[See BMCR 2016.02.47 for a response by Brunilde Ridgway.]
Introduced in the first essay of this catalogue by the editors, Daehner and Lapatin, and running like a thread throughout the exhibition, is the assumption that the repetitive casting of a series of bronzes in the indirect lost-wax technique from a single artist’s model was the norm in the production of large-scale Greek sculpture. But the notion of the serial casting of bronze statues in Greek antiquity is highly controversial. Its endorsement in areas of the catalogue and exhibition, and in parts of the film presented with the exhibition at the National Gallery in Washington DC (through March 20, 2016), and its appearance in Brunilde Ridgway's review, confuses the story line of a selection and presentation of ancient bronzes that is otherwise outstanding in scope and appeal. 1
The eminent French scholar Claude Rolley consistently and firmly rejected the theory that Greek bronzes were routinely cast in multiples and that replication in serial production was characteristic of Greek art. Rolley pointed out that there were only exceptional reasons for duplicating statues in the Greek period, citing as an example the Archaic marble twins Cleobis and Biton in Delphi, a duplication dictated by the subject. He called Carol Mattusch’s long- promoted idea of the re-use of the same original model for the two early fifth-century Riace statues “bizarre.”2
In a discussion of the Riace statues in her essay in the “Power and Pathos” catalogue Mattusch expands this (113): “…they were surely based on molds taken from a single model. But that was only the starting point: in the wax working-models, their feet were placed in slightly different positions; and their hair, faces, and beards, were markedly differentiated…” The differences in the two statues are sufficiently major to have encouraged some scholars to assign them different dates. Such radical reworking of a wax model is magical thinking. Even the most minor alteration necessitates labor- intensive carving into hardened wax and/or adding softer wax for modeling. Truer to the artists’ creative process in all malleable media for working out compositions in three dimensions is direct modeling of whatever is required. For the Riace statues, the artist (or artists), together with the workshop, created two original wax models from which they took two sets of piece molds to make the two wax “inter-models” to invest and cast.3
If the Hellenistic period was “an early ‘age of mechanical reproduction’ ” (Daehner and Lapatin, 33) and the bronze industry was “geared to the production of multiples” (Mattusch, 111) one would expect some replication in a category of small figurative bronzes for which there was high demand, like the hollow cast fulcrum attachments produced to decorate the leaning headrests of Hellenistic and early Roman beds. In 1989, over 400 of these bronze fixtures, produced in the late 3rd century BC to the first half of the 1st century AD, were catalogued; each one was uniquely cast. 4
Rolley dedicated the last chapter of his 1999 book on Greek sculpture to the issue of copies, repeating his firm objections to contemporary replicas of an original Greek bronze.5 In the late Hellenistic period figural bronzes were sometimes produced in pairs for display in Roman households. But in those known to me, each bronze was made in a separate process that often included some direct casting, the two works reveal major or minor differences, and both are of equal quality.6
In her review, with characteristic generosity, Ridgway comments on all the essays and many of the catalogue entries, adding bibliography, gently disapproving “the natural tendency to identify facial features as portraits of specific personages, or to attribute statues to the hand of major masters,” and applauding “the wide-ranging” dates and that “the meaning of ‘Hellenistic’ oscillates between chronology (traditionally 323-31BC) and modes of expression.” Because she accepts without question the editors’ promulgated theory of the serial replication of Greek bronzes, however, we should look carefully at the three bronzes found in the sea, the “comparative display(s)” of which Ridgway says exemplify “the easy replication of bronze types and therefore their relative lack of ‘originality’ in our sense of the term.” Such an examination would be especially pertinent as she says these comparisons “validate one of the most important purposes of the entire exhibition.”
The first comparative display Ridgway discusses is the bronze Dionysos herm from the Mahdia shipwreck with the signature of Boethos of Kalchedon as artist (Cat. 45) and the similar herm in the Getty Museum (Cat. 46). As curator at the J. Paul Getty Museum in the 1970s, Jirí Frel purchased this herm, which is without provenance, from a consortium of Swiss dealers in Geneva. Frel believed the Getty herm to be “a replica of the herm found in the Mahdia shipwreck.” A new feature, he said, was the two-stepped cast socle, or plinth, soldered to the base of the herm, which the Mahdia herm does not have. Frel thought the surface of the Getty herm showed “a patination of the first century AD,” effectively dating it to the Imperial period.7
The dimensions of the two herms are very close, but their means of production were very different. The Mahdia herm (Cat. 45) was directly cast, as was made clear in the two studies by the German curator Frank Willer in the 1994 publication of the bronzes and marbles from the Mahdia shipwreck.8 In contrast, the Getty herm (Cat. 46) was indirectly cast from molds taken from a model in a single operation.9
Overlooking the differences in their production techniques in his catalogue entry for the Getty herm (Cat. 46), Daehner concludes that the correspondence of their measurements, together with similar alloys with “elevated” levels of the trace element cobalt, indicates the two herms “were produced at the same time and in the same workshop, in fact, from the same batch of metal.” (285). But are these features of the two herms really reason enough for Ridgway to also assume they “must be attributed to the same workshop, perhaps even the same pour.” ?
Cobalt is a trace element in raw copper and in lead and it is not unusual to find it in small quantities in the alloys of bronzes. In the Catalogue Appendix prepared by Jeffrey Maish we find that both the Mahdia and the Getty herms have more than a trace of cobalt (at most 0.28% and 0.15% respectively) and, measured by another analytical method, so does the Getty Youth (Cat. 12) (0.1290% Co in the sample with the highest amount.)10 Analyzed by the same method as that used for the herms, a large satyr statuette from the Mahdia shipwreck has 0.16% Co (F 209) and most of the other Mahdia bronzes also contain cobalt.11
The Mahdia herm’s headdress of broad fillets is draped over all but two areas of the head—where a textured surface suggests braids—arranged in broad loops of varying lengths with areas of deep undercutting in a tour-de-force of composite illusionism. The entire ensemble was made free hand with ribbon-shaped strips of wax. In the back, a negative oval space created by a fillet’s long loop is defined by a penetration through the bronze that is completely open. This void, the most obvious of two such openings, can be seen clearly when viewing the herm from its sides. But it is not easy to see in most published photographs. Available images in print that show it clearly are a view of the left side of the herm, an X-radiograph, and the images made by 3-D computerized tomography.12
As Willer wrote in one of his 1994 publications, the rigid plaster piece-molds of antiquity used for indirect casting could not have accommodated what was required for this elaborate headdress to be reproduced in bronze. Willer’s logical conclusion was that the Mahdia herm, with the head joined to its shaft—which was constructed from four sheets of wax—had been directly cast.
Willer believed that a hollow wax shell of the head was prepared before the complicated assembly of ribbons was added. But a simpler and less laborious way, and the time-honored practice for creating a wax model to be directly cast, would have been for the artist to model some preliminary forms in the moist clay of the core for guidance in the preparation of the wax head. When dry, the clay would be coated with a layer of wax as thick as the walls of the eventual bronze were intended to be, and the artist would proceed to model the head with its complex headdress. The clay core—secured in place with iron chaplets—would shrink in the casting, as was normal, and would be removed afterwards. As a direct cast, the Mahdia herm is unique, encapsulating the sculptor’s original idea as worked out in wax, which was lost, in the more permanent material of bronze.
It is conceivable that the model for the indirectly cast Getty herm was, in fact, the Mahdia herm. Along with the correspondence of other dimensions, the measurements and simple volumes of the faces of the two herms are almost identical. But the Getty herm’s face has a lifeless surface, and the elements of the beard are drawn rather than modeled. The curls hanging down the back of the shaft have a similarly two-dimensional quality, and the two areas of the top of the head not covered by fillets are without texture. Most telling, the undercut details that make the Mahdia herm’s headdress so lively and would be impossible to replicate in molds made for an indirect cast, are filled in. The Getty herm’s headdress has an ungainly rounded protrusion in the back that exactly follows the outline of that area of the Mahdia herm’s headdress where the loop of a fillet creates the oval opening. It is possible that this awkward form on the Getty herm head was determined by a mold’s accommodation of that void.13
Included in a poster on the casting of the Getty herm prepared by Jeffrey Maish for display at the 19th Congress on Ancient Bronzes held at the Getty Museum in October, 2015, is an image of an overlay of two two-dimensional computer tomography (CT) cross sections of the heads of the two herms. One image reveals a cross section of the walls of the head of the Mahdia herm that appears to be at the area where the ribbon creates a void; the overlaid 2D-CT image of the Getty herm shows the walls of a cross section of the Getty herm head at the corresponding level. The relationship between them calls for careful study and interpretation.14
Also in his poster, Maish draws new attention to one of the “unique and more puzzling casting features of the Getty herm,” vertical “tubular remains attached to the inner bronze walls.” These were studied by David Scott and Jerry Podany in 1990 and mentioned again by Scott in 2011, where they are described as four partially hollow bronze rods that traverse the length of the shaft.15 Maish concludes that they constituted an interior gating system, which is an anomaly in ancient bronzes. (Wax rods, to establish “gates” to regulate the flow of bronze during casting, were attached to the exterior of the wax model in antiquity; we know this from bronze fragments found in ancient casting areas.) The tubes in the interior of the Getty herm were briefly discussed in one of his 1994 publications by Willer, who quoted the conclusion of Edilberto Formigli that interior casting gates do not appear in bronzes until the Renaissance.16
Another of Ridgway’s examples of “replication” is of a very different nature. The dramatic juxtaposition in the exhibition at the Getty Museum of two bronze copies of a fourth-century BC statue of a bronze athlete cleaning a strigil (Cat. 40 and Cat. 41), together with the bronze head of a third copy (Cat. 42)—all of different dates over a period possibly as long as three centuries—was highly informative, but their relationship to Ridgway’s definition is ambiguous at best. (The bronze from Croatia, found in the waters of the Northern Adriatic, is absent from the exhibition in Washington D.C.). 17
These bronzes are further proof of a practice well known but, due to the melting down and re-use of bronze throughout antiquity, not often illustrated: that late Hellenistic and Roman copies of original Classical Greek bronzes were made in bronze as well as in marble, probably widely. What they conspicuously suggest is that the original fourth century BC athlete—at rest, cleaning his strigil after some contest—was available to be copied over a long period of time and that, whoever its artist, it must have had some renown.
Ridgway’s final juxtaposition is a pair of “Archaistic” Apollos who stand in the same pose with similar arm gestures—a pattern known from Archaic images in many media. Recent study of the lead inscription found inside the Apollo that was discovered off the coast of Italy near Piombino in 1832 (Cat. 47) confirmed that it was made in Rhodes, and, with the inlaid silver inscription on its foot suggesting it was an offering to Athena, that it was probably erected in the sanctuary of Athena Lindia “during the last quarter of the 2nd century BC.” (Sophie Descamps-Lequime, 290).18 The Apollo from a house in Pompeii (Cat. 48, “1st century BC-1st century AD”) was produced as a useful household item, holding elaborate bronze tendrils supporting a tray onto which were placed lamps, or perhaps dishes for food. The two statues have considerably more than superficial differences.19
In the Epilogue to the third edition of her book Archaic Greek Youths, Gisela Richter suggested that the more complicated movements that artists sought to represent by the fifth century BC may well have increased the use of bronze over marble, allowing the sculptor to “pursue his naturalistic aims with greater ease” by modeling in clay or wax than by carving directly into stone.20 It is virtually a truism today that naturalism developed significantly during the Hellenistic period, and the use of bronze undoubtedly facilitated this. But there was an important corollary: a desire to depict emotional and psychological human truths, and to reveal the individual, an objective that was, in effect, a new “canon of truth.” The concept is brilliantly illuminated by Gianfranco Adornato’s catalogue essay, which opens with the epigram by the poet Poseidippos of Pella describing the statue of Philitas by Hekataios (49-59).
Fortunately prevailing over the confounding and limiting propaganda about serial production, the greatest strength of this exhibition, suggested by its title, Power and Pathos, is elsewhere. It is an emphasis on those bronzes that express the remarkable Hellenistic artistic innovations conveying this new “canon of truth.” They are not difficult to find: the head from Delos (Ca. 29), the North African man (Cat. 28), the portrait of the Roman equestrian (Cat. 39), the Arundel head (Cat. 27), the portrait of a man from the Bibliotèque nationale (Cat. 32), and two bronzes not in the Washington D.C. exhibition: the head of Seuthes III (Cat. 9) and the Seated Boxer (Cat. 18), among others.
1. There are basically two methods of casting hollow bronzes by the lost-wax technique, the direct and indirect. In the direct technique, most successful for smaller formats, a model was built up of beeswax over clay that would stay in place during the casting as the “core,” shrinking sufficiently from the heat to be extracted from the hollow bronze sculpture afterwards. Wax rods were attached to convenient areas of the wax form, and it would be covered (“invested”) with a heavy layer of clay, leaving the tips of the rods free. When the investment was dry, the assembly was heated in a furnace, firing the clay and burning out the wax (thus lost.) The investment was now a hollow ceramic mold. Into this, through one of the holes—“sprues” or “gates”—left by the wax rods, molten bronze was poured. Others of the holes allowed the heated air to be released. After the investment was cool, the mold was broken off to reveal the form of the wax model, now transformed into bronze. The bronze rods that had been formed in some of the holes—“sprues” or “gates”—were cut off and the surface finished with hand tools. The indirect method involved casting a second wax model made in molds taken from the original model made by the artist. This second model was called an inter-model, and it was this model, not the original, that was cast after being supplied with a clay core, following the same procedures above. The first difficult step was making the molds from the original model in order to create the inter-model. In antiquity (and later) plaster molds were made in pieces, the number of pieces depending on the size of the sculpture and the skill of the workshop (typically the head, the torso, the arms and the legs). And see Formigli, below n. 3. Once achieved, the plaster piece molds were assembled as hollow shells for the interiors to be lined with molten wax that was either brushed on or swished around for the desired thickness of the bronze wall with the excess wax poured out. Freeing the wax inter-model from the rigid plaster piece molds was the second difficult step and it is likely that in Greek antiquity these molds were destroyed in the process. The assembled wax inter-model was then filled with clay as a core to ensure that the final bronze would have walls of a uniform thickness. As with direct casting, the core would be fixed in place by chaplets—fine iron wire that penetrated the wax walls into the core—wax rods added as gates or “sprues,” the assembly invested, wax melted out, molten bronze poured in, etc. Today, molds made of rigid plaster contain inner molds that are usually made of latex, polyurethane rubber, or silicone to take the negative impression of the original model and make the extraction of inter-molds much easier. According to Haynes, highly elastic materials for this purpose existed as early as the 19th century. Haynes, D.,The Technique of Greek Bronze Statuary, Mainz/Rhein, 1992, 62, n. 31. Also see Hayes for a more detailed description of these two bronze casting options, 24-33.
2. Rolley, C., Les bronzes grecs, Fribourg/Paris, 1983, 19, 33; id., “Les bronzes grecs et romains: recherches récentes,” Rev. Arch., 2/1990, 407; id., La sculpture grecque, 1, Des origins au milieu du Ve siècle, Paris, 1994, 349 (Riace bronzes). Also Haynes, above n. 1, 41, n. 25.
3. Formigli, E., “La tecnica di construzione delle statue di Riace,” in Borrelli, L.V. and P. Pelagatti (eds.), Due bronzi da Riace. Rinvenimento, restauro, analisi, ed ipotesi di interpretazione. Bollettino d’arte 3, serie speciale Rome, Vol. 1, 1984, 107-142. Rolley believed they were made by two artists in the same workshop. Rolley, above n. 2, La sculpture grecque, 1, 349.
4. Faust, S., Fulcra. Figürlicher und ornamentaler Schmuck an antiken Betten, Munich, 1989. Two mule heads following the same prototype with only minor differences that may have been produced in the same workshop differ in size by 6.4 cm. Faust, Cat. 82 and 324. Rolley pointed out that even the horizontal handles of the same bronze hydria are frequently different in dimension. Rolley, above n. 2,“Les bronzes grecs,” 19. What seems too often disregarded in this discussion is the facility with which artists can reiterate, particularly in a malleable material like clay or wax, an idea they have once resolved, i.e., repeat a model for another casting, whether by the direct or indirect method.
5. Rolley, C., La sculpture grecque, 2. La period classique, Paris, 1999, 406-410.
6. E.g., two bronze lamps in the forms of running figures from the Mahdia shipwreck, one an Eros, the other a hermaphrodite, clearly cast independently from different models. Hiller, H., “Zwei bronzene Figurenlampen,” in Hellenkemper-Salis, G., Das Wrack. Der antike Schiffsfund von Mahdia, Rheinisches Landesmuseum Bonn, Cologne, 515-530. The large statuettes of boys with pyramidal hats in draped and wrinkled Eastern costume found together in Egypt in 1912, now in the Metropolitan Museum of Art, NY, and the Walters Art Gallery, Baltimore, are apparent twins, probably lamp holders for simultaneous display. Mattusch, C. The Fire of Hephaistos. Large Classical Bronzes from North American Collections, Cambridge MA, 1996, 259-262, maintaining that the statuettes “shared the same master molds” and were only two in “the production of a series.” But the complicated surface of each of these sculptures must have required either some direct casting or multiple unyielding plaster piece molds that were almost certainly destroyed in the process of removing the wax inter-models for casting.
7. Frel, J., “Some Greek Sculpture in Malibu,” J. Paul Getty Museum Journal 8, 1980, 5. Bronze Herm. 79. AA.138. Stewart later dated it to the first or second century AD, calling it a “coarser version” of the Mahdia herm. Stewart, A.,Greek Sculpture. An Exploration. New Haven and London, 1990, 229, fig. 849-850.
8. Willer, F., “Die Restaurierung der Herme,” in Hellenkemper-Salies, above n. 6, 953-958 with col. pl. 33; id.,“Zur Herstellungstechnik der Herme,” in Hellenkemper-Salies, above n. 6, 959-970. See the comparative drawing of the faces of the two herms, 986, fig. 17. For the similar dimensions of the two herms, also Mattusch, C., “ Bronze Herm of Dionysos,” in Hellenkemper-Salis, G., above n. 6, 431-450; and Mattusch, C., The Fire of Hephaistos. Large Classical Bronzes from North American Collections, Cambridge MA, 1996, 186-191, where they are described as “two ‘editions’ in a series of castings molded from one basic model.” (190)
9. Scott, D.A. and J. Podany, “Ancient Copper Alloys: Some Metallurgical and Technological Studies of Greek and Roman Bronzes,” in Small Bronze Sculpture from the Ancient World, Malibu, 1990, 47-55. Scott, D. A., Ancient Metals: Microstructures and Metallurgy, Vol. 1, Los Angeles/Hertfordshire, 2011, 134-135.
10. Catalogue Appendix, herms, 321-322, analyzed by EDXRF; Getty Youth, 313, analytical method ICP-MS. Appendix, 311, for descriptions of techniques.
11. The satyr, Hellenkemper-Salies, above, n. 6, 1053, table 1, no. 2180 (F 209), C. Klages, C., “Die Satyrstatuetten,” 531- 538. Other bronzes, analyzed by EDXRF, e.g., lamps (up to 0.14%), candelabra (up to 0.13%), Hellenkemper-Salies, above n. 6, 1053-1061.
12. Mattusch, above n. 8, “Bronze Herm,” 436, fig. 9; Willer, above n. 8, “Die Restaurierung,” 956, fig. 5 (X-radiograph); Goebbels, J., B. Illerhaus and D. Meinel, “Computertomographie an ‘Agon’ und Herme,” in Hellenkemper- Salies, above n. 6, 988-989, fig. 9-10.
13. The best photographs for making these comparisons can be found in Mattusch, above n. 8, “Fire of Hephaistos,” 188-189: Getty herm, above, figs. 3b,c,d; Mahdia herm, below, Cat. 3, figs. 1b,c,d. Plaster cast replicas of both the Mahdia herm and the Mahdia Eros were made in Tunis in the early 1940s. In an attempt to put them together as a group, the cast of the herm was placed on a two-stepped base to support the Eros’ right leg, which had been incorrectly restored. Söldner, M., “Der sogenannte Agon,” in Hellenkemper-Salies, above n. 6, 401, fig. 3-4; 424, ns. 16,18. The casts are lost.
14. Investigation of the Getty herm head by 3D-CT imaging would also be useful, as would scrutiny of the surface of the Getty herm for indications of seams left by molds at their assembly, which Willer advised in 1994. Willer, above n. 8, “Herstellungstechnik,” 969. The poster should be published with the collected papers of the Congress. I am grateful to the author for sending me a copy so I could make accurate references to it.
15. Scott, above n. 9, 134-135. Like Willer (below, n. 16), Scott suggested they might have been made of organic material, like reeds. Scott and Podany, above n. 9.
16. Willer, above n. 8, “Herstellungstechnik,” 969. Formigli, E., “Zur Form und Gusstechnik des Jünglings vom Magdalensberg,” in Griechische und römische Statuetten und Grossbronzen. Akten der 9. Tagung über antike Bronzen, Wien, Vienna, 1988, 35-38. It was established in 1988 that the Magdalensberg Youth was a copy made in 1551 of a first century BC bronze original discovered in 1502, now lost. The Getty herm cannot be a Renaissance replica, because the Mahdia herm was at the bottom of the Mediterranean sea from the date of the shipwreck, concluded from the dates of the associated clay amphorae to have been about 90-60 BC, until its retrieval from the sea off the coast of Tunisia in 1907.
17. The Ephesos copy, excavated in 1896:first half of the 1st century AD (Cat. 40); the Croatian bronze found in the Northern Adriatic in 1997:the 1st century BC (Cat. 41); the head, recognized as ancient on its 16th century bust and purchased at auction in 2000 in New York by Timothy Potts, now Director of the Getty Museum: the 2nd BC to 1st century AD (Cat. 42).
18. It was Ridgway whose meticulous 1967 study of the Piombino Apollo proved it a Hellenistic pastiche of earlier styles. Ridgway, B.S., “The Bronze Apollo from Piombino in the Louvre,” Antike Plastik 7, 1967, 43-75.
19. To these compare the Piraeus Apollo, whose pose differs only in the reversed position of the legs; dated tentatively by Rolley to 150-100 BC. Rolley, above n. 2, Les bronzes grecs, 36. In the third and last of her comprehensive volumes on Hellenistic sculpture, Ridgway nevertheless suggested that the two statues, both defined by her term “Archaistic,” “could possibly come from the same workshop.” She does not repeat that in this review. Ridgway, Hellenistic Sculpture III, Madison, WI, 2002, 147-148.
20. Richter, G.M.A., Kouroi. Archaic Greek Youths. London and New York, 1970, 148.