The Three Skulls
1902–6
Watercolor, with graphite, on ivory wove paper; 48 × 62.8 cm (18 7/8 × 24 3/4 in.)
The Art Institute of Chicago, Olivia Shaler Swan Memorial Collection, 1954.183
FWN 1987
Skulls and Other Objects
While this watercolor would seem to be a straightforward representation of three skulls arranged on a floral rug, the history and identity of these objects is complex and stands in need of careful analysis.1
John Rewald, who was instrumental in the preservation of Cezanne’s Lauves studio, states that “the four skulls [Cezanne] used as models” are kept in his “studio at Les Lauves.”[1] In fact, five skulls are preserved at the Atelier des Lauves. Three of these are intact and were on display at the time of writing in 2024; two in need of conservation were also made available.[2] It is likely that Cezanne owned at least one skull by the mid-1860s, when he began to draw and paint individual skulls,[3] but no witness recalls Cezanne owning any skulls until the 1880s. The artist’s niece Marthe Conil was the first to record some in his possession in a possibly unreliable account of a visit she made to the Grand Salon at the Jas de Bouffan, which Cezanne had turned into a studio after his father’s death in 1886. She recalled seeing “three skulls on the mantelpiece” there while Cezanne was working on his Mardi Gras (1888; FWN 668).[4] In July 1902 Borély saw “three human skulls facing each other” on “a narrow center table” in the bedroom of the rue Boulegon apartment.[5] But although both Émile Bernard and Ambroise Vollard recalled seeing the oil painting related to Chicago’s watercolor, Three Skulls on a Patterned Carpet (fig. 13.1), in Cezanne’s Lauves studio in 1904 and 1905, only the former gave any hint that he saw real skulls there, when he stated that Cezanne “needed time to advance [a painting], and he found it in front of skulls, green fruit or paper flowers.”[6]2
Fig. 13.1
Paul Cezanne (French, 1839–1906). Three Skulls on a Patterned Carpet, 1904 (FWN 875). Oil on canvas; 54.5 × 65 cm (21 3/8 × 25 5/8 in.). Kunstmuseum Solothurn (Dübi-Müller-Stiftung), Switzerland, C 80.2. Photo: Paul Smith, July 2022.
Rumors abound in Aix-en-Provence as to where Cezanne obtained his skulls (including some about clandestine arrangements with gravediggers). Cezanne’s close friend, the polymath Antoine-Fortuné Marion, discovered ancient skulls while excavating a neolithic cave settlement in the hills around Saint-Marc in 1866, but there is nothing to suggest that Cezanne obtained his skulls from him.[7]3
The most distinctive skull in the watercolor Three Skulls is the one at the back (fig. 13.2). This represents the skull in the Lauves studio with damage to its (proper) left zygomatic bone, which has left a gap in the eye socket (fig. 13.3).[8] This gap is also visible in Three Skulls on a Patterned Carpet, now in the Kunstmuseum Solothurn, Switzerland (fig. 13.4). In addition, the cranial sutures in this skull are only partially fused, making the vertical, metopic suture into the pronounced line that can be seen in Pyramid of Skulls (fig. 13.5). Cezanne did not represent this feature in The Three Skulls or Three Skulls on a Patterned Carpet, however. The skull at the left of the watercolor (fig. 13.6) corresponds to a real counterpart that retains several back teeth on both sides (fig. 13.7). But although there is significant bone loss in the center of the upper jaw, this is not apparent in the watercolor or the oil painting. The distinctively squarish shape of the eye sockets (known as “aviator sunglasses”) in the actual skull is apparent in both works, however, as is the skull’s slightly pointed shape. The skull at right in The Three Skulls (fig. 13.8) is probably modeled on the one in the studio that has only a few teeth remaining in its lower left jaw and a jawline that slopes down to right (fig. 13.9). This skull also has the pronounced anterior nasal spine visible in the watercolor. 4
Cezanne told Vollard that “a skull is a beautiful thing to paint,” but the interest skulls held for the artist went beyond their formal properties.[9] For one thing, Cezanne’s paintings of them sit squarely within the genre of vanitas paintings popularized in the Netherlands in the early seventeenth century, which itself derives from a more ancient tradition in which skulls carry the message of memento mori.[10] In Cezanne’s first vanitas painting, Skull and Candlestick (fig. 13.10), a skull is accompanied by wilting roses and an extinct candle, emblems of decay and the transience of life, and a book, symbolizing the worthlessness of human knowledge.[11] Similarly, the skull in Cezanne’s Young Man and Skull (fig. 13.11) relates to those contemplated by Christian saints, particularly Jerome and Francis, in several paintings in the Louvre.[12]5
Fig. 13.10
Paul Cezanne. Skull and Candlestick, c. 1866 (FWN 707). Oil on canvas; 47.5 × 62.5 cm (18 11/16 × 24 5/8 in.). Private collection, Zurich.
Fig. 13.11
Paul Cezanne. Young Man and Skull, 1896–98 (FWN 693). Oil on canvas; 130 × 97.5 cm (51 3/16 × 38 3/8 in.). Barnes Foundation, Philadelphia, BF 929.
The rug with a floral pattern in both The Three Skulls and Three Skulls on a Patterned Carpet appears as well in several paintings and watercolors of the period.[13] It also features—inverted—in a photograph taken by Robert Doisneau in 1945 (fig. 13.12). In some works, it appears alongside another rug whose floral motifs are integrated within a diamond-shaped diaper pattern.[14] (John Rewald photographed this second rug around 1935.[15] It can also be found on its own in several works.[16]) The sole surviving segment of the rug in The Three Skulls is a rectangle of lightweight felt a few inches across (fig. 13.13) with a remnant of a floral pattern.6
Fig. 13.12
Robert Doisneau (French, 1912–1994). Paul Cezanne’s Studio in 1945. Gamma-Rapho/Getty Images.
Fig. 13.13
The last remaining segment of the rug in Cezanne’s studio at Les Lauves, 2023. Atelier Cezanne, Aix-en-Provence. Photo: Paul Smith.
The vertical, leftward-slanting, largely blank area on the right-hand side of The Three Skulls is difficult to decipher until the Solothurn oil is consulted, when it becomes clear that it represents the side of what seems to be a shallow box. The curved shape at the bottom of the watercolor is also unintelligible without interrogating the painting, where it appears to stand for an ornate border on the rug.[17]7
Time and Place
There is nothing to indicate when or where The Three Skulls was executed. Stylistic evidence has been used, but this is somewhat imprecise, as it relies solely on visual analysis and comparisons.[18] There is, however, a wealth of evidence concerning the date of Three Skulls on a Patterned Carpet in Solothurn. But this information is sometimes contradictory and does not directly illuminate the relationship between the two works.8
Georges Rivière, the artist’s contemporary and the father-in-law of Paul Cézanne fils, states that this watercolor was made around 1898 in the studio Cezanne occupied in the Villa des Arts on the rue Hégésippe-Moreau in Paris—but his recollections are sometimes unreliable.[19] A vividly colored rug is visible in both the painting and the watercolor. It also appears in two or possibly three other works, all of which feature a dark-red wall that is known to have been part of Cezanne’s Hégésippe-Moreau studio.[20] But the only rug known to have been in this studio was described by Vollard as “faded.”[21] And significantly, no painting of skulls with a red wall in the background survives, so Cezanne evidently did not paint any of his still lifes featuring skulls in Paris.9
Three Skulls on a Patterned Carpet depicts a cool-colored wall, as do several other still-life paintings in which the same rug appears.[22] One of these, Bouquet of Flowers (FWN 881), appears to have been painted in Cezanne’s studio at 23, rue Boulegon in Aix around 1902, since the poet Léo Larguier recalls seeing “a large bouquet of artificial flowers” in this “attic” studio that year.[23] And in July 1903 the artist told the amateur painter Jules Borély that he had “a very good painted study” of skulls “in his attic” studio on the rue Boulegon that was related to the three human skulls his visitor had seen in the apartment below. Borély was nevertheless unable to view the painting because the maid had lost the key.[24] This “study” could have been Three Skulls on a Patterned Carpet or possibly Pyramid of Skulls (see fig. 13.5).[25]10
While these recollections make it probable that Cezanne began the oil in his rue Boulegon studio, there is firm evidence that he continued to work on it in his studio at Les Lauves in 1904 and 1905, perhaps even in the absence of the skulls themselves.[26] The artist and writer Émile Bernard recalled seeing “a canvas depicting three skulls on an Oriental carpet” in Cezanne’s “studio outside the city” when he first visited the painter in February 1904.[27] When he visited again in March 1905, he found that “the skulls were nailed to the wall, abandoned.”[28] However, Vollard remembered that when he went to Aix at the end of 1905 there “was a still life on the easel which [Cezanne] had started several years earlier, depicting skulls on an Oriental carpet.”[29] By way of confirmation, the poet Joachim Gasquet, who was close to Cezanne at the time, recounted that even “at the end” of his life, the artist’s incurable romanticism “made him pile … skulls … on rich carpets in blue still lifes.”[30] (For his part, Rivière states that Cezanne painted a “Skull on a Carpet” in his Lauves studio in Aix in 1903. Although no corresponding oil painting survives, there is a contemporary watercolor of this subject, Skull on a Drapery [1902–6; FWN 1986]).[31] These various sources imply that Cezanne worked on Three Skulls on a Patterned Carpet for some three years in total. Cezanne’s process was notoriously “slow and painful,” as Vollard put it.[32]11
There is, however, one feature of the scene in The Three Skulls and Three Skulls on a Patterned Carpet that has to be a permanent fixture of the room in which these works were begun, namely the chair rail or dado on the background wall that divides the compositions into a cooler, lighter section above and a darker, warmer one below. This object, or one very much like it, reappears in several other works, notably Still Life with Blue Pot (fig. 13.14).[33] There is no dado, however, in any of the works known to have been painted on the rue Hégésippe-Moreau. It may have been in the rue Boulegon studio, but photographs taken after Cezanne’s death show that it was wallpapered after he vacated it and that it lacked a chair rail.[34] Nor was there ever a chair rail in the Lauves studio, or in any part of the building it occupies.[35] Nevertheless, the cumulative evidence indicates that The Three Skulls was made in Cezanne’s studio on the rue Boulegon. This could suggest that it was conceived and executed as a preliminary study for Three Skulls on a Patterned Carpet. The idea that Cezanne’s watercolors had this subsidiary status has been conventional wisdom in the field, but several scholars in recent years have called it into question.12
Fig. 13.14
Paul Cezanne. Still Life with Blue Pot, 1900–1906 (FWN 1970). Watercolor and graphite; 48.1 × 63.2 cm (18 15/16 × 24 7/8 in.). The J. Paul Getty Museum, Los Angeles, 83.GC.221.
Oil and Watercolor
One obvious rationale for making a preliminary watercolor is that the medium’s manipulability allows the artist to work out how to frame a motif and to find an optimal viewpoint before proceeding to finalize a composition in an oil.[36] Moreover, even a large and relatively “finished” watercolor like The Three Skulls can be made in a matter of hours (sometimes over several days), while a fully realized oil such as Three Skulls on a Patterned Carpet could take hundreds of hours over several years. This fact suggests that Cezanne used the repeated contours in The Three Skulls to, among other things, rough out the outline of the skulls as a preliminary to working on them in earnest. They could also indicate that he experimented with their positioning or with his viewpoint prior to settling them slightly to the right in the corresponding oil.[37]13
Several scholars have nonetheless asserted that some key divergences between the Chicago watercolor and the Solothurn oil indicate that Cezanne’s watercolors are independent works of art that represent motifs of their own.[38] Three Skulls on a Patterned Carpet is significantly squarer, or about 10 percent deeper, than The Three Skulls. This is because both have standard-size supports. The oil is painted on a no. 15 figure (portrait) canvas,[39] while the watercolor occupies a full Raisin sheet (see Technical Summary in the Technical Report). Cezanne was not constrained to adopt the particular size of canvas he chose for the oil, however. Notably, he could have chosen a no. 15 paysage (landscape) canvas (50 × 65 cm), which almost exactly reproduces the proportions of the Raisin sheet. His choice must have been motivated, therefore, by a desire to include the ornate border of the rug in the composition.14
In his 1907 memoir, Émile Bernard casts further light on how Cezanne employed watercolor. He recalls that the artist would “start with shadow, by making a stain that he covered with a second, more overflowing one, then a third, until all these hues, forming screens, modeled the object at the same time as coloring it.” Bernard continues: “I understood immediately that there was a law of harmony that guided his work, and that all these modulations had a direction determined beforehand in his mind. In short, he proceeded in the same way as ancient tapestry makers must have, making related colors follow one another until they met their contrasting complementary.”[40] This suggests that Cezanne produced The Three Skulls to map out a set of harmonious color relationships according to a predetermined scheme, while in Three Skulls on a Patterned Carpet he exploited the flexibility that only oil allowed to continually adjust the harmony more empirically, and painstakingly. This distinction is simplistic, however, particularly since Bernard writes that “Cézanne prepared scales on his palette for a whole gradation of colors, and he proceeded by applying them” to the canvas in his oils.[41]15
Nevertheless, a watercolor like The Three Skulls, with its thin paint and many blank areas, is visually similar in these respects to oils of the kind that Bernard describes as merely “initial scales” of colors, whose progress “time had interrupted.”[42] Once again, though, Bernard’s remarks cannot be taken to suggest that the watercolor was equivalent to the lay-in stage of the oil. This is because he also reported that Cezanne recommended laying in a painting with “almost neutral tons” and only subsequently “heightening the scale.”[43]16
Three Skulls on a Patterned Carpet is much darker, muddier, and narrower in tonal range than the Chicago watercolor as a result of having been so heavily reworked. A watercolor like The Three Skulls does not yield a “preview” of a corresponding oil, in other words, since its medium generates a chromatically purer, more contrasty, and more luminous harmony. Similar considerations have led Carol Armstrong to suggest that The Three Skulls and Three Skulls on a Patterned Carpet “might be variations of each other in different media rather than preliminary study and final picture.”[44] And conservators at the Art Institute of Chicago have argued that “Cezanne frequently painted the same subject in both watercolor and oil as a means of examining his subjects through the unique properties of each.”[45]17
Cezanne clearly did intend some watercolors to be independent works of art equal in status to oil paintings, inasmuch as they have no counterparts in the other medium. Armstrong, for example, has argued that Cezanne’s Still Life with Blue Pot (see fig. 13.14) “is what the French call a tableau, a fully-fledged, finished picture in its own right, as distinct from a sketch (croquis), or a study (étude).”[46] She bases this claim on the large size of the work, its technical accomplishment, and the complexity of its layered surface: all features exhibited by the Chicago watercolor as well. It seems perfectly possible, then, that Cezanne could have envisaged and made The Three Skulls as a self-sustaining work of art and only subsequently decided to explore how the motif would look in oil because he enjoyed the particular qualities of the medium.18
Armstrong goes even further in speculating that perhaps “the oil yielded the watercolor.” In the same spirit, Elisabeth Reissner has suggested that “the thin and translucent paint application that watercolor makes possible” in The Three Skulls may have enabled Cezanne to paint “thinly” (as John Rewald noted) in Pyramid of Skulls.[47] Indeed, there are passages in Three Skulls on a Patterned Carpet that indicate a similar kind of intermedial interchange.19
On balance, therefore, there is no conclusive evidence as to whether the Chicago watercolor or the Solothurn oil was begun first. 20
Technique
Cezanne employed graphite pencil and watercolor (both transparent and opaque), as well as the white paper support itself, in The Three Skulls (see Media in the Technical Report).[48] In places, these are visually inextricable, particularly when the watercolor is viewed from a distance. The pencil lines stand out more clearly when the watercolor is seen close up, and better still in a reflected-infrared image (fig. 13.15).[49]21
Fig. 13.15
Reflected-infrared (Sinar rePro, 0.9–1.05 μm) image of Cezanne’s The Three Skulls (1902–6) revealing Cezanne’s preliminary graphite drawing. The Art Institute of Chicago, 1954.183.
When viewed from the distance at which they were made (arm’s length), Cezanne’s watercolors exhibit a “sketchy” or unfinished appearance, as noted by art historian Matthew Simms.[50] Most obviously, their color often tails off toward the periphery of the sheet, although not as dramatically in The Three Skulls as in some other works. Sometimes this leaves the white paper in these areas signifying nothing; or it could simply be a record of Cezanne’s lack of interest in articulating these areas further, as in the wall and the area below the rug.[51] White paper often fulfills a more “active” role, however, in the more central areas of many watercolors. Its most simple function is to stand for highlights or white surfaces, but it can also indicate perceptual uncertainty on the artist’s part.[52]22
Pencil, too, normally serves at least two distinct purposes in Cezanne’s watercolors: It delineates “contours,” which stand for the edges of objects, and it generates “hatchings,” which serve as shading.[53] Conventional hatchings are difficult to identify in The Three Skulls, however, possibly because they tend to morph into contours in the artist’s later works.[54] But there are loose indications of shading in the parallel, diagonal pencil lines at the top right of the two skulls center and right (see figs. 13.2, 13.8) and in the right eye socket of the one at right. Pencil contours are more widely distributed in The Three Skulls, but—uncharacteristically—few if any are “broken,” as they frequently are in other watercolors (as are many of the contours applied in watercolor).[55] The equivocal, repeated pencil contours at the edges of all three skulls nonetheless attest to a similar “instability” or “flux” in Cezanne’s perceptual process.[56] It is thus worth recalling in this context that the artist told Vollard while painting his portrait, “the edge escapes me.”[57] This statement may owe something to Jean Désiré Régnier’s De la lumière et de la couleur chez les grands maîtres anciens of 1865, which argued that contours “vacillate” when we fixate on the edges of objects.[58]23
The looseness and imprecision of Cezanne’s contours can also be explained perceptually as a technique that de-emphasizes drawing and modeling in favor of optical sensation.[59] It may also be that very loose drawing of the kind Cezanne used in the skull at the left (see fig. 13.6), in particular, was meant to correspond to the blurry information Cezanne would have picked up when he squinted at the motif.[60]24
Cezanne’s use of watercolor in The Three Skulls is no less multipurpose, according to Simms. In some places, the artist used strokes of color to define contours, either to “reinforce” those already stated in pencil or to shift them elsewhere.[61] Thus, the blue contours on the skull at right (see fig. 13.8) correspond most closely to earlier pencil lines, while the predominantly red contour on the left of the skull at left (see fig. 13.6) attempts to settle the issue (left by the pencil lines) of precisely where the skull’s edge lies. Elsewhere, painted contours denote edges on their own, as in the top of the skull at the back (see fig. 13.2).[62] This use of color is consistent with Cezanne’s employment of blue contours in his paintings merely as expedients, to hold edges in place before they were modeled in color.[63] Cezanne did this because he believed that contours did not exist in principle, but that “the contour of objects is given by the oppositions and contrasts that result from the coloring particular to them.”[64]25
Color also assumes the role of shading in Cezanne’s watercolors.[65] When it does, it sometimes coincides with pencil hatching, but at other times color conflicts with the hatching or works independently as a result of the “non-alignment” between the two media.[66] There is little pencil hatching in The Three Skulls, but where there is, there is some misalignment of this kind, particularly in the skull at right where the watercolor adds significantly to the shading of the eye sockets (see fig. 13.8). Different areas of watercolor are also misaligned, and hence either supplement one another or conflict in the information they supply. Notably, a broken blue contour that issues from the lower right of the right-hand skull emphatically announces an edge in the rug that earlier touches did not, whereas areas of local color simply refuse to correspond to watercolor contours in the lower edge of the red floral motifs below the skull at left. But although the media here are physically discontinuous, colors become assimilated to the contours around them optically because our spatial resolution is better for lines like this than it is for color.[67]26
Simms argues that Cezanne’s use of pencil and watercolor amounts to operating “two systems of notation”: a “tonal” system geared toward “haptic” sensation, and a “chromatic” system which favors “optical” sensation.[68] This argument has been criticized accordingly for not giving enough weight to Cezanne’s “interwoven acts of drawing and painting.”[69] Simms does speak of Cezanne’s use of watercolor as a “hybrid medium,” however, and of “the dialogue between drawing and color” in his watercolors.[70] And he acknowledges that Cezanne worked back and forth between the two different materials, sometimes going over watercolor with pencil.[71] This tactic can be difficult to identify because of the transparency of watercolor; and it may be that Cezanne drew the repeated contours on the right-hand skull’s right side over areas of previously applied color.27
It could seem odd that the two white skulls in the foreground do not have the distinctly yellowish color seen in the corresponding oil, where only the skull at the rear is near white. But technical analysis has revealed that Cezanne did use yellowish wash, particularly on the skull at the left, which has now faded (see Technical Summary in the Technical Report). There are nevertheless passages in The Three Skulls that are both physically and perceptually “blank, not active.”[72] Blank areas of this kind arise because Cezanne normally worked locally in watercolor, concentrating on defining edges, shading, and coloring in relatively isolated regions, whereas in oil he worked globally, attempting to create depth by adjusting all the colors in the painting to one another. By this account, when he worked from a watercolor to the corresponding oil, he had to “translate an (initial) denotation system based on (pencil) lines into the wholly different system wherein color defines spatial relationships.”[73]28
“Translation” here is perhaps too weak a term, as the differences between the two media are sometimes very considerable. It is apparent from comparing the Chicago watercolor with the Solothurn oil that Cezanne was far more committed in the latter to the global treatment of color relationships.[74] Oil also facilitates reworking much more readily than watercolor, and hence the surface of the painting is thickly painted overall, and in places very heavily. Indeed, the background pushing through the gap in the zygomatic bone of the rear skull, which is rendered with a few blue and reddish strokes in the watercolor (see fig. 13.2), needed paint more than 1 cm thick in the oil (see fig. 13.4). And while Cezanne was content to sketch out the color relationships between the top of the rearmost skull and the wall in the watercolor (fig. 13.16), the paint in the corresponding area of the oil is a thick enamel created by numerous thin applications of paint, which are designed to harmonize with the colors on either side of the edge (fig. 13.17).[75] Similarly, the richly colored shadows in the eye sockets and nasal cavities of the skulls in the watercolor become dark and chromatically muted in the oil, as Cezanne continually reworked the color relationships in this work.29
Fig. 13.16
Detail of top of skull at rear in Cezanne’s The Three Skulls (1902–6). The Art Institute of Chicago, 1954.183.
Fig. 13.17
Detail of top of skull at rear in Cezanne’s Three Skulls on a Patterned Carpet (1904). Kunstmuseum Solothurn (Dübi-Müller-Stiftung), Switzerland, C 80.2. Photo: Paul Smith.
Cezanne clearly spent a great deal more time attempting to resolve the color relationships in the oil than in the watercolor. In fact, Bernard recalled that during the month he spent in Aix in 1904, the painting “changed in color and form almost every day.”[76] And despite telling Bernard that he could not attain what he called “réalisation” (satisfactory completion) in this work, Cezanne did tell Vollard the following year that he was coming close to achieving precisely this.[77] In the last analysis, therefore, it was only in the flexible medium of oil that Cezanne could hope to achieve his ultimate ambition. 30
These differences notwithstanding, it is instructive to notice just how closely Cezanne followed the technique of The Three Skulls in one portion of Three Skulls on a Patterned Carpet, namely the red floral motif at bottom center. It is evident that Cezanne achieved the local color he desired in this area of the watercolor by superimposing several layers of reddish pigment, some verging toward orange and others faintly bluish, some quite transparent and others more opaque (fig. 13.18; see also Technical Summary in the Technical Report).[78] Although oil does not readily permit the same range of transparency, it is clear nonetheless that Cezanne did try to achieve similar effects in the Solothurn painting. Several layers of differently colored reddish paint can certainly be discerned in the corresponding area of this work (fig. 13.19), some of them exploiting the semitransparency of thinly applied oil paint. Sometimes, in other words, Cezanne did not “translate” from The Three Skulls when working on his oil so much as pay homage to it.31
Fig. 13.18
Detail of the floral motif at bottom center in Cezanne’s The Three Skulls (1902–6). The Art Institute of Chicago, 1954.183.
Fig. 13.19
Detail of the floral motif at bottom center in Cezanne’s Three Skulls on a Patterned Carpet (1904). Kunstmuseum Solothurn (Dübi-Müller-Stiftung), Switzerland, C 80.2. Photo: Paul Smith, 2022.
- I am grateful to the Leverhulme Trust for the award of a Major Research Fellowship, which facilitated much of the research for this essay. See John Rewald, The Paintings of Paul Cézanne: A Catalogue Raisonné (Harry N. Abrams, 1996), 1:88, no. 81. Rewald also notes (1:493, no. 822), with regard to Pyramid of Skulls (see fig. 13.5), that Cezanne “took these skulls (all without jawbones) to his studio at Les Lauves, erected in 1902, where they remain.” On Rewald’s role in saving the studio, see Jayne Warman, “A Dual Legacy,” in Société Paul Cézanne, Atelier Cézanne, 1902–2002: Le centenaire (Actes Sud, 2002), 166–73.
- Five skulls were recorded in inventories taken between 1954 and 1969, and by Michael Fraisset in 1998. Four lacked a “machoire inférieure,” and one had “pariétal gauche perforé de 2 trous” (this was split across the nose at the time of writing). See Michel Fraisset, Les vies silencieuses de Cézanne (Office du Tourisme, Aix-en-Provence, 1999), n.p., nos. 161–65. The damaged skulls are being restored.
- See Walter Feilchenfeldt, Jayne Warman, and David Nash, eds., The Paintings, Watercolors and Drawings of Paul Cezanne: An Online Catalogue Raisonné, accessed June 1, 2025, http://www.cezannecatalogue.com, no. 707 (hereafter FWN).
- M[arthe] C[onil], “Quelques souvenirs sur Paul Cézanne par une de ses nièces,” Gazette de beaux-arts, 6e période, 56, no. 1102 (Nov. 1960): 300. Cezanne began painting Mardi Gras in Paris in 1888 and sold it in 1891. See Rewald, The Paintings, 1:408, no. 618; and FWN 668.
- Jules Borély, “Cézanne à Aix,” L’art vivant, no. 37 (July 1, 1926): 493: “J’ai quitté Cézanne après avoir passé encore une demi-heure chez lui dans sa froide maison. Dans sa chambre, sur une étroite table de milieu, j’avisai trois crânes humains confrontés, trois beaux ivoires polis.”
- Émile Bernard, “Souvenirs sur Paul Cézanne et lettres inédites,” Mercure de France 70, nos. 247 and 248 (Oct. 1 and 16, 1907): 626: “Il avait besoin de temps pour pousser, et il le trouvait devant des crânes, devant des fruits verts ou des fleurs de papier.”
- For Marion, see Nina Athanassoglou-Kallmyer, Cézanne and Provence: The Painter in His Culture (University of Chicago Press, 2003), 157; and André Dombrowski, Cézanne, Murder, and Modern Life (University of California Press, 2013), 50. On skull motifs on Celto-Ligurian sculptures found in the region, see Mary Tompkins Lewis, Cézanne’s Early Imagery (University of California Press, 1989), 40–43.
- I am indebted to Laura Perramond for retrieving this skull from storage and for holding it while I photographed it. She has also confirmed my identification of the skulls, drawing on extensive unpublished research on their condition. I must also thank Fabienne Morucci, Elodie L’Huillier, and Camille Malta for facilitating my research at the Atelier Cezanne.
- Ambroise Vollard, Paul Cézanne (Ambroise Vollard, 1914), 148: “Que c’est beau à peindre, un crâne!’
- See Lewis, Cézanne’s Early Imagery, 37, 233–34n27; and Carol Armstrong, Cézanne in the Studio: Still Life in Watercolors, exh. cat. (J. Paul Getty Museum, 2004), 66. Compositionally, The Three Skulls has prototypes going back to fifteenth-century engravings. See also Theodore Reff, “Cézanne: The Severed Head and the Skull,” Arts Magazine 58, no. 2 (Oct. 1983): 96.
- See Lewis, Cézanne’s Early Imagery, 233n27, on a prototype of this image in the painting by Jacques de Claeuw (now attributed Simon Renard de Saint-André) in the Musée des Beaux-Arts in Marseille. See also Reff, “The Severed Head,” 90.
- These include works by Georg Pencz, Jan Cornelisz. Vermeyen, and Guido Reni. See also Reff, “The Severed Head,” 95.
- FWN 870, 883, 884, and 1986.
- FWN 879, 881, and 885.
- See Rewald, The Paintings, 1:493, no. 824.
- FWN 534, 541, 871, 1970, and 1971.
- I am indebted to Kristi Dahm for this observation. There is a similar border in FWN 871 and 1970 (which also shows a chair rail). See Armstrong, Cézanne in the Studio, 65.
- See John Rewald, Paul Cézanne: The Watercolours; A Catalogue Raisonné (Thames and Hudson, 1983), 246, no. 611: “The prevalent hues of this work point to Cézanne’s Lauves studio, since they are closely related to those of the landscape watercolours he made in the vicinity of that studio between 1902 and his death.”
- Georges Rivière, Le maître Paul Cézanne (Floury, 1923), 222: “Vers 1898, Trois têtes de mort: Représentées sur un tapis; peinture exécutée à Paris, rue Hégésippe-Moreau.” On Cezanne’s occupation of this studio, see Cezanne to Egisto Fabbri, May 31, 1899, in Paul Cézanne: Correspondance, ed. John Rewald (Grasset, 1978), 270. The artist’s presence there in 1898 is implied by the mention of a painting store he kept “dans une des rues qui avoisinent la place Blanche.” See Saint-Georges de Bouhélier, “Simplicité de Cézanne,” L’echo de Paris, June 15, 1936, 1, quoted in Robert William Ratcliffe, “Cézanne’s Working Methods and Their Theoretical Background” (PhD diss., Courtauld Institute of Art, University of London, 1960), 390n76.
- FWN 879 (possibly), FWN 883, and FWN 1986 (possibly).
- See Vollard, Paul Cézanne, 105, quoted in Ratcliffe, “Cézanne’s Working Methods,” 403n144: “dans un coin de l’atelier … il y avait eu, depuis toujours, un vieux tapis jeté par terre, et qui n’avait plus couleur de tapis.” For the portrait of Vollard, see fig. 7.11.
- These include FWN 884 and 885.
- See Léo Larguier, Cézanne, ou, Le drame de la peinture (Denoël et Steele, 1936), 24–26, quoted in Ratcliffe, “Cézanne’s Working Methods,” 403n144: “Il avait encore un autre atelier, dans les combles de sa maison, rue Boulegon, un grenier… . Entre deux croisées, sur une table de bois blanc, il y avait un grand bouquet de fleurs artificielles.”
- Borély, “Cézanne à Aix,” 493: “Il me parlait d’une très bonne étude peinte qu’il avait quelque part, en son grenier.” See Ratcliffe, “Cézanne’s Working Methods,” 29, which suggests this work may have been Three Skulls (c. 1898; FWN 873).
- See Rewald, The Paintings, 1:493, no. 822, for the suggestion that Pyramid of Skulls was made in Cezanne’s “small studio” in the rue Boulegon since it was “painted from very close-up.” Cf. Léo Larguier, Le dimanche avec Paul Cézanne (L’Édition, 1925), 107, which contradicts this claim in its description of the studio: “C’était un grenier qu’on avait aménagé sous les toits, une immense pièce, avec de petites fenêtres.”
- On the chronology of Cezanne’s Lauves studio, which he occupied from September 1902, see Cezanne, Correspondance, 286–87, 290, 292; Fraisset, Les vies silencieuses, n.p.; and Bruno Ely, “A Place of Truth,” in Société Paul Cézanne, Atelier Cézanne, 27–41.
- Bernard, “Souvenirs,” 393–94: “Après le repas nous allâmes à l’atelier hors la ville, il m’introduisit enfin parmi ses tableaux, dans sa propre salle de travail … Il était à l’ouvrage d’une toile représentant trois têtes de mort sur un tapis d’Orient. Il y avait un mois qu’il y travaillait tous les matins, de six heures à dix heures et demie.”
- Bernard, “Souvenirs,” 621: “Les têtes de mort étaient clouées au mur, abandonnées.”
- Vollard, Paul Cézanne, 147–48: “Sur le chevalet on voyait une nature morte commencée plusieurs années auparavant et représentant des cranes sur un tapis d’Orient.”
- Joachim Gasquet, Cézanne (Bernheim-Jeune, 1921), 19: “Et c’était lui encore qui, même à la fin … lui faisait empiler, en de bleues natures mortes, ces crânes … sur de riches tapis.”
- Rivière, Paul Cézanne, 224: “1903. Tete de mort sur un tapis. Toile peinte dans l’atelier des Lauves.”
- Vollard, Paul Cézanne, 95: “lent et pénible.”
- See also FWN 543–45, 1947, 1987, and 1971. On this chair rail, see Ratcliffe, “Cézanne’s Working Methods,” 403n144.
- Michael Fraisset, email message to author, Apr. 30, 2024.
- The present narrow border immediately above the painted maroon skirting in the Lauves studio was added between 1970 and 1972 by Marianne Bourge; there appears to have been wallpaper on the ground floor. I am grateful to Michel Fraisset for this information.
- Cezanne had a different reason for making watercolor studies unrelated to oils, according to Maurice Denis’s journal entry for Oct. 21, 1899, which states that Vollard had told him: “Pour s’entraîner à peindre dès le matin … il fait une aquarelle en plein air; il pretend être ainsi tout disposé à bien voir le lendemain.” Maurice Denis, Journal (La Colombe, 1957), 1:157, quoted in Françoise Cachin et al., Cézanne, exh. cat. (Philadelphia Museum of Art, 1996), 422.
- I am indebted to Kristi Dahm for this observation. For more on Cezanne’s use of repeated contours, see cat. 2 and cat. 11.
- I am grateful to Fabienne Ruppen and Kristi Dahm for bringing these arguments to my attention. See the entry by Joseph Rishel in Cachin et al., Cézanne, 260; Matthew Simms, Cézanne’s Watercolours: Between Drawing and Painting (Yale University Press, 2008), 103; and Elisabeth Reissner, “Ways of Making: Practice and Innovation in Cezanne’s Painting in the National Gallery,” National Gallery Technical Bulletin 29 (2008): 11.
- See Walter Feilchenfeldt, “On Sizes and Subjects,” in Rewald, The Paintings, 1:16. Cezanne’s local supplier of artist’s materials was Justin Gabet, a childhood friend and painter. See Gustave Coquiot, Paul Cezanne (Paul Ollendorff, 1919), 133: “Et, pour ne pas être pris au dépourvu par les moyens matériels de son ‘métier,’ il prenait toujours ses toiles chez la veuve Tacussel, place Saint-Honoré,—et ses châssis chez Gabet, menuisier, établi dans une impasse de la rue Boulegon.” See also Cezanne, Correspondance, 135–136; and FWN 631.
- Bernard, “Souvenirs,” 396: “Il commençait par l’ombre et avec une tache qu’il recouvrait d’une seconde plus débordante, puis d’une troisième, jusqu’à ce que toutes ces teintes, faisant écrans, modelassent, en colorant, l’objet. Je compris de suite que c’était une loi d’harmonie qui guidait son travail et que toutes ces modulations avaient une direction fixée d’avance, dans sa raison. Il procédait en somme comme ontdùrobserver les tapissiers anciens, faisant se suivre les couleurs apparentées jusqu’à ce qu’elles rencontrassentleur contraste dans l’opposition.”
- Bernard, “Souvenirs,” 398: “Cézanne avait des gammes faites sur sa palette pour toute unegradationde teintes, et qu’il procédait en les appliquant.”
- Bernard, “Souvenirs,” 399: “des gammes commencées et que le temps avait interrompues.”
- Bernard, “Souvenirs,” 614: “tons presques neutres,” “serrant davantage les chromatismes.” See also cat. 8.
- Armstrong, Cézanne in the Studio, 90.
- Kim Muir et al., “‘A Harmony Parallel to Nature’: Color, Form, and Space in Cezanne’s Watercolors and Oil Paintings,” first published in Cezanne, ed. Achim Borchardt-Hume et al., exh. cat. (Art Institute of Chicago, 2022), 48.
- Armstrong, Cézanne in the Studio, 75–76.
- Elisabeth Reissner, “Technical Study Within Art Historical Scholarship: ‘Meaning in Making’ with Particular Reference to the Works of Paul Cézanne” (PhD diss., University of London, 1960), 257; and Rewald, The Paintings, 1:19. See also Elisabeth Reissner, “Transparency of Means: ‘Drawing’ and Colour in Cézanne’s Watercolours and Oil Paintings in the Courtauld Gallery,” in The Courtauld Cézannes, ed. Stephanie Buck et al., exh. cat. (Courtauld Gallery; Paul Holberton, 2008), 49–71.
- See Borchardt-Hume et al., Cezanne, 141.
- See Muir et al., “’A Harmony Parallel to Nature,’” 48–49. Viewing distance makes a difference because we favor the low-spatial frequency information supplied by large areas of color from farther away, whereas finer high-spatial frequency information falls below the threshold at which we can resolve it; conversely, we can resolve fine detail better than color from closer-to by picking up on high-spatial frequency information that is now above this threshold.
- See Simms, Cézanne’s Watercolours, 92–93.
- On this use of paper, see Simms, Cézanne’s Watercolours, 90, 99.
- See Simms, Cézanne’s Watercolours, 87, 99.
- Simms, Cézanne’s Watercolours, 104–11.
- I am indebted to Fabienne Ruppen for this observation.
- Simms, Cézanne’s Watercolours, 105.
- Simms, Cézanne’s Watercolours, 105.
- Vollard, Paul Cézanne, 105.
- Jean Désiré Régnier, De la lumière et de la couleur chez les grands maîtres anciens (Paris, 1865), 52: “On ne peut pas regarder le contour d’un objet et voir en même temps le fond sur lequel il se détache, sans que le contour de l’objet qu’on regarde semble vaciller.” See also Ratcliffe, “Cézanne’s Working Methods,” 336–39. On Cezanne’s copy of Régnier, see Jean de Beucken, Un portrait de Cézanne (Gallimard, 1955), 304; and Ratcliffe, “Cézanne’s Working Methods,” 330–51, which suggests that Cezanne read the book between 1877 and 1885.
- See Simms, Cézanne’s Watercolours, 112–13.
- On Cezanne’s habit of squinting,see cat. 1. See Michael J. Land and Benjamin W. Tatler, Looking and Acting: Vision and Eye Movements in Natural Behaviour (Oxford University Press, 2009), 38, which describes how information of this kind creates a sense of the perception flowing by preventing the eye from fixating.
- See Simms, Cézanne’s Watercolours, 87, 99, 106.
- On this type of contour, see Simms, Cézanne’s Watercolours, 111.
- For a discussion of this use of contour, see cat. 2.
- Cezanne quoted in Larguier, Le dimanche, 135.
- See Simms, Cézanne’s Watercolours, 105–6.
- Simms, Cézanne’s Watercolours, 108–9.
- See John Mollon, “Seeing Colour,” in Colour: Art and Science, ed. Trevor Lamb and Janine Bourriau (Cambridge University Press, 1995), 132.
- Simms, Cézanne’s Watercolours, 111–12.
- Armstrong, Cézanne in the Studio, 140n2.
- Simms, Cézanne’s Watercolours, 199–200.
- See Simms, Cézanne’s Watercolours, 111.
- Paul Smith, “Cézanne: Finished/Unfinished,” Apollo 152, no. 464 (Oct. 2000): 62.
- Smith, “Cézanne: Finished/Unfinished,” 62. A “denotation system” is a system of physical marks and the psychologically active “picture primitives” it generates in the form of lines, hatchings, or colors, which correspond to “scene primitives,” or elements of the perceived world in the form of edges, shading, and hue. See John Willats, Art and Representation: New Principles in the Analysis of Pictures (Princeton University Press, 1997), 96–100. Cf. Simms, Cézanne’s Watercolours, 106: “Cézanne’s color translates his drawing into a new visual language.”
- On global harmony in Cezanne, see cat. 8.
- On this technique, see cat. 2.
- Bernard, “Souvenirs,” 395: “a changé de couleur et de forme presque chaque jour.”
- Bernard, “Souvenirs,” 394: “‘Ce qui me manque, me disait-il devant ses trois têtes de mort, c’est la réalisation’”; and Cezanne quoted in Vollard, Paul Cézanne, 148: “‘Comprenez un peu, je touche à la réalisation!’”
- See Muir et al., “‘A Harmony Parallel to Nature,’” 50–51.
- “Standards for French paper formats … varied dependent on region, mill, or brand.” Fabienne Ruppen, “On Margins and Versos: The Hidden Interrelationships Among Cézanne’s Works on Paper,” in Cezanne: Metamorphoses, ed. Alexander B. Eiling, exh. cat. (Staatliche Kunsthalle Karlsruhe; Prestel, 2017), 96n36. Ruppen also notes that the Canson et Montgolfier Raisin format can be found in the various product catalogues in the archives of the Musée des Papeteries Canson et Montgolfier.
- Fabienne Ruppen, conversation with author, October 24, 2023.
- Adrien Chappuis, The Drawings of Paul Cézanne: A Catalogue Raisonné (New York Graphic Society, 1973), 1:10.
- Though the techniques used to analyze the pigments cannot discriminate between the two types of copper-arsenate pigments available in the nineteenth century—emerald green and Scheele’s green—the former is more likely, since the use of Scheele’s green is rare and emerald green has been confirmed in Cezanne’s paintings at other institutions. See Francesca Rosi et al., “A Non-Invasive XRF Study Supported by Multivariate Statistical Analysis and Reflectance FTIR to Assess the Composition of Modern Painting Materials,” Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 71, no. 5 (2009): 1655–62, https://doi.org/10.1016/j.saa.2008.06.011; and Faith Zieske, “An Investigation of Paul Cézanne’s Watercolors with Emphasis on Emerald Green,” Book and Paper Annual 14 (1995): 105–15. However, recent investigations have further highlighted the complexities related to the identification of copper-arsenic compounds. See Leonardo Pantoja Munoz, “Hidden in Plain Sight: Revisiting the Synthesis, Characterization, Degradation and the Intricate Relationship Between Scheele’s Green and Emerald Green,” Heritage Science 12, no. 94 (2024): https://doi.org/10.1186/s40494-024-01192-7.
- No inorganic or organic elements were detected with XRF, reflectance FTIR, or HSI analysis. See Clara Granzotto and Maria Kokkori, “1954-183_Cezanne_ARP_XRF,” Dec. 4, 2023; and Giovanni Verri, “1954-183_Cezanne_FTIR_HSI,” Jan. 8, 2021, on file in the department of Conservation and Science, Art Institute of Chicago.
- Marjorie Shelley, “Cézanne as Draftsman: Sketchbooks and Graphite Drawings,” in Dita Amory et al., Madame Cézanne, exh. cat. (Metropolitan Museum of Art, 2014), 118.
- The presence of the pigments was inferred using the following methods: cobalt blue (XRF, HSI), indigo (HSI), emerald green (XRF, HSI), yellow ochre (XRF, HSI), vermillion (XRF, HSI), an animal-based red lake (HSI), and lead white (XRF, HSI). Analysis was carried out on selected areas and may not include all pigments present in the watercolor. See Granzotto and Kokkori, “1954-183_Cezanne_ARP_XRF”; and Verri, “1954-183_Cezanne_FTIR_HSI.”
- For pigment analysis of Cezanne’s watercolors in other collections, see Faith Zieske, “Paul Cézanne’s Watercolors: His Choice of Pigments and Papers,” in The Broad Spectrum, ed. Harriet Stratis and Brit Salvesen (Archetype, 2002), 89–101; Reissner, “Transparency of Means,” 49–71; and Laura Neufeld, “Belle Formule: Materials and Methods in Cézanne Watercolors,” in Cézanne Drawing, ed. Jodi Hauptman and Samantha Friedman, exh. cat. (Museum of Modern Art, 2021), 203–4.
- Maria Kokkori, “Cezanne’s Materials,” app. 2 in Borchardt-Hume et al., Cezanne, 228–29.
- Standards for texture and thickness are given in Print Council of America, The Print Council of America Paper Sample Book: A Practical Guide to the Description of Paper (Print Council of America, 1996).
- See Clara Granzotto and Maria Kokkori, “1954-183_Cezanne_ARP_XRF,” Dec. 4, 2023.
- Fabienne Ruppen, “Tackling Cézanne’s Paper: On the Reconstruction of Loose Sheets,” in Walter Feilchenfeldt, Yuval Etgar, and Fabienne Ruppen, Reconstructing Cezanne, exh. cat. (Ridinghouse; Luxembourg and Dayan, 2019), 32.
- “It stands to reason that he bought his drawing paper à la main, that is in batches of twenty-five sheets, since it would have been significantly less expensive and more efficient than purchasing single sheets.” Fabienne Ruppen, “Mapping Cezanne Drawings and Watercolors on Paper from Émile Desloye,” in Actes des journées de’étude “Les Filigranes, une marge á explorer” (organisée le 20 Octobre 2018) et “Histoire de papier et de la papeterie—Actualités de la recherche” (organisée le 11 octobre 2019), ed. Claude Laroque, Valérie Lee and Maryse Pierrard (Site de l’HICSA, 2020), 216, https://hicsa.pantheonsorbonne.fr/sites/default/files/2023-09/livre_laroque_2020_05.pdf.
- See note 79 above.
- Shelley, “Cézanne as Draftsman,” 118–19.
- Lionello Venturi, Cézanne: Son art, son œuvre (Éditions Paul Rosenberg, 1936), 1:383, no. 1131; and a letter from Walter Feilchenfeldt of Apr. 19, 2002, in the curatorial file.
- Jacques Seligmann Gallery, Water Colors by Cézanne (Jacques Seligmann Gallery, 1933), no. 3; and Feilchenfeldt letter, Apr. 19, 2002.
- Venturi, Cézanne, 1:383, no. 1131; and Feilchenfeldt letter, Apr. 19, 2002.
- Feilchenfeldt letter, Apr. 19, 2002.
- Feilchenfeldt letter, Apr. 19, 2002.
- Feilchenfeldt letter, Apr. 19, 2002.
Entry by Paul Smith; Technical Report by Kristi Dahm, “Cat. 13 The Three Skulls, 1902–6,” in Cezanne Paintings and Watercolors at the Art Institute of Chicago, ed. Gloria Groom and Kim Muir (Art Institute of Chicago, 2026), https://doi.org/10.53269/9780865593251/18. © 2026 by The Art Institute of Chicago. This work is licensed under a CC BY-NC 4.0 license: creativecommons.org/licenses/by-nc/4.0/.