Between piece molds and lost wax: the casting of a diatrete ornamentation in early China rethought

Abstract

For a long time, research on the pre-Han (before 202 BCE) Chinese use of the lost-wax process was targeted at demonstrating that it was of no use at all. That position was severely shaken when the zun-pan set of Zeng Hou Yi (“Marquis Yi of Zeng,” d. 433 BCE) was unearthed at Hubei Suizhou in 1978, along with subsequently excavated finds or rediscovered artifacts carrying similarly interpenetrating openwork. Since the early 2000s, however, there has been a persistent revival of a research orientation against the use of the lost-wax process in Bronze Age China. After surging first in Chinese academia, the new tide took some time to arrive in the West. A response to Donna Strahan’s prestigious lecture, “Debating the Use of Lost-wax Casting in Ancient China,” this article presents my thinking about lost wax versus piece molds in early Chinese casting practices. Strahan’s studied “diatrete” ornamentation in the Metropolitan Museum of Art, to me, can be securely identified as lost-wax cast thanks to some telltale traces. Through this critical investigation, I invite my fellow art historians, archaeologists, and historians of metal technology to emerge from the “doubting of lost wax” era to tackle new challenges in early Chinese archaeo-metallurgy.

Introduction

The essentially three-dimensional character of some Eastern Zhou (770–256 BCE) bronze openwork ornamentation has evoked both praise and wonder. How early Chinese craftsmen accomplished this three-dimensional structure is not easy to discern, and the observed correlation of design and technology may not have been firm in practice. For instance, certain vessels called “diatrete” by Helen Loveday, with an external openwork shell set over the internal container, exhibit this thorny problem—some of them were exclusively cast using the piece-mold (i.e., section-mold, see Fig. 1) technique; the rest, according to Loveday, may have been cast using lost wax (see Figs. 2 and 3), the most prevalent means of investment casting (Loveday, 2002, pp. 102, 131–33).1 In this diatrete group, a spherical tripod in the Metropolitan Museum of Art (hereafter “the Met,” Fig. 4), though far from the finest, is among the most curious pieces in the debate over “the use of lost-wax casting in ancient China” (as Donna Strahan described it in her 2015 Forbes lecture).2 Compared to some scholars, Strahan is less unyielding in her doubts about the pre-Han (before 202 BCE) adoption of lost wax in early China. Her relatively nuanced position makes Strahan a useful interlocutor for clarifying the issue of lost-wax use, which has implications that extend to the status of early Chinese civilization.

Fig. 1 [Images not available. See PDF.]

Mold diagram, the piece-mold process to cast a bronze ding.

Probably cast upside down, this vessel could be underpinned by the bulky clay core shown at the bottom right. In many such cases in early China, the finished vessel was substantially a bronze replica of the decorated model. Hence, the central item in the diagram can act for either thing; first imagine that it represents the model. Wet clay was compressed around it, dry-hardened, and peeled off in three fitted sections, and thereafter recombined around an appropriate core. Maintaining its distance from that inner core desired for the thickness of casting (e.g., by metal spacers), the piece-mold assembly was now ready for the casting of the final bronze. Note: mortises and tenons, often incised into the edges of mold pieces, should have been incorporated to this schematic drawing (script adapted from Bagley, 1990, pp. 9–10, and Bagley, 2009, p. 44; diagram after Wan, 1976, p. 27). This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of NGV Melbourne; copyright © Wan Chia-pao, all rights reserved.

Fig. 2 [Images not available. See PDF.]

The direct lost-wax process in one of its simplest forms.

Schematic drawing of the main steps in the manufacture of a South American tunjo: founders first shaped a wax model of that small figure (left). Clay, along with essential additives, was invested all around the model, and in this way an “investment mold” was built (middle). The wax was melted out of the baked mold, and the molten metal thereafter filled the cavity left by the wax; once the metal had gone hard, the roughcast product could have been brought to light (right). After M. Martinón-Torres and M.A. Uribe-Villegas 2015, p. 379, Fig. 1. Drawing by Ana Rivera Uribe (Bogotá, Colombia). This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of Marcos Martinon-Torres; copyright © M. Martinón-Torres and M.A. Uribe-Villegas, all rights reserved.

Fig. 3 [Images not available. See PDF.]

The indirect lost-wax process in one of its typical schemes today.

The application of permanent piece molds (the “master mold”) could facilitate serial production of wax models. After Craddock, 2015, p. 73, Fig. 23b. This figure is covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of Paul Craddock; copyright © The British Museum, S. La Niece and A. Simpson, all rights reserved.

Fig. 4 [Images not available. See PDF.]

The globular diatrete tripod.

Warring States period (475–221 BCE). Height 21.9 cm. The Metropolitan Museum of Art, 1947.25. After Chase, 1991, p. 73, no. 35; Strahan, 2019, p.17, Fig. 14. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permissions of The Metropolitan Museum of Art and Donna Strahan; copyright © The Metropolitan Museum of Art, Corey J. Tucker, all rights reserved.

Echoing Erlitou a millennium later?

Regarding the lost-wax problem, Donna Strahan devotes all her case studies to artifacts bearing three-dimensional openwork interlaces. The choice to set aside other categories of design that could potentially be fashioned by lost wax3 is understandable. One of the main reasons for suspecting that an artifact was lost-wax cast lies in its shape: despite the practical concerns (e.g., systematic venting for elaborate casting; see Peng, 2018a, pp. 108–109), lost wax is appealing to theorists because it challenges the limits of structural complexity. Interlaced openwork, if equipped with a daunting three-dimensionality, naturally draws scholarly attention to the possibility of lost-wax casting. Not all pre-Han investment castings, however, follow what is assumed to be the usual pattern.4 It is also worth mentioning that, apart from the zun-pan set of Zeng Hou Yi (“Marquis Yi of Zeng,” d. 433 BCE, Fig. 5), which is inescapable for any lost-wax study of early China, Strahan’s target pieces are from either the Met or the Freer Gallery. This selection may have been made based on first-hand access, and so overgeneralization is a risk. Strahan concludes that neither the Zeng zun-pan nor any of the samples she inspected directly seem to be sufficient to tell “whether they were made by piece-mold casting or by lost-wax casting, or both” (Strahan, 2019, p. 26). Behind this truth-seeking prudence, astute readers can sense Strahan’s (2019) general inclination against lost wax, which is most notably revealed by her judgment on the Met tripod—the so-called “incense burner”—when discussing its diatrete shell (i.e., the “sphere”):

The sphere is enclosed in two openwork hemispheres that make up an outer decorative shell. They have an elaborate three-dimensional openwork design of intertwined snakes…. The decoration is typical of Eastern Zhou vessels with interlaced zoomorphic forms in horizontal bands extending around its circumference. Bagley believes that this type of a linear decoration is unlikely to occur in lost-wax casting because a mold made by the lost-wax method is closed and can be worked only on the wax model. He notes that the pattern used here is constrained by the piece-mold section. The design was drawn only on one of the mold sections. It is therefore confined by the arc spanned by that section. The side boundaries of the design correspond to the vertical lines of the mold joins. It was easier and more natural to draw pattern units that were complete within a single mold section. The decoration that resulted on the cast object was then composed of self-contained units whose boundaries clearly showed the divisions of the mold assembly. This casting method constrained how a bronze was decorated (Bagley, 1987, 18).

(Strahan, 2019, pp. 18–19)

Fig. 5 [Images not available. See PDF.]

The Zeng Hou Yi zun-pan set from the Tomb of Marquis Yi of Zeng (d. 433 BCE).

Height of zun 30.1 cm. Height of pan 23.5 cm. Adapted from Notis and Wang, 2017, p. 13, Fig. 13. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of Springer Nature; copyright © Hubei Museum, all rights reserved.

This assessment is somewhat convoluted, as Robert Bagley on the page Strahan cites was not remarking on the Met tripod or anything close to it. Bagley was discussing something from approximately a millennium earlier: the first known two-dimensional adornment of Chinese ritual bronzes presented by an Erlitou jue (Fig. 6), dated to the sixteenth century BCE (Bagley, 1987, p. 18). In the Chinese bronze assemblage, nothing specific from the Met sphere can be stylistically or artistically tracked back to the Erlitou surface decoration. Confusingly, Strahan calls the “interlaced zoomorphic” pattern of the Met shell a “linear decoration,” perhaps because the outstanding “intertwined snakes,” highlighted by the hollowed-out space, have a somewhat linear movement. Alternatively, it may be that Strahan is using “linear” in the sense that the interlaced pattern of the Met openwork “forms in horizontal bands extending around its circumference”—if so, the label “linear” has little identifying value, since this fashion “is typical of Eastern Zhou vessels,” as Strahan realizes, and is not peculiar to the Met sphere alone. There appear to be no grounds for associating the formal counterpart, prototype, or even stimulus of the Met interlace with the Erlitou “linear” ornament. With no mention of Erlitou or any indication of the immense temporal jump she is making, Strahan rushes to the anachronistic association of the Met openwork and the Erlitou pattern: “Bagley believes that this type of a linear decoration is unlikely to occur in lost-wax casting….” This phrasing may give inattentive readers the impression that behind Strahan’s argumentation is a trustworthy assessment that lost wax was not used in casting the Met shell. Bagley’s insight, though accurate in its own setting, is misused by Strahan to support a different argument.

Fig. 6 [Images not available. See PDF.]

The bronze jue (left) from the third stratum of Erlitou (Henan Yanshi, ca. 16th century BCE), Height 22.5 cm.

Its thread-relief decoration (right, in both photograph and rubbing) only appears on the side of the pitcher opposite the grip. After Zhongguo, 1996, pl. 7; Fong, 1980, p. 74, figs. 17–18. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permissions of Luoyang Museum and Robert W. Bagley; copyright © Luoyang Museum and Robert W. Bagley, all rights reserved.

Observations on the so-called “linear” decoration were, of course, made long before Bagley’s. In 1937, for instance, J. Leroy Davidson had remarked on a he (“ho”) jug in the Museum Für Völkerkunde, Berlin, carrying what he called a “thread-like” relief. With Orvar Karlbeck’s publication on Anyang molds in mind (Karlbeck, 1935), Davidson proposed the possibility of “incis[ing] the mould with a fine graver” to achieve the linear pattern that was eventually raised from the metal surface; behind this, as Davidson conjectured, may stand “the first style of ornamental bronze in China of which the Berlin ho may be the earliest example” (Davidson, 1937, pp. 34 and 51). Given that visually more “primitive” thread reliefs (e.g., the Erlitou jue pattern) had not yet been unearthed at the time, Davidson’s conjecture was fair. In 1953, the surface adornment of pre-Zhou (before ca. 1050 BCE) bronzes was described by Max Loehr in a sequence of five styles, with the thread relief décor as the first (Loehr, 1953, p. 45). On the invention of Loehr’s Style I pattern, Bagley revisited Davidson’s discourse on the Berlin he: for what he called the “incised” thread relief, Davidson took the foundrymen’s access to the inside of mold into serious consideration; yet, perhaps affected by the prevailing conviction, Davidson did not dismiss the possibility of lost wax, whose investment mold cannot be opened up before casting (Davidson, 1937, p. 34; Bagley, 1990, pp. 10–11). Bagley’s attack on lost wax, cited by Strahan, can be seen as an amendment to Davidson’s thesis and the relevant discourses on the rise of the “linear” surface ornamentation—characterized by Loehr as Style I décor—of Chinese bronzes.5

Could the Met openwork interlace be categorized under Loehr’s Style I? Certainly not, as Loehr’s five styles were only meant to group pre-Zhou metalwork. Is there then a chance that the Met “type of a linear decoration,” as Strahan calls it, descends from or revives Loehr’s Style I? The answer is still no, as there is no detectable resemblance or connection between the two, on whatever scale we observe them. Interlace, openwork, or interlaced openwork—whatever that might refer to when Strahan calls it “typical of Eastern Zhou vessels”—is a different story, unrelated to the thread décor in relief. If the Met openwork interlace has nothing to do with the Erlitou thread-relief décor, then Strahan has no grounds to turn to Bagley’s argument on the birth of Loehr’s Style I “linear” decoration.

Notwithstanding, there is a “parallel” between the Eastern Zhou sphere and Erlitou jue that may explain Strahan’s perplexing association of the two. Unfortunately, that parallel is a fiction born from a voguish false belief about early Chinese bronze casting. According to Bagley’s scenario in drawing up the Erlitou jue pattern, “When the moment came to execute the decoration, the craftsman had the mold sections in front of him, but the bronze vessel did not yet exist: at that moment he must have been thinking less about the finished vessel than about the mold sections, and he was content to decorate only one of them, the largest one” (Bagley, 1990, p. 14). This is a vivid version of what Strahan paraphrases above, but can any point of this explication be applied to the Met openwork a millennium later? Strahan thinks so, as she seems to believe that Bronze Age founders starting from Erlitou had no need to individually decorate more than a single mold section or its corresponding range on the model. She is persuaded that once there was a need to expand surface embellishment beyond that limit, pattern replication could be conveniently done as she imagines it was for the Met openwork: “This description of linear decoration [by Bagley on the Erlitou jue] clearly coincides with what is found in the Metropolitan incense burner. The model for each hemisphere was a perfect dome, probably made of clay and produced by a rigid profile jig. After the dome-shaped model was made, only one-quarter of it was decorated” (Strahan, 2019, p. 19). Here, Strahan is apparently visualizing four “mold” sections of the Met hemisphere as serially achieved from one and the same decorated model unit. The Erlitou-led decorative consideration of the self-contained mold sections in Bagley’s explanation (Bagley, 1987, p. 18; Bagley, 1990, p. 14) is modified by Strahan into an idealistic self-replicating scheme of decoration. This scheme could resort to what we may call the “fractional model,” such as Strahan’s asserted quarter-model for the Met shell. This new point, which is a central proof in her argument against lost wax, will be examined below in more detail.

Identical piece-mold cast sections?

Strahan is, as far as I know, the first to use observed “identical” units to argue against lost wax, though discussions centered on identicalness have a long history in the study of Chinese bronzes.6 In the era when lost wax was still universally thought to have shaped pre-Han castings, some scholars already noticed that the early Chinese bronzes were almost never identical (Collins, 1931, p. 37). After the piece-mold theory entered the mainstream, students of early Chinese bronzes remained intrigued by the millennium-long absence of duplicate castings in the records. Many accordingly turned to Shi Zhangru’s famous but unverified proposal that the decorated model was pared down to serve as the core, explaining the shortfall of identical bronzes (Shi, 1955, p. 109; for a critical analysis, see Bagley, 1987, pp. 60 and 617). A few scholars even denied the Bronze Age use of decorated models when explaining the lack of metal duplicates: “when producing a series of identically-shaped vessels, why did they apply a slightly different decoration each time instead simply repeating the original design?” (Nickel, 2006, p. 38). The non-appearance of identical bronzes would be intelligible, Lukas Nickel argues, if we admitted that “Throughout the Bronze Age undecorated models stood at the beginning of the casting process. Decoration was always applied directly to the outer mold and was usually designed in a way that would fit the mold sections” (Nickel, 2006, p. 5).

It seems that Nickel, like Strahan, did not consider lost wax a promising option for early Chinese bronze casting. This is not, however, the reason his argument—which was stunning even among the piece-mold suggestions—shocked academia. The real point is that if Nickel’s view were adopted, surface ornaments as exquisite as those on the Four-ram zun or Winthrop fangyi (e.g., see Bagley, 1990, pp. 6 and 128) would all be taken to have been graved into the mold straightforwardly; it is not possible to visualize such delicate patterns inversely, not to mention the extraordinary difficulty of manually executing them in negative. For this reason alone, it would be reasonable to dismiss Nickel’s argument with a sentence, though Robert Bagley has taken the trouble to systematically disprove his points (Nickel, 2006, pp. 5–35; Bagley, 2009, pp. 39–84). In the same year that Nickel published his claim, Dong Yawei made an analogous argument about the invariable use of undecorated models, but with a more palpable parallel to Strahan’s later perspective in its underlying logic. Dong suggested that there had been an established tradition of using fractional models since Erlitou: for a jia vessel with Loehr’s Style I relief, only a one-third undecorated model needed to be—or ought to be, if we follow Dong more accurately—prepared; clay was packed against that single fractional model serially and removed with three identical sections before the pattern was cut into the mold. This style of operation underwent no essential change, Dong claimed, and the undecorated and fractional model was in dominant use in every stage of the Chinese Bronze Age (Dong, 2006, pp. 4–6, 41–49). Dong takes for granted that efficiency is a paramount concern, but if his narrative were accurate, one would expect that duplicate parts making up the whole that are congruent in shape would be pervasive in pre-Han castings. But no objects with precisely identical duplicates have ever shown up from the early Chinese Bronze Age (tentatively by c. 1050 BCE), even though any such object would instantly be the center of attention, not to mention that foundry attestation to fractional models is currently absent from the entirety of Bronze Age China (the era, say, from the early second to the late first millennium BCE).

Strahan’s claim about the Met shell bears a glaring resemblance to Dong’s sweeping assertion that models were made not integrally but merely as a portion of the whole. Yet their perspectives, even limited to the overestimated piece-mold casting, differ in notable ways: unlike Dong or Nickel, Strahan does not insist on the constant use of undecorated models from Erlitou to Eastern Zhou, though her comparison of these two distant endpoints of the Chinese Bronze Age is imprudent. Specific to the hemisphere of the Met shell, Strahan constructs a baffling procedure in which all ornamental units came from the same décor that had been initially applied to only a quarter of the overall model. Reminiscent of the partially decorated Erlitou jue for this step, Strahan’s ornament is nonetheless executed on the model rather than in the mold; in addition, her fractional décor will eventually turn into four replicates and make up a full circle. Strahan’s “decorated” fractional model, which is reminiscent of Dong’s dubious vision, is thought to serially produce mold units to be finally aggregated as a whole. The decorative subdivision or compartmentation, a hallmark of early Chinese piece-mold castings, is achieved here in Strahan’s refashioned manner. Perhaps for the sake of coordination—that is, to ensure the interlinkage of the “identical” mold quarters into a perfect circle—Strahan’s decorated one-fourth model is formed as part of a complete entity (a proposal she does not explain).9 Before stepping into the labyrinth of her technical construct, let us first examine Strahan’s (2019) piece-mold evidence, concentrating on the bottom hemisphere of the Met shell:

This can be seen on the incense burner when digitally overlaid. Each quarter of the hemisphere is identical, aligning perfectly with one another. The exterior horizontal rows and interior vertical rows of each quarter are identical matches to one another, including their slight irregularities (…). Traces of vertical mold seams can be found between the quarters with slight discontinuities (…). The remarkable precision to which they align suggests that the same mold was used to produce each quarter (…).

If this object was lost-wax-cast, how could it have been done? Since each outer section is identical, the same model or pattern was used. Was this made by pushing wax or resin into the mold to make four exact duplicates? Any wax pressed onto the model or outer mold would have to be removed and placed onto the core or inner mold. But this thin design, if removed from a mold, would be too flimsy. There would be distortions, pinching, slumping, or other indications of where sections were joined. There are none. Every row, vertical and horizontal, front and back, lines up perfectly.

(Strahan, 2019, p. 20)

Strahan’s “digitally overlaid” photographs have only been obtained, or exhibited to us, from one side of the openwork (Strahan, 2019, p. 21, figs. 17 and 19). At least half of the information is ignored or hidden, and it is the more telltale half: for any ornamentation or artwork, the invisible interior is often technically more revealing, as no suppression or cover-up of technology—“inherently ugly and unpleasant,” in our popular misapprehension10—is needed there. In a perusal of the Met warehouse,11 I managed to observe the inner surfaces of the hemisphere, which carry a variety of dissimilar three-dimensional undulations (Fig. 7). If digitalized, the inner units seem unlikely to overlap even when examined with Strahan’s reduced approach. In fact, the identicalness of Strahan’s inspected four quarters, even we choose to reluctantly concede it, does not substantiate her piece-mold proposal. In particular, it does not in principle rule out lost wax, the possibility of which is dismissed by Strahan because the “thin design” of wax is “too flimsy” to be “removed from a mold” (Strahan, 2019, p. 20). This idea about the flimsiness of wax is not new; it is shared by several deniers of lost wax to rule out its pre-Han use.12 We are left with the impression that the globally universal material for ancient investment casting failed only in China. In fact, if well processed, wax can be both pliable and tough enough to, for instance, maintain the visually far more “precarious” pre-Columbian votive rafts (which sometimes carry massed frail tunjos, like the one illustrated in Fig. 2).13

Fig. 7 [Images not available. See PDF.]

Details of the interior of the lower hemisphere of the Met shell, revealing piled C-shaped forms and absolute undercuts of the fully three-dimensional openwork.

Author’s photographs, courtesy of The Metropolitan Museum of Art. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of The Metropolitan Museum of Art; copyright © The Metropolitan Museum of Art, all rights reserved.

Another reason for Strahan to side with piece molds lies in her lack of success in detecting any of the traces lost wax calls for: “There are none”—a resolute judgment—when it comes to “distortions, pinching, slumping, or other indications of where sections were joined” (Strahan, 2019, p. 20). In her case studies, nonetheless, Strahan does discern jarring “stepped C-shaped” patterns (Fig. 7) on the surfaces of the Met sphere and a finial in the Smithsonian’s Sackler Gallery (Fig. 8): “most researchers say they are lost-wax-cast. I am not convinced that lost-wax casting could produce this repeated regular pattern…. If wax had been used, there would be more distortion and evidence of pinching or slumping” (Strahan, 2019, p. 22; italics mine). It seems that the proof is never enough for Strahan when it comes to lost wax, though her tone is less decisive when confronted with the C-shaped irreconcilables. By a stricter standard, however, the clues to wax use are too plentiful to dismiss: what originally formed the C-shaped forms appear to have been semisolid slices laminated in somewhat staggered piles. With evident softness and glutinosity, these slices give the impression of having been comfortably bended, folded, and adhered one another (Peng, 2020, pp. 69–70). Seen together, the C-shaped slice edges present wavy wrinkles that seem to have been left by a thick semifluid, which arguably may be a possibility for certain instances (Zhang, 2007, p. 90; Huang et al., 2014, p. 33).14 Yet a similar effect could have been produced by piling up waxy pieces with a higher degree of melting on their edges (e.g., for better gluing?), a particular likelihood for the present Met example.15

Fig. 8 [Images not available. See PDF.]

Bronze finial (left) and its details (right), ca. 6th–5th century BCE, 3 × 6.5 cm.

The Dr. Paul Singer Collection of Chinese Art of the Arthur M. Sackler Gallery, Smithsonian Institution (RLS1997.48.417). This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of National Museum of Asian Art, Smithsonian Institution; copyright © National Museum of Asian Art, Smithsonian Institution, Arthur M. Sackler Gallery, The Dr. Paul Singer Collection of Chinese Art, S2012.9.2515 (object and detail), all rights reserved.

To me, Strahan’s resistance to the feasibility of lost wax in understanding the stepped C-shaped traces is specious: “Creating these designs would be much more complicated if carried out in wax. It would be easier to weave wax strands into a basket than to have created exterior horizontal rows and interior vertical rows” (Strahan, 2019, p. 22). Rather than the Met sphere, the C-shaped forms of which are considered in a rather confused piece-mold context that requires effort to explain, Strahan is commenting here on the Sackler finial she examined. The logic is somewhat bewildering: if it would be “more complicated” to create the C-shaped patterns “in wax,” why would it be “easier” (and by how much?) to achieve them in clay, or any other material that was available? The answer Strahan provides, against the expectation of most readers, is neither an alternative modeling medium nor a different mold-making approach, but an exclusive design she deems appropriate for shaping in wax. It seems that in Strahan’s imagination, the sole form that wax naturally assumed in the decorators’ hands was “a basket.” As a matter of fact, that form does appear to have been assumed in some reasonable lost-wax instances, such as an openwork frame (variously called “hu cover” or object “base,” compare Wang, 2002, pp. 71–72; Strahan, 2019, p. 22)16 excavated at Jiangsu Pixian (Fig. 9). Though inspired by basketry, such a construction would not have arisen without experimentation with waxy materials: “When craftsmen obtained the wax, it would have been natural to think of rolling it and forming it into strips and then using them to fabricate something like the current design” (Peng, 2020, p. 136). Interestingly, this Pixian openwork frame, manifesting one of the conspicuous technical styles of lost wax, also bears the C-shaped forms.17

Fig. 9 [Images not available. See PDF.]

Details of the openwork frame excavated from the Liulin site, Jiangsu Pixian, Warring States period (475–221 BCE): exteriors (left, three-dimensional openwork of serpent interlaces) and interiors (right, revealing absolute undercuts and piled C-shaped forms).

Bottom right image after Wan, 2013, p. 75, Fig. 6; the rest photographs offered by Su Rongyu. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permissions of Nanjing Museum and Su Rongyu; copyright © Nanjing Museum and Su Rongyu, all rights reserved.

The most obvious witness to the C-shaped micro-features, which is curiously omitted by Strahan, is the always-spotlighted Zeng zun-pan set. In 2007, by studying the fragments fallen from the pan appendages, Zhang Changping remarked on the ripple-like marks presumed to have been caused by the flow of liquefied wax (Zhang, 2007, p. 90). That causal explanation, which is perhaps true in part, deserves continued attention. In 2014 I conducted a further examination, arranged by Zhang and guided by Yung-ti Li, and found upon unraveling it that it was mainly the lamination of staggered C-shaped slices that brought about the wavy wrinkles on the surface of the pan interpenetrating rods. When pressing the sticky sheets for adhesion, two ancient fingerprints have been preserved (Fig. 10) on a waxy medium, an unmistakable proof of handmaking (Peng, 2020, pp. 115–16, 317–21, figs. 71a–72d; Strahan, 2019, pp. 20 and 22). Extensive evidence of fusing waxy pieces with handmade marks has long been noted in, for instance, the spectacular jin table from Henan Xichuan Xiasi, a cemetery dated to the mid-sixth century BCE (Peng, 2018a, pp. 107–108).18 Though carrying no tangible fingerprint, the Met shell possesses a no less determining warrant for lost wax: with the C-shaped laminations and other features of lumpiness, the Met sphere interiors present a craggy—and certainly unintentional—three-dimensionality that is unlikely to occur in piece-mold casting. If this cragginess was not left by lost wax for technical reasons, why would piece-mold casters have taken needless pains to work on hidden areas nobody would see? (Peng, 2020, p. 69). And even if they had done so, the painstaking efforts of the casters would have turned into unpleasant irregularities and rough textures that would be out of touch with aesthetics. The logic of the piece-mold proposals does not hold up.

Fig. 10 [Images not available. See PDF.]

Fingerprints (left) left on the C-shaped forms of the bronze rods (right) broken from a handle-shaped appendage of the Zeng Hou Yi pan.

Photographs by Yung-ti Li and the author. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of Hubei Museum; copyright © Hubei Museum, all rights reserved.

Rethinking Strahan’s construct and “mold seams”

If we do not concur with Strahan’s (2019) piece-mold stand on the casting of the Met shell, then we have no obligation to dive into the details of her concrete proposal. That said, the thought-provoking scheme she constructs for assembling mold sections for casting four quarters of the lower hemisphere of the Met shell is still noteworthy:

…. The model for each hemisphere was a perfect dome, probably made of clay and produced by a rigid profile jig. After the dome-shaped model was made, only one quarter of it was decorated…. A closer study of the bottom hemisphere (…) reveals that the design is made up of two registers. The design of the openwork upper register is composed of horizontal bands of snakes with a repeating head-tail pattern. The bands are connected to each other on the interior by vertical stepped C-shaped forms. These vertical C-shaped forms are often sloppy…. The lower register is composed of a repeating diagonal, twisted snake-like design. This register is less three-dimensional than the upper register…. Because the design of the openwork upper register is more complicated and both registers were cast together, its process of casting requires more consideration. This is how the upper register could be piece-mold-cast. The exterior horizontal design including the simpler lower register was first created in the positive on a single quarter of the model, then fired.

Then, an impression was taken of the design, creating a negative image of the snakes (mold piece A). Next, an impression of that impression was taken, so the snakes are positive again (mold piece B). On this impression, the snakes are carved away leaving slight depressions in their place. These depressions provide good locations for holes. Then, holes were made through the piece wherever connections were needed. These two pieces were reassembled for ease of handling. The whole assembly was turned over and the C-shaped vertical connections were carved in between the holes, spanning over one horizontal row. This creates two mold pieces: one for the exterior design (A) and one for the interior (B). At this point, three more mold pieces were pulled from the initial model with the horizontal snakes; and three negative molds were made from mold B with the vertical C-shaped connections. Then, the pieces are assembled to form a complete dome with the lower sections included. Triangles are carved out at each mold seam on A and used as pouring gates and vents. The assembled molds were fired, a core was then added, and it was cast creating the openwork dome. Since each quarter was made from the same molds (A and B), each quarter is identical to the others.

(Strahan, 2019, pp. 19–20)

Strahan’s mold-making procedure, curvy and abstruse, can be summarized in the following way: (1) one quarter of a dome with its shape desired for the selected Met hemisphere was adorned as a one-fourth decorated model; (2) this fractional model was applied to produce four pairs of outer mold pieces (series “A”) and inner mold pieces (series “B”) with a potentially hyper-challenging transfer process; (3) by combining pieces from the A and B series, the exterior mold and interior mold were formed, before they were fitted together as an entire assembly (i.e., “complete dome”).19 But Strahan’s suggested construct, in addition to being highly complex, also contains some inaccurate technical descriptions. For example, in the last sentence of Strahan’s cited passage: “Since each quarter was made from the same molds (A and B), each quarter is identical to the others.” At the first glance, readers will likely be puzzled: because working clay molds were scorched by molten metal, even those that had not been smashed could seldom be reused for casting a highly delicate ornamentation; how then could “the same mold” have been used “to produce each quarter” of the final bronze?20 After checking the context, we deduce that not all metal units are claimed to have been cast in succession from the same mold set; what Strahan intends to say may be that all four pairs of mold pieces—which together comprise a single mold assemblage—came from the same decorated fractional model. With her meticulous but far-fetched details, Strahan seems to believe that wet clay or whatever she considers apt must work better than the “too flimsy” wax, and likewise that the molding substance must survive her overdone transfer process and, at the same time, guarantee that all of the three-dimensional interlaces of the Met shell emerge intact (Strahan, 2019, pp. 19–20). The reason is obvious: if any damage occurs in any way to the design, how could Strahan set the seal on the identicalness of the openwork sections? Considering the historical and three-dimensional concerns that I have laid out, the use of the fractional-model in replicating identical units—steps 1 and 2 of Strahan’s procedure, in my summary—are extremely unlikely.

My own hands-on examination in the Met warehouse offers another clue. I ascertained that the Met shell layers at intervals do not touch each other, generating physical gaps between them (Figs. 7 and 11). These interspaces enclosed within the original model would not have permitted the removal of the mold in undamaged pieces, however extraordinary the artist’s skill might have been. Bringing about a fully undercutting structure, such trapped spaces have been termed “absolute undercuts,” an often-characteristic sign of investment casting (Peng, 2020, pp. 64 and 67). When describing the transfer procedure in her scheme cited above, however, Strahan seems to be indifferent to any difficulty in withdrawing the “mold piece A” from the “model,” and thereafter the “mold piece B” from the “mold piece A.” Neither the deadlocking undercuts nor any three-dimensional interferences are mentioned by Strahan in her transfers of interlaced ornamentation from “the positive” to the “negative,” and then to the “positive again.” Indeed, when it comes to piece molds, all problems seem to disappear in Strahan’s eyes, and taking an “impression” repetitively “of the design” seems to pose no difficulty. In contrast, “any wax” (“or resin”) could not even be in a less daunting manner “removed and placed onto the core” without “distortions” or the like in Strahan’s narrative. However, such distortions are equally or more likely to have occurred to the three-dimensional clay interlaces in Strahan’s multi-transfer plan (Strahan, 2019, pp. 19–20). If Strahan really fails to distinguish the feature of absolute undercuts, she cannot be expected to justify how to overcome the challenge they pose to her suggested means of transferring the decoration. To the piece-mold advocates who are now aware of that problem, a response is necessary: it should not have been too hard, of course, to apply the muddy or slippery clay into the openwork rooms of the model locked by absolute undercuts. Yet when the clay turned hard, how could the artist have detached the mold in decent pieces from those trapped spaces while keeping every detail undisturbed?

Fig. 11 [Images not available. See PDF.]

Details of the exterior of the upper hemisphere of the Met shell, revealing absolute undercuts of the fully three-dimensional openwork.

Author’s photographs. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of The Metropolitan Museum of Art; copyright © The Metropolitan Museum of Art, all rights reserved.

Fig. 12 [Images not available. See PDF.]

Diatrete incense burner reported in 1996 around the Yaojiagang site, Shaanxi Fengxiang (left), and the two-dimensional openwork inside its square base (right).

Warring States period (475–221 BCE). Height 35.3 cm. Left image after Tokyo National Museum, 2004, p. 61 (cited in Strahan, 2019, p. 17, Fig. 13); right images offered by Su Rongyu. This figure is not covered by the Creative Commons Attribution 4.0 International License. Reproduced with permission of Fengxiang Museum; copyright © Fengxiang Museum, all rights reserved.

The astounding fabrication of the Met shell urges us to be cautious with the challenges that Strahan downplays. Strahan tells us that the interlaced design was first engraved on the model (positive) and impressed to mold piece A (negative), before “an impression of that impression was taken” as mold piece B (positive). However, she does not specify what portion of the entire ornamentation was executed on the original model or how many decorative elements were added in later processes. If we follow the logic of Strahan’s construct, further embellishment should have been strictly prohibited on the transferred “impressions,” because any such individualization on the mold pieces would ruin the supposed identicalness of bronze units. But if the artist were finishing every detail—of the “horizontal bands of snakes with a repeating head-tail pattern” and the “repeating diagonal, twisted snake-like design”—on the model, what could be done about the scattered absolute undercuts? This attempt would be impossible for every pragmatist aiming at model-mold transfers. This dilemma is ultimately inescapable, and it directly threatens or even contradicts Strahan’s overall scheme. While she does not explicitly talk about practical feasibility, Strahan is far from incautious when it comes to piece molds; to me, her suggestion about the post-transfer stage implies certain technical anxieties or fears. That modus operandi, attaining a certain three-dimensionality of the twisting interlaces in negative by excavating “holes” into the molds and connecting them, is not completely new: except for Strahan’s attempted rationalization of C-shaped traces as “connections” “carved in between the holes,” the other implausible ideas (e.g., hole-digging and interlinking in the mold) were proposed earlier by Wang Jinchao and thoroughly rebutted.21

Compared with Wang, Strahan is less detailed only about this specific implement, and therefore her proposition is less “falsifiable,” as Karl Popper would say. But looking at the lamination of staggered C-shaped slices (Fig. 7), no reader should agree that they were mold-cut into shape, just as no one would believe that the vividly intertwined ophidian bodies were dug out eventually. Could we reasonably expect that an exquisite weave-like artwork was made partially by digging and drilling tools, with its three-dimensionality accomplished blindly in negative? To my mind, another critical problem for Strahan is the failure to raise any motive for her unnecessarily convoluted method to have been used. In other words, why would the Met shell casters have formulated such a superfluous and inconvenient mold-making approach? Strahan has a motivation for devising the method: she is making every effort to reproduce an artifact she has studied. But what was the motivation of the Met shell casters? Were they told by their patron: “Show me something completely fanciful—not only the result, but also the whole performance?” Such a scenario is unlikely, and we are left with the feeling that Chinese founders would have gone to any length to steer clear of lost wax. Some insist that this was the case, in a last-ditch attempt to save early China from the charge that it acquired a potentially nonnative idea (i.e., lost wax).22 But Strahan (2019), who is less prejudiced, prioritizes another observation as her reason to prove the piece-mold conception:

[The lower register of the lower hemisphere of the Metropolitan incense burner] was probably formed of a front and back mold with core extensions through the design to hold the sections in place. Vertical mold seams in some places can be seen where the front and back mold sections came together (…). A few holes in the openwork are partly blocked by metal fins, where metal seeped out between the mold seams during casting. The disk-shaped bottom plate was also part of the openwork design using core extensions in its five holes to keep the front and back molds in place….

(Strahan, 2019, p. 19)

Strahan’s discovery here, which raises her confidence in piece molds, is the feature she calls the “mold seams” of the metal shell (Strahan, 2019, pp. 18 and 20, figs. 18 and 20). This phrase, problematic in this context, was popular among early writers who did not distinguish between mold-join gaps (“mold seams”) and the corresponding piece-mold cast marks on the final metalwork. In a 1929 catalog, for instance, W. Perceval Yetts clearly associated what he labeled mold “seams” with the indirect lost-wax process (Yetts, 1929, pp. 35–38). In 1935, Yetts restated his opinion by arguing that sectional molds from Anyang were largely used in “shaping the wax models,” and those with burn marks were dismissed as investment mold fragments (Yetts, 1935, p. 473). Yetts, who was influential in academia, was rebutted by Rutherford John Gettens: “He [Yetts] had observed the ‘seams’ on ancient bronzes, which are now recognized as mold joins.” Yetts’s mold “seams” were therefore renamed “mold-join marks”—or “mold marks” for short—in Gettens’s milestone study of The Freer Chinese Bronzes (Gettens, 1969, pp. 23, 60–67). Over half a century later, Yetts’s term is somehow still in use, although his opinion has long been abandoned.

Although the term “mold seams” is remembered and favored by the “anti-lost-wax” camp today,23 Yetts’s indirect lost-wax proposal seems to have been entirely forgotten. That far-flung suggestion, which has already been invalidated as an explanation for most pre-Han bronzes, remains a theoretical possibility for special cases. In other words, lost wax is not theoretically contradicted by the appearance of “mold marks” (i.e., “mold seams,” as they are called by Yetts and Strahan), and one of the reasons is that its indirect process (Fig. 3) takes technical advantage of piece molds. As Bagley notes, “The direct lost-wax method (…) is today seldom used because it does not allow duplicate castings and because of the risk involved in destroying the original model. The indirect lost-wax method, employed in Greek foundries as early as the seventh century B.C., preserves the original model intact by forming a section mold on it and then using the section mold to produce a wax model or models” (Bagley, 1990, p. 20, note 2).24 Considering the intensified Eurasian integration during the first millennium BCE, it would not be outside the realm of possibility to find indirect lost wax relevant to the Mediterranean tradition present in China earlier than the currently known Imperial Qin testimonies (221–206 BCE).25

Perhaps more to the point, the boundary between the direct and the indirect lost-wax processes is not always clear-cut. Consider the masses of interpenetrating patterns appended to the rim of Zeng Hou Yi zun (Fig. 5), for example: might these components, which are similar in appearance, originally have been wax pieces mass-shaped within some permanent “mother” molds? The committed lost-wax objectors do not think so, and they often cite “mold seams” as an argument to bolster various piece-mold schemes that involve no wax use.26 By contrast, certain lost-wax advocates suggest that the wax elements may have been fashioned in series from moldboard and assembled following a preset design.27 The latter suggestion is reminiscent of various indirect lost-wax fashions in “traditional” China in the mass production and replication of ornamental wax for the easier construction of wax models (Tan 1989, pp. 16–18; Wang et al., 2014, pp. 115–16). When linking these ethnographical cases of no clear origins to the millennia-apart Met shell, the cogency of such an analogy can be questioned;28 yet in a Chinese context these analogous examples at least empirically extenuate the mold-mark issue to some degree. Besides ethnography, replication experiments also favor the use of wax along with piece or bivalve (two-piece) molds over the mold-drilling formation for some controversial castings. With Wang Jinchao’s essentially unsuccessful reproductions of the Zeng Hou Yi zun ornamentation and a stylistically similar zhan handle in mind (Peng, 2020, pp. 106–108, 302–30329), I am curious about whether any better replication of the Met shell can be achieved by following Strahan’s or Wang’s atypical piece-mold scheme.30 Given these ethnographic and experimental clues as well as suggestive information from its allied castings, we can posit that no necessity of mold marks lowers the likelihood of a lost-wax scenario for the Met shell.

Indeed, we need to demolish the idea that “the lost-wax method” is a single simple thing. The problem is the word “the”: the term “the lost-wax method” lumps together a variety of different operations. From investigating modern foundries, we understand that what goes on there is complicated and variable. Every new kind of object is a new problem for those casters, and they adapt their processes in whatever way is necessary to solve the problem. In other words, the foundrymen are not guided by textbook abstractions or diagrams; they do what they need to do to cast the desired artifact. Some sophisticated casters freely combine the direct and the indirect lost-wax processes along with piece molds.31 Only when we discard the layman’s oversimplification of lost wax can we understand that the mold marks on a piece’s castings—such as those on the Met shell, though Strahan disagrees—are nothing special. Like the interpenetrating ornamentation of the Zeng zun-pan or Xiasi jin, the “mold seams” Strahan has cited as proof of piece-mold casting could just as feasibly have been generated when constructing the wax model or shaping its components (Peng, 2020, p. 42).32

Final Remarks

Strahan’s assiduous study attempts to give both a detailed description and a convincing explanation of the Met shell. The first goal is admirably reached, but not the second. Guided by her piece-mold construct, we find intriguing details that we have little chance to make sense of. In particular, the three-dimensional openwork of the Met shell, involving absolute undercuts and intricately entwined rods, can hardly have been achieved by piece molds alone. We can also ask why, if the casters had depended entirely on piece molds, they would have created a curious three-dimensional interiority (invisibility) of the diatrete shell that is technically difficult to achieve but visually unnecessary and functionally useless. If only the piece-mold process had been used, the interior of the Met shell would be conveniently flattened, like the two-dimensional surface shown in Fig. 12. This surface is inside the openwork base of an incense burner from Shaanxi Fengxiang, another Eastern Zhou diatrete vessel with a three-dimensional sphere that manifests the technical style of lost wax. The contrasting features of openwork in a single set—an excellent synthesis of two casting techniques—most tellingly distinguishes the lost-wax interlaces that we also see on the Met shell from a purely piece-mold design.33

For a long time, research on the pre-Han use of lost wax in China has been driven primarily by the discovery that, in a tradition with massive metalwork casting, lost wax was rarely used. For some scholars, it was curious that the Chinese cast the highly sophisticated bronzes invariably by the cumbersome method of “direct casting into piece-moulds,” as they had always heard that “the ‘finest castings’ were achieved by means of the cire-perdue process” (e.g., Barnard, 1961, pp. 108–109). That position was severely shaken when the Zeng Hou Yi zun-pan (Fig. 5) was unearthed at Hubei Suizhou in 1978, along with subsequently excavated or rediscovered finds carrying similarly interpenetrating openwork. In a 1979 conference, lost-wax use was formally acknowledged by the China Foundry Association in casting the rim appendages of the Zeng zun-pan. Because of that “consensus” reached by dozens of experts and accepted by the official report of the Zeng Hou Yi Tomb (see Hubei, 1989, pp. 229–231; Hua, 2010, pp. 89–90), in the following two decades or so it was nearly impossible to forget lost wax when looking at the zun-pan. Since the early 2000s, however, a number of scholars have once again turned against the hypothesis that lost wax was used in the Chinese Bronze Age (e.g., Wang, 2002; Zhou et al., 2006; Zhou et al., 2007; Dong et al., 2008; Wang, 2008; Wang, 2009; Zhou, 2009; Zhou et al., 2009; Zhou et al., 2009; Wang et al., 2010; Wang and Wang, 2014; Zhou and Huang, 2015a, b; Zhou and Huang 2016; Huang and Zhou, 2018). After surging first in China, this new tide of doubt, despite a resisting headwind (e.g., Zhao, 2006; Zhang, 2007; Huang, 2008; Tan, 2008; Hua, 2010), spilled over from the academic world to the media and the public.34 It took some time to arrive in the West, and Donna Strahan is among several open-minded scholars who are inviting this vigorous wave.35 The claims against early Chinese lost-wax use have been now, to me, falsified by the diatrete Met shell with sufficient clarity. With this essay, I hope “researchers can finally put aside previous disputes and move forward to more culturally significant questions” (Liu, 2022, p. 179). In the rare confirmed cases of early Chinese lost-wax utilization, for example, how exactly was this process used, and why? Behind the knowledge of lost wax ultimately stands the broader entanglement of technique and art—two manifestations of the same creativity—that cry for both a Weberian disenchantment and a Foucauldian archaeology in our scholarly discourses.

Acknowledgements

This research was sponsored by the Research Grants Council of the Hong Kong Special Administrative Region, China (RGC Early Career Scheme, Project No. 24609620) and the CUHK Direct Grants for Research (Project Nos. 4051178 and 4051217). I would like to express my gratitude to SIT Shui Yan for her remarkable research assistance over the years, and to Allison Van Deventer and the two anonymous reviewers for their insightful comments on the draft of my essay. I am deeply grateful to the curators in the Metropolitan Museum of Art for their warm hospitality. The Met warehouse is a paradise for art historians conducting object studies.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Competing interests

The author declares no competing interests.

Ethical approval

This article does not contain any studies with human participants performed by any of the authors.

Informed consent

This article does not contain any studies with human participants performed by any of the authors.

1

For definitions and comparisons of these casting processes, see Peng, 2020, pp. 5–22.

2

Published as Strahan (2019) and translated as Xu (2021)^.

3

For details, see Peng, 2020, pp. 143–176.

4

A notable sample as such refers to dozens of bronze waterfowl in the third century BCE. For details, see Shao, 2019^.

5

For more details, see Bagley, 1987, pp. 16–18.

6

In the present paper, following Li et al., “identical” is used rigorously with respect to the “size, shape, and decorative patterns” of metalwork; whatever is called “identical” is supposed to have “originated from the same decorated model.” See Li et al., 2021, p. 110.

7

See notes 174 and 188.

8

See figs. 1A and 10.

9

For details, see Strahan, 2019, pp. 19–20.

10

This is a critical description by the late Cyril Smith. See Smith, 1981, p. 325.

11

This object study (October 31, 2016) was supported by Zhixin Jason Sun and Hwai-ling Yeh-Lewis.

12

For example, see Zhang, 2002, pp. 66–67; Zhou et al., 2007, pp. 45–46.

13

For one such example, see Peng, 2018b, p. 345, Fig. 9.33. On processing wax, sometimes mixed with local ingredients (e.g., resin, tallow), see Hua, 2010, pp. 87–88, 94–95. For more details, see Cowan, 1908^; Warth, 1947^.

14

For further discussions, see Zhou and Huang, 2015a, 1635; Zhou and Huang, 2015b, 154; Peng, 2020, pp. 113–116.

15

The purpose of this design—more likely to be functional than ornamental if not simply a byproduct—is not entirely clear. For possibilities to confirm, see Peng, 2020, p. 116.

16

For its illustration, “Detail of Nanjing base interior, showing stepped form similar to that in figs. 20 [i.e., interior of the Met lower hemisphere], 22, and 23 [i.e., Sackler finial interior],” see Strahan, 2019^, p. 23, Fig. 25.

17

For discussions, see Wan, 2013, pp. 72–73, and 75, Fig. 6; Strahan, 2019, p. 22; Peng, 2020, pp. 78–79, 136.

18

For the jin restorers’ first-hand observations, see Ren and Wang, 1987^.

19

Cf. Peng, 2020, p. 68.

20

This expression appears more than once in Strahan’s paper. See Strahan, 2019, p. 20.

21

Cf. Strahan, 2019, pp. 19–20, and various papers by Wang Jinchao (e.g., Wang, 2002; 2008; 2009). For proof against Wang, see Peng, 2018a, pp. 105–108.

22

For example, see Zhou et al., 2006, pp. 80–85; Dong et al., 2008, pp. 97–107; Wang, 2009, pp. 278–87; Zhou and Huang, 2015a, 1629–36 (cf. Peng, 2020, 47, note 20). Robert Bagley, personal communications, August 2014, December 2017, and July 2018.

23

The Chinese version of “mold seams” is fanfeng. For the use of it by those against lost wax in early China, see Zhou et al., 2006^; Zhou et al., 2007^; Dong et al., 2008^.

24

For more details on indirect lost-wax casting and its Greek use, see Peng, 2020, pp. 9–10.

25

Bronze waterfowl from the mausoleum of the First Emperor of Qin (mausoleum constructed 247–208 BCE) exactly manifest Hellenistic-style practices of lost wax (e.g., chaplets and armatures in use) with some evidence—impressions of brushes presumably for coating wax, for instance—pointing to its indirect fashioning. See Peng, 2020, pp. 167–170, and 391–394, figs. 105ab, 106ab.

26

For example, see Zhou et al., 2006^; Zhou et al., 2007^; Dong et al., 2008^; Zhou and Huang, 2015a. One of the leading lost-wax objectors laudably reflected on the improper “use the presence or absence of mold seams as a criterion” for the lost-wax judgment in 2009; six years later, however, he came back to his initial insistence: “….these casting seam lines are the principal characteristic for the identification of piece-mold casting….Mold seam lines and parting faces are the essential characteristics that distinguish piece-mold casting from lost-wax casting.” Cf. Zhou, 2009, p. 126; Zhou and Huang, 2015a, p. 1632.

27

For example, see Hua, 1999, p. 173; Tan, 2008, pp. 118–124; Hua, 2010, p. 94.

28

Though labeled as “traditional,” could any traits of these methods be tracked back to the pre-Han “tradition(s)”? For ^a critique, see Zhou, 2009, p. 44.

29

Compare figs. 63a and 63b.

30

Though Wang Jinchao is not cited when describing her approach, elsewhere Strahan notes Wang’s contribution in a sentence: “Wang Jinchao made reproductions of Zenghouyi [i.e., Zeng Hou Yi] parts using piece-molds to demonstrate how the vessel could have been created (…), and he points out how regular in size and shape the pieces are (Wang, 2002).” See Strahan 2019, pp. 12, and 13, Fig. 10. Wang’s unconvincing replication can be immediately sensed by its comparison with the lost-wax reproduction (cf. Huang, 2008, pp. 115–17, figs. 11–15).

31

In antiquity, such lost-wax complications are best demonstrated by the Classical or Hellenistic bronzes, such as the renowned Riace warriors of the fifth century BCE. Robert Bagley, personal communications, March and August 2014. See also Peng, 2020, pp. 7, 9–10.

32

To add a layer of complexity: some uncritically accepted “mold marks” (e.g., the irregularly defined ridges on the well-known Xiasi jin table, the first half of sixth century BCE) may not be transferred from the mold seams directly in metal or initially in wax. In other words, they may be neither the piece-mold cast marks nor mold seams impressed to the wax model, but the wax-fusing marks instead. See Li, 1994, p. 41, and 44; Peng, 2020, pp. 52–54, and 92, note 24.

33

For details of the argumentation, see Peng, 2020, pp. 71–74.

34

For example, see CCTV-9 2011, National Treasure 1.3, “Zeng Hou Yi zun-pan” (part 1, released on January 3). CNTV video, https://tv.cctv.com/2011/09/08/VIDE1355587710362454.shtml (accessed July 23, 2022); CCTV-9 2011, National Treasure 1.3, “Zeng Hou Yi zun-pan” (part 2, released on January 4). CNTV video, https://tv.cctv.com/2011/09/09/VIDE1355587262494255.shtml (accessed July 23, 2022).

35

For a more prudent stance in the West, see Notis and Wang, 2017^.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

Bagley R (1987) Shang ritual bronzes in the Arthur M. Sackler Collections. Arthur M. Sackler Foundation, Washington D.C

Bagley, R. Shang ritual bronzes: casting technique and vessel design. Arch Asian Art; 1990; 43, pp. 7-20.

Bagley, R. Anyang mold-making and the decorated model. Artibus Asiae; 2009; 69, 1 pp. 39-90.2009AIPC.1103..39B

Barnard, N. Bronze casting and bronze alloys in ancient China; 1961; Canberra, Australian National University:

Chase WT (1991) Ancient Chinese bronze art: casting the precious sacral vessel. China Institute in America, New York

Collins, W. The corrosion of early Chinese bronzes. J Inst Met; 1931; 45, pp. 23-55.

Cowan T (1908) Wax craft, all about beeswax: its history, production, adulteration, and commercial value. S. Low. Marston & Co. Ltd, London

Craddock, P. The metal casting traditions of South Asia: continuity and innovation. Indian J Hist Sci; 2015; 50, 1 pp. 55-82. [DOI: https://dx.doi.org/10.16943/ijhs/2015/v50i1/48112]

Davidson, J. Toward a grouping of early Chinese bronzes. Parnassus; 1937; 9, 29–34 51.

Dong Y (2006) Fanzhu qingtong 範鑄青銅. Beijing Yishu yu Kexue Dianzi Chubanshe, Beijing

Dong, Y et al. Zaitan Zeng Hou Yi zun-pan de zhuzao gongyi 再探曾侯乙尊盤的鑄造工藝 (Re-exploration of the casting process of Zeng Hou Yi zun-pan). Zhongyuan Wenwu 中原文物; 2008; 1, pp. 97-107.

Fong, W. The Great Bronze Age of China: an exhibition from the People’s Republic of China; 1980; New York, Metropolitan Museum of Art:

Gettens RJ (1969) The Freer Chinese bronzes. vol II: Technical studies. Smithsonian Institution, Washington

Hubei S (1989) Zeng Hou Yi mu 曾侯乙墓 (The tomb of Marquis Yi of Zeng). Wenwu Chubanshe, Beijing

Hua J (1999). Zhongguo gudai jinshu Jishu: Tong he tie zaojiu de wenming中國古代金屬技術: 銅和鐵造就的文明 (Ancient Chinese metal technology: A civilization made of copper and iron). Daxiang Chubanshe, Zhengzhou

Hua, J. Zhongxifang shilafa zhi tongyi: Jianping ‘xian Qin bu cunzai shilafa’ yi shuo中西方失蠟法之同異—兼評先秦不存在失蠟法一說 (The similarities and differences of Chinese and Western lost-wax methods—with additional comments on how the lost-wax method did not exist in pre-Qin China). Kaogu考古; 2010; 4, pp. 87-96.

Huang, F et al. Hubei Suizhou shi Wenfengta dongzhou mudi 湖北隨州市文峰塔東周墓地 (The Eastern Zhou cemetery of Wenfengta in the Suizhou city of Hubei). Kaogu 考古; 2014; 7, pp. 18-33.

Huang, J. Zeng Hou Yi zun-pan caiyong shilafa gongyi zhuzao wuyong zhiyi: yu ‘Zhongguo qingtong shidai bu cunzai shilafa zhuzao gongyi taolun 曾侯乙尊盤采用失蠟法工藝鑄造毋庸置疑—與中國青銅時代不存在失蠟法鑄造工藝討論 (There is no doubt that the Zeng Hou Yi zun-pan was cast by the lost-wax method—discussion of how the lost-wax casting process did not exist in the Chinese Bronze Age). Jianghan Kaogu 江漢考古; 2008; 4, pp. 109-117.

Huang, W; Zhou, W. Suizhou Wenfengta Dongzhou qingtong pan loukong er de zhuzao gongyi 隨州文峰塔東周青銅盤鏤空耳的鑄造工藝 (On the craft in casting the openwork handles of the Eastern Zhou bronze pan from Suizhou Wenfengta). Gugong Bowuyuan Yuankan 故宮博物院院刊; 2018; 1, pp. 120-129.

Karlbeck, O. Anyang moulds. Bulletin of the Museum of Far Eastern Antiquities; 1935; 7, pp. 39-60.

Li, J. Xichuan Chunqiu Chu mu tongjin shila zhuzaofa de gongyi tantao 淅川春秋楚墓銅禁失蠟鑄造法的工藝探討 (Discussion on the lost-wax method of casting the bronze jin in the Xichuan tomb of Chu during the Spring and Autumn period). Wenwu Baohu yu Kaogu Kexue 文物保護與考古科學; 1994; 6, 1 pp. 39-45.

Li, K et al. Decorated models, replication, and assembly lines for bronze industrial production in 500 B.C.E China. Early China; 2021; 44, pp. 109-142. [DOI: https://dx.doi.org/10.1017/eac.2021.9]

Liu, S. Metalworking in Bronze Age China: the lost-wax process by Peng Peng (review). Asian Perspect; 2022; 61, 1 pp. 177-179. [DOI: https://dx.doi.org/10.1353/asi.2022.0009]

Loehr, M. The bronze styles of the Anyang period (1300–1028 B.C.). Arch Chinese Art Soc Am; 1953; 7, pp. 42-53.

Loveday, H. Diversity in Eastern Zhou bronze casting: a look at a group of openwork vessels. J East Asian Archaeol; 2002; 4, 1 pp. 101-142. [DOI: https://dx.doi.org/10.1163/156852302322454512]

Martinón-Torres, M; Uribe-Villegas, MA. Technology and culture in the invention of lost-wax casting in South America: an archaeometric and ethnoarchaeological perspective. Camb Archaeol J; 2015; 25, 1 pp. 377-390. [DOI: https://dx.doi.org/10.1017/S0959774314001164]

Martinón-Torres, M; Uribe-Villegas, MA. The prehistoric individual, connoisseurship and archaeological science: the Muisca goldwork of Colombia. J Archaeol Sci; 2015; 63, pp. 136-155. [DOI: https://dx.doi.org/10.1016/j.jas.2015.08.014]

Nickel, L. Imperfect symmetry: re-thinking bronze casting technology in ancient China. Artibus Asiae; 2006; 66, 1 pp. 5-39.

Notis, M; Wang, D. Ancient Chinese bronze casting methods: the dilemma of choice. MRS Adv; 2017; 2, 33-34 pp. 1743-1768.[COI: 1:CAS:528:DC%2BC2sXhtFygsrfJ] [DOI: https://dx.doi.org/10.1557/adv.2017.298]

Peng, P. Was lost-wax casting adopted in pre-Qin Chinese bronze art?. Orientations; 2018; 49, 1 pp. 103-111.

Peng P (2018b) The lost-wax casting in Bronze Age China: art, technology, and social agency. Ph.D. dissertation, Princeton University

Peng P (2020) Metalworking in Bronze Age China: the lost-wax process. Cambria Press, Amherst

Ren, C; Wang, C. Henan Xichuan Xiasi Chunqiu yunwen tong jin de zhuzao yu xiufu 河南淅川下寺春秋雲紋銅禁的鑄造與修復 (Casting and restoration of the cloud-scroll-patterned bronze jin in the Spring and Autumn period from Henan Xichuan Xiasi). Kaogu 考古; 1987; 5, pp. 474-478.

Shi, Z. Yin dai de zhutong gongyi 殷代的鑄銅工藝 (The bronze casting process of the Yin Dynasty). Zhongyang Yanjiuyuan Lishi Yuyan Yanjiusuo Jikan 中央研究院歷史語言研究所集刊; 1955; 26, pp. 95-129.

Smith, C. A search for structure: selected essays on science, art, and history; 1981; Cambridge, MIT Press:

Strahan D (2019) Debating the use of lost-wax casting in ancient China. Forbes Lecture on Scientific Research in the Field of Asian Art. Lecture given on January 15, 2015 at the Freer Gallery of Art. Smithsonian, Washington DC

Shao A (2019) Qin Shihuangdi lingyuan chutu caihui qingtong shuiqin zhizuo gongyi ji xiangguan wenti yanjiu 秦始皇帝陵園出土彩繪青銅水禽製作工藝及相關問題研究 (Research on the production process and related problems of painted bronze waterfowl unearthed in the cemetery of the first Qin emperor). Kexue Chubanshe, Beijing

Tan D (2008) Qianyi gudai shila zhujian de bianren—Jianping Zhou Weirong deng zhi wupan shuo 淺議古代失蠟鑄件的辨認—兼評周衛榮等之誤判說. In: Qingtong Wenhua Yanjiu 青銅文化研究, vol 5. Huangshan Shushe, Tongling, pp 118–124

Tan D (1989). Canlan de zhongguo gudai shila zhuzao燦爛的中國古代失蠟鑄造(Splendid ancient Chinese lost wax casting). Shanghai Kexue Jishu Wenxian Chubanshe, Shanghai

Tokyo N (2004) Treasures of ancient China. Catalogue for the eponymous special exhibition in the Tokyo National Museum, September 28 to November 28, 2004. Asahi Shinbunsha, Tokyo

Wan C (1976) A comparative study of the casting of bronze Ting-cauldrons from Anyang and Hui-hsien. In: Ancient Chinese bronzes and Southeast Asian metal and other archaeological artifacts. National Gallery of Victoria, Melbourne, pp 17–46

Wan, L. Nanjing bowuyuan cang Zhanguo toukong panlongwen qingtongqi 南京博物院藏戰國透空蟠龍紋青銅器 (The bronze ware of the Warring States period with three-dimensional dragon-interlaced openwork in the Nanjing Museum). Dongnan Wenhua 東南文化; 2013; 6, pp. 72-76.

Wang, J. Tan Zeng Hou Yi zun-pan de zhuzao gongyi 談曾侯乙尊盤的鑄造工藝 (On the casting technology of Zeng Hou Yi zun-pan). Dongnan Wenhua 東南文化; 2002; 1, pp. 70-74.2002stma.conf...W

Wang, J. Nifanfa fuyuan Xu gong Ning loukong panhuiwen qingtong shijian yanjiu 泥範法復原許公寧鏤空蟠虺紋青銅飾件研究 (Study on the clay-[piece-]mold reproduction of the Duke Ning of Xu’s openwork ornament with interlaced hui-serpentine patterns). Nanfang Wenwu 南方文物; 2008; 4, pp. 34-36.

Wang J (2009) Zhongguo qingtong shidai loukong qingtongqi nifan zhuzuo gongyi yanjiu 中國青銅時代鏤空青銅器泥範鑄作工藝研究 (Research on the clay-[piece-]mold craft of casting openwork bronzes in Bronze Age China). In: Qingtong Wenhua Yanjiu 青銅文化研究, vol 6. Huangshan Shushe, Tongling, pp 278–287

Wang, J; Wang, W. Shiyan kaogu─Zhongguo qingtong shidai toukong qingtongqi nifan zhuzao gongyi qiushi 實驗考古─中國青銅時代透空青銅器泥範鑄造工藝求實 (Experimental archeology: Seeking the truth of clay-mold bronze casting craft in Chinese Bronze Age). Ku Chin Lun Heng 古今論衡; 2014; 26, pp. 4-34.[COI: 1:CAS:528:DC%2BC2MXpvFOlsg%3D%3D]

Wang, W et al. Sanwei donghua li de gudai chuantong gongyi 三維動畫裏的古代傳統工藝 (Ancient traditional crafts in 3D animation). Jianghan Kaogu 江漢考古; 2010; 3, pp. 113-122.

Warth, A. The chemistry and technology of waxes; 1947; New York, Reinhold:

Xu D (2021) Zhongguo gudai qingtongqi shilafa zhuzao zhi bianlun 中國古代青銅器失蠟法鑄造之辯論 (Debating the use of lost-wax casting in ancient China). Dongfang Kaogu 東方考古 18(0):224–245

Yetts W (1929) The George Eumorfopoulos Collection: catalogue of the Chinese & Corean bronzes, sculpture, jades, jewellery and miscellaneous objects, vol 1. E. Benn, London

Yetts, W. Recent finds near An-yang. J R Asiatic Soc; 1935; 67, 3 pp. 467-474. [DOI: https://dx.doi.org/10.1017/S0035869X00087165]

Zhang, G. Zhongguo zuizao ‘shilafa’: Chunqiu zhongqi ‘shewang gaiguan longhu fang hu’ de zhufa lunzheng 中國最早‘失镴法’:春秋中期‘蛇網蓋冠龍虎方壺’的鑄法論證 (The earliest ‘lost-pewter method’ in China: demonstration of the casting method of the ‘serpentine-latticed crown for the dragon-and-tiger square hu cover’ in the middle Spring and Autumn period). Dongnan Wenhua 東南文化; 2002; 1, pp. 63-69.

Zhang, C. Guanyu Zeng Hou Yi zun-pan shifou caiyong shilafa zhuzao de shuping 關於曾侯乙尊盤是否采用失蠟法鑄造的述評 (Commentary on whether the Zeng Hou Yi zun-pan was cast with the lost-wax method). Jianghan Kaogu 江漢考古; 2007; 4, pp. 85-90.

Zhao, S. Chunqiu shiqi shilafa zhuzao gongyi wenti tantao 春秋時期失蠟法鑄造工藝問題探討 (Discussion on the lost-wax casting craft in the Spring and Autumn period). Zhongyuan Wenwu 中原文物; 2006; 6, pp. 84-91.

Zhongguo Q (1996) Zhongguo qingtongqi quanji1 中國青銅器全集 (Complete compilation of Chinese bronzes), vol 1. Wenwu Chubanshe, Beijing

Zhou, W. Shila gongyi de qiyuan yu shila zhuzao de gongyi tezheng—Jian tan shila gongyi wenti yanjiu de jinzhan yu yiyi 失蠟工藝的起源與失蠟鑄造的工藝特徵—兼談失蠟工藝問題研究的進展與意義 (The origin of lost-wax technology and the process characteristics of lost-wax casting—also on the research progress and significance of lost-wax craft). Nanfang Wenwu 南方文物; 2009; 4, pp. 39-45.

Zhou, W; Huang, W. Lost-wax casting in ancient China: new discussion on old debates. J Miner Met Mater; 2015; 7, pp. 1629-1636. [DOI: https://dx.doi.org/10.1007/s11837-015-1462-0]

Zhou, W; Huang, W. Shilun qingtong shidai toukong qingtongqi de gongyi tese—Jian tan shila zhuzao wenti 試論青銅時代透空青銅器的工藝特色—兼談失蠟鑄造問題 (On the technical characteristics of the three-dimensional openwork bronzes in the Bronze Age—also on the problem of lost-wax casting). Zhongguo Guojia Bowuguan Guankan 中國國家博物館館刊; 2015; 1, pp. 147-155.

Zhou W, Huang W (2016). "Shila shizhi fa" shangque "失蠟失織法"商榷(Discussion on the method of “lost-wax and lost textile”). Nanfang Wenwu南方文物 1: 59–63

Zhou, W et al. Zhongguo qingtong shidai bu cunzai shilafa zhuzao gongyi 中國青銅時代不存在失蠟鑄造工藝 (Lost-wax casting process did not exist in Bronze Age China). Jianghan Kaogu 江漢考古; 2006; 2, pp. 80-85.

Zhou, W et al. Zai lun ‘shila gongyi bushi Zhongguo qingtong shidai de xuanze’ 再論失蠟工藝不是中國青銅時代的選擇 (Re-discussion of how lost-wax craft was not the choice of the Chinese Bronze Age). Nanfang Wenwu 南方文物; 2007; 2, pp. 41-48.

Zhou W et al. (2009) New research on lost-wax casting in ancient China. In: Mei J, Rehren T (eds). Metallurgy and civilisation: Eurasia and beyond. Archetype, London, pp 73–78