When you look at Van Gogh’s painting The Yellow House you might well wonder what is going on with those big heaps of sand. You can see them in the lower left corner on the edge of Place Lamartine – the square in front of the Yellow House – and also running up Avenue de Montmajour towards the railway viaduct in the background.

Vincent van Gogh, The Yellow House (The Street), Arles, 27-29 September 1888, oil on canvas, 71,3 × 91 cm, Van Gogh Museum, Amsterdam (Vincent van Gogh Foundation)
The artist himself does not mention the subject in his letters, nor is it referred to very often in the literature on the painting. It was suggested in 2000 that the sand probably had something to do with work on a gas pipe, as Van Gogh wanted to have gaslighting installed in the Yellow House.
While researching the new museum catalogue, Paintings 3: Arles, Saint-Rémy, Auvers-sur-Oise, however, the Van Gogh Museum discovered an entirely different and more satisfying explanation, courtesy of a historical postcard.
The English artist Julian Trevelyan (1910–1988) was one of the first to ponder the significance of the sand. In an appreciative article on Van Gogh’s Yellow House in 1948, he mentioned ‘piles of earth from the road excavations’, without speculating as to what sort of works these might have been.
In the catalogue for his exhibition Van Gogh in Arles in 1984, meanwhile, the Van Gogh expert Ronald Pickvance suggested that the ‘mounds of earth’ are likely to have been heaps of ‘cut grass’, although this does not seem entirely plausible. Not until Light! The Industrial Age 1750–1900, held at the Van Gogh Museum in 2000, did the catalogue actually present a serious hypothesis for the first time.
Was it about installing gas lighting?
According to this, prior to the arrival in Arles of his artist-friend Paul Gauguin, Van Gogh had gas installed at the Yellow House so that they would be able to continue working after dusk. The piles of sand were thus silent witnesses to the related roadworks.
It is an ingenious suggestion that was widely accepted, but is it correct? We know from Van Gogh’s letters that he did not go ahead with the installation until he was sure precisely when Gauguin would come to Arles. And by that time, he had already finished The Yellow House, which he painted on 27–29 September.
Gauguin kept him in the dark for months regarding his arrival and did not finally confirm until around 10 October that he intended to travel to Arles at the end of the month. He proposed at the same time that they should draw designs for lithographs in the evenings.
Van Gogh was pleased that this would give him a little more time; he wrote to his brother Theo that ‘I want to try to have everything in good order and ready to receive him the day he comes’ [letter 701, Van Gogh’s underlining]. His preparations included having gas installed so they would be able to work in the evening. This would cost 25 francs, so Vincent asked Theo for extra cash. On 17 and 21 October, he notified Gauguin and Theo respectively that the studio and the kitchen had been connected [letter 706 and letter 709].
Detail of The Yellow House, with the lamppost on the left
Arles already had gas street-lighting
Installing gas in the house would anyway not have required the whole street to be dug up: most of the work would have occurred inside the building. The gas mains had already been laid between 1849 and 1881 and the whole of Arles had gas street-lighting, as witnessed by the lamppost on the corner of the square and the lamp a little further along on the wall above the pavement terrace.
The bars around the square – including Joseph Ginoux’s night café, where Van Gogh rented a room before moving into the Yellow House – also had interior gaslighting. It is noticeable, moreover, that while heaps of sand are visible in the painting, there is no sign of the trench you would expect if gas pipes were being laid.
Fig. 1: Arles, entrée de la Ville, Porte de la Cavalerie, 1902–03. Picture postcard, private collection
A picture postcard provides insight
A picture postcard dating from 1902–03 sheds light on the kind of work that was actually being carried out in 1888 (fig. 1).
It shows the Porte de la Cavalerie, the northern gateway into Arles, opposite the Yellow House, with a man holding a shovel standing in the middle of the street at the end of a pile of sand. The latter runs parallel to the pavement, in a very similar way to the mound shown in the painting.
What we see in the photograph is the municipal sanitation service cleaning up the dung – mostly that of horses – deposited every day in the street; the droppings are most clearly visible in the foreground. The street behind the man, who seems to be taking a quick break, has already been cleaned and spread with a fresh layer of sand to absorb the urine and loose manure left by the passing animals.
Sand layers on busy roads
In other words, Van Gogh’s Yellow House, with its similar piles of sand, also records the cleaning of Avenue de Montmajour, the city’s main northern thoroughfare.
Most of Arles and its outskirts were paved around 1865 with caladoun – small, specially cut pieces of typical Provençal stone. From then on, until the early twentieth century, busy roads were customarily provided with a layer of sand that was regularly cleaned and refreshed. Compacted sandy roads were treated in the same way – a practice that would disappear entirely with the advent of modern traffic.
Once upon a time, people would have immediately recognized these piles of sand, which also explains why no one felt the need to comment on them. The same does not apply, however, to later generations. Armed with this knowledge, once you start looking you can spot sand-covered roads like this in other old postcards too, with or without the dung.
There is another one from Arles, for instance, in which the Yellow House can also be seen (fig. 2). Or a postcard dating from around 1905 of the street in front of the Auberge Ravoux in Auvers-sur-Oise, in which we make out a relatively fresh layer of sand (with clear wheel tracks and as yet only a few droppings), along the edges of which the paved road surface is just about visible (figs. 3 and 4).
Fig. 2: The Yellow House and Avenue Montmajour, c. 1905. Picture postcard, private collection
Fig. 3: Auvers-sur-Oise – La rue de la Gare, c. 1905. Picture postcard, private collection
Fig. 4: Detail of the edge of the street
A nice little detail is that in August 1888 Arles municipal council took the decision to outsource the cleaning of its streets, ditches and public conveniences (‘balayage et vidange’) to a new company, as reported in the local weekly paper L’Homme de Bronze.
The council voted unanimously to accept the tender submitted by the ‘Compagnie Générale des vidanges de Marseille’, the terms of which included an obligation on the city’s part to provide ‘storage for dung and sweepings’. Mayor Jacques Tardieu was given a free hand to make the arrangements. We, as viewers of Van Gogh’s painting, are thus eyewitnesses to the Compagnie’s earliest activities.
Teio Meedendorp
Researcher, Van Gogh Museum
Literature
Letter numbers refer to: Leo Jansen, Hans Luijten, Nienke Bakker (eds.), Vincent van Gogh – The Letters, Amsterdam/The Hague/Brussels 2009 (www.vangoghletters.org)
Archives: Archives nationales, 136 AQ, Société diverses filiales de Gaz de France, E (La Compagnie du Gaz d’Arles), and Les Archives communales d’Arles, letter to the Van Gogh Museum, 4 November 1997.
Andreas Blühm, Louise Lippincott, Light! The Industrial Age 1750–1900: Art, Science, Technology and Society, exhib. cat. Amsterdam (Van Gogh Museum), Pittsburgh (Carnegie Museum of Art), 2000, p. 202.
Ronald Pickvance, Van Gogh in Arles, exhib. cat. New York (The Metropolitan Museum of Art), 1984, p. 76.
Jean-Maurice Rouquette, Paul Allard, et al., Arles : histoires, territoires et cultures, Paris 2008, pp. 867, 932.
Julian Trevelyan, ‘Van Gogh’s ‘The Yellow House’, II - The Picture’, The Listener, 8 January 1948, p. 59.
Annie Tuloup-Smith, Rues d’Arles. Qui êtes-vous ?, Amis du Vieil Arles 2003, p. 13.
L’Homme de Bronze – Journal de l’arrondissement d’Arles, no. 467, 25 September 1888, p. 2 – report of the council meeting, 18 August 1888
The Musée de la Camargue confirmed the identification (correspondence between the author and curator Estelle Rouquette, 21 August 2019).