Interregional Obsidian Exchange During the Late Initial Period and Early Horizon: New Perspectives from Campanayuq Rumi, Peru

INTERREGIONAL OBSIDIAN EXCHANGE DURING THE LATE INITIAL PERIOD AND EARLY HORIZON: NEW PERSPECTIVES FROM CAMPANAYUQ RUMI, PERU

Yuichi Matsumoto, Jason Nesbitt, Michael D. Glascock, Yuri I. Cavero Palomino, and Richard L. Burger

Campanayuq Rumi is a large civic-ceremonial center located in the south-central highlands of Peru. Founded in the late Initial Period (1100–800 BC), Campanayuq Rumi became an important center within the Chavín Interaction Sphere in the Early Horizon (ca. 800–400 BC). In particular, Campanayuq Rumi is signiﬁcant because of its geographical proximity to Quispisisa, the most important and widely circulated obsidian source during the Early Horizon. Portable X-ray ﬂorescence (pXRF) was used to geochemically source a sample of 370 obsidian artifacts from Campanayuq Rumi. Though obsidian from Quispisisa dominates the assemblage throughout the site’s history, diachronic analysis indicates that the diversity of obsidian sources increases markedly in the Campanayuq II Phase (700–450 BC). The pXRF data lead us to conclude that Campanayuq Rumi was the locus of obsidian distribution to other locations in highland and coastal Peru within the Chavín Interaction Sphere, and functioned as a regional center of worship and interaction.

Campanayuq Rumi es un centro cívico-ceremonial de grandes dimensiones localizado en la sierra centro-meridional del Perú. El sitio fue construido durante el período Inicial tardío (1100–800 aC) y se transformó en un centro ceremonial de gran importancia dentro de la esfera de interacción Chavín durante el horizonte Temprano (800–450 aC). Campanayuq Rumi es un sitio estratégico particularmente significativo por su proximidad a las canteras de Quispisisa, la fuente de obsidiana más importante en los Andes centrales, cuyos productos circularon extensivamente durante el horizonte Temprano. Se utilizó un espectrómetro de fluorescencia de rayos X portátil (pXRF, por sus siglas en inglés) para identificar geoquímicamente la proveniencia de una muestra de 370 artefactos de obsidiana recuperados en el sitio. Los resultados de estos análisis revelan que la obsidiana proveniente de Quispisisa constituye el mayor porcentaje de la muestra durante toda la historia del sitio. El análisis diacrónico indica también que la diversidad de fuentes de obsidiana incrementó considerablemente en la fase Campanayuq II (700–450 aC). Los datos de pXRF permiten llegar a la conclusión de que Campanayuq Rumi fue un lugar de distribución de obsidiana hacia otros lugares en la sierra y la costa del Perú dentro de la esfera de interacción Chavín, funcionando también como centro regional de veneración e interacción.

he Early Horizon (800–400 BC) was char- 2012; Figure 1). This network was centered at acterized by important socioeconomic the urban and ceremonial center of Chavín de Tchanges in the Central Andes of Peru. Huántar and promoted long-distance exchange, During this time, much of ancient Peru was partially through the expansion of religious closely linked in a widespread exchange network ideology (Burger 1988, 2008; Patterson 1971). referred to as the Chavín Horizon or Chavín Exotic goods comprised crucial elements of Interaction Sphere (Burger 1988, 1993, 2008, the political economy of the Chavín Interaction

Yuichi Matsumoto Faculty of Literature and Social Sciences, Yamagata University, Yamagata, Japan (Corresponding author: [email protected]) Jason Nesbitt Department of Anthropology, Tulane University, New Orleans, LA 70118, USA (Corresponding author: [email protected]) Michael D. Glascock Research Reactor Center, University of Missouri-Columbia, Columbia, MO 65211, USA Yuri I. Cavero Palomino Facultad de Letras y Ciencias Humanas, Universidad Nacional Mayor de San Marcos, Lima, Peru RichardL.Burger Department of Anthropology, Yale University, New Haven, CT 06511, USA Latin American Antiquity, page 1 of 20 Copyright © 2017 by the Society for American Archaeology doi:10.1017/laq.2017.64

1 Downloaded from https://www.cambridge.org/core. IP address: 183.91.99.232, on 20 Nov 2017 at 07:55:20, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/laq.2017.64 2 LATIN AMERICAN ANTIQUITY

Figure 1. The location of Campanayuq Rumi and other Chavín-related centers mentioned in the text. (Color online)

Downloaded from https://www.cambridge.org/core. IP address: 183.91.99.232, on 20 Nov 2017 at 07:55:20, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/laq.2017.64 [Matsumoto et al.] INTERREGIONAL OBSIDIAN EXCHANGE DURING THE LATE INITIAL PERIOD 3

Sphere (Burger 2008, 2013), with obsidian fig- shared religious ideology (Burger 1988). The uring prominently among the exchanged exotic core of this area was centered in the north materials. Geochemical analyses over the last and north-central highlands, where several large, four decades demonstrate that the important regional centers, including Chavín de Huántar, obsidian quarries are situated over a large area Kuntur Wasi, and Pacopampa interacted as peer- of the south-central highlands (Figure 2), some polities (sensu Renfrew 1986; see also Burger 600 kilometers to the south of Chavín de Huán- 1993). Within this network a variety of exotic tar (Glascock et al. 2007). One source known goods rapidly appeared, including gold jewelry, as Quispisisa was used intensively, comprising foreign pottery, cinnabar pigment, and obsidian roughly 90% of all obsidian circulated through- tools, as well as Spondylus and Strombus shells out the Chavín Interaction Sphere (Burger and from coastal Ecuador (Burger 1988, 2008, 2013; Glascock 2009). The significant distances tra- Burger and Matos 2002; Burger, Lane, and versed to obtain obsidian and other exotic goods Cooke 2016; Cooke et al. 2013;Druc2004; raises the question of precisely how these mate- Lothrop 1941; Onuki 1995;Rick2005, 2008; rials were distributed and acquired within this Sayre 2010; Seki et al. 2010). Coinciding with interaction sphere. To address this question this intensification of interaction is strong evi- requires the study of obsidian procurement from dence for the development of social hierarchies. the perspective of the south-central highlands For instance, excavations at both Kuntur Wasi rather than from obsidian consumers in the north. and Pacopampa have revealed burials of elite In this article, we detail the changing role of the individuals accompanied with gold ornamenta- south-central highlands with respect to obsidian tion, cinnabar, and Strombus shells (Onuki 1995; exchange in the Chavín Interaction Sphere. To Seki et al. 2010). illustrate these changes and their significance, The evidence for the importance of long- we present the results of a recent portable X-ray distance trade is, however, best exemplified at florescence (pXRF) sourcing study of 370 obsid- the urban and ceremonial center of Chavín de ian artifacts from the civic-ceremonial center of Huántar. Chavín de Huántar had a long history Campanayuq Rumi. Campanayuq Rumi is criti- of development that began around 1000 BC, cal to understanding the circulation of obsidian reaching its maximal size in the Janabarriu because it is one of the largest known centers in Phase, around ca. 700 BC (Burger 1984, 2008). the south-central highlands of Peru and is located Chavín de Huántar consisted of a large cere- in close proximity to the Quispisisa quarry, as monial core that included a U-shaped platform well as other important south-central highland complex that functioned as a temple (Burger obsidian sources. Furthermore, Campanayuq 1992; Kembel 2008). Archaeological excava- Rumi has a long history of occupation, permitting tions in the temple core demonstrate that since a diachronic perspective on changing patterns its foundation around 1000 BC, Chavín de of obsidian exploitation between circa 1100 and Huántar was a center that received a significant 450 BC. As a result, we were able to docu- amount of foreign goods from distant locations ment how Campanayuq Rumi became embed- (Lumbreras 2007;Rick2005, 2008; Sayre et al. ded in the Chavín Interaction Sphere through 2016). time. Chavín de Huántar also has a large residential occupation with strong evidence for craft pro- The Chavín Interaction Sphere and duction activities and status distinctions (Burger Long-Distance Exchange 1984;Rick2005; Sayre 2010). Archaeological excavations indicate that the residential com- The Chavín Interaction Sphere was a cosmopoli- ponent grew significantly during the Janabarriu tan system of interacting polities spread over Phase (ca. 700–400 BC) to cover an area of a large area of the north and central Peru- at least 40–50 ha. The explosive growth of vian highlands, coastal areas, and tropical forest Chavín de Huántar corresponded with increased regions (Burger 2008, 2012; Contreras 2011). elite demand for the exotic materials circulated While these polities exhibited a great deal of within the Chavín Interaction Sphere of the Early regional diversity, they were also linked through Horizon described above.

Figure 2. Key obsidian localities in the south-central highlands of Peru. (Color online)

Obsidian and the Chavín Interaction Sphere production (Miller and Burger 1995; Rosenfeld Diachronic studies suggest that the exploita- and Sayre 2016). Ethnographic accounts suggest tion of obsidian at Chavín de Huántar changed that obsidian would have been suitable for wool- through time (Burger et al. 1984; Burger, Lau, processing activities. Bandelier (1904:446) notes Ponte, and Glascock 2006). For instance, during that obsidian was a preferred material for Aymara the Urabarriu Phase (ca. 950–800 BC), obsidian herders to shear llamas, while Flores-Ochoa use at Chavín de Huántar was rare. The paucity (1979:94) points out that contemporary herders of obsidian suggests that obsidian did not form in isolated communities in south-highland Peru a significant component of Chavín de Huántar’s employed shards of bottle glass for shearing their political economy during its earliest occupation animals. phases. Two geochemical sourcing studies suggest In the Janabarriu Phase (ca. 700–400 BC) that roughly 90% of the obsidian from Chavín obsidian was found in significant quantities at de Huántar came from the Quispisisa source Chavín de Huántar, implying close relations with (Burger et al. 1984; Burger, Lau, Ponte, and the south-central highlands during this time. Glascock 2006; Contreras and Nado 2018). The Obsidian was found in ceremonial and temple source area of Quispisisa obsidian was a massive deposits, suggesting that the material was used quarry found near the towns of Sacsamarca and as an offering and/or ritual paraphernalia (Mesía Huanca Sancos in the Department of Ayacucho Montenegro 2014). The symbolic value of obsid- (Burger and Glascock 2000a, 2002). A recent ian could be related to its exotic provenance survey of one locality, Cerro Jichja Parco, doc- (Helms 1993), as well as its color and reflective umented more than 30 ancient quarrying pits. glass-like properties. Some pits are as large as 80 m in diameter, a testa- There are also strong indicators that during the ment to the intensity of exploitation of this source Janabarriu Phase, obsidian formed a significant in the past (Tripcevich and Contreras 2011, component of the political economy. Investiga- 2013). Smaller amounts of obsidian come from tions in Chavín de Huántar’s residential sectors the Jampatilla, Potreropampa, and Alca sources showed that obsidian was ubiquitous, replacing (Burger et al. 1984; Contreras and Nado 2018). chert, quartzite, and basalt as the dominant lithic All of these sources are significantly greater material (Burger 1984). Obsidian was distributed than 600 km to the south of Chavín de Huántar. widely in high quantities and was a common arti- The long distances involved raise the question fact in both high- and low-status households. It of how obsidian was acquired. To address this was recently argued that this distribution pattern problem requires a consideration of the situation (sensu Hirth 1998) implies that a form of market of the south-central highlands at sites closer to exchange system operated at Chavín de Huántar the sources of archaeological obsidian. in the Janabarriu Phase (Burger 2013:325–327). Obsidian was used to manufacture bifaces and Archaeological Research at Campanayuq scrapers, though the vast majority of artifacts Rumi consisted of utilized flakes and debitage (Burger 1984:190–192; Burger, Lau, Ponte, and Glas- One of the largest late Initial Period and Early cock 2006; Contreras and Nado 2018). Burger Horizon settlements in the south-central high- (1984:238) suggests that the flakes were used lands is Campanayuq Rumi, a civic-ceremonial in crafting activities such as processing animal center in the Department of Ayacucho (Figure 1). hides in the residential sectors of Chavín de The site is located at an altitude of 3,600 m Huántar. It is also plausible that obsidian flakes asl, just above the town and former Inca admin- were used as tools for shearing camelids in istrative center of Vilcashuaman. Importantly, the production of wool textiles. The increase Campanayuq Rumi is only 52 km to the east of in obsidian at Chavín de Huántar correlates the Quispisisa source area. with an intensification in the use of domestic Campanayuq Rumi is composed of a monu- camelids, especially llamas, which were valued mental core of artificial stone platforms arranged for human consumption, transport, and wool in a U-shaped pattern that covers an area of

Table 1. Chronology of Campanayuq Rumi and Other Key Sites Mentioned in the Text.

Chavín de Huántar Chavín de Huántar Pacopampa Kuntur Wasi Campanayuq Rumi (Burger) (Rick) Cal. BC Sotera Huaraz 250 Huaraz Copa 500 Janabarriu Campanayuq II Pacopampa II Kuntur Wasi 800 Chakinani Janabarroide Campanayuq I Idolo Urabarriu 1000 Pacopampa I Pre-Plaorm 1200 Pandanche

1500

Figure 3. Map of Campanayuq Rumi and the surrounding residential settlements.

3.5 ha (Figures 3 and 4), and is surrounded suggest that late Initial Period and Early Horizon to the north and south by at least two resi- settlement covered at least 12 ha. dential sectors (Matsumoto 2010;Matsumoto Using absolute and relative dating, the occu- and Cavero Palomino 2010; Matsumoto et al. pation of the site has been divided into three 2013, 2016). While modern construction activity phases (Table 1). During this time, Campanayuq makes it difﬁcult to estimate the full extension Rumi grew from a small village in the Initial of the residential areas, recent investigations Period into one of the largest ceremonial and

Figure 4. Map of the monumental core of Campanayuq Rumi exhibiting a U-shaped pattern (T1, T2, T3, T4, T6, P1, P2, P6, and P9 indicate excavation units).

residential centers of its time in the south- millennium BC (Bauer et al. 2010:53–55; Gross- central highlands (Matsumoto 2010;Matsumoto man 1972, 1983). and Cavero Palomino 2010). In the following sections we provide a brief overview of how the The Campanayuq I Phase (950–700 BC) site changed through time. The Campanayuq I Phase marks a radical change in which the site transformed into a Pre-Platform Phase (1100–950 BC) ceremonial center that was tied in important The ﬁrst phase of occupation at Campanayuq ways to Chavín de Huántar. Architecture, Rumi is referred to as the Pre-Platform Phase ceramic style, and absolute dates support this (1100–950 BC). The Pre-Platform Phase cor- view (Matsumoto 2010; Matsumoto and Cavero responds to an occupation that pre-dates the Palomino 2010; Nesbitt and Matsumoto 2014). construction of temple architecture (Matsumoto By this time, Campanayuq Rumi displays a scale 2010). Excavations in both the monumental core of monumentality that is signiﬁcantly greater and occupational contexts suggest that during the than other public centers in the Ayacucho region, Pre-Platform Phase there was a small village- such as Chupas or Wichqana (Cruzatt 1971; level Initial Period settlement at Campanayuq Lumbreras 1981). Rumi. If this was the case, Pre-Platform Phase The monumental core of Campanayuq Campanayuq Rumi was probably one of a num- Rumi forms a U-shaped pattern with four ber of small-scale settlements that characterized stone masonry platforms surrounding a sunken the south-central highlands during the second rectangular plaza. This architectural pattern

resembles the architectural layout of Chavín as well as the Early Paracas style of the south de Huántar. Shared characteristics between the coast (see Dulanto 2015; García and Pinilla 1995; two centers include the presence of at least one Isla and Reindel 2006). gallery within Campanayuq Rumi’s southern This radical change in ceramic assemblage platform that is similar to some of the gallery was associated with modifications in the mon- complexes for which Chavín de Huántar is umental architecture. For instance, Campanayuq famous (Matsumoto 2010:424–425). In addition, Rumi incorporated an innovative masonry tech- excavations in 2016 uncovered a sunken circular nology of cut and polished stone referred to as plaza on the summit of the central platform, ashlar construction (Figure 5). This new tech- which is highly similar to the one found at nique is especially important because it was Chavín de Huántar (Lumbreras 2014). popularized by the Chavín de Huántar temple The combination of a U-shaped arrangement (e.g., Kembel 2008). It is also during this time and the use of an interior gallery indicate a close that the residential settlement of Campanayuq relationship with Chavín de Huántar, despite a Rumi was occupied more intensively (Figure 3) distance of over 600 km separating these two cen- (Matsumoto et al. 2013, 2016). ters (Matsumoto 2010). Given the greater scale The changes in material culture observed and elaboration of the Chavín de Huántar public during the Campanayuq II Phase serve as a proxy architecture, as well as earlier radiocarbon dates for significant socioeconomic transformations (Kembel and Haas 2015), it seems reasonable to occurring at the site. Personal ornaments such assume that from the outset, the public center as stone beads, bone ornaments with incised at Campanayuq Rumi was emulating Chavín iconography, ear spools, and gold jewelry sug- architectural conventions. gest that social organization became much more Despite the strong similarities observed in hierarchical during this phase (Matsumoto 2010, monumental architecture, the ceramic styles that 2012; Matsumoto et al. 2013, 2016). It is also dur- characterize the Campanayuq I Phase do not ing the Campanayuq II Phase that the percentage show clear affiliation to those of contempo- of obsidian artifacts increases from nearly 80 to rary late Initial Period (or Urabarriu Phase) almost 90% (see discussion below). Chavín de Huántar (see Burger 1984). Rather, the ceramic assemblage of the Campanayuq Summary I Phase was composed of styles that exhibit The data presented above reveal how the char- affinities with pottery documented in the south- acter of Campanayuq Rumi changed through central highlands and parts of the south coast time from a small village to a major civic- (Matsumoto 2010, 2018a). For instance, stylistic ceremonial center. Because Campanayuq Rumi linkages exist between Campanayuq Rumi and emulated Chavín architectural conventions as Pirwapukio in the Mantaro Valley (Browman a ceremonial center during the Campanayuq I 1970), Muyo Moqo in Andahuaylas (Bauer et al. Phase, the relationship between these two cen- 2010:51–52; Grossman 1972), Hacha in the ters seems to have been religious rather than Acari Valley (Neira and Cardona 2000–2001; economic, and might have caused significant— Riddell and Valdéz 1987–1988; Robinson 1994), but infrequent—interactions. During the Cam- and Marcavalle in Cuzco (Mohr-Chavez 1977). panayuq I Phase, Campanayuq Rumi was a ceremonial center of regional importance within The Campanayuq II Phase (700–450 BC) the south-central highlands and upper val- Several changes in architecture, ceramic styles, leys of the south coast. Though Campanayuq and social organization occur during the Cam- Rumi likely benefited from its links to Chavín panayuq II Phase (Matsumoto 2010). Around de Huántar, it also maintained much of its 700 BC, the pottery of the previous Campanayuq independence. I Phase was largely replaced by pottery that In the Campanayuq II Phase, Campanayuq shows strong similarities to the ceramic style of Rumi exhibits much stronger influence from the contemporary Janabarriu Phase at Chavín de Chavín de Huántar, both in terms of material cul- Huántar (Burger 1984; see also Rick et al. 2010), ture and social organization. Based on Richard

Figure 5. Image of the central platform showing the central staircase and ashlar construction. (Color online)

Burger’s model of a pilgrimage network centered Formal artifacts were rare at Campanayuq Rumi, at Chavín de Huántar (Burger 1988, 1993), consisting of unifacial tools such as scrapers, Matsumoto has hypothesized that Campanayuq perforators, and drills, as well as bifaces, and Rumi could have been incorporated into this projectile points (Figure 6). In addition, there is religious network and functioned as a branch a great deal of lithic material, including primary shrine (Matsumoto 2010). As we will argue later flakes with cortex, that point to some lithic in this paper, one of the reasons that Campanayuq production in residential contexts. However, this Rumi became embedded within this pan-regional will only be confirmed by a more detailed study interaction sphere was due to its proximity to the of the lithic assemblage in the future. highly prized Quispisisa obsidian source. pXRF Methodology and Sampling Obsidian Use at Campanayuq Rumi The pXRF analysis presented in this paper was Obsidian was the most important lithic material undertaken in August of 2014 at the UNSAAC- used during the Campanayuq I and Campanayuq Yale International Center for the Study of Machu II Phases. Matsumoto (2010:294, 574–593) Picchu and Inca Culture in Cuzco, Peru. Obsidian studied a sample of 2,846 stone artifacts and artifacts from Campanayuq Rumi were analyzed concluded that there was a marked increase using a Bruker Tracer Series III-SD pXRF spec- in obsidian usage in the ceremonial core from trometer. The pXRF spectrometer used in this 76.2% in the Campanayuq I Phase to 89.2% in study was a prototype designed for application at the Campanayuq II Phase. In both phases, obsid- high altitudes. Some of the previous attempts to ian was pervasive and was found in significant use the standard production models of the Bruker quantities in the monumental/ceremonial core, pXRF spectrometer at high altitude failed due to ritual offering deposits, and residential areas an electrical discharge that damaged the X-ray (Matsumoto 2010, 2012). tube. The vast majority of obsidian artifacts at Cam- All measurements with the pXRF were per- panayuq Rumi consisted of flakes and shatter. formed with an operating voltage of 40keV and

Figure 6. Examples of obsidian implements from Campanayuq Rumi. a–c are biface projectile points; d is a drill. (Color online)

current of 30 microamps. Measurement times from Pre-Platform, Campanayuq I, and Cam- were 60 seconds per sample. While other studies panayuq II contexts. Obsidian was selected by have employed longer measurement times (Craig site sector, including samples from the mon- et al. 2007; Rademaker et al. 2013), the XRF that umental/ceremonial core, and from residential we used had a sufﬁciently high count rate that sectors located to the north and south of the permitted us to reduce the counting time to 60 temple core. Obsidian also was taken from seconds and still obtain reliable data. The XRF a circular structure in the South Sector that was calibrated for analysis of obsidian by using includes a number of deep pits that were used the University of Missouri Research Reactor for the placement of various types of offerings (MURR) suite of 40 obsidian source specimens (Matsumoto et al. 2016). Eleven artifacts were and their associated concentration data (Glas- sampled from the Pre-Platform Phase, 93 from cock and Ferguson 2012). The elements for Campanayuq I Phase and 266 from Campanayuq which the concentrations could be determined in II Phase contexts. Artifacts consisted of debitage obsidian include Mn, Fe, Zn, Ga, Rb, Sr, Y, Zr, including ﬂakes (n = 289), shatter (n = 49), and Nb, and Th. cores (n = 6); and complete or semicomplete tools such as projectile points (n = 24), scrapers Obsidian Sample from Campanayuq Rumi (n = 1), and drills (n = 1). The sample analyzed in this study consisted of 370 obsidian artifacts. As mentioned earlier Results of the pXRF Analysis in this paper, the understanding of obsidian exchange and its changing relationship to larger Obsidian at Campanayuq Rumi came from at interaction spheres requires a diachronic per- least six, and perhaps seven, sources includ- spective. For this reason, obsidian was sampled ing Quispisisa, Potreropampa, Anillo, Alca,

Figure 7. Bivariate scatterplot of rubidium (Rb) versus strontium (Sr) for obsidian artifacts from Campanayuq compared to sources in Peru. Sources are displayed as 90% conﬁdence ellipses.

Table 2. Obsidian Counts Overall by Phase (n = 370).

Phase Pre-Platform Campanayuq I Campanayuq II Total Source Quispisisa 11 (100%) 83 (89.25%) 215 (80.83%) 309 (83.5%) Alca-1 0 3 (3.23%) 16 (6.02%) 19 (5.14%) Alca-5 0 2 (2.15%) 0 2 (0.54%) Anillo 0 0 3 (1.13%) 3 (0.81%) Potreropampa 0 0 14 (5.26%) 14 (3.78%) Jampatilla/Lisahuacho 0 0 10 (3.76%) 10 (2.70%) Puzolana 0 5 (5.38%) 8 (3.01%) 13 (3.51%) TOTAL 11 93 266 370

Puzolana, and Jampatilla and/or Lisahuacho ian artifacts come from the Quispisisa quarries (Figure 7). Of particular signiﬁcance was the located 52 km to the west of Campanayuq Rumi. dominance of obsidian from the Quispisisa These data therefore suggest that from the time source, which comprised just under 84% (n = of its foundation in the second millennium BC, 309) of the total sample from all three phases the population of Campanayuq Rumi was provi- of occupation (Table 2; see also Supplemental sioned with high-quality Quispisisa obsidian. Table 1). Campanayuq I Phase (950 BC–700 BC) The Pre-Platform Phase (ca. 1100–950 BC) The sources of obsidian artifacts changed during Though the sample from the Pre-Platform Phase the Campanayuq I Phase. Quispisisa obsidian is quite small (n = 11), it nonetheless provides continued to dominate, making up close to 90% important insight into obsidian use at Cam- (n = 83) of the total sample (n = 93). Neverthe- panayuq Rumi prior to its transformation into less, obsidian from other quarries appears, albeit a large civic-ceremonial center. All 11 obsid- in low quantities.

Over 5% (n = 5) of the artifacts come southwest part of the Department of Apurímac, from Puzolana, a diffuse source that consists near the town of Chalhuanca, and 104 km to the of volcanic deposits containing relatively small southeast of Campanayuq Rumi (Burger, Fajardo nodules in comparison with obsidian from the Rios, and Glascock 2006). An additional 4% Quispisisa source region. Two localities with of the obsidian comes from either Jampatilla nodules of sufficient size to produce tools are or Lisahuacho, which are 62 km and 102 km known; one location is situated just to the south to the south of Campanayuq Rumi, respectively of the city of Ayacucho, approximately 60 km (e.g., Burger, Schreiber, Glascock, and Ccencho to the north of Campanayuq Rumi (Burger and 1998). The pXRF cannot discriminate between Glascock 2000b), and the second location, Arco the Jampatilla and Lisahuacho sources, but this Punco, is just over 8 km to the southeast of uncertainty can be resolved in the future using Campanayuq Rumi. Arco Punco has nodules instrumental neutron activation analysis or a measuring 5–7 cm in diameter, which is adequate more permanent, elaborate XRF facility. for fabricating usable flakes. Given its proximity Finally, 1% of the sample comes from the to Campanayuq Rumi, we hypothesize that the Anillo source. This source is located in the head- Puzolana artifacts were probably obtained from waters of Acari, relatively close to the Alca quar- Arco Punco rather than the source situated near ries, more than 200 km from Campanayuq Rumi the city of Ayacucho. Because of the proximity of (Tripcevich 2016:213; Tripcevich and Contreras Arco Punco, it is noteworthy that the exploitation 2013:30). Thus far, Campanayuq Rumi is the of Puzolana obsidian was so rare at Campanayuq only known archaeological site with artifactual Rumi. materials from this source. Apparently it was not More surprisingly, low amounts of obsidian an important locality for obsidian procurement, from the Alca source (n = 5) appear. Situated at least during the late Initial Period and Early in the headwaters of the Cotahuasi River and Horizon. more than 200 km to the south of Campanayuq Rumi, Alca is a large obsidian deposit. Research Discussion at Alca has shown the obsidian from this locality to be geochemically complex, consisting of six distinct subsources distributed over an area of Diachronic Perspectives on Obsidian Use at at least 330 km² (Rademaker et al. 2013;see Campanayuq Rumi also Burger, Asaro, Trawick, and Stross 1998; Examination of obsidian sourcing from Cam- Jennings and Glascock 2002). At Campanayuq panayuq Rumi shows diachronic changes in Rumi, obsidian comes from the Alca-1 (n = 3) obsidian consumption. When taken together with and Alca-5 (n = 2) subsources. other evidence, changes in obsidian consumption can help shed light on the changing identi- Campanayuq II Phase (700–450 BC) ties that characterize the first half of the first The Campanayuq II Phase is marked by sub- millennium BC and show how Campanayuq stantial diversification of obsidian procurement Rumi became incorporated into a much wider at Campanayuq Rumi. Though obsidian from interaction sphere associated with the expansion the Quispisisa source still makes up the vast of Chavín religious ideology. majority of the assemblage, it comprises a lower While the data from the Pre-Platform Phase proportion (just under 81%: n = 215) of the are limited, they are significant in revealing that assemblage than in the Campanayuq I Phase. from its inception, Campanayuq Rumi, even as a There is likewise a slight increase in obsidian village-level settlement, had a special connection coming from the Alca source discussed earlier with Quispisisa. Further, it is interesting to note (6%: n = 16). that obsidian was not commonly distributed It is also during the Campanayuq II Phase outside of the south-central highlands during that obsidian from three or possibly four new the second millennium BC. The limited circu- sources appears. Five percent of the obsidian lation of obsidian seems to reflect a pattern of sample comes from Potreropampa, located in the local distribution observed at other contemporary

south-central highland archaeological sites (Kel- Despite the small quantities of obsidian from lett et al. 2013). minor sources, Quispisisa obsidian dominates, In the Campanayuq I Phase, Quispisisa con- pointing to continuity in the close relationship tinued to be the predominant exploited source of between Campanayuq Rumi and Quispisisa. We obsidian, but small proportions of obsidian from argue that in the Campanayuq I Phase, Cam- other sources appear for the first time. Among panayuq Rumi became an agent in pooling these are local sources such as Puzolana, as obsidian and exchanging it with other south- well as obsidian from the distant Alca source. ern trading partners in the upper valleys of The presence of Alca obsidian indicates that the south coast and other contemporary settle- Campanayuq Rumi was involved in a wider ments in Cotahuasi, as well as Andahuaylas and interaction sphere situated in the south-central Apurímac. highlands. At the same time, obsidian from These conclusions are supported by other Lisahuacho and Potreropampa are absent at lines of archaeological evidence. Comparative Campanayuq Rumi at this time. This pattern analysis of the ceramic styles of the Campanayuq suggests that obsidian was principally exchanged I Phase show that the pottery assemblage shared with local trading partners in the south-central stylistic affinities primarily with contemporary highlands. For instance, recent obsidian charac- assemblages throughout the south-central and terization studies suggest that contemporary set- southern highlands, and to some extent the tlements in the Andahuaylas region, such as Way- south coast (Matsumoto 2018a). Thus, there is waka (Grossman 1972, 1983), obtained obsidian an emerging picture that throughout the Cam- from nearby sources such as Potreropampa and panayuq I Phase, Campanayuq Rumi appears to Jampatilla, with lesser quantities from Quispi- have been one of the major centers within an sisa (Burger, Schreiber, Glascock, and Ccen- interaction sphere that was associated with the cho 1998; Burger, Fajardo Rios, and Glascock south-central highlands. 2006:112; Kellett et al. 2013). The alterations in obsidian procurement in Campanayuq Rumi during the Early Horizon the Campanayuq I Phase correspond with the Though Quispisisa remains the most important transformation of Campanayuq Rumi into a large source, artifacts from as many as six additional ceremonial center. Though Campanayuq Rumi obsidian sources were used during the Cam- was related to Chavín de Huántar during this panayuq II Phase. It is worth stressing that the phase, there appears to be little evidence for the diversity of obsidian at Campanayuq Rumi is high-volume exchange of obsidian outside of the greater than any contemporary center of the Early south-central highlands. As mentioned earlier, Horizon, including Chavín de Huántar, which obsidian is a rare commodity in the Urabarriu has obsidian from four sources (Quispisisa, Alca, Phase at Chavín de Huántar. Furthermore, there Potreropampa, and Lisahuacho; see Burger et al. is no obsidian found in other contemporary 1984; Burger, Lau, Ponte, and Glascock 2006; northern highland centers such as Kuntur Wasi Contreras and Nado 2018). and Ingatambo during the late Initial Period The diversification of obsidian at Cam- (Sakai and Shimizu 2002; Yamamoto 2014). panayuq Rumi serves as a proxy for open net- It is notable that during the Campanayuq I works of interaction during the first millennium Phase there is an increase in the number of BC. Obsidian from distant sources such as Alca sources represented in the obsidian assemblage. or Anillo, which came to Campanayuq Rumi in Local Puzolana obsidian occurs in low percent- small quantities, can be interpreted as indicators ages; this shows that this material, though located for ephemeral interactions in which outsiders in close proximity to Campanayuq Rumi, was visited Campanayuq Rumi as pilgrims or traders, eschewed as an important location to obtain thus producing a cosmopolitan atmosphere rad- obsidian. This is probably a result of the rela- ically different from that of the Pre-Platform tively small size of the nodules that were avail- village. able at any of the different Puzolana source areas This argument is strengthened by other lines (Burger, Bencic, and Glascock 2016). of archaeological evidence. Matsumoto (2010,

2018a, 2018b; Matsumoto and Cavero Palomino the interaction spheres to the north during the 2010) observed that during the Campanayuq II Early Horizon (Burger et al. 2000:304). Phase there was an influx of ceramic material that shows contact with more distant commu- Campanayuq Rumi and the Chavín Interaction nities than in the Campanayuq I Phase. Pottery Sphere during this time includes the sudden appearance As discussed earlier, there is a dramatic increase of Paracas-related assemblages from the south in obsidian at Chavín de Huántar in the Janabar- coast of Peru. At the same time that Paracas riu Phase assemblage, and obsidian tools were an pottery appears at Campanayuq Rumi, there important component of the crafting economy. is a corresponding influx of obsidian at Early This demand for obsidian was linked with a Horizon Paracas settlements (DeLeonardis and number of socioeconomic changes at Chavín de Glascock 2013; see also Burger and Asaro 1977), Huántar, including marked growth of the urban suggesting exchange relationships between these settlement, the presence of craft specialists, and two regions. Equally significant is the appearance emergence of an elite class (Burger 1984, 1992; of pottery styles with iconographic motifs that Sayre 2010). are clearly linked to the Janabarriu Phase of We argue that by the Campanayuq II Phase, Chavín de Huántar (Burger 1984;Matsumoto Campanayuq Rumi developed into an important 2010; Matsumoto and Cavero Palomino 2010). “gateway community” (Hirth 1978) that served Taken together, the opening up of interaction net- as a central redistributive point to provide a works reflects how Campanayuq Rumi, and parts reliable flow of obsidian to Chavín de Huántar of the south-central highlands more generally, and other centers during the Early Horizon (Mat- became embedded within the Chavín Interaction sumoto 2010). Campanayuq Rumi was situated Sphere during the early first millennium BC. on an advantageous trade route that connected While we stress the importance that the diver- this site to the Quispisisa obsidian quarries (cf. sity of sources has for understanding changing Contreras 2011;Matsumoto2010). interaction patterns, it is worth pointing out the Demand for Quispisisa obsidian therefore information we can infer from sources absent necessitated intensified connections with the from the assemblage. For example, the sam- south-central highlands. Given the close con- ple from Campanayuq Rumi does not contain nection with Quispisisa, and its role as the any obsidian from the Chivay source, which is preeminent Early Horizon center of the south- located in the headwaters of the Colca Valley central highlands, we posit that Campanayuq in the Department of Arequipa, some 340 km Rumi directed the long-distance trade of obsidian to the southeast of Campanayuq Rumi (Burger, to centers such as Chavín de Huántar in the Asaro, Stross, and Salas 1998:210; Tripcevich north. This argument is based on several lines and MacKay 2011). Research indicates that of evidence. during the Early Horizon, Chivay obsidian was First, one of the unique characteristics of occasionally used at sites in the Cuzco region Campanayuq Rumi is its imitation of the archi- and much more heavily consumed at sites such as tectures and layout of Chavín de Huántar. Sim- Qaluyu and Chiripa near Lake Titicaca (Burger, ilarities between the two sites include a U- Asaro, Stross, and Salas 1998:210; Burger et al. shaped plan and the incorporation of galleries 2000; Tripcevich 2007:229, Figure 3–12). The and a sunken circular plaza into the monumen- absence of Chivay obsidian therefore points to a tal architecture. In contrast, many of the other southern boundary of interaction during the Early highland civic-ceremonial centers involved in Horizon, in which Campanayuq Rumi was part of the Chavín Interaction Sphere, such as Atalla, an interaction sphere oriented toward the south- Kuntur Wasi, and Pacopampa, have architectural central and north-central highlands, rather than characteristics that are quite distinct from Chavín the southern highland regions of Cuzco or Titi- de Huántar. The fact that Campanayuq Rumi so caca. Conversely, sites in the southern highlands clearly emulates Chavín de Huántar indicates a did not exploit Quispisisa obsidian, supporting unique relationship between these two centers. the argument that this region was isolated from One of the reasons for this close relationship was

rooted in Campanayuq Rumi’s access to high- The paucity of obsidian at these centers would quality Quispisisa obsidian. appear to demonstrate the fall-off patterns more The broader pattern of obsidian distribution indicative of down-the-line exchange (Renfrew in the central highlands in the region situated 1975). between Campanayuq Rumi and Chavín de Long-distance movement of obsidian from Huántar is a second line of evidence suggest- the south-central highlands was facilitated by ing direct trade between the two regions. For llama caravans. At Campanayuq Rumi, faunal instance, Burger (1980) argued that the puna analysis shows that domesticated llamas were a grasslands of Junín formed a conduit for camelid critical component of the economy since Cam- caravans transporting obsidian from the south- panayuq I times (Matsumoto 2010:344–353). ern highlands to the north. Quispisisa obsidian In the central highlands, llamas were used as artifacts were found in low quantities at remote pack animals for the transportation of bulk goods cave sites in this region, suggesting that cara- such as obsidian during the late Initial Period vans passed through these areas, resulting in the and Early Horizon (Browman 1975; Tripcevich exchange of obsidian as they moved goods to the 2010). As pack animals, llamas can carry north. between 25 and 40 kilograms, depending on A possible exception to this pattern is the the length of the trip (Flores-Ochoa 1979:95). civic-ceremonial center of Atalla near Huan- The distance between Campanayuq Rumi and cavelica, where abundant obsidian was reported the Quispisisa quarries at Cerro Jichja Parco on the surface (Burger and Matos 2002). Geo- is roughly 60 km. If llama caravans traveled chemical analysis of two ﬂakes demonstrates between 15 and 20 kilometers per 10-hour that they came from Quispisisa (Burger and day (Flores-Ochoa 1979:95; see also Tripcevich Matos 2002:169). Recent excavations at Atalla 2007), then a round trip would take somewhere by Michelle Young have recovered large quan- between 8 and 12 days along the high puna tities of artifactual obsidian in Early Horizon routes that link the two localities. Caravan expe- contexts at the site (Michelle Young, personal ditions between Campanayuq Rumi and Chavín communication 2017). It is possible that this de Huántar consisted of an approximately 600- obsidian arrived from traders from Campanayuq kilometer journey (Contreras 2011:386, Table 2) Rumi, who used settlements such as Atalla as which would have been a round-trip travel time stopover points on the way to Chavín de Huántar. of at least 30 days. Atalla itself was an important node in the Chavín The type of trade that we describe in this Interaction Sphere and its occupants apparently paper raises the question of how the movement of were consumers of the high-quality volcanic goods and people over such long distances was glass. We therefore hypothesize that the spatial structured. Burger (1988) argued that much of the patterns of obsidian utilization speak to a prob- Chavín expansion could be explained through the able direct trading relationship between Cam- widespread acceptance of a common religious panayuq Rumi and Chavín de Huántar, rather ideology. Shared religion aided in the rapid than down-the-line exchange. For instance, at diffusion of exotic goods and people from distant Chavín de Huántar we do not see the kinds of localities (Burger 1988, 1993, 2008, 2012, 2013). fall-off in obsidian quantity that would suggest Use of religious mechanisms in long-distance down-the-line trade (Renfrew 1975). Obsidian trade has been documented in other parts of is found in high quantity at Chavín de Huántar, the world. Among the precolonial Igbo of West and was distributed widely in the temple core Africa, traders were known to travel signiﬁcant and the residential occupation (Burger 1984, distances under the protection of powerful ora- 2013; Burger, Lau, Ponte, and Glascock 2006; cles (Northrup 1978). Given the architectural and Contreras and Nado 2018). This contrasts with iconographic evidence, Campanayuq Rumi and the low numbers of obsidian found at coastal and other highland centers may have served as branch highland centers situated to the north of Chavín shrines (Burger 1988;Matsumoto2010). Fur- de Huántar, such as Cerro Blanco, Kuntur Wasi, thermore, long-distance exchange may have been and Pacopampa (Burger and Glascock 2009). aided by common language. Recent research

suggests that the Chavín Interaction Sphere sodalite, strengthened preexisting connections partly drove the expansion of Aymara, which with communities in the south-central highland may have served as a prestigious and/or religious region. We posit that Campanayuq Rumi was lingua franca spoken over a wide swath of critical to supplying Chavín de Huántar with highland Peru during the Early Horizon (cf. the high-quality obsidian that formed a crucial Burger 2012; Heggarty and Beresford-Jones component of its ritual and political economy. 2010; Hornborg 2014). These connections stimulated important socioeconomic changes at Campanayuq Rumi, Conclusions and the south-central highlands more broadly. While the relationship between Campanayuq This paper shows the value of geochemical Rumi and Chavín de Huántar was significant, characterization studies of large assemblages obsidian analysis likewise confirms the contin- of obsidian and analyzing the results from a ued and broadened importance of Campanayuq diachronic perspective (e.g., Hirth et al. 2013). Rumi as part of a regional interaction sphere By focusing on a single site we have been able centered within the south-central highlands, one to demonstrate how Campanayuq Rumi shifted that included portions of Arequipa, Apurímac, from involvement in a local interaction sphere to and the Peruvian south coast. The existence of an interregional network centered in the north- a regional interaction sphere in south-central central highlands. Peru during the Early Horizon is paralleled by In the Pre-Platform and Campanayuq I a regional interaction sphere in northern Peru Phases, Campanayuq Rumi obtained obsidian that incorporated portions of highland Cajamarca almost exclusively from the Quispisisa obsidian and the north coast. These regional interaction source. We believe that the preponderance of spheres articulated with each other to consti- obsidian from this source is a proxy for the close tute the pan-regional Chavín Interaction Sphere. ties Campanayuq Rumi had with the Quispi- They can be viewed as constituting two of its sisa locality. During the Campanayuq I Phase, best-documented components. Campanayuq Rumi appears to have been one of Furthermore, our conclusions should prompt the largest civic-ceremonial centers in the south- a rethinking of regions once thought peripheral central highlands. The access that Campanayuq to developments in the north-central highlands Rumi had to this high-quality obsidian was a of Peru during the late Initial Period and Early factor in the site’s preeminence in a localized Horizon and a better appreciation of the role interaction sphere centered in the south-central of the prehistoric communities of the south- highlands during the late Initial Period. central highlands in producing and transporting It is in the subsequent Campanayuq II Phase the exotic raw materials that were the hallmark that we witness a major change in obsidian of the Chavín Interaction Sphere. At the same use. This period is notable for the diverse time, the transformative power of the Chavín obsidian sources present at Campanayuq Rumi. cult and the economic forces associated with it The presence of six minor sources serves as a must be taken into account in order to under- proxy for visitors who came to this large civic- stand the socioeconomic changes that occurred ceremonial center on pilgrimages or for other at Campanayuq Rumi during the first millennium purposes. Further, these patterns also point to BC. an expansion of the interaction networks that Acknowledgements. The excavations at Campanayuq Rumi characterize the Chavín Interaction Sphere. As were financed by the Coe Fund, Albers Fund, and Williams part of the widening interaction spheres of this Fund of Yale University, The Roger Thayer Stone Center for time period, the ties between Campanayuq Rumi Latin American Studies, Tulane University, JSPS KAKENHI and Chavín de Huántar were particularly pro- Grant-in-Aid for Young Scientists (B) 25770282 (2013), nounced. As Chavín de Huántar’s economy was and JSPS KAKENHI Grant-in-Aid for Young Scientists (A)15H05383 (2014–). The pXRF analysis in Cuzco was growing during the Janabarriu Phase and social carried out as part of the archaeometry workshop in the stratification was increasing, demand for exotics Casa Concha with the financial support of Yale University as from the south, including obsidian, cinnabar, and part of the ongoing educational exchange program between

the archaeology programs of Yale and UNSAAC. Costs for 1993 The Chavín Horizon: Stylistic Chimera or Socio- analysis of the artifacts were also covered by a National economic Metamorphosis? In Latin American Hori- Science Foundation grant (1415403) to the Archaeometry zons, edited by Donald Rice, pp. 41–82. Dumbarton Lab at the University of Missouri Research Reactor. Spe- Oaks Research Library and Collection, Washington, D.C. cial thanks to Edison Mendoza Martinez, Ruben Cisneros 2008 Chavín de Huántar and Its Sphere of Inﬂuence. In Cardenas, Jhon Rene Huamani Diaz, Hilda Mirtha Bellido Handbook of South American Archaeology, edited by Huaña, Jorge Ronald Sulca Flores, Maribel Gutierrez Solier, Helaine Silverman and William H. Isbell, pp. 681–703. and Roy Gutiérrez Silva for working in the excavations Springer, New York. at Campanayuq Rumi. Our deepest appreciation goes to 2012 Central Andean Language Expansion and the Daniel Contreras, Lucy Salazar, Yuji Seki, Kurt Rademaker, Chavín Sphere of Interaction. Proceedings of the Bri- Christopher Rodning, Marcello Canuto, Tatsuya Murakami, tish Academy 173:133–159. Oxford University Press, Michelle Young, Alina Alvarez Larraín, Bebel Ibarra, Rachel Oxford. Horowitz, Ryan Hechler, and Rachel Johnson for providing 2013 In the Realm of the Incas: An Archaeologi- us with valuable suggestions for shaping the ideas presented cal Reconsideration of Household Exchange, Long- Distance Trade, and Marketplaces in the Pre-Hispanic in this article. Central Andes. In Merchants, Markets, and Exchange in the Pre-Columbian World, edited by Kenneth Data Availability Statement. The materials excavated and G. Hirth and Joanne Pillsbury, pp. 319–334. Dumbarton collected by the Proyecto de Investigaciones Arqueoló- Oaks Research Library and Collection, Washington, gicas, Campanayuq Rumi, Vilcashuaman-Ayacucho 2007– DC. 2008 and 2013 are curated in the facilities of the Minis- Burger, Richard L., and Frank Asaro terio de Cultura in Ayacucho, Peru. All detailed scientiﬁc 1977 Trace Element Analysis of Obsidian Artifacts from data are recorded in reports in possession of the authors, the Andes: New Perspectives on Pre-Hispanic Economic Interaction in Peru and Bolivia. Lawrence Berkeley and they are available by request to Yuichi Matsumoto Laboratory Report 6343:1–88. ([email protected]). Burger, Richard L., Frank Asaro, and Helen Michel 1984 Appendix E: The Source of the Obsidian Artifacts Supplemental Materials. For supplementary material at Chavín de Huántar. In The Prehistoric Occupa- accompanying this paper, visit https://doi.org/10.1017/laq. tion of Chavín de Huántar, Peru, edited by Richard 2017.64 L. Burger, pp. 263–270. University of California Press, Supplemental Table 1. Obsidian samples from Cam- Berkeley. panayuq Rumi. Burger, Richard L., Frank Asaro, Fred Stross, and Guido Salas 1998 The Chivay Obsidian Source and the Geological Ori- gin of Titicaca Basin Type Obsidian Artifacts. Andean References Cited Past 5:203–224. Burger, Richard L., Frank Asaro, Paul Trawick, and Bandelier, Alfred F. Fred Stross 1904 Aboriginal Trephining in Bolivia. American Anthro- 1998 The Alca Obsidian Source: The Origin of Raw pologist 6:440–446. Material for Cuzco Type Obsidian Artifacts. Andean Bauer, Brian, Lucas C. Kellett, and Miriam Aráoz Silva Past 5:182–202. 2010 The Chanka: Archaeological Research in Burger, Richard L., Catherine M. Bencic, and Michael Andahuaylas (Apurimac), Peru. Cotsen Institute of D. Glascock Archaeology Press Monograph Vol. 69. University of 2016 Obsidian Procurement and Cosmopolitanism at California, Los Angeles. the Middle Horizon Settlement of Conchopata, Peru. Browman, David Andean Past 12:21–44. 1970 Early Peruvian Peasants: The Culture of a Cen- Burger, Richard L., Kris E. Lane, and Colin A. Cooke tral Highland Valley. PhD dissertation, Department of 2016 Ecuadorian Cinnabar and the Prehispanic Trade in Anthropology, Harvard University, Cambridge, Mas- Vermillion Pigment: Viable Hypothesis or Red Herring. sachusetts. Latin American Antiquity 27:22–35. 1975 Trade Patterns in the Central Highlands of Peru in Burger, Richard L., and Ramiro Matos the First Millennium BC. World Archaeology 6:322– 2002 Atalla: A Center on the Periphery of Chavín Hori- 329. zon. Latin American Antiquity 13:153–177. Burger, Richard L. Burger, Richard L., Katharina Schreiber, Michael D. Glas- 1980 Trace-Element Analysis of Obsidian Artifacts from cock, and José Ccencho Pachamachay, Junín. In Prehistoric Hunters of the High 1998 The Jampatilla Obsidian Source: Identifying the Andes, edited by John W. Rick, pp. 257–261. Academic Geological Source of Pampas Type Obsidian Artifacts Press, New York. from Southern Peru. Andean Past 5:225–229. 1984 The Prehistoric Occupation of Chavín de Huántar, Burger, Richard L., Fidel A. Fajardo Rios, and Michael Peru. University of California Press, Berkeley. D. Glascock 1988 Unity and Heterogeneity within the Chavín Horizon. 2006 Potreropampa and Lisahuacho Obsidian Sources: In Peruvian Prehistory, edited by Richard W. Keatinge, Geological Origins of Andahuaylas A and B Type pp. 99–144. Cambridge University Press, Cambridge. Obsidians in the Province of Aymares, Department 1992 Chavín and the Origins of Andean Civilization. of Apurimac, Peru. Ñawpa Pacha: Journal of Andean Thames and Hudson, London. Archaeology 28:109–127.

Burger, Richard L., and Michael D. Glascock Flores-Ochoa, Jorge A. 2000a Locating the Quispisisa Obsidian Source in the 1979 Pastoralists of the Andes: The Alpaca Herders Department of Ayacucho, Peru. Latin American Antiq- of Paratía. Institute for the Study of Human Issues, uity 11:258–268. Philadelphia. 2000b The Puzolana Obsidian Source: Locating the García, Rubén, and José Pinilla Geologic Source of Ayacucho Type Obsidian. Andean 1995 Aproximación a una Secuencia de Fases con Past 6:289–308. Cerámica Temprana de Paracas. Journal of the Steward 2002 Tracking the Source of Quispisisa Type Obsidian Anthropological Society 23:43–81. from Huancavelica to Ayacucho. In Andean Archaeol- Glascock, Michael D., and Jeffrey R. Ferguson ogy Vol. 1: Variations in Sociopolitical Organization, 2012 Report on the Analysis of Obsidian Source Samples edited by William H. Isbell and Helaine Silverman, pp. by Multiple Analytical Methods. Prepared for Bruker 341–368. Kluwer Academic/Plenum Publishers, New Elemental. Report on file, University of Missouri- York. Columbia. 2009 Intercambio Prehistórico de Obsidiana a Larga Glascock, Michael D., Robert J. Speakman, and Richard Distancia en el Norte Peruano. Revista del Museo de L. Burger Arqueología, Antropología e Historia 11:17–50. 2007 Sources of Archaeological Obsidian in Peru: Burger, Richard L, George F. Lau, Víctor M. Ponte, and Descriptions and Geochemistry. In Archaeological Michael D. Glascock Chemistry: Analytical Techniques and Archaeological 2006 The History of Prehispanic Obsidian Procurement Interpretation, edited by Michael D. Glascock, Robert in Highland Ancash. In La Complejidad Social en la J. Speakman, and Rachel S. Popelka-Filcoff, pp. 522– Sierra de Ancash: Ensayos Sobre Paisaje, Economía y 552. American Chemical Society, Washington, DC. Continuidades Culturales, edited by Alexander Herrera, Grossman, Joel Carolina Orsini, and Kevin Lane, pp. 103–120. PUNKU 1972 Early Ceramic Cultures of Andahuaylas, Apurimac, Centro de Investigación Andina, Lima. Peru. PhD dissertation, Department of Anthropology, Burger, Richard L., Karen L. Mohr Chávez, and Sergio University of California, Berkeley. J. Chávez 1983 Demographic Change and Economic Transforma- 2000 Through the Glass Darkly: Prehispanic Obsidian tion in the South-Central Highlands of Pre-Huari Peru. Procurement and Exchange in Southern Peru and North- Ñawpa Pacha 21:45–126. ern Bolivia. Journal of World Prehistory 14:267–362. Heggarty, Paul, and David Beresford-Jones Contreras, Daniel A. 2010 Agriculture and Language Dispersals: Limita- 2011 How Far to Conchucos? A GIS Approach to Assess- tions, Refinements, and an Andean Exception? Current ing the Implications of Exotic Materials at Chavín de Anthropology 51:163–191. Huántar. World Archaeology 43: 380–397. Helms, Mary W. Contreras, Daniel A., and Kristin Nado 1993 Craft and the Kingly Ideal: Art, Trade, and Power. 2018 Interpreting Geochemically Characterized Obsidian University of Texas Press, Austin. from Chavín de Huántar, Peru. In The Archaeology of Hirth, Kenneth Circulation, Exchange and Human Migration: Tech- 1978 Interregional Trade and the Formation of Prehistoric niques, Cases, Evidence, edited by David Peterson Gateway Communities. American Antiquity 43:35–45. and John Dudgeon. Equinox Publishing Ltd., Sheffield, 1998 The Distributional Approach: A New Way to England. In press. Identify Market Behavior Using Archaeological Data. Cooke, Colin A., Holger Hintelmann, Jay J. Ague, Richard Current Anthropology 39:451–476. Burger, Harald Biester, Julian P. Sachs, and Daniel Hirth, Kenneth, Ann Cyphers, Robert Cobean, Jason De R. Engstrom León, and Michael D. Glascock 2013 Use and Legacy of Mercury in the Andes. Environ- 2013 Early Olmec Obsidian Trade and Economic Orga- mental Science & Technology 47:4181–4188. nization at San Lorenzo. Journal of Archaeological Craig, Nathan, Robert J. Speakman, Rachel S. Popelka- Science 40:2784–2798. Filcoff, Michael D. Glascock, J. David Robinson, M. Hornborg, Alf Steven Shackley, and Mark. S. Aldenderfer 2014 Political Economy, Ethnogenesis, and Language 2007 Comparison of XRF and PXRF for Analysis of Dispersals in the Prehispanic Andes: A World-System Archaeological Obsidian from Southern Peru. Journal Perspective. American Anthropologist 116:810–823. of Archaeological Science 34: 2012–2024. Isla, Johny, and Marcus Reindel Cruzatt, Augusto. 2006 Una Tumba Paracas Temprano en Mollake Chico, 1971 Horizonte Temprano en el Valle de Ayacucho. Valle de Palpa, Costa Sur del Peru. Zeitschrift für Anales Científicos 1:603–631. Archäologie Außereuropäischer Kulturen 1:153–181. DeLeonardis, Lisa, and Michael D. Glascock Jennings, Justin, and Michael D. Glascock 2013 From Queshqa to Callango: A Paracas Obsidian 2002 Description and Method of Exploitation of the Assemblage from the Lower Ica Valley, Peru. Ñawpa Alca Obsidian Source, Peru. Latin American Antiquity Pacha 33:163–192. 13:107–118. Druc, Isabel C. Kellett, Lucas C., Mark Golitko, and Brian S. Bauer 2004 Ceramic Diversity in Chavín de Huántar, Peru. Latin 2013 A Provenance Study of Archaeological Obsidian American Antiquity 15:344–363. from the Andahuaylas Region of Southern Peru. Journal Dulanto, Jahl of Archaeological Science 40:1890–1902. 2015 Puerto Nuevo: Redes de Intercambio a Larga Kembel, Silvia Distancia Durante la Primera Mitad del Primer Milenio 2008 The Architecture at the Monumental Center of Antes de Nuestra Era. Boletín de Arqueología PUCP 17 Chavín de Huántar: Sequence, Transformations, and (2013):103–132. Chronology. In Chavín: Art, Architecture and Culture,

edited by William J. Conklin, and Jeffrey Quilter, pp. Mohr Chávez, Karen L. 35–81. Cotsen Institute of Archaeology, University of 1977 Marcavalle: The Ceramics from an Early Horizon California, Los Angeles. Site in the Valley of Cuzco, Peru and Implications Kembel, Sylvia, and Herbert Haas for South Highland Socio-economic Interaction. PhD 2015 Radiocarbon Dates from the Monumental Architec- dissertation, Department of Anthropology, University ture at Chavín de Huántar, Perú. Journal of Archaeolog- of Pennsylvania, Philadelphia. ical Method and Theory 22:345–427. Neira, Maximo, and Augusto Cardona Lothrop, Samuel K. 2000–2001 El Periodo Formativo en el Área de Are- 1941 Gold Ornaments of Chavín Style from Chongoyape, quipa. In Andes 3: Boletín de la Misión Arqueológica Peru. American Antiquity 6:250–266. Andina, edited by Mariusz Ziólkowski and Luis Augusto Lumbreras, Luis G. Belan Franco, pp. 27–60. Universidad de Varsovia, 1981 The Stratigraphy of the Open Sites. In Prehistory Varsovia. of the Ayacucho Basin, Peru, Volume II: Excavations Nesbitt, Jason, and Yuichi Matsumoto. and Chronology, edited by Richard S. MacNeish, Angel 2014 Cupisnique Pottery at the South Highland Site of Garcia Cook, Luis G. Lumbreras, Robert K. Vierra, and Campanayuq Rumi: Implications for Late Initial Period Antoinette Nelken-Terner, pp. 167–198. University of Interaction. Peruvian Archaeology 1:47–61. Michigan Press, Ann Arbor. Northrup, David 2007 Chavín: Excavaciones Arqueológicas (2 volumes). 1978 Trade Without Rulers: Pre-Colonial Economic Universidad Alas Peruanas, Lima. Development in South-Eastern Nigeria. Clarendon 2014 Excavaciones en la Plaza Circular y el Atrio del Press, Oxford. Lanzón en Chavín de Huántar. Antamina, Lima. Onuki, Yoshio (ed.) Matsumoto, Yuichi 1995 Kuntur Wasi y Cerro Blanco: Dos Sitios del Forma- 2010 The Prehistoric Ceremonial Center of Campanayuq tivo en el Norte del Perú. Hokuesen-Sha, Tokyo. Rumi: Interregional Interactions in the South-Central Patterson, Thomas C. Highlands of Peru. PhD dissertation, Department of 1971 Chavín: An Interpretation of its Spread and Inﬂu- Anthropology, Yale University, New Haven, Connecti- ence. In Dumbarton Oaks Conference on Chavín, edited cut. by Elizabeth P. Benson, pp. 29–48. Dumbarton Oaks 2012 Recognizing Ritual: The Case of Campanayuq Research Library and Collection, Washington, DC. Rumi. Antiquity 86:746–759. Rademaker, Kurt, Michael D. Glascock, Bruce Kaiser, 2018a South of Chavín: Interregional Interactions David Gibson, Daniel R. Lux, and Martin G. Yates between the Central Highlands and South Coast during 2013 Multi-Technique Geochemical Characterization of the Late Initial Period and Early Horizon. In New the Alca Obsidian Source, Peruvian Andes. Geology Perspectives on Early Andean Civilization: Interaction, 41:779–782. Authority, and Socioeconomic Organization during the Renfrew, Colin 2nd and 1st Millennia B.C., edited by Richard L. Burger, 1975 Trade as Action at a Distance: Questions of Integra- Yuji Seki, and Lucy C. Salazar. Yale University Publi- tion and Communication. In Ancient Civilization and cations in Anthropology, New Haven, Connecticut, in Trade, edited by Jeremy A. Sabloff and C.C. Lamberg- press. Karlovsky, pp. 3–59. University of New Mexico Press, 2018b Paracas en la Sierra: Interacción Temprana entre Albuquerque. la Sierra Centro-sur y Costa Sur. Peruvian Archaeology 1986 Introduction: Peer Polity Interaction and Socio- 3, in press. political Change. In Peer Polity Interaction and Socio- Matsumoto, Yuichi, and Yuri Cavero Palomino political Change, edited by Colin Renfrew and John 2010 Una Aproximación Cronológica del Centro Cere- F. Cherry, pp. 1–18. Cambridge University Press, Cam- monial de Campanayuq Rumi, Ayacucho. Boletín de bridge. Arqueología PUCP 13:323–346. Rick, John W. Matsumoto, Yuichi, Yuri Cavero Palomino, and Roy Gutié- 2005 The Evolution of Authority and Power at Chavín de rrez Silva Huántar, Peru. In Foundations of Power in the Prehis- 2013 The Domestic Occupation of Campanayuq Rumi: panic Andes, edited by Kevin J. Vaughn,Dennis Ogburn, Implications for Understanding the Initial Period and and Christina A. Conlee, pp. 71–89. Archaeological Early Horizon of the South-Central Andes of Peru. Papers of the American Anthropological Association, Andean Past 11:167–211. No. 14, Arlington, Virginia. Matsumoto, Yuichi, Jason Nesbitt, Yuri Cavero, and Edi- 2008 Context, Construction, and Ritual in the Develop- son Mendoza ment of Authority at Chavín de Huántar. In Chavín: 2016 Actividades Rituales en Áreas Circundantes al Art, Architecture, and Culture, edited by William Con- Centro Ceremonial de Campanayuq Rumi, Vilcashua- klin and Jeffrey Quilter, pp. 3–34. Cotsen Institute of man, Ayacucho. In Actas del I Congreso Nacional de Archaeology, University of California, Los Angeles. Arqueología, Vol. 2, pp. 99–104. Ministerio de Cultura, Rick, John W., Christian Mesía, Daniel Contreras, Silvia R. Lima. Kembel, Rosa M. Rick, Matthew Sayre, and John Wolf Mesía Montenegro, Christian 2010 La Cronología de Chavín de Huántar y Sus Impli- 2014 Festines y Poder en Chavín de Huántar Durante cancias para el Formativo. Boletín de Arqueología el Periodo Formativo Tardío en los Andes Centrales. PUCP 13 (2009):87–132. Chungara: Revista de Antropología Chilena 46:313– Riddell, Francis A. and Lidio M. Valdéz 343. 1987–1988 Hacha y la Ocupación Temprana del Valle de Miller, George R., and Richard L. Burger Acari. Gaceta Arqueológica Andina 16:6–10. 1995 Our Father the Cayman, Our Dinner the Llama: Ani- Robinson, Roger mal Utilization at Chavín de Huántar, Peru. American 1994 Recent Excavations at Hacha in the Acari Valley, Antiquity 60:421–458. Peru. Andean Past 4: 9–37.

Rosenfeld, Silvana A., and Matthew P. Sayre 2010 Exotic Goods and Socio-political Change in the 2016 Llamas on the Land: Production and Consumption South-Central Andes. In Trade and Exchange: Archae- of Meat at Chavín de Huántar, Peru. Latin American ological Studies from History and Prehistory, edited by Antiquity 27:497–511. Carolyn D. Dillian and Carolyn L. White, pp. 59–73. Sakai, Masato, and Masaaki Shimizu Springer, New York. 2002 Kuntur Wasi shinden ni okeru sekki to sek- 2016 The Ethnoarchaeology of a Cotahuasi Salt Cara- izai: kaigan tono kouryuu to soen, sekki koubou van: Exploring Andean Pastoralist Movement. In The wo megutte. In Andes senshi no jinruigaku teki Archaeology of Andean Pastoralism, edited by José kenkyuu-Kuntur Wasi iseki no hakkutsu, H11-13 nendo M. Capriles and Nicholas Tripcevich, pp. 211–229. kagaku kenkyuuhi hojokin kenkyuu seika houkoku, University of New Mexico Press, Albuquerque. edited by Yasutake Kato, pp. 175–190. Saitama, Tripcevich, Nicholas, and Daniel A. Contreras Japan. 2011 Quarrying Evidence at the Quispisisa Obsidian Sayre, Matthew P. Source, Ayacucho, Peru. Latin American Antiquity 2010 Life Across the River: Agricultural, Ritual, and 22:121–136. Production Practices at Chavín de Huántar, Peru. PhD 2013 Archaeological Approaches to Obsidian Quar- dissertation, Department of Anthropology, University of ries: Investigations at the Quispisisa Source. In Min- California, Berkeley. ing and Quarrying in the Ancient Andes, edited by Sayre, Matthew P., Melanie J. Miller, and Silvana A. Rosen- Nicholas Tripcevich and Kevin Vaughn, pp. 23–44. feld Springer, New York. 2016 Isotopic Evidence for the Trade and Production Tripcevich, Nicholas, and Alex MacKay of Exotic Marine Mammal Bone Artifacts at Chavín 2011 Procurement at the Chivay Obsidian Source, Are- de Huántar, Peru. Archaeological and Anthropological quipa, Peru. World Archaeology 43: 271–297. Science 8:403–417. Yamamoto, Atsushi Seki, Yuji, Juan Pablo Villanueva, Diana Alemán, 2014 Activities and Strategies for Ceremonial Architec- Mauro Ordoñez, Masato Sakai, Walter Tosso M., ture in the Andean Formative: Settlement Patterns and Kinya Inokuchi, Araceli Espinoza, and Daniel Morales Inter-regional Interactions in the Huancabamba Valley, 2010 Nuevas Evidencias del Sitio Arqueológico Paco- Northern Peru. Annual Papers of the Anthropological pampa en la Sierra Norte del Perú. Boletín de Arque- Institute 4:1–33. (in Japanese) Anthropological Insti- ología PUCP 12:69–95. tute, Nanzan University, Nagoya, Japan. Tripcevich, Nicholas 2007 Quarries, Caravans, and Routes to Complexity in the South-Central Andes. PhD dissertation, Department of Submitted March 7, 2017; Revised April 5, 2017; Accepted Anthropology, University of California, Santa Barbara. September 29, 2017