(Photo Jahrbuch der Berliner Museen 2004)


A group of researchers from fields as far apart as Art History, Sculptural Restoration, Computer Tomography and Nuclear Analysis joined forces in 2004 in order to find answers to some basic questions concerning Daumier’s RATAPOIL sculpture, which have been occupying collectors and curators for years: For what reason do two obviously different versions of Ratapoil plasters exist, and what does the “inner life” of each plaster tell us about their origin and possible change over time?

The “Jahrbuch der Berliner Museen” of 2004 (Gebr. Mann Verlag, Berlin) published the relevant findings (in German only) in their annual edition on pages 249 – 283 under the title

„Honoré Daumiers RATAPOIL und die Untersuchungen der beiden Gipsexemplare”. The authors responsible for the study were: J. de Caso, University of California Berkeley; B. Maaz, Alte Nationalgalerie in Berlin; E. Papet, Musée d’Orsay, Paris; A. Badde, Sculpt. Restoration Berlin; A. Cascio, Sculpt. Restoration, Paris; B. Illerhaus, BAM Berlin; D. Kushel, Buffalo State College.

The outcome is an extensive, photographically documented report of greatest interest to Daumier amateurs and researchers alike, and especially to the proud owners of one of the 55 (plus 2) original Ratapoil bronze sculptures. Unfortunately, there are only very few printed copies of this report available, which can however be purchased using following information: Jahrbuch der Berliner Museen, Neue Folge 46, 2004, Gebr. Mann Verlag Berlin 2005

Informationen: ISBN 3-7861-2405-1 and ISSN 0075.

The first section of the report covers the historical, well-known background, acknowledging the initial existence of only one unbaked clay model, which served as basis for the later plaster model(s) and consequently their bronze reproductions. It was agreed that the complexity of the plaster models indicated that Daumier may probably have been assisted by an experienced sculptor. Possible candidates would have been his sculptor friend A. Préault or, more likely, A.V. Geoffroy-Dechaume (as previously suggested by Gobin). It was he who had assisted Daumier already in 1850-1852 in creating plaster models of the “Fugitifs”.

The report notes on p. 254, footnote 22 (but this still needs further confirmation) that the original clay model may still have been part of the Henry Bing Collection in Paris about 1930, as mentioned in the MOMA exhibition catalogue and photographed as a bronze version. After that date no traces of the clay remained. If however Alexandre’s observations from 1888 are correct, Daumier, being aware of the fragility of the clay, may have decided as early as 1851 to have a plaster version done in order to preserve a more solid copy of his model. This first plaster can thus be considered an ‘original model’. It is consequently well possible that Bing owned a plaster, but most likely not the clay model.

The creation of the clay can with some certainty be dated around March 1851 (and not 1850 as suggested by Gobin and Cherpin) when Michelet saw it in Daumier’s studio. After that it was hidden during the Napoleon III era (1852-1870) and only as late as 1878/79 did the first plaster appear in an exhibition. After Daumier’s death it was given to Geoffroy-Dechaume to be exhibited again in 1888 together with Dantan’s sculptures. In the same year Arsène Alexandre published a fotomechanical reproduction, an autotypie of a photography of the plaster in his Daumier biography. A first order for a bronze sculpture followed in 1890/91 from A. Dayot for the State Collection, now at the Orsay. For this purpose, Geoffroy-Dechaume’s plaster was transferred on April 17, 1891 to Jean Pouzadoux, a specialist in this field, who produced a plaster copy from the original Dechaume model. This second plaster would then be used as a basis for the bronze ordered with the Siot-Decauville foundry.

Up to 1891 it seems that Art Historians agreed on the existence of only one plaster. More recent findings however (Gobin in 1952 and others) suggest the existence of two plasters, strikingly similar and both of impeccable provenance. The authors went to great length in explaining that Pouzadoux’s plaster copy may have been preserved, thus representing the second plaster model. Both served as basis for the various later bronze editions. A number of questions immediately arise: is one plaster a copy of the other? Which was the first one? Which version was closest to the original clay modeled by Daumier himself?

Unfortunately both plasters had repeatedly been repaired and partially remodeled over time after having been used as models for the later bronze editions. Traces of this “treatment” are still clearly visible today.

The report also came to the conclusion that in all probability two Siot-Decauville bronzes (au sable) existed in the first edition: one sample from 1891 was ordered by the State and kept at the Orsay and a second one was delivered in 1896 to the Musée des Beaux-Arts in Marseille, Daumier’s city of birth. An additional second edition of 20 pieces was produced by the same foundry in 1892. It seems unclear whether for this series the chef modèle was used or Pouzadoux’s second plaster.


The study group decided on the use of two different examination techniques: Computer Tomography as a three dimensional screening method for the research of the physical structure and of possible changes in the plasters, suggesting that future projects could eventually link each plaster to its respective bronze copy. Differences which had occurred during the molding / casting process could thus be made visible. This method was applied for the “Geneva version”.

The other technique used was Xeroradiography. It was applied for the second plaster, the „Buffalo version“.


The Geneva Version

This plaster is covered by an isolating protective wax coating applied by Valsuani´s prepatory work for his 1960 bronze edition (…). It has since

yellowed / darkened. Traces of the tools are still visible on the surface of the plaster. The entire plaster surface is covered by seamlines stemming from the casting process in a small-piece mold. Differences in surface levels of these sections suggest that the small-piece mold had been produced directly on the original clay model.

Images from Computed Tomography showed that the head, which was originally hollow, had been attached to the neck and chest with various metallic rods, which were encased in an initially moist plaster block. Once the block had dried up it served as a support to the head and the fragile neck sections. While most of the chest area is hollow, the legs, initially also shaped as empty hulls, were also supported with metal rods during the manufacturing process. They were inserted into an added, moist, plaster filling, which after some time dried up and gave the legs a rigid support. The walking stick (or cudgel) had accidentally broken off during previous handlings (about 1984) but was refitted. Various damages at the feet are still visible, possibly caused by Valsuani’s work or may have happened during transport.


The Buffalo Version

This plaster served as the cast model (Giesserform) for the Siot-Decauville bronze sculptures. It is still occasionally assumed in the literature that the Buffalo plaster may be considered the model taken from the original clay. The plaster was purchased in 1954 in Paris for the Albright Knox Art Gallery in Buffalo. This model had also been used by Rudier in 1929 for the edition of 20 samples, as noted by E. Fuchs in “Der Maler Daumier”. Both Rudier and Siot-Decauville were experts in the “au sable”-method , but not in the “lost wax” process. Yet, strange as it may seem, Ratapoil-sculptures produced by Rudier in the lost wax process do exist (e.g. Städelsches Kunstinstitut, Frankfurt), a fact which has been confirmed by the author’s own research (note by the Daumier Register: research showed that Rudier’s „Louis XIV“ for example was also done after the lost wax process; this for Rudier atypical work was presumably sub-contracted to another foundry).

The Buffalo plaster is covered with tan layers, heavily mottled (Schellack) and with varying deep brown stains. Like in the Geneva version, traces left behind by tools are clearly visible and numerous corrections and adjustments have been added to the plaster over the years. Some of the lines found are reminding of a textile imprint. The moustache/beard, a characteristic sign of recognition of Ratapoil’s physiognomy, has become less pointed, ‘heavier’ and shorter than in the Geneva version. The surface of the later bronze samples showed numerous changes and “softenings”. Xeroradiographic images revealed that entire sections of the legs and the beard had been removed, reconstructed and refitted. In total some twelve iron and steel rods were detected inside Ratapoil. The photography clearly shows the difference of material between the original structural support and the one added at a later date. A very large part of the figure has been “rearmed” at various occasions to ascertain its structural integrity.

When a plaster model is molded from the original unbaked clay model, the latter will customarily be destroyed. The new plaster model will show on its surface traces of seams. Because the surface of the Buffalo plaster has been worked over frequently, hardly any of these seams are visible. Only one small seam can be seen, which is also visible in the same position on the Geneva model. Thus a final proof of the Buffalo plaster being the original mold from Daumier’s clay model is no longer possible (contrary to Wasserman’s opinion).


The report comes to the following result: The Geneva version with its numerous small sections must have been molded on a ‘flexible’ material and can certainly not be considered a surmoulage from a previous plaster. The

Geneva version cannot definitely be considered a surmoulage from a previous plaster, as the numerous small sections of the mold in which the plaster was cast were possibly formed directly on a `flexible´ material, i.e. unbaked clay.

The surface of the Buffalo version has unfortunately been reworked repeatedly at so many occasions that it is impossible to confirm that it was molded from the original clay.

When comparing the plasters to the various bronze editions one will observe that some details from the original Siot-Decauville version, which showed up in the Buffalo plaster, had disappeared in the Rudier bronzes of 1929. (This observation differs with Wasserman’s on p. 169) The 15 Valsuani bronzes produced in 1960 from the Geneva plaster still show these details initially visible in the 1890 version. It seems that the cleaner and relatively untouched surface of the Geneva version may indeed reproduce / represent Daumier’s original clay design.

Assuming that both plasters have initially been molded from the identical base/mold would it be safe to assume that both plasters had initially been cast in the same mold? Further research into Geoffroy-Dechaume’s oeuvre may possibly answer this question. There is however no doubt that both plaster figures have been modeled with great craftsmanship and expertise and have served as models for later bronze editions.

Finally it should be pointed out that the Museum’s report not only supplies in-depth information going far beyond the above observations , but shows also some highly interesting and ‘revealing’ photography.

We also wish to remind all Daumier amateurs that the Kunsthaus in Zürich, Switzerland, starting on December 7, 2007 will open an exciting Daumier bicentennial exhibition, where the Geneva Ratapoil will be exhibited. The lithographic counterparts of this statue can also be seen, as well as a large selection of drawings and a rare lithographic stone.