Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports

journal homepage: www.elsevier.com/locate/jasrep

Obsidian sourcing and dynamic trade patterns at Izapa, Chiapas, Mexico:100 BCE–400 CE

Rebecca R. MendelsohnSmithsonian Tropical Research Institute, Center for Tropical Paleoecology and Archaeology, Ave. Roosevelt, Balboa Ancon, Edific. Tupper #401, Panamá, República dePanamá

A R T I C L E I N F O

Keywords:Obsidian sourcingTrade networksCollapseDynamic economiesMesoamericaPacific coast

A B S T R A C T

Recent research on Mesoamerican economies has demonstrated that Prehispanic trade networks were not static,but constantly evolving. Relationships between peoples in different settlements and regions shift as social, po-litical, and economic circumstances change. One question that results from these findings is how significantevents such as sociopolitical decline or abandonment affect exchange networks. This study investigates changesin obsidian procurement during a period of widespread upheaval across southern Mesoamerica, ca. 100–250 CE,from the perspective of a surviving capital on the Pacific coast of Chiapas, Mexico. X-Ray Fluorescence (XRF)analysis using a Bruker Tracer III–V was conducted for obsidian collected from Terminal Formative(100 BCE–250 CE) and initial Early Classic (250–400 CE) domestic contexts to evaluate how obsidian procure-ment patterns at the ancient city of Izapa were affected by events in neighboring regions. Obsidian results arecompared to known shifts in sociopolitical relationships as observed through ceramics, burials, and art styles.Data indicate that residents of Izapa shifted trade from east to west over the Terminal Formative to Early Classicperiod transition. Patterns correspond with sociopolitical turmoil at the Guatemalan highland capital ofKaminaljuyú and the rise of the central Mexican city of Teotihuacan. Results also indicate, however, that eco-nomic relationships may continue even in places where political affiliations have been severed. Results suggestthat datasets for different artifact classes, when viewed independently, introduce additional nuance into thequestions of collapse and resilience of ancient societies.

1. Introduction

Recent advancement in archaeological techniques has allowed re-searchers to reevaluate the organization of ancient economies. Whereaspre-modern economic activity was once relegated to a few economictypes, such as reciprocity and redistribution (Polanyi, 1957) and theritual-regal model for urban economies (Fox, 1977; Sanders andWebster, 1988), archaeologists have now documented a greater rangeof ancient economic systems. Dynamic cycles of economic growth anddecline have been recognized, as well as waves of globalization andregionalization, where economic shifts occur simultaneously acrossgreat regions (Frank, 1993; Hodos, 2017; Jennings, 2011; Ur, 2010).The introduction of sourcing techniques such as X-Ray fluorescence(XRF) and neutron activation analysis (NAA) have become essentialtools for tracking interregional and intraregional exchange, data whichhave been used to demonstrate a complex array of past economic ac-tivities (Aoyama, 2001; Braswell and Glascock, 2002; Moholy-Nagyet al., 2013; Neff et al., 1989, 1994; Nichols et al., 2002).

Recent research on Mesoamerican economies has demonstrated that

Prehispanic trade networks were not static, but constantly evolved inresponse to new technologies and changing social, political, and eco-nomic circumstances (Braswell and Glascock, 2002; Golitko andFeinman, 2015; Nichols et al., 2002; Stark et al., 2016). Much of thework devoted to Mesoamerican economies has emphasized trade duringhorizon periods, or eras of widespread interaction, associated withprominent cultural groups, such as the Olmec, Teotihuacanos, LateClassic Maya, and Aztecs (e.g. Blanton et al., 1978; Blomster et al.,2005; Braswell, 2003a; Smith and Berdan, 2003; Stoltman et al., 2005).Fewer scholars have directly addressed how the collapse of ancientcities affected ancient trade networks (but see Masson et al., 2010; Pooland Loughlin, 2015; Reese-Taylor and Walker, 2002).

The following study investigates obsidian trade during an earlyperiod of widespread disruption, between 100 CE and 250 CE, whenmany of southern Mesoamerica's first cities declined. Changing patternsof obsidian importation are evaluated for the Pacific coastal capital ofIzapa, Chiapas, Mexico using X-Ray fluorescence (XRF) sourcing ofobsidian recovered from domestic contexts. Documentation of domesticobsidian for a capital city that survived the Formative period “collapse”

https://doi.org/10.1016/j.jasrep.2018.06.002Received 4 February 2018; Received in revised form 21 May 2018; Accepted 1 June 2018

E-mail address: MendelsohnR@si.edu.

Journal of Archaeological Science: Reports 20 (2018) 634–646

2352-409X/ © 2018 Elsevier Ltd. All rights reserved.

T

helps to define how the 100–250 CE decline of prominent southernMesoamerican cities impacted interregional exchange systems.

In addition to tracking regional economic shifts, this study exploreswhether effects of collapse can be better understood by separating da-tasets associated with different materials. Do obsidian procurementpatterns support the ceramic, monument, and burial data that suggestan abrupt decline in Izapeños' (residents of Izapa) external relationshipsfollowing the 100–250 CE events?

Obsidian sourcing results generally corroborate previously observedshifts in ties between Izapeños and peoples of neighboring regions.Changing rates of imported blades of El Chayal source obsidian supportdata that relationships between Izapeños and peoples at capitals withinthe “Miraflores sphere” of highland Guatemala and western El Salvador(Demarest and Sharer, 1986) surge during the Terminal Formativeperiod (100 BCE–250 CE) but were severed by the onset of the EarlyClassic period (250–400 CE). Obsidian blades recovered from the Sierrade Pachuca source (hereafter, Pachuca) in Jaritas phase (250–400 CE)contexts now represent the earliest known appearance of this centralMexican obsidian at Izapa and suggest that new relationships wereestablished to the west by this time. The initiation of these western tiescould have been prompted either by the collapse of Izapeños' easterntrade partners, including the highland Guatemalan capital of Kami-naljuyú, or the rise of the central Mexican city of Teotihuacan. Morelikely, Izapeños shifted their import patterns in response to multiplesuch events.

The continued appearance of some El Chayal obsidian at Izapa after250 CE, however, suggests that, while political relationships may haveceased with Miraflores capitals such as Kaminaljuyú, economic activitymay have continued between these two zones. Results suggest thatsocial, political, and economic datasets, when viewed independently,introduce additional nuance into questions of the collapse and re-generation of ancient societies.

2. Formative period collapse in southern Mesoamerica

In Mesoamerica, the transition between the Formative and Classic

periods, between 100 BCE and 400 CE, was a time of great socialchange, including the apogee and subsequent decline of several pow-erful Formative period cities. During the first half of this period, from100 BCE to 100 CE, archaeologists and art historians have documentedthe spread of low relief sculptural traditions with writing and the de-pictions of rulers, the establishment of kingly burial practices, and thedevelopment of a widespread network for the procurement of foreigngoods and materials (Freidel and Schele, 1988; Guernsey, 2006;Inomata and Henderson, 2016; Reese-Taylor and Walker, 2002). Manyof the earliest Mesoamerican cities, such as El Mirador, Kaminaljuyú,and Takalik Abaj were at their height of artistic production andmonumental construction. Levels of interregional interaction also ap-pear to have been high, with mounting evidence for a widespreadceramic horizon and exchange network (Inomata et al., 2014;Mendelsohn, 2018a; Mendelsohn, 2018b; Reese-Taylor and Walker,2002).

By the close of the Formative period, many prominent cities wereabandoned, while others experienced significant disruptions. For theperiod between 100 and 250 CE, archaeologists have documented po-pulation declines, cessations in construction activities and monumentproduction, signs of warfare, environmental stress, and abrupt changesin artifact assemblages, all pointing to collapse or turmoil at manysouthern Mesoamerican cities at this time (Doyle, 2017: 112–129;Estrada-Belli, 2011; Grube, 1995; Love, 2012; Popenoe De Hatch, 1998;Reese-Taylor and Walker, 2002). As more datasets for the Formative toClassic period transition are becoming available, it is increasingly clearthat disruption at the close of the Formative period was variable.

Among the regions affected was the Pacific coast. Archaeologistshave documented an apparent depopulation of coastal settlements atthe close of the Formative period. Surveys of coastal Guatemala haverevealed limited Early Classic activity in this zone (Coe and Flannery,1967: 6; Shook, 1965: 185–186). Michael Love's research at the Gua-temalan coastal site of El Ujuxte suggests that this capital collapsedshortly after 100 CE (Love, 2002, 2007, 2018). Unlike the GuatemalanPacific coast, however, data from the Izapa polity suggest continuity(Lowe et al., 1982: 145; Rosenswig and Mendelsohn, 2016). Izapa

Fig. 1. Map of sites mentioned in the text and obsidian sources identified at Izapa: 1. Tajumulco, 2. San Martin Jilotepeque, 3. El Chayal, 4. Ixtepeque, 5. Pachuca.Map by the author.

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

635

appears to have followed a different trajectory than neighboring cities.Instead, signs of turmoil appear to occur 200 years earlier, between100 BCE and 100 CE (Lowe et al., 1982: 139, 194).

2.1. Izapeños and their neighbors during the Formative to Classic periodtransition

The archaeological site of Izapa is located in the Soconusco region ofChiapas, Mexico (Fig. 1). It lies within a Pacific coastal region of Me-soamerica that is defined for its high level of rainfall and known forimportant Mesoamerican products, such as cotton, rubber, birdfeathers, and cacao (Gasco and Voorhies, 1989). Cacao was a particu-larly important export for this region and is represented in art of theLate Formative period (Guernsey et al., 2018).

Early excavations at Izapa focused on culture historical questions,including documentation of the famous “Izapan” style low reliefmonuments (Drucker, 1948; Guernsey, 2006; Norman, 1976; Parsons,1986; Smith, 1984) and developing a chronology for the site (Drucker,1948; Lowe et al., 1982). The largest-scale excavations to date wereundertaken by the New World Archaeological Foundation in the 1960s(Lowe et al., 1982). Subsequent excavations in the Formative periodcore have been directed by Hernando Gómez Rueda (1995) and RobertRosenswig (Rosenswig et al., 2018). Rosenswig's regional settlementsurvey and recent LiDAR (light detection and ranging) mapping of thesite have helped to put events at Izapa in broader spatial and temporalcontext (Rosenswig and Mendelsohn, 2016; Rosenswig et al., 2013).

Following their 1960s excavations, Lowe et al. (1982) developed thefirst preliminary culture-historical view of Izapa. Construction of thecity's primary ceremonial center began with the creation of the prin-cipal mound (Mound 30a) in the Duende phase, between 850 and750 BCE. Growth at Izapa continued throughout the Formative period,with the greatest era of monumental construction and stone sculptureerection proposed for the Guillen phase (300–100 BCE) of the LateFormative period (Lowe et al., 1982: 23, 133).

With the onset of the Terminal Formative period, radical changesoccurred at Izapa. During the Hato phase (100 BCE–100 CE), reportssuggest the cessation of large-scale construction activities and stonemonument production, the introduction of new burial practices, and theinitiation of a second ceremonial center to the north (Lieske, 2013;Lowe, 1993; Lowe et al., 1982: 194, 315). This era of apparent dis-ruption was followed by an age of stability (Lowe et al., 1982: 141).Construction activities resumed at the site during the Itstapa phase(100–250 CE), and, while the city did not maintain its former grandeurin scale and monument carving, Izapa would continue to be occupiedfor another 1000 years (Lowe et al., 1982; Rosenswig and Mendelsohn,2016).

Lowe used ceramic styles to document shifts in Izapeños' trade andforeign relationships over the history of the city. Results suggest thatfrom 750 to 500 BCE, during the latter part of the Middle Formativeperiod, Izapan potters participated in the widespread “Modified Olmec”horizon, where “splotchy” orange ceramics were encountered fromcentral Mexico to El Salvador (Clark and Cheetham, 2005; Lowe, 1977).By the Late Formative period, he observed a regionalization in ceramicstyles. Ties between Izapa and the Gulf Coast region diminished (Poolet al., 2018), though the presence of a continuous orange-slippedceramic tradition along the Pacific coast suggests strong social ties weresustained within this zone (Mendelsohn, 2018a). The onset of theTerminal Formative period marked an era of re-integration. The in-troduction of modal changes on Hato phase (100 BCE–100 CE) cera-mics, such as Usulutan-style resist decoration and the introduction oftripod nubbin supports (Clark and Cheetham, 2005), suggests a closerelationship with peoples of the Miraflores ceramic sphere in highlandGuatemala and El Salvador (Demarest and Sharer, 1986; Lowe et al.,1982: 139).

But by the onset of the Early Classic period, links to sites associatedwith the Miraflores sphere to the southeast were virtually nonexistent.

In fact, limited Jaritas phase (250–400 CE) ceramic ties are observedwith any neighboring area, indicating a period of relative regionaliza-tion (Lowe et al., 1982: 145). Similarities in ceramic styles that do existappear to be restricted to the western Soconusco, suggesting a west-ward shift in social relationships (Mendelsohn, 2017: 264–268; Pfeiffer,1983: Vol. II, 224–246). This shift was likely associated with “collapse”or disruption at early capitals in the Miraflores zone and elsewhere inthe Maya lowlands. While these data are a useful baseline for inter-preting foreign relationships, using ceramic styles to predict economicrelationships can be misleading, as style is influenced by a complex setof factors, including ethnicity, religious affiliation, political alliances,and economic connections (Bowser, 2000; David et al., 1988; Hegmonand Kulow, 2005; Plog, 1983).

3. Obsidian and the Soconusco economy

Obsidian is a volcanic glass that was commonly used as a cuttingmaterial in Mesoamerica. Because it is only available in select volcanicdeposits, obsidian was not locally available in most regions ofMesoamerica and had to be imported through long-distance exchange.As a result, obsidian is often considered a good indicator of economicactivity (Aoyama, 1996; Braswell, 2003b; Stark et al., 2016). Obsidianprovenance studies indicating the volcano from which an individualpiece was obtained are helpful for determining distribution networks ofthis important material (e.g. Asaro, 1978; Cobean et al., 1971; Sidryset al., 1976).

Early obsidian studies in Mesoamerica emphasized the political andreligious aspects of trade in this resource (Clark, 1987; Sidrys, 1976).Clark (1987), for example, saw early blade-making (that is, trade inblade cores accompanied by the local production of prismatic blades) asa political rather than economic act. The adoption of blade technologyin the Early Formative period, he argues, was used by “paramountchiefs” to strengthen their political position; obsidian blades were usedas political currency (Clark, 1987: 278). Sidrys (1976) also con-ceptualized obsidian as a status good for the Late Formative and EarlyClassic periods, suggesting it reflected political and religious ties.During the subsequent Classic period, however, obsidian blades becamemore widely available and more began to serve as utilitarian goods(Sidrys, 1976).

The recent increase in obsidian studies from non-elite domesticcontexts (e.g. DeLeón et al., 2009; Hirth, 1998; Stark et al., 2016), haveallowed archaeologists to evaluate the role of obsidian as an economicresource, rather than as ritual or political currency. In his distributionalmodel, Hirth (1998) outlines how the composition of household as-semblages will differ depending on whether obsidian is being dis-tributed by elites or is obtained through marketplace exchange. Datafrom DeLeón et al. (2009) challenge the established model that obsidianindustries evolved from trade in finished blades in the Formative periodto widespread trade in cores in the Classic period. Identifying correlatesfor whole-blade trade, processed-blade trade, and local-blade produc-tion, they found that there was considerable regional variability in themanner in which the obsidian blade trade evolved from the Middle toLate Formative periods (400 BCE–100 CE). Stark et al. (2016) observethat, in the coastal lowlands zone, additional cutting materials such aschert are not available. Obsidian flakes and blades in this region, theyargue, would have been utilitarian, rather than luxury goods.

Here, obsidian results are compared with the social and politicaldata inferred from ceramic styles and elite burial goods to determinewhether different datasets illustrate the same shifts in Izapeños' inter-regional relationships over the Formative to Classic period transition.Investigation into obsidian consumption at Izapa, a center that survivedthe Terminal Formative collapse, allows us to determine how tradenetworks may have been impacted by the decline of important southernMesoamerican cities between 100–250 CE.

Located in a zone with no local access to this material, all obsidianfound at Izapa needed to be imported through long-distance travel or

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

636

exchange. The obsidian sources closest to Izapa are located inGuatemala (Fig. 1). The nearest source, Tajumulco, is located justacross the modern-day border to Guatemala and represents a prominentpoint on the horizon at the site. Despite being the closest source toIzapa, Tajumulco obsidian (technically, ignimbrite) has a more granulartexture and coarse matrix and is generally considered unsuitable for theproduction of pressure blades (Clark et al., 1989). The closest obsidiansources to Izapa with material of sufficient quality for blade-makingcome from other Guatemalan volcanoes, including San Martin Jilote-peque, El Chayal, and Ixtepeque. Additional sources of Mesoamericanobsidian have been identified in central Mexico, including the Sierra dePachuca source, famous for its distinctive green color and its associa-tion with the central Mexican city of Teotihuacan (Spence, 1996).

Because obsidian is not found locally at Izapa, changes in con-sumption patterns of this material can track shifting economic inter-actions through time. Temporal trends of obsidian consumption pat-terns of the Soconusco region were previously compiled by Clark et al.(1989) using visual sourcing. Pooling materials from several sites in theregion, including Izapa (Table 1), they suggest temporal trends in ob-sidian exchange for the full range of occupation in the area, for theArchaic to Colonial periods (Clark et al., 1989).

Golitko and Feinman (2015) have since investigated dynamicchanges in obsidian consumption patterns at a pan-regional scale.Previous results from Izapa, derived from the Clark et al. (1989) study,are placed into a broader regional context using Social Network Ana-lysis. Golitko and Feinman compare networks for eight distinct timeperiods. They find that, during the Late and Terminal Formative periods(250 BCE–250 CE), sites in the Soconusco region “are relatively per-ipheral to the overall network structure,” primarily linking sites in theGuatemalan highlands with the Maya lowlands (Golitko and Feinman,2015: 221). However, they observe that Late and Terminal Formativecities in the Soconusco region play an important role in bridging easternand western spheres of obsidian procurement, overtaking a role pre-viously served by centers in the Gulf Coast. The bridging role of So-conusco centers in Golitko and Feinman's network models is maintainedduring the Early and Middle Classic periods (300–600 CE), particularlythrough the western Soconusco city of Los Horcones (García-DesLauriers, 2007).

While both studies are helpful as preliminary impressions of ob-sidian consumption and trade at Izapa, several problems are presentwith the data that was available for these studies. First, becausescreening was not undertaken during 1960s New World ArchaeologicalFoundation excavations (Lowe et al., 1982: 84), quantitative results ofthe Clark et al. (1989) study are not reliable. Without a consistent re-covery methodology, equal rates of obsidian collection across archae-ological contexts cannot be assumed. Second, samples are from ex-cavations in the two Izapa ceremonial centers. Recovery from suchspecialized contexts may be more indicative of religious obsidian use orelite consumption patterns, rather than a representation of the domesticobsidian economy at Izapa. Third, small sample sizes were available forthe Terminal Formative and Early Classic periods. Results, therefore,may not be an accurate reflection of the Izapa economy for these per-iods. In the Golitko and Feinman (2015) study, Terminal Formativeperiod results were excluded from the social network analysis preciselyfor this reason. However, the lack of data from a regional capital such asIzapa for their Terminal Formative period network, may lead to an

underrepresentation of the role of Soconusco-region peoples in greaterMesoamerican economies during this time. Indeed, results from a recentcompilation of obsidian quantification data by Stark et al. (2016),suggest that, because of their position at a regional capital, Izapeñosmay have had preferred access to obsidian, particularly during the EarlyClassic period.

4. The Izapa Household Archaeology Project (IHAP)

The obsidian data presented here from the Izapa HouseholdArchaeology Project (IHAP) expand on previous results from the Izapaceremonial core with data recovered from the southern periphery of thesite. The IHAP results represent the first systematically collected ob-sidian data from Izapa, recovered through screened excavations indomestic contexts. Increased sample sizes resulting from the excavationof new Terminal Formative and initial Early Classic contexts at thiscapital city are valuable as a reference point for understanding how theTerminal Formative collapse affected Mesoamerican exchange spheres.

The IHAP sought to document domestic activity at Izapa. The pro-ject consisted of 25 excavation units at seven mounds in the southernperiphery of the site (Fig. 2). Significant contexts included TerminalFormative construction and occupation at Mound 255 and Early Classicdomestic debris at Mound 260 (Mendelsohn, 2017: 117–164). To-gether, these contexts represent a combination of construction fill,primary occupation, and midden deposits associated with the Hato(100 BCE–100 CE), Itstapa (100–250 CE), and Jaritas (250–400 CE)phases, which span the Formative period to Classic period transition atthe site. Obsidian recovered from the IHAP was collected throughconsistent screening. Fill from all excavated lots was dry-screened forartifact recovery using ¼ inch mesh screens. The volume of the soilremoved from each lot was also recorded, so that artifact densitiescould be calculated in proportion to the volume of material excavated.

Subsequent provenance analyses of the obsidian were undertaken toinvestigate several questions. From which sources did Izapeños obtaintheir obsidian? How do the results from southern Izapa compare toprevious results from the ceremonial centers? Do patterns of obsidianprocurement shift over the Terminal Formative to Early Classic periodtransition, in association with the 100–250 CE collapse? Are patterns ofobsidian importation correlated with ceramic ties? Sourcing using XRFwas employed to address these questions.

Hypotheses were derived from previous interpretations of ceramicties. If economic relationships were to correspond with social and po-litical ties, as determined by ceramic data and art styles, then the fol-lowing patterns might be expected. For the Terminal Formative period(100 BCE–250 CE), when ceramic styles at Izapa share close ties withsites in the Miraflores sphere of highland Guatemala and western ElSalvador (Demarest and Sharer, 1986; Lowe et al., 1982: 139), in-creased rates of El Chayal and Ixtepeque obsidian are projected. Higherrates of El Chayal obsidian would indicate an increased relationshipwith the Guatemalan highland capital of Kaminaljuyú, where bladeproduction workshops of this material have been identified and, somehave even argued, maintained control of this source (Hay, 1978;Kaplan, 2002, 2011; Love, 2007: 297–298). Importation of obsidianfrom the Ixtepeque source would indicate strengthening ties with ElSalvador, as this was the predominant source of obsidian for settle-ments in that region (Braswell et al., 1994; Neivens and Demarest,

Table 1Previous visual sourcing results of Izapa obsidian. Data was drawn from materials excavated during the 1960s New World Archaeological Foundation project (Loweet al., 1982).

n SMJ El Chayal Tajumulco Ixtepeque Source

Middle-Late Formative 404 64.9% 1.7% 33.4% 0.0% Clark et al., 1989: Table 12–1Terminal Formative 43 69.0% 4.0% 27.0% 0.0% Clark and Lee Jr., 2013: 10Early Classic 15 53.3% 13.3% 33.3% 0.0% Clark et al., 1989: Table 12–1

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

637

1986).Patterns of obsidian procurement were expected to shift with the

onset of the Early Classic period (250–400 CE), as trade networks wouldbe affected by the collapse of prominent cities between 100–250 CE.Based on the break in ceramic ties between Izapa and the Mirafloresregion (Lowe et al., 1982: 145), and subsequent increase in stylisticsimilarities with the western Soconusco (Mendelsohn, 2018a; 1983:Vol. II, 224–246), a decline in obsidian from Guatemalan sources andan increase in the appearance of sources from Mexico was projected.

The archaeological contexts selected to address these questionsconsisted of obsidian recovered from two mounds at the southernperiphery of the site, Mounds 255 and 260 (Fig. 2). Results from Mound255 and its associated off-mound excavation unit represent the

Terminal Formative period (100 BCE–250 CE), while obsidian fromMound 260 represent the onset of the Early Classic period Jaritas phase(250–400 CE).

Results from Mound 255 included obsidian recovered from three1x3m vertical excavation units, Suboperations 104a, 104b, and 104d,as well as Suboperation 104c, the shallow 2x3m horizontal extensionwest of Suboperation 104a. Mound 255 was constructed during theHato phase (100 BCE–100 CE) of the Terminal Formative period, withboth carbon and offerings dating to this era (Mendelsohn, 2017:118–135, Mendelsohn, 2018b). Materials from as early as the Cuadrosphase (1350–1200 BCE) of the Early Formative period, however, werealso recovered in construction fill. Known technological changes andthe metric measurements of blades from Mound 255 help to narrow

Fig. 2. Location of IHAP excavations at the southern periphery of Izapa. Top base map from Lowe et al., 1982: inset. LiDAR map of IHAP excavation zone courtesy ofRobert Rosenswig and the Izapa Regional Settlement Project (Rosenswig et al., 2013).

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

638

down the temporal range of some of this material to the TerminalFormative period (see below). Nevertheless, results from Mound 255are the most tentative presented and should be re-evaluated as addi-tional Terminal Formative data become available from Izapa.

The second context where Terminal Formative materials were re-covered was Suboperation 104e, a 1×2m vertical off-mound ex-cavation unit immediately south of Mound 255 (Fig. 3). This unit re-vealed signs of construction activity as well as a layer of domesticdebris. Suboperation 104e contained Terminal Formative ceramics as-sociated with the Hato and Itstapa phases, with a layer (Levels 8–10) ofItstapa phase (ca. 100–250 CE) occupation discovered atop a possiblefloor (Mendelsohn, 2017: 139–145; Mendelsohn, 2018b). Because theoccupation at Suboperation 104e is believed to be slightly later and ismore securely associated with the Terminal Formative period, it wasconsidered separately.

Mound 260 (Fig. 4) was a small, one-meter-high structure occupiedexclusively during the Jaritas phase (250–400 CE). Excavated depositswere domestic in nature, including a trash pit and fired clay feature(Mendelsohn, 2017: 147–164; Mendelsohn, 2018b). Obsidian fromMound 260 was recovered from the three vertical excavations unitsatop this mound: Suboperation 105a (1× 3m), 105b (1×3m), and

105d (1× 2m). Obsidian was also recovered from Suboperation 105c,a small extension (40× 65 cm) excavated to recover material from thetrash deposit discovered at the corner of Suboperation 105a. BecauseJaritas phase results derive from this one mound, it is possible that theyare not representative of all Jaritas phase obsidian consumption at thesite. As with the Terminal Formative period contexts, results should beupdated as additional data becomes available. Nevertheless, with theirsystematic collection and increased sample sizes, the results fromMound 260 currently represent the best Early Classic economic dataavailable at Izapa. Because Mound 260 obsidian derives from a non-elite domestic trash pit, it is assumed that materials from Mound 260represent functional utilitarian goods rather than ritual activity.

5. XRF analysis

While visual sourcing is often reliable for this region for those with atrained eye (such as John Clark, who conducted the prior visual ana-lyses of the Izapa obsidian), previous studies of the efficacy of visualsourcing have revealed that Ixtepeque and “black” obsidians derivedfrom central Mexican sources may be under-identified through visualsourcing alone (Braswell et al., 2000). Because the identification of

Fig. 3. IHAP excavations at Mound 255. Profile and representative obsidian illustrated for off-mound unit Suboperation 104e. Contour map courtesy of the IzapaRegional Settlement Project (Rosenswig et al., 2013).

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

639

obsidian from Ixtepeque and Mexican sources was an important ob-jective for answering the research questions, I elected to conduct XRFanalysis of the IHAP obsidian. To test the above hypotheses, all obsidianblades from Terminal Formative and Early Classic contexts (n=121)and select deposits of non-blade obsidian dating to these periods(n=442) underwent compositional analysis (Supplemental Tables 1and 2).

Provenance data for blades were obtained through XRF analysis atthe New York State Museum, where a Bruker Tracer III-V model wasused. The machine was set to heavy metals mode, using a filter con-taining 6 μm of Copper, 1 μm of Titanium, and 12 μm of Aluminum.Samples were run using the S1PXRF program at 40 kV and 22microamps, matching the settings used for the obsidian calibration previouslyanalyzed on the instrument. Following the methodology developed byMoholy-Nagy et al. (2013), samples were run for 90 s each and controlsamples were run at the beginning and end of each day, so that anychanges in the readout of the machine could be observed (seeMendelsohn, 2017: 292–295 for additional methodological details).Each blade was visually sourced prior to XRF analysis to test the rate ofefficacy for my visual identifications.

Parts per million (ppm) concentrations were generated for Mn, Fe,

Zn, Ga, Th, Rb, Sr, Y, Zr, and Nb (Supplemental Table 1) using theobsidian calibration provided for the machine (GL1.cfz). These valueswere then run through a hierarchical cluster analysis. Control samplesfrom each day of analysis were included to assure accurate clusteringand aid in the identification of the source materials for the statisticalgroups. Multiple iterations of the hierarchical cluster analysis werecompleted using different subsets of elements. The test which best se-parated the control samples into the correct groupings included fourelements: Rubidium (Rb), Strontium (Sr), Yttrium (Y), and Zirconium(Zr). Group membership probabilities were also calculated usingMahalanobis distance for Rb and Sr to confirm that the group assign-ments produced by the cluster analysis were accurate. Results were thesame for both tests.

The statistical analyses resulted in five main groups (Figs. 5,6). Thechemical composition of Group 1 was the most distinct and denoted theSierra de Pachuca source of green obsidian in the state of Hidalgo incentral Mexico. Group 2 consisted solely of control samples of the Ix-tepeque source. Obsidian in Group 3 came from the nearby Tajumulcosource. Group 4 represented obsidian from El Chayal. Group 5 includedobsidian from the San Martin Jilotepeque source.

To make the data more comparable with previous Izapa sourcing

Fig. 4. IHAP excavations at Mound 260. Profile and representative non-blade obsidian illustrated for Suboperations 105a and 105c. Contour map courtesy of theIzapa Regional Settlement Project (Rosenswig et al., 2013).

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

640

results, sourcing of a sub-sample of the non-blade obsidian recoveredfrom the IHAP excavations was also undertaken (SupplementalTable 2). Because of the limited time available with the machine,contexts that were of the greatest temporal security for the TerminalFormative and Early Classic periods were targeted for analysis. For theTerminal Formative period, this included obsidian from Suboperation104e, the off-Mound unit south of Mound 255, with Terminal Formativeceramics and a layer of Itstapa phase occupation. For the Early Classicperiod contexts, obsidian from Suboperation 105a and the associatedSuboperation 105c extension, which contained the best domestic de-posits recovered during the project, were targeted for analysis.

The sourcing methodology for the non-blade obsidian sourcing wasdifferent than that of the blades. Because visual sourcing for the bladeshad produced a 90.4% success rate (Mendelsohn, 2017: 301), non-bladeobsidian pieces were visually sorted into groups by source. Thesegroups were then tested with shorter (30 s.) runs on the XRF. Pieceswith the least secure visual identifications were targeted for testing.

Correct or incorrect associations were determined by the form of thespectra produced by the machine, rather than by elemental con-centrations. Multiple pieces were tested until confidence was reachedthat all fragments were accurately assigned. Counts and weights of eachgroup were then modified, as needed.

5.1. Terminal Formative period (100 BCE–250 CE) results

Results for the Terminal Formative period (Table 2) were derivedfrom the excavations at Mound 255 and the associated off-mound unitSuboperation 104e (Fig. 3). Blades recovered from on-mound excava-tions (Suboperations 104a, 104b, 104c, 104d) derived from Hato phase(100 BCE–100 CE), or early Terminal Formative period, constructionfill. The non-blade obsidian at Mound 255 could date to any periodbetween 1350 BCE and 100 CE and was, therefore, not targeted foranalysis. Prismatic blade technology, however, is not present in theSoconusco region until after 900 BCE (Clark, 1987; Jackson and Love,1991), so it can be safely assumed that the 74 blades recovered fromMound 255 represent obsidian from the Middle Formative periodthrough the Terminal Formative Hato phase, or 900 BCE–100 CE. Re-sults for the blades from Mound 255 indicate that 45 (60.8%) camefrom the San Martin Jilotepeque source, 25 (33.8%) came from ElChayal, and 4 more (5.4%) were associated with Tajumulco.1 No bladesof Ixtepeque or Pachuca obsidian were recovered from Mound 255.

While the blades from construction fill on Mound 255 could re-present any period from 900 BCE–100 CE, their width distributions mayoffer additional insight into the temporal association of these pieces(Fig. 7). The size of the larger blades of El Chayal source obsidian iscomparable to Terminal Formative examples found in neighboring re-gions. Clark and Lee Jr. (2007: 117) suggest that wide (up to 31mm)Terminal Formative blades of El Chayal obsidian from the inlandChiapas site of El Cerrito date to this period and were likely imported ascompleted products. Given their comparable widths and evidence thatMound 255 was constructed during the Hato phase, it is likely that thelarge El Chayal blades recovered from fills in Mound 255 are associatedwith Hato phase (100 BCE–100 CE) occupation and economic activityin this zone.

Obsidian blades obtained from the Suboperation 104e unit are moresecurely dated to the Terminal Formative period. Of the eight bladesrecovered from Suboperation 104e (Table 2), four of them were fromthe El Chayal source (50%), three came from San Martin Jilotepeque(37.5%), and one derived from Tajumulco (12.5%). The two bladesrecovered from the Itstapa phase occupation layer in Levels 8–10 wereboth from the El Chayal source. As with the Mound 255 results, noIxtepeque or Pachuca blades were recovered from Suboperation 104e.The absence of Ixtepeque obsidian in Terminal Formative contextscame as a surprise, given the heightened connections between Izapaand El Salvador observed through ceramic burial goods dating to theTerminal Formative period (Lowe, 1993).

For the non-blade results from Suboperation 104e, Tajumulco ob-sidian chunks dominated the assemblage, at 81.0%. San MartinJilotepeque (7.6%) and El Chayal (10.5%) obsidian was present inroughly equal proportions. One piece (1.0%) could not be assigned withconfidence. For the Itstapa phase debris in Levels 8–10, however, onlytwo non-blade pieces of San Martin Jilotepeque obsidian was recordedfrom the total of 34 non-blade pieces and three blades recovered.Instead, El Chayal blades and Tajumulco chunks were most prevalent in

Fig. 5. Elemental plots showing the division between the five obsidian sourcespresent at Izapa. Control samples are labeled by source: SMJ=San MartinJilotepeque, CHY=El Chayal, TAJ=Tajumulco, IXT= Ixtepeque,PACH=Sierra de Pachuca. Plot generated by the author using the GAUSSprogram developed at the University of Missouri Research Reactor (MURR).

Fig. 6. Close-up of elemental plot without the Pachuca samples. Control sam-ples are labeled by source: SMJ=San Martin Jilotepeque, CHY=El Chayal,TAJ=Tajumulco, IXT= Ixtepeque. Plot generated by the author using theGAUSS program developed at the University of Missouri Research Reactor(MURR).

1 The appearance of Tajumulco blades was surprising given previous reports that thismaterial could not be used for blade production (Clark et al., 1989). Upon renewed in-spection, it became clear that I had been more liberal in the pieces I assigned as blades atthe beginning of the analysis (i.e. for the Suboperations at Mound 255 and Mound 260)than I was as the analysis progressed. The pieces assigned as Tajumulco blades all hadtriangular or trapezoidal cross sections and parallel sides but were not third series pres-sure blades (Clark and Bryant, 1997). This identification contributed to the appearance ofTajumulco “blades” in the Terminal Formative and Early Classic assemblages.

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

641

Levels 8–10.The Mound 255 and Suboperation 104e results suggest that a small

spike in El Chayal obsidian may have occurred with the onset of theTerminal Formative period at Izapa. Of the blades recovered fromMound 255, 33.8% of them were sourced to El Chayal. This rate issignificantly higher than previous results reported for both the Middleto Late Formative periods (1.7%) or the Terminal Formative period(4%) at Izapa (Clark and Lee Jr., 2013: 10; Clark et al., 1989).

For IHAP contexts associated with both the Terminal Formative andinitial Early Classic periods, El Chayal obsidian was recovered almostexclusively as wide blade fragments. For the non-blade obsidian ana-lyzed, only 11 pieces (10.6%) of El Chayal obsidian were recoveredfrom the Terminal Formative off-mound unit Suboperation 104e(Table 2). While the non-blade obsidian for the Mound 255 fill was notanalyzed using XRF, qualitative observations suggest that the over-whelming majority of El Chayal obsidian in the Mound 255 fill wasassociated with blade fragments.

The wider width of the El Chayal blades recovered from TerminalFormative contexts at Izapa, together with the relative lack of El Chayalproduction debris, suggest that blades of El Chayal obsidian were likelynot produced at Izapa during the Terminal Formative period, at least inthe southern zones of the site. In their discussion of blades from theChiapas depression site of El Cerrito, Clark and Lee Jr. (2007: 117)propose that the size and regularity of the El Chayal blades recoveredsuggest evidence of craft specialization. The lack of manufacturingdebris of El Chayal obsidian at El Cerrito, however, suggested that the

specialized production of El Chayal blades was not undertaken at thatcenter. Likewise, the Terminal Formative evidence from Izapa, with thepresence of wide blade fragments of El Chayal obsidian and overall lackof production debris from the same source, suggests that completed ElChayal blades were being imported by Izapa residents during theTerminal Formative period.

5.2. Initial Early Classic period (250–400 CE) results

Obsidian recovered from domestic occupation at Mound 260 re-presented the beginning of the Early Classic period. This was the eraimmediately following the decline of many southern Mesoamericancapitals, when a regionalization of ceramic styles is documented. TheJaritas phase (250–400 CE) obsidian from Mound 260 included 39blades (Table 2). Of these, 20 (51.3%) were derived from the SanMartin Jilotepeque source, seven (17.9%) came from the El Chayalsource, five (12.8%) were obtained from the Pachuca source in centralMexico, and six (15.4%) came from the Tajumulco source. One finalblade (X12) was unassigned due to an erroneous result associated withthe intense curvature of the piece. The visual source assigned to thispiece, San Martin Jilotepeque, is assumed to be the correct designation.

For the non-blade obsidian from Early Classic contexts, I elected toanalyze the Suboperation 105a unit, which contained the best single-occupation domestic deposits recovered during the project. Obsidianfrom the Suboperation 105c extension, excavated to recover vesselsassociated with the Jaritas phase trash pit found in the southeastern

Table 2Obsidian sourcing results. SMJ=San Martin Jilotepeque, CHY=El Chayal, TAJ=Tajumulco, IXT= Ixtepeque, PACH=Sierra de Pachuca.

IHAP results n SMJ CHY TAJ IXT PACH # Unassigned

Terminal Formative period contexts (100 BCE–250 CE)Mound 255: Early Terminal Formative (Hato Phase)+mixed Formative period fill

Blades 74 45 (60.8%) 25 (33.8%) 4 (5.4%) 0 (0.0%) 0 (0.0%) 0 (0.0%)Non-blade obsidian⁎ – – – – – – –Pooled 74 45 (60.8%) 25 (33.8%) 4 (5.4%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

Off-mound unit Suboperation 104e: Mixed Terminal Formative (Hato and Itstapa phases)Blades 8 3 (37.5%) 4 (50.0%) 1 (12.5%) 0 (0.0%) 0 (0.0%) 0 (0.0%)Non-blade obsidian 104 8 (7.6%) 11 (10.5%) 85 (81.0%) 0 (0.0%) 0 (0.0%) 1 (1.0%)Pooled 112 11 (9.8%) 15 (13.4%) 86 (76.8%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

Initial Early Classic period contexts (250–400 CE)Mound 260: Jaritas phase

Blades 39 20 (51.3%) 7 (17.9%) 6 (15.4%) 0 (0.0%) 5 (12.8%) 1 (2.6%)Non-blade obsidian⁎⁎ 338 112 (33.1%) 6 (1.8%) 220 (65.1%) 0 (0.0%) 0 (0.0%) 0 (0.0%)Pooled 377 132 (35.0%) 13 (3.4%) 226 (59.9%) 0 (0.0%) 5 (1.3%) 1 (0.3%)

⁎ Not tested.⁎⁎ Non-blade obsidian tested for Suboperations 105a and 105c (the extension) only. Results are pooled for both units.

0

2

4

6

8

10

12

14

16

18

20

4-5 6-7 8-9 10-11 12-13 14-15 16-17 18-19 20-21 22-23

Coun

t

Width (mm)

Mound 255

SMJ El Chayal Tajumulco

Fig. 7. Blade widths from Terminal Formative contexts (Suboperations 104a, 104b, 104c, 104d, and 104e), plotted by obsidian source (left). Example of a large bladeof El Chayal obsidian recovered from Mound 255 (right).

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

642

corner of Suboperation 105a, was also analyzed. The results of non-blade obsidian from Early Classic contexts in Suboperations 105a and105c (Table 2) demonstrated a different profile than the TerminalFormative results. Rates of Tajumulco obsidian also dominated the as-semblage, though at slightly lower rates than Terminal Formativecontexts, with 67.1% percent in Suboperation 105a and 42.9% in theSuboperation 105c extension. The proportion of El Chayal obsidian,however, dropped significantly in Jaritas phase contexts. El Chayalobsidian represented only 1.9% of the non-blade material in Sub-operation 105a (n=338), down from 10.6% in Suboperation 104e(n=104). The proportion of San Martin Jilotepeque obsidian increasedsubstantially from 7.7% in the Terminal Formative period to 31% in theSuboperation 105a context. San Martin Jilotepeque obsidian also ac-counted for over half (57.1%) of the obsidian recovered from theSuboperation 105c trash pit.

These findings suggest that Izapeños in this section of the site eitherhad reduced access to El Chayal obsidian by the Jaritas phase, or thattheir preferences for obsidian procurement had changed. The highpercentage of San Martin Jilotepeque obsidian associated with theJaritas phase trash pit, and the large flakes of this material, includingsome with traces of cortex, (Mendelsohn, 2017: Fig. 7.3), suggest thatSan Martin Jilotepeque obsidian was being imported to Izapa as coresand that the flaking of this material by Jaritas phase residents tookplace at Mound 260. The five blades of Pachuca obsidian recoveredfrom Mound 260 currently represent the earliest known appearance ofthis central Mexican source at Izapa. Izapeños may have supplementeda decline in the availability of completed El Chayal blades with finishedblades from this newly available source.

Yet, despite its reduced rates, the continued appearance of El Chayalobsidian at Mound 260 is significant. Seven blades (17.9%) and sixpieces (1.8%) of non-blade obsidian from this source were recoveredfrom Jaritas phase contexts. While these rates of El Chayal obsidian arelower than in the Terminal Formative contexts, their presence at Mound260 suggests that distribution networks of El Chayal obsidian did notcompletely collapse with the 100–250 CE decline. This is a departurefrom the ceramic data, which suggest regionalization during the Jaritasphase and a notable decline in stylistic ties with peoples of theMiraflores ceramic sphere. The implications of this discrepancy arediscussed further below.

6. Discussion

Though preliminary, the Terminal Formative and Early Classicperiod obsidian provenance data from the IHAP offer important cluesabout trade at Izapa before and after the widespread decline of southernMesoamerican cities between 100 and 250 CE. Changing proportions ofEl Chayal obsidian are observed over the Formative to Classic periodtransition, mirroring the increase and subsequent decrease in ceramicties between Izapa and the Miraflores ceramic sphere of highlandGuatemala and western El Salvador. By the Jaritas phase (250–400 CE),San Martin Jilotepeque becomes the predominant source ofGuatemalan obsidian used in southern Izapa. Completed blades ofSierra de Pachuca obsidian are also introduced to the assemblage.Despite these changes, the continued availability of El Chayal obsidianat Izapa during the Jaritas phase suggests that obsidian distributionnetworks may have operated independently of political ties, at leastafter the decline of prominent Terminal Formative kingdoms.

The IHAP results suggest that the availability of El Chayal blades toconsumers at Izapa increased during the Terminal Formative period.Three different possibilities, or a combination thereof, might reasonablyexplain this result. First, residents of Izapa may have shifted their re-liance from a highland trade network to a coastal trade network.Second, changes in the obsidian industry at the San Martin Jilotepequesource may have meant that Izapeños' preferred (or at least, traditional)obsidian source became more difficult to obtain. Finally, the increasingimportation of ready-made blades from the El Chayal source may have

been associated with a burgeoning export industry of El Chayal bladesat Kaminaljuyú.

The first possibility is a shift in the reliance of different distributionnetworks from highland to coastal during the Terminal Formativeperiod. Clark proposes different highland and coastal procurementnetworks for San Martin Jilotepeque and El Chayal obsidian, respec-tively, dating to the Early Formative period (Clark and Lee Jr., 2007;Clark and Salcedo Romero, 1989). By the Late Formative period, Love(2011) posits three exchange spheres, highland distribution of SanMartin Jilotepeque, coastal circulation of El Chayal, and an easterndispersal of Ixtepeque obsidian.

In this scenario, the Terminal Formative increase in the importationof El Chayal obsidian at Izapa would indicate increasing use of thecoastal trade network during that period, while the Early Classic use ofSan Martin Jilotepeque obsidian could indicate a return to obsidianprocurement through highland networks. While it is possible thatIzapeños shifted their reliance to a coastal network for obsidian pro-curement during the Terminal Formative period, it seems unlikely thatIzapeños suddenly gave up their established system for procuring SanMartin Jilotepeque obsidian from highland networks unless other po-litical and economic factors were at play.

One such factor may have been the decreasing availability of ob-sidian from the San Martin Jilotepeque source. Braswell (2002) hassuggested, based on his survey results in the San Martin Jilotepequezone, that the region surrounding this source was largely abandonedduring the Late and Terminal Formative periods, from 300 BCE to250 CE. The area was later heavily reoccupied during the Early Classicperiod, from 250–600 CE (Braswell, 2002).

The settlement patterns of San Martin Jilotepeque presented byBraswell correspond remarkably well to the obsidian data from Izapa.During the Middle to Late Formative (Clark et al., 1989) and EarlyClassic periods, obsidian from the San Martin Jilotepeque source atIzapa appears to be at its highest levels. Clark and colleagues suggestthat macro-cores of San Martin Jilotepeque obsidian were distributed toregional capitals such as Izapa at this time (Clark and Lee Jr., 2007;Clark et al., 1989). During the Terminal Formative period, the rate ofSan Martin Jilotepeque obsidian dropped and the importation of ElChayal blades increased. At Izapa, as elsewhere in Mesoamerica, thedecline in San Martin Jilotepeque obsidian during the Terminal For-mative period may have been associated with a depopulation of thearea surrounding this source.

The final possibility is that an incentive arose to import El Chayalobsidian. The availability of ready-made blades for export entering themarket may have provided one such motivation. Increasing speciali-zation in blade production, where specialists at Kaminaljuyú wereproducing blades for export, is recorded beginning in the LateFormative period (500–200 BCE), with output greatly increasing by theEarly Classic period (200–400 CE) (Hay, 1978). While comparable datais not available for the Terminal Formative period (200 BCE–200 CE) inbetween, Hay estimates that during the interceding period, “productionfor export was differentiated from the remainder of Kaminaljuyu'seconomy, and became centralized and more highly specialized” (Hay,1978: 142). Clark and Lee Jr. (2007) have suggested that the wideTerminal Formative blade segments imported into inland Chiapas, re-covered at sites such as El Cerrito, were likely imported from specia-lized workshops at Kaminaljuyú. Another possibility is importationfrom nearby coastal centers, such as El Ujuxte, where prismatic bladecores of El Chayal obsidian have been recovered (Michael Love, per-sonal communication 2018), though evidence of sufficient productionto suggest an export industry has not yet been documented for that site.

Given that the El Chayal blades recovered from IHAP contexts ap-pear to have been imported as final products, and that the production ofEl Chayal blades for export has been documented at Kaminaljuyú (Hay,1978), the Terminal Formative obsidian data at Izapa appear to supportincreased trade between consumers at Izapa and blade makers fromKaminaljuyú. An increased economic relationship between residents of

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

643

Izapa and Kaminaljuyú during the Hato phase would support theceramic and burial data at Izapa, which suggest an increase in tiesbetween Izapeños and residents of the Miraflores ceramic sphere, in-cluding peoples of Kaminaljuyú, at this time (Lowe, 1993; Lowe et al.,1982: 139). Though the dating of Izapan style monuments is less cer-tain, this art style could also have been introduced to Kaminaljuyúduring the Terminal Formative period.

Given these close ceramic ties to the Miraflores sphere, the lack ofIxtepeque obsidian in Terminal Formative contexts was surprising. NoIxtepeque obsidian was recovered from Mound 255 or the adjacent off-mound unit, Suboperation 104e. At present, the only documentedcontexts of Ixtepeque obsidian at Izapa are from Terminal Classic (or“Early Postclassic”) deposits (Clark et al., 1989). This suggests that,while Hato phase ceramics from Izapa indicate participation in a hor-izon style reaching at least as far as Chalchuapa, and pots from ElSalvador were recovered from offerings at Izapa, the obsidian sourcecharacteristic of western Salvadoran sites is not found in domesticcontexts at Izapa during the Terminal Formative period. The presenceof pots imported from El Salvador in elite burials at Izapa, but lack ofIxtepeque obsidian during this period, suggest that Izapeños' relation-ships with the peoples of El Salvador was social and/or political, but noteconomic during the Terminal Formative period.

The Jaritas phase pattern of higher rates of obsidian from the SanMartin Jilotepeque source more closely approximate the Middle to LateFormative results for Izapa observed by Clark et al. (1989). This patternsuggests one of three things about the Jaritas phase economy at Izapa.The first possibility is that, after a fleeting period during the TerminalFormative period when El Chayal obsidian was popular at Izapa, Iza-peños returned to their previous procurement system of San MartinJilotepeque obsidian, perhaps in association with a repopulation in thesource zone (Braswell, 2002). The second possibility is that El Chayalblades were no longer available to the residents at Izapa or were ac-cessible in diminishing quantities due to sociopolitical turmoil at thehighland capital of Kaminaljuyú. The final possibility is that El Chayalobsidian was still popular at Izapa but may have become more re-stricted to the elite at the site, and, therefore, did not trickle down tocommoners in peripheral zones, such as the residents of Mound 260.

Of these possibilities, the decline in the availability of ready-madeEl Chayal blades due to a break in political ties between Izapa andKaminaljuyú, currently seems the most likely. Ceramic and burial datasuggest a sharp break in political and social ties between Izapa and sitesassociated with the Miraflores ceramic sphere after the TerminalFormative period (Lowe et al., 1982: 145; Mendelsohn, 2018a). Thismay have been associated with turmoil at the highland Guatemalancapital of Kaminaljuyú, where evidence for desiccation of the LagunaMiraflores, monument destruction, ritual termination events, and evena possible population displacement have been presented for the periodbeginning around 150 CE (Arroyo, 2015; Popenoe De Hatch, 1998;Popenoe de Hatch et al., 2002). While specialized production of ElChayal blades at Kaminaljuyú appears to have increased during theEarly Classic period (Hay, 1978), sociopolitical disruption documentedat the Guatemalan highland capital may have impacted the distributionof these blades to the Pacific coast.

The presence of obsidian from both San Martin Jilotepeque and ElChayal at Mound 260, however, suggests that Jaritas-phase Izapeñoswere still obtaining obsidian from Guatemala through trade networks.While this statement may seem self-evident given the proximity of theobsidian sources in Guatemala, it is at odds with the ceramic data,which suggest a marked break in imported ceramics and shared ceramicstyles with highland Guatemala by the onset of the Early Classic period.These results indicate that, even while social relationships with cities inthe Miraflores ceramic sphere dissolved by this time and potters de-veloped increasing regionalization in ceramic styles, mechanisms forthe distribution of Guatemalan obsidian were still in operation. TheJaritas phase results from the IHAP suggest that the obsidian economyduring the initial Early Classic period may have been conducted

independently of social and political ties.Another new finding from the IHAP obsidian analyses is that green

obsidian from the Sierra de Pachuca source in Hidalgo, Mexico, wasreaching Izapa by the first half of the Early Classic period. The dis-covery of the green obsidian from a Jaritas phase (250–400 CE) do-mestic context currently represents the earliest recorded use of Pachucaobsidian at the site. Initial results did not record Pachuca obsidian untilthe Late Classic period (Clark et al., 1989), though findings of this greenobsidian at Izapa in Loros phase (500–600 CE) contexts have since cometo light (Clark and Lee Jr., 2018). Five blades of Pachuca obsidian wererecovered during the IHAP excavations at Mound 260. Given thatMound 260 was occupied exclusively within the Jaritas phase, all thesepieces can be confidently dated to this era. The discovery of Pachucaobsidian from a Jaritas phase domestic context at Izapa suggests thatsome level of exchange was occurring between Izapeños and Teoti-huacanos before 400 CE.

The Mound 260 results provide support for Bove and MedranoBusto's (2003) proposal that Teotihuacan contact may have been earlieron the Pacific coast than in the adjacent highlands, and that these earlyactivities did not involve conquest or control by Teotihuacanos. As atthe Guatemalan coastal site of Balberta (Bove, 1993), there is nothing inthe Jaritas phase ceramic complex or burial data at Izapa to suggestparticularly close ties – certainly not conquest – by Teotihuacanos atthis time. While economic data from the Jaritas phase are still limited,the discovery of Pachuca obsidian in non-elite domestic contexts atMound 260 would appear to support Bove and Medrano Busto's (2003)projection of an initial relationship between Teotihuacanos and re-sidents of the Pacific coast based on mutual exchange.

7. Conclusion

Obsidian provenance data from the IHAP suggest that, whereasmany capital cities collapsed at the close of the Formative period(100–250 CE), disrupting trade networks, residents at Izapa respondedby importing obsidian from other easily-obtainable sources. Theshifting patterns of obsidian consumption over the Formative to Classicperiod transition suggest that Izapeños capitalized on multiple optionsfor trade, accounting for waning exports of pre-made El Chayal bladesto this region by increasing their reliance on other available options.The pattern apparent from the IHAP results is that when importantcities and trade networks collapsed, Izapeños imported obsidian fromother sources.

This flexibility was likely made possible by the interest of manyMesoamerican cultures, from the Olmec, to the Maya, to theTeotihuacanos, to the Aztecs, in products from the Pacific coast. Despitefrequent depiction of this area as peripheral (e.g., Parsons, 1978), dueto the Soconusco region's role as a producer of valuable commoditiessuch as cacao, the Pacific coast was never truly in danger of becomingmarginal in the Mesoamerican world system. The very central role ofthe Pacific coast in Mesoamerican commerce allowed economic flex-ibility for Izapa residents.

Whereas many centers collapsed with the sociopolitical turmoil atthe southern Terminal Formative capitals, Izapa residents appear tohave responded by adjusting their trade ties from east to west. Theincreased presence of Pachuca obsidian during the Early Classic Jaritasphase suggests that this transition may have had as much to do with theIzapeños' desire to capitalize on the expanding economy atTeotihuacan, as it did with declining capitals to the southeast. Theability of Izapa residents to import goods from several distinct networksmay have given the population a competitive advantage. In the case ofthe Formative to Classic period transition, shifts in obsidian importationallowed Izapeños to continue business as usual as they sustained oc-cupation at the site for the next 1000 years.

Yet the continued appearance of Guatemalan obsidian at Izapa after250 CE, suggests that economic relationships over the Formative toClassic period transition may have been conducted independently from

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

644

social and political ties. The presence of El Chayal obsidian in initialEarly Classic contexts, in particular, suggests that Izapeños' economicrelationships may have continued with peoples of the Miraflores city ofKaminaljuyú despite a decline in ceramic ties and evidence for politicalturmoil at the distant capital. Obsidian sourcing results from Izapa thuspresent a cautionary tale against assumptions that economic systemswill collapse as sociopolitical systems falter. Archaeologists in-vestigating the collapse and resilience of ancient societies will reveal amore complete picture of important cultural transitions if datasets fromdifferent artifact classes are evaluated independently.

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jasrep.2018.06.002.

Acknowledgements

The Izapa Household Archaeology Project was conducted with thehelp of excavation and lab supervisors Gabriela Galicia Castillo, RoymaGutiérrez Garcia, Saskia Kucknicki, and Kelly Rich. I am thankful to LisaAnderson and Julie Weatherwax of the New York State Museum for useof the XRF and to Marilyn Masson and William Lanford (University atAlbany) for providing the control samples used. I thank Hector Neff,Jeffrey Ferguson, and Michael Glascock for helping me to identify andresolve some of the problems I encountered during the initial XRFanalysis. This paper has also benefitted from comments by MarilynMasson, John Clark, and Robert Rosenswig, Jeffrey Ferguson, andMichael Love. Funding: This work was supported by a National ScienceFoundation Doctoral Dissertation Research Improvement Grant(#1349916), a Fulbright-García Robles award (Mexico, 2013–2014), aDeCormier award from the Institute for Mesoamerican Studies, and anexcavation permit from the Consejo de Arquología of Mexico's InstitutoNacional de Antropología e Historia.

References

Aoyama, K., 1996. Exchange, Craft Specialization, and Ancient Maya State Formation: AStudy of Chipped Stone Artifacts from the Southeast Maya Lowlands. Doctoraldissertation. University of Pittsburgh.

Aoyama, K., 2001. Classic Maya state, urbanism, and exchange: chipped stone evidence ofthe Copán Valley and its hinterland. Am. Anthropol. 103, 346–360.

Arroyo, B., 2015. El Altiplano Central Maya, Kaminaljuyú y Sus Vecinos. Arqueol. Mex.134, 50–55.

Asaro, F., 1978. High-precision chemical characterization of major obsidian sources inGuatemala. Am. Antiq. 43, 436–443.

Blanton, R.E., Kowalewski, S.A., Feinman, G.M., 1978. The Mesoamerican world-system.In: Review (Fernand Braudel Center). 15. pp. 419–426.

Blomster, J.P., Neff, H., Glascock, M.D., 2005. Olmec pottery production and export inancient Mexico determined through elemental analysis. Science 307, 1068–1072.

Bove, F.J., 1993. The Terminal Formative-Early Classic transition. In: Bove, F.J.,Medrano, B.S., Lou, P.B., Arroyo, L.B. (Eds.), The Balberta Project: The TerminalFormative-Early Classic Transition on the Pacific Coast of Guatemala. University ofPittsburgh Memoirs in Latin American Archaeology, No. 6. University of Pittsburgh,Department of Anthropology, Pittsburgh, pp. 177–194.

Bove, F.J., Medrano Busto, S., 2003. Teotihuacan, militarism and Pacific coastalGuatemala. In: Braswell, G.E. (Ed.), Teotihuacan and the Maya: Reinterpreting EarlyClassic Interaction. University of Texas Press, Austin, pp. 45–79.

Bowser, B.J., 2000. From pottery to politics: an ethnoarchaeological study of politicalfactionalism, ethnicity, and domestic pottery style in the Ecuadorian Amazon. J.Archaeol. Method Theory 7, 219–248.

Braswell, G.E., 2002. Praise the Gods and pass the obsidian? The organization of ancienteconomy in San Martín Jilotepeque, Guatemala. In: Masson, M.A., Freidel, D.A.(Eds.), Ancient Maya Political Economies. Altamira Press, New York, pp. 285–306.

Braswell, G.E., 2003a. The Maya and Teotihuacan: Reinterpreting Early ClassicInteraction. University of Texas Press, Austin.

Braswell, G.E., 2003b. Obsidian exchange spheres. In: Smith, M.E., Berdan, F.F. (Eds.),The Postclassic Mesoamerican World. University of Utah Press, Salt Lake City, pp.131–158.

Braswell, G.E., Glascock, M.D., 2002. The emergence of market economies in the ancientMaya world: obsidian exchange in Terminal Classic Yucatán, Mexico. In: Glascock,M.D. (Ed.), Geochemical Evidence for Long-Distance Exchange. Bergin and Garvey,Westport, pp. 33–52.

Braswell, G., Andrews, E.E.W.V., Glascock, M.D., 1994. The obsidian artifacts of Quelepa,El Salvador. Anc. Mesoam. 5, 173–192.

Braswell, G.E., Clark, J.E., Aoyama, K., McKillop, H.I., Glascock, M.D., 2000. Determiningthe geological provenance of obsidian artifacts from the Maya region: a test of theefficacy of visual sourcing. Lat. Am. Antiq. 11, 269–282.

Clark, J.E., 1987. Politics, prismatic blades, and Mesoamerican civilization. In: Johnson,J.K., Morrow, C.A. (Eds.), The Organization of Core Technology. Westview Press,New York, pp. 259–284.

Clark, J.E., Bryant, D.D., 1997. A technological typology of prismatic blades and debitagefrom Ojo de Agua, Chiapas, Mexico. Anc. Mesoam. 111–136.

Clark, J.E., Cheetham, David, 2005. Cerámica del Formativo de Chiapas. In: Garrión,Merino, Leonor, Beatriz, García Cook, Ángel (Eds.), Le Producción Alfarera en elMéxico Antiguo I. Instituto Nacional de Antropología e Historia, México, pp.285–433.

Clark, J.E., Lee Jr., T.A., 2007. The Changing Role of Obsidian Exchange in CentralChiapas. In: Lowe, L.S., Pye, M.E. (Eds.), Archaeology Art, and Ethnogenesis inMesoamerican Prehistory: Papers in Honor of Gareth W. Lowe. Papers of the NewWorld Archaeological Foundation, No. 68. Brigham Young University, Provo, pp.109–159.

Clark, J.E., Lee Jr., T.A., 2013. Minor excavations in Lower Izapa. In: Papers of the NewWorld Archaeological Foundation No. 75. Brigham Young University, Provo.

Clark, J.E., Lee Jr., T.A., 2018. A Touch of Teotihuacan at Izapa: The Contents of TwoBurials from Group F. Anc. Mesoam. 29 (2) (in press).

Clark, J.E., Salcedo Romero, T., 1989. Ocós obsidian distribution in Chiapas, Mexico. In:Bove, F., Heller, L. (Eds.), New Frontiers in the Archaeology of the Pacific Coast ofSouthern Mesoamerica. Arizona State University Anthropological Research PapersNo. 39, Tempe, pp. 15–25.

Clark, J.E., Lee Jr., T.A., Salcedo, T., 1989. The distribution of obsidian. In: Voorhies, B.(Ed.), Ancient Trade and Tribute: Economies of the Soconusco Region ofMesoamerica. University of Utah Press, Salt Lake City, pp. 268–284.

Cobean, R.H., Coe, M.D., Perry, E.A., Turekian, K.D., Kharkar, D.P., 1971. Obsidian tradeat San Lorenzo Tenochtitlan, Mexico. Science 174, 666–671.

Coe, M.D., Flannery, K.V., 1967. Early cultures and human ecology in south coastalGuatemala. In: Smithsonian Contributions to Anthropology. 3 Smithsonian Press,Washington.

David, N., Sterner, J., Gavua, Kodzo, 1988. Why pots are decorated. Curr. Anthropol. 29,365–379.

DeLeón, J.P., Hirth, K.G., Carballo, D.M., 2009. Exploring Formative period obsidiantrade: three distribution models. Anc. Mesoam. 20, 113–128.

Demarest, A.A., Sharer, R.J., 1986. Late Preclassic ceramic spheres, culture areas, andcultural evolution in the southeastern highlands of Mesoamerica. In: Urban, P.A.,Schortman, E.M. (Eds.), The Southeast Maya Periphery. University of Texas Press,Austin, pp. 194–223.

Doyle, J.A., 2017. Architecture and the origins of Preclassic Maya politics. CambridgeUniversity Press, New York.

Drucker, P., 1948. Preliminary Notes on an Archaeological Survey of the Chiapas Coast.Middle American Research Records. 1. pp. 151–169.

Estrada-Belli, F., 2011. The First Maya Civilization: Ritual and Power Before the ClassicPeriod. Routledge, New York.

Fox, R., 1977. Urban Anthropology. Prentice Hall, Englewood Cliffs.Frank, A.G., 1993. Bronze Age world system cycles. Curr. Anthropol. 34, 383–429.Freidel, D.A., Schele, L., 1988. Kingship in the Late Preclassic Maya lowlands: the in-

struments and places of ritual power. Am. Anthropol. 90, 547–567.García-Des Lauriers, C., 2007. Proyecto Arqueológico Los Horcones: Investigating

Teotihuacan Presence on the Pacific Coast of Chiapas, Mexico. Ph.D. Dissertation.Department of Anthropology, University of California, Riverside.

Gasco, J., Voorhies, B., 1989. The role of the soconusco as an Aztec tributary. In:Voorhies, B. (Ed.), Ancient Trade and Tribute: Economies of the Soconusco Region ofMesoamerica. University of Utah Press, Salt Lake City, pp. 48–94.

Golitko, M., Feinman, G.M., 2015. Procurement and distribution of Pre-HispanicMesoamerican obsidian 900 BC-AD 1520: a social network analysis. J. Archaeol.Method Theory 22, 206–247.

Gómez Rueda, H., 1995. Exploración de Sistemas Hidráulicos en Izapa. In: Laporte, J.P.,Escobedo, H. (Eds.), VIII Simposio de Investigaciones Arqueológicas en Guatemala,1994. Museo Nacional de Arqueología y Etnología, Guatemala, pp. 6–16.

Grube, N., 1995. The emergence of lowland Maya civilization: The transition from thePreclassic to the Early Classic. In: Grube, N. (Ed.), Acta Mesoamericana. Vol. 8 VerlagAnton Saurwein, Germany.

Guernsey, J., 2006. Ritual & Power in Stone: The Performance of Rulership inMesoamerican Izapan Style Art. University of Texas Press, Austin.

Guernsey, J., Love, M., Mendelsohn, R., Romero, S., Turner, A., 2018. Late FormativeFeline Pedestal Sculptures and an Iconography of Cacao. Mexicon, (In Press).

Hay, C.A., 1978. Kaminaljuyu Obsidian: Lithic Analysis and the Economic Organization ofa Prehistoric Mayan Chiefdom. Unpublished Ph.D. dissertation. Department ofAnthropology, Pennsylvania State University, University Park.

Hegmon, M., Kulow, S., 2005. Painting as agency, style as structure: innovations inMimbres pottery designs from Southwest New Mexico. J. Archaeol. Method Theory12, 313–334.

Hirth, K.G., 1998. The distributional approach: a new way to identify marketplace ex-change in the archaeological record. Curr. Anthropol. 39, 451–476.

Hodos, T., 2017. The Routledge Handbook of Archaeology and Globalization. Routledge,New York.

Inomata, T., Henderson, L., 2016. Time tested: re-thinking chronology and sculpturaltraditions in Preclassic southern Mesoamerica. Antiquity 90, 456–471.

Inomata, T., Ortiz, R., Arroyo, B., Robinson, E.J., 2014. Chronological revision ofPreclassic Kaminaljuyú, Guatemala: implications for social processes in the SouthernMaya area. Lat. Am. Antiq. 25, 377–408.

Jackson, T.L., Love, M.W., 1991. Middle Preclassic obsidian exchange and the introduc-tion of prismatic blades at La Blanca, Guatemala. Anc. Mesoam. 2, 47–59.

Jennings, J., 2011. Globalizations and the Ancient World. Cambridge University Press,New York.

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

645

Kaplan, J., 2002. From under the volcanoes: some aspects of ideology of rulership at LatePreclassic Kaminaljuyú. In: Love, M., Popenoe de Hatch, M., Escobedo, H.L. (Eds.),Incidents of Archaeology in Central America and Yucatán: Essays in Honor of EdwinM. Shook. University Press of America, Lanham, pp. 311–357.

Kaplan, J., 2011. Miraflores Kaminaljuyu: corpse and corpus delicti. In: Love, M., Kaplan,J. (Eds.), The Southern Maya of the Late Preclassic: The Rise and Fall of an EarlyMesoamerican Civilization. University Press of Colorado, Boulder, pp. 237–286.

Lieske, R., 2013. Izapa Group B: Excavations, Burials, and Offerings. M.A. thesis.Department of Anthropology, Brigham Young University, Provo.

Love, M., 2002. Ceramic chronology of Preclassic period Western Pacific Guatemala andits relationship to other regions. In: Love, M., Popenoe de Hatch, M., Escobedo, H.L.(Eds.), Incidents of Archaeology in Central America and Yucatán: Essays in Honor ofEdwin M. Shook. University Press of America, Lanham, pp. 71–73.

Love, M., 2007. Recent research in the southern highlands and Pacific coast ofMesoamerica. J. Archaeol. Res. 15, 275–328.

Love, M., 2011. Cities, states, and city-state culture in the Late Preclassic Southern Mayaregion. In: Love, M., Kaplan, J. (Eds.), The Southern Maya of the Late Preclassic: theRise and Fall of an Early Mesoamerican Civilization. University Press of Colorado,Boulder, pp. 47–75.

Love, M., 2012. The development of complex societies in Formative-period PacificGuatemala and Chiapas. In: Nichols, D.L., Pool, C.A. (Eds.), The Oxford Handbook ofMesoamerican Archaeology. Oxford University Press, New York, pp. 200–214.

Love, M., 2018. Kaminaljuyú, chronology, and ceramic analysis: an alternate view. Lat.Am. Antiq. 29 (2), 260–278.

Lowe, G.W., 1977. The Mixe-Zoque and competing neighbors of the early Lowland Maya.In: Adams, R.E.W. (Ed.), The Origins of Maya Civilization. University of New MexicoPress, Albuquerque, pp. 197–248.

Lowe, G.W., 1993. La Gran Importancia de las Ofrendas del Preclásico Terminal en Izapa.In: Segundo y Tercer Foros de Arqueología de Chiapas. Instituto Chiapaneco deCultura, Tuxtla Gutiérrez, pp. 179–210.

Lowe, G., Lee Jr., T.A., Espinoza, E.M., 1982. Izapa: an introduction to the ruins andmonuments. In: Papers of the New World Archaeological Foundation 31. BrighamYoung University, Provo.

Masson, M.A., Hare, T.S., Peraza Lope, C., 2010. Postclassic Maya Society regenerated atMayapán. In: Schwartz, G.M., Nichols, J.J. (Eds.), Beyond Collapse. University ofArizona Press, Tucson, pp. 188–207.

Mendelsohn, R.R., 2017. Resilience and Interregional Interaction at the EarlyMesoamerican City of Izapa: The Formative to Classic Period Transition. Ph.D.dissertation. Department of Anthropology, University at Albany, SUNY, Albany, NY.

Mendelsohn, R.R., 2018a. The Formative to Classic period transition at Izapa: updatesfrom the Izapa Household Archaeology Project. Anc. Mesoam. 29 (2) (in press).

Mendelsohn, R.R., 2018b. The chronology of the Formative to Classic period transition atIzapa: a reevaluation. Lat. Am. Antiq. 29 (2), 239–259.

Moholy-Nagy, H., Meierhoff, J., Golitko, M., Kestle, C., 2013. An analysis of pXRF ob-sidian source attributions from Tikal, Guatemala. Lat. Am. Antiq. 24, 72–97.

Neff, H., Bishop, R.L., Bove, F.J., 1989. Compositional patterning in ceramics from Pacificcoastal and highland Guatemala. Archaeomaterials 3, 97–109.

Neff, H., Bove, F.J., Robinson, E.J., Arroyo, L.B., 1994. A ceramic compositional per-spective on the Formative to Classic period transition in southern Mesoamerica. Lat.Am. Antiq. 5, 333–358.

Neivens, M., Demarest, A.A., 1986. Technical Comments on Trace Element Sourcing ofSanta Leticia Obsidian Artifacts. In: Demarest, A.A. (Ed.), The Archaeology of SantaLeticia and the Rise of Maya Civilization. Middle American Research Institute,Publication 52. Tulane University, New Orleans, pp. 217–218.

Nichols, D.L., Brumfiel, E.M., Neff, H., Hodge, M., Charlton, T.H., Glascock, M.D., 2002.Neutrons, markets, cities, and empires: a 1000-Year perspective on ceramic pro-duction and distribution in the postclassic basin of mexico. J. Anthropol. Archaeol.21, 25–82.

Norman, V.G., 1976. Izapa Sculpture, Part 2: Text. Papers of the New WorldArchaeological Foundation No. 30. Brigham Young University, Provo.

Parsons, L.A., 1978. The peripheral coastal lowlands and the Middle Classic Period. In:Pasztory, E. (Ed.), Middle Classic Mesoamerica: A.D. 400–700. Columbia University

Press, New York, pp. 25–34.Parsons, L.A., 1986. The origins of Maya Art: monumental stone sculpture of

Kaminaljuyu, Guatemala, and the Southern Pacific Coast. In: Studies in Pre-Columbian Art & Archaeology, Number 28. Dumbarton Oaks Research Library andCollection, Washington, D.C..

Pfeiffer, L., 1983. Pottery Production and Extralocal Relations at Rio Arriba, Chiapas,Mexico. Ph.D. dissertation. Department of Anthropology, University of California,Santa Barbara.

Plog, S., 1983. Analysis of style in artifacts. Annu. Rev. Archaeol. 12, 125–142.Polanyi, K., 1957. The economy as instituted process. In: Polanyi, K., Arensburg, C.M.,

Pearson, H.W. (Eds.), Trade and Market in the Early Empires. Regnery, Chicago, pp.243–270.

Pool, C.A., Loughlin, M.L., 2015. The evolution of a resilient polity in the Olmec heartlandof Mexico. In: Tres Zapotes, Faulseit, R.K. (Eds.), Beyond Collapse: ArchaeologicalPerspectives on Resilience, Revitalization, and Transformation in Complex Societies.Southern University Press, Carbondale, pp. 287–309.

Pool, C.A., Loughlin, M.L., Ceballos, P.O., 2018. Transisthmian ties: Epi-Olmec and Izapaninteraction. Anc. Mesoam. 29 (2) (in press).

Popenoe de Hatch, M., 1998. Los K'iche's-Kaqchikeles en el Altiplano Central deGuatemala: Evidencia Arqueológica del Periodo Clásico. Mesoamérica 35, 93–115.

Popenoe de Hatch, M., Ponciano, E., Barrientos, Q.T., Brenner, M., Ortloff, C., 2002.Climate and technological innovation at Kaminaljuyu, Guatemala. Anc. Mesoam. 13,103–114.

Reese-Taylor, K., Walker, D.S., 2002. The passage of the Late Preclassic into the EarlyClassic. In: Masson, M.A., Freidel, D.A. (Eds.), Ancient Maya Political Economies.Altamira Press, New York, pp. 87–122.

Rosenswig, R.M., Mendelsohn, R.R., 2016. Izapa and the Soconusco Region, Mexico, inthe first millennium A.D. Lat. Am. Antiq. 27, 357–377.

Rosenswig, R.M., López-Torrijos, R., Antonelli, C.E., Mendelsohn, R.R., 2013. Lidarmapping and surface survey of the Izapa state on the tropical piedmont of Chiapas,Mexico. J. Archaeol. Sci. 40, 1493–1507.

Rosenswig, R.M., Culleton, B., Kennett, D., Lieske, R., Mendelsohn, R., Nuñez Cortés, Y.,2018. The early occupation of Izapa: recent excavations, new Middle Formativedating and ceramic analyses. Anc. Mesoam. 29 (2) (in press).

Sanders, W.T., Webster, D., 1988. The Mesoamerican urban tradition. Am. Anthropol. 90,521–546.

Shook, E.M., 1965. Archaeological survey of the pacific coast of Guatemala. In: Willey, G.(Ed.), Archaeology of Southern Mesoamerica: Part One, Handbook of MiddleAmerican Indians, Volume Two. University of Texas Press, Austin, pp. 180–194.

Sidrys, R.V., 1976. Classic Maya obsidian trade. Am. Antiq. 41, 449–464.Sidrys, R., Andresen, J., Marcucci, D., 1976. Obsidian sources in the Maya area. J. New

World Archaeol. 1, 1–13.Smith, V., 1984. Izapa Relief Carving: Form, Content, Rules for Design, and Role in

Mesoamerican Art History and Archaeology. Studies in Pre-Columbian Art &Archaeology, Number 27. Dumbarton Oaks Research Library and Collection,Washington, D.C.

The Postclassic Mesoamerican World. In: Smith, M.E., Berdan, F.F. (Eds.), University ofUtah Press, Salt Lake City.

Spence, M.W., 1996. Commodity or gift: Teotihuacan obsidian in the Maya region. Lat.Am. Antiq. 7, 21–39.

Stark, B.L., Boxt, M.A., Gasco, J., González-Lauck, R.B., Hedgepeth-Balkin, J.D., Joyce,A.A., King, S.M., Knight, C.L.F., Kruger, R., Levine, M.N., Lesure, R.G., Mendelsohn,R., Navarro-Castillo, M., Neff, H., Ohnersorgen, M., Pool, C., Raab, L.M., Rosenswig,R., Venter, M., Voorhies, B., Williams, D.T., Workinger, A., 2016. Economic growth inMesoamerica: obsidian consumption in the coastal lowlands. J. Anthropol. Archaeol.41, 263–282.

Stoltman, J.B., Marcus, J., Flannery, K.V., Burton, J.H., Moyle, R.G., 2005. Petrographicevidence shows that pottery exchange between the Olmec and their neighbors wastwo-way. Proc. Natl. Acad. Sci. 102, 11213–11218.

Ur, J.A., 2010. Cycles of civilization in northern Mesopotamia, 4400–2000 BC. J.Archaeol. Res. 18, 387–431.

R.R. Mendelsohn Journal of Archaeological Science: Reports 20 (2018) 634–646

646