Leonardo excelled in the representation of horses

Contemporaries were also unanimous that Leonardo excelled in the representation of horses. He had so designed the subject as to give this admitted superiority full play and Vasari tells us how these horses played a leading part in the drama, "for rage, hatred and revenge", he says, "are seen in them no less than in the men". Of this, also, we are in a measure able to judge, as some of his horse studies for the battle have come down to us.

But when all this had been granted, a Florentine of that date would have decided in favour of Michelangelo, on account of the matchless beauty of his nudes. Leonardo felt this, and during this period made a number of magnificent nude studies. It is impossible to imagine better drawings of the nude than that; as an actual study of a muscular torso Michelangelo could hardly have excelled it, and in fact the heroic pose and treatment shows Michelangelo's influence. But although Leonardo could master the nude when he chose, he was not prepared to make it the main subject of a composition. It gave too little opportunity for his love of fantastic invention, his unexpected imagery--in short for the expression of that anti-classical side of his character to which I have more than once referred. The Anghiari cartoon may have had a classic firmness and coherency; but in spirit, like the Adoration, it was a romantic masterpiece--a precursor of Tintoretto, Rubens and Delacroix. Thus it was outside the main current of its time. Contemporaries could not ignore its marvellous qualities of drawing and design, and we know that it influenced the mannerist painters of the next generation. But Michelangelo with his severe concentration on the nude, his passionate research for noble and expressive form, seemed to offer a firmer and, perhaps, a shorter way to excellency, and it was the Battle of Cascina which captivated the younger artists of the time.

On 30 May 1506, Leonardo asked and obtained leave from the Signoria to return to Milan for three months, at the urgent request of the governor, Charles d'Amboise, Lord of Chaumont; and before the time was out Amboise wrote asking that it might be prolonged "since we still have need of the master Leonardo to furnish us with a certain work which we have had him begin". What was the work so urgently required and so discreetly left unnamed? Several scholars of Leonardo have suggested that it was the London version of the Virgin of the Rocks, and although this hypothesis cannot be proved, it disposes so neatly of a difficult problem without violence to any of the known facts, that it may be accepted until denied by some new discovery.

The main facts in a tangle of evidence are as follows. On grounds of style the Louvre picture is undoubtedly the one painted in 1483 and the National Gallery version was painted considerably later. Yet the National Gallery picture was bought from the Confraternity whereas the Louvre picture was in the French royal collection at an early date. At some point, clearly, the original version was sold to France, and a second painted to take its place. Only one picture is mentioned in all the documents relating to Leonardo's lawsuit with the Confraternity, so the second version can hardly have been painted till after the last of these documents dated 27 April 1506. Now it is precisely on 30 May 1506, that Leonardo is summoned back to Milan to work on an unnamed commission. Hence the theory that, the lawsuit safely over, Amboise sent this acknowledged masterpiece of Leonardo's early period to adorn his King's collection, promising the Confraternity that Leonardo would paint a replica in a more mature style.

If this hypothesis is correct, we must sympathise with the Confraternity. For Leonardo, with the Battle of Anghiari still unfinished, Isabella d'Este still unsatisfied, and his scientific studies growing daily more absorbing, was unlikely to carry out much of the work with his own hand. He found time, however, to bring the composition more into conformity with the high Renaissance desire of unity of effect. Thus the figures have been made larger in relation to the panel (the National Gallery panel is actually a fraction smaller, but gives the reverse impression) and the distracting motive of the Angel's hand pointing at the Infant John has been suppressed. Possibly its Florentine insistence on the Forerunner was distasteful to a body dedicated to the Immaculate Conception. An analogous change has been made in the heads, which have been redesigned under the influence of Leonardo's later theories of painting. The types have lost their Gothic freshness and naturalism, and the modelling is carried out with darker shadows. Contrary to the best critical opinion I believe that Leonardo also had some part in its execution. Many of the details are drawn with a delicacy quite beyond a pupil, and with Leonardo's own feeling for living tissue. Unfortunately almost every figure has been damaged at some point by coarse repainting, and this makes it hard to say where Leonardo's work ends and a pupil's begins. One passage is intact, the Angel's head, which departs entirely from the Paris version, and which achieves real beauty of a kind made familiar in the work of the Milanese school, but certainly deriving from Leonardo himself. No one who has looked at it closely can doubt who was responsible for the mouth and chin, and the characteristic curves of the golden hair. The most difficult part to explain is the Virgin's head. It is essentially un-Leonardesque--heavy and devoid of inner life: yet the execution is delicate and shows the marks of thumb and palm in the thin paint, which are to be found in all the best passages in the picture.

Although much can be said in praise of the National Gallery Virgin of the Rocks, it falls far short of the Louvre picture in every kind of beauty and must be largely pupil's work, which pupil we do not know. He is generally supposed to be Predis, owing to his documentary connection with the commission for the earlier version, but long before 1506 he had ceased to be in any sense a pupil of Leonardo, and there is no reason why he should have been employed on the replica when he had not even touched the original. The style is completely foreign to him. The pupil whose style seems to bear most resemblance is the author of a Virgin and Child at Zürich, in which the Virgin's head is almost a replica of that in the London Virgin of the Rocks. The Zürich picture is signed FRLTA, and Suida has suggested that this is the Spaniard Fernando de Llanos, who is more than once recorded as having assisted Leonardo on the Battle of Anghiari. But unfortunately Fernando returned to Valencia in 1506, so that if my hypothetical date of the London Virgin of the Rocks is correct he cannot have been its author.