Van Gogh Museum Journal 1995

[p. 62]

cat. 8
Basket of apples, Nuenen, September 1885 The X-ray shows a bouquet of honesty in a pot and smoker's requisites belonging to Van Gogh's father

[p. 63]

Radiographic images of Vincent van Gogh's paintings in the collection of the Van Gogh Museum
Sjraar van Heugten

After learning to paint in Holland and Antwerp, Vincent van Gogh developed into a truly modern artist in the course of a two-year sojourn in Paris. During those years of constant study and growth he is known to have painted over a number of works, apparently because he was dissatisfied with them. Raking light can occasionally give a rough idea of the earlier picture.¹ At times this image corresponds to traces of brushstrokes beneath the paint bearing no relation to the composition on the surface. At other times there is no sign of the old impasto, even though the radiograph leaves no doubt that the artist had used the canvas before. The opposite is also possible: in the Still life with fruit in the Van Gogh Museum, brushstrokes can be seen with the microscope which evidently belong to an earlier scene, although there is no trace of the scene in the radiograph.²

To determine which canvases in the collection of the Van Gogh Museum were reworked by Vincent, all the pictures dating from the period in question, some 130 in all, were examined with X-rays.³ In 20 cases - some 15% - works that had been covered over came to light. Two of these were simply earlier stages of the same composition (cat. nos. 1 and 2), while 17 others differ entirely from that which can be seen on the surface. One of the canvases stands alone: doubts raised about its attribution to Van Gogh were confirmed.

X-rays have also revealed hidden works in other collections.⁴ If the results of the Van Gogh Museum's research are any indication, however, there must be much more beneath the surfaces of the 400 paintings Vincent produced before he left for Arles. The importance of these

1 See numbers 1, 4, 6, 7, 8, 14 and 17 in the following catalogue.

2 The still life is F 383 JH 1339. Pigments appear in a radiograph in direct proportion to the density of the metals - e.g. lead, mercury, cobalt-they contain. Such pigments permit relatively little radiation to reach the film and therefore seem lighter. Less dense pigments permit more radiation to penetrate and thus look darker as a result. If there are no dense pigments in a given picture then of course they will not show up in a radiograph. The lack of such pigments in the now visible surface of several paintings described below (cat. nos. 3, 6, 10, 14, 16 and 17) makes it that much easier to read what lies below.

3 The radiographs were made by the Röntgentechnische Dienst of Rotterdam during several sessions at the Van Gogh Museum.

4 Several of these works are discussed below, in connection with pictures revealed by X-rays: see cat. nos. 4, 5, 6 and 18.
X-ray examination showed that a large painting with several peasant cottages (SP 1669 JH 825, present whereabouts unknown) conceals a ploughman Van Gogh painted in Nuenen as part of a wall decoration for Antoon Hermans; see Alan Bowness, ‘A Van Gogh discovery,’ The Burlington Magazine 111 (1969), 299-300. The diminutive Study of two hands (F 66 JH 743, private collection) is painted over the depiction of a cottage with a woman in peasant dress; see Sjraar van Heugten, ‘Detail studies,’ in Louis van Tilborgh et al., exhib. cat. The potato eaters by Vincent van Gogh, Zwolle 1993, pp. 105-07 and no. 29 (without ill. of the radiograph). The Self-portrait in the Wadsworth Atheneum in Hartford, Connecticut (F 268 JH 1299) covers a picture of a woman at the spinning wheel which the artist must have painted in Nuenen. This information notwithstanding, the portrait's authenticity was recently challenged by Roland Dorn and Walter Feilchenfeldt, ‘Genuine or fake?: on the history and problems of Van Gogh connoisseurship,’ in Tsukasa Kodera (ed.), The mythology of Vincent van Gogh, Tokyo 1993, pp. 290 and 296. Both the work and the radiograph are illustrated on p. 298.
Van Gogh painted the Still life with two sunflowers in the Kunstmuseum in Bern (F 376 JH 1331) over the portrait of a man wearing a straw hat, seated at a table and holding a glass in his left hand; see Anne Trembley, ‘Vincent van Gogh: Sonneblumen 1887,’ Berner Kunstmitteilungen 258/259 (1987-88), pp. 1-10; she raises the possibility that the concealed work is actually a self-portrait. To obtain a better image of it, radiographs, infrared reflectograms and photographs with light shining through the canvas from behind were made.

[p. 64]

findings is self-evident: they shed valuable light on Van Gogh's subject matter, working methods, and early development. For example, the artist's correspondence tells us that he tried his hand several times at the theme of a shepherd with his flock. Though this motif does not occur in his mature oeuvre, X-rays show that he did indeed treat it more than once (cat. nos. 5 and 6). By the same token, we now know that certain works the artist mentioned in his letters are not lost, as has always been assumed, but simply covered with paint. Thanks to the radiographs, we have at least some idea of what the portrait of the shepherd, the woman at the spinning wheel, and the still life with the vase of honesty looked like (cat. nos. 4, 5 and 8).

Vincent never stated in his letters why he used a canvas a second time. Indeed, only in the case of Cottage at nightfall (‘La Chaumière’) did he even admit to painting a picture over another one (cat. no. 6). Financial necessity must certainly have played a role. The monthly allowance his brother Theo sent him would have been more than enough, had it not been for his high productivity and the cost of his models - a constant drain on funds. ‘The dark side of painting is the paint bill,’ the artist sighed in a letter of circa 7 December 1885 [475/388]; on another occasion he complained that ‘My paint bill is [...] such that I have to be a bit careful about starting on new things in a larger format, especially since the model will cost me rather a lot’ [454/372].

That Van Gogh was occasionally forced to recycle his canvases does not mean masterpieces were lost. He painted over the works described in the following catalogue either because they were studies of no commercial or artistic value, or because he was simply unhappy with them. Given what is known about his attachment to his work, Vincent would not have discarded anything of artistic interest to him. In Arles, when he really could not afford paint supplies, he turned to drawing, which was much less expensive. During the previous years of study in Nuenen and Paris, the painter was so eager to learn that he would not have been able to abide such delays; hence the care with which he used (and occasionally reused) his canvases.

In the case of approximately half of the works in the following catalogue, reasons can be given as to why Van Gogh opted to destroy them. For example, the picture of the Protestant church at Nuenen (cat. no. 2) initially contained all too obvious references to a painting by Millet. Or, as in the Two women in the peat-field with a wheelbarrow (cat. no. 1), he was unhappy with the composition. His various treatments of shepherds (cat. nos. 5 and 6) apparently fell short - and, indeed, the compositions are rather dull and repetitive. The flower still life in memory of his father (cat. no. 8) was simple at first; adding a new element to the composition seems to have spoiled it in the end. Something similar may have happened to the view on Montmartre (cat. no. 14): the artist probably overworked it, and then had no choice but to discard it altogether. As for the study of a plaster copy after Michelangelo's Young slave (cat. no. 12), Van Gogh doubtless decided the subject had too much space; he may also have been unhappy with the figure's proportions. The physiognomies of the shepherd and the spinner gave him trouble as well (cat. nos. 4 and 5); he seems to have overestimated his capabilities in choosing a rather large format for both (cat. nos. 4 and 5).

[p. 65]

1 Two women in the peat-field with a wheelbarrow
Drenthe, October 1883
The X-ray shows the first version of this canvas,
October 1883
Oil on canvas, 27.5 × 36.5 cm
F 19 JH 409

As revealed by the X-ray, at an earlier stage the composition of this canvas contained more figures and more piles of peat. The artist painted over several parts of it, as can be seen to some extent with the naked eye. The original composition was virtually identical to a letter sketch from Drenthe, but lacked the area to the left of the wheelbarrow in the drawing (fig. 1a).¹ As the artist wrote to his brother, ‘I hope to make something of the women in the peat-field in the sketch on the back and am returning to that same field’ [396/331]. The initial design of our canvas may possibly have been the result of that plan. The X-ray shows that Vincent subsequently took it up again, simplifying the composition and also altering the colour scheme. The sky behind the women was originally pink, but Van Gogh changed this picturesque sunset into a depressing grey twilight. The composition revealed by the radiograph recurs as a detail in a drawing from Drenthe, where wheelbarrow, women, and piles of peat appear like silhouettes on the horizon.²

fig. 1a
Vincent van Gogh, sketch in letter 396/331,
6-7 October 1883, Amsterdam, Van Gogh Museum
(Vincent van Gogh Foundation)

1 1 Based on observation with the naked eye, I previously suggested a link between this sketch and the painting; see exhib. cat. Vincent van Gogh and his time: Van Gogh & Millet from the Vincent van Gogh Museum and the H.W. Mesdag Museum, Tokyo (Seiji Togo Memorial Yasuda Kasai Museum of Art) 1993, no. 1.

2 2 F 1095 JH 406.

[p. 66]

2 Leaving the church at Nuenen,
Nuenen, winter-spring 1884.
The X-ray shows an earlier version of this canvas,
January 1884.
Oil on canvas, 41.5 × 32 cm
F 25 JH 521

As in the case of the Drenthe landscape with toiling women, the radiograph indicates that changes were made to the original design of Leaving the church at Nuenen, and that here, in particular, these have important implications regarding the content of the work.¹

fig. 2a Vincent van Gogh, sketch in letter 492/355, c. 24 January 1884, Amsterdam, Van Gogh Museum (Vincent van Gogh Foundation)

1 The X-ray was made several years ago. The conclusions drawn from it were published by Evert van Uitert in Evert van Uitert and Michael Hoyle (eds.), The Rijksmuseum Vincent van Gogh, Amsterdam 1987, no. 14. The X-ray first appeared in Louis van Tilborgh (ed.), exhib. cat. Van Gogh and Millet, Amsterdam (Van Gogh Museum) 1988-89, p. 61.

[p. 67]

Where now a group of churchgoers stand, one of whom wears a veil, Van Gogh first painted a man with a spade over his shoulder. He appears in a drawing of about the same time, now in the Kröller-Müller Museum, Otterlo, and in a letter sketch of circa 24 January 1884 (fig. 2a) 1429/355].² As it was initially planned, the canvas quoted far more literally than is now the case from one of Van Gogh's favourite works by Jean-François Millet. The French artist's Church at Gréville, which at that time hung in the Musée du Luxembourg, likewise shows a man with a spade over his shoulder walking past a church. Having apparently decided that the quotation was in fact too literal, Vincent changed the design, also adding leaves to the originally barren trees. Judging from the paint surface, the underlying layer had undoubtedly dried by the time the foliage was painted.³ Van Gogh, who was very fond of autumnal effects, apparently found this a more suitable setting for the aggrieved worshippers.

3 Still life with birds'nests
Nuenen, September/October 1885
The X-ray shows part of an interior with a loom,
Nuenen, between January and July 1884
Oil on canvas, 31.5 × 42.5 cm
F 109r JH 942

Van Gogh seems to have gotten as much out of this particular canvas as he possibly could. First of all, he used both sides: the front for the Still life with birds' nests, painted while he was still in Nuenen, and the back for a self-portrait after moving to Paris. The X-ray reveals that it was originally used for yet another painting: it formed part of a larger picture of an interior with a loom.

Between January and July 1884, Van Gogh was fascinated by the theme of the weaver and his loom. The larger work from which the present canvas was cut must have been painted during this period. A corner of a window can be seen in the upper left of the X-ray, and part of a loom with a lamp suspended above it in the lower right. We may assume that a drawing in the Van Gogh Museum depicts the same machine: it has the characteristic long beams resting on wooden supports, and a lamp hanging in the same place (fig. 3a).¹

2 The drawing is F 1117 JH 446.

3 It is not clear exactly when Van Gogh altered the composition. Presumably, and in light of the autumnal effect, an October dating can be upheld. In his Vincent van Gogh: a guide to his work and letters, Amsterdam 1993, pp. 20-21, however, Jan Hulsker recently noted that the canvas was intended for the artist's mother. He would have needed her permission to change it, which, given his strained relationship with his parents in the autumn of 1884, would have been ‘out of the question.’ Hulsker therefore argues that Van Gogh reworked the canvas already in January of that year, ‘whether of his own free will or at her [i.e. his mother's] request.’ However plausible Hulsker's reasoning may be, January is not the only date that can be inferred from it. I prefer to maintain a somewhat wider margin by placing the final version of the canvas in the winter-spring of 1884.

1 Van Gogh painted only one other picture with a weaver showing a loom with beams supported in this way, a work which is now in Boston (F 29 JH 471). There, however, the position of the lamp is different.

[p. 68]

The artist must have let this canvas sit for a long time. The Still life with birds' nests dates from September 1885, at least a year after he painted his weaver scenes. Since the thick layer of paint must have already hardened, it would have been difficult to conceal. Indeed, the impasto of both the window and the loom is visible; the paint of the supporting beam even comes to the surface. This probably explains why Van Gogh decided to reuse the canvas for smaller, less important studies instead of for a large new composition with which the impasto would have interfered. He even exploited the dark, underlying layer: instead of painting over the twigs jutting out from the birds' nests, he used their colour in the new painting.

The size of the original work is difficult to gauge with certainty. All four edges have suffered and may have been trimmed. Representing the entire loom would have required a canvas approximately three times the present width of 31.5 cm. Two of the artist's other canvases with weavers are indeed 93 cm wide.² Since one of them is 68.5 cm high and the other 61 cm, our canvas may have originally been about 65 cm in height; a canvas of that size would have been one of the larger formats in Vincent van Gogh's oeuvre.

fig. 3a
Vincent van Gogh, Weaver at his loom, February 1884, Amsterdam, Van Gogh Museum (Vincent van Gogh Foundation)

4 Jar, brass bowl and potatoes
Nuenen, September 1885.
The X-ray shows a shepherd with staff,
Nuenen, October 1884.
Oil on canvas, 65.5 × 80.5 cm
F 51 JH 925

The Jar, brass bowl and potatoes is painted over a large, three-quarter-length portrait of a man wearing a cloak with large buttons and holding a staff in his hands. The portrait can almost certainly be identified with that of the shepherd about whom Vincent wrote Theo circa 22 October 1884: ‘I am working on a figure of a shepherd wearing a large cloak, of the same format as the spinning woman. [...] If I have some luck with that shepherd, it will become a figure with something of the very old Brabant. Well, he's not yet finished and we shall see how he turns out’ [468/382]. The spinner in question is in all probability one of the two renderings of a woman at the spinning wheel, which are likewise only known as radiographic images. One of them lies beneath the work discussed under the following number in the present catalogue, which measures 78.5 × 65.5 cm. The other is concealed by The Vicarage garden under snow in the Norton Simon Museum in Pasadena (fig. 4a), which is 80.5 × 58.2 cm.¹ The shepherd with staff is indeed approximately ‘of the same format,’ namely 80.5 × 65.5 cm.

The shepherd with staff was one of Van Gogh's first works in the series of portraits and figure studies of Brabant farmers he began working on in October 1884. He would continue working on them until the spring of 1885, preparing himself for the Antwerp academy. ‘But if I first paint some thirty more heads here, I will get more out of Antwerp-and on those 30 heads I am now starting, or rather, I have already started with a large bust of a shepherd,’ he wrote in late October 1884 [469/383]. As a three-quarter-length portrait of a strikingly large format, the work has no equivalent in Van Gogh's Dutch years. And, as the artist later realised - witness the overpainting - he may have overestimated his proficiency at the time: the man's proportions are unfortunate, the head in particular being inordinately small.

The shepherd here heralds the portraits of Patience Escalier that Van Gogh would paint at Arles in August 1888. It especially resembles the version in a private collection which shows Escalier leaning on a stick (fig. 4b).

2 F 24 JH 500 and F 37 JH 501, respectively.

1 The X-ray of the work in Pasadena (F 194 JH 603) was published in a number of newspapers in 1972. It has also appeared in several scholarly publications: René Berger and Lloyd Eby (eds.), Art and technology, New York 1986, pp. 15-27; Martha M. Op de Coul, ‘Nuenen and its environs in the work of Vincent van Gogh,’ in Evert van Uitert (ed.), exhib. cat. Van Gogh in Brabant: paintings and drawings from Etten and Nuenen, Bois-le-Duc (Noordbrabants Museum) 1987, pp. 92-101 (ill. p. 98, reversed). And it was included in the press release about the exhibition Vincent van Gogh: painter, printmaker, collector, Pasadena (Norton Simon Museum of Art) 1990.

[p. 69]

fig. 4a
X-ray of Vincent van Gogh's The vicarage garden under snow, January 1885, Pasadena, The Norton Simon Museum The X-ray shows a woman at a spinning wheel, 1884-85

fig. 4b
Vincent van Gogh, The peasant: portrait of Patience Escalier, August 1888, Private collection

[p. 70]

5 Two baskets of potatoes
Nuenen, September 1885
The X-ray shows two works: a shepherd with his flock and a
woman at the spinning wheel, Nuenen, 1884-85
Oil on canvas, 65.5 × 78.5 cm
F 107 JH 933

This canvas with a dark still life with baskets of potatoes was used twice previously by Van Gogh. Upside down, vis-à-vis the still life, the X-ray discloses a shepherd with his flock in the centre of the composition and, in the lower right corner, a dog. If we then rotate the canvas 90 degrees, with some difficulty we can see a woman at a spinning wheel. The radiograph becomes clearer if we compare it to that of the canvas from Pasadena (see cat. no. 4), which depicts another spinning woman beneath The vicarage garden under snow (fig. 4a). The compositions are virtually identical, even if the two works are reverse images of one another. The canvas in the Van Gogh Museum has the spinning wheel in the lower left corner. The spinner herself is difficult to distinguish, being all but obscured by sheep and potatoes, but the back of her chair can be seen on the far right.

Dating the two works is rather difficult. Judging from the letter quoted in cat. no. 4, the spinning woman was painted in October 1884 or somewhat earlier, but there are no clues to help us place the shepherd with flock (see also cat. no. 6). The X-ray reveals, however, that the shepherd image is painted over the barely legible and certainly scraped-down representation of the spinning woman.

Recent research carried out at Amsterdam's Rijksmuseum revealed that the Farming village at twilight (F 190 JH 492) in that collection covers a still life with two baskets of potatoes, and that the composition closely resembles the present work. The Rijksmuseum's canvas, however, is wider and shorter (57 × 82 cm) - a peculiar oblong format for a still life, perhaps, but perfectly suitable for a landscape. Van Gogh may well have adapted the format when he decided to paint the landscape over the previous image.

As a result of this discovery, the Rijksmuseum's picture should be redated. On stylistic grounds, it is usually situated in 1884 (late May of that year, according to Hulsker). Yet the artist's correspondence tells us that he was painting still lifes with baskets of potatoes in September 1885. He must therefore have made Farming village at twilight after that date - probably in early October since the trees still have their leaves.¹

1 I am indebted to Wouter Kloek, Head of the Rijksmuseum's Painting Department, for pointing out to me that another representation is visible in the surface of the paint. Martin Bijl, Head Restorer of the Rijksmuseum, kindly made an X-ray of the work on my behalf.

[p. 71]

6 Cottage at nightfall (‘La Chaumière’)
Nuenen, May 1885
The X-ray shows a shepherd with flock,
Nuenen, 1884-May 1885
Oil on canvas, 65.5 × 79 cm
Signed lower left: Vincent
F 83 JH 777

In a letter to Theo of early June 1885, Vincent wrote that he had sent his brother a number of works, including The old church tower at Nuenen (‘Cimetière de paysans’) and The cottage. Later on, he reported that his first attempt at The old church tower had failed, whereupon he scraped off what he had done and started over again. ‘Now, and the other - that of the cottage - was originally a shepherd’ [510/411].

Only vague traces of an earlier painting are visible in either the impasto or the X-ray of The cottage. It must not have been long before Van Gogh destroyed the first painting, and the paint he used may have contained few pigments dense enough to block X-rays. By comparison, the ‘shepherd’ he covered with The cottage must have already been rather dry: the radiograph gives a clear image of it that is also still faintly visible in the impasto,¹ without, however, adversely affecting The cottage. The X-ray also shows why Van Gogh did not hesitate to paint a new picture over the shepherd with flock: the shepherd - who stands exactly in the centre of the canvas with a crook or branch over his shoulder - and his flock are rendered with thinly applied,



sketch-like areas of paint. Having apparently realised the work was going poorly, he simply opted not to finish it. The theme of the shepherd with his flock, so popular with the Barbizon painters and their Dutch followers of the Hague School, evidently interested Van Gogh more than anyone suspected. He chose it as an allegory of autumn in his design for a wall decoration with the four seasons, commissioned by Antoon Hermans, a goldsmith from Eindhoven. Unfortunately, that large sketch, known only from a photograph, is now lost. He painted over three other renderings of

fig. 6a X-ray of Vincent van Gogh's Still Life with pottery and two bottles, September-October 1885, Pasadena, The Norton Simon Museum.
The X-ray shows a shepherd with flock, 1884-85.

1 This was previously noted by Evert van Uitert in Van Uitert and Hoyle, op. cit. (cat. no. 2, note 1), p. 138.

[p. 72]

the theme as well: they are to be found underneath the Two baskets of potatoes in the Van Gogh Museum (see cat. no. 5) and the Still life with pottery and two bottles in the Norton Simon Museum in Pasadena (fig. 6a). The composition of all three works is very similar - the shepherd is always at the centre of the canvas, surrounded by his flock-but the Pasadena canvas is less ambitious than the others: its format is relatively small (40 × 56 cm) and the picture itself is less impressive than the two in the Van Gogh Museum. The shepherd with his flock beneath The cottage was far more monumental: the principal motif is placed prominently in the foreground and fills virtually the entire image.

7 Still life with three beer mugs
Nuenen, September-October 1885
The X-ray shows a portrait of a peasant woman,
Nuenen, April 1885
Oil on canvas, 32 × 43 cm
F 49 JH 534

The Still life with three beer mugs is generally grouped together with the other still lifes Van Gogh painted in November 1884. Yet the likeness of a peasant woman underneath this work doubtless belongs to the long series of peasant portraits Vincent produced between December 1884 and May 1885. The radiograph is so clear that the sitter is recognisable as the woman who sat for another portrait in the same group.¹ That work, which is now in the Van Gogh Museum, can be securely dated to late March-early April 1885; the portrait seen in the present radiograph was probably painted at the same time. The dating of the Still life with three beer mugs should therefore be revised: we now know it is one of the still lifes Van Gogh painted in September-October 1885, and that it could not have been made in November 1884.

1 F 160 JH 722.

[p. 73]

8 Basket of apples
Nuenen, September 1885
The X-ray shows a bouquet of honesty in a pot and smoker's requisites belonging to Van Gogh's father,
Nuenen, April 1885
Oil on canvas, 45 × 60 cm
F 99 JH 930

Although Vincent van Gogh's relationship with his father had been troubled for years, the preacher's sudden death on 27 March 1885 came as a heavy blow. Still mourning the loss, on 5 April he wrote to Theo about a new canvas: ‘a still life with honesty in the same style as the one you took with you. It's a bit larger, however, and the objects in the foreground are a tobacco pouch and a pipe of Pa's. If you'd like to have it, of course you're welcome to it’ [493/398]. The importance he attached to the still life is suggested by the accompanying letter sketch, one of the few he made in colour (fig. 8a). Before the vase lies what appears to be the tobacco pouch. The pipe, to its right, leans against another pouch, presumably designed to hold other smoking accessories. Van Gogh must have painted the work in memory of his father, whose attributes are as commonplace as they are meaningful. The artist loved pipes and portrayed himself on many occasions with one in his mouth. Similar objects serve the same symbolic purpose in the famous Van Gogh's chair (London, National Gallery); a sheet of paper with some tobacco and a pipe on a simple wooden chair function as tokens of the painter.

The assumption has always been that the painted still life was lost. The X-ray of Basket of apples, a work painted in September 1885, shows that this is indeed the case, but that the loss is less definitive than once thought. Van Gogh had also previously used that canvas for a still life, every detail of which is recognisable in the X-ray: the Cologne pot with honesty, the tobacco pouch lying before it (albeit less clear than in the little sketch), and the pipe and pouch on the right. To the left of the vase lie several flowers (apparently asters) that have fallen out of it; Vincent must have added them later, as they do not appear in the detailed sketch he made for Theo. Their addition suggests that Van Gogh was unhappy with the original composition, but the new element evidently failed to solve the problem. He may have decided that the fallen flowers - whether or not he was aware that they are a traditional symbol of transience - struck a note that was too ominous for what was otherwise a meaningful, yet restrained picture; he therefore reused the canvas for the Basket of apples.

fig. 8a
Vincent van Gogh, sketch in letter 493/398, c. 5 April 1885, Amsterdam, Van Gogh Museum (Vincent van Gogh Foundation)

[p. 74]

9 Three bottles and pottery
Nuenen, September-October 1885
The X-ray shows a peasant woman preparing a meal,
Nuenen, probably June 1885
Oil on canvas, 39.5 × 56 cm
F 53 JH 538

The profile of a peasant woman wearing a bonnet can be seen to the right of centre, near the top of this radiograph. The rest of the image is more difficult to discern. With one arm outstretched, the woman seems to be bending forward. Steam appears to rise behind her arm, to the left of her face, recalling the scenes that Van Gogh painted and drew in June 1885 of women boiling water or preparing meals by the hearth. With the discovery of this work, the usual dating of the Three bottles and pottery to the winter of 1884-85 will have to be revised. Like the Still life with three beer mugs (cat. no. 7), it must be one of the still lifes Van Gogh painted in September-October 1885.

[p. 75]

10 Self-portrait with pipe
Paris, spring-summer 1886
The X-ray shows a woman with loose hair,
probably Antwerp, December 1885-February 1886
Oil on canvas, 46 × 38 cm
F 180 JH 1194

With some difficulty, we may here distinguish a woman's portrait. Her loose hair appears strikingly white in the radiograph, meaning the paint is highly absorbent. Although X-rays cannot tell us the colour of whichever pigment absorbs the radiation, we may tentatively assume it is white lead or a yellow pigment containing lead, and that the model had fair hair.¹ As a type, she is related to the women of presumably loose morals whom Van Gogh portrayed in Antwerp.² She may, in fact, be the blonde he referred to in one of his letters: ‘I am looking for pieces such as for instance those blonde heads in Sainte Thérèse en Purgatoire. Precisely for the sake of Rubens I am searching for a blonde model’ [550/439]. Van Gogh probably took this portrait with him to Paris, later scraping down the canvas in order to use it for the self-portrait.

1 A number of the pigments Van Gogh employed have a very high absorption factor. Among these are the white lead carbonate (lead white), the yellow lead chromate (chromo yellow) and lead antimonide (Naples yellow), and the red mercuric sulphide (vermilion cinnabar).

2 See F 206 JH 972 and F 207 JH 979.

[p. 76]

11 A pair of shoes
Paris, spring 1886
The X-ray shows a townscape (?), place and date unknown
Oil on canvas, 37.5 × 45 cm
Signed upper left: Vincent
F 255 JH 1124

Only the general outline of the picture revealed by the X-ray is visible; the reproduction is even less clear. Unfortunately, it sheds no light on the contested dating of the still life.¹ A large, two-storey house can be distinguished. There may also be other buildings, in which case the work could possibly be a townscape. The lack of any further topographical details precludes a more specific identification, and Antwerp, Paris, Amsterdam or Eindhoven (near Nuenen) are good possibilities.

1 Though the canvas is usually assigned to the artist's Parisian period, one scholar has suggested it was executed in Nuenen: Mark Roskill, Van Gogh, Gauguin and French painting of the 1880s: a catalogue raisonné of key works, Ann Arbor 1970, pp. 9-10.

[p. 77]

12 Glass with roses
Paris, summer-autumn 1886
The X-ray shows a plaster cast after Michelangelo's Young slave, Paris, spring-summer 1886
Oil on pasteboard laid down on multiplex, 61 × 46 cm
F 218 JH 1144

At the apartment he shared with his brother Theo in Paris, Van Gogh kept various plaster statuettes on hand, models of ancient and Renaissance sculptures he had copied in accordance with academic tradition. Several of these models were passed down in the family and are now preserved in the Van Gogh Museum. One of them is modelled on one of the two versions of Michelangelo's Dying slave in the Louvre, of which there is no known painted or drawn copy by Van Gogh.¹

Vincent must have also had a statuette after another of Michelangelo's slaves, the so-called Young slave in the Galleria dell'Accademia, Florence. Until now the copy was only known from a rather rubbed drawing, but the same statuette appears in the X-ray of Glass with roses.² Compared to the other copies he painted after such statuettes, the artist here surrounded it with too much space. The work has other problems, such as the inordinate size of the arm over the figure's head, which is probably what prompted him to abandon it.

1 For the plaster statuettes now preserved in the Van Gogh Museum, see Van Uitert and Hoyle, op. cit. (cat. no. 2, note 1), nos. 3.2, 3.3 (there simply entitled Plaster statuette of a male nude), 3.5 and 3.6.

2 To judge from the X-ray and the drawing, Van Gogh's statuette was more finished than the Florentine original is in reality.

[p. 78]

13 Plaster statuette of a female torso
Paris, spring-summer 1886
The X-ray shows the lower half of a plaster statuette of a
female torso, Paris, spring-summer 1886
Oil on pasteboard laid down on multiplex and pasteboard,
35 × 27 cm
F 216j JH 1059

The X-ray of this small study after a plaster torso on a round pedestal shows the same plaster model, but rotated 90 degrees, seen from behind, and cut off at the waist. The cardboard was probably twice as large originally and was simply bisected by the artist for further studies.

[p. 79]

14 Self-portrait at the easel
Paris, July-August 1886
The X-ray shows a windmill and terrace on Montmartre,
Paris, June-July 1886
Oil on canvas, 46.5 × 38.5 cm
F 181 JH 1090

The dark Self-portrait at the easel is painted over what may perhaps be considered Van Gogh's earliest view of a windmill on Montmartre. Through gaps in the uppermost layer of paint and impressions in the impasto, a vague image of it had already emerged.¹ The X-ray brought more details to light, but the work is still difficult to read. In the lower left corner of the canvas a house or shop with a door on the far left can be discerned (this can actually be seen more clearly with raking light). To the right of the door is a small window and what appears to be a text painted on the wall. Behind this structure stands a wooden windmill that seems to be crowned by a terrace. If this is indeed the case, the mill must represent the Moulin le Blute-fin, also known as the Moulin de la Galette. On the right side of the composition is a terrace on poles with a flag, with a man and woman standing below it. There was in fact a terrace with a view near the Moulin le Blute-fin, but it was adjacent to the windmill and not, as in this case, separated from it by a building. This makes it difficult to identify the topography; we can only be certain that a windmill on Montmartre is involved. Van Gogh may have painted the canvas on the spot, decided he was not happy with it, and then begun working on a somewhat different scene. Whatever happened, the X-ray shows that the composition was forced and overcrowded, which is probably what prompted the artist to abandon it and reuse the canvas for the self-portrait.

The likeness was generally considered one of the earliest pictures Vincent painted in Paris, but even Nuenen and Antwerp were not ruled out. With the discovery of the Montmartre scene, however, we can now be certain it was painted in the French capital. The artist is thought to have produced the first of his Montmartre scenes in the summer or early autumn of 1886. Van Gogh's difficulties with the present composition suggest it is the earliest of these townscapes, and was therefore made in June-July. The date of the self-portrait is equally uncertain: it was presumably painted in July-August.

1 The heavy craquelure and paint loss led to the recent restoration of the canvas.

[p. 80]

15 Nude girl, seated
Paris, summer-autumn 1886
The X-ray shows a still life with a glass of flowers,
Paris, summer 1886
Oil on canvas, 27 × 22.5 cm
F 215 JH 1045

Beneath the study of the nude girl, who probably posed for Van Gogh and his fellow students in Cormon's studio, is a still life with a glass or small vase of flowers, possibly hyacinths. The brushstrokes to the left of the vase are illegible. One can imagine that this still life was of little commercial or artistic interest to Van Gogh, and that he therefore decided to use the canvas for another study.

[p. 81]

16 Self-portrait with dark felt hat
Paris, August-September 1886
The X-ray shows a study after a nude model
Paris, August 1886
Oil on canvas, 41.5 × 32.5 cm
F 208a JH 1089

The dating of this self-portrait has not, until now, been entirely resolved. Stylistically, it would seem to be an experiment: instead of painting with a loaded brush as he was wont to do, Van Gogh adopted a strikingly smooth manner, without the usual structuring of the background or much, if any, modelling in the clothing. As for colour, the canvas recalls the artist's earlier, dark palette, but lacks the expressive brushwork of his peasant portraits from Nuenen and peasant types from Antwerp.

Given its experimental character, the self-portrait was probably painted in Paris. The radiograph now enables us to date it even more precisely. The concealed full-length study of a female nude could have been executed in Antwerp. Vincent received training in life drawing at that city's Academy; we know, however, from letter 558/447 that he only occasionally painted from the model. It therefore seems more likely that the work was painted in Cormon's studio. Since it was of no particular interest to Van Gogh, and had no commercial value, it is not surprising that he decided to use the canvas again for a self-portrait study. It is not known when exactly Van Gogh studied with Cormon in 1886, but it was probably between late summer and November. Given the darkness of the self-portrait, the earliest possible dating - August-September 1886 - seems plausible.

[p. 82]

17 Self-portrait with pipe and glass
Paris, early 1887
The X-ray shows a portrait of a woman with her hair up
and a bare breast, Paris, probably late 1886
Oil on canvas, 61 × 50 cm
Signed and dated upper left: Vincent 87
F 263a JH 1199

The Self-portrait with pipe and glass conceals an representation of a woman with her left breast bared. Her right breast and shoulder are apparently covered by a shawl or some other garment. Unlike the academic nude study beneath the self-portrait with felt hat (cat. no. 16), this is a strikingly intimate study of the model. The relatively large format betrays the artist's intention to create a work of some allure: it is one of the largest bust-length portraits he would paint in either Antwerp or Paris; the Portrait of a man with skullcap and the late Parisian Self-portrait with easel measure only a few centimetres more. The Antwerp portrait of a woman with a red ribbon in her hair is of virtually the same size.¹

Since the self-portrait painted over the woman's likeness is dated ‘87’ we know it was executed in the French capital, where Van Gogh lived from March 1886 till February 1888. We may assume the portrait of the woman had dried by the time he superimposed his self-portrait; otherwise he would have first scraped it off. Consequently, the impasto of the first work is clearly visible on the surface of the second, which is thinly applied and very cracked in some of the areas where it covers earlier image. Indeed, near the forehead of the artist the underlying portrait shows through. The woman's patently brown hair - which she wears up, to judge from the X-ray - blends with both the background and Van Gogh's own hair, which therefore seems to stand somewhat higher than in reality.

Though nothing is known about the woman's identity, her hairstyle, apparently smiling mouth, naked breast, and highly particularised likeness call to mind the oval nude portrait in the Barnes Collection and the identical sheet in the Van Gogh Museum. The same woman, with a broad nose and her hair worn up, seems also to have sat for a portrait in Basel.²

1 The Portrait of a man with skullcap (F 289 JH 1203) measures 65.6 × 54 cm, the Self-portrait with easel (F 522 JH 1356) 65 × 50.5 cm. The Antwerp portrait of the woman (F 207 JH 979) is 60 × 50 cm.

2 The number of the Barnes nude is F 330 JH 1214, that of the drawing in the Van Gogh Museum F 1404 JH 1213. The portrait in Basel is F 357 JH 1216.

[p. 83]

18 Woman at a table in the Café du Tambourin
Paris, February-March 1887
The X-ray shows a portrait of a woman,
Paris, 1886-early 1887
Oil on canvas, 55.5 × 46.5 cm
F 370 JH 1208

The radiograph of the woman in the Café du Tambourin exposes a rather large, bust-length portrait of a woman with a round face, full lips, and what seems to be a short hairstyle. Below her neck, at the same level as the table top in the surface painting, is a light spot that must form part of her clothing. Van Gogh presumably scraped the canvas carefully: even though the woman at the table is thinly painted, there is no trace of the underlying impasto. This probably explains the rather atypical, seemingly hesitant touch revealed by the X-ray. In this respect, the work resembles a female likeness recently discovered beneath a view of Montmartre in the Bridgestone Museum of Art, Tokyo.¹

1 F 273 JH 1116. For more on this work see Kodera Tsukasa, ‘Van Gogh's windmills on Montmartre: the significance of scientific examinations in art historical research,’ in exhib. cat. Hidden faces, hidden portraits: the scientific examination of works of art, Tokyo (The Bridgestone Museum of Art) 1993, pp. 7, 8 and 16-20. On p. 44 there is an English summary of the article, which was published in Japanese.

[p. 84]

19 View from Vincent's room in the Rue Lepic
Paris, spring 1887
Oil on canvas, 46 × 38 cm
F 341 JH 1242

Judging from the radiographic image, this townscape appears to conceal another composition bearing little if any relation to it. Despite its illegibility it is included here for the sake of completeness.

[p. 85]

20 Anonymous artist
Still life with bottle of wine, two glasses and a plate with bread and cheese
place and date unknown
The X-ray shows a portrait of a woman, place and date unknown
Oil on canvas, 37.5 × 46 cm
F 253 JH 1121

The authenticity of this still life, which entered the Museum as part of the Van Gogh family collection (to which it had always belonged), was recently challenged.¹ From a stylistic point of view it is indeed difficult to situate the canvas in Vincent's oeuvre. Only the handling of the bread is somewhat reminiscent of his free brushwork, whereas the rest is distinguished by a manner that is reserved, if not lacklustre. There is nothing of interest in the background or tablecloth, while the glasses, plate, and knife are arranged against them with no more than technical competence.

The X-ray gives support to the doubts raised about the attribution of the still life. Beneath the surface lies a woman's portrait that would seem to have little in common with Van Gogh's art. To be sure, the putative scraping of the canvas makes it difficult to say anything about the technique; yet brushstrokes are still clearly visible in the woman's garment. Though the artist was apparently working rapidly, again one cannot help but notice a certain dullness: the strokes lack both the strength and structure so characteristic of Vincent. Thus, the radiograph confirms that the canvas is not by Van Gogh.

1 Dorn and Feilchenfeldt, op. cit. (introduction, note 4), pp. 278, 280.