Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports

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

ED-XRF analysis of obsidian artifacts from Yanawilka, a settlement oftransplanted laborers (mitmaqkuna), and implications for Inca imperialism

Di Hua,⁎, M. Steven Shackleyb

aHopkins-Nanjing Center, School of Advanced International Studies, Johns Hopkins University, ChinabDepartment of Anthropology, University of California, Berkeley, United States

A R T I C L E I N F O

Keywords:Obsidian provenanceEd-XRFIncasLithic technologyPolitical economy

A B S T R A C T

Perhaps the most ambitious social policy carried out by the Incas, the mitmaq program resettled one third to onequarter of subject populations for the purposes of control and producing for the state. Ethnohistoric sourcessuggest that the relocated people, called mitmaqkuna, were given access to fertile lands and enjoyed elevatedsocial status and freedoms bestowed to them by the Incas. Until now, these claims have not been tested witharchaeological evidence. This paper evaluates the ethnohistoric claims through the geochemical analysis of 84obsidian artifacts from a probable mitmaqkuna agricultural labor colony called Yanawilka, located inVilcashuamán province, Peru. There is evidence that access to obsidian was restricted for the inhabitants ofYanawilka. The obsidian was mostly from the Quispisissa source, but the relative scarcity, small flake size, andevidence for conservation of raw material suggests that access to this high-quality source was limited and notdue to direct procurement.

1. Introduction and research questions

The mitmaq, or resettlement policy of the Inca, affected the Andeansocial landscape more than any other state policy and affected ap-proximately a quarter to a third of the subject population (D'Altroy,2014: 373). The Inca are renowned for their divide-and-control policiessuch as the mitmaq system and encouraging distinction, especiallythrough dress, among groups (D'Altroy, 1992; Rowe, 1982). The mit-maqkuna, subjects of the mitmaq policy, were transplanted from theiroriginal homeland by the Inca to control rebellious areas and to meetstate economic demands (D'Altroy, 1992: 188; Rowe, 1946: 269–270).The Inca fragmented the political landscape by resettling small groupsof people from different political and ethnic loyalties together in thesame area as a form of divide-and-control (Cieza de León, 1959, pp.56–63). The environs of the Inca provincial capital of Vilcashuamánwere populated by mitmaqkuna from approximately a dozen differentethnic groups (Salas, 2002). The Incas may have socially elevated mit-maqkuna groups by granting them economic advantages such as accessto fertile land (Alconini, 2013: 278). According to the sixteenth-centurychronicler Sarmiento de Gamboa (2010: 146), the Inca emperor TopaInca, credited with expanding mitmaq system created by his father IncaPachacuti, gave the mitmaqkuna “more privileges and freedom” com-pared to non-mitmaqkuna subject populations. On the other hand, theInca were also known for tightly controlling traffic on their extensive

road network (Cieza de León, 1959: 127; Hyslop, 1984; Polo deOndegardo, 1873: 169, and others have shown how obsidian exchangewas politically mediated under Inca rule (Ogburn et al., 2009;Yacobaccio et al., 2002, 2004).

Through the geochemical analysis of 84 obsidian artifacts from theprobable mitmaqkuna settlement of Yanawilka, we assess whether theethnohistoric sources regarding the privileges and freedom of mitmaq-kuna groups extended to the regional procurement and/or exchange ofobsidian. Based on AMS Radiocarbon dating, shallow stratigraphy, andthe lack of Spanish colonial artifacts and ecofacts, Yanawilka was oc-cupied for about a hundred years from approximately 1430 to 1530 CE,probably by an ethnic group the Incas called the Conde (see Section 2for a fuller discussion). Yanawilka is positioned only 700m from amajor royal Inca road and is close to both the royal palace atPumaqocha-Intihuatana (3 km) and the Inca provincial capital ofVilcashuamán (6 km) (Fig. 1). The proximity of Yanawilka to this majorartery of the Inca Empire would have meant that there was constanttraffic passing by, providing Yanawilka inhabitants the possibility toengage in regional trade. Whether they were able to do so may havedepended on Inca permission. The roads around Vilcashuamán werecarefully monitored by the Inca; the sixteenth-century chronicler Ciezade León (1959: 127) stated that troops patrolled these roads and wereprovisioned by the extensive storehouses of maize in Vilcashuamán.According to the late sixteenth/early seventeenth-century mestizo

https://doi.org/10.1016/j.jasrep.2018.01.018Received 25 July 2017; Received in revised form 16 November 2017; Accepted 10 January 2018

⁎ Corresponding author.E-mail addresses: dhu15@jhu.edu (D. Hu), shackley@berkeley.edu (M.S. Shackley).

Journal of Archaeological Science: Reports 18 (2018) 213–221

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

T

chronicler Guamán Poma de Ayala (1615: 546), the Inka made travelerswear special badges to show that they had permission on the roads. TheCondes were considered allies of the Inca, and they were transplantedto Inca Vilcashuamán province to secure this geopolitically strategicarea (Salas, 2002). The Condes were one of the groups who formed partof the Inca armies' campaigns in the Vilcashuamán area (Cieza de León,1984 [1553]: second part, chapters 34, 36, 47, 48). If the Inca were togrant any mitmaqkuna group special privileges or freedoms, it would bethe Condes of Vilcashuamán province.

We use three lines of inquiry to assess whether the Condes mit-maqkuna of Yanawilka had unrestricted access to obsidian. First, avariety of obsidian sources used often signals diverse exchange part-ners, implying some degree of regional integration and interaction(Mills et al., 2013; Shackley, 2005: 143–144). If the number of sourcesof obsidian that the inhabitants of Yanawilka used is significantlysmaller than the total number of nearby high-quality sources, thensome restriction of access was likely.

Second, the relationship between geographic distances to thesources and the relative proportions of the sources used can be used toinfer the political aspects of exchange (Shackley, 2005: 154–155). Byanalyzing any significant deviations from the standard distance decaymodel, also known as the “Law of Monotonic Decrement,” we can assesswhether there was significant political mediation of exchange (Levineet al., 2011; Molyneaux, 2002: 144; Renfrew, 1984: 136; Shackley,2005: 137, 154–155; Torrence, 1986: 15–16). Ogburn et al. (2009)suggested that political boundaries, as in the case of the Inca frontier,may have prevented certain sources from being exploited despiteproximity. If the obsidian artifact proportions at Yanawilka were con-sistent with the standard distance decay model, then political mediationof exchange is less likely. Conversely, if certain obsidian sources werenot represented at Yanawilka despite being high-quality and relativelyclose, then political mediation is more likely.

Third, the extensive use of bipolar reduction and the high propor-tion of small flakes of obsidian from relatively nearby sources char-acterized by large nodules can indicate scarcity (Torrence, 1986: 128).Bipolar reduction is a conservative reduction technique often employedwhen there is limited access to large nodules (Bayman and Shackley,1999: 843). Limited direct access to large nodules at Yanawilka would

imply that the inhabitants were not well integrated in obsidian ex-change or procurement networks despite their proximity to a majorroad. Another line of evidence to assess whether obsidian scarcity atYanawilka was due to general Inca political mediation is to compareobsidian artifact densities with other settlements under Inca rule, suchas Pulapuco (Abraham, 2010) and settlements that were not under staterule, such as pre-Inca Achanchi and Luisinayoc (Kellett et al., 2013). IfYanawilka has high obsidian density compared to the other sites, thenaccess to obsidian was not restricted. If, however, Yanawilka had lowerthan expected obsidian density, then restriction of access to obsidianwas likely.

2. Identification of Yanawilka as a mitmaqkuna settlementthrough ethnohistoric and archaeological evidence

Identifying mitmaqkuna settlements archaeologically is difficult forseveral reasons (Alconini and Malpass, 2010: 281, 293–295; D'Altroy,2014: 376–377). First, it is difficult to determine whether a settlementwas composed of non-locals solely from the archaeological record. Non-local artifacts, especially ceramics, do not prove that the inhabitants arenot from the area (Van Buren and Presta, 2010: 187). Second, evenwhen one can prove the inhabitants were not locals, for example fromstrontium isotope or biodistance evidence (e.g., Andrushko et al., 2006,2009; Bethard, 2013; Haun and Cock Carrasco, 2010; Turner et al.,2009), it is difficult to distinguish the various reasons for the migration,whether imperially sanctioned or locally motivated. Even if the mi-gration could be identified as imperially sanctioned, distinguishingbetween yanakuna retainer populations and mitmaqkuna populations,both non-local peoples transplanted by the Inca for labor, is proble-matic. Third, mitmaqkuna often settled in an area for multiple genera-tions, so it would be even more difficult after the first generation todistinguish, through artifacts and strontium isotopes, the mitmaqkunafrom their local neighbors. Fourth, the distinction between local re-settlement (i.e., from hilltops to valleys) and long-distance resettlement(most often associated with the mitmaq policy) is fuzzy. Although cul-turally hybridized architecture, foodways, and material culture can besupportive of the identification of mitmaqkuna colonies (e.g., Makowskiand Vega Centeno, 2004; Marcus and Silva, 1988), cultural hybridity by

Fig. 1. Location of Yanawika in relation to nearby major LateHorizon (1437–1532 CE) settlements.

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

214

itself cannot be used as proof.One way to identify mitmaqkuna settlements is to study areas, such

as most areas of the Inca province of Vilcashuamán, where the originalpopulation was completely depopulated by the Inca and then settledwith transplanted mitmaqkuna populations (Valdez and Vivanco, 1994).With this logic, all domestic settlements dated to the Late Horizon(1437–1532 CE) without clear Inca architectural canon in that areashould be mitmaqkuna settlements. Because this information comesfrom ethnohistoric sources, we must employ other lines of evidence tocorroborate the identification of a mitmaqkuna settlement. Multiplelines of evidence converge to show that the site of Yanawilka in Po-macocha was a mitmaqkuna settlement inhabited by a people the Incascalled the Conde.

First, AMS radiocarbon dating and ceramic evidence indicate thatYanawilka was first settled in the Inca Late Horizon. The earliestradiocarbon date, (1424–1439 CE 1σ), was embedded in the floor ofone of the domestic structures excavated (Y3). Diagnostic Inca ceramicswere found underneath large foundation stones of one of the excavatedstructures (Y1), indicating the structure dated to after the beginning ofthe Inca Late Horizon.

Second, the material culture of Yanawilka indicates a humble do-mestic settlement of around one to two hundred people, not a settle-ment where the ruling Inca would have lived. There were around sixtydomestic structures, and they did not manifest Inca architectural canonand masonry, being mostly circular or ovoid and expediently con-structed of stone and possibly adobe. Excavations and site surveyyielded very few diagnostic Inca ceramics, and most of the ceramicswere utilitarian. Some of the decorated ceramics exhibited non-Incastyles.

Third, ethnohistoric evidence strongly supports that the area sur-rounding Yanawilka was settled by mitmaqkuna identified as the Condes(de Carabajal, 1965 [1586]; Piel, 1995; Salas, 1998; Santillana, 2012).They spoke Quechua (de Carabajal, 1965 [1586]: 217; Salas, 2002: 64).In the post-Spanish conquest sixteenth century, Yanawilka belonged tothe Conde mitmaqkuna community of Vischongo (Hu, 2016: AppendixD; Piel, 1995), which supports that the Condes settled Yanawilka duringthe Late Horizon. In the land titles of Pomacocha, which originallybelonged to the community of Vischongo, “Yanavilca” is listed as one ofthe boundary markers of the former “patrimonio de los Incas,” or the“patrimony of the Incas” which the Condes of Vischongo had claimed itas their own through traditional usufruct rights (Hu, 2016: Appendix D:6–8, 13–15). Given its proximity and well documented history withVischongo, Yanawilka was almost certainly inhabited by Conde

mitmaqkuna, who were probably originally from the Arequipa region tothe south (Fig. 2). Yanawilka was appropriated from the Condes ofVischongo under the pretext that Yanawilka belonged to the Inca andtherefore should now belong to the Spaniards (Salas, 1998: 120–121;Hu, 2016: Appendix D: 7–8). Unlike the yanakuna who normally livedat Inca settlements to serve the daily needs of the Inca, the mitmaqkunahad their own settlements and government. Furthermore, according tothe sixteenth century Relaciones Geograficas de Indias, all the originalinhabitants of Vilcashuamán province were forcibly migrated elsewhereto make room for mitmaqkuna groups, with the exception of the Tan-quihua (de Carabajal, 1965 [1586]: 219). The Tanquihua were not theinhabitants of Yanawilka, however, because they lived in communitiesfar from Yanawilka. Thus, all non-Inca domestic settlements near Ya-nawilka from the Late Horizon should be mitmaqkuna settlements.

Although some ambiguities exist regarding whether the Condes ofYanawilka were from different original communities, their politicalorganization post-Spanish conquest indicates that they saw themselvesas part of a larger group with the ethnonym Conde. Although the Condeethnonym was likely imposed by the Inca, the Condes of Vilcashuamánprovince adopted it.

3. Methodology

3.1. Obsidian source assignments through Energy-Dispersive X-RayFluorescence analysis

A Thermo Electron Quant'X Energy Dispersive X-Ray Fluorescence(ED-XRF) machine was used to analyze mid-Zb condition elements ofTi-Nb, Pb, and Th. The X-ray tube operated at 30 kV with the current setautomatically and used a 0.05mm (medium) palladium (Pd) filter. Eachsample was analyzed at 200 s ‘livetime’ and in an air path atmosphere.Trace element data are reported in parts per million (ppm) for titanium(Ti), manganese (Mn), iron (Fe), zinc (Zn), rubidium (Rb), strontium(Sr), yttrium (Y), zirconium (Zr), niobium (Nb), (Shackley, 2011: 203).Shackley calibrated the Quant'X machine using international rockstandards certified by the National Institute of Standards and Tech-nology (NIST), the U.S. Geological Survey (USGS), the Canadian Centrefor Mineral and Energy Technology, and the French Centre de Re-cherches Pétrographiques et Géochemiques (Shackley, 2011: 204).Energy calibration was checked each analysis day using a copper diskand the USGS standard RGM-2 was used to check accuracy with eachanalysis batch. The instrumentation and calibration have shown goodaccuracy over numerous analyses and results are comparable to both

Fig. 2. Map of the approximate area of the Condes' heartland in relation to obsidian sources, the Inca capital of Cuzco, and Yanawilka. Condes' heartland extent based on Julien, 1991:107–108 and Trawick, 2003: 47–48.

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

215

XRF and NAA analyses among laboratories for most elements (Shackley,2011: 33). Most recently, the re-characterization of the Quispisisasource obsidian was carried out using the same instrument (Tripcevichand Contreras, 2011).

The optimum sample size is a minimum of 10mm in diameter and1.2–2.5 mm in thickness, at which “no statistically measurable elementdistortions are observed” (Davis et al., 2011: 61). This level of preci-sion, however, is not necessary for accurate source assignment becauserelative element proportions are not affected, and samples withminimum dimensions of 8mm in diameter and 0.5mm in thicknessshould yield results appropriate for source assignment (Davis et al.,2011: 62). Our sampling strategy was to analyze all available obsidianartifacts above the 8mm diameter and close to the 0.5mm thicknessthreshold; as a result, 84 obsidian artifacts from all excavation unitswere analyzed. The 84 obsidian artifacts constitute 66% (84/128) bycount and 94% (128 g/136 g) by mass of the total obsidian artifactsexcavated from four excavation units (Y1, Y2, Y3, Y4) from Yanawilka.Detailed descriptions of the excavation units and the obsidian artifactsrecovered from them is described in Hu (2016). A bivariate plot of Srversus Rb distinguishes most major South-central Andean obsidiansources. When identification using the bivariate plot is not clear, otherelements such as Fe (ppm), Y (ppm), Zn (ppm), Zr (ppm), and Nb (ppm)were considered (Table 1) (Shackley, 2005; Glascock et al., 2007). Theobsidian artifacts were manually washed to remove excess dirt beforeanalysis.

3.2. Calculating anisotropic travel times to obsidian sources usingNaismith's rule

To calculate the distance decay model for Yanawilka obsidian arti-facts, we took into account the rough terrain of the Andes. Instead ofusing Euclidean distance, which obscures differences in travel times invaried topography, we generated cost surfaces of travel times from the2014 30-meter resolution Shuttle Radar Topography Mission digitalelevation model (NASA) using the r.walk function in the GRASS version0.1 plugin of QGIS version 2.8.2 software. The r.walk function is basedon Naismith's rule, an alternative to Tobler's Hiking Function.Naismith's rule generated slightly better estimates of true hiking timescollected from 120 people, and showed good overall concordance withTobler's Hiking Function (Irtenkauf, 2014). Di Hu's hiking experienceand knowledge of locals' travel times in Vilcashuamán province accordswell with the cost surface generated from Yanawilka as the startingpoint. We then conducted regression analysis to find the strength of thelinear relationship between travel time and proportion of Yanawilkaobsidian represented at the closest high-quality sources. We did notinclude the Puzolana source in our analysis because most of its nodulesdo not measure more than a couple centimeters in the largest dimensionand because the nodules are diffuse in their geologic matrix, making it a

poor candidate for intensive exploitation (Burger and Glascock, 2000).

3.3. Identifying scarcity

The lines of evidence we use to identify obsidian scarcity atYanawilka are the importance of bipolar reduction, the smaller thanexpected size of debitage compared to the size of source nodules, andthe low density of obsidian. Bipolar reduction is evident when flakes aresheared, have diffuse or multiple bulbs of percussion/point of impact orripple direction on the same surface, and when their inner and outersurfaces are not distinct (Ahler, 1989: 210; Kooyman, 2000: 56; Kuhn,1995: 97–98; Odell, 2004: 61). Orange-slice shaped flakes (with cortexas the ‘rind’) or splintered flakes are also generally indicative of bipolarreduction (Shott, 1999). Bipolar reduction can produce flakes of varioussizes and shapes (Shott, 1989). Debitage consist of angular debris,flakes, and blades. For the analytical purposes of this paper, the“blades” category was lumped with debitage: the single blade found atYanawilka is incidental in that it happened to be a complete flake withthe formal properties of a blade. Flakes are defined by having anidentifiable interior surface, and angular debris have uncertain dorsalto ventral orientation (the interior surface is not identifiable) (Shott,1994: 70). Debitage size was assessed through measuring the length,width, thickness, and mass, and obsidian density was measured innumber of obsidian artifacts per cubic meter.

4. Results

4.1. Obsidian source assignments

A bivariate plot of strontium and rubidium aided in assigning thevast majority of obsidian artifacts to obsidian sources (Fig. 3). Quispi-sisa obsidian is predominant at Yanawilka, comprising at least 83% ofthe obsidian artifacts analyzed (Table 2). The next most frequent ob-sidian source utilized at Yanawilka was Jampatilla (10.7%). Becausethe Jampatilla and Lisahuacho sources were only characterized by sixsamples each, we expect that their true 95% confidence ellipses differsignificantly from the present chemical characterization (Glascocket al., 2007). For four of the samples that fell outside of the Jampatillaconfidence ellipse (#63, 73, 84, 104), high yttrium concentrations wereused to assign them to Jampatilla rather than to Lisahuacho (Glascocket al., 2007: 541). One sample (#78) is most likely from an unknownsource due to its high rubidium concentration; nevertheless, #78's otherelemental concentrations are largely consistent with Jampatilla. Thereare three unassigned samples (#81, 132, 139) that could either be Alca-2 or Quispisisa, judging from manganese (Mn) concentrations (Sup-plementary data).

Table 1Obsidian sources and their elemental compositions (ED-XRF) (Glascock et al., 2007: table II; Rademaker et al., 2013: supplement). Alca-1, Alca-2, and Alca 3 have combined means andstandard deviations from Glascock et al., 2007 and Rademaker et al., 2013. Tripcevich and Contreras (2011) have analyzed 34 samples from the Quispisisa source, but elementalcompositions have not yet been published (± =1 standard deviation).

Attribute Alca-1 Alca-2 Alca-3 Jampatilla Lisahuacho Potrero-pampa Quispisisa

N 169 5 32 6 6 6 16K (%) 3.70 ± 0.20 3.57 ± 0.34 3.60 ± 0.21 3.68 ± 0.10 4.03 ± 0.08 3.96 ± 0.26 3.73 ± 0.03Ti (ppm) 831 ± 161 1142 ± 96 1378 ± 96 1164 ± 63 1389 ± 59 534 ± 63 836 ± 17Mn (ppm) 439 ± 52 428 ± 83 398 ± 71 610 ± 43 472 ± 17 547 ± 43 332 ± 12Fe (ppm) 5371 ± 271 6961 ± 178 8118 ± 324 8692 ± 301 8547 ± 135 4593 ± 167 5623 ± 39Zn (ppm) 40 ± 4 46 ± 5 49 ± 7 110 ± 7 85 ± 8 42 ± 3 35 ± 1Ga (ppm) 17 ± 2 15 ± 2 15 ± 2 11 ± 1 18 ± 1 21 ± 3 16 ± 1Rb (ppm) 134 ± 4 141 ± 5 129 ± 5 158 ± 3 150 ± 2 170 ± 6 181 ± 1Sr (ppm) 82 ± 10 142 ± 13 233 ± 19 252 ± 16 310 ± 8 88 ± 7 121 ± 5Y (ppm) 13 ± 2 16 ± 2 15 ± 2 27 ± 2 14 ± 3 16 ± 3 19 ± 2Zr (ppm) 101 ± 9 141 ± 10 159 ± 12 169 ± 10 198 ± 6 91 ± 9 104 ± 3Nb (ppm) 13 ± 2 11 ± 3 12 ± 1 27 ± 3 15 ± 4 13 ± 3 11 ± 1

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

216

4.2. Distance decay model

The correlation at Yanawilka between source proportion and traveltime to source is very weak (r2= 0.147, p= .524), implying that po-litical influences on obsidian exchange existed (Table 3). Even thoughJampatilla is closer to Yanawilka, there was an overwhelming pre-ference for Quispisisa obsidian even though the quality of obsidian atboth sources is good. Although not prohibitively distant and of high-quality, the Lisahuacho and Potreropampa sources were not exploited,at least to any appreciable degree, at Yanawilka.

4.3. Frequency of bipolar reduction in obsidian artifacts and flake sizedistribution

Obsidian was relatively scarce at Yanawilka and productionmethods reflected economizing scarce resources. Production certainlyoccurred on site due to the predominance of debitage (flakes and an-gular debris) in the assemblage (Table 4). Furthermore, the small size ofthe debitage, with 83% (83/99) weighing a gram or less, shows thatproduction occurred within the structures. Half of the obsidian artifacts(64/128) have cortex on the dorsal side or platform. The high pro-portion of obsidian artifacts with cortex also supports production oc-curring on site. Evidence of bipolar reduction is present at all of thestructures and occurs on 13% (17/128) of obsidian artifacts. The bi-polar classification is conservative, as we only included flakes that ex-hibited clear signs of bipolar reduction (at least two identifying attri-butes). The actual proportion of debitage that was created throughbipolar reduction is undoubtedly higher, because a lot of bipolar re-duction shatter does not leave clear evidence (Barham, 1987: 48).Furthermore, there is no significant difference in dimensions and massof debitage with and without evidence of bipolar reduction, making itpossible that some of the debitage classified as non-bipolar neverthelesswere a result of bipolar reduction (Table 5). Surprisingly, bipolar re-duction is evident on obsidian from sources with large nodule sizes,such as Quispisisa (Tripcevich and Contreras, 2011) (Table 6). Hu has

Fig. 3. Bivariate plot of strontium (Sr) and rubidium (Rb) con-centrations of 84 obsidian artifacts from Yanawilka super-imposed on 95% confidence ellipses of obsidian sources.

Table 2Source provenance of obsidian artifacts at Yanawilka.

Unit Alca-3 Jampatilla Quispisisa Unassigned Unknown Total

Y1 1 (4.5%) 6 (27.3%) 13 (59.1%) 1 (4.5%) 1 (4.5%) 22 (100%)Y2 1 (9.1%) 10 (90.9%) 11 (100%)Y3 1 (6.7%) 13 (86.7%) 1 (6.7%) 15 (100%)Y4 1 (2.8%) 34 (94.4%) 1 (2.8%) 36 (100%)Total 1 (1.2%) 9 (10.7%) 70 (83.3%) 3 (3.6%) 1 (1.2%) 84 (100%)

Table 3Frequency distribution of obsidian source provenance at Yanawilka (1460–1532 CE) andtravel times to source, r2= 0.147, p=.524.

Source Frequency Percent Travel time to source

Jampatilla 9 10.7 21.2 hQuispisisa 70 83.3 25.3 hPotreropampa 0 0 32.2 hLisahuacho 0 0 33.4 hAlca-3 1 1.2 64.9 hTotal 80 95.2

Table 4Distribution of obsidian lithic classes among Yanawilka units.

Unit Angulardebris

Blade Core Flake Flake tool ProjectilePoint

Total

Y1 22 1 19 2 1 45Y2 4 4 3 11Y3 7 1 2 6 3 19Y4 15 25 12 1 53Total 48 1 1 50 23 5 128

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

217

personally seen some Quispisisa nodules at the primary source mea-suring 40 to 50 cm in the largest dimension. The prevalence of bipolarreduction on obsidian from a source with large nodules is a classic in-dicator of resource scarcity and minimal direct procurement. Utilizingbipolar reduction shows that the inhabitants of Yanawilka were keen toconserve a scarce resource.

The density of obsidian at Yanawilka was low. Even a smaller-sizednodule from Quispisisa could more than account for all the mass ofQuispisisa obsidian excavated at Yanawilka. Large debitage was rare,and the vast majority clustered around 1.4 cm in length and 1.0 cm inwidth, implying that the original cores, preforms, or source materialstarted out small (Table 7). Although the inhabitants of Yanawilka hadaccess to the Jampatilla and Quispisisa sources, they did not have greatquantities of obsidian from those sources. Although some obsidian re-duction undoubtedly occurred outside of the structures, we do not ex-pect it to be significantly different than inside the structures because weexcavated a small area outside of Y1 and did not find any debitage, onlyone intact obsidian point. Survey of Yanawilka also confirmed the lowdensities of obsidian on site (Hu, 2016: 74).

5. Comparisons with nearby Late Intermediate Period and LateHorizon sites

5.1. Distance decay model comparison with Achanchi and Luisinayoc

To contextualize the distance decay model of obsidian at Yanawilka,we compare Yanawilka (ca. 1450–1532 CE) to the Chanka sites ofAchanchi (1227–1315 CE 1σ) and Luisinayoc (1045–1221 CE 1σ)(Kellett et al., 2013). There are three reasons for comparing Yanawilkato these two sites. First, by comparing Yanawilka to settlements thatpre-dated state hegemony, we can better understand the role that theInca may have played in access to obsidian. Second, no other majorobsidian provenance study has been carried out in the area for eitherthe Late Intermediate Period (1000–1450 CE) or the Late Horizon(1450–1532 CE). Third, the differences in travel times from the sites toa particular obsidian source are less than 8 h (Table 8). Therefore, travel

times alone should not prevent any one source from being potentiallyexploited.

Although the absolute number of obsidian sources exploited doesnot differ significantly among Yanawilka and the Chanka sites, there aresubstantial differences in emphasis (Table 9; Fig. 4). At Yanawilka,Quispisisa is the overwhelming preference, and at the Chanka sites ofAchanchi and Luisinayoc, the vast majority of the obsidian used comesfrom the Lisahuacho and Potreropampa obsidian sources, which are lessthan 5-10 km away from each other. Because the obsidian sourcespresent at Achanchi and Luisinayoc all came from either south or westof Andahuaylas as opposed to the east toward Cusco and Arequipa,Kellett et al. (2013: 1898) suggest that the Chanka obsidian patternsreflected stronger cultural ties to the Apurimac and Ayacucho regions.This is consistent with ethnohistorical sources that say the Chanka werea confederation of allied groups of the Apurimac and Ayacucho regionswho opposed Inca expansion (Bauer et al., 2010).

There was no Lisahuacho and Potreropampa obsidian among theartifacts analyzed from Yanawilka, which is surprising given that thosesources comprised 80.8% of the total assemblage of Achanchi andLuisinayoc. Potreropampa had been a major source of archaeologicalobsidian in the area from 2500 BCE–1400 CE (Kellett et al., 2013:1899), and it is roughly equidistant from the Andahuaylas area as it isfrom the Vilcashuamán area. The dominance of Quispisisa at Yanawilka(83.3%) and of Lisahuacho/Potreropampa at Achanchi and Luisinayoc(80.8%) show a stark difference in obsidian source preference. A pos-sible explanation is that since Lisahuacho and Potreropampa were lo-cated in the Chanka territories, the Inca may have discouraged theCondes at Yanawilka from having direct economic relations with theInca-unfriendly Chanka groups. Alternatively, the Condes themselvesmay not have traditionally pursued obsidian exchange with the groupsin the Apurimac Chanka area.

At Yanawilka, distant Alca obsidian is present at very low quan-tities, with only 1 out of 84 artifacts analyzed being from this source.The Alca obsidian source is located in the Condes homeland, and itspresence at Yanawilka could be a result of weak or infrequent exchangewith groups from their homeland. Another possibility is that the Alcaobsidian (specifically Alca-3, found around Cerro Aycano) was carriedfrom the homeland as part of the initial migration. Direct procurementafter initial migration is less likely given the low quantities, small sizeof the debitage, and distance of the Alca obsidian source. Just as theWari Empire had spread and intensified the use of Quispisisa obsidian(Burger et al., 2000: 343–344, 351; Burger et al., 2016; Kellett et al.,2013: 1899), the dominance of Quispisisa obsidian at Yanawilka maybe a result of similar Inca imperial interventions in interregional ex-change. Unlike the Wari, however, Inca hegemony may not havetranslated into the intensification of obsidian exchange and use by theirsubjects. Obsidian source provenance research at other Late Horizonand Late Intermediate Period sites is needed to test this hypothesis.

The combined source proportions at Achanchi and Luisinayoc areconsistent with the standard distance decay model (r2= 0.768,p= .051), which is expected if no state intervention in regional ex-change was present (Table 10). The pattern at Yanawilka is moreconsistent with controlled traffic than the pattern at Achanchi and

Table 5Attributes of all obsidian artifacts with and without evidence of bipolar reduction.

Bipolarevidence?

N Mean ofmass (g)

Mean oflength (cm)

Mean ofwidth (cm)

Mean ofthickness (cm)

Yes 17 0.9 1.6 1.1 0.4No 111 1.1 1.6 1.1 0.4Total 128 1.1 1.6 1.1 0.4

Table 6Evidence of bipolar reduction on Jampatilla and Quispisisa obsidian artifacts.

Bipolar evidence? Jampatilla Quispisisa Total

Yes 1 13 14No 8 57 65Total 9 70 79

Table 7Attributes of debitage (angular debris, blade, and flakes).

Lithicclass

N Mean and SDof mass (g)

Mean and SDof length (cm)

Mean and SDof width (cm)

Mean and SDof thickness(cm)

Angulardebris

26 1.1 ± 0.9 1.3 ± 0.5 0.9 ± 0.4 0.4 ± 0.2

Blade 1 0.3 1.6 0.8 0.2Flake 32 1.1 ± 1.8 1.6 ± 0.9 1.0 ± 0.4 0.3 ± 0.2Total 59 1.1 ± 1.4 1.4 ± 0.7 1.0 ± 0.4 0.4 ± 0.2

Table 8Anisotropic travel times from archaeological sites to obsidian sources calculated using2014 SRTM digital elevation model and QGIS GRASS plugin's r.walk function (Naismith'srule).

Source Yanawilka Achanchi Luisinayoc

Quispisisa 25.3 h 41.7 h 41.9 hJampatilla 21.2 h 36.5 h 36.8 hAlca-3 64.9 h 54.0 h 52.7 hLisahuacho 33.4 h 26.1 h 25.1 hPorteropampa 32.2 h 25.6 h 24.5 hMean travel times (all) 35.4 h 36.8 h 36.2 h

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

218

Luisinayoc.

5.2. Comparisons of obsidian artifact densities at Achanchi, Luisinayoc,Pulapuco, and Achanchi

To assess how representative the scarcity of obsidian at Yanawilkawas during the Late Horizon under Inca hegemony, we compare ob-sidian densities of Pulapuco, a settlement of Lucanas people with anInca administrative core, to Yanawilka (Abraham, 2010). Pulapuco'sdistance from the major obsidian sources discussed is comparable toYanawilka and the Chanka sites, so the low density is not a function ofdistance. Both settlements had relatively low densities of obsidiancompared to the pre-Inca Chanka sites of Achanchi and Luisinayoc(Table 11). We suspect that the low obsidian densities at Pulapuco and

Yanawilka are common, if not the norm, among sites in areas of firmInca hegemony. We interpret the low density of obsidian at Yanawilkaas a result of restriction of interregional traffic of obsidian. Such controlof traffic is likely a result of Inca imperial policies, as other scholarshave noted. Unfortunately, other than the sites of Pulapuco, Achanchi,and Luisinayoc, there is a lack of published information on obsidiandensities at nearby sites, which precludes further comparisons.

6. Discussion and conclusions

This article represents the first major obsidian provenance study inthe core of the Inca Empire and presents intriguing preliminary pat-terns. Diverse lines of evidence show that access to obsidian at

Table 9Source provenance of obsidian artifacts at Achanchi and Luisinayoc.

Site Jampatilla Lisahuacho Potreropampa Quispisisa Unknown Total

Achanchi 1 (6.3%) 15 (93.8%) 16 (100%)Luisinayoc 1 (2.8%) 18 (50%) 8 (22.2%) 5 (13.9%) 4 (11.1%) 36 (100%)Total 1 (1.9%) 19 (36.5%) 23 (44.2%) 5 (9.6%) 4 (7.7%) 52 (100%)

Fig. 4. Obsidian sources utilized by Yanawilka (LH),Achanchi and Luisinayoc (LIP). Line thickness is propor-tional to the square root of the percentage of obsidian arti-facts at the site from the source.

Table 10Frequency distribution of obsidian source provenance at Achanchi (1227–1315 CE 1σ)and Luisinayoc (1045–1221 CE 1σ), and travel time in hours to source (averaged timesfrom Achanchi and Luisinayoc), r2= 0.768, p= .051.

Source Frequency Percent TravelfromAchanchi

Travel fromLuisinayoc

Achanchi/Luisinayocaveraged

Potreropampa 23 44.2 25.6 h 24.5 h 25.1 hLisahuacho 19 36.5 26.1 h 25.1 h 25.6 hJampatilla 1 1.9 36.5 h 36.8 h 36.7 hQuispisisa 5 9.6 41.7 h 41.9 h 41.8 hAlca-3 0 0 54.0 h 52.7 h 53.4 h

Table 11Obsidian artifact densities at four sites (Abraham, 2010; Kellett, 2010). Sites are orderedchronologically from youngest to oldest.

Site No. ofobsidian

Volume excavated(m3)

Density (No. ofobsidian/m3)

Yanawilka (Condes,LH)

128 25.93 4.94

Pulapuco (Lucanas,LH)

120 33.54 3.58

Achanchi (Chanka,LIP)

228 9.06 25.17

Luisinayoc (Chanka,LIP)

156 8.63 18.08

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

219

Yanawilka was politically mediated and consistent with what otherscholars have noted about the highly controlled nature of interregionaltraffic under the Inca (Murra, 1980; Yacobaccio et al., 2002). Access toobsidian was restricted at Yanawilka compared to the pre-Inca Chankasites of Achanchi and Luisinayoc both qualitatively (restriction of accessto certain sources) and quantitatively (amount of obsidian).

Despite being in an area of high economic connectivity, the in-habitants of Yanawilka seem to have been economically marginalized.They did not seem to have high-status goods such as fancy ceramics ormetal objects, and their domestic structures were expediently con-structed (Hu, 2016). The obsidian provenance data at Yanawilka doesnot contradict ethnohistoric claims and previous archaeological re-search on Inca regulation and division of the larger political landscapeto their advantage. The obsidian at Yanawilka only came from onedirection: south and southwest. Notably, obsidian did not come fromobsidian sources in Chanka territory, even though those sources are notmuch further away. Before the Inca, the Ayacucho area, which includedInca Vilcashuamán province, had ties to the Andahuaylas area (Kellettet al., 2013). Obsidian was scarce at Yanawilka despite its proximity tomultiple major high quality sources of obsidian. By controlling majortraffic on the royal Inca roads, the Inca presumably were also themediators of a significant, if not majority, portion of interregional in-teraction. The literature that states that certain mitmaqkuna groupswere socially and economically elevated by the Inca did not apply to theCondes at Yanawilka. The agricultural focus of Yanawilka might be afactor in their economic marginalization, because at other sites, likeMilliraya, the mitmaqkuna were craft specialists and may have beenelevated by the Inca (Alconini, 2013; Spurling, 1992). An importantimplication of the obsidian provenance study at Yanawilka is that eventhe mitmaqkuna afforded special status and considered friendly by theInca may have been restricted in interethnic interactions, both locallyand to other regions. The Inca strategy was to minimize political alli-ances over the wider social landscape that would pose a threat to theirrule.

Distinguishing whether the Inca intended to control obsidian accessor whether the difficulty in accessing large quantities of obsidian from avariety of different sources was a side-effect of the Inca reshuffling thesocial landscape is difficult. The scarcity of obsidian at Yanawilka couldalso have been a result of the Inca fragmenting the larger politicallandscape, which would have disrupted the previous obsidian exchangenetworks. That there was obsidian available, albeit scarce, implies thatthe Inca did not or could not implement a fully restrictive economicpolicy. Future research could explore intra-site variation in obsidiansource access at Yanawilka to test whether households were evenlyaffected by restricted access. This article contributes a significant ad-vance in our understanding of obsidian exchange in the Late HorizonPeru, especially of a core imperial area like Vilcashuamán. The data canbe used in future social network analyses of multiple sites in the areaonce more provenance studies are done.

Acknowledgements

The authors gratefully acknowledge Tim Teague and the Earth andPlanetary Science XRF laboratory for assistance. The excavations atYanawilka and subsequent laboratory analyses were funded by theFulbright -Hays Doctoral Dissertation Research Abroad Fellowship,Wenner -Gren Dissertation Fieldwork Grant (#8265), the Stahl Fund,and a Ford Foundation Dissertation Fellowship. Excavations atYanawilka were carried out under the Ministry of Culture-PeruResolución Directoral No 085-DGPC-VMPCIC/MC and the exportationof obsidian artifacts from Peru to the US was carried out under theResolución Viceministerial No. 038-2014 VMPCIC-MC. We also aregrateful to Lucas C. Kellett for providing most of the shapefiles of theobsidian source locations. We thank the four anonymous reviewers whogreatly improved the focus and clarity of the article.

Appendix A. Supplementary data

Supplementary data associated with this article can be found in theonline version, at doi:https://doi.org/10.1016/j.jasrep.2018.01.018.These data include the Google map of the most important areas de-scribed in this article.

References

Abraham, S.J., 2010. Provincial Life in the Inca Empire: Continuity and Change atPulapuco, Peru. Unpublished Ph.D. Dissertation. University of California, SantaBarbara.

Ahler, S.A., 1989. Experimental knapping with KRF and mid-continental cherts: overviewand applications. In: Amick, D.S., Mauldin, R.P. (Eds.), Experiments in LithicTechnology. BAR International Series. Archaeopress, Oxford, pp. 67–99.

Alconini, S., 2013. El territorio kallawaya y el taller alfarero de milliraya: Evaluación dela producción, distribución e intercambio interregional de la cerómica inka pro-vincial. Chungará 45, 277–292.

Alconini, S., Malpass, M.A., 2010. Toward a better understanding of Inka provincialism.In: Malpass, M.A., Alconini, S. (Eds.), Distant Provinces in the Inka Empire: Toward aDeeper Understanding of Inka Imperialism. University of Iowa Press, Iowa City, pp.279–299.

Andrushko, V.A., Pino, E.C.T., Bellifemine, V., 2006. The burials at Sacsahuaman andChokepukio: a bioarchaeological case study of imperialism from the capital of theInca Empire. Ñawpa Pacha 28, 63–92.

Andrushko, V.A., Buzon, M.R., Simonetti, A., Creaser, R.A., 2009. Strontium isotopeevidence for prehistoric migration at Chokepukio, Valley of Cuzco, Peru. Lat. Am.Antiq. 20, 57–75.

Barham, L.S., 1987. The bipolar technique in southern Africa: a replication experiment. S.Afr. Archaeol. Bull. 42, 45–50.

Bauer, B.S., Kellett, L.C., Aráoz, M., 2010. The Chanka: Archaeological Research inAndahuaylas (Apurimac), Peru. Los Angeles. The Cotsen Institute of Archaeology,University California, Los Angeles.

Bayman, J.M., Shackley, M.S., 1999. Dynamics of Hohokam obsidian circulation in theNorth American Southwest. Antiquity 73, 836–845.

Bethard, J.D., 2013. The Bioarchaeology of Inka Resettlement Practices: Insight fromBiological Distance. Unpublished Ph.D. Dissertation. University of Tennessee,Knoxville.

Burger, R.L., Glascock, M.D., 2000. The Puzolana obsidian source: locating the geologicsource of Ayacucho type obsidian. Andean Past 6, 289–307.

Burger, R.L., Chávez, K.L., Chávez, S., 2000. Through the glass darkly: prehispanic ob-sidian procurement and exchange in southern Peru and northern Bolivia. J. WorldPrehist. 14, 267–362.

Burger, R.L., Bencic, C., Glascock, M.D., 2016. Obsidian procurement and cosmopoli-tanism at the Middle Horizon settlement of Conchopata, Peru. Andean Past 12,21–44.

de Carabajal, P., 1965. Descripción Fecha de la Provincia de Vilcas Guaman Por el IllustreSeñor Don Pedro de Carabajal, Corregidor y Justicia Mayor Della, Ante Xpristobal deGamboa, Escribano de su Juzgado, en el Año de 1586. In: Jiménez de la Espada, M.,Martínez Carreras, J.U. (Eds.), Relaciones Geográficas de Indias, Biblioteca deAutores Españoles. Real Academia Española, Madrid, pp. 205–219.

Cieza de León, P., 1959. The Incas. In: The Civilization of the American Indian Series.University of Oklahoma Press, Norman.

Cieza de León, P., 1984. Crónica del Peru, Colección Clásicos Peruanos. Fondo Editorialde la Pontificia Universidad Católica del Perú: Academia Nacional de la Historia,Lima.

D'Altroy, T.N., 1992. Provincial Power in the Inka Empire. Smithsonian Institution Press,Washington.

D'Altroy, T.N., 2014. The Incas, Second edition. Wiley Blackwell, Chichester, England.Davis, M.K., Jackson, T.L., Shackley, M.S., Teague, T., Hampel, J.H., 2011. Factors

Affecting the Energy-Dispersive X-Ray Fluorescence (EDXRF) analysis of archae-ological obsidian. In: Shackley, M.S. (Ed.), X-ray Fluorescence Spectrometry (XRF) inGeoarchaeology. Springer, New York, pp. 45–63.

Glascock, M.D., Speakman, R.J., Burger, R.L., 2007. Sources of archaeological obsidian inPeru: descriptions and geochemistry. In: Glascock, M.D., Speakman, R.J., Popelka-Filcoff, R.S. (Eds.), Archaeological Chemistry: Analytical Techniques andArchaeological Interpretation. American Chemical Society, Washington, DC, pp.522–552.

Guamán Poma de Ayala, F., 1615. El primer nueva corónica y buen gobierno [WWWDocument]. The Guaman Poma Website. http://www.kb.dk/permalink/2006/poma/info/en/frontpage.htm, Accessed date: 4 August 2016.

Haun, S.J., Cock Carrasco, G.A., 2010. A bioarchaeological approach to the search formitmaqkuna. In: Malpass, M.A., Alconini, S. (Eds.), Distant Provinces in the InkaEmpire: Toward a Deeper Understanding of Inka Imperialism. University of IowaPress, Iowa City, pp. 193–220.

Hu, D., 2016. Labor under the Sun and the Son: Landscapes of Control and Resistance atInka and Spanish Colonial Pomacocha, Ayacucho, Peru. Unpublished Ph.D.Dissertation. University of California, Berkeley.

Hyslop, J., 1984. The Inka Road System. Academic Press, Orlando.Irtenkauf, E., 2014. Analyzing Tobler's Hiking Function and Naismith's Rule Using Crowd-

Sourced GPS Data. The Pennsylvania State University. https://gis.e-education.psu.edu/sites/default/files/capstone/Irtenkauf_596B_20140430.docx, Accessed date: 1November 2017.

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

220

Julien, C.J., 1991. Condesuyo: The Political Division of Territory under Inca and SpanishRule. Bonner Amerikanistische Studien, Bonn.

Kellett, L.C., 2010. Chanka Settlement Ecology: Hilltop Sites, Land Use and Warfare inLate Prehispanic Andahuaylas, Peru. Unpublished Ph.D. Dissertation. University ofNew Mexico.

Kellett, L.C., Golitko, M., Bauer, B.S., 2013. A provenance study of archaeological ob-sidian from the Andahuaylas region of southern Peru. J. Archaeol. Sci. 40,1890–1902.

Kooyman, B.P., 2000. Understanding Stone Tools and Archaeological Sites. University ofCalgary Press, Calgary.

Kuhn, S.L., 1995. Mousterian Lithic Technology: An Ecological Perspective. PrincetonUniversity Press, Princeton.

Levine, M.N., Joyce, A.A., Glascock, M.D., 2011. Shifting patterns of obsidian exchange inPostclassic Oaxaca, Mexico. Anc. Mesoam. 22, 123–133.

Makowski, K., Vega Centeno, A.M., 2004. Estilos regionales en la Costa Central en elHorizonte Tardío: Una aproximación desde el valle del Lurín. 33. Bulletin de l'Institutfrançais d'études andines, pp. 681–714.

Marcus, J., Silva, J.E., 1988. The Chillón Valley “Coca Lands”: archaeological backgroundand ecological context. In: Rostworowski, M. (Ed.), Conflicts Over Coca Fields inSixteenth-century Peru, Memoirs of the Museum of Anthropology. University ofMichigan, Ann Arbor, pp. 1–32.

Mills, B.J., Clark, J.J., Peeples, M.A., Haas, W.R., Roberts, J.M., Hill, J.B., Huntley, D.L.,Borck, L., Breiger, R.L., Clauset, A., Shackley, M.S., 2013. Transformation of socialnetworks in the late pre-Hispanic US Southwest. PNAS 110, 5785–5790.

Molyneaux, B.L., 2002. Exploring the Landscapes of long-distance exchange: evidencefrom Obsidian Cliffs and Devils Tower, Wyoming. In: Glascock, M.D. (Ed.),Geochemical Evidence for Long Distance Exchange. Bergin and Garvey, Westport,Connecticut, pp. 133–151.

Murra, J.V., 1980. The Economic Organization of the Inka State. JAI Press, Greenwich,Connecticut.

Odell, G.H., 2004. Lithic analysis. In: Manuals in Archaeological Method, Theory, andTechnique. Kluwer Academic/Plenum Publishers, New York.

Ogburn, D., Connell, S., Gifford, C., 2009. Provisioning of the Inka army in wartimeobsidian procurement in Pampamarca, Ecuador. J. Archaeol. Sci. 36, 740–751.

Piel, J., 1995. Capitalismo Agrario en el Perú. IFEA, Lima.Polo de Ondegardo, J., 1873. Report by Polo de Ondegardo of the lineage of the Yncas,

and how they extended their conquests. In: Markham, C.R. (Ed.), Narratives of theRites and Laws of the Yncas, Works Issued by the Hakluyt Society. Hakluyt Society,London, pp. 151–171.

Rademaker, K., Glascock, M.D., Kaiser, B., Gibson, D., Lux, D.R., Yates, M.G., 2013. Multi-technique geochemical characterization of the Alca obsidian source, Peruvian Andes.Geology 41, 779–782.

Renfrew, C., 1984. Approaches to Social Archaeology. Harvard University Press,Cambridge.

Rowe, J.H., 1946. Inca culture at the time of the Spanish conquest. In: Steward, J.H. (Ed.),Handbook of South American Indians. Smithsonian Institution Press, Washington, pp.183–330.

Rowe, J.H., 1982. Inca policies and institutions relating to the cultural unification of the

empire. In: Collier, G.A., Rosaldo, R., Wirth, J.D. (Eds.), The Inca and Aztec States,1400–1800: Anthropology and History. Academic Press, New York, pp. 93–118.

Salas, M., 1998. Estructura Colonial Del Poder Español En El Perú: Huamanga (Ayacucho)a Través De Sus Obrajes: Siglos XVI-XVIII. Pontificia Universidad Católica del Perú.Fondo Editorial, Lima.

Salas, M., 2002. Advenedizos y traspuestos: los mitmaquna o mitimaes de Vilcashuamánen su tránsito de los tiempos del Inka al de los “Señores de los mares”. 6. Boletín deArqueología PUCP, pp. 57–78.

Santillana, J.I., 2012. Paisaje sagrado e ideología inca: Vilcas Huamán. Fondo Editorial dela Pontificia Universidad Católica del Perú. Lima, Perú.

Sarmiento de Gamboa, P., 2010. History of the Incas. Works Issued by the HakluytSociety. Ashgate, Farnham, Surrey, Burlington, VT.

Shackley, M.S., 2005. Obsidian: Geology and Archaeology in the North AmericanSouthwest. University of Arizona Press, Tucson.

Shackley, M.S. (Ed.), 2011. X-ray Fluorescence Spectrometry (XRF) in Geoarchaeology.Springer, New York.

Shott, M.J., 1989. Bipolar industries: ethnographic evidence and archaeological im-plications. N. Am. Archaeol. 10, 1–24.

Shott, M.J., 1994. Size and form in the analysis of flake debris: review and recent ap-proaches. J. Archaeol. Method Theory 1, 69–110.

Shott, M.J., 1999. On bipolar reduction and splintered pieces. North AmericanArchaeologist 20, 217–238.

Spurling, G.E., 1992. The Organization of Craft Production in the Inca State: The Pottersand Weavers of Milliraya. Unpublished Ph.D. Dissertation. Cornell University,Ithaca, NY.

Torrence, R., 1986. Production and Exchange of Stone Tools: Prehistoric Obsidian in theAegean. New studies in Archaeology. Cambridge University Press, Cambridge, NewYork.

Trawick, P.B., 2003. The Struggle for Water in Peru: Comedy and Tragedy in the AndeanCommons. Stanford University Press, Stanford, CA.

Tripcevich, N., Contreras, D.A., 2011. Quarrying evidence at the Quispisisa obsidianSource, Ayacucho, Peru. Lat. Am. Antiq. 22, 121–136.

Turner, B.L., Kamenov, G.D., Kingston, J.D., Armelagos, G.J., 2009. Insights into im-migration and social class at Machu Picchu, Peru based on oxygen, strontium, andlead isotopic analysis. J. Archaeol. Sci. 36, 317–332.

Valdez, L.M., Vivanco, C., 1994. Arqueologia de la Cuenca del Qaracha, Ayacucho, Peru.Lat. Am. Antiq. 5, 144–157.

Van Buren, M., Presta, A.M., 2010. The organization of Inka silver production in Porco,Bolivia. In: Malpass, M.A., Alconini, S. (Eds.), Distant Provinces in the Inka Empire:Toward a Deeper Understanding of Inka Imperialism. University of Iowa Press, IowaCity, pp. 173–192.

Yacobaccio, H.D., Escola, P., Lazzari, M., Pereyra, F.X., 2002. Long-distance obsidiantraffic in Northwestern Argentina. In: Glascock, M.D. (Ed.), Geochemical Evidence forLong Distance Exchange. Bergin and Garvey, Westport, Connecticut, pp. 167–203.

Yacobaccio, H.D., Escola, P.S., Pereyra, F.X., Lazzari, M., Glascock, M.D., 2004. Quest forancient routes: obsidian sourcing research in Northwestern Argentina. J. Archaeol.Sci. 31, 193–204.

D. Hu, M.S. Shackley Journal of Archaeological Science: Reports 18 (2018) 213–221

221