INTRODUCTION
It has been more than 60 years since Alfonso Caso (1932) discovered the spectacular Tomb 7 at the hilltop center of Monte Alban in the Valley of Oaxaca (Fig. 2). The excavators of this Postclassic-period tomb (see Table 1) were especially impressed by the quantity of gold and jade objects that accompanied the burials. Far less interest was shown in the shell objects. Yet, included with the more than 500 exotic ornaments in the tomb were necklaces made of hundreds of shell beads, especially red ones crafted from the spiny oyster (Spondylus). Other necklaces made of small whole shells were used to ornament breastpieces of jaguar skin. There were ornamental shell armbands, earpieces, plaques (little pieces of cut shell; Pinctada mazatlanica was the primary species) that were used in mosaics, and perforated shells that served as eyes in mosaics of turquoise. Placed on top of the tomb was an offering of jade ornaments, shell fragments, and a conch-shell trumpet (Marcus 1983a:283).
Earlier, during the Classic period dedication of the great South Platform on Monte Albán’s Main Plaza, stone boxes with nearly identical offerings were placed underneath at least three of the corners (Acosta 1958-59:27; Marcus 1983b:175-79). While each box contained a necklace of 7 jade beads, the principal contents were shells-5 large and 5 small spiny oyster shells and 10 tent olive shells (Oliva).
The Tomb 7 and South Platform finds attest to the high value that the prehispanic inhabitants of highland Oaxaca, and Mesoamerica more generally, placed on shell. Certain kinds of shell, particularly the red spiny oyster, were especially esteemed. Marine shell ornaments were traded widely, had great symbolic importance, and often were deposited in high-status contexts. Yet, because marine shell has generally been recovered as whole pieces or finished ornaments from dedicatory offerings and funerary contexts, there has been little discussion of the production of shell ornaments until recently. Relatively little is known about the technologies utilized (see Suarez 1981 for a notable exception), the range of goods produced, the species used to make specific ornaments, or the scale and context of the production activities—who the artisans were, where they worked, and what they did with their products. Recent evidence from an area of prehispanic shell working at the edge of the modern town of Ejutla de Crespo in Oaxaca is helping to change this picture.
SHELL PRODUCTION IN OAXACA
In the landlocked Valley of Oaxaca, marine shell from both the Pacific and Atlantic coasts was imported as early as the Early Formative period. In the 1970s Kent V. Flannery and Joyce Marcus documented shell working at the Formative period village site of San Jose Mogote, north of Monte Alban. Several extensively excavated houses at San Jose Mogote contained areas of 1-2 square meters littered with flint chips, chert knives and drills, fragments of cut shell, and shell ornament fragments that were broken in the process of manufacture (Flannery and Winter 1976:39; see also Parry 1987).
San Jose Mogote represents an early site for shell ornament production. Most of the shell was from the Pacific Coast, but a significant minority was imported from the Atlantic. Pearl oyster (Pinctada mazatlanica) and spiny oyster (Spondyins) were the most frequently worked species. The most common ornaments were shell pendants, both perforated whole shells and thin pieces carved in a variety of forms, and flat disk beads.
The Formative-period shell assemblage at San Jose Mogote differs somewhat from the shell recovered from later Classic and Postclassic contexts at Monte Alban. Although Pinctada and Spondylus were important species used for ornamentation at both San Jose Mogote and Monte Alban, Atlantic shell species are rare at Monte Alban. In addition, the shell-ornament assemblages vary between the two sites, with Monte Alban having a greater relative abundance of various bead forms and multiple-piece mosaics. These differences raised questions concerning the nature of shell exchange and ornament manufacture in the Valley of Oaxaca during the later prehispanic periods. Given the lack of known shell ornament production sites in Oaxaca for the Classic and Postelassie periods, we also wondered how shell working in these later periods may have differed from shell working in the earlier period represented by San Jose Mogote.
EJUTLA RESEARCH
The hundreds of archaeological sites located and mapped during this recent Ejutla Valley survey (Feinman and Nicholas 1990) have added greatly to_ our knowledge of the region. One of the largest and moist impressive sites is the prehispanic settlement situated beneath the modern town of Ejutla de Crespo, where Diguet had noted the tomb. Although the site has been disturbed by modern occupation, several 10-meter-high prehispanic mounds are still visible in the center of town (Fig. 4). The site was occupied from the Late Formative through the Postelassie; however, the major phase of occupation was the Classic period.
During the survey we discovered a dense concentration of shell debris in several plowed fields at the eastern edge of Ejutla de Crespo. Such finds are rare in landlocked highland areas. Artifact collections from this several-hectare area included shell fragments with obvious signs of work (such as cut marks and perforations), several broken and unfinished shell ornaments, and an unusual abundance of heavily worn stone tools, including obsidian blades. The most well-represented shell taxa on the surface were varieties that generally were used for ornamentation rather than food in prehispanic Mesoamerica.
In 1990, we initiated the first of four seasons of intensive field study at Ejutla. We focused our attention on the area where dense surface concentrations of shell had been encountered previously and on the shell itself, hoping to find answers to a number of questions. What range of items were made and what species of shell were used? What was the nature of the technology? At what scale and in what social context was the craft carried out? Were the items made for local use or were they traded to the neighboring Valley of Oaxaca?
In 1990 and 1991 we concentrated primarily on the excavation of midden deposits composed largely of shell debris, broken pottery, and obsidian and chert tools (Fig. 4). In 1992 and 1993 we focused on the exposure and definition of a nearby prehispanic structure (Fig. 3), which included a small sub-floor tomb where four individuals and a dog were interred. During the course of the investigation, at least four ceramic firing areas (pit kilns) were also excavated. In total, a significant proportion of a prehispanic household unit has been excavated, which includes residential and work areas, the pit kilns, and associated midden areas.
Based on surface observations, it is not surprising that we found remains of shell ornament manufacture. However, archaeological indicators for several other craft activities, including ceramic vessel and figurine manufacture, lapidary arts, and possibly spinning, also were recovered (Feinman et al. 1993). Even though we do not detail these other craft activities here, they help place shell-working at the site in a broader context.
SHELL WORKING AT EJUTLA
In total, more than 24,000 pieces of marine shell have been collected at the Ejutla site. Roughly 5 percent are finished or partially finished ornaments (Fig. 5) or small unmodified whole shells that could have been perforated to be strung as ornaments. An additional 35 percent show very clear indications of modification, such as drilling, string-cut surfaces and edges, and abraded surfaces (Fig. 6). The rest of the shell material consists of broken pieces of varying sizes and minute pieces of chipping debris. No complete shells from large marine species were recovered. Yet the wide range of shell parts represented in the debris indicates that most of the shell was brought into Ejutla as whole (or nearly whole) specimens, and that primary breakage and working of the shells occurred on site.
More than 90 taxa of marine shell have been identified at the Ejutla site; however, only 7 genera account for 95 percent of the identifiable shell. Four of these genera are bivalves (pelecypods): nacreous pearl oysters (Pinctada), jewel boxes (Champ), spiny oysters (Spondylus), and ark shells (Anadara); three are snails (gastropods): giant limpets (Patella), limpets (Acmaea), and conch shells (Strombus). Across prehispanic Mesoamerica, each of these taxa was used ornamentally (and not generally for food). In the Ejutla excavations, the genus Pinctada alone accounts for approximately 60 percent of the recovered shell by weight.
Most of the Ejutla marine material is from larger species (Strombus, Pinctada, Patella, Spondylus), the shells of which were cut and shaped to make various ornaments. Many of the taxa present in low quantities were small gastropods, including olive shells (Agaronia, Oliva, and Olivella), nutmegs (Cancellaria), horn shells (Cerithidea), cowries (Cypraea), keyhole limpets (Fissure), sea buttons (Jenneria), periwinkles (Littorina), marginellas (Marginalia), nerites (Nerita), dye shells (Thais), and turret shells (Turritella). These shells were generally perforated and strung whole. Almost the entire corpus of Ejutla shell represents Pacific varieties. Only two Atlantic species were present, including one Cypraea cinerea and several Marginalia apicina.
SHELL ORNAMENTS
The prehispanic artisans of Ejutla crafted a range of ornaments from the Pacific Coast shell (Fig. 7). These forms included small plaques that were used in mosaic inlay, disks, beads, pendants, bracelets, and some miscellaneous forms of unknown use. In many cases, certain ornaments were made from only one or a few kinds of shell. Nacreous pearl oyster (Pinctada mazatlanica), the most abundant shell species at the Ejutla site, was used to make the most commonly encountered ornaments, small plaques for mosaics (Table 2, Fig. 8). Many shell disks were cut from Pinctada, but depending on the technique used (see below), they were also made from conch and other non-nacreous shell. While bracelets usually were cut from giant limpets (Patella ?Mexicans), a small percentage were crafted from Pinctada. Beads and pendants were made from a variety of shell species, but rarely from Pinctada.
Pearl oyster plaques were cut into a variety of shapes and sizes. Rectangular, trapezoidal, and triangular shapes were especially common. The more finished of these pieces had very straight, smooth cut edges (Fig. 9). These pieces could have been used for mosaic inlays, such as those recovered from Tomb 7 at Monte Alban, or sewn onto cloth. Some cut nacreous shell may have been used as incrustations to decorate the teeth of ceramic figurines and urns.
A much less common type of plaque was cut from the outer wall of certain large gastropods. These non-nacreous pieces were always in the form of small triangles (Fig. 10), and may also have been intended for mosaic inlay.
After small plaques, the second most common ornaments were shell disks (Fig. 11). Few entirely finished disks were found. Some circular disks were perforated, evidently to be used as beads, but most were not. Disks may also have been employed in the construction of elaborate mosaics, such as those of shell and turquoise found in Tomb 7.
Beads, both finished and unfinished, were the third most frequently recovered ornaments. In the shell from Tikal, a distinction has been drawn between “natural” and “formed” beads (Moholy-Nagy 1989:141). If the shape of the original shell is still identifiable, then the head is considered natural. In contrast, if the shape of the shell has been obliterated in the shell-working process, then the bead is considered formed. The Ejutla beads generally fit Moholy-Nagy’s second category, with types including tiny spherical beads, larger cylindrical ones, and tubular beads (Fig. 12) (see Moholy-Nagy 1989 for similar types). At Ejutla, the majority of small beads were made from Spondylus or Chama, while large beads were crafted from Strombus and other large gastropods.
Pendants, bracelet fragments, and blanks were found with less frequency in the Ejutla collections. Themajority of the pendants were formed by cutting small tabular pieces (Fig. 13) from the walls of large shells (Spondylus shells were the most frequently identifiable taxa; see Moholy-Nagy 1989:141). Other pendants were made by perforating whole gastropod shells (especially Oliva; see Fig. 14). While many bracelet fragments (Fig. 15) and debris from bracelet manufacture were recovered, no complete bracelets were found during the excavations.
SHELL-WORKING TECHNIQUES
Obsidian and other chipped stone tools were found in close subsurface association with the shell debris, particularly in midden deposits. These tools most likely were used to craft shell ornaments. Thousands of tiny stone flakes, which appear to have been produced during use and retouching of the tools, were recovered from excavated midden strata where we found the densest shell debris.
The obsidian at the site included many heavily worn blades. These spent blades would have been effectively dulled by repeatedly cutting and working the hard, abrasive shell. Chert artifacts were also abundant, coMmonly in the form of small, solid micro-drills that have been linked to the perforation of beads and pendants elsewhere in Oaxaca (Parry 1987) and in other parts of the Western Hemisphere (e.g., Mester 1985:107; Yerkes 1989:115).
Careful analysis of the shell artifacts and debris indicates that perishable materials—string and, probably, cane—were also used to modify shell in spite of its hardness. Many of the small plaques and disks were made using these materials. The small tabular shapes frequently were cut from the walls of large nacreous shells using string in conjunction with water and an abrasive such as sand. This method often leaves a small lip on the bottom surface where the shell snaps before being cut completely through. This lip was smoothed away on more finished plaques, but remained in evidence on less finished examples. Many pieces of shell debris at Ejutla also had very smooth string-cut edges or incomplete string cuts across one or more surfaces. In some instances, the nature of these cuts allows us to determine the sequence or steps of manufacture.
Hollow tubular drills, most likely made of cane, were utilized to extract small circular disks from larger pieces of shell (see Fig. 11). The average diameter for the drilled disks was 11 millimeters; more than three-quarters were between 8.5 and 12.5 millimeters. The nacreous disks were cut from large Pinctada shells using these hollow drills. As with string-cut shell, a small lip often remains on the side away from the initial cut.
Another prevalent technique used to make ornaments was to abrade the edges of small pieces of shell against a hard surface to create a smooth rounded edge (Suarez 1981: lamina 14). While many shell disks were made using the hollow drill, a greater number of disks, from both nacreous and hard matte-white shell, were formed by this abrading process. Not surprisingly, the variation in diameter was much greater for the abraded disks than the drilled disks. Abrasion was also used to make most of the shell beads. Many river cobbles were found in the midden and in association with the structure, and these often bore linear marks from abrasion wear.
The production process was not the same for all beads. The crafting of large, thick, formed beads involved more steps and a more labor-intensive process than the production of whole-shell or flat disk beads. To manufacture the more cylindrical beads, an artisan would have to cut a thick piece of matte shell (using string) from either the lip or cohunella of a large gastropod or the thick upper wall of certain large pelecypods and then abrade the shell into the desired shape (Suarez 1981: laminae 8, 9, 12). The risk of failure during perforation also was greater for the thicker, formed beads. Perhaps due to this risk of breakage, the large cylindrical beads generally were perforated before effort was devoted to final shaping and polishing. In contrast, holes were drilled into smaller, miniature, and flat beads after they were finely shaped and smoothed.
CONCLUDING THOUGHTS
The shell workers at Ejutla used a variety of techniques to make a diverse range of ornaments, including small plaques, disks, beads, pendants, and bracelets. The majority of shell ornaments recovered were either unfinished or broken, as one might expect in a production context. Likewise, most of the other shell recovered was chipping debris and discarded pieces, the by-products of the manufacturing process.
The Ejutla artisans were involved in a range of craft activities. In some cases similar techniques appear to have been used on different classes of materials. For example, in the midden adjacent to the structure we found onyx drill plugs that would have been the byproducts of using a hollow tubular drill to perforate stone. The diameter of most of the plugs was comparable to that of the majority of shell disks that were cut with a tubular drill. Although no finished complete stone ornaments were recovered in the Ejutla excavations, the same hollow-drill technique that was used to fashion the shell disks could have been employed to produce lapidary craft items, possibly small bowls (see Diehl 1983:101-2) or earspools. Involvement in multiple crafts may have enabled this household of Ejutla artisans to concentrate more heavily on non-agricultural production than would have been possible if they focused exclusively on shell or any other single craft.
The excavated structure may not have been the only one involved in multiple craft activities. The intensive study of the Ejutla site located a 5-hectare area with surface shell of which only a small portion was excavated. Included in this larger area was at least one dense concentration of shell artifacts, onyx drill plugs, and building stone that appears to be the location of a second residential structure devoted to a range of craft activities.
Shell ornament manufacture in Classic-period Ejutla was practiced in a household context, as it was in San Jose Mogote during the Formative period. Yet in contrast to San Jose Mogote, shell working in Ejutla appears to have been practiced at a higher intensity and in conjunction with a range of other crafts, and seems to have focused almost exclusively on Pacific marine species.
The volume of shell debris at Ejutla eclipses that found at other known sites in highland Oaxaca. Yet the low volume of finished shell artifacts at the site in general, and particularly on the house floor and in the tomb (where only a single shell bead was recovered), indicates that the majority of the finished ornaments were not consumed by their producers. Given the general rarity of evidence for shell working elsewhere in the valleys of Ejutla and Oaxaca (the sole exceptions are San Jose Mogote [Flannery and Winter 1976; Marcus 1989] and Monte Alban [Blanton 1978]), it seems likely that shell ornaments crafted in Ejutla were transported to other surrounding highland settlements.
Many of the ornament forms and shell species found in Ejutla are strikingly similar to those found at Monte Alban. Although we cannot definitely demonstrate that specific shell artifacts recovered at Monte Albán were made in Ejutla, these artifactual similarities and the rarity of other known shell production sites in the vicinity makes this hypothesis tenable. Firmer assessment, however, awaits a fuller inventory and analytical study of the Monte Alban shell assemblage.
ACKNOWLEDGMENTS
The National Science Foundation (BNS 8919164, BNS 91-05780, and SBR-9304258 to the senior author) has provided the principal support for the Ejutla excavations. Additional funding has been generously awarded by the H. J. Heinz Foundation, the National Geographic Society, the Graduate School of the IJniversity of Wisconsin-Madison, and Arvin B. Weinstein. The Instituto Nacional de Antropologia e Historia has provided essential permissions and support. We also wish to thank the members of our 1990-1993 field and laboratory crews for their diligent efforts and endurance.