Sculptures

Henri Matisse is most familiar through paintings and cut-outs, but sculpture was a sustained part of his practice. He modelled bodies, busts and reliefs, returned to subjects over years, and sometimes placed his own sculptures inside paintings. Sculpture gave him a way to test weight, balance and contour in three dimensions. It also produced independent works, meant to be seen as objects rather than regarded only as exercises for canvases. The Metropolitan Museum of Art observes that more than half of Matisse's sculptures were completed between 1900 and 1910; later series carried some of the same questions into the 1920s, 1930s and beyond.^[1]

This guide follows the forms and the making of representative works, from The Serf and La Serpentine to the Jeannette busts and the four Back reliefs. The dates attached to sculpture need particular care. A model could be made in clay, preserved in plaster and cast in bronze much later. One bronze in a museum may have a different casting date, surface or edition mark from another version of the same model. That material history is part of the artwork's story, not a technical footnote.^[2][3]

What counts as a Matisse sculpture?

Much of Matisse's sculptural work began by modelling: adding, pressing and changing a material such as clay to form a body or head. A plaster could preserve a state of that model or become the surface he worked on further. A bronze cast is a metal object made from a mould, often by a specialist founder. These terms identify different steps and sometimes different surviving objects. A catalogue may rightly say “1908; this cast c.1930” because the first date belongs to the sculptural conception and the second to the particular bronze in front of us.^[3]

The Met's Seated Figure, Right Hand on Ground makes the distinction concrete. Its model dates to 1908, while the museum's bronze was cast around 1930 and is marked 1/10 in an edition. Calling that physical object simply a “1908 bronze” would conceal more than two decades between the original modelling and its metal form. Conversely, calling the sculpture a “1930 invention” would erase Matisse's 1908 work. A collection record should be read for both dates, the material and the particular cast's identifier.^[3]

Bronze also has a surface. It can be finished, worked and given a patina, changing its colour and the way light catches raised or recessed areas. The Baltimore Museum of Art's technical research for Matisse: Painter as Sculptor compared casts of the same editions and found notable differences in construction, finish, patination and even size. Photographs can flatten those differences. A single familiar image is therefore not a complete account of every cast of a named sculpture.^[2]

Not all objects that look sculptural belong to the same category. Matisse's late cut-outs can present a body as a strongly shaped colour form, yet their materials are gouache-painted paper and a support. A sculpture may appear inside a painted still life as a depicted object. The paintings directory follows those canvases; here the focus is on Matisse's modelled forms and their surviving casts. The relationship between media is important precisely because they are distinct things.^[2]

How did the early sculptures begin?

The most intense early period falls around the first decade of the twentieth century. The Baltimore Museum of Art's 2007 exhibition assembled sculptural works conceived across a long span, 1899–1950, yet placed The Serf, Madeleine I and II and other early figures at the start of the story. The Met's career account says the majority were made between 1900 and 1910. These broad surveys are a stronger guide to activity than the old site's unsupported exact tally of bronzes, because a “work” can mean a model, a plaster, a later cast or a numbered variation depending on the catalogue's purpose.^[2][1]

The Serf was developed in Paris between 1900 and 1904, according to MoMA's record. It is a standing male nude, built as a weighty body rather than polished into an academic ideal. The Baltimore exhibition paired it with Matisse's painted Male Model, connecting both works to an Italian professional sitter who had also worked for Rodin. That documented connection lets us compare a body constructed in paint with one constructed in mass without claiming the two objects are identical studies or assigning the man a biography beyond the record.^[4][2]

The work's title may suggest social narrative, but its most immediate fact is formal: the torso stands heavily, its surface uneven and its limbs arranged around a compact vertical figure. The viewer can move around the bronze, seeing how shoulders, back and chest change from side to side. In the painting, the eye instead moves across a flat image whose colours and edges build an impression of depth. Both media investigate the same bodily presence through different constraints. The Serf artwork page follows the individual sculpture; the Male Model page provides the painted comparison.^[4]

At roughly the same time, Matisse made Madeleine I and Madeleine II, variations on a female figure dating to 1901–03 in Baltimore's exhibition account. The relationship between them is significant. BMA technical research indicates that Matisse could take a mould from a wet clay state and use the resulting forms as a basis for further development. A later numbered version was therefore a chance to rethink a body, not simply an extra bronze pulled from an unchanged original. This method recurs more dramatically in the Back series.^[2]

Why did Matisse study an écorché?

An écorché is an anatomical figure shown as if the skin were removed, exposing the underlying muscles. Matisse made his L'Écorché, d'après Puget model in 1903. The Fitzwilliam Museum explains a historical correction in its title: the earlier sculptural source was believed in Matisse's time to be by Michelangelo, but was subsequently attributed to the French sculptor Pierre Puget. Calling Matisse's work “after Puget” reflects that later attribution; it does not mean Matisse was necessarily naming Puget when he studied the figure.^[5]

The model is an act of looking at another sculpture, but it is also a test of how a twisted, seated body can be held together. The figure bends forward, with arms and legs making strong diagonals around the torso. Muscle forms can be read as anatomy, yet the compression of the pose is a compositional problem in its own right. It asks how one mass can turn, push and rest against another. The Écorché page examines the image and its uncertainty about which individual bronze it shows.^[5]

The Fitzwilliam's object demonstrates the date problem once more. Its 1903 date belongs to Matisse's original model. The museum's HM 8/10 bronze was cast at an unknown later date by the Valsuani foundry. Other numbered bronzes have their own casting histories. The 1903 model is thus not a single, self-evident 1903 metal object. The museum also records the plaster form appearing in Matisse's 1911 still lifes. When a sculpture enters a painting in this way, a work he made becomes one of the things he arranges and looks at anew.^[5]

What changed in La Serpentine?

MoMA dates La Serpentine to autumn 1909 at Issy-les-Moulineaux and records its bronze as object 624.1939. It is a small figure compared with the large Back reliefs, but its pose is expansive: limbs turn around an open centre, and the body's shifting directions are part of its energy. The work makes sense from more than one viewpoint, a quality that a frontal photograph alone cannot fully show. Its physical gaps matter as much as its solid areas.^[6]

Compared with the standing weight of The Serf, La Serpentine invites a different route for the eye. An arm may lead into a shoulder, then down through a bending leg and into the base; the spaces under and beside those limbs help make the movement legible. The name encourages us to notice a winding line, but the sculpture is not just a three-dimensional outline. Its twists carry volume, with a torso whose changing planes catch light as the viewer moves. The La Serpentine page follows the individual object and its relation to other early poses.^[6]

These early works show why calling Matisse's sculpture a brief detour from painting is misleading. He did not merely add volume to a painted silhouette. A sculpture occupies shared physical space with the viewer; a bronze surface reflects light differently from oil paint; a bodily pose must remain coherent from more than one angle. Yet the subjects and formal questions can return on canvas. Baltimore's cross-media exhibition made that repeated exchange its organizing theme.^[2]

How did the Jeannette heads become a series?

Matisse made five busts of Jeanne Vaderin, commonly called Jeannette I–V, between 1910 and 1916. MoMA says the first two were made directly from the model and provided starting points for the later variations. The sequence does not simply record a sitter ageing or move evenly from a realistic head to an abstract one. Matisse took already-made forms as material for new decisions: features could be widened, compressed, lengthened or joined more strongly to the neck and torso.^[7]

Jeannette V makes this especially clear. MoMA dates it to summer 1916 and says Matisse began from a plaster of Jeannette III rather than the immediately preceding Jeannette IV. He removed one hair mass, built up the brow, worked lines down from the hairline and stretched the neck. The route from version III to V is evidence that “series” did not mean copying each bronze in simple numeric order. It could be a branching investigation in which an earlier state remained available for a new ending.^[7]

Standing before several busts would let a viewer compare how much a portrait needs to retain to remain connected to its sitter. Hair, brow, nose and chin change their balance. In the later heads, the face can feel more architectural: a few strong masses carry a character that a greater number of small details might not. That does not require inventing Matisse's opinion of Jeanne or treating abstraction as a loss of knowledge about her. It directs us to what he physically changed between versions. The drawings directory provides a parallel path into his handling of human likeness through line.^[7]

The principle of variation continued with the three Henriette busts of 1925–29, identified in Baltimore's exhibition overview. They belong to a later period and a different sitter; they should not be folded into the Jeannette sequence simply because both are groups of female heads. In each series, repetition offered time to reconsider a form. The changes had a sculptural logic—mass, surface, support, profile—as well as any connection with contemporary portraits on paper or canvas.^[2]

Why do the four Back reliefs matter?

The Back series is Matisse's long study of a standing figure viewed from behind. A relief projects from a ground rather than standing free on every side. MoMA dates Back I to spring 1908–late 1909. Its gallery account says Matisse kept the clay wet for more than a year, cast it in plaster in June 1909, and then changed his working method: he could add new plaster or cut hardened areas with tools. He returned to the subject repeatedly across roughly twenty-one years, retaining plaster states that became the basis for bronzes called Back I–IV.^[8]

The series is a study in the human figure and in the material possibilities of relief. At the beginning, a viewer can find a comparatively detailed body, with shoulders, spine and weight on two legs. Across later versions those observations become broader structural decisions. The head may be drawn into the rising column of the body; the back's division and the relation of arms to torso become more emphatic. The final result does not simply “simplify” by deleting things. It rebuilds the figure around fewer, larger relations that occupy the relief's tall field.^[8]

Baltimore's technical examination found that Matisse used moulding methods to keep developing the Back models while preserving earlier states. This is important because four numbered reliefs can look, in a book, like a tidy line from detail to abstraction. The actual process involved material transfer, return and reworking. MoMA's current records date the fourth state to about 1931, while Baltimore's exhibition overview uses the broader 1909–30 span for the series. The exact dating of an individual state should follow its current owner record; the larger point is that Matisse sustained the problem through decades.^[2][8]

The works also invite a different comparison with a painting of a back. A painted contour can imply a body while preserving the flatness of canvas. A relief gives a literal projection, even if it remains tied to its background. Shadows change as light or the viewer moves. A photograph taken head-on can understate that depth and the roughness of the worked surface. These are reasons to consult object measurements and several views where available, rather than rank the four versions only by how “realistic” their front photographs look.

How did sculpture feed back into painting?

Matisse sometimes depicted his own sculpted figures among the objects of a painting. Baltimore's exhibition account points to Still Life with Geranium and Still Life with Plaster Figure, both 1906, as examples. The Fitzwilliam identifies the plaster Écorché in later 1911 still lifes. These appearances do not turn the paintings into mere records of studio inventory. A sculpted figure could change scale and role when represented in pigment, standing beside flowers, fruit, walls or textiles whose colours reshape how it is seen.^[2][5]

The reciprocal relation also runs from painted or drawn questions toward sculpture. The Serf and Male Model investigate a related sitter in different media. Reclining bodies and seated poses recur in bronzes and canvases across later decades, as Baltimore's survey emphasizes. But “related” should not mean “same work.” A particular painting may have its own subject, date, surface and ownership, just as a bronze may have its own cast history. The strongest comparisons name the objects and explain the specific shared problem—torso, pose, balance or the scale of a figure in a surrounding field.^[2]

MoMA's exhibition of The Red Studio provides a vivid example of works occupying more than one role. Matisse painted the studio in 1911 with other artworks inside it; the museum later reunited surviving objects represented in that canvas, including sculptures. For the viewer of the painting, a small form is part of the red composition. Seen separately in a gallery, it is also an object with its own material, modelling date and possible later casting history. Neither view cancels the other.^[9]

The late paper figures add a final comparison. Baltimore's Painter as Sculptor exhibition ended with cut-outs such as Blue Nude I of 1952, drawing a line from modelled bodies to Matisse's pared-down coloured shapes. This is a comparison across decades and media, not a claim that a blue paper nude is a bronze or that it was produced by casting. The cut-outs directory lets a reader follow that later method on its own terms, while the sculptures show how long Matisse had been testing the body's form.^[2]

Which sculpture pages can I explore here?

Begin with Study of Puget's L'Écorché, The Serf and La Serpentine. They give three different early problems: an interpretation of an anatomical model, a standing body developed over several years, and a turning figure whose open spaces matter to its movement. For the later Jeannette and Back series, the museum records linked in the references provide individual objects and more views.

Read each sculpture's title beside its model date, material, cast date if known, edition and present collection. That compact habit prevents a surprising number of errors. It also makes the forms more interesting: The Serf is a long-worked early standing figure, La Serpentine a compact exploration of bodily turning, the Jeannettes a branching portrait series, and the Backs a decades-long relief investigation. Together they show Matisse treating sculpture as an active way of thinking, one that kept meeting his painting and drawing without losing its own physical demands.