meanings, allusions and desires that prompted Leonardo to return to the painting almost twenty years later.
Although it was started in Florence at the turn of the new century, the painting followed Leonardo to Milan after 1506. Here it was again worked on, but not finished. Leonardo treated it, as usual, as a living being on which the transformations of light and colour were observed and discovered over time. The development of the painting up unil 1513, just before Leonardo’s departure for Rome, is witnessed by a number of copies, which then remained in Milan, made by pupils (they certainly used Leonardo’s cartoon). The most interesting copy from this point of view is the one that is now at the Hammer Museum of Los Angeles [Plate 60], and until 1810 had been in the church of San Celso, Milan, where it was copied continuously by many artists. This practice has given rise to a strange phenomenon of copying the copy, since it was located in a church open to the public while the original was inaccessible in the collections of the French monarchy.
As a result there are many more copies of the painting at its 1513 stage than there are of the painting that was finished by Leonardo between 1513 and 1519. A comparison of the two allows us to highlight the differences in pictorial quality between the work undertaken by pupils under his guidance and the final painting and to understand how Leonardo, in spite of having attained a perfect definition of the composition in 1513, insisted for another five years on a few small details (the differences) that, at first glance, seem unessential but that proved decisive to making the master’s painting a perfect masterpiece, while the painting by his pupils remains a most beautiful painting. The variations made by Leonardo during his Roman years concern the landscape – the tree on the left of the group disappeared – changes to the rear part of Mary’s robe and the removal of both her and Saint Anne’s sandals. Moreover, the colours of the robes are radically muted in order to give greater luminosity and depth to the scene.
In the 1513 version, the tree on the left encroaches on the two seated women, squeezing them into a narrow visual corridor that creates a sense of claustrophobia. In the following years Leonardo not only removed this tree to give more light to the scene and to lead the viewer’s eye to the magnificent backdrop of the landscape, but also opened the tree on the right by adding glimpses of the sky between the leaves and by making it as transparent as a stage wing made of lace. In the later version the landscape is divided cleanly between the upper part, all blue, where the mountains fade and merge in the distance with the water and the sky, and the lower part, which is misty, rocky and almost arid in its golden reflections. One of the most successful landscapes in Italian art, it has such immediacy that it seems to anticipate impressionist painting. With the lightest of tonal variations, Leonardo paints the snow-capped mountains, adding small touches of white on the tops and transparent layers [velature] of shade on the rocky flanks. The landscape that replaces the tree on the left has slightly different colours, and its later execution is evident. The blue of the mountains on the right, whose tectonic structure is defined almost with Quattrocento-like precision, makes away for inconsistent browns that look like seething froth rather than like eroding rock. Water runs through and penetrates the whole landscape, like a flood from which the two women and the child are saved, seated on a hill beyond which the world of nature is being changed before our very eyes. The plateau on which the scene is set is coloured, in marked contrast with the background – which consists of brown and golden ground deprived of any plants on the left, while on the right the landscape is less archaic and even shows signs of human presence: a small bridge, built over a river that runs between the rocks (but whose waters are white, not blue as in the distance). The spatial divide between the plateau where the women sit and the rest of the world could not be wider – were it not for the fact that, as in the first version of The Virgin of the Rocks, water flows also across the foreground, beside the naked feet of Saint Anne, in a clear stream that reveals the pebbles on its bed.
The other major transformation can be seen on the Virgin’s left-hand side, for which a study drawing has survived and is now in the Windsor collection [Plate 61]. This, too, might seem a trivial detail prompted by Leonardo’s maniacal obsession with minutiae, but this is not so: he realized that in the first version of the painting Mary’s body was tipped too far forwards and her movement to grasp the child lacked harmony and communicated a sense of precariousness to the viewer. By painting that large knot that holds up the folds of the roselake dress, the the artist rebalanced the whole figure – as if he added a counterweight to a scale that was off balance in order to bring the two pans into line.
This variation, on which the artist reflected for years, proves critical to giving the composition that sense of perfect mechanical equilibrium that makes it so fascinating. But the general rebalancing of the composition is also linked to the particularly pale colours used for the final version. The main colours are ultramarine blue for Mary’s mantle and laquer colour [color lacca] for her overdress. All that can be seen of Saint Anne is her arm covered in grey fabric, the same colour that is barely visible on her lap, on which the Virgin is sitting. The tones of these colours, and even of the lapis lazuli, which is generally so deep, are very pallid, consumed by light and reduced to little more than a vibration, in order not to overpower the delicate hues of the complexions caressed by shadows. While in Saint John Leonardo plunges the image into shadow, in Saint Anne he attempts to consume it in light, lightening the drama of the shadows so much that on both women’s necks they are reduced to a faint golden hue that in no way conceals the physiognomy.
Here again the drawing is invisible and the definition of the image is linked to a subtle variation in colour. Moreover, colour defines the age difference between the two women, who otherwise share the same abstract physiognomy that had by now become Leonardo’s norm. The viewer can tell that Saint Anne is older only by her darker skin, as if she had spent longer in the sun, but otherwise nothing detracts from the regularity of her features, which are completely identical to her daughter’s. In this delicate play of chiaroscuro vibrations, the garments are reduced to transparent veils, as if they, too, were made of shadow, albeit a coloured shadow. Mary’s overdress, painted using delicate pink lake pigment [lacca rosa], shades into white over her bosom, creating a revolutionary effect of insubstantiality. No painter at this time had reduced matter to its luminous essence, and here Leonardo reaches a high point of astounding novelty. Centuries would pass before painting would abandon itself to the effects of pure colour and light without the mediation of drawing. Even the transparent veil that covers the Virgin’s arm was unimaginable before its use here.
The success of these effects, contrary to what might be expected, is linked to the rapidity of Leonardo’s brushstrokes. The fact that he waited for years to define a detail did not mean that he worked by mixing colours on the palette and continuously changing the colour, even at a distance of months. Leonardo observed and waited until he could capture the right light, but when that moment came he was quick to mix the colour and dab the painting, leaving a small shadow of pigment that miraculously achieved the desired effect. In the mountains on the left-hand side of the painting that were finished during his Roman period, the manner of execution is very clear, as is that of the Virgin’s arm. The mountain arises out of a quick dab of brown, alongside an equally quick and confident touch of blue or white, and the transparency of the thin layers produces that luminous, atmospheric effect, at the sight of which we are prompted to squint as if we were in a real landscape, in order to bring the distant outlines into focus. Had the artist wasted time by adding further touches of colour, this miraculous freshness that turns matter into light would have been lost.
Lastly, another alteration that, again, might seem insignificant is the removal of the women’s footwear. The sandals that appear in the Milanese version (and in all five of the known versions copied from this painting)20 have the crossed laces typical of late fifteenth-century and early sixteenth-century Madonnas. This type of footwear was copied from Roman bas-reliefs and statues and provided an embellishment suited to a religious image. But Raphael had already removed these sandals from his Florentine Madonnas (Madonna of the Meadow, Canigiani Madonna, and Holy Family with the Lamb): leaving the Virgin with bare feet gave the image a certain timelessness while revealing the influence of classical sculpture, which was increasingly used as a reference in sacred art. Raphael gave bare feet to his Sistine Madonna, too, which was painted before 1513 and seen by Leonardo on his arrival in Rome.
This classicizing choice influenced Leonardo who, while continuing to work on his Saint Anne, decided to remove (or leave out) the sandals worn by both women. Moreover, the appropriateness of the image in a devotional sense might have been a consideration in the early version, which was almost certainly intended for a religious institution; but, once the painting no longer had a particular recipient and became instead a mobile laboratory on which Leonardo continued to study the effects of light and colour, its religious significance was less important. The artist emphasized the character of the painting as a naturalistic metaphor, the symbiosis between human generations in movement and natural transformation (of the rocks and the water). And the bare feet brought Saint Anne and Mary closer to the pagan images that populated the walls of Rome’s palaces and churches. Even this simple act of removal likened the painting to an ideal vision that recalled the beauty and harmony of the universe. The infant Christ, whom Pietro da Novellara had described as ready to sacrifice himself to fulfil the destiny of humanity, seems to ask his mother’s permission to climb onto the lamb, a creature whose docility is transferred, through empathy – as in the case of Cecilia Gallerani’s ermine, or in that of the horses at Anghiari – to the child; and its gentle submission heightens the tender emotion of the scene.
Although unfinished in some areas, or perhaps precisely because of this, the painting arouses a deeply emotional and aesthetic response, as the numerous copies that were made of it demonstrate. In purely pictorial terms, this is Leonardo da Vinci’s most successful attempt to exalt the material substance of nature and human bodies through colour and to conjure a vision of pure light that is mobile and difficult to grasp, yet goes straight to the viewer’s heart.