`Sunset at Montmajour`: a newly discovered painting by Vincent van
Transcription
`Sunset at Montmajour`: a newly discovered painting by Vincent van
‘Sunset at Montmajour’: a newly discovered painting by Vincent van Gogh by LOUIS VAN TILBORGH, TEIO MEEDENDORP and ODA VAN MAANEN ‘Sunset at Montmajour’: a newly discovered painting by Vincent van Gogh by LOUIS VAN TILBO RGH, TEIO MEEDENDORP and ODA VAN MAANEN W HEN THE MUCH-AWAITED , revised edition of De la Faille’s catalogue raisonné of Van Gogh’s œuvre was published in 1970 it was not to universal acclaim. In this Magazine Ronald Pickvance spoke glumly of a ‘progress report’, and John Rewald thought that if there was ‘a more confusing way to present a catalogue, the editors apparently couldn’t find it’.1 But, pace the grumblers, it did have its virtues, the most important being that ‘at least 100 paintings and drawings’ had been discovered.2 Although that was a sizable number, it turned out that the additions to the œuvre were far from complete, because even more works have been rescued from oblivion since then.3 However, no matter how fascinating and important such discoveries are for a proper understanding of the œuvre, when one takes a closer look at what has been added since De la Faille’s original book of 1928 one finds that there have been very few real surprises. In retrospect, what he had missed back then was not all that much: The Tarascon coaches of 1888, some pen drawings from Arles, and one from Saint-Rémy.4 So the words ‘absolutely sensational’ are no less than fitting for the recent discovery of an unpublished and ambitious painting in a private collection from ‘the zenith, the climax, the greatest flowering of van Gogh’s decade of artistic activity’ – his Arles period (Fig.56).5 It shows a wild, rocky area to the east of the ruined abbey on the hill of Montmajour. There are numerous holm (or holly) oaks, and even today the area abounds in these distinctively Mediterranean trees with their holly-shaped leaves and twisted trunks (Fig.54).6 At top left are the ruins bathed in the late afternoon sun, which have made it possible to pin down the precise spot and Van Gogh’s orientation. He was looking west on a plateau about a kilometre from Montmajour, just south of the road that winds its way to Fontvieille (Fig.57).7 Through the bushes on the right is a glimpse of the farmland that stretches north-west to the Rhône. Van Gogh had realised the potential of this untouched area around Montmajour as a subject almost as soon as he arrived in Arles in early 1888, but because it was too cold and windy at the time he had put off recording the ‘beautiful things’ there.8 It was not until the end of May that he began to explore the area in drawings (Figs.55 and 71), but he seems to have been also thinking of capturing some of the subjects in oils. Possibly inspired by Theo’s recent description of Monet’s Under the pines: evening (Fig.58), he thought that a sunset at this spot, with the orange- This article was translated from the Dutch by Michael Hoyle. All the data on the pigments in Sunset at Montmajour are from the technical report by Muriel Geldof, Luc Megens and Maarten van Bommel of the Cultural Heritage Agency of the Netherlands (RCE), project number 2012-014 (hereafter RCE report). Unless otherwise stated, all the samples in this article were prepared in cross-section and examined with the light microscope and Scanning Electron Microscope with Energy Dispersive Xray Analysis (SEM-EDX) by Muriel Geldof. The indication of pigments with handheld X-ray Fluorescence Spectrometry (XRF) was performed by Luc Megens, and analyses with High Performance Liquid Chromatography (HPLC) by Maarten van Bommel. Don H. Johnson and Robert G. Erdmann were instrumental in analysing the types of canvas used for Sunset at Montmajour and The rocks in the framework of the Thread Count Automation Project (TCAP). The information about the history of the painting and Christian Nicolai Mustad was supplied by the former owner, who wishes to remain anonymous, as do the present owners. Many thanks to David Bomford and Helga Aurisch of the Museum of Fine Arts, Houston, who gave us the opportunity to examine The rocks, and to Melissa Gardner for the identification of the pigments in the ground of that painting using polarised light microscopy (PLM) of dispersed pigment samples and XRF. We are also grateful to Tina Tan and Maureen Eck for their kind help with the photography and photomicrographs of The rocks; to Petra Pettersen, Munch Museum, Oslo, Turid Aakhus and Nils Messel, both of the Nasjonalmuseet, Oslo, for providing us with information on Mustad as a collector. Our thanks, too, to Walter Feilchenfeldt, Sjraar van Heugten, Evert van Uitert, and our colleagues in the Van Gogh Museum: Monique Hageman, Ella Hendriks, Leo Jansen, Hans Luijten and Marije Vellekoop. 1 J.-B. de la Faille: The works of Vincent van Gogh. His paintings and drawings, Amsterdam 1970 (hereafter the catalogue numbers in this book are referred to as F followed by a number); the quotations are from R. Pickvance: ‘The new De la Faille’, THE BURLINGTON MAGAZINE 115 (1973), p.175, and J. Rewald: Post-Impressionism. From Van Gogh to Gauguin, New York 1978, p.538. 2 Pickvance, op. cit. (note 1), p.175, to which he added that it was ‘difficult not to 696 o c to be r 20 1 3 • c lv • t he b u rl in g t on ma g a z i ne 54. Holm oaks near Aureille, just south of the Alpilles. (Photograph 2013). 55. La Bruyère, by Vincent van Gogh. 1888. Pencil, reed pen in purple ink (now brown in most places), 31.3 by 48.1 cm. (Van Gogh Museum, Amsterdam, Vincent van Gogh Foundation). A NEWLY DISCOVERED PAINTING BY VAN GOGH 56. Sunset at Montmajour, by Vincent van Gogh. 1888. Canvas, 93.3 by 73.3 cm. (Private collection). yellow rays that are so characteristic of the south of France, would make a superb picture. ‘I saw a red sunset that sent its rays into the trunks and foliage of pines rooted in a mass of rocks, colouring the trunks and foliage a fiery orange while other pines in the further distance stood out in Prussian blue against a soft blue-green sky – cerulean. So it’s the effect of that Claude Monet. It was superb. The white sand and the seams of white rocks under the trees took on blue tints’.9 When Van Gogh returned to the site at the beginning of July, evidently planning to crown his reconnaissances on paper with ambitious paintings, it was perhaps no coincidence that his first subject was just such a fiery sunset. ‘Yesterday, at sunset’, he wrote to Theo on 5th July, ‘I was on a stony heath where very small, twisted oaks grow, in the background a ruin on the hill, and wheatfields in the valley. It was romantic, it couldn’t be more so, à la Monticelli, the sun was pouring its very yellow rays be more approximate, let alone precise’, for reasons that emerged in his critical review. 3 This is not the place to list them in full, but see J. Hulsker: The new complete Van Gogh: paintings, drawings, sketches, Amsterdam and Philadelphia 1996 (hereafter cited as JH), Addenda, pp.484–86; M. Bailey: ‘Real discoveries’, The Art Newspaper (July–August 1997), pp.24–25; M. Vellekoop: ‘A newly discovered drawing by Van Gogh’, THE BURLINGTON MAGAZINE 150 (2008), pp.106–09; L. van Tilborgh and E. Hendriks: ‘Dirk Hannema and the rediscovery of a painting by Vincent van Gogh’, THE BURLINGTON MAGAZINE 152 (2010), pp.393–405; and T. Meedendorp: ‘Van Gogh in training: the idiosyncratic path to artistry’, in M. Vellekoop et al., eds.: Van Gogh’s studio practice, New Haven and London 2013, pp.40 and 41. 4 F 478a, F 1482a, F 1507a, F 1518a, F 1525a, F-JH Add. 3 and F 1728. 5 Which is how Ronald Pickvance summarised the traditional view of this period in the Van Gogh literature; see his exh. cat. Van Gogh in Arles, New York (Metropolitan Museum of Art) 1986, p.11. 6 The Latin name is Coccifera ilex, and we cannot resist mentioning that the identification was made on the basis of acorns gathered from the same spot. With thanks to Hans den Nijs of the Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam. 7 This was discovered on a visit to the location. Far more of the north front of the ruins is visible from a position north of the road, and the distance between the tower on the left and the rectangular structure behind it is far greater than in the painting. The ruins are seen at more of an acute angle from a spot south of the road, with the result that those two sections move closer together. Van Gogh must have been close to the road, but it is not visible in the picture because he was standing on a slightly elevated plateau. 8 L. Jansen, H. Luijten and N. Bakker: Vincent van Gogh. The letters, The Hague 2009, which are constantly being updated at www.vangoghletters.org (hereafter cited as Letters), no.583. 9 Letters, no.615. the burling ton maga zine • clv • o cto ber 2 013 697 A NEWLY DISCOVERED PAINTING BY VAN GOGH 57. Detail from the Carte de l’état-major de France 1820–66. (Institut national de l’information géographique et forestière (IGN), France). over the bushes and the ground, absolutely a shower of gold. And all the lines were beautiful, the whole scene had a charming nobility. You wouldn’t have been at all surprised to see knights and ladies suddenly appear, returning from hunting with hawks, or to hear the voice of an old Provençal troubadour. The fields seemed purple, the distances blue. And I brought back a study of it too, but it was well below what I’d wished to do’.10 Up until now that study has been associated with The rocks, a work related to Monticelli’s style and subject-matter that is now in the Museum of Fine Arts, Houston (Figs.60 and 59).11 However, that scene of a tree in the rocky area around Montmajour does not match Van Gogh’s description. The missing elements are the abbey ruins, the wheatfields in the valley, the many small, twisted oaks, and the ‘very yellow rays over the bushes and the ground’, and since they are all present in Sunset at Montmajour this must be the study that Van Gogh was referring to. On top of that, he said that the result was ‘well below what I’d wished to do’, and that does not fit The rocks either, which he rated highly, as we know from the pen drawing that he made after it and sent to Emile Bernard.12 In the spring Van Gogh had ruthlessly destroyed works that he felt to be below par,13 but Sunset at Montmajour escaped that fate. It was one of the paintings from the previous period that his friend Paul-Eugène Milliet took to Theo in Paris in mid-August. Van Gogh liked the subjects but felt that the execution was not always quite up to the mark, with Sunset at Montmajour being the most striking example. ‘The roll that he’ll bring you contains 36 studies; among them there are many with which I’m desperately dissatisfied, and which I’m sending you anyway because it will still give you a vague idea of some really fine subjects in the countryside. For example, there’s a quick sketch I made of myself laden with boxes, sticks, a canvas, on the sunny Tarascon road; there’s a view of the Rhône, in which the sky and the water are the colour of absinthe, with a blue bridge and black figures of ruffians; there’s the sower, a washing-place and still others, not at all successful and unfinished, especially a large landscape with brushwood’ (emphasis added).14 It had always been uncertain which that last picture was,15 but the newly discovered work now clears up that problem. Sunset at Montmajour was Van Gogh’s third painting after The harvest of early June on the large ‘Portrait’ 30 size canvas (92 by 73 cm.),16 which he had rarely used before, so he evidently had high hopes for the subject. It is a commercial, ready-primed canvas that matches the type he used a week later for The rocks, as the technical examination revealed.17 Almost all the pigments are the ones he habitually had on his palette at this time.18 A typical one is the cobalt blue, which is of the type he began to use from the summer of 1887 onwards.19 Van Gogh’s method is clearly visible in both the structuring and detailing of the scene, which seems to have been executed in 10 13 Letters, no.636. This was first suggested in J. Hulsker: Van Gogh door Van Gogh. De brieven als commentaar op zijn werk, Amsterdam 1973, p.139. Pickvance, op. cit. (note 5), p.118, then believed that this was ‘almost certainly’ correct, whereupon it was adopted unconditionally in C. Ives: exh. cat. Vincent van Gogh. The drawings, Amsterdam (Van Gogh Museum) and New York (Metropolitan Museum of Art) 2005, p.237; and Letters, no.636, note 3. 12 Van Gogh sent the drawing to Bernard on 15th or 20th July (F 1554; Letters, nos.641, note 1, and 643, note 1), and it seems that The rocks was made shortly before. The Houston picture was not only associated with his letter of 5th July, but also with his letter of around 13th July, in which he spoke of his very last attempt to paint rather than draw during the mistral (Letters, no.639, note 6), and the rapid, almost stenographic brushwork seems to bear that out. It is known from letter no.610 that Van Gogh had seen one of Monticelli’s paintings of the same subject: ‘You remember that fine landscape by Monticelli that we saw at Delarebeyrette’s, of a tree on some rocks against a sunset’. 11 698 o c to be r 20 1 3 • c lv • t he b u rl in g t on ma g a z i ne 58. Under the pines: evening, by Claude Monet. 1888. Canvas, 73 by 92.1 cm. (Philadelphia Museum of Art). See Letters, no.606, note 3. Letters, no.660. 15 Letters, no.660, note 6: ‘There is no known painting that fits this description’. For the other works, see note 52 below. 16 Although Van Gogh must have occasionally used a size 30 canvas in April and May 1888, none of those works has survived (see Letters, nos.602 and 606, note 3), so the earliest extant pictures in that format date from June and are, in chronological order, F 412, F 425 and F 465. The stretcher of Sunset at Montmajour is not the original working frame, because a narrow strip of ground of 0.3 to 1 cm. wide can be seen around the edges on the picture side, which could indicate that the visible area was enlarged at some stage. 17 In the framework of the Thread Count Automation Project – see L. van Tilborgh et al.: ‘Weave matching and dating of Van Gogh’s painting: an interdisciplinary approach’, THE BURLINGTON MAGAZINE 154 (2012), pp.112–22 – automatic thread counts were made from high-resolution digital scans of X-radiographs covering the whole picture area, which for Sunset at Montmajour resulted in an average horizontal 14 A NEWLY DISCOVERED PAINTING BY VAN GOGH 59. Tree on a rock at Ganagobie, by Adolphe Monticelli. After 1875. Panel, 44.8 by 34.8 cm. (Tanimoto Hiroaki Collection, Japan). 60. The rocks, by Vincent van Gogh. 1888. Canvas, 55 by 65.7 cm. (Museum of Fine Arts, Houston, gift of Audrey Jones Beck). three stages. He began with an initial lay-in of the composition in quite a bold bluish green made up of viridian, emerald green and a little cobalt blue (Fig.62), and this use of a bright underlayer is known from other paintings of his.20 The scene was then worked up alla prima almost in one go, with the third stage being completed in the studio. As usual this involved adding small accents to the touch-dry paint, ‘to adjust the workmanship a little, to harmonize the brushstrokes’,21 to quote Van Gogh’s own words about his standard practice at the end of the painting process (Figs.63 and 64). Although there are of course differences in colour because of the subjects depicted, the palette of Sunset at Montmajour is very similar to that of Wheatfield (Van Gogh Museum, Amsterdam) and Newly mown lawn with a weeping tree (private collection) of mid-June and early July respectively. Characteristic details include the liver colour of the clouds, and the ‘vert véronèse’ bluegreen of the sky, which is identical to that in many other landscapes from that summer in Arles.22 The types of discolouration are also typical of Van Gogh’s work. At the bottom edge, for instance, there is a red brushstroke that has turned a greyish purple (Fig.65), and many of the greys elsewhere, in the trunks of 61. The public garden (‘The poet’s garden’), by Vincent van Gogh. 1888. Canvas, 73 by 91 cm. (Art Institute of Chicago, Mr and Mrs Lewis Larned Coburn Memorial Collection). thread density of 15.9 threads per cm. (weft) and an average vertical thread density of 17.6 threads per cm. (warp), and for The rocks in an average horizontal thread density of 17.6 threads per cm. (warp) and an average vertical thread density of 15.9 threads per cm. (weft). A sample of the ground of Sunset at Montmajour shows two layers containing lead white, calcium carbonate, barium sulphate, probably lithopone and some yellow ochre. The upper layer contains more lead white and less calcium carbonate than the lower one. A sample from the ground of The rocks showed a similar composition and build-up of two layers (RCE report). Initial pigment analyses were performed by Melissa Gardner (Museum of Fine Arts, Houston). 18 Pigments indicated with handheld XRF and confirmed by SEM-EDX (in italics) or HPLC (italic and underlined): lead white, zinc white, emerald green, viridian, cobalt blue, ultramarine, Prussian blue, lead chromate, yellow ochre, vermilion, a little barium sulphate, a little zinc yellow (?), calcium sulphate (probably filler), an organic red pigment (redwood) on a substrate containing tin and possibly aluminium, an organic red pigment (cochineal) on a substrate containing tin and calcium and (probably) starch particles (Ø c. 5 µm) (RCE report). As far as we know these types of organic red pigment have not previ- ously been identified in paintings by Van Gogh from his Arles period, but there is a lack of comparative material owing to the limited analyses of other paintings from the period. See also M. Geldof et al.: ‘Van Gogh’s palette in Arles, Saint-Rémy and Auvers-sur-Oise’, in Vellekoop et al., op. cit. (note 3), pp.238–55. 19 Semi-quantitative analysis on the cross section showed that the amount of nickel is relatively low in the cobalt blue used for Sunset at Montmajour (average weight percentage nickel to cobalt of 2.7 x 10-2), and furthermore the pigment contains a little phosphorus (RCE report). For the use of this type of cobalt blue in Van Gogh’s postParis period, which probably corresponds to the paint sold by Tasset et l’Hôte, see M. Geldof and L. Steyn: ‘Van Gogh’s cobalt blue’, in Vellekoop et al., op. cit. (note 3), pp.256–67. 20 See, for example, Van Tilborgh and Hendriks, op. cit. (note 3), p.402. 21 Letters, no.635. 22 Analyses with XRF indicated that zinc white, emerald green, a little cobalt blue and a little viridian were used for this paint (RCE report). For the clouds see F 261 and F 575, and for skies containing Veronese green, see F 411, F 428, F 472, F 428, F 429 and F 430. the burling ton maga zine • clv • o cto ber 2 013 699 A NEWLY DISCOVERED PAINTING BY VAN GOGH the holly oaks for instance, were originally redder or more purple, while the rocks in the foreground may have been more purplish. In addition, much of the chrome yellow has darkened, most noticeably in a brushstroke of which part has broken off (Fig.66).23 Van Gogh’s hand can be recognised in the diversity of the brushstrokes and the creaminess of the paint, as well as in the rapidity and liveliness with which it was applied. There are many characteristic details. For instance, the gaily brushed zigzagging strokes for the tall tree that almost reaches the top of the picture, the straight strokes for the smaller bushes and tussocks, and the sinuously rendered trunks of the holly oaks are also found in The rocks (compare Figs.56 and 60), while the short strokes for the leaves of the trees on the horizon are almost identical to those in The public garden (‘The poet’s garden’) (Fig.61). And as usual, Van Gogh used one of his broadest brushes for the sky. In the weeks leading up to this painting Van Gogh had already tried to adapt his style to the rather uncomfortable conditions created by high winds, and in order to record the fleeting effect of a sunset he was now ‘forced to lay the paint on thickly, à la Monticelli’, to take a later quotation out of context.24 The paint is indeed very impastoed, particularly in the crowns of the holm oaks, and although this is more typical of his later work from Saint-Rémy, he did start to experiment with it in 1888. This makes Sunset at Montmajour a continuation, as it were, of Boats at sea at Saintes-Maries (Pushkin Museum, Moscow), and a precursor of a picture such as The green vineyard (Kröller-Müller Museum, Otterlo). There are several impressions of a canvas structure in the high, flattened impasto, particularly on the right, and this is another familiar feature in Van Gogh’s paintings. In this period he would take his canvases from their strainers, stack them on top of each other and then roll them up and send them to Theo.25 This explains not only the impressions left by a canvas in the impasto but presumably also the presence of tiny traces of brown-red and yellow paint on the back of the unlined canvas. Although Sunset at Montmajour vanished from sight for over a century it transpires that it does have a traceable history, and it is possible to offer an explanation for its long absence from the literature. Van Gogh described it himself in two letters of 1888, and its next documentary sign of life comes only two years later, when it was listed as ‘soleil couchant à Arles 30’ under number 180 in the Catalogue des œuvres de Vincent van Gogh in Theo’s collection that Andries Bonger drew up in 1890.26 The subject and size of the picture match that description, but the clinching piece of evidence is simply that the number 180 is written on the back of the canvas (Fig.67). The handwriting resembles that on Wheatfield with setting sun, now in the Kunstmuseum, Winterthur (Fig.68), which is also marked with a Bonger number.27 It is numbered 100 on the list and was given the title ‘Arles (Soleil couchant) 30’, which makes it likely that both numbers, which were probably written by packers, were added to prevent confusion between the two works when they were sent off to exhibitions.28 Confirmation that Theo’s widow, Jo van Gogh-Bonger, did indeed send the painting out into the world comes from other references. She lent it for the exhibition of Van Gogh’s works in the Panorama building in Amsterdam in 1892–93, which was curated by the artist Richard Roland Holst. A reviewer of that large show mentioned a painting of ‘dry stony ground with thin trees’, which can hardly be a reference to anything other than Sunset at Montmajour.29 The picture was also in a selling exhibition in March 1901 mounted by an artists’ society in Utrecht called Voor de Kunst, a fact we know from a list that Jo van Gogh-Bonger made of works she had lent, where it appears under its Bonger number of 180 but with the not entirely incomprehensible title of ‘autumn landscape’.30 Although it is not documented, Sunset at Montmajour changed hands shortly afterwards. The family archive contains no further documents relating to number 180, and the fact that Jo sold it can be deduced from Meier-Graefe’s book on modern art of 1904, in which he noted that the French merchant collector Maurice Fabre owned seven Van Goghs, including ‘Groupe d’arbres avec nuages mouvementés’ (‘Group of trees with scudding clouds’).31 Until now no one knew which painting that was,32 but since that title is very acceptable as a description of Sunset at Montmajour, with its lively sky and striking group of trees, this must be it. Turning to the question of when Fabre could have acquired it from Jo van Gogh-Bonger’s collection, the only possible occasion could have been a small selling exhibition held in April and May 1901 in the Paris apartment of the art critic and dealer Julien Leclercq.33 Leclercq had approached Theo’s widow in April and asked her to send five or six paintings, including two or three landscapes of her own choice, and Sunset at Montmajour was almost certainly one of them.34 It must just have returned from its outing to Utrecht, so sending it straight on would have been a simple option.35 Fabre had several other pictures by Van Gogh which he gradually disposed of, the last one probably around 1908.36 However, there is no record anywhere of the sale of Sunset at Montmajour. It vanished without trace after the mention in 23 many of them have since been lined, any such inscriptions are now hidden from sight. 29 J.W. van Dijckveldt: ‘Tentoonstelling Vincent van Gogh’, Amsterdammer (1st January 1893): ‘. . . dorre steengronden met schraalharde boomen’. 30 Lijst van de schilderijen in ‘Voor de Kunst’ Utrecht, no.17, which was offered for sale at 900 guilders (Utrecht, Utrechts Archief, Archief Vereeniging voor de Kunst, no.777-2: 7). It was described in a review by ‘Gio.’: ‘Vincent van Gogh’, Algemeen Handelsblad (19th March 1901), which is included in Andries Bonger’s scrapbook (Amsterdam, Van Gogh Museum Library, BVG 3117, p.51): ‘Toch is er veel moois in [. . .] een sterk verlicht, als verguld herfstlandschap’ (‘But much in it is beautiful [. . .] a brightly lit autumn landscape, as if gilded’). 31 J. Meier-Graefe: Entwicklungsgeschichte der modernen Kunst: vergleichende Betrachtung der Bildenden Künste, als Beitrag zu einer neuen Aesthetik, Stuttgart 1904, I, p.120, note 1. 32 On this, see Feilchenfeldt, op. cit. (note 26), p.299. See note 36 below for the works in Fabre’s collection. 33 Fabre and Leclercq knew each other, for the former lent six works to the Van Gogh exhibition at Bernheim-Jeune that Leclercq organised in March 1901; see exh. For discolourations of chrome yellow in paintings by Van Gogh, see L. Monico et al.: ‘The degradation process of lead chromate in paintings by Vincent van Gogh studied by means of Synchrotron X-ray Spectromicroscopy and related methods’, Analytical Chemistry 83/4 (2011), pp.1224–31. 24 Letters, no.689. 25 It is known from Letters, no.660, that Sunset at Montmajour was rolled up with other works for dispatch. 26 Until now it was not known which picture this was; see W. Feilchenfeldt: Vincent van Gogh. Die Gemälde 1886–1890. Händler, Sammler, Ausstellungen. Frühe Provenienzen, Wädenswill 2009, p.288. The number 30 is an identification of the format (‘Portrait’ 30 = 92 by 73 cm.). 27 E. Hendriks: ‘Condition report F465’, 15th January 2008, written jointly with Jan Jedlicka (former paintings conservator at the Kunstmuseum, Winterthur); on file at the Van Gogh Museum, Amsterdam. 28 There must be more of these numbers on the back of paintings by Van Gogh (F 730, for instance, has Bonger number 221, and F357 has number 30), but because 700 o c to be r 20 1 3 • c lv • t he b u rl in g t on ma g a z i ne A NEWLY DISCOVERED PAINTING BY VAN GOGH 62. Photomicrograph of Fig.56 showing the layin of the composition with diluted washes of green paint. 63. Photomicrograph of Fig.56 showing the wetin-wet mixing of the green paint with the paint of the sky and final touches of green added onto the touch-dry paint. 64. Photomicrograph of Fig.60 showing similar wet-in-wet mixing of the green paint with the paint of the sky and final touches of green added onto the touch-dry paint. (Photograph Museum of Fine Arts, Houston). 65. Photomicrograph of Fig.56 showing the discolouration of the red lake paint in an abraded area. 66. Photomicrograph of Fig.56 showing the discolouration of the chrome yellow paint, where part of the impasto has broken off, slightly enhanced by the absence of varnish. Meier-Graefe’s book, and only reappeared in 1970 in the estate of Christian Nicolai Mustad (1878–1970), a Norwegian industrialist in Oslo who, with his four brothers, joined his father as co-owner of the company O. Mustad & Søn in 1905. According to the family, he started collecting contemporary art three years later, mainly work by Edvard Munch, but he also had fine pictures by Van Gogh, Cézanne, Pissarro and Degas (Figs.69 and 70).37 The family has no documents recording these purchases, but interestingly they do say that Mustad bought Sunset at Montmajour at an early date, possibly even in 1908. If true, some considerable time passed before he bought his second work by a French master – in 1918. That was Cézanne’s Seated peasant, which is cat. Exposition d’œuvres de Vincent van Gogh, Paris (Galerie Bernheim-Jeune) 1901, nos.24–29. 34 Leclercq promised to return them by mid-May, when he was going abroad (letter to Jo Bonger, 15th April 1901; Amsterdam, Van Gogh Museum Archive, no.B4142V/1984), but the correspondence stops there and Leclercq died suddenly in October. For Leclercq and his communications with Theo’s widow, see M. Supinen: ‘Julien Leclercq: a champion of the unknown Vincent van Gogh’, Jong Holland 6 (1990), pp.5–14. 35 In an earlier letter Leclercq had asked about the exhibition catalogue, which did not in fact exist; Amsterdam, Van Gogh Museum Archive, no.B4140V/1984. 36 Between c.1899 and 1908 Fabre acquired and sold the following paintings by Van Gogh: F 317, F 419, F 434, F 445, F 486, F 587, F 613, F 664, F 669, F 750 and F 800; see Feilchenfeldt, op. cit. (note 26). Feilchenfeldt did not associate F 750 and F 800 with Fabre, but in January 1908 they were both hanging in an exhibition in Galerie Druet, with Fabre identified as the owner; see exh. cat. Vincent van Gogh, Paris (Galerie Druet) 1908, nos.5 (‘Village au ravin’; F 750) and 9 (‘Paysage à Auvers’; F 800). For the latter work, see also Meier-Graefe, op. cit. (note 31), p.120, note 1, although he spoke of ‘Vue d’Auvers’. F 750 can be traced back to Jo’s collection – see Feilchenfeldt, op. cit. (note 26), p.250 – and we believe that Fabre bought it at the same exhibition in Leclercq’s apartment. Feilchenfeldt believed that Harry Graf Kessler was the next owner after Johanna van Gogh-Bonger, and that this German collector also acquired F 800, but that seems unlikely. In May 1908, both works and Fabre’s F 613 were auctioned in Paris; sale, Hôtel Drouot, Paris, 16th May 1908, lots 24–26. 37 Mustad eventually owned thirty-one paintings by Munch; see G. Woll: Edvard Munch. Complete paintings. Catalogue raisonné, New York 2009, nos.60, 80, 93, 94, 104, 113, 115, 117, 158, 226, 266, 284, 324, 339, 340, 358, 369, 381, 388, 469, 510, 514, 698, 1083, 1126, 1129, 1195, 1296, 1299, 1341 and 1361. His earliest documented purchases are nos.369 and 381. In 1937 Mustad donated seven Munchs to the Nasjonalmuseet, Oslo, and on his death in 1970 it turned out that he had bequeathed a further ten to the Museum; see S. Willoch: Edvard Munch. Charlotte og Christian Mustads gave, Oslo 1970. the burling ton maga zine • clv • o cto ber 2 013 701 A NEWLY DISCOVERED PAINTING BY VAN GOGH 68. Detail in transmitted light of the number ‘100’ from the Bonger list inscribed on the reverse of Sunset: wheatfield near Arles, by Vincent van Gogh. 1888. Canvas, 74 by 93 cm. (Kunstmuseum,Winterthur). 67. Reverse of Fig.56 showing the number ‘180’ from the Bonger list inscribed on the canvas. now in the Metropolitan Museum of Art, New York, and it was only then that he really acquired a taste for modern foreign art.38 Van Gogh’s Street in Saintes-Maries and Park in Asnières then followed, probably in the early 1920s, while pictures by Renoir and Sisley were acquired before 1922, and Degas’s A woman ironing, which is now in the Neue Pinakothek, Munich, was bought before 1929.39 According to the family’s recollection, Mustad and his wife were advised by Jens Thiis (1870–1942), who in 1908 was appointed Director of the Nasjonalmuseet, Oslo.40 Thiis was a fan not only of Munch but also of modern French art, as demonstrated by his purchases for the Museum of Van Gogh’s Self-portrait and Cézanne’s Still life in 1910, so it is very possible that the young collector asked the advice of this key figure in the local art world when he decided to start buying modern art from beyond Norway.41 Thiis is known to have visited Paris in 1908 and 1910, and it may have been after one of those trips that he advised Mustad to buy Sunset at Montmajour, possibly because of its affinity with View of Arles from Montmajour (Fig.71), which his Museum had acquired in 1905.42 Nevertheless, it is still incomprehensible how what seems to have been the first picture by Van Gogh to enter a private Norwegian collection managed to escape mention in the literature 38 Nils Messel, chief curator of the Nasjonalmuseet, Oslo, has told us that to the best of his knowledge Mustad did not have any works by leading masters before the First World War. What is interesting, and perhaps significant (see the main text), is that when he again wanted to buy modern French art, he used Walter Halvorsen as his adviser, on whom see H. Spurling: The unknown Matisse. A life of Henri Matisse, London 1998, II, pp.190, 193, 195 and 218, and Halvorsen’s letter of 18th May 1966 to Mustad’s personal secretary (private collection). 39 F 315 and F 420 are two works by Van Gogh that Mustad seems to have bought from the Danish collector Christian Tetzen-Lund in the early 1920s; see K. Monrad: ‘The merchant with the sharp eye and unlimited ambition’, in idem, ed.: exh. cat. Henri Matisse. Four great collectors, Copenhagen (Statens Museum for Kunst) 1999, pp.137–55. Never before, incidentally, has F 420 been associated with Tetzen-Lund. In 1917 he lent it to the exhibition Fransk Konst. Fran 1800-Talet at the Nationalmuseum, Stockholm, where it was no.60: ‘Hus I Saintes Maries. – Les mas de Saintes Maries, Tetzen Lunds samling’. In 1922 Mustad offered Thannhauser works by ‘Renoir, Sisley, Van Gogh’ (Cologne, Zentral Archiv des Internationalen Kunsthandels E.V. ZADIK, Galerie Thannhauser Kundelkartei [client card], Berlin, M, 702 o c to be r 20 1 3 • c lv • t he b u rl in g t on ma g a z i ne on Van Gogh for so long, even if one assumes that Mustad was reluctant to lend works to exhibitions. However, the family have an explanation for this. According to a story handed down in the family, the French ambassador to Sweden visited Mustad soon after he had bought the picture and suggested that it was a fake, or had been wrongly attributed. The collector promptly banished it to the attic. He evidently attached more importance to that verdict than to his adviser’s eye, and was clearly very annoyed, because the banishment was permanent; he never wanted to see the landscape again, and later photographs of his home confirm that it did not hang among his other pictures (Figs.69 and 70). To put all this into historical perspective: expressing doubts about what later turned out to be perfectly genuine works by Van Gogh was not unknown around 1910. The art world was jittery at the time, possibly because of a rise in the number of forgeries in circulation, and as a result owners felt uncertain. For example, the Swiss collector Fritz Meyer-Fierz started to ask Jo van GoghBonger for certificates of authenticity two years after buying four Van Goghs from her.43 He did not want to run the slightest risk, because art dealers and connoisseurs were now bombarding each other with contradictory opinions. For example, in 1912 Bernheim-Jeune started questioning the authenticity of Van Gogh’s Fishing boats on the beach at Saintes-Maries,44 and when one of the versions of the intriguing portrait of Mme Ginoux changed hands in 1914 the art dealer Perls later recalled that there had been ‘rumours, especially among artists’,45 one of whom was the German painter and Matisse pupil Hans Purrmann.46 The fact that Mustad attached so much weight to his guest’s opinion would seem to mean that the ambassador was an authority qualified to pass judgment. That cannot have been the case,47 so we must allow for the possibility that the story became 1927–37, Sign. A077_XIX_0195_001), although no transaction took place, but he must already have had one or both of the works from Tetzen-Lund’s collection, because he would not have offered Sunset at Montmajour for sale, for reasons that will become clear below. The Degas and the Cézanne were said to be in Mustad’s possession in 1929; see P. Jamot: ‘L’Art français en Norvège, galerie nationale d’Oslo et collections particulières’, La Renaissance 12/2 (1929), pp.84 and 95. 40 See O. Mæhle: Jens Thiis. En kunstens forkjemper, Oslo 1970; and N. Messel: ‘French paintings on Norwegian soil’, in T. Gunnarsson et al.: exh. cat. Impressionism and the north. Late 19th-century French avant-garde art and the art in the Nordic countries 1870–1920, Oslo (Nasjonalmuseet) and Copenhagen (Statens Museum for Kunst) 2002–03, pp.231–40; and H. Larsson: Flames of the south. On the introduction of Vincent van Gogh to Sweden, Eslöv 1996, passim. 41 Doubt was later cast on the authenticity of the self-portrait F 528; for a discussion of this, see M.I. Lange: ‘The provenance of Vincent van Gogh’s “Self-portrait” in Oslo’, THE BURLINGTON MAGAZINE 148 (2006), pp.113–16. 42 For the date of purchase, see Ives, op. cit. (note 11), p.162. 43 They were F 459, F 625, F 766 and F 785; see also W. Feilchenfeld and R. Dorn: A NEWLY DISCOVERED PAINTING BY VAN GOGH 69. Christian Nicolai Mustad’s house, Oslo. Photograph 1960s. (Private collection). 70. Christian Nicolai Mustad’s house, Oslo. Photograph 1960s. (Private collection). distorted on its way down from one generation to the next. That is not inconceivable, because if one substitutes the Norwegian consul in Paris for the French ambassador in Sweden, then understanding begins to dawn. Since 1906 the consul had been the entrepreneur Auguste Pellerin (1853–1929) who, as a major collector of late nineteenth-century art, was certainly regarded as an authority.48 He owned many works by Cézanne, and by Van Gogh too; he was also the owner of Astra Margarine, which was in direct competition in Norway with the Mustad family factory. If such a foreign competitor who was also very knowledgeable about art sowed doubts about the authenticity of Sunset at Montmajour, then one can well understand that Mustad would certainly have been upset and that despite the status and reliability of his own adviser he would have wanted to forget his foray into buying modern French art as soon as possible, and as a result did not even take the trouble to discover whether his guest’s view was shared by others. Most of Mustad’s collection was appraised for sale after his death in 1970, and the heirs also showed the exile from the attic to the art dealer Daniel Wildenstein. He ‘believed it was false’ and suggested that ‘it could maybe have been made by a German painter’.49 Later owners also took the trouble to check whether there was any truth in the old attribution to Van Gogh. In 1991 they contacted the Van Gogh Museum and, although the picture was felt to be interesting, it was eventually decided that it was not by Van Gogh.50 That may be a painful admission, given that the same Museum is now attributing it to Van Gogh, but that is not incomprehensible, and nor was the initial rejection. In Bonger’s inventory, Sunset at Montmajour is described as a work from Arles, something that would have been known to the later owners, up to and including Mustad, but it seems as if none of them knew the exact location. Not one of the descriptions refers to the only specifically topographical element in the scene: Montmajour Abbey. It is certainly not easy to recognise (there is no other work by Van Gogh in which it is so small and seen from the west), and after doubts had been expressed about the work’s authenticity the puzzle about the subject would have lent added weight to its rejection. It is telling that, according to the story handed down in the family, Wildenstein accompanied his suggestion about a German artist with the observation that the building was a castle rather than a ruined abbey. Admittedly, recognition of a subject should have no influence on the assessment of a style, but psychologically it is a tour de force to arrive at an attribution if you cannot really identify the scene.51 On top of that, despite the many stylistic affinities with ‘Genuine or fake? – on the history and problems of Van Gogh connoisseurship’, in T. Kōdera and Y. Rosenberg: The mythology of Vincent van Gogh, Tokyo, Amsterdam and Philadelphia 1993, p.266; and J.F. Heijbroek and E.L. Wouthuysen: Kunst, kennis en commercie. De kunsthandelaar J.H. de Bois 1878–1946, Amsterdam and Antwerp 1993, p.31. 44 F 1429; and Heijbroek and Wouthuysen, op. cit. (note 43), pp.39–40. The Dutch connoisseur Bremmer expressed his doubts about F 688 in 1910, after Mrs KröllerMüller bought it from Cassirer, to which the German dealer rightly reacted in a letter to Jo van Gogh-Bonger by asking ‘Ist der Herr verrückt geworden?’ (‘Has the gentleman gone mad?’); see Amsterdam, Van Gogh Museum letter archive, B4075V/1989, cited in J. ten Berge et al.: The paintings of Vincent van Gogh in the Kröller-Müller Museum, Otterlo 2003, p.390. 45 F 489, noted in H. Perls: Warum ist Kamilla schön? Von Kunst, Künstler und Kunsthandel, Munich 1962, p.119: ‘. . . munkelten besonders die Mahler’. 46 Purrmann had earlier doubted the authenticity of both F 520 and F 814; see ibid., pp.119–20; for him as a Matisse pupil, see Spurling, op. cit. (note 38), pp.14–15. 47 The French ambassadors to Sweden in this period were Henri Allizé (1907–09), Jules-Albert Defrance (1909–10) and Napoléon-Eugène-Emile Thiébault (1910–18). With thanks to the French embassy in Stockholm and François Fensterbank of the Centre des Archives Diplomatiques de Nantes. 48 Pellerin made a financial contribution towards the purchase of Cézanne’s Still life for Oslo’s Nasjonalmuseet in 1910; see Messel, op. cit. (note 40), p.232. It is also useful to know that he was a friend of Matisse – see Spurling, op. cit. (note 38), p.193 – and may have been aware of the doubts of the latter’s pupils about the authenticity of some of Van Gogh’s works (see note 46 above). 49 Letter from the owner to the Van Gogh Museum, 6th and 8th November 1991 (Amsterdam, Van Gogh Museum Archive). 50 Letters to the then owner of 18th, 26th and 28th November 1991 and, after having seen the work, in a letter of 18th December 1991 (Amsterdam, Van Gogh Museum Archive): ‘we think that the picture in question is not an authentic Van Gogh’. 51 The spot was recognised in 2011 by an artist friend of the former owner who had grown up in Arles. He then pointed out the similarity of the painting to the description in letter 636, whereupon the Van Gogh Museum was again asked to give its verdict. the burling ton maga zine • clv • o cto ber 2 013 703 A NEWLY DISCOVERED PAINTING BY VAN GOGH 71. View of Arles from Montmajour, by Vincent van Gogh. 1888. Reed and quill pen and ink over graphite on paper, 59.5 by 70 cm. (Nasjonalmuseet, Oslo). 72. Path to the beach, by Vincent van Gogh. Sketch in his letter no.369 to Theo van Gogh, 29th and 30th July 1883. (Van Gogh Museum, Amsterdam, Vincent van Gogh Foundation). other works from this period, it is perhaps not immediately straightforward to spot Van Gogh’s hand throughout the whole of Sunset at Montmajour because large parts of the foreground and middle ground are not as well observed as is usual. On the right, in particular, to which the eye is led, the strokes are barely descriptive at all, making this passage more reminiscent of the late work of Lovis Corinth, say, than of Van Gogh. There is a similar abstraction in the strip of sand and boulders on the left, which he wanted to illuminate with his ‘shower of gold’ but which is so overwhelmed with dark green, reddish brown and orange-yellow strokes and dots as to make it almost unrecognisable. Van Gogh clearly did not know how to handle either passage and just filled them in willy-nilly, possibly as darkness was falling. That failure would explain why he was so disappointed with Sunset at Montmajour and regarded it as the most prominent example of works that were ‘not at all successful and unfinished’ that he sent to Theo in August.52 However, something that undoubtedly contributed to Van Gogh’s disparagement of his work was that he must have had very high expectations for the subject. He had painted several sunsets in June,53 but the one at Montmajour was his very first attempt to do so by showing the sunlight reflecting off the landscape, which enabled him to display a colourist’s invention by evoking the strong golden glow typical of sunsets in the south of France. In order to draw the light sky and the dark foreground closer together he not only made the fields in the vista lighter than they were in reality,54 but also painted yellow-orange dots and blobs all over, even where the late evening sunlight could not have reached, such as the shaded side of the bushes on the right. However, the uncontrolled manner prevents the dots and blobs from functioning properly as lighting effects. They look as if they are part of the foliage and, although the darkening of the yellow-orange will have contributed to that, one can well understand why the scene was once described as an ‘autumn landscape’.55 However important it might have been, the convincing suggestion of the reflection of a sunset on the landscape was actually a secondary aim. What Van Gogh really wanted to depict was the intimacy of nature at this spot, untouched by modern life. The ‘virginity of nature’ brought him peace and awakened existential, religious feelings, and they, after all, lay at the heart of his calling as an artist.56 Sunset at Montmajour was thus an extension of a work such as his watercolour of a path to the beach near The Hague (Fig.72), where ‘the stillness, nature alone’ of the windy stretch had struck him so forcibly in 1883: ‘sometimes it’s precisely those spots that one needs to achieve calm’.57 He linked this particular feeling with Jacob van Ruisdael’s Bush in the Musée du Louvre, Paris (Fig.73), a work he knew well and copied in drawings at the beginning of his career.58 What he admired in Ruisdael’s painting was ‘that dramatic quality that causes one to find a je ne sais quoi in it’, and hence ‘expresses that moment and that place in nature where one can go alone, without company’.59 And although Van Gogh referred to Monticelli in his letter about Sunset at Montmajour in order to stress the immemorial Provençal nature of the area, the place awakened in him, above all, the drama of Ruisdael’s Bush, for to his mind the 52 after Ruisdael’s Bush. 59 Ibid., no.361. 60 Ibid., no.639. 61 Ibid., nos.635 and 636. 62 Drawings preceded his earlier size 30 paintings (see note 16 above) with the exception of F 465, but that is such a simple composition that it was unnecessary to do that kind of preparatory work. 63 See Letters, no.639, and also no.658 of 9th August, in which he reported that the Letters, no.660; he mentioned F 448, F 426, F 422 and F 427. F 422, F 465, F 427. 54 In his description of the spot he said that the fields were ‘purple’; see Letters, no.636. 55 See note 30 above. 56 The quotation is from Letters, no.369. 57 Ibid. 58 Ibid., no.361. See nos.157–59 for Van Gogh’s copies after the etching by Daubigny 53 704 o c to be r 20 1 3 • c lv • t he b u rl in g t on ma g a z i ne A NEWLY DISCOVERED PAINTING BY VAN GOGH 73. The bush, by Jacob van Ruisdael. c.1647. Canvas, 168 by 82 cm. (Musée du Louvre, Paris). 74. Olive trees, by Vincent van Gogh. 1888. Reed pen and ink on paper, 48 by 60 cm. (Musée des Beaux-Arts, Tournai). whole region around Montmajour was redolent of the spirit of the Dutch painter. ‘I’ve already told you more than once how much the Camargue and the Crau – apart from a difference in colour and the clearness of the atmosphere – make me think of the old Holland of Ruisdael’s day’.60 Although Van Gogh must have been thinking of painting in this area since May, he began this particular scene by accident. He had planned to go and explore the Camargue at the beginning of July, but when that was cancelled because his guide pulled out he dashed off to Montmajour instead,61 which was just as primitive. That train of events may explain why this first attempt at capturing the area was such a failure. Ambitious paintings need time to ripen and grow to fruition in the mind, and Van Gogh seems not to have given himself that period of incubation. He had discovered that there was a painting in the oaks bent by the wind with the view of the abbey ruins and the fields in the distance, but unlike his equally large Harvest he had not reconnoitred the site first in drawings but immediately set about making a painting, and had to face the attendant problems as he was working.62 When Van Gogh returned to the area three days later in order to carry on with his campaign he blamed his lack of control over his brush on the mistral.63 He dropped the idea of painting and started to make large drawings of the area, but still with the model of Ruisdael in his mind’s eye.64 The control and command of the scenes that he drew with his reed pen is unparalleled (Fig.74), and it is difficult to see this in isolation from his disappointment over the lack of precision and style in Sunset at Mont- majour. It is as if he was out for revenge, but was forced to take it in another medium. However, he did not forget the point of departure, for at the end of his campaign of drawing he tried once again to paint the subject of the holly oaks in the mistral at sunset, and the result was The rocks (Fig.60). By then he had learned more and adjusted his sights accordingly. He chose a smaller format, simplified the subject and composition, kept the brushstrokes as terse as possible, and largely omitted the reflection of the sun’s rays. Van Gogh was a little dismissive of his achievement in The rocks, and that is a little odd, given the undeniable attractiveness of the painting, but it is understandable.65 The rocks was born of an attempt to discover whether even in a high wind he could retain some precision, whereas Sunset at Montmajour was painted to achieve the highest ambition of all: to show that there was ‘no pettiness to lofty feelings’, as Fromentin said of Ruisdael.66 Like his seventeenth-century predecessor, Van Gogh wanted to convince as ‘one of the solitary ramblers who fly from the town, frequent the outskirts, sincerely love the country, who feel it without exaggeration and describe it without phrases, who are made uneasy by distant horizons, charmed by large plains, affected by a shadow, and enchanted by a ray of sunshine’.67 Although the high winds coupled with the short preparation time had prevented him from producing a truly persuasive work, the shortcomings of his painting perhaps strike us less today than the seduction of his ambition to lift the scene to such a high poetic level, ‘because nature like this can sometimes awaken in a mind things that would otherwise never have woken’.68 mistral had been blowing uninterruptedly for six weeks. 64 For this series of five drawings, to which he wanted to add a sheet that he had already sent to Theo in May (Fig.71), see M. Vellekoop and R. Zwikker: Vincent van Gogh. Drawings. Arles, Saint-Rémy & Auvers-sur-Oise 1888–1890. Van Gogh Museum. Volume 4, Zwolle and Amsterdam 2007, pt.1, pp.135–44. The switch from paintings to drawings has also been explained as a consequence of Van Gogh’s need to economise – see Ives, op. cit. (note 11), p.236 – but the discovery of this new painting makes that explanation less plausible. 65 ‘I’ve started a painting too – but no means of doing it with the mistral, absolutely impossible’; Letters, no.639. Van Gogh was by no means as dissatisfied as this quotation suggests; see note 12 above. 66 E. Fromentin: Les maîtres d’autrefois: Belgique, Hollande, Paris 1876, here cited after the English translation: The Masters of Past Time, Oxford 1981, p.143. 67 Ibid., pp.144–45. 68 Letters, no.401. the burling ton maga zine • clv • o cto ber 2 013 705 Reprinted from OCTOBER 2013, NO . 1327 VOL . CLV