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

The paper addresses whether complex Eastern Zhou bronze openwork, especially diatrete structures, could have been produced solely with piece-mold casting or whether lost-wax casting was employed before the Han period. It focuses on a globular diatrete tripod at the Metropolitan Museum of Art, previously discussed by Donna Strahan, and interrogates assumptions linking certain linear or banded decorations exclusively to piece-mold methods. The author frames the debate within broader questions of technique, design constraints, and cultural implications for understanding early Chinese metallurgical innovation, arguing that the Met object provides a critical test case to reassess the role of lost-wax casting in pre-Han China.

The article situates the debate in a long historiography: early scholarship often denied pre-Han lost-wax use, but the 1978 discovery of the Zeng Hou Yi zun-pan revived acceptance of lost-wax techniques for interpenetrating openwork. Since the early 2000s, a renewed skepticism emerged in China (e.g., Wang, Zhou, Dong), later influencing Western discussions. The author critiques Strahan’s (2019) interpretation of the Met diatrete as piece-mold-based, noting her reliance on parallels to much earlier Erlitou linear decorations discussed by Bagley (1987, 1990), which the author argues are inapplicable to Eastern Zhou interlaced three-dimensional openwork. Earlier observations on "linear" or thread-like reliefs (Davidson, 1937; Loehr, 1953) are reviewed, as is Bagley’s critique of lost wax for early surface ornament. The paper also engages debates on identicalness and replication: Nickel (2006) and Dong (2006) proposed undecorated or fractional models and mold-applied decoration, views the author and others (Bagley, 2009) find implausible given the complexity and variability of Bronze Age bronzes. The terminology and evidentiary value of "mold seams" are traced from Yetts (1929, 1935) to Gettens (1969), clarifying that such marks can also arise in indirect lost-wax procedures. Broader technical syntheses are referenced (Notis and Wang, 2017) alongside ethnographic and experimental studies, positioning the Met diatrete within a nuanced technological landscape where both piece molds and lost wax could be combined.

• Object-focused technical analysis: The author conducted hands-on examination of the Metropolitan Museum object, especially its interior surfaces, to identify diagnostic fabrication traces (e.g., absolute undercuts, internal three-dimensional undulations, laminated or stepped C-shaped forms).
• Comparative assessment: The Met diatrete is compared with other Eastern Zhou openwork artifacts (e.g., Zeng Hou Yi zun-pan, Sackler finial, Liulin openwork frame, Fengxiang incense burner base) to distinguish traits consistent with lost-wax vs piece-mold techniques, including differences between fully three-dimensional and two-dimensional openwork treatments.
• Critical evaluation of opposing reconstructions: Strahan’s proposed piece-mold workflow (fractional decorated model; multi-stage impressions A and B; drilling and carving connections between holes) is dissected for technical feasibility, especially given undercuts and the practicality of removing rigid molds from complex interlaced geometries without damage.
• Materials reasoning: The properties of wax (pliability, toughness when properly processed) are discussed against claims of flimsiness, including analogies to pre-Columbian lost-wax smallworks and ethnographic accounts of wax processing/mixes.
• Process taxonomy: The analysis distinguishes between direct and indirect lost-wax casting and considers hybrid techniques, noting that mold marks can originate in wax model construction as well as in metal casting.
• Use of prior experimental and ethnographic work: Published replications and traditional practices are referenced to gauge plausibility of proposed methods and to contextualize observed surface and structural clues.

• The Met diatrete shell exhibits absolute undercuts and internal three-dimensional interlacing that are highly unlikely to result from piece molds alone; such trapped spaces would prevent mold removal without damage.
• Interior surfaces show laminated, stepped C-shaped forms consistent with stacked or fused wax slices (with softened or semi-molten edges), supporting a lost-wax construction of the openwork connectivity.
• Claims of four identical quarters derived from a fractional model are unconvincing; interior variability and three-dimensional undulations contradict strict identicalness, and the proposed multi-transfer clay mold route would be impractically complex and vulnerable to distortion.
• The presence of mold marks does not exclude lost wax; in indirect lost-wax or hybrid processes, marks can be imparted during wax model formation or component assembly.
• Comparative evidence: The Fengxiang incense burner’s base features two-dimensional openwork consistent with piece molds, contrasting with the Met shell’s fully three-dimensional interior, reinforcing the lost-wax attribution for the Met diatrete.
• Broader context suggests that lost-wax use in pre-Han China, though rare, did occur for complex interpenetrating ornamentation (e.g., Zeng Hou Yi zun-pan), and hybrid practices combining piece molds and lost wax should be expected.
• The article refutes specific elements of Strahan’s piece-mold scenario (decorated quarter-model replication; drilled-and-carved C-shaped connectors) as technically implausible given absolute undercuts and removal challenges.

By identifying absolute undercuts and wax-like lamination traces inside the Met diatrete shell, the study directly addresses the core question of whether such complex openwork can be achieved with piece molds alone. The technical evidence favors lost-wax casting (likely with indirect or hybrid elements) and undermines reconstructions dependent on fractional decorated models and multi-stage clay impressions. Recognizing that mold marks can originate in indirect lost-wax procedures dissolves a commonly cited objection to lost-wax attribution. The findings situate the Met piece within a broader Eurasian and Chinese context where artisans flexibly combined methods to solve specific casting problems. This reevaluation has significance for understanding innovation, transmission, and workshop practices in Eastern Zhou metallurgy, encouraging a shift from categorical denial of lost wax toward nuanced, evidence-based attributions.

The paper reinterprets the Metropolitan Museum diatrete shell as a lost-wax (likely indirect or hybrid) casting, based on diagnostic undercuts and wax-lamination traces inconsistent with pure piece-mold production. It critiques alternative piece-mold reconstructions as technically implausible and argues that mold marks are compatible with lost-wax workflows. The study contributes to resolving a long-standing debate on pre-Han lost-wax use and invites scholars to move beyond blanket skepticism to investigate how and why lost wax was selectively employed in Bronze Age China. Future research directions include: detailed reconstructions of workshop practices for diatrete openwork, systematic experimental replications testing hybrid methods, closer comparative studies of interiors vs exteriors across artifact classes, and exploration of technological transmission pathways linking Chinese and wider Eurasian traditions.

• The study relies on qualitative, object-centered visual and tactile examination rather than newly generated analytical datasets.
• Conclusions are drawn from a limited number of directly examined artifacts (notably the Met object) and comparative published exemplars; broader surveys and systematic replications would strengthen generalizability.
• The work critiques alternative reconstructions but does not present a full experimental reproduction of the Met shell by the proposed lost-wax method within this article.