Old ceremonial masks and knives are popular symbols of pre-Hispanic Peruvian culture. Examples adorn the covers of books on Peru and serve as emblems for some Peruvian institutions. These precious metal artifacts are often attributed, even by knowledgeable persons, to the Incas or to their coastal rivals, the Chimú. Yet many of them are not Incan or Chimú at all: they were created much earlier by the Sicán culture, which was centered in the Lambayeque region of northern Peru and flourished from the ninth to the 14th centuries.

The Middle Sicán era, between 900 and 1100 C.E., produced enormous quantities of precious metal artifacts, many showing extraordinarily high craftsmanship. We and our colleagues from several disciplines have scrutinized the metalwork from one Middle Sicán trove in an attempt to reconstruct the technology and organization of precious metal production and to define the meaning of those products within the culture. We determined that the scale and the range of metal use by the people of the Middle Sicán was unprecedented in the pre-Hispanic New World. That culture's extensive production of arsenical copper ushered the bronze age into northern Peru. Gold alloys were the most prestigious media for political, social and religious expression. In fact, we suspect that metallurgical production was a prime mover of Middle Sicán cultural developments.

Ambiguity and ignorance have traditionally shrouded precious metal artifacts of the Middle Sicán. Almost all of those in private and public collections were looted from tombs within what is today the Poma National Historical Sanctuary in the mid-La Leche Valley, about 800 kilometers north of Lima. The modern period of intensive grave robbery began in the 1930s. Treasure hunters sank vertical prospecting pits into likely spots, then dug horizontal tunnels outward. With the discoveries of more rich tombs, the extent of the looting continued to increase through the 1940s and 1950s. It culminated in the late 1960s, when a bulldozer was employed for a year to remove the surface soil so that outlines of the tomb pits could be seen more easily. Looting took place sporadically until the mid-1970s, effectively hindering any long-term scientific study of the regional prehistory. When one of us (Shimada) began fieldwork in 1978, he counted more than 100,000 looters' holes and hundreds of long bulldozer trenches on aerial photographs of the Poma sanctuary.

The lack of contextual information for those looted artifacts greatly limits understanding of their sociopolitical, religious and economic significance. Moreover, looters and collectors often took questionable and undocumented measures to "restore" stolen artifacts. Pigments, feathers and ancient tool marks on gold objects could have been removed by careless cleaning. "Missing" inlay pieces or bangles were often arbitrarily replaced. As a result, the appearance of objects cannot be taken at face value, which limits the information that can be drawn from them. Any attempt to understand the objects, their cultural significance and the techniques used in their manufacture is therefore best founded on those artifacts scientifically recovered from intact tombs.

Sampling a Tomb
THE OPPORTUNITY to gain just such an understanding came about with the first scientific excavation of the tomb of a member of the Middle Sicán elite at Huaca Loro, an adobe platform mound in the Poma sanctuary. The tomb was apparently one of a string left by the Middle Sicán, some of which had already been looted, along the east and south bases of Huaca Loro. Shimada recognized it during a survey of Batn Grande in 1978. He planned its excavation over the next 10 years as part of his broader sampling of Sicán tombs for the elucidation of that culture's social organization. Preparations included assembling a group of specialists and piecing together a Sicán cultural chronology, as well as the performance of other background research.

In particular, Shimada needed to make certain that the groundwater level was sufficiently low to allow safe excavation. He also gave a series of public talks on the scientific value of the planned tomb excavation to local residents, who were weary of tomb looting. The chamber was finally excavated, under Shimadas supervision, by the Sicán Archaeological Project between October 1991 and March 1992.

The central person buried in the tomb was a 40- to 50-year-old man who had been one of the elite. He was accompanied by the bodies of two women and two children who apparently had been sacrificed. The six-month excavation yielded approximately 1.2 tons of diverse grave goods packed in a burial chamber roughly three meters on a side at the bottom of an 11-meter vertical shaft. By weight, metal objects and scrap account for nearly three quarters of the grave goods. Most of the objects, according to a systematic analysis of some 1,000 samples by Adon Gordus, professor of chemistry at the University of Michigan at Ann Arbor, appear to be 12- to 18-karat gold-silver-copper alloys roughly equivalent to the gold commonly used today for jewelry. Some objects and nearly all the scrap are tumbaga, gold-silver-copper alloys that are high in copper and have a few to 10 karats.

Arranged concentrically, the objects surrounded the body of the man, which was thoroughly painted with cinnabar (an intensely red paint made up of mercuric sulfide and a binder). The body was seated and placed in an inverted position [see (1) in box on opposite page]. The head, with its three sets of attached ear ornaments and its large gold mask, was detached from the body and rotated 180 degrees so that his head was right side up and facing west. A mantle (its cloth long since decayed) onto which nearly 2,000 gold foil squares had been sewn was laid on the floor underlying the inverted body. Placed on, around and underneath the body were a staff with gold and tumbaga ornaments on top of a wooden shaft, a gold headdress with a sculptural representation of an animal head, a pair of gold shin covers, a pair of meter-long tumbaga gloves (one holding a gold cup with a silver rattle base), a gold ceremonial knife (or tumi) and a cluster of six magnificently made pairs of gold earspools. His chest was bedecked with a nearly 10-centimeter-thick layer of beads (made of sodalite, amethyst, quartz crystal, turquoise, amber, fluorite, calcite, shell and other materials). Farther away, near the edges of the chamber, were about 500 kilograms of tumbaga scrap and more than 250 kilograms of arsenical copper implements.

By far the most impressive find within the tomb is Gold Cache 1, which we discovered at the northwest corner of the burial chamber. Inside a rectangular box lined with woven mats were at least 60 major objects, most of gold sheet, the balance of silver or tumbaga. This cache contains a peculiar mixture of ritual paraphernalia and personal ornaments: five crowns, four headbands, at least 12 tumi-shaped head ornaments, at least six sets of gold feather head ornaments, three tumbaga fans and 14 large disks that were either ornaments for staffs or the backs of headdresses. At the bottom were the largest objects: four sets of parabolic headdresses that would have been set on top and in front of the crown.

Two of the seven niches carved into the walls of the burial chamber also contained metal objects. A pit dug into the largest niche, in the east wall, was packed with an estimated 1,500 bundles of naipes. These arsenical copper sheet objects of uniform shape and size may have served as currency. Each bundle consisted of 12 or 13 naipes. The pit also held two silver alloy tumi knives, thousands of small gold foil squares and at least two dozen tumbaga masks identical in shape to the large gold mask on the buried man but smaller, technically inferior and less ornamented. A second niche contained Gold Cache 2--another collection of gold ornaments and ritual objects. We also recovered from this grave more than 50 kilograms of diverse stone and shell beads, the carved wooden frame of a litter, about three kilograms of cinnabar and 21 ceramic vessels. Notable contents included clusters of 179 thorny oyster (Spondylus princeps) and 141 Conus (C. fergusoni) tropical seashells, probably imported from coastal Ecuador. These shells are known to have held considerable symbolic value relating to the concepts of life force, water and fertility.

Our study differed from earlier ones in two important respects: we had access to materials from an intact tomb, and we had the benefit of the insights of metalworking specialists. Most published studies have focused on single aspects of the manufacture or use of pre-Hispanic precious metal artifacts. Such studies relied primarily if not exclusively on laboratory analyses of looted objects. Also, the investigators who have made inferences about the manufacturing techniques have rarely had personal knowledge of metalworking as a craft. The reconstructions emerging from those narrow, academic studies are therefore tenuous and biased. What this area of archaeology needs is more comprehensive studies of scientifically excavated samples from a multitude of analytical and interpretive perspectives. Only then will we gain an in-depth appreciation for the organization of metal production and the meaning of metal products within a culture--a "holistic vision" of metallurgy.

One of us (Griffin), a skilled metalsmith and longtime conservator, had experience with many pre-Hispanic collections. Because of her background, she was able to elucidate much about Sicán precious metal production by examining the Huaca Loro artifacts. Specialists in related fields helped us to interpret the evidence of other metal samples, feathers, beads and remains, including human skeletons, associated with the metal objects.

From its documented beginning around 1500 B.C.E., Andean metallurgy emphasized the use of sheet metal fashioned from ingots with stone anvils and hammers, in contrast to gold works in Central America and Colombia, where lost-wax casting was the main technique. Gold crowns and other ornaments found at various sites are all essentially gold sheets decorated with repoussé and cutout designs. In terms of the dimensions, smoothness, consistent thickness and overall quantity, the objects from the Huaca Loro tomb are prime examples of this sheet-metal tradition.

The primary tools for making sheet and wire are handheld stone hammers and companion anvils, chisels and chasing tools made of arsenical copper and tumbaga. The hammers are commonly of magnetite, hematite or fine-grained basalt; they range from very tiny to the size of a mans fist. One hammer face is usually domed for stretching the metal, and the other face is flat for planishing--removing the shallow dimples left by the stretching blows. When properly executed, this technique yields a flat, smooth sheet. Though simple to describe, the task requires considerable time and skill. Because the sheet-metal maker must forge the gold while it is cold, the metal must be annealed regularly to prevent stress cracks.

Many gold objects in the Huaca Loro grave attest to remarkable expertise in sheet making. Consider, for example, the long borders on four parabolic head ornaments. Two of them are two-meter-long continuous strips of forged metal with an even width (around 4.5 centimeters) and thickness (about 0.15 millimeter)--virtuoso performances in sheet making. The mask (46 by 29 centimeters) that covers the face of the man buried in the tomb is another tour de force. It was fashioned from a sheet about 0.6 millimeter thick. The metal, composed of 52 percent gold, 31 percent silver and 17 percent copper (about 12.5 karats), had to be thin to keep down the mask's weight (only 677 grams) yet thick enough to allow the large naturalistic nose to be raised some four centimeters from the center without causing stress cracks. The high silver content, in addition to the gold of the alloy, provided the malleability necessary for the job. Silver-copper sheet metal was used for the whites of the eyes with a large pierced amber and emerald bead representing the iris and pupil of each eye, respectively. Our analysis indicates that the amber and emeralds were most likely imported from Colombia. Overall, the mask represents an elegant solution to the technical challenges as well as aesthetic and symbolic considerations of the Middle Sicn culture.

More than a dozen tumi-shaped headdresses provide additional illustrations of expert sheet-metal making. The tang, or stem, of each headdress had to be narrow but sufficiently stout (about a millimeter thick) to stay upright when inserted into a turban or crown socket. At the top the sheet is only about 0.15 to 0.18 millimeter thick. The smith knew just how much to planish the sheet to give it the right amount of springiness so it would wave with each movement of the head but not crack or bend.

The same is true for six sets of gold feathers that are believed to have been part of an elaborate headdress. The sets we have studied consist of 11 or 12 feathers, each about 20 to 21 centimeters long and two centimeters wide. Each feather tapers in thickness from the stem (about 0.10 millimeter) to the upper tip (about 0.07 millimeter). The sets as a whole have a fan shape: the central feather is straight, and those to the sides are increasingly more curved to the right or left. The component feathers are mechanically joined by straps and slots near their stems. Below the straps, each feather has a slight ridge along the longitudinal axis to provide some rigidity. Preserved on the stems are the imprints of fine threads that apparently stitched them to some cloth backing. The overall design and structure, as with the tumi-shaped headdresses, allowed the feathers to sway gently with head movements while staying rigid and light enough to mount securely on the headdress. In general, Middle Sicán precious metalworking emphasized movement as much as color (not only of gold but those of feathers, inlays and paints) and sound.

The six pairs of gold earspools found near the southeast corner of the tomb demonstrate a level of technical mastery rarely seen in pre-Hispanic gold objects. They display a constellation of design features that may well represent a single school of goldsmiths. The varied and advanced metalworking effects, not usually found together, include forged wire, true filigree, excellent finish and polish, and protobrazing. Protobrazing is a superbly simple method of joining gold or silver alloys that utilizes either the copper in those alloys or the verdigris (copper acetate) in an organic glue. The pieces to be joined are heated over charcoal in a reduction atmosphere; at the right temperature a new alloy forms where the metals touch.

To prevent the earspools from developing a torque while being inserted into the earlobes, the metalsmiths had made them from a relatively heavy gauge metal (about 0.35 to 0.55 millimeter thick). The front flange was domed by striking it from the back with a hammer against a shallow depression, probably in wood. The resultant domed flange greatly increased the twist strength while adding needed depth to the design of the piece.

In three of the pairs the metalsmiths implemented complex decorative designs by means of mechanical solutions that are simple and elegant. One such solution is the "tab and slot" joining of the sheet metal that forms the central rod connecting the front and back flanges. In several pairs the smiths also used wire structural supports to create the illusion that the central elements of the design "float" within the frame of the front flange.

One pair of the earspools, which might be called a "sketch" or trial piece, shows how the Sicán goldsmiths gradually refined the mechanical solutions to problems posed by an intricate new design. In this pair, very fine gold wire was used to anchor a circular frame to the X-brace below it. Other pairs, which were presumably made later, show the use of permanent protobraze joins for the same purpose.

An additional indication of careful planning comes from the gold mask, which has its own pair of large earspools anchored directly on the metal earlobes by straps and slots. Not only do the size and shape of the earlobes match those of the back flanges of the earspools, but the slots for all three straps on each ear also match. The slots were punched through both pieces simultaneously by the same vertical strokes. It is quite likely that the mask and all the earspools were manufactured according to high-quality standards in the same workshop.

Not all of the objects found in the Huaca Loro tomb are flawlessly finished. One example is the small double-bottomed beaker (about 12 centimeters high and 10 centimeters in diameter) found in the hand of a ceremonial gold glove. The base of the beaker is made of a raised silver sheet ornamented with cutout designs. It fit onto the base of the gold beaker and was intended to contain rattle stones. Part of the rim of the silver base is melted, and the base shows an "orange peel" texture, the result of overheating. The lower part of the gold beaker displays a distinctly gray semicircular coating that extends up the side, where a part of the silver base rim is missing. Those features suggest that part of the rim melted accidentally while the silver base was being joined to the bottom of the gold beaker. The cup was placed in the brazier upside down. It would not take more than a few seconds of overheating for the silver to melt and create the observed features. The base of the gold beaker was covered by a flash of melting silver as it flowed down toward the heat source. In general, the flawed pieces are most informative as to the manufacturing techniques and process.

Sound Practice
OTHER TECHNICAL DETAILS on the beaker are also informative. For example, it is decorated with three chased representations of a Sicán lord. To create the image, the goldsmith used a tool called a tracer about three to four millimeters in width. That tool was too wide and its edges too sharp for executing the round chin of the figure's face; as a result, the chin line is ragged. The metalworker made no attempt to correct the error when the chasing was finished on the outside of the beaker. This kind of mistake generally indicates haste or an apprentice worker.

The goldsmith would have stopped frequently to anneal the piece to prevent stress cracks from developing. How could he have known when to anneal? Immediately after annealing, the metal emits a dull sound when struck with a hammer. After repeated blows, the pitch of the sound becomes much higher, rising from a thuk sound to a think. With experience, one can tell from the pitch when it is time to anneal.

The impressive scale of sheet making during the Middle Sicán can best be seen in the 500 kilograms of scrap piled along the edges of the burial chamber at Huaca Loro. These piles are apparently not unique: local old-time looters recall finding similar quantities of scrap in other tombs nearby. In addition, we documented the extensive use of tumbaga sheets to line the interior of the gigantic Middle Sicán tomb at Huaca Las Ventanas. That tomb measured 15 by 15 meters at the mouth and three by three meters at the bottom, which was about 11 meters below the surface. Rectangular sheets of set dimensions were carefully placed side by side on the interior surface. They were then covered with cotton cloth that was painted with elaborate polychrome religious images and scenes. The total surface area of the sheets lining this tomb may have exceeded 100 square meters.

The scrap is essentially small pieces left over from sheet-metalworking processes and rejects from manufacturing mishaps. It includes, for example, a partially used tumbaga ingot, square gold foils with poorly executed perforations, broken wires and bells, and sheet-metal trimmings that still retain the outline of the cutout pieces. Such scrap would have been carefully saved for recycling into new ingots.

The scrap clearly represents an enormous investment of manpower and materials. Its presence in the tomb testifies to the political power of the person buried there. In one experiment, using ancient stone hammers, Griffin needed about a day and a half to produce a uniformly thin sheet 10 by 15 centimeters in size from a 30-gram gold nugget. Moreover, the ancient Sicán metalworkers added another step: they treated tumbaga sheets with acid, which dissolved some of the copper near the surface. As a result, the tumbaga sheets had an appearance that approximated that of 24-karat gold. This process is generally known as depletion gilding. The metalsmiths then burnished the sheets, which imparted an excellent finish and compacted the layer of spongy gold left by the depletion gilding. In our opinion, this compacted layer is the peeling gold often seen on gilded tumbaga pieces. Some researchers have proposed that this gilt was deposited electrochemically, but none of the examinations of the sheets conducted by us and others using microscopes and electron microprobes can find any evidence to support that idea.

Sicán culture must have employed a sizable corps of master sheet makers who produced sheet goods for various applications. The remarkable degree of control over forging and finishing seen in these objects argues persuasively that those activities were in the hands of full-time specialists. These master sheet makers would have been assisted by perhaps dozens of apprentices who would have carried out repetitive and time-consuming tasks, such as burnishing or the early stages of remelting scraps to prepare ingots for making sheets.

This master-apprentice arrangement is clearly visible in the manufacturing stages of other objects. One crown in particular shows well-done chasing and perforations in the front but uneven hammer blows and perforations, as well as fine-scribed guidelines on the back. It is likely that the front was begun by a master who showed an apprentice how the remainder was to be done and then went on to another task.

This type of workshop would probably have required a series of multiroom shops, each with a fair number of apprentices and a sizable output. Sheet making, which entails long hours of rhythmic hammering periodically interrupted by annealing, most likely took place in a well-ventilated room. Polishing was probably done in a separate, well-protected room, because airborne sand and other contaminants would have wreaked havoc with the polishing efforts. Significantly, multiroom adobe structures atop the north platform of Huaca Loro and northeast of Huaca Las Ventanas have benches, split-level floors, and many dispersed spots where one can find slag fragments, droplets of copper alloys and evidence that fire was used there. Those two areas were probably centers of metalworking.

Making metal sheet requires great finesse. The shaping and ornamentation of gold objects would have been in the hands of even more consummate master specialists. Because of their exceptional quality, innovative designs and technical distinctiveness, the mask and earspools in the Huaca Loro tomb are probably the products of only one or two masters. Other gold objects, we suspect, were manufactured in other workshops. Although those workshops may have performed different functions, it is unlikely that they were isolated: part of the apprenticeship training would have depended on frequent association with the masters. The apprentices were no doubt given tasks, such as making bangles, that were instructive but did not pose too many technical challenges.

Some of the observed technical variation may reflect the goldsmiths personal styles. Many of the gold objects from Huaca Loro are nearly identical in size and shape but were clearly made in different ways. For example, on some rattles, bangles were attached to the "floating" circular bands by protobrazed wires, whereas others were attached by wire loops. Some of the sharp gold nails used on a dart thrower were cut in a sawtooth pattern from hand-forged wire with a chisel; others were cut from the end of wire that had been filed to a conical shape.

Such observations lead us to conclude that the production of metal objects was organized into task-specific work groups, which in turn were based on a nested hierarchy of masters, apprentices and other supportive personnel. Precious metalworking must not be viewed in isolation from other crafts. Considerable effort had to go into the procurement and preparation of feathers, cinnabar, hematite and other materials that covered the metal objects. Turquoise, shells, bitumen and other substances were needed for inlays. Resins and pitch had to be prepared to make adhesives. Cloth had to be woven as a backing material. We know that arsenical copper was produced on a large scale at specialized settlements close to the mines. All these activities need to be considered to appreciate the impressive magnitude and complexity of the production of sumptuary goods during the Middle Sicán.

Meaning of Metalwork
FOR THE SICN PEOPLE to have invested so much effort in metalworking, metal objects must have held strong meaning for them. We have developed some working hypotheses about what that meaning was. Gold objects seem to have been the aesthetic locus of Middle Sicán art--they embodied the highest standards for artistic expression in the culture. And it is among the gold objects that we find the most explicit expressions of the important Middle Sicán icons and scenes. Ceramic decorations, in comparison, present only partial or simplified versions of these portrayals.

Differential access to a range of metals seems to have marked the social strata. Approximately five dozen excavated burials can be grouped into those that contain no metal objects, those that contain only arsenical copper, those that have arsenical copper and tumbaga items, and those that have gold in addition to those other materials. Tumbaga, along with gold and silver, seems to have been used to symbolize political power or high social status and to convey religious messages. In terms of the scale of production and the range of use, tumbaga appears to top the list of precious metals. Yet it was secondary to gold in the perception of the Middle Sicn elite. The personal ornaments immediately surrounding the central body at Huaca Loro were all gold. The tumbaga objects were placed at the periphery of the burial chamber, and their use was probably auxiliary.

In other words, the gold objects were reserved for the personal use (including ornamentation and ritual paraphernalia) of the highest elite, whereas gilded tumbaga was used to decorate items associated with them as well as the objects used by lower-echelon elites. Tumbaga allowed those lower in status to emulate their social superiors. Gilded tumbaga with relatively low gold content would have been a most practical substitute for meeting the broad demand for rich gold-colored sheet metal.

Many of the precious metal objects found in the tomb were probably used together in public settings for ostentatious displays to impress onlookers. The full ceremonial regalia of the important person buried in the Huaca Loro tomb offers a vivid example.

Depending on the ritual to be conducted, he would have worn various headdresses--sometimes a crown decorated with sets of gold feathers or tumi-shaped ornaments, sometimes a large parabolic headdress in addition to the crown. The upper perimeter and draping sides of that parabolic headdress would have been decorated with colorful bird feathers and bangles that reached almost to the shoulders. Over his face he would have worn a gold mask. He was probably carried on a wooden litter decorated with the carved heads of mythical animals. The litter was likely to have been flanked by people waving long tumbaga fans and preceded by someone holding a staff or standard almost two meters high, which was bright and colorful with gold and feathers. With each step, each breath of air, the bangles, gold feathers and other delicately articulated metal objects would have been set in motion to create a dazzling visual and auditory effect. It is not hard to be entranced by the thought of that luminous figure--or by thoughts of what future studies of Sicán artifacts may yet tell us about that lost culture.

IZUMI SHIMADA and the late JO ANN GRIFFIN joined forces to investigate the precious metal artifacts of the Sicn. Shimada has conducted fieldwork on the north coast of Peru for the past 30 years and maintains interests in ancient technologies and the evolution of complex societies. A native of Japan, Shimada received his Ph.D. from the University of Arizona. He joined the department of anthropology at Southern Illinois University in 1994. Griffin had for almost 30 years been a goldsmith and conservator specializing in pre-Hispanic metallurgy, working with some of the largest public and private collections of pre-Hispanic gold in the world.