Revealing ancient technology: a high-energy X-ray computed tomography examination of a Mesopotamian copper alloy head

Abstract

Although the origins of lost wax casting extend back into the 5th millennium BCE, it was not until the development of hollow core casting that life-sized metal sculptures could be produced. Based on existing evidence, the earliest adoption of this technique, which involves the inclusion of a clay core within a wax model, occurred in Iraq (Mesopotamia) during the Early Dynastic III period (ca. 2600–2350 BCE). To date, only one hollow core casting from the succeeding Akkadian period (ca. 2350–2150 BCE)—the Sargon Head in the collection of the Iraq Museum—has been studied from a technical point of view. The recent attribution of The Metropolitan Museum of Art's Head of a ruler to this formative period of hollow core lost wax casting provided the impetus for its examination by high-energy X-ray computed tomography—the most practical technique for an object that is continuously on display that could image in 3D the interior morphology of this sculpture given the considerable thickness of its metal walls. This scan revealed a markedly different style of production than the Sargon Head. Although further research on early castings is required to determine the chronological implications of the differences observed and to elucidate more generally the early development of hollow casting technology, the scan of the Head of a ruler provides evidence of some of the challenges encountered and problem-solving strategies used in the casting process.

Introduction

Determining the origins of specific metallurgical innovations in the ancient world is challenging given the chance nature of archaeological finds and the fact that metal objects were often melted down for reuse. Even the rationale for pursuing this effort is questionable since different cultures may have independently developed similar techniques when faced with the same needs and constraints. Nevertheless, based on existing evidence it appears that the earliest largescale and highly naturalistic representations of human and animal figures in metal were cast in Mesopotamia in the middle of the 3rd millennium BCE thanks to the adoption of lost wax casting over a ceramic core [1, 2].

Lost wax casting had been practiced for several centuries before this time but there was a limit to the size and articulation of objects that could be produced in solid form due to the risk of casting flaws. For this reason, largescale metal sculptures—such as the impressive lintel relief and adjacently installed freestanding bull statues at the Ninhursag Temple at Tell Al-‘Ubaid—were made of metal cladding nailed over forms of bitumen-covered wood ([3], p. 77). Wrought copper could not match the modeling potential of wax, however, and ceramic cores provided a solution to the scale limitations of solid lost wax casting. It appears that a desire to overcome these limitations rather than economy in the use of metal was the primary motivating factor in developing this new technology.

For over half a century, the International Congress on Ancient Bronzes has highlighted the importance of undertaking technical investigations to advance knowledge about casting in the ancient world [4]. To date, the focus of its attention has largely centered on casting during the Classical Greek period (ca. 480–323 BCE) and later. Due to a number of factors that precluded their accessibility, the much earlier examples of Mesopotamian largescale lost wax casting in the collection of the National Museum of Iraq—including the so-called Sargon Head that was excavated at the site of Nineveh in northern Iraq and the partially preserved Bassetki Statue that was discovered during road construction near Dohuk—have not received the same level of scrutiny [5, 6] (Fig. 1). Both of these sculptures date to the Akkadian period (2334–2154 BCE) during the Early Bronze Age. Josef Riederer, a key figure in the development of heritage science as a discipline, was able to radiograph the Sargon Head and take samples for compositional analysis when it came on loan to Berlin in 1979. However, the results of his examination were only published in a brief summary by Strommenger, and the radiographs he took are published here for the first time [7]. No technical details have been published regarding the Bassetki Statue.

Fig. 1

Above: Three sides of the Sargon Head, Iraq Museum IM11331. Photo from Public Domain. Below: Bassetki Statue, Iraq Museum IM77823. Image by Osama Shukir. Muhammed Amin FRCP(Glasg), CC BY-SA 4.0 < https://creativecommons.org/licenses/by-sa/4.0 > , via Wikimedia Commons

In this context, the recent reattribution of the Head of a ruler in the collection of The Metropolitan Museum of Art (The Met) to approximately the same period as these two sculptures was a significant development. The Head of a ruler depicts a man with a long beard wearing an elaborately braided hairstyle with intertwined hairbands. (Fig. 2) This sculpture has no known archaeological provenience. When acquired from Joseph Brummer, a prominent art dealer with offices in Paris and New York, in 1947, this head was thought to have been found in Iran even though its visual attributes pointed to ancient Mesopotamia as its probable cultural context. For many years, scholars including Muscarella sought ways to reconcile these two geographies ([8], pp. 368–374). However, recent research has shown that the Iranian provenience derived, like the head itself, from the art market and that the statue likely represents a Mesopotamian ruler [9]. A stylistically similar stone sculpture fragment from the site of Tello in Iraq and an analysis of the Head of a ruler’s visual attributes allow us to further narrow down its production to the late third millennium BCE during the Akkadian or Ur III periods.

Fig. 2

© The Metropolitan Museum of Art, photos by Anna-Marie Kellen

Three views of the Head of a ruler, Mesopotamia, Iraq, ca. 2300–2000 BCE. Copper alloy, 13 9/16 × 8 3/8 × 9 3/16 in. (34.4 × 21.3 × 23.3 cm). The Metropolitan Museum of Art, Rogers Fund, 1947 (47.100.80) Image

This reattribution provided the Ancient Near East Department at The Met with a rationale to seek out the most advanced imaging techniques available (Fig. 2). During a brief period between its removal from The Met’s galleries and before being sent on loan to the National Museum of Korea, the head was brought to the Center for X-ray Analytics at the Swiss Federal Laboratories for Materials Science and Technology (Empa) where it was investigated using high-energy X-ray computed tomography. This scan revealed the considerable challenges encountered by ancient casting innovators seeking to produce sculptures of considerable artistry and power.

Casting copper sculptures in Mesopotamia

A brief survey of the materials and process involved in Mesopotamian copper alloy casting provides the background needed to interpret the features revealed in the CT scan of the Head of a ruler.

In antiquity, metal casting was a dramatic experience for the senses in which pyrotechnical knowledge gained from practical experimentation was used to transform raw materials into symbolically meaningful works of art. Based on recorded observations from ancient Egypt and other cultures, image making was viewed as a generative process—what Bahrani has called “the alchemical production of being through art”—and there is evidence to suggest that in Mesopotamia sculptures were imbued with their own agency and power ([10], p. 163) Metalworking in Mesopotamia was also a domain associated with the gods, from whom knowledge and skills derived, and metalworking expertise was celebrated in royal inscriptions. As suggested by Winter, the technical mastery particularly inherent metal casting—not to mention the access to the valuable resources it required—enhanced the perceived beauty and affective power of metal sculptures to ancient audiences [11].

Modern analytical surveys such as the Sumerian Metals Project undertaken at the University of Pennsylvania have revealed that early Mesopotamian metalworkers were using essentially pure copper with minor amounts of other metals—chiefly arsenic and nickel—that were impurities in the ore that was smelted/reused or could have been intentionally added ([12], pp. 250–251; [13], pp. 165–169; [14], pp. 18–20). Since the Mesopotamian plain lacked mineral resources, copper had to be imported from the surrounding mountainous regions ([15], pp. 220–232). Pure copper has a high melting point and readily absorbs oxygen and hydrogen when molten, resulting in the formation of oxides and pockets of steam that can impede metal flow through a mold during casting ([16], pp. 269–274; [17], pp. 241–242). Alloying with tin and/or lead helps to lower the temperature required and reduce oxygen absorption. While both of these metals were known in Mesopotamia in the during the mid-to-late 3rd millennium BCE, tin was not imported in large quantities into the region until centuries later, and the use of copper-tin alloys did not become commonplace until the mid-second millennium BCE ([12], pp. 251–254). In addition, the advantages conferred by adding lead to copper were either not known or rejected due to the alterations that added lead caused in appearance and/or working properties ([12], p. 294).

Alloying with small amounts of arsenic can confer some advantages over casting pure copper [18], but an inductively coupled plasma spectrometric analysis of a metallic sample obtained in 1992 from the inside of an opening under the beard of the Head of a ruler indicated the alloy used to cast this head contained only 0.69% arsenic and 0.48% nickel as the main constituents besides copper. This composition is closely comparable to the analyses by atomic absorption of the samples that Riederer took from the Sargon Head ([7], p. 115). Casting essentially pure copper without flaws remains a challenge for casting engineers today [19]. Campbell has noted the importance of designing a good system to supply metal to the mold, and Taleb has outlined the requirements needed to avoid flaws, which include filling a mold in the smallest time possible, avoiding turbulence in the flow of molten metal into the mold in order to avoid air entrainment, and ensuring that enough molten metal reaches the mold cavity ([16], pp. 417–446; [19], pp. 15–16).

Having a ceramic core in place reduces both the amount of metal needed and the length of time it takes to fill all parts of the mold before solidification begins. By reducing the thickness of castings, the shrinkage and ‘hot tearing’ that can occur between quickly chilled surfaces and more slowly crystallizing internal zones can be avoided ([16], pp 417–419; [17], p. 145 and 166). A core also reduces the risk that there is not enough molten metal to fill the mold completely—a flaw known as ‘shorting’ ([17], p. 250). Finally, a sufficiently porous core can absorb some of the gases that are evolved during casting, which is especially helpful if the mold is not adequately vented.

No significant metal sculpture has been found in Mesopotamia since Braun-Holzinger’s survey of relevant castings was published [3]. Based on this survey, it appears that a Vessel Stand with ibex support dated to the Early Dynastic III period (ca. 2600 to 2350 BCE)—exhibited for many years at The Met—is one of the oldest known examples of the use of a core in casting a human or animal figure by direct lost wax casting (Fig. 3). Although it has no known excavation history, this vessel stand is stylistically related to a group of stands with solid cast human figures that were excavated at sites in the Diyala region of Iraq from Early Dynastic III levels ([3], plates 10–11, pp. 20–23). While not particularly large, the ibex figure—which measures approximately 20 cm tall and 24 cm wide—presented a casting challenge, given the long, slender legs and the horned head were intended to be cast along with the trunk of the ibex. The base and the upper vessel stand were cast separately.

Fig. 3

© The Metropolitan Museum of Art, photo by Anna-Marie Kellen

Left: Vessel Stand with ibex support formerly in the collection of The Metropolitan Museum of Art; center: radiograph of the ibex in figure with core supports indicated by arrows and circle around flaw in head area; right: cutaway model of a mold of a Medieval aquamanile showing feeders (red) and feeders and fents (blue) attached. Left image:

With the help of conventional radiography, which allows one to distinguish between the hollow core and areas of solid metal, one can reconstruct how the ibex figure was modelled and prepared for casting following the direct lost wax technique (Fig. 3). A clay form slightly smaller than the final shape of the animal’s trunk was made by hand and covered with wax. Legs and a head made of solid wax were then attached to the wax covering the clay core. Metal core supports were driven through the wax and into the core while solid rods made of wax were affixed to the model. These rods provided the channels that supplied molten metal to the mold and the vents that allowed air and gases to escape as it was being filled. This entire assemblage was then covered with an investment of clay.

After the investment had dried, the mold was fired, causing the clay to harden and the wax to pour out of a channel provided by one of the tubes. At this point, the core was held suspended in place by the core supports that bridged the gap between the core and the mold. Before the advent of iron, such core supports were made of copper. During casting, metal filled all of the spaces previously occupied by the wax. After cooling, the mold was broken off, and the projecting channels and core supports now cast in metal were cut down to the surface of the casting. Linear details were added or improved by chasing, which was relatively easy to do in copper. The surface likely was polished with abrasives, but due to subsequent corrosion it is not known if an artificial patina was applied or if the metal was allowed to oxidize naturally.

The optimal disposition of feeder and vent lines posed a challenge to ancient metalworkers as they began to cast larger and more complicated objects for the first time. In this case, poor feeder and vent placement may have caused the head of the ibex not to form properly in the initial casting. As indicated in the radiograph, it had to be added as a separate component secured in place by what appears to be a metal plate and a rivet. Wartke’s study of two late third millennium BCE copper alloy statuettes excavated at Assur—one of the first to use CT scanning in the study of Mesopotamian metal sculptures—demonstrates the lengths that ancient metalworkers would go to salvage a flawed casting [20]. In one of the figures, an entirely new face was cast over the poorly-cast original one—a fact that has been made more apparent by injudicious cleaning in modern times. Although a modern artisan might have scrapped this figure and started over, it should be noted the resources available to early Mesopotamian metalworkers were limited and that the original wax model was lost in the direct lost wax process. Indirect casting, which was not practiced until later periods, involved using a model that could be used to cast multiple wax shells that were subsequently filled with a clay core.

Even with a core in place, therefore, casting flaws and high internal porosities—caused by the release of absorbed oxygen when the copper cooled and solidified—were common in ancient metal sculptures. Given the Mesopotamian tendency to manage risk through the use of rituals, it is possible that the various steps involved in casting were accompanied by prescribed actions similar to those preserved for the creation of Mesopotamian protective spirits in wood and in clay [21]. If the casting of figures was seen as a ritually significant generative act, there may have been an incentive to make the best of the outcomes by addressing flaws rather than completely starting over.

Method

Rationale for high-energy X-ray analysis of the Head of a Ruler

When the Met acquired an X-ray tube capable of generating up to 320kV in the early 1970’s, attempts were made to image the interior of the Head of a ruler, but the radiographs obtained suggested that only an oblong, vertically orientated void was present. Subsequent removal of an old fill in a hole under the proper left jawline indicated the existence of a larger empty core that had been missed by conventional radiography in which radio-opaque features can be superimposed. Following this discovery, it was decided to take The Met’s head to a GE testing facility in Cincinnati in 1992 for industrial CT scanning using an X-ray tube capable of generating up to 420 kV. Given the limitations of the technology at that time, only isolated slices of limited resolution were obtained. While these slices confirmed the existence of a sizable central hollow, they lacked clarity similar to those in Wartke’s study due to excessive attenuation and scattering (Fig. 4).

Fig. 4

Left: profile view radiograph of the Head of a ruler made in 1973 and right: horizontal CT slice made in 1992

The reattribution of the Head of a ruler in 2022 provided the impetus to obtain a complete, high-resolution scan. Given the thickness of its metal wall, it was recognized that computed tomographic techniques using either high-energy X-rays or neutrons would provide the 3D volumetric analysis being sought. Thermal and cold neutrons have been employed to investigate metal sculptures [22,23,24,25,26] and copper alloy coffins from Egypt from the British Museum [27], providing a comparable resolution to high-energy X-ray CT. They also can offer a high sensitivity for the analysis of corrosion products, which can be of interest to archaeologists and conservators [28,29,30,31]. In addition, Neutron Bragg-edge transmission imaging can also provide crystallographic phase information over large volumes of the sample and has been demonstrated in the context of analyzing Japanese swords [32].

The main drawback of using thermal and cold neutron imaging in a museum context, however, is the radioactive activation of the artifact under examination, which requires a period of post-imaging isolation and storage. Using well-known online activation calculators (e.g. of NIST https://www.ncnr.nist.gov/resources/activation/ or from TUM Germany https://webapps.frm2.tum.de/activation//), the activity of the pure copper in our specimen would result in estimated cooling times ranging from days to weeks. In addition, activation of minor constituents such as the 0.48% Ni found in the ICP analysis as well as trace elements in the considerable mass of a sculpture such as the Head of a ruler could result in an unpredictably long half-life. Given that this head is regularly on view and would only be available for a short period of time, neutron imaging had to be removed from consideration.

High-energy or fast neutron imaging has been used to penetrate dense materials with less risk of activation, but this technique is only available at a small number of facilities and still lags with respect to spatial resolution behind CT imaging with X-rays [33, 34]. The latest developments with fast neutron converter screens demonstrate some proof of concept for high resolution imaging, but this has only been demonstrated to date with considerably smaller objects (see e.g. [35]).

Medical CT units, utilizing X-rays that do not induce radioactive activation, have been used to investigate the internal structure of art objects made primarily of organic materials with relatively low X-ray energy requirements such as the Ptolemaic cartonnage masks studied by Vandenbeusch et al. [36], fabrics and wooden sculptures studied by Bossema et al. [37], and Central African sculptures examined by Bouttiaux and Ghysels [38]. A CT system with specially designed detection array and a tube capable of generating up to 450kv was used in the J. Paul Getty Museum to scan a Roman bronze cupid, but museums do not have the resources to maintain such setups permanently, and the tube used in that case was not sufficiently powerful for this project [39]. Since that time, significant developments have occurred in the design of compact linear accelerators while the spatial resolution of detectors has been refined, allowing for highly detailed imaging even of dense metal objects [40]. It should also be noted that a broad range of non-destructive techniques and spectroscopic methods have been developed that could reveal information regarding elemental composition, crystalline phases, and residual stress and strain in parts or in small gauge volumes of the metal with different penetration strength and applicability. These include neutron and X-ray diffraction methods, X-ray fluorescence and X-ray absorption spectroscopy, prompt-gamma or general neutron activation analysis and neutron resonance capture analysis [34, 41, 42]. Given time constraints in this case, however, the investigation was limited to resolving questions raised by the earlier CT scan slices.

Swiss Empa’s Center for X-ray Analytics was one of the first facilities to apply high-energy CT scanning to the study of archaeological material [43, 44]. Given its experience and ability to provide museum-level security, it was decided to have the Head of a ruler CT scanned at this location. Recently, the Fraunhofer EZRT High Energy facility has announced the availability of two high-energy photon CT systems for the study of cultural material that would also have been suitable for this study [45, 46].

Scanning parameters

The Empa high-energy scanner is a linear accelerator (LINAC)-based system consisting of a dual energy 4/6MeV LINAC from U.S. Photon with a focal spot size of 2 mm centered on an array of 29 detector modules placed along an arc of a radius of 2.7 m. Each module from Detection Technology is composed of 128 channels made of 10 mm thick CdWO4 scintillators for an efficient detection of the high-energy X-rays with a 0.4 mm pitch. Both source and detector are horizontally collimated with 2 mm slits made of tungsten plates. Such fan-beam scan arrangement with heavy collimation enables the suppression of scattered radiation in the detector and thus a superior image quality with the trade-off of longer overall acquisition times for the slice-by-slice scanning of 3D objects [43, 44] (Fig. 5). The scan conditions are summarized in Table 1. The raw projections have been corrected for dark-current and subsequently flat field normalized. Outliers and ring artefacts have been filtered by a selective 2D median filter applied on the sinogram such as "Remove Outliers" filter of ImageJ [47], using a strongly asymmetric filter kernel of size 1 × 15 pixels implemented in MATLAB. The reconstruction was done by standard fan-beam filtered back projection method [48] implemented in an in-house software package. The reconstructed CT slices in 32 bit TIFF format have been visualized and further processed by the commercial software VG Studio Max 3.5 (Volume Graphics). This included the analysis of closed porosity (defects) in the metal walls of the sculpture based on threshold segmentation.

Fig. 5

Left: LINAC CT system with the detectors on the right. Right: VG Studio Max 3.5 (Volume Graphics) using Empa Platform for Image Processing. Photos by Empa

Table 1 Summary of the scan conditions

Results and discussion

The CT scan revealed that the front of the head is very thick—with a thickness ranging between 30 and 37 mm—while the back of the head is considerably thinner, going down to a thickness of approximately 6 mm. Six large core supports—essentially tapered spikes—measuring between 23 and 26 mm long—and with a square cross section of approximately 5 mm wide where they intersect with the metal wall—were inserted through the wax and into the core: two on the front and two on each side (Fig. 6). Virtually all of the features of the face and beard, therefore, were modeled in thick layer of wax that was replaced by metal during casting. The substantial mass of metal in the face accounts for a significant proportion of the weight of this sculpture. The scan also indicated that the large cracks visible on the proper right side of the face exist only in the corrosion layer and fortunately do not extend into the metal. By providing the three-dimensional boundaries of this head’s internal cavity, the CT scan has enabled the reconstruction of the shape of the Head of a ruler’s original ceramic core once the unintentional pockets of trapped gas are removed. This core—which was strikingly modern in its simplified abstraction—included only a slightly raised ridge for the brow line and the nose and did not include most of the beard (Fig. 7). By contrast, the profile radiograph of the Sargon Head reveals that the shape of its core followed more closely the exterior contours of the head (Fig. 8). As a result, the wall of that casting is thinner and more consistent in its thickness overall, which would have been easier to cast. One can also see that pairs of core supports—found by Riederer to have been made using very pure copper—were used in the Sargon Head unlike the large ones found in The Met’s head.

Fig. 6

Left: CT slice taken at the center of the profile of the Head of a ruler in which the differing thickness of the metal wall is visible; right: 3D reconstruction of the Head of a ruler with the back of the head removed and looking towards the front in which all six core supports can be seen

Fig. 7

3D CT reconstruction of the internal void of the Head of a ruler in yellow, which corresponds to the shape of the original ceramic core. Grey shaded overlay corresponds to the wax that was applied and modelled over the core and replaced by metal during casting

Fig. 8

Frontal and profile view radiographs of the Sargon Head taken at the Rathgen-Forschungslabor in 1979. Compare the profile view with that of the Head of a ruler above in Fig. 4. Provided by the Staatliche Museen zu Berlin—Vorderasiatisches Museum

The most unusual feature of the Sargon Head’s fabrication is the way that the ears were made. According to Riederer, the ears were cast first and attached to a metal plate spanning the head on which the clay core was formed ([7], p. 114). He did not specify how the ears were attached to the plate, and this detail is not visible in the radiographs. Most likely they were mechanically fixed, cast on, or attached by a version of flow welding in which molten metal was introduced to the space between the ears and the plate ([49], pp. 94–95). Such careful attention to the ears may reflect the fact that this head was intended to embody the agency of the ruler it represented and needed well-formed ears to hear. Given the difficulty of casting thin projections flawlessly during this period, it may have been safer to make the ears first. Originally the eyes (as with the Head of the ruler) were animated with inlaid materials, possibly shell or colored stones, that are no longer preserved.

The enduring power of the Sargon Head as a living image is reflected in the fact that its ears were hacked off, its eyes gouged out and the ends of its beard were broken off late in its long life. Nylander has proposed that these damages were inflicted by invading armies in the late seventh century BCE—a millennium and a half after its manufacture—when they sacked Nineveh the Assyrian capital where the head was discovered [50]. The Head of a ruler most likely projected the same life-like power to ancient audiences.

One of the earlier CT slices of The Met’s head contained a black line surrounding the proper right ear suggesting that it may have been precast like those of the Sargon Head and set into the wax model. While the 3D scan at Swiss Empa indicated that this was not the case, it did reveal significant porosity near the proper left ear which almost prevented it from being cast. The beard, which was also meant to be cast as a solid projection from the rest of the head, was not so fortunate as a large section on the proper right side failed to cast. No evidence has been found to indicate if any compensation was made to restore this important symbol of power in antiquity (Fig. 9).

Fig. 9

Casting flaws in the Head of a ruler. Left: vertical slice through 3D model showing large porosity, which almost prevented proper left ear from casting; right: missing area on proper right side of beard seen from the back of the 3D reconstruction

The CT scan revealed a high degree of porosity overall which is largely concentrated on the front of the head. While a majority of the pores have a volume below 200 mm³ (Fig. 10), several large pockets of gas were found trapped against the core while another extends from the back of the proper left eye socket to the core. A large pore above the corner of the proper right eye was covered with a 20.54 × 10.2 mm metal patch that is now buried beneath the thick oxidation crust (Fig. 11). Fortunately, the porosity of the metal does not appear to have adversely affected the surface. To some degree surface porosity can be reduced by controlling the solidification rate of metal during cooling, but it is not known if ancient metalworkers exploited this possibility.

Fig. 10

Pore analysis of the Head of a ruler by VG StudioMax 3.5. Material shown with 70% transparency. Note: only closed porosity/defects (with no connection to outside/inside of statue) are shown

Fig. 11

Vertical slice of the Head of a ruler (viewed from inside the head) revealing metal patch buried in corrosion over proper right eye

It was once thought that an area of surface deformation and rough texture located at the back of the head could have been the spot where a bun of gathered hair was attached such as that found on the Sargon Head (Fig. 12). Visual evidence from the scan, however, suggests that it is an area where another significant casting flaw occurred. It appears that the wax model may have been so thin in this area that the metal fronts cooled and solidified before joining to close the wall of the head. On the interior one can see a flow of metal with a raised edge and a central concentric pattern (Fig. 13) suggesting that metal was added in a second casting that flowed into the cavity and pooled against the back wall while the head was placed horizontally with the face up. The back of the core originally filled the space now taken up by this added metal and had to be at least partially removed before it was cast. This explains why the 3D cavity reconstruction looks like it was scooped out at the back creating raised edge, which would not make sense given the steps involved in direct lost wax casting as described above. Nonmetallic material still adhered to the inside wall of the face may be remnants of the original core that were not removed in this repair process.

Fig. 12

Left: Back of the Head of a ruler seen in natural light; Right: the same in 3D reconstruction

Fig. 13

Left: 3D reconstruction of internal cavity with raised edge indicated by arrows; right: view looking towards the back of head with raised edge similarly indicated

Finally, the CT scan helped to confirm that the line on the proper left side of the head above the ear is an intentional feature that was made in the wax model (Fig. 14). While the purpose of this line is not certain, we know from Pliny as well as statue fragments excavated in Athens that ancient copper alloy sculptures were occasionally gilded using gold foil that was secured in place by inserting the ends of the foil into grooves cut into the surface ([51], pp. 2–3). The drawback of using gold foil instead of leaf is that fine surface details—such as the braiding of the hair on this head—could be obscured by the thicker metal. In this case, given that the line was placed in an area where it ran largely across the hairbands crisscrossing the hair, rather than the hair itself, it is possible that gold foil strips were burnished down only onto the bands.

Fig. 14

3D CT reconstruction of the upper proper left side of the Head of a ruler where an intentional incision was made, perhaps to secure gold foil

Three videos can be found under Supplementary materials in the online version. The first illustrates some of the findings of the 3D scan. The second is a 3D animation of the CT slices along the coronal and sagittal isoplanes, while the third is an animation of the slices along the transverse plane beginning at the bottom of the sculpture.

Conclusions

The LINAC CT scanning system at Swiss Empa provided a highly detailed 3D model of one of the earliest life-sized metal sculptures produced in the ancient world. Condition information regarding the corrosion layers, metal walls, and extant core material was also obtained. This information will help to ensure the preservation of this important example of early hollow core casting and will be used to inform didactic material available to the public in The Met’s galleries and on its website.

The production of the Head of a ruler was strikingly different from that of the Sargon Head. With its comparatively thin walls, numerous core supports, and separately cast ears, the latter sculpture seems to reflect the work of artisans who were aware of the pitfalls of copper casting and thus avoided major casting flaws. While the Head of a ruler was modelled with considerable skill and arguably greater naturalism, its execution in metal was considerably less successful. Casting a partially thick-walled sculpture in what appears to have been a poorly vented mold was risky, and it resulted in significant casting flaws that almost doomed an effort on which considerable material and artistic resources had already been expended. The Met’s head seems to have been produced by an artist who was used to sculpting in a solid material such as stone or clay and either was not fully aware of the specific needs of casting in metal or was collaborating with those with limited casting expertise.

With little comparative material available, it is not possible at this point to ascribe chronological significance to the difference in fabrication between both heads. Nevertheless, the details revealed by the X-ray computed tomography scan along with the composition of the metal corroborate the reattribution of the Head of a ruler to the formative period of life-sized hollow core casting in Mesopotamia proposed by Eppihimer [9]. It should be noted that the Bassetki Statue—which weighs approximately 150 kg despite its fragmentary condition—also appears to have been cast with very thick metal walls even though a ceramic core was used. Hopefully this sculpture will receive the technical study it deserves in the future.