[For Part 1, of which this is a continuation, see https://villacastagnadaylesford.com.au/2018/11/12/views-of-the-colosseum-from-the-north-1-luigi-rossinis-panorama/
The ex-Dunraven Panini is probably the earliest version of Panini’s Colosseum compositions, with a date that has been read as 1731, 1734 and 1735, and is now confirmed as 1734 (Fig. 1). It is without question his prime Colosseum veduta.
Many of Panini’s subsequent paintings are capricci or vedute ideate (recognisable monuments combined in an ideal space, usually based on the Campo Vaccino-Colosseum area). The ex-Dunraven painting is unusual in that it can be considered to be primarily a veduta. Many of Panini’s early topographical views derive from those by Gaspar van Wittel (Vanvitelli), but we need to be careful about what we mean by ‘derive’. When Panini came to paint topographical subjects he was undoubtedly aware of van Wittel’s painting, the only consistent body of painted Roman views. He was also aware of the long iconographic tradition were always temptated to follow, even copy, these precedents, rather than start from scratch. Van Wittel is unusual, if not unique, in using a camera obscura set up on site, with the gridded image on the camera obscura screen carefully transposed to large drawings, many of which survive (see Part 3). It is likely something of the kind occurred when van Wittel first created his Colosseum composition, as he did with his view of Piazza S. Pietro.  Panini could have chosen to simply copy these paintings, and in the case of his first Piazza S. Pietro paintings he may have done just this. His Dunraven Colosseum, however, differs greatly from van Wittel’s, and is clearly quite freshly observed; indeed, it is quite his most interesting view of the Colosseum in terms of its topographical interest and the vitality of its details. There is no doubt in my mind that this painting is based on studies from life. There are, however, few, if any, drawings by Panini of complete compositions, and his practice seems to have involved studies – or notations – of details (we find some of these in the British Museum sketchbook).
Sightlines of Panini’s Dunraven Colosseum. A. The Colosseum. (Right and left refer to the point of view of the painting.) Looking at the Dunraven Colosseum (Fig. 4) starting at the point where the upper arcades have collapsed at the right there are 16 arches visible to the left of this point before they curve out of sight. Starting at the same point, the point where the arcades are lined up with the centre of one of the four arcs that form the curvature of the Colosseum, helpfully marked by Nolli, is the pier at the left side of the third bay (Figs 2, 3). This line gives you one line running through the viewpoint. By drawing a tangent to the pier at the left of the sixteenth arch provides another. These intersect at point A1, which is the viewpoint of the Colosseum component of the picture.
The door in the houses lining the road to the left appears to be taken from the same viewpoint, and seems to be the gate in the garden wall (F1c) rather than the door in the building (F1a) (Fig. 5). You can see its little tile roof, the rounded top of the wall, and two bollards. The building (F1a) is on the bend in the road, so strictly speaking (according to Nolli) ought to be visible. But if we look at van Wittel (Fig. 6), there is a hint of further curvature (as well as a rise in the road). Van Wittel emphasises the rustication of the gateway, which is hardly visible in Panini. Healso shows the bollards.
Sightlines of Panini’s Dunraven Colosseum. B. Distances. Looking to the right of the collapsed upper stories, the structure is hard to read because the arcades on different levels merge, but there are two steps back in levels (Fig. 2). From the line of the edge of upper outer bays, there are 3 bays of the middle row of arches, as Nolli confirms, followed by 8 bays (if I have not miscounted) of the inner row of arches before the Colosseum curves around at the right. This provides one of four points than can generate two converging sight lines: (a) the edge of the Colosseum to the left edge of the monastery of SS. Giovanni e Paolo, and (b) the left edge of the Arch of Constantine line to the apse of SS. Giovanni e Paolo. This points to a viewpoint further to the left than the first (A2). This is necessarily so, as SS. Giovanni e Paolo is not visible from point A1.
This suggests that Panini made two separate drawings from life, one of the Colosseum and the street, the other of SS. Giovanni e Paolo and the Colosseum.
The presence of the high ground in the foreground, which is the remains of the platform of the Temple of Venus and Rome, confirms that his viewpoint was near point A2: the corner of the temple platform in Fig. 3 is exactly on the sightline left edge of Arch of Constantine–apse of SS. Giovanni e Paolo. Moreover, mound 2 is in the right place for such a viewpoint. The fact that these relationships works so well suggests that all of the side to the right of the Colosseum, foreground and background, is based on a single drawing.
Height. Panini’s horizon is lower than van Wittel’s, near the bottom of the first arcade and at the level of the temple platform, and a little higher than the garden wall at F1c. He shows road 5 below the temple platform. The viewpoint (K2) would therefore have been at the level of the temple platform. As Rossini’s view (Fig. 7, also Fig. 3) shows, the nearest high point across road 5 that would provide a viewpoint with the corner of the temple platform (K2) in line with the Arch of Constantine/SS. Giovanni e Paolo is on wall C, to the west of building G.
Conclusion. It seems likely that Panini’s Dunraven painting is a topographical paintings made from studies made on site, but that instead of using a single ‘camera obscura’ viewpoint like van Wittel, Panini made studies from two slightly different points (A1 and A2), one of the Colosseum and the street, the other of the view towards SS. Giovanni e Paolo from a little beyond the platform of the Temple of Venus and Rome, probably on wall C.
Panini and van Wittel. Because of his camera obscura technology, Van Wittel places distant elements in their correct topographical relationships, and can convey a great deal of information about the area between fore- and middle grounds to effectively map the space. Panini, by contrast, works from drawings—to judge from the British Museum sketchbooks, little more than notations—of self-contained units that are combined plausibly but not necessarily as accurate views from a particular viewpoint. He takes the basic distribution of elements and renders individual buildings accurately enough, but manipulates their positions compositionally and spatially to get the effect he wanted. Because of this the veduta ideata came naturally to him, and even his most assertively topographical pictures tend in this direction. Because the viewpoints used in the Dunraven Colosseum are relatively low, his composition resolves itself into a series of vertical planes, whereas van Wittel describes a horizontal plane extending out from his viewpoint, even if he pictorialized the foreground a little to avoid looking down on it.
Compared to van Wittel, Panini, at least in the Dunraven picture where he is new to the subject, picks up on numerous little pictorial incidents that enliven what could easily become a monotonous subject. The rendering of the arches in particularly lively (including the numbering of each bay with Roman numerals, which needs further investigation), as are the views through to glowing pink inferno-like interiors (Fig. 8). It is as if Panini has painted the gesticulating figures in front to draw our attention to them. And the landscape details are delightfully naturalistic: the view of SS. Giovanni e Paolo looks as if it were painted by Annibale Carracci (Fig. 9). Another delightful detail is the little bridge of Roman brickwork near point L (Fig. 10), presumably from some Roman ruin, which I am inclined to believe is topographically factual though I have not seen it elsewhere (it’s site is hidden by mound 2 in van Wittel).
 Giovanni Paolo Panini (1691–1765 Rome), Rome, a View of the Colosseum and the Arch of Constantine, 1734. Oil on canvas, 99 x 135.5 cm. Signed with initials and dated on the step, lower left: I.P. P. 1734. Arisi 224. Provenance: Earl of Dunraven, Limerick; with Herner Wengraf, London, 1972; Rome, Finarte, 12–13 December 1973, lot 55, reproduced pl. XXIX; anonymous sale (‘The Property of a Gentleman’), London, Sotheby’s, 1 November 1978, lot 43, reproduced in colour (unsold); private collection.
 Christoph Lüthy, ‘Hockney’s secret knowledge, Vanvitelli’s camera obscura’, Early Science and Medicine, vol. 10, issue 2, 2005, pp. 315–339.