This dramatic view of the interior of the impressive remains of the third-century Roman baths, which are now incorporated into the Musée de Cluny in Paris, is one of two versions of the composition that were presumably produced following Girtin’s stay in France in the winter of 1801–2 (the other being TG1896). Unlike the multiple versions of earlier architectural subjects, such as the west front of Peterborough Cathedral (TG1017 and TG1018), which were painted at different times and were subject to variations in staffage, this and a second watercolour of the ruined frigidarium (which was then used by a cooper to store barrels) are virtually identical. There was at least one occasion when Girtin produced replicas of watercolours for different patrons (TG1562 and TG1563), but in this case he seems to have abandoned the work at the last moment as a result of an accident and then made the second identical version; the large vertical drip of colour in the lower centre is surely unsightly enough to have forced the artist to start again. The idea that this is the first version of the composition is supported by an inscription on the back of the drawing, which presumably records the name of a purchaser (‘C Cutter’) and a date (‘Ja.^ny 28 1803’). This matches the receipt of £3 18s from a ‘Mr Cutter’ dated 21 February 1803 that is listed by the artist’s brother, John Girtin (1773–1821), in his newly discovered accounts (Chancery, Income and Expenses, 1804; Smith, 2017–18, p.40).¹ All of this suggests that the watercolour was left in Girtin’s studio, unsold at his death, and the likelihood is that it was produced on commission but abandoned and then replaced. 

Our understanding of the work has been further enhanced by the startling discovery that, rather than being the outcome of an on-the-spot sketch executed by Girtin in 1801–2, as had always been assumed, Girtin’s watercolour is actually a copy, either from a watercolour by Jean-Baptiste-François Génillion (1750–1829) (see figure 1) or, more probably, after a print made by François-Denis Née (1732–1818) (see the source image above). Ephemeral details such as the carriage, the cloth hanging from poles above and the disposition of the barrels all replicate the print, confirming that, although the spectacular Roman ruins were no more than a few minutes’ walk away from Girtin’s lodgings in Paris, he must have based his watercolours on this secondary source. This discovery also helps to explain the problem Girtin’s watercolour has with the perspective of the frigidarium, which has the arch to the right coming out at an angle when it is actually on the same plane as the back wall, and it elucidates the curious nature of the wooden wheel form seen above the balcony, which turns out to be a misinterpretation of the print, which shows a smaller lower arch with vertical decorative brickwork above. Née’s etching was included in the Voyage Pittoresque de la France (vol.3, 1786), where it is titled Vue de Restes du Palais des Thermes, and there is a possibility that the artist actually acquired this and other prints of architectural subjects in Paris in 1801–2 (La Borde and others, 1781–1800). John Girtin records that amongst the sundry articles he found in the artist’s studio after his death were ‘French prints of Shipping bound and 5 D.o. Landscapes unbound’ (Chancery, Income and Expenses, 1804). I suspect that Née’s image was amongst those ‘French prints’, though the fact that it was published in 1786 means that one cannot entirely rule out what, at one time, would have seemed like an unthinkable alternative – namely, that this and some of the other Paris views might predate Girtin’s French trip, especially as neither of the two versions of the composition is dated. Nonetheless, the probability is still that they were painted in London later in 1802, not least because, as the paper historian Peter Bower has noted, both supports used by Girtin were English-made. In this case, the laid cartridge paper, which was also used for Durham Cathedral and Castle (TG1074), was attached to a ply sheet of white wove paper, probably made by Robert Edmeads (unknown dates) and Thomas Pine (unknown dates) at Great Ivy Mill near Maidstone (Smith, 2002b, p.106; Bower, Report). This was presumably done by John Girtin, seeking to transform a late discarded work into a more attractive commodity for sale. 

A third version of the composition was once in the collection of Leonard Gordon Duke (1889–1971) and was sold in his sale in 1970, sadly without an illustration (Exhibitions: Sotheby’s, 16 July 1970, lot 81). The low price it attracted, £20, suggests that the market did not agree with Duke’s attribution of the work to Girtin. 

(?) 1802

Paris: The Ruins of the Roman Baths, Hôtel de Cluny

TG1896

1794

The West Front of Peterborough Cathedral

TG1017

1794

The West Front of Peterborough Cathedral

TG1018

(?) 1800

The South Front of Chalfont Lodge, Seen from across the Lawn

TG1562

(?) 1800

The South Front of Chalfont Lodge, Seen from across the Lawn

TG1563

1799

Durham Cathedral and Castle, from the River Wear

TG1074