OBSIDIAN AND FORMATIVE POLITICAL ECONOMIES

From the Formative period on, obsidian was the most widely used material for making cutting and piercing implements in Mesoamerica, in addition to being used for finely crafted jewelry, insignia, ritual paraphernalia, and other lapidary items. In most parts of Mesoamerica, obsidian tools and debitage also constitute the largest class of lithic artifacts, the exceptions occurring in places such as parts of Oaxaca and the Maya lowlands where high-quality chert was available. Natural occurrences of obsidian, however, are restricted to those cases where rapid cooling of magma or lava flows resulted in an amorphous, glassy structure with little to no crystal development. As is well known, conditions for the production of voluminous obsidian deposits are unusual, even in volcanic terrains. For example, the volcanic Tuxtla Mountains that rise from the southern Gulf lowlands of Mexico produced extensive basalt flows but no obsidian deposits. As a result, exploitable obsidian deposits are restricted to a limited (though not small) number of discrete sources. Currently, about 40 obsidian sources are distinguished within Mesoamerica. These sources lie in two large regions of volcanism, one containing some 28 sources stretching from Jalisco and Nayarit through central Mexico to Veracruz and the other in the Guatemala highlands (eight sources) and nearby Honduras (one source, at Lake Yojoa) (Cobean Reference Cobean2002:Figure 1.1). Other sources are known beyond the geographic limits of Mesoamerica in north and west Mexico, and it is possible that as-yet unexplored sources exist in Oaxaca, Guerrero, and/or Chiapas (Cobean Reference Cobean2002:31).

Outside of these source zones, obsidian had to be acquired as an exotic good, often over hundreds of kilometers, along routes that ran through the territories of multiple polities, either by direct procurement, through some form of exchange, or as tribute. The use of obsidian for a variety of utilitarian, military, and symbolic functions; the technological skill required to produce prismatic blades and lapidary items; the wide distribution of obsidian beyond source zones; and the limited geographic occurrence of sources offer multiple possibilities for elite involvement in and control over obsidian production and distribution systems. Consequently, obsidian has figured in explanations for the Formative period emergence of administrative hierarchies and political elites for decades, whether framed within an adaptationist model (in the sense of Brumfiel and Earle Reference Brumfiel, Earle, Brumfiel and Earle1987; but see also Cobean et al. Reference Cobean, Coe, Perry, Turekian and Kharkar1971; Pires Ferreira Reference Pires-Ferreira and Flannery1976; Rathje Reference Rathje1971, Reference Rathje and Leone1972) or in more strictly political terms (for example, Clark Reference Clark, Johnson and Marrow1987).

For the Formative period Gulf lowlands, Rathje (Reference Rathje and Leone1972) argued that sociopolitical hierarchies developed among the Olmecs in response to the need to organize long-distance trade for utilitarian goods not available locally, including obsidian, basalt, and salt. Coe and Diehl (Reference Coe and Diehl1980b:148) likewise argued that control over the exchange of obsidian and basalt played an important (but not exclusive) role in supporting elites, stating, “Once these imports arrived at San Lorenzo, the elite redistributed them to their kin, festival guests, and allies. The effects of withholding these essentials would have been realized by both the givers and the recipients, and this inevitably led to the formation and consolidation of economic power by the San Lorenzo elite.” Clark (Reference Clark, Johnson and Marrow1987) later observed that, while obsidian exchange and expedient reduction technologies involving direct percussion and/or bipolar reduction were widespread, blade technology was adopted on a region-by-region basis and appeared to be associated with chiefly centers. Clark noted that there was little evidence for Early Formative exchange of blade cores, including at San Lorenzo, where it appeared finished prismatic blades were imported as luxury goods used to reinforce chiefly status. In contrast, Middle Formative chiefs at La Libertad and in the Central Depression of Chiapas adopted blade core technology and patronized the craft, distributing blades to supporters. Clark (Reference Clark, Johnson and Marrow1987:283) concluded that the spread of prismatic blade technology required the achievement of a threshold of organizational complexity comparable to that of a complex chiefdom, as well as sufficient political motivation by elites to sponsor the production of blades and the acquisition of suitable material.

Describing Formative period obsidian procurement in the Mixtequilla, Stark and associates (Reference Stark, Heller, Glascock, Elam and Neff1992:231) suggested that relatively open community relations and down-the-line exchange accounted for the greater diversity of Formative period obsidian sources but hinted that the general adoption of prismatic blades in the Classic period and the shift to one dominant source zone (Zaragoza-Oyameles) was consistent with achievement of Clark's organizational threshold, albeit in a later time frame. Santley and associates (Reference Santley, Barrett, Glascock and Neff2001:59) indicate that prismatic blades occur in low frequencies in Formative assemblages in the Tuxtlas. They do not specify whether those blades were assigned to the Early, Middle, or Late Formative period, but Santley (Reference Santley2007:Figure 3.3) reports the presence of prismatic blades, macrocore reduction debitage, and pressure cores from Early Formative contexts at Matacapan, which at the time consisted of a small village and several hamlets. There is some question as to whether the Olmec ceramic assemblage at Matacapan is more indicative of the Early or Middle Formative period, but in either case, the assemblage offers a contrast to the pattern described by Clark and would suggest that prismatic blade production in some cases escaped elite control. Santley and associates (Reference Santley, Barrett, Glascock and Neff2001:59) nevertheless posit individual exchange partnerships between the source region's producers and clients, with an (apparently) minor role for local “elites” in managing the network of exchange partners.

Evidence from beyond the southern Gulf lowlands supports some degree of elite involvement in the procurement of strategic resources. Discussions of Formative interregional interaction underscore the extensive exchange of obsidian and frequently posit obsidian procurement in combination with other materials (especially greenstone and iron ore), as an incentive for Gulf Olmec involvement in other regions (Hirth Reference Hirth1978; Pool Reference Pool2007:216; Tolstoy Reference Tolstoy, Sharer and Grove1989). The Early Formative site of Cantón Corralito offers a case in which Olmecs may have established an enclave at an important trade node (Cheetham Reference Cheetham2010), although the degree to which heartland elites may have controlled the enclave's activities is uncertain (Pool Reference Pool2007:192–194). The themes and distribution of Middle Formative Olmec-style monuments and rock carvings, especially, suggest the formation of alliances between Olmec elites and those of polities along routes to the sources of these materials (Castro and Cobean Reference Castro Lainez and Cobean1996; Doering Reference Doering2002; Grove Reference Grove and Grove1987, Reference Grove and Rice1993; Pool Reference Pool2007:240). Such efforts evoke Hirth's (Reference Hirth1996:224) “matrix control” principle in which elites seek to control networks by placing themselves at key positions to control the production, accumulation, and flow of resources—they may also have sought to create social contexts (marriages, feasts, alliance renewals, etc.) that promoted the accumulation and exchange of goods like obsidian.

INFERRING CONTROL

How do we infer elite control? As this brief review of research on Formative obsidian production and exchange suggests, we must first ask, control over what? Not just control over which materials or products, but control over which aspects of the procurement-production-distribution sequence. We may ask that question broadly, with respect to control over source regions, production technology, transport, or the formulation of social and religious sanctions, but each of these encompasses particular sets of knowledge, skills, tools, and social relations that may offer opportunities for control or its subversion. Another approach is to analyze political economies in terms of how they applied general principles or strategies, such as those described by Hirth (Reference Hirth1996) as the accumulation, context, matrix-control, and ideological principles; the sets of “corporate” and exclusionary strategies discussed by Blanton et al. (Reference Blanton, Feinman, Kowalewski and Peregrine1996); or the staple and wealth finance systems proposed by D'Altroy and Earle (Reference D'Altroy and Earle1985). Such frameworks provide useful insights into how political actors may have employed economic relations toward political ends, but they still hinge on empirical demonstration of control over particular components of the economic system.

Demarest (Reference Demarest2014) makes the important point that the conceptual distinctions among “control,” “supervision,” and “participation” must be maintained, with control implying power over actors, supervision involving oversight and direction of participants, and participation involving the physical execution of some aspect of the activity. We would add that these are distinctive but not mutually exclusive categories. Thus supervision is one way by which control may be exercised, although supervisors themselves may respond to the orders of superiors. Most arguments for elite control over production come down to a demonstration of the proximity of craft production facilities to elite residential or administrative structures, and by inference elite or state supervision of those activities, as in discussions of attached, as opposed to independent, specialization (Brumfiel and Earle Reference Brumfiel, Earle, Brumfiel and Earle1987; Costin Reference Costin and Schiffer1991; Pool Reference Pool and Pool2003; but compare with Hirth Reference Hirth2008:443–444).

Some words of caution regarding propinquity arguments are necessary, however. First, physical proximity may argue strongly for supervision over the activities conducted in a facility, as well as control over access to the facility, but unless producers resided at such attached production locations, it's hard to say what they did while they were away; whether they worked similar materials or created similar products elsewhere, or whether their labor was mobilized seasonally or intermittently for production at those locations. Second, identifying one or a few cases of physically attached production does not necessarily imply that elites controlled all such production, as John Clark (Reference Clark and Hirth2003:53) has observed. At Tres Zapotes, for example, we have found evidence for contemporaneous crafting of ceramics, basalt, and obsidian in elite and nonelite contexts (Pool Reference Pool and Hirth2009). Third, a lack of proximity does not necessarily imply an absence of control. Indeed, elite power is all about increasing the ability to exercise control over more people at ever-greater distances, and it is not always necessary that clients make their crafts under a patron's watchful eye for such control to be effective.

Some have therefore argued that the production of highly crafted, labor intensive luxury or wealth items in nonelite residential contexts also constitutes attached specialization (Lewis Reference Lewis1996). On the other hand, we have identified cases at Tres Zapotes in which quotidian items were produced in facilities physically attached to elite structures (Pool Reference Pool and Pool2003). Here the problem is one of disentangling the product and the location of production from control over production. Doing so is important if we are to recognize novel arrangements with respect to production and patronage; distinguishing supervision as one means of control moves us in that direction. All of this is to say that, though propinquity may provide a strong argument for elite supervision of that production unit when we find it, it is certainly incomplete. To argue that elites controlled production of a class of goods by locating producers under their gaze (or producing items themselves as, for example, has been argued for Aguateca, Guatemala [Aoyama Reference Aoyama2007; Emery and Aoyama Reference Emery and Aoyama2007]), we must also look for (and fail to find) evidence for production of those goods in nonelite (residential or nonresidential) contexts.

The exercise of control also involves determining rights of allocation—that is, who has access to a good and how much. For valuable items, it is reasonable to posit that those in control of valuables will have greater access to them, and greater access will be reflected by proportionately greater representation of the good in contexts associated with that segment of society. In essence, this is the basis for distributional arguments for elite control over a resource: elites will have more and better examples of a valued good than will nonelites, with value being determined by overall scarcity, exoticness, labor expenditure, and such subjective characteristics as quality of material and skill of execution.

Keeping these considerations in mind in our evaluation of the role of obsidian in the political economy of Tres Zapotes and the southern Gulf lowlands, we expect elite control over obsidian procurement to result in one or more of the following patterns:

1. (1) If elites exclusively controlled obsidian procurement, the number of sources represented at centers should equal or exceed the number at hinterland sites (compare with Hirth Reference Hirth2008:443–444).Footnote ¹ Demonstration of greater source diversity in centers is not, however, conclusive demonstration of elite control because inhabitants of centers could also have access to more sources through periodic fairs or incipient markets. If procurement was household-based, we might also expect a greater diversity of sources due to the aggregate effect of individualized procurement arrangements among a larger populace.

2. (2) Where particular technologies require specific material characteristics, elite control may be expressed in monopolization of these “higher quality” (most suitable) resources for elite use. For example, for prismatic blade production on obsidian, an amorphous texture lacking crystalline inclusions or internal fracture planes is frequently cited among “high quality” characteristics.

3. (3) Under classic models of chiefly redistribution, obsidian pooled by elites at centers and redistributed to lower-order centers and hinterland sites should exhibit similar proportions of different sources within sites or intrasite contexts of similar rank (Hirth Reference Hirth2008:444; Pires-Ferreira Reference Pires-Ferreira and Flannery1976).

4. (4) It follows from (1), (2), and (3) that the obsidian sources present in villages, lower-order centers, and nonelite residences should be a subset of those present in elite contexts at higher-order centers, again, if classic redistribution models apply.

5. (5) Supervision of a specific obsidian technology would be supported by the presence of production debitage characteristic of that technology in elite contexts and its absence in nonelite contexts (although we acknowledge that elite control may also be exercised at a distance). To the degree that specific sources are favored for a particular reduction technology, we would expect this pattern to appear for some sources and not others.

A reduced degree or absence of elite control should be reflected in divergences from these patterns. In particular, if elites are unable to effectively control procurement-distribution networks at a particular scale (supraregionally, regionally, or within their hinterlands), we would expect greater variation in the number of sources and relative frequencies of specific sources, the number of sources would not be a simple function of site rank or intrasite social rank, hinterland sites and domestic contexts might gain access to alternative obsidian sources not represented in centers or elite contexts, and production residues associated with specific obsidian technologies will not be exclusive to elite contexts.

ARCHAEOLOGICAL BACKGROUND FOR TRES ZAPOTES

Tres Zapotes is famous as the site of the first discovery of a colossal Olmec head (Melgar Reference Melgar1869, Reference Melgar1871) and the second earliest Long Count inscription in Mesoamerica (Stirling Reference Stirling1939, Reference Stirling1940). The modern study of Formative period Gulf Coast archaeology was launched there in 1939 by Stirling's (Reference Stirling1943) joint National Geographic–Smithsonian Institution expedition (see also Drucker Reference Drucker1943; Weiant Reference Weiant1943). Archaeological survey, surface collection, and excavation we have conducted at Tres Zapotes since 1995 have documented the following site history, spanning two millenia.

Tres Zapotes was founded in the Early Formative period, growing to a village of 17 ha by 1000 cal b.c. (Pool et al. Reference Pool, Ceballos, Rodríguez and Loughlin2010). Over the course of the Middle Formative period (1000–400 cal b.c.), the site center grew to 80 ha, surrounded by another 70 ha of adjoining settlement concentrations. At least 11 Olmec monuments were set at Tres Zapotes during the Middle Formative period, including two colossal heads, two stelae, five figures in the round, two tenoned busts, and a carved serpentine column (Pool Reference Pool, Clark, Guernsey and Arroyo2010). The number of monuments, their thematic content, and the areal extent of Tres Zapotes, second only to La Venta among Middle Formative Gulf Olmec sites, marks it as a regional Olmec center (Pool and Ortiz Ceballos Reference Pool, Arnold and Pool2008). The sculptural focus on the ruler as an individual, the arrangement of monuments in a way that emphasizes the centralization of political power, and greater use of exotic greenstone in ritual displays and mortuary programs suggest that polity governance and the political economy emphasized exclusionary principles (Pool Reference Pool, Clark, Guernsey and Arroyo2010; see also Blanton et al. Reference Blanton, Feinman, Kowalewski and Peregrine1996).

Alone among Gulf Olmec centers, Tres Zapotes continued to flourish in the Late Formative period, expanding to cover 500 ha. Major programs of formal mound construction were initiated at the Epi-Olmec center, focused on four widely dispersed complexes, each of which replicated a basic layout consisting of a plaza oriented east-west, bounded on the north by a long, loaf-shaped, administrative/elite residential mound and on the west by a pyramidal temple mound, with a low mound on the centerline of the long axis of the plaza; other mounds were added as the complexes grew (Pool Reference Pool, Arnold and Pool2008). The functional and formal redundancy of the architectural layouts, the lack of a single, clearly dominant center, and the chronological overlap in their construction and use suggest that political power was shared among faction leaders (Pool Reference Pool, Arnold and Pool2008). Additionally, a thematic shift in stone monuments away from an individualizing emphasis on the ruler suggests that governing strategies at the polity level emphasized collective (or “corporate” [Blanton et al. Reference Blanton, Feinman, Kowalewski and Peregrine1996]) as opposed to exclusionary themes (Pool Reference Pool, Clark, Guernsey and Arroyo2010). In concert with these changes (whether as cause or consequence) we would expect the Late Formative political economy of Tres Zapotes to deemphasize the monopolization of prestige goods and to broaden access to strategic resources. These expectations are met in a reduction in the prevalence of exotic materials other than obsidian in our excavations and the redundant production of utilitarian craft items in elite and nonelite contexts (Knight Reference Knight and Pool2003, 2012; Pool Reference Pool and Pool2003).

During the Protoclassic period (a.d. 1–300), architectural layouts at Tres Zapotes diversified through a combination of new construction, shifts in mound function, and resetting of associated monuments—trends which continued in the Classic period (Pool Reference Pool, Arnold and Pool2008, Reference Pool, Clark, Guernsey and Arroyo2010). As factional leaders began to distinguish themselves through novel architectural programs at Tres Zapotes, leaders at settlements like El Mesón, Alvarado, and La Mojarra reemphasized exclusionary themes of dominance and individualizing claims to rulership (Loughlin Reference Loughlin2012; Pool and Loughlin Reference Pool and Loughlin2006). Although Tres Zapotes retained its regional prominence and most of its areal extent in the Protoclassic period, it began a slow decline, falling to about 200 ha by the beginning of the Classic period around a.d. 300. By a.d. 900 Tres Zapotes lay abandoned, to be reoccupied only briefly in the Early Postclassic period (Drucker Reference Drucker1943; Pool Reference Pool1995).

The location of Tres Zapotes, on the ecotone between the Papaloapan drainage system to the west and the Sierra de los Tuxtlas to the east, was ideally suited to take advantage of both riverine and montane resources. One montane resource, basalt, is plentiful in the Tuxtlas and occurs in fine-grained varieties that can produce a sharp edge when flaked. Kruszczynski's (Reference Kruszczynski2001) survey of basalt exploitation in the area, however, indicates that this common local source was used for different tasks from obsidian. At the same time, cryptocrystalline alternatives, such as chert, are scarce in the southern Gulf lowlands. As a result, the inhabitants of Tres Zapotes relied almost exclusively on imported obsidian throughout the Formative and Classic periods for all cutting and scraping activities. Since the valleys and streams of the Ríos Papaloapan, Salado, and Grande to the west of Tres Zapotes provide the most direct and gradual route between the southern Gulf lowlands, the Basin of Mexico, and the Valley of Oaxaca (Drennan et al. Reference Drennan, Fitzgibbons, Dehn and Isaac1990:197), Tres Zapotes was better positioned than most other sites in the southern Gulf lowlands to acquire obsidian from sources in the central and western highlands of Mexico.

The three closest and most commonly utilized obsidian sources for the southern Gulf lowlands are those at Pico de Orizaba, Guadalupe Victoria, and Zaragoza-Oyameles, all located 150–200 km to the northwest along the eastern edge of the Trans-Mexican Volcanic Belt, straddling the Veracruz/Puebla border (Ferriz Reference Ferriz1985, Siebert and Carrasco-Núñez Reference Siebert and Carrasco-Núñez2002). Surface and excavation data from the Mixtequilla (Stark et al Reference Stark, Heller, Glascock, Elam and Neff1992) indicate that obsidian from Guadalupe Victoria and Pico de Orizaba were the most common sources used there during the Early and Middle Formative periods, while Zaragoza-Oyameles was most common in the Protoclassic and Classic periods, coinciding with the establishment of Cantona as a regional center near the Zaragoza-Oyameles source area in eastern Puebla (García Cook Reference García Cook, Sanders, Mastasche and Cobean2003, García Cook and Merino Carrión Reference García Cook and Carrión1998). A similar diachronic pattern was observed in a nonrandom sample of excavated obsidian from the Sierra de los Tuxtlas (Santley et al. Reference Santley, Barrett, Glascock and Neff2001). In both samples, however, materials also were identified from sources in the central and western Mexican highlands at Paredón, Otumba, Zacualtipan, and Pachuca and from the highland Guatemalan source of San Martín Jilotepeque. Prior to the study reported here, the degree to which Tres Zapotes participated in these and similar provisioning networks through time was uncertain.

Whether Tres Zapotes was provisioned with obsidian from the three closest sources or from farther afield, it had to be transported over hundreds of kilometers from sources in the Mexican and Guatemalan highlands. As such, the energy involved in the transportation of obsidian to the southern Gulf lowlands—and the reduction technologies embedded in the commodity—may have imbued it with a value whose control could have been beneficial to local elites. It is within this framework of political economy that we investigate the role of obsidian in the models of centralization and confederacy at Tres Zapotes.

LITHIC TECHNOLOGY AND OBSIDIAN SOURCING AT TRES ZAPOTES

The Tres Zapotes Archaeological Project (PATZ) investigated a sequence of site occupation spanning some 2,000 years, from the Early Formative to the Early Classic period (1400 b.c.–a.d. 600) (Pool Reference Pool2005). As noted above, Late Classic and Postclassic occupations also have been identified at Tres Zapotes but were not a focus of this study. The project excavated 41 stratigraphic units grouped into 12 Operations and Suboperations, sampling all formal plaza groups and three domestic areas with evidence of craft production. This resulted in a chipped-stone assemblage of 5,728 artifacts, with all but 15 made from obsidian. Industries represented in the assemblage consisted almost entirely of bipolar and core-blade reduction technologies. Less than 1% of the assemblage was the product of bifacial reduction technology. Of the 5,728 artifacts recovered, 29% (n = 1,636) came from civic-ceremonial and elite contexts and 27% (n = 1,562) came from nonelite, domestic contexts. The remaining 44% of artifacts came from the plowzone, from mixed contexts, or from contexts with questionable temporal affiliation. All obsidian artifacts collected so far have been analyzed by Charles Knight, ensuring consistency in visual characterization and technological classification.

Following established procedures for the Tuxtla region, Knight's visual characterization of artifacts distinguished black, gray, clear, and green obsidian, as well as finer variations in color, texture, and fabric referred to as “subcolors” (Barrett Reference Barrett2003; Santley et al. Reference Santley, Barrett, Glascock and Neff2001).Footnote ² The Olmec epoch (Early and Middle Formative periods) was dominated by clear obsidian. Olmec reduction technology at Tres Zapotes was principally bipolar, but core-blade reduction became increasingly prevalent in both elite and nonelite contexts with time (Knight Reference Knight2009; Knight and Pool Reference Knight and Pool2008). In the Early Formative period, prismatic blades were made from black obsidian and were limited to elite contexts. By the end of the Middle Formative period, prismatic blades and their production indicators occurred in both elite and nonelite contexts, in clear and, to a much lesser degree, black obsidians. Without chemical analysis it was difficult to say if the Early and Middle Formative clear and black obsidians represented an “international” variety of sources, such as those identified at San Lorenzo (Cobean et al. Reference Cobean, Coe, Perry, Turekian and Kharkar1971, Reference Cobean, Vogt, Glascock and Stocker1991; Hirth et al. Reference Hirth, Cyphers, Cobean, de León and Glascock2013), or if they represent more local or “provincial” obsidian exchange networks centered on the Guadalupe Victoria, Pico de Orizaba, or Zaragoza-Oyameles sources of eastern Puebla and western Veracruz.

Core-blade technology became a relatively common Epi-Olmec technology in the Late Formative period before the major exchange network in black obsidian was established, which source analyses elsewhere in the southern Gulf lowlands suggest was from the Zaragoza-Oyameles source (Esquivias Reference Esquivias2003; Knight and Glascock Reference Knight and Glascock2009; Santley et al. Reference Santley, Barrett, Glascock and Neff2001; Stark et al. Reference Stark, Heller, Glascock, Elam and Neff1992). By the Protoclassic period, black obsidian predominated in commoner contexts where core-blade technology all but replaced bipolar technology. Visual determinations suggest the existence of significantly different obsidian procurement strategies in the Protoclassic period from previous occupation periods and between elite and nonelite contexts. One such system may be that which provisioned a single nonelite locus of the site with polyhedral cores exhibiting ground and scratched platforms, a technology heretofore unseen at the site (Knight Reference Knight2009). All examples with this platform treatment are black and may have come from one source, possibly as a single core. If, however, they came from multiple sources, the platform grinding would have had to occur at some common distribution point away from the sources, perhaps by itinerant craftsmen. Evidence for the removal of these blades from cores with ground platforms exists at Tres Zapotes, but no direct evidence for platform grinding has been recovered, such as platform preparation flakes, pecking tools, or stone slabs with characteristic abrasion marks (see Hirth and Flenniken [Reference Hirth, Flenniken, Hirth and Andrews2006] for examples).

PREVIOUS OBSIDIAN STUDIES AT TRES ZAPOTES AND IN ADJACENT AREAS OF THE SOUTHERN GULF COAST

Over the course of the Formative period, the southern Gulf lowlands were provisioned by at least 13, and possibly as many as 16, obsidian sources (Table 1). The number of sources utilized varied greatly over time and space, however. At San Lorenzo, for example, 11 sources have been identified in the Early Formative assemblage (not including two still “unidentified sources” and “unknown” sources [Cobean et al. Reference Cobean, Coe, Perry, Turekian and Kharkar1971, Reference Cobean, Vogt, Glascock and Stocker1991; Hirth et al. Reference Hirth, Cyphers, Cobean, de León and Glascock2013]). At San Andrés, Tabasco, Formative deposits yielded obsidian from a total of nine sources, with only two sources assigned to the Early Formative period (Doering Reference Doering2002). For Formative settlements in the Tuxtla Mountains, Santley and associates (Reference Santley, Barrett, Glascock and Neff2001) identified six sources, with three to four being utilized in any one period. In the south-central Gulf lowlands, Stark and associates (Reference Stark, Heller, Glascock, Elam and Neff1992) identified six known sources in the Formative obsidian assemblage of two low mounds (Mounds 693 and 985) in the Mixtequilla area, as well as a possible additional source.

Table 1.

Formative period obsidian source utilization in the southern Gulf lowlands

X = present, integer = rank in sample (1 = highest rank), percentage values for Tres Zapotes and San Lorenzo indicate prevalence in random samples.

Source Abbreviations: UM, Ucareo-Zinapecuaro, Michoacan; SH, Sierra de Pachuca, Hidalgo; OM, Otumba, Mexico; PP, Paredón, Puebla; ZP, Zaragoza, Puebla; AV, Altotonga, Veracruz; PV, Pico de Orizaba, Veracruz; GP, Guadalupe Victoria, Puebla; ECG, El Chayal, Guatemala; SMG, San Martín Jilotepeque, Guatemala; IG, Ixtepeque, Guatemala; TG, Tajumulco, Guatemala; SL-A', San Lorenzo source A'; SL-D, San Lorenzo source D; SL-E, San Lorenzo source E; UID, Unidentified sources.

References:

¹Cobean et al. Reference Cobean, Coe, Perry, Turekian and Kharkar1971:Table 2 (presence/absence). Subsequent analysis reassigned Group A/A' to Paredón, B/B' to Ucareo-Zinapecuaro, and C/C' to Zaragoza-Oyameles (Cobean Reference Cobean2002:53, personal communication 2013; Hirth et al. Reference Hirth, Cyphers, Cobean, de León and Glascock2013:2785, footnote 5).

²Doering Reference Doering2002:Tables 5 and 10 (rank based on weight in visually sourced sample).

^3aSantley et al. Reference Santley, Barrett, Glascock and Neff2001:Table 3a (phase-specific samples only).

^3bSantley et al. Reference Santley, Barrett, Glascock and Neff2001:Table 3b (includes samples assigned to more than one phase).

⁴González 2005, personal communication.

⁵Stross et al. Reference Stross, Hester, Heizer, Jack and Taylor1976.

⁶Stokes Reference Stokes1999.

⁷Stark et al. Reference Stark, Heller, Glascock, Elam and Neff1992:229. Includes Late Preclassic (600–300 b.c.) and Terminal Preclassic (300 b.c.–a.d. 300). Pachuca (SH) was visually identified.

⁸Heller Reference Heller and Stark2001:161–163, Table 9.2. See notes for Reference 7 (Stark et al. Reference Stark, Heller, Glascock, Elam and Neff1992).

⁹Hirth et al. Reference Hirth, Cyphers, Cobean, de León and Glascock2013:Tables 4–8 (percentages calculated from reported frequencies).

In one of the early obsidian sourcing studies, Hester et al. (Reference Hester, Jack and Heizer1971b) conducted XRF source analysis from a sample of 865 unprovenienced obsidian artifacts excavated in 1939 and 1940 by Matthew Stirling at Tres Zapotes. Their analysis determined that 93.4% of the sample was Zaragoza-Oyameles obsidian. In addition, obsidian from Guadalupe Victoria, Pachuca, Pico de Orizaba, and three unknown sources were identified, one of which has since been identified as Zinapecuaro (Healan Reference Healan1997). Unfortunately, the results cannot be used to understand diachronic and spatial variation in the Tres Zapotes obsidian economy.

Knight had a sample of 100 provenienced artifacts from excavated contexts at the Protoclassic to Classic period site of Palo Errado, located 10 km southwest of Tres Zapotes (Figure 3), sourced using INAA (Knight and Glascock Reference Knight and Glascock2009). The results identified clear obsidians from Zaragoza-Oyameles, Pico de Orizaba, and Guadalupe Victoria; black obsidians from Zaragoza-Oyameles, Ucareo-Zinapecuaro, and Otumba; and green obsidian from Pachuca (Figure 1). Obsidian from Guatemala has not been identified at Palo Errado but has been recovered from Middle and Late Formative contexts in the Sierra de los Tuxtlas (Santley et al. Reference Santley, Barrett, Glascock and Neff2001), Classic period contexts in the Hueyapan region (Esquivias Reference Esquivias2003), and Early and Middle Formative occupations at the Olmec centers of La Venta and San Lorenzo (Cobean et al. Reference Cobean, Coe, Perry, Turekian and Kharkar1971, Reference Cobean, Vogt, Glascock and Stocker1991; Hirth et al. Reference Hirth, Cyphers, Cobean, de León and Glascock2013; Jack and Heizer Reference Jack and Heizer1968).

Figure 3.

Map of the southern Gulf Coast of Mexico, with sites discussed in the text.

SAMPLING AND ANALYTICAL PROCEDURES

Obsidian artifacts from the PATZ excavations were visually classified based on color, including characteristics of transparency and translucency. Four base colors were distinguished (black, clear, green, and gray). Color categories also were divided into fourteen “subcolors” that refer to secondary visual characteristics and account for the full range encountered in the analysis. The presence and size of inclusions, surface texture, and other characteristics also were noted. The visual categories assigned to the PATZ obsidian match those assigned to the sample submitted to the University of Missouri Research Reactor (MURR) for INAA testing from Palo Errado (Knight and Glascock Reference Knight and Glascock2009).

After the initial visual classification, a stratified random sample of 180 excavated obsidian artifacts was selected for XRF at MURR. These were supplemented by a judgmental sample of 24 artifacts, raising the total to 204 artifacts (a .04% sample overall). The random sample was stratified between Olmec (Early and Middle Formative) and Epi-Olmec (Late Formative and Protoclassic) occupations to obtain a statistically valid representation of obsidian source use through time. The figure of 90 specimens from each sampling stratum was chosen to provide a confidence level of .99 that a missing source constituted less than 5% of the obsidian recovered (Drennan Reference Drennan1996:Table 18.1). The nonrandom sample was chosen to ensure that all visually distinct categories were sampled and to determine the source of several anomalous prismatic blades and flakes with ground platforms. The blades and flakes with platform grinding all come from a single Protoclassic occupation and represent one of the earliest examples of platform grinding technology in Mesoamerica (Coe and Diehl [1980a:248] report a possibly earlier example from the Nacaste phase).

All artifact samples were analyzed by an energy-dispersive XRF spectrometer. The spectrometer was a Bruker Tracer III-V handheld portable. It is equipped with an air-cooled rhodium target anode X-ray tube with a 140 µm Be window and a thermoelectrically cooled Si-PIN diode detector. The detector has a resolution of 180 eV for the 6.4 keV peak from iron. The beam dimensions are approximately 2 × 3 mm. In order to measure the elements in this study (K, Ti, Mn, Fe, Zn, Ga, Rb, Sr, Y, Zr, and Nb), the X-ray tube was operated at 40 kV and a count rate of about 1,250 counts per second, which varied according to sample size. Measurement times on all samples were 180 seconds. Peak deconvolution and element concentrations were accomplished using the Bruker spectral analysis package. The instrument was calibrated using data from a series of well-characterized source samples in the MURR reference collection, including eleven Mesoamerican sources (El Chayal, Ixtepeque, San Martín Jilotepeque, Guadalupe Victoria, Pico de Orizaba, Otumba, Paredón, Sierra de Pachuca, Ucareo, Zaragoza, and Zacualtipan) and three Peruvian sources (Alca, Chivay, and Quispisisa). These sources were previously analyzed by INAA and XRF in several different analytical laboratories to establish consensus values.

RESUMEN

Reportamos los resultados de la caracterización química de artefactos de obsidiana provenientes de Tres Zapotes por medio del análisis de florescencia de rayos-X (XRF). Este estudio de la identificación de yacimientos de obsidiana es el primero para este centro olmeca y epi-olmeca en que la muestra proviene de contextos seguros asignados a los periodos formativo temprano, formativo medio, formativo tardío, y protoclásico. Empleamos una estrategia de muestreo al azar estratificado para seleccionar 180 artefactos de obsidiana del conjunto excavado, supliéndolos con otro 24 especímenes de categorías visuales escasos. Entonces logramos caracterizar con mayor confianza los cambios en la importancia relativa de distintos yacimientos de obsidiana dentro de la economía política de Tres Zapotes a través de la transición clave desde la sociedad olmeca a la epi-olmeca. También pudimos extender nuestras observaciones a la muestra completa de 5,713 artefactos que caracterizamos visualmente, incluyendo los 2,695 piezas provenientes de contextos de fechas seguros que examinamos en este artículo. Nuestro estudio confirma la ausencia de obsidiana de Otumba y de los yacimientos guatemaltecos dentro del conjunto excavado olmeca. Este resultado apoya un modelo de obtención de obsidiana por redes traslapados pero autónomos por los sitios olmecas de la costa del Golfo. La presencia de obsidiana del yacimiento guatemalteco de San Martín Jilotepeque en contextos epi-olmecas puede relacionarse con el re-establecimiento de contactos trans-istmeños, mientras que el aumento en la obsidiana del yacimiento de Zaragoza-Oyameles, ubicado en el este del estado de Puebla, marca la empieza de una tendencia de largo tiempo. Aunque la representación más uniforme de yacimientos de obsidiana en contextos epi-olmecas es consistente con la transición hipotética desde una economía política olmeca exclusiva hacía un sistema más cooperativa epi-olmeca asociada con el repartimiento de poder entre facciones elites en Tres Zapotes, ni los elites olmecas ni epi-olmecas monopolizaron un yacimiento particular de obsidiana o una tecnología especial para trabajarla.