The view is based on an on-the-spot sketch by John Robert Cozens (1752–97), inscribed ‘Castle of St. Elmo – Naples’, and is dated ‘Novr.10’ (see figure 1) (Bell and Girtin, 1935, no.310). The drawing is found in the fourth of the seven sketchbooks from Cozens’ second Italian trip, which saw the artist travel to Naples in 1782 in the company of his patron William Beckford (1760–1844). It is unlikely that the Monro School watercolour was copied directly from the sketch by Cozens, however. An album put together by Sir George Beaumont (1753–1827), now in the collection of the Yale Center for British Art, New Haven, includes more than seventy tracings from on-the-spot drawings in the first three of the sketchbooks, which provided the basis for more than thirty Monro School works. There are only five tracings from the next three books, but there is no reason to think that others did not exist, and it was presumably from these lost copies by Cozens that as many as thirty-five more watercolours were produced by Girtin and Turner, including this view of the massive walls of the castle. We can be reasonably sure that this was the case here for two reasons. Firstly, the drawing is reversed from both the on-the-spot sketch and the finished studio watercolour that Cozens also produced (see figure 2), suggesting that Girtin’s copy was made from a tracing on a piece of thin paper that he simply turned over so that the design was inverted. Secondly, the Monro School watercolour includes a significant copying error, with the shaded area to the left in the on-the-spot sketch rendered as a prominent rocky outcrop in the copy. What was clearly an area of dark shadow in the on-the spot sketch must have looked quite different in a simple outline tracing. Whilst Cozens, having seen and sketched the building, would have known that the wall was perpendicular throughout its height, Girtin and Turner misinterpreted the limited information at their disposal.

The album containing this drawing was sold in 1833 as the work of Turner, but the cataloguer of the Turner Bequest, Alexander Finberg, thought that Girtin alone was responsible for the watercolours, whilst more recently Andrew Wilton has established their joint authorship (Finberg, 1909, vol.2, p.1232; Wilton, 1984a, pp.8–23). Identifying the division of labour within Monro School drawings is considerably helped, as here, when the colour washes by Turner leave large areas untouched in order to create highlights, so that Girtin’s pencil work is clearly evident, particularly in the figures. In practice, Girtin did little more than trace the general outlines of the composition and it was left to Turner to obscure the essentially mechanical task of replication, though in this case there is an air of incompleteness about the work. The outline is elaborated with just the two tones of the one colour, which means that it lacks much of the drama and power evoked by the Cozens watercolour (see figure 2). Comparing the two drawings refutes any suggestion that the Monro School drawings were copied from Cozens’ watercolours, something that is still occasionally claimed, and, equally, it reminds us that the ‘finished drawings’ that Monro commissioned were finished only in comparison with the slight materials that he was able to borrow.