Tripcevich PhD Diss

Quarries, Caravans, and Routes to Complexity:
Prehispanic Obsidian in the South-Central Andes

Nicholas Tripcevich, 2007
Ph.D. Dissertation, Anthropology
University of California, Santa Barbara Committee in charge:
Mark Aldenderfer
, Michael Jochim , Katharina Schreiber , Keith Clarke
  PDF Version
(27 Mb)
Original format and pagination.  


  Web Version Choose from a chapter below, or from the right menu. Content appears in site searches.  


Ch0 Introduction

Cover Page


Santa Barbara



Quarries, Caravans, and Routes to Complexity:

Prehispanic Obsidian in the South-Central Andes



A Dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

in Anthropology




Nicholas Tripcevich


Committee in charge:

Professor Mark Aldenderfer, Co-Chair

Professor Michael Jochim, Co-Chair

Professor Katharina Schreiber

Professor Keith Clarke

March 2007


Signature Page


                 Quarries, Caravans, and Routes to Complexity:

                 Prehispanic Obsidian in the South-Central Andes


                                 Copyright © 2007


                                 Nicholas Tripcevich



                       The dissertation of Nicholas Tripcevich is approved.



                        Keith Clarke



                        Katharina Schreiber



                        Michael Jochim, Committee Co-Chair



                        Mark Aldenderfer, Committee Co-Chair


                                         March 2007




Acknowledgements go first to the people of the Colca valley who have kindly welcomed the foreigners who hike through their lands and camp on the volcanoes that ring the valley. I would like to acknowledge the support of my dissertation committee, Mark Aldenderfer, Katharina Schreiber, Michael Jochim, and Keith Clarke, who provided valuable guidance in this process. I am particularly grateful to my advisor Mark Aldenderfer who shares my fondness for mountain environments and the people who thrive there. My background was primarily in geography when Mark admitted me to the program at UCSB, and I am grateful that he recognized the contribution that a geographical emphasis can make to anthropological research. I would like to thank Katharina Schreiber who has terrific enthusiasm for the Andes and for prehistory, Michael Jochim has been willing to provide guidance and feedback at crucial moments during this research, and Keith Clarke whose knowledge of geography and enthusiasm for mobile geographical technologies were inspiring. Although not on my committee, Charles Stanish has been consistently encouraging and has provided valuable insights and a useful regional perspective.

This dissertation research was made possible by a doctoral dissertation improvement grant from the Nation Science Foundation and a UCSB graduate division dissertation fellowship. The fieldwork was conducted with a permit from the Instituto Nacional de Cultura - Perú. The UC Santa Barbara Anthropology Department and Letters and Science Information Technology, and the GIS division at URS Corporation/Santa Barbara have provided me with part-time work and have kindly put up with my odd schedule during this writing up process.

The Titicaca Basin is an excellent introduction to the Andes as it possesses enduring Andean cultural traditions, rich archaeology, and a collegial community of archaeologists that work there. My thanks to the many colleagues who have shared their insights and their data from the Titicaca Basin research, and to Cecilia Chávez and Edmundo de la Vega at the Puno house. Ever supportive, Karen Doehner contributed valuable advice and generously translated key documents. This research project is regional in scope and it inevitably has resulted in debts of gratitude to many people whose data and observations have contributed to this exploration of Chivay obsidian and regional interaction.

Striking out from the Titicaca Basin to work in Arequipa was made easier by the warm community and valuable facilities that I found there. The Centro de Investigaciones Arqueológicas de Arequipa (CIARQ) was a important base for field operations and a hospitable environment for lab analysis, and special thanks go to Karen Wise and Augusto Cardona for establishing and continuing this valuable facility. I am especially grateful for Augusto’s logistical support for this project, his enthusiasm and expertise in Arequipa archaeology, and for his hosting of memorable ceviche parties at CIARQ. The INC offices in Arequipa, in particular Marco Lopez, Cecilia Quequesana, and Lucy Linares, helped to make the bureaucratic hurdles less difficult.

This project was made possible by the energetic participation of a committed team of researchers and students from many countries. The enthusiasm of co-director Willy Yépez Alvarez made fieldwork at the high altitude source a pleasure, and made subsequent lab work in Arequipa and coordination with the INC proceed smoothly. Important contributions to the field research were made by Cheyla Samuelson and Alex Mackay who generously contributed months, even years, away from their own research goals to greatly advance this project and the data that has resulted from it. The theoretical and methodological merit of the lithics analysis in this project is largely due to Alex Mackay’s keen field observations, efficient lab procedures, and unparalleled speed with the digital calipers. Saul Morales has been a valuable friend and collaborator nearly every year since I first met him in Juli, Peru eight years ago. As Saul anticipated in 1999, he was an important contributor to this field project, and he was also an astute cultural mediator who relies sometimes on the axiom “ en Perú, todo es posible”. A sincere thank you is also due to all of the other tireless participants of the Upper Colca project, a list that includes Ross Burley, Mirza del Castillo, Erik Erwin, Brian Finucane, Guillermo Flores, Tamara Flores Ramos, Melissa Joubert, Chris Lagen, Adan Lacunza, José Ubeda, and Daniel Zimbler.

A great deal of credit for turning this geographer into an anthropologist goes to my colleagues at UCSB, in particular Elizabeth Klarich and Nathan Craig. I would like to thank Liz Klarich for her friendship and because her explanations of theoretical and regional subjects have been invaluable. Liz has been a reliable source of sound advice about grad school since I first visited Santa Barbara through to consultation about organizing this dissertation. Nathan Craig is a never-ending source of ideas for original ways to apply GIS to anthropological problems, for training in meticulous excavation work, and for demonstrating how tough and focused work in the field can be enjoyable.

I would like to thank Cynthia Herhahn for demonstrating how to supervise a project and for field training when I worked for her in Juli, my first year in Peru. Thanks also to Cynthia for joining us in 2003 in the Colca for a three day high altitude survey/backpacking death march, and thanks for follow-up discussions of my research as I wrote up this work. I am grateful also to Cindy Klink who provided advice and shared results from her work in the Ilave and Huenque valleys. Justin Jennings has been a good colleague as he has shared material from his work in Cotahuasi as well as valuable insights about working in the Arequipa. I would like to thank Steve Wernke for his friendship and collegiality, and for his advice about research and the community in the Colca valley. Other colleagues that have provided advice and encouragement include Matthew Bandy, David Browman, Michele Buzon, Augusto Cardona, Christi Conlee, José-Antonio Chávez, Miriam Doutriaux, Tobias Fischer, Sarah Fowler, Anabel Ford, Martín Giesso, Michael Glassow, Paul Goldstein, Ian Lindsay, Michael Malpass, Daniel Sandweiss, Ericka Simborth, Bruce Owen, Felix Palacios, Kurt Rademaker, Frank Spera, Waldo Tobler, Tiffiny Tung, Hendrik Van Gijseghem, Kevin Vaughn, and Ryan Williams. In the Colca valley, Timoteo and Adriana Valdivia, Eliseo Vilcahuaman Panibra, la Familia Espinel, la Familia Sotormayor, Saturnino Ordóñez, and Padre Franz Windischhofer all contributed in important ways to the success of this project.

Acknowledgement is due as well to the efforts of Richard Burger and Sarah Brooks and their collaborators in geochemistry for locating and first documenting the obsidian deposits in the Colca valley. Both researchers provided me with useful advice in carrying out this project. Michael Glascock at M.U.R.R. and M. Steven Shackley at UC Berkeley have both been generous with their time in educating me about geochemical analysis, and in analyzing obsidian source samples from Peru. Spatial data and imagery that were essential to this work was acquired from several agencies free-of-cost, and thanks are due to NASA, USGS, NIMA, Digital Perú, DigitalGlobe (GoogleEarth), and the IGS base station in Arequipa for GPS correction. Mark Aldenderfer was generous with equipment and I wish to thank him for the loan of GPS equipment and cameras for several seasons in a row, as well as for his truck that served as our field vehicle.

Finally, I would like to thank my parents, Robert Tripp and Susan Ervin-Tripp, for their interest, encouragement, and support through all stages of my research. My brother Alexander and sister Katya were likewise supportive. I would like to acknowledge the kind advice and direction from Laura Nader, family friend, and advocate for an integrative view of anthropology. My deepest love and gratitude goes to my wife Cheyla Samuelson who has contributed emotional support, encouragement, and effort to advancing my Ph.D. project from the beginning. She enthusiastically participated in the rigorous 5000m portion of the research, and her keen intellect and constructive observations helped considerably in directing this project. Cheyla spotted more than her share of projectile points during survey, and thanks to her concern for the crew while camping at the obsidian source we can say that “no students were permanently damaged in the making of this dissertation”!



Regional studies of obsidian artifacts in the south-central Andes have shown that over 90% of the analyzed artifacts from the Lake Titicaca Basin belong to a single geochemical obsidian type. A decade ago researchers identified the geological origin of this obsidian type as the Chivay / Cotallalli source, located 180km west of Lake Titicaca above the Colca valley in Arequipa at 71.5355° S, 15.6423° W (WGS84), and at 4972 meters above sea level. This research project focused on the obsidian source and adjacent lands within one day’s travel from the source. The project included a 33 km2 survey, 8 test units, and in-depth lithic attribute analysis. Mobile GIS (Arcpad) was used extensively during survey. A substantial quarry pit and an obsidian workshop were examined closely, as were consumption sites in nearby areas. The results of this study found that the earliest diagnostic materials at the source date to the Middle Archaic (8000 – 6000 BCE) and that intensification of obsidian production occurred earlier than previously recognized, at circa 3300 BCE Increased obsidian production appears to have been focused on the acquisition of large (> 20cm) and homogeneous obsidian nodules, although the formal tools produced with obsidian were predominantly small projectile points. It is argued that the acquisition of large, homogenous nodules was prioritized because the production potential of large nodules was highest, and because obsidian was associated with competitive display among early aggrandizers. The timing and economic associations of obsidian production and circulation suggest that the possession of large obsidian pieces in the Titicaca Basin was a demonstration of social connections to distant resources, and to regional trade networks that emerged with regularized camelid caravan transport networks. Obsidian artifacts were not inherently “prestige goods”, rather it is suggested here that obsidian was the least perishable of a number of cultural goods distributed by an expanding network of caravans that linked communities in the region. The acquisition and consumption of these cultural goods was a demonstration of economic connections and cultural influence during the dynamic period of incipient social inequality between the Terminal Archaic (3300–2000 BCE) through the Middle Formative (1300–500 BCE)


Table of Contents

Table of Contents

The pagination below applies to the PDF and the printed version of this document.

With the web version a Table of Contents is shown on the right side of the browser.







Chapter 1 – Like Salt or Like Gold?.. 1

1.1. Overview.... 1

1.1.1. Prehispanic Economy. 7

1.1.2. Chemical characterization work in the Andes. 12

1.2. Structure of the Dissertation... 14

1.2.1. Organization of presentation.. 14

1.2.2. Digital data availability. 14

1.2.3. Dates. 15

1.2.4. Spatial data.. 15

1.2.5. Photographs and scale. 17

1.3. Conclusion... 18

Chapter 2 – Theoretical Approaches to 
Economy, Exchange, and Raw Material Sources.. 19

2.1. Introduction... 19

2.2. Anthropological approaches to economy and exchange.. 21

2.2.1. Economies. 21

2.2.2. Transfer of goods and exchange value. 28

2.2.3. Transfer of goods and socio-political complexity. 46

2.2.4. Definitions of exchange. 60

2.2.5. Exchange and social distance. 69

2.2.6. Territoriality and access to raw material sources. 76

2.2.7. Discussion.. 79

2.3. Chemical provenience and exchange.. 81

2.3.1. Quantitative approaches to regional exchange. 82

2.3.2. Other Distance Decay studies. 86

2.3.3. Site-oriented studies of exchange. 92

2.3.4. Discussion.. 98

2.4. The View from the Quarry... 100

2.4.1. The specialization and efficiency framework for quarry studies. 102

2.4.2. Analysis of a production system.. 105

2.4.3. Specialization at a Mexican Obsidian workshop.. 108

2.4.4. A contextual approach to Neolithic axe quarries in Britain.. 110

2.4.5. Discussion.. 118

Chapter 3 –  The Regional Context of 
Chivay Obsidian Research... 121

3.1. Andean Economy and Exchange.. 126

3.1.1. Economic organization and trade in the Andes. 128

3.1.2. Economy and exchange in mountain environments. 135

3.2. Long distance trade.. 156

3.2.1. Household-level caravans. 159

3.2.2. Incentives for early caravan formation.. 162

3.2.3. Exchange between herders and farmers. 164

3.2.4. Types of products carried by caravans. 166

3.2.5. Caravan travel distances and speeds. 168

3.2.6. Circuit mobility and role of the periphery. 172

3.2.7. Compadrazgo relationships and commerce. 175

3.2.8. Discussion.. 180

3.3. Regional patterns and major sources of obsidian... 183

3.3.1. Obsidian and larger geographical associations. 188

3.4. Chivay Obsidian Consumption Contexts.. 191

3.4.1. Asana.. 191

3.4.2. Qillqatani rock shelter. 193

3.4.3. Sumbay. 200

3.4.4. The Ilave Valley and Jiskairumoko.. 202

3.5. Andean Obsidian Distributions through Time.. 206

3.5.1. Archaic Foragers in the South-Central Andes. 211

3.5.2. Early Agropastoralist obsidian distributions. 231

3.5.3. The Late Prehispanic. 255

3.6. Obsidian Use in the South-Central Andes.. 271

3.6.1. Variability in Andean obsidian use. 272

3.6.2. Symbolic significance of obsidian.. 284

3.7. Models for the Procurement and Circulation of
Chivay Obsidian in Prehistory... 291

3.7.1. Direct acquisition Model 292

3.7.2. Multiple Reciprocal Exchanges (Down the line) Model 296

3.7.3. Independent Caravans Model 300

3.7.4. Elite-Sponsored Caravans Model 305

3.8. Summary... 307

Chapter 4 –  Regional Geography and Geology 
of the Upper Colca Project Area... 309

4.1. The geography of the Colca Valley study area... 310

4.1.1. Climate across the study area.. 313

4.1.2. Lower elevation biotic zones: Study Area Blocks 3 and 6.. 320

4.1.3. High Puna: Block 2 survey and adjacent Blocks 4 and 5.. 327

4.1.4. The Chivay Source: Block 1.. 331

4.2. Tectonic geology... 332

4.3. Formations in the Upper Colca Valley... 335

1.       A geological descent of the upper Colca drainage. 337

4.3.1. Yura and associated sedimentary strata.. 343

4.3.2. Oligocene and Miocene lavas. 344

4.3.3. Pliocene lavas – Barroso group.. 347

4.3.4. Pleistocene – Ampato group.. 350

4.3.5. Holocene stratovolcanoes. 350

4.3.6. Glaciation.. 350

4.4. How obsidian is formed... 355

4.4.1. Chemical differentiation.. 357

4.4.2. Obsidian color. 358

4.5. Pliocene (Barroso group) obsidian in the Colca valley... 359

4.5.1. Chivay obsidian source observations. 360

4.6. Other Tertiary obsidian in the Colca valley... 366

4.6.1. Tripcevich source sampling work.. 369

4.7. Conclusion... 373

Chapter 5 –  Research Methods and 
Data Recording Strategies.. 375

5.1. Introduction... 375

5.1.1. Locus rather than Site-oriented survey methods. 376

5.1.2. Data recording in both categorical forms and field journals. 377

5.2. Geographical context.. 379

5.2.1. Geographical datum and regional data sets. 380

5.2.2. Regional datasets. 382

5.3. Data recording approach... 382

5.3.1. Introduction.. 382

5.3.2. Organization, sampling, and inference. 383

5.3.3. Aggregating by Sites versus Loci 385

5.3.4. Site and Loci recording structure. 386

5.3.5. The Primary Key: ArchID... 387

5.3.6. Site classification.. 389

5.3.7. Linking Field and Lab data: an example. 390

5.3.8. Theoretical relevance of the provenience system.. 391

5.4. Survey Strategy... 393

5.4.1. Goals of Survey and Testing.. 393

5.4.2. Surveys types: Prospection, Statistical, and Spatial Structure. 394

5.4.3. Surveyor interval and sampling.. 395

5.4.4. Survey design.. 396

5.4.5. Testing the effectiveness of the B3 survey strategy. 400

5.5. Mobile GIS for archeological survey... 402

5.5.1. Standard survey practice. 402

5.5.2. The contribution of mobile GIS.. 404

5.5.3. Hardware configuration.. 405

5.5.4. Defining loci and sites. 407

5.5.5. Attribute Forms. 408

5.5.6. Variability within a Locus. 410

5.5.7. Sampling High-density Loci 414

5.5.8. Collection during survey. 415

5.5.9. Other Data Types. 415

5.5.10. Processing steps with mobile GIS.. 416

5.5.11. Implications of Mobile GIS for Fieldwork.. 418

5.6. Test Excavation Methodology... 420

5.6.1. Excavation procedures. 420

5.6.2. Proveniencing of excavated materials. 421

5.6.3. Proveniencing for database management and quantitative analysis. 423

5.7. Lab analysis.. 424

5.7.1. Phase I Lab work.. 425

5.7.2. Phase II Lab work.. 426

5.8. Sampling in Upper Colca work... 431

5.8.1. Sampling during survey. 432

5.8.2. Sampling during excavation.. 432

5.8.3. Sampling during lab analysis. 433

5.9. Derivative indices.. 434

5.9.1. Lithic density raster index. 436

5.9.2. Cumulative Viewshed Analysis and Exposure Index. 439

5.9.3. Bifacial Thinning Flake index. 443

5.10. Database structure and 1:M labeling.. 445

5.10.1. The All ArchID Centroids file. 446

5.10.2. VB Script for One-to-Many labeling.. 448

5.11. Conclusion... 450

Chapter 6 –  Survey Results from 
Research in the Upper Colca... 451

6.1. Introduction to data presentation... 451

6.1.1. Data presentation and cartographic conventions. 453

6.2. Obsidian variability in the study area... 457

6.2.1. Material type by survey block.. 457

6.2.2. Production Indices. 472

6.2.3. Projectile Points and obsidian variability. 475

6.3. Survey Results: Archaic Foragers Period (9000–3300BCE). 481

6.3.1. Block 1 – Archaic Source and adjacent high puna.. 484

6.3.2. Block 2 – Archaic San Bartolomé. 510

6.3.3. Block 3 – Archaic Callalli and adjacent valley bottom areas. 534

6.4. Survey Results: Early Agropastoralists Period (3300BCE – AD400)  578

6.4.1. Block 1 – Source. 584

6.4.2. Block 2 – San Bartolomé. 611

6.4.3. Block 3 – Callalli 623

6.5. Survey Results: Late Prehispanic Period (AD400 - AD1532). 635

6.5.1. Block 1 – Source. 637

6.5.2. Block 2 – San Bartolomé. 651

6.5.3. Block 3 – Callalli 663

6.6. Chapter summary discussion... 677

Chapter 7 –  Results and Analysis of  Data
from Test Excavations
.. 679

7.1. Goals of the analysis of production... 680

7.2. General Indices of Production... 685

7.3. Test excavation units.. 689

7.4. Block 1 Test Excavations at Maymeja Q02-2. 692

7.4.1. Q02-2, Test Unit 2.. 692

7.4.2. Q02-2, Test Unit 3.. 707

7.4.3. Summary Interpretation of A03-126 Workshop.. 748

7.5. Block 2 Test Excavations at Pausa A02-39. 751

7.5.1. A02-39, Test Unit 1 and 2.. 751

7.5.2. A02-39, Test Units 3 and 4.. 752

7.6. Block 3 Test Excavations at Taukamayo A02-26. 762

7.6.1. A02-26, Test Unit 1.. 762

7.6.2. Test Unit 2.. 773

7.7. Comparisons between Blocks with surface and excavated data. 774

7.7.1. Size of flaked stone artifacts. 774

7.8. Conclusion... 776

Chapter 8 –  Major Findings from
the  Upper Colca Project
.. 778

8.1. Introduction... 778

8.2. Review of major findings.. 778

8.2.1. Archaic Foragers (10,000 – 3300 BCE). 780

8.2.2. Early Agropastoralist 782

8.2.3. Late Prehispanic. 786

8.2.4. Discussion.. 790

8.3. Production and interaction... 790

8.3.1. Lithic raw material use in the vicinity of the Chivay source. 790

8.3.2. Quarrying for non-local consumption.. 792

8.3.3. Site visibility at Maymeja and warfare in the Colca area.. 794

8.3.4. Quarry pit and associated workshop activities. 797

8.3.5. Pottery and lithic production.. 805

8.3.6. Projectile Points. 806

8.3.7. Regional caravans and local interaction.. 809

8.3.8. Models of regional obsidian circulation.. 815

8.4. Theoretical implications from obsidian procurement evidence   826

8.4.1. Herder status and herd size. 827

8.4.2. The changing scale of regular interaction.. 828

8.4.3. Aggrandizing behavior. 830

8.5. Future research... 833

8.5.1. Other Andean sources. 833

8.5.2. Technological improvements. 834

8.6. Conclusion... 835

Appendix A –  Arch ID and Lot ID Registry.. 837

A.1. All Arch ID (Centroids) values.. 837

A.2. Lot ID Registry... 865

Appendix B –  Selected Obsidian Sources
in Southern Peru... 870

B.1. Aconcagua Obsidian Source.. 870

B.2. Alca Obsidian Source.. 870

B.3. Chivay Obsidian Source.. 871

B.4. Uyo Uyo Obsidian Source.. 874

Appendix C –  Obsidian Artifact Examples
and  Representations from the Andes.. 875

References Cited –.. 883


List of Figures

List of Figures

Pagination for printed and PDF versions.

Figure 1.1. Larger study region with modern towns and roads.. 3

Figure 1-2. Tobler's (1993) Hiking Function models foot travel velocity as a function of slope. 16

Figure 2-1. Varieties of reciprocal exchange (after Sahlins 1972: 199). 31

Figure 2-2. Modes of exchange from Renfrew (1975:520). 66

Figure 2-3. Network configurations. 68

Figure 2-4. Model for inference about prehistoric exchange (from Torrence 1986: 5). 70

Figure 2-5. Log-Log fall-off curve of obsidian density (Sidrys 1976: 454). 84

Figure 2-6. Stages of production from quarry, local area, and region (after Ericson 1984: 4). 106

Figure 3-1. Chronologies discussed in the text. 124

Figure 3-2. Contemporary types of Andean zonation (Brush 1977: 12). 138

Figure 3-3. Diagram of institutions of Andean complementarity (from Salomon 1985: 520).. 140

Figure 3-4. Subsistence exchange for products by ecozone versus single-source, diffusive goods. 146

Figure 3-5. Known south-central Andean obsidian sources used in prehistory with least cost paths (Tobler’s function on SRTM DEM data) from the three major sources to sites in the region. 184

Figure 3-6. Cumulative frequency graph showing three major Peruvian obsidian sources by consumption site altitude. 187

Figure 3-7. Qillqatani data showing percentage of bifacially flaked tools and percentages of debris made from obsidian per assemblage by count (from Aldenderfer 1999c, in prep.). 195

Figure 3-8. Sumbay pitchstone projectile points. 201

Figure 3-9. Comparison of projectile point counts in the Ilave Valley and the Upper Colca. 204

Figure 3-10. Projectile Point Typology with Titicaca Basin chronology (from data in Klink and Aldenderfer 2005) 210

Figure 3-11. Chivay type obsidian distributions during the “Archaic Foragers” time (circa 10,000 – 3,300 BCE). 214

Figure 3-12. Chivay type obsidian distributions during the “Early Agropastoralists” time (3,300 BCE – AD 400). 232

Figure 3-13. Chivay type obsidian distributions during the “Late Prehispanic” time (AD 400 – 1532). 256

Figure 3-14. Type 5d projectile points from a Terminal Archaic level at Asana and Early Formative levels at Qillqatani. 278

Figure 4-1. View of the volcanic Chivay source area above the town of Chivay in the Colca valley. 310

Figure 4-2. Survey blocks in the Upper Colca study area are shown with modern production zones described in Table 4-1. 312

Figure 4-3. Temperature by Altitude at mid- and high-altitude Arequipa meteorological stations.. 315

Figure 4-4. Comparison of Temperature highs, means, and lows for August and January). 316

Figure 4-5. Mean monthly temperatures (°C) from data in ONERN (1973). 316

Figure 4-6. Precipitation by Altitude (left), precipitation in the study area. 317

Figure 4-7. Latitude against barometric pressure. 319

Figure 4-8. Exterior and interior of probable colonial pyrotechnical structure. 323

Figure 4-9. Tuber cultivation at 4200 masl surrounded by large tuff outcrops. 325

Figure 4-10. ASTER scenes with the terminus of volcanic breccia outcrop. 328

Figure 4-11. View of San Bartolomé (Block 2) area during the dry season t. 330

Figure 4-12. Select raw material sources in the central Andes. 333

Figure 4-13. Rock groups in the Colca region (based on Palacios et al., 1993). 336

Figure 4-14. Legend showing geological map units in maps that follow. 340

Figure 4-15. Geological map units and 2003 Survey Block boundaries (in gray) for the Upper Colca project study area. Legend shown in preceding figure. 340

Figure 4-16. Geological map units shown on ASTER scene from 28 Sept 2000; legend is shown in Figure 4-14. In general, red pixels show areas of photosynthesizing vegetation (bofedales). 342

Figure 4-17. Andesitic Tacaza deposits with breccias and tuff outcrops in the Quebrada de los Molinos drainage. The Chivay obsidian source in later Barroso deposits is found high above, on the right side of the photo. 345

Figure 4-18. The lower section of the Castillo de Callalli is known as “Cabeza de León”. 346

Figure 4-19. Detail of Chivay source, Maymeja area with INGEMMET geological map. 348

Figure 4-20. Glacial polish and striations (aligned towards camera) on lava flows adjacent to Maymeja workshop on the southern end of the Maymeja area. 354

Figure 4-21 (a). Small box in lower-right gully shows Q02-1, an obsidian flow eroding out of ashy-pumaceous soils below western arm of Cerro Ancachita. (b). This obsidian is of limited use for tool making because it contains vertical, subparallel fractures. 361

Figure 4-22. Obsidian gravels exposure in tephra soils east of Cerro Hornillo. 363

Figure 4-23. Geological map units with Uyo Uyo sampling locations. See Figure 4-14 for legend. Selected archaeological sites in the main Colca Valley shown in blue. 367

Figure 4-24. Bivariate plot showing Dysprosium against Manganese for Tripcevich 2005 samples. 370

Figure 4-25. Map showing locations of Colca valley obsidian source samples analyzed by MURR in 2005. 371

Figure 5-1. Criteria in designing regional survey from three stage research proposal including obsidian source survey, testing program, and concluding with the river valley survey. 397

Figure 5-2. GPS tracks from edges of most survey routes. 399

Figure 5-3. An example of a pedestrian survey line following a river terrace at a 15 meter interval.. 403

Figure 5-4. Mobile GIS implementation with ESRI Arcpad 6.. 406

Figure 5-5. (a) Arcpad screen showing a large site with loci and points.. 409

Figure 5-6. Maps for two different hypothetical sites recorded in less than one hour.. 411

Figure 5-7. (a) Structure of the archaeological Shapefiles with names and descriptions.. 412

Figure 5-8. (a) Example of proveniencing for four 50cm quads within a 1x1m unit. 422

Figure 5-9. Showing some of the percussive metrics, platform metrics, and measures of retouch invasiveness for ventral side (after Clarkson 2002) used in the Phase II analysis.. 429

Figure 5-10. Line of sight across hilly terrain results in specific cells and targets being in view or out of view. 440

Figure 5-11. Viewshed is not necessarily reciprocal, as the individual and the left can see the person on the right, but the opposite is not true. 441

Figure 5-12. Cumulative Viewshed using 5000 random observers and 10 km viewing distance. 442

Figure 5-13. The All_ArchID_Centroids table provides a single reference layer for all the ArchID numbers used in the projecta. 447

Figure 6-1. Artifactual lithic material types in the Upper Colca Project study region. 458

Figure 6-2. Proportion of obsidian for four colors (shades) of glass, by count. 462

Figure 6-3. Proportion of obsidian material as Ob1 and Ob2 (heterogeneities), by count. 464

Figure 6-4. Photographic comparison of the homogeneous Ob1 obsidian and the Ob2 obsidian with heterogeneities. 466

Figure 6-5. Complete projectile point weights and lengths by material type for the entire project area. Series 5 projectile points are excluded and chalcedony is combined with chert. 480

Figure 6-6. Projectile Point weights and lengths (when not broken) by material type for Block 1 and adjacent high puna areas of Blocks 4 and 5. Series 5 projectile point types excluded. 484

Figure 6-7. Possible Chivay obsidian source camps during Archaic Forager times. 486

Figure 6-8. Middle Archaic obsidian projectile point from Maymeja area [A03-184]. 489

Figure 6-9. (a) Small rock shelter at “Molinos 2” [A03-580] is filled with debris from heavy runoff. (b) A density of flakes, predominantly of obsidian, are found in profile in the flood channel. 494

Figure 6-10. Lithic scatters are associated with shelter provided by large boulders located along moraines [A03-539]. One surveyor that is visible in blue provides scale. 496

Figure 6-11. Obsidian scatters were found along the base of these viscous lava flows at 5040 masl [A03-291]. 497

Figure 6-12. Map of project area showing Reconnaissance Blocks 4 and 5 with diagnostic projectile points and logistical sites from the Archaic Foragers period. 501

Figure 6-13. Rock shelter [A03-229] passes behind the collapsed margin of the grey lava flow in the center of the photo. Walls are built to partially close off both of the two entrances. 502

Figure 6-14. Multicomponent site of A03-910 "Collpa" among crystalline tuff outcrops.. 506

Figure 6-15. Photo of moraines and bofedales of Escalera south-west of Chivay source. 509

Figure 6-16. Projectile Point weights and lengths (when complete) by material type for Block 2. 512

Figure 6-17. Archaic Forager sites in Block 2: Major sites described in the text. 514

Figure 6-18. A03-1050 consists of four rock shelters: A, B, C, and D. 517

Figure 6-19. Chiripascapa [A03-1014], Archaic Foragers occupation. 518

Figure 6-20. Huañatira [A03-900] and vicinity, Archaic Foragers occupation. 524

Figure 6-21. Rock shelter of Huañatira with circular mortuary feature visible inside. One meter scale showing on tape resting on rock along dripline (see also Figure 6-76). 527

Figure 6-22. Block 2 Archaic Foragers component from lithic evidence. 530

Figure 6-23. Projectile Point weights and lengths (when complete) by material type. 536

Figure 6-24. Archaic Period in Blocks 3, 6 and 5 (valley): Sites described in the text. 538

Figure 6-25. Block 3: Early, Middle, and Late Archaic projectile points and Aceramic sites. 539

Figure 6-26. Site of Kakapunku [A03-1000]. 541

Figure 6-27. The entrance of A03-1001 with looted cist tomb visible inside the rock shelter. 542

Figure 6-28. Cueva de Quelkata was dynamited by Majes Project road crews in 1976. Two people are visible for scale inside the shelter to the left of center. 545

Figure 6-29. Quelkata in 1975 redrawn from Chavez (1978: 20). Red dotted line shows estimated border of the portion of the rock shelter that remains after the dynamiting for road widening. 545

Figure 6-30. Quickbird satellite image of Quelkata area in 2005. The shelter is under the tuff outcrops immediately to the left of the bridge. Data courtesy of Google / Digital Globe. 547

Figure 6-31. A 65cm column of cultural deposits still exists at the back of Quelkata at a height of approximately 1.8m above datum (at head height between the gravels and the tuff in photo). 549

Figure 6-32. Profile of extant deposits at Quelkata shown in terms of relative collection units. 551

Figure 6-33. Cueva de Mollepunco (A02-3). 556

Figure 6-34. Petroglyph of camelid on wall of Mollepunco. 556

Figure 6-35. View westward from “Callalli 11” across quebrada shows the eastern edge of Callalli. 561

Figure 6-36. Callalli 11 [A03-599] on terrace on the east edge of the town of Callalli. 562

Figure 6-37. Chalcedony projectile point [A03-790] from Anccasuyo [A03-785]. 563

Figure 6-38. Map of the Sullullumba [A03-806], an Archaic Foragers site showing lithic loci on terrace edge 40m above the annual flood line. Image courtsey of Quickbird DG / Google. 567

Figure 6-39. Sullullumba [A03-806] along terrace edge above the upper Río Colca. Yellow notebook and scale bar are visible in the center-rear of photo. 568

Figure 6-40. Projectile points from Sullullumba [A03-806]. 570

Figure 6-41. The high density lithic locus A03-808 consisted of aphanytic volcanics and chert. 570

Figure 6-42. Block 6 Challacone - Ichocollo overview map. 573

Figure 6-43. Ichocollo complex along Quebrada Ichocollo. 574

Figure 6-44. Photo looking northwest across Ichocollo creek. 576

Figure 6-45. Testing at Quarry Pit Q02-2 with 1x1 test unit Q02-2u3. Snow remains in the pit. 587

Figure 6-46. A03-269 and A03-734 "Camino Hornillo" and modern trail system. 588

Figure 6-47. Apacheta (cairn) close to junction of the Escalera route with Cerro Hornillo. 589

Figure 6-48. Camino Hornillo showing (a) A03-268 and (b) A03-734 segments. Photos were digitally modified to highlight the route in red. 591

Figure 6-49. Block 1 possible Early Agropastoral settlement pattern with Series 5 projectile points. 594

Figure 6-50. A03-126 "Maymeja 1" workshop and vicinity. 596

Figure 6-51. View of A03-126 "Maymeja" from north. Terraced area A03-334 on upper level. Test Unit Q02-02 is just right of the orange bucket. Project tents are visible in corral A03-127. 597

Figure 6-52. (a) Workshop area of "Maymeja 1" showing proximity of bofedal, (b) Testing Q02-2U3, with the quarry pit [Q02-2] visible among light ash 600m uphill in the background. 598

Figure 6-53. Base of structure [A03-335] is formed by fifteen large, partially buried stones and measures 2.5m in diameter. 601

Figure 6-54. A03-126, 209, 275 Maymeja workshop and vicinity. 604

Figure 6-55. View looking south from above at A03-201 "Saylluta" with excavation of test unit Q02-2u1 under way in top center of corral by the orange bucket. Site datum is on top of the large boulder to the south of Q02-2u1 (just above test unit in the photo). 605

Figure 6-56. Blocks 4 and 5 showing Series 5 projectile point distribution. 609

Figure 6-57. Block 2 - Early Agropastoral occupation. 612

Figure 6-58. Pausa [A02-39] showing raised oval structures, lithic concentrations, large rock forming wall bases, and test unit locations. Site mapped with Topcon total station and dGPS. 618

Figure 6-59. Circular structure A03-557 extends from 1m behind the tape to just below the largest rocks at the back of the photo. 620

Figure 6-60. Testing u3 and u4 on north edge of structures A03-558 and A03-559. This photo is taken from above, from the base of the lava flow. 620

Figure 6-61. Early Agropastoralist settlement in Block 3. 624

Figure 6-62. Taukamayo [A02-26], a multicomponent site partially destroyed by a landslide. 627

Figure 6-63. Overview of Taukamayo [A02-26] on slump along base of hillside. Grey box shows area detailed in Figure 6-64, below. 628

Figure 6-64. Taukamayo [A02-26] detail showing two test units locations on cutbank margins of the creep area. Excavators are visible on right-side at A02-26u1 and provide scale for photo. 629

Figure 6-65. Non-local incised and stamped pottery from Taukamayo [A02-26], in the 2003 provenience it is A03-679. and A03-679.2. 631

Figure 6-66. Sixteen large andesite hoes were found at Taukamayo [A02-26]. 634

Figure 6-67. Block 1 Late Prehispanic features. 639

Figure 6-68. Three levels of terracing in [A03-275] below glacier polished rhyolite flow. Yellow tape shows 1m. 641

Figure 6-69. (a) Cutstone masonry [A03-339] from site A03-275 close to the workshop area at the Chivay obsidian source. Yellow tape shows 50cm. (b) Rim sherds from an 18cm diameter Inka-Collagua plate were found adjacent to this corner. 642

Figure 6-70. Blocks 1 and 4 overview showing possible route of Late Prehispanic canal. 647

Figure 6-71. Blocks 4 and 5 Reconnaissance - Late Prehispanic component. 649

Figure 6-72. Non-local sherd with geometric elements akin to Middle Horizon or LIP styles [A03-1056.1]. 653

Figure 6-73. Block 2 diagnostic artifacts from the Middle Horizon and Late Intermediate Period. 654

Figure 6-74. Huañatira [A03-855]. Corrals, structures, and artifacts scatters wrap around the base of the lava escarpment. A small figure is visible on the right edge for scale. 656

Figure 6-75. Huañatira A03-817 and A03-855 multicomponent site showing corral structures. 658

Figure 6-76. Cist tomb with mortared stonework. 659

Figure 6-77. Block 2 diagnostic artifacts from the Late Horizon. 662

Figure 6-78. Block 3 Middle Horizon and Late Intermediate Period features. 665

Figure 6-79. Recently looted cocoon-type interment from A03-815. 667

Figure 6-80. Block 3 Late Horizon diagnostic materials. 670

Figure 6-81. Callalli Antiguo [A03-662] and surroundings. 672

Figure 6-82. A03-662 north sector of Callalli Antiguo agricultural sector. 673

Figure 6-83. Base of walls of structure A03-663 in Callalli Antiguo. 675

Figure 6-84. Collapsed wall of A03-663 structure. 675

Figure 6-85. Cores at Callalli Antiguo [A03-662] and vicinity compared with all of Block 3. 676

Figure 7-1. Principal lithic reduction in the northern African Gumu-sana assemblage (from B. A. Bradley 1975: Fig. 1). 683

Figure 7-2. Bifacial sequences diverge based on the original nodule form with arrows showing percussion direction (Pastrana and Hirth 2003: 204-205). 684

Figure 7-3. Radiocarbon dates on charcoal from test excavations in 2003 showing uncalibrated and calibrated dates as BCE from OxCal v3.9 (Ramsey 2003). 691

Figure 7-4 Map of quarry pit showing topographic surface acquired using a total station. 693

Figure 7-5. (a) Perspective view of quarry pit, (b) Perspective view bisected with an inferred natural slope. 694

Figure 7-6. The quarry [Q02-2] and workshop [A03-126] are 600m apart. In this image snow blankets most south-facing slopes but the quarry pit debris pile is visible protruding through the snow mantle. 695

Figure 7-7. Diagram of Q02-2 quarry pit and position of the Q02-2u2 test unit 698

Figure 7-8. Q02-2, view of quarry pit and u2 test unit from the north. 699

Figure 7-9. Q02-2u2 quarry pit south and west profile diagrams showing excavation levels. 700

Figure 7-10. South profile of Q02-2u2 at quarry pit at the top of levels 11/12. 700

Figure 7-11. Location of test unit Q02-2u3 in site A03-126 workshop. 710

Figure 7-12. Q02-2u3 workshop test unit west and north profile diagrams showing strata. 712

Figure 7-13. Q02-2u3 west profile at Maymeja workshop site [A03-126]. 712

Figure 7-14. Proportions of Tech Classes in excavation levels by count in Q02-2u3. 714

Figure 7-15. Retouched artifacts by Excavation Level. 719

Figure 7-16. Q02-2-u3 Cores by Level showing changes length and weight. 721

Figure 7-17. Canonical discriminant chart for core clusters from Q02-2u3. 724

Figure 7-18. Boxplots showing three clusters of Q02-2u3 cores with size measures. The tabular shape of C2 cores is apparent in their thinness relative to their length and width. 726

Figure 7-19. An example of a C1 core [L03-162.303], arrows indicate percussion. 727

Figure 7-20. An example of a C2 core [L03-162.305]. 728

Figure 7-21. An example of a C3 core [L03-162.280] showing a small area of thin cortex. 728

Figure 7-22. Clustered Cores from Q02-2u3 workshop test unit showing counts by level. 729

Figure 7-23. Q02-u3: Graph showing means of complete cores and cortical flakes. 731

Figure 7-24. Q02-2u3: Graph showing means of measures on complete obsidian flakes. 734

Figure 7-25. Canonical discriminant chart for complete flake clusters from Q02-2u3. 738

Figure 7-26. An example of a flake from the F1 cluster [L03-162.21]. 740

Figure 7-27. An example of a flake from the F2 cluster [L03-162.118]. 740

Figure 7-28. An example of a flake from the F3 cluster [L03-162.66]. 740

Figure 7-29. Graph showing Flake clusters by excavation level. 742

Figure 7-30. A02-39u3 (upper unit in photo) and u4 (lower) at basal levels. Top of photo is east. 752

Figure 7-31. Circular hearth (F1) in A02-39u4, level 5. Top of photo is north, unit is 1m on a side. 753

Figure 7-32. A02-39 "Pausa" – north, with Test Units 3 and 4, ovals, and inferred features. Compare with larger map showing ovals in Figure 6-58. 754

Figure 7-33. A02-39u4 Pausa test unit west profile diagram showing strata and levels. 757

Figure 7-34. A02-39u3/u4 Bar graph showing counts of lithics classes by excavation level. 758

Figure 7-35. Cumulative Frequency of A02-39 u3 & u4 showing Level 5 reduction against reduction for all levels. 760

Figure 7-36. Test unit A02-26u1 is a 1x1m test unit placed on the edge of a creeping landslide with an irregular extension area (1x) on the south edge of the unit. Top of the photo is north. 763

Figure 7-37. A02-26u1 Taukamayo test unit, north profile diagram showing strata and levels. 765

Figure 7-38. A02-26u1 Technical Class by Excavation Level. 768

Figure 7-39. A02-26u1 Material Types by Excavation Level. 768

Figure 7-40. Obsidian at Taukamayo A02-26u1, Ob1 and Ob2 compared. 769

Figure 7-41. Cumulative frequency comparing obsidian and non-obsidian flakes. 771

Figure 7-42. Flake metrics from Quarry area (Block 1) as compared with local consumption areas 774

Figure 8-1. View westward from obsidian production area A03-210 towards main Colca valley. 795

Figure B?1. Nodule from the Aconcagua source. Photo courtesy of Mark Aldenderfer. 870

Figure B?2. Nodules of Alca obsidian with brownish tint. 871

Figure B?3. Nodule of Chivay obsidian from Maymeja weighing 1750g and 23.3 cm long. 872

Figure B?4. (a, L03-161.341) nodule from quarry pit area. 873

Figure B?5. Nodule of Uyo Uyo obsidian with irregularities perhaps caused by air bubbles in the magma. 874

Figure C-1. Obsidian projectile points from the Upper Colca survey project.. 876

Figure C-2. Obsidian point (Cat. 609) from Terminal Archaic context at Jiskairumoko. 876

Figure C-3. Bifacially flaked obsidian “Wari style” points or knives from surface contexts. Photos courtesy of Bruce Owen. 877

Figure C-4 Rollout of Tiwanaku q'ero showing bows and black-tipped arrows (top) near the rim of vessel (Posnansky 1957: XXa). 878

Figure C-5. Early Nasca ritual obsidian knife hafted to painted dolphin palate (Disselhoff 1972: 277). 879

Figure C-6. Representations of black-tipped projectiles are diagnostic to Nasca B1 and B2 ceramics (Carmichael et al. 1998: 151). 880

Figure C-7. Paracas textile with figure holding black-tipped projectiles (Lavalle and Lang 1983: 95) 880

Figure C-8. Wooden hand mirror with obsidian inlay, Wari (Lavalle and Perú 1990: 185). 881

Figure C-9. Obsidian point hafted on harpoon found in a Paracas necropolis, Ica, Peru (Engel 1966: 180c). 881

Figure C-10. An obsidian point embedded in a human lumbar vertebra (from Ravines 1967: 230). 882

Figure C-11. Obsidian Point penetrating through arm muscle near left humerus found at Carhua in Ica, Peru (Engel 1966: 212). 882

List of Tables

List of Tables

Pagination from printed and PDF versions.

Table 2-1. Three categories of exchange goods. 39

Table 2-2. Economic processes in political evolution. 47

Table 2-3. Characteristics of reciprocity and redistribution (from Renfrew 1975: 8). 65

Table 2-4. Household composition of raw materials should vary with different types of exchange.. 95

Table 2-5. Measurement indices for procurement system (after Ericson 1984: 4). 105

Table 3-1. Reported llama caravan loads, distances, and times. 166

Table 3-2. Three major Peruvian obsidian sources showing average and maximum distances and times. 183

Table 3-3. Obsidian in the north and south Titicaca Basin by counts and percents. 186

Table 3-4. Asana obsidian samples, collections, and associated 14C samples by level (from Aldenderfer 1998b: 131, 157, 163, 209, 268; Frye et al. 1998). 189

Table 3-5. Qillqatani excavation levels, radiocarbon dates, and obsidian samples. 191

Table 3-6. Qillqatani periods by tools (bifacially flaked) and debitage (all other lithics), in obsidian and non-obsidian categories. 193

Table 3-7. Counts of obsidian debitage at Qillqatani by weight (g). 194

Table 3-8. Sumbay, SU-3 Pit 5. Strata with obsidian samples and associated artifacts (Máximo Neira Avendaño 1990: 32-33). 199

Table 3-9. Temporal organization of data. 205

Table 3-10. Examples of obsidian use in the south-central Andes (part 1). 272

Table 3-11. Examples of obsidian use in the south-central Andes (part 2). 273

Table 3-12. Models of procurement and exchange for Chivay obsidian. Compare terms with those used in Figure 2-2 and Figure 3-3. 290

Table 4-1. Andean ecological zones with approximate local elevation values for each zone. 308

Table 4-2. Equatorial bulge and effects on barometric pressure (in torr) at 15° and 60° latitude. From data in (J. B. West 1996: 1851). Equivalent altitude column shows altitude at 60° latitude with equivalent pressure to the value shown at 15° latitude. 317

Table 4-3. Coordinates and names of select raw material sources in the central Andes. Coordinates in WGS1984 datum. 332

Table 4-4. Characteristics of Obsidian. 355

Table 4-5. Peruvian obsidian source samples submitted to MURR by Tripcevich in 2002 and 2005. Coordinate datum is WGS84. 370

Table 5-1.The two reference ellipsoids used in Peruvian and Bolivian cartography (NIMA 1977). 379

Table 5-2. Three parameter cartographic transformations for UTM coordinates from PSAD 1956 (La Canoa) to WGS 1984 (Dana 1998; Mugnier 2001; 2006: 496; NIMA 1977). 379

Table 5-3. Types of archaeological survey described by Banning (2002: 27-38). 392

Table 5-4. Digital data sources used in developing the survey strategy. 394

Table 5-5. Sites and isolates from 100% survey strip that would not have been encountered using the regular Block 3 survey strategy. 399

Table 5-6. Locus and Site artifact density definitions. 405

Table 5-7. Sources of inconsistent data during the 2003 project. 414

Table 5-8. Measures on flaked stone artifacts during Phase II lithics analysis. 425

Table 5-9. Proportion of analysis by projectile point typological group. 427

Table 5-10. Vector and raster layers used in analysis and derived raster output. 433

Table 5-11. Loci to GRID conversion values. 436

Table 5-12. Script for labeling through a One-To-Many relate in ArcMap 9.1. 447

Table 6-1. Abbreviations for lithics used in maps, figures, and tables. 451

Table 6-2. Example of a map abbreviation label for a diagnostic lithic. 452

Table 6-3. Abbreviations for ceramics used in maps, figures, and tables. 453

Table 6-4. Example and explanation of a map label for a diagnostic ceramic. 453

Table 6-5. Counts of artifactual lithics material types throughout the study region. 456

Table 6-6. Lengths of complete obsidian artifacts with > 30% cortex. 458

Table 6-7. Obsidian artifact color (shade) by survey block surface collection. 460

Table 6-8. Obsidian artifact material type by Survey Block surface collections. 462

Table 6-9. Projectile Points made from obsidian containing heterogeneities (Ob2). 465

Table 6-10. Ratio of Obsidian Projectile Points with heterogeneities. 465

Table 6-11. Obsidian: mean sizes of complete Ob1 and Ob2 artifacts, by Survey Block. 467

Table 6-12. Obsidian production system indices for surface survey. 472

Table 6-13. ArchID numbers for diagnostic projectile points, Series 1-4 only. 474

Table 6-14. Projectile point mean weights by point type and material type. 475

Table 6-15. All obsidian projectile points by survey block. 476

Table 6-16. Aggregated Projectile Point Styles by Material Type for the project area. 477

Table 6-17. Classifications for Archaic components of sites. 481

Table 6-18. Diagnostic Projectile points Series 1-4 from Blocks 1, 4 and high altitude areas of Block 5, identifed by ArchID number. 488

Table 6-19. All Diagnostic Projectile Points from Blocks 1, 4, and Block 5 upper puna. Weights included for unbroken points only. 489

Table 6-20. Environmental characteristics of Archaic Foragers logistical camps in Block 1. 490

Table 6-21. Lithic artifacts from site A03-229 excluding eleven Series 5 projectile points. 501

Table 6-22. Lithic artifacts from site A03-777 excluding 1 Series 5 projectile point. 502

Table 6-23. Lithic artifacts from site A03-910 excluding five Series 5 Ob1 projectile points. 505

Table 6-24. Complete Cores at Collpa [A03-910]. 506

Table 6-25. Diagnostic Projectile points Series 1-4 from Block 2 identifed by ArchID number. 509

Table 6-26. Loci in Chiripascapa. 513

Table 6-27. Diagnostic Series 1 through 4 projectile points from Chiripascapa [A03-1014]. 519

Table 6-28. Representative proportions of material types by projectile point styles. 519

Table 6-29. All lithic artifacts from Chiripascapa. 520

Table 6-30. Non-consecutive ArchID numbers at Huañatira [A03-900], an Archaic Foragers site. 524

Table 6-31. Diagnostic Series 1 through 4 projectile points from Huañatira [A03-900]. 526

Table 6-32. All lithic artifacts from Huañatira. 526

Table 6-33. Environmental characteristics of selected Archaic Foragers sites in Block 2. 530

Table 6-34. Diagnostic Projectile points Series 1-4 from Blocks 3, 6 and the valley portion of Block 5 identifed by ArchID number. 532

Table 6-35. Loci in Kakapunku [A03-1000]. 539

Table 6-36. Rock shelters at Kakapunku [A3-1000], dimensions in meters. 539

Table 6-37. Selected Lithics from Kakapunku [A03-1000]. 541

Table 6-38. Counts of Material types by Artifact Form from 1977 Quelkata surface collections. 551

Table 6-39. Loci in Pokallacta [A03-1074]. 555

Table 6-40. Rock shelters at Pokallacta [A3-1074], dimensions in meters. 555

Table 6-41. Selected Lithics from Pokallacta [A03-1074]. 556

Table 6-42. Environmental characteristics of aceramic sites. 557

Table 6-43. Loci at Sullullumba [A03-806]. 566

Table 6-44. Site sizes in Ichocollo [A03-32]. 569

Table 6-45. Bifacially Flaked Lithics from Ichocollo by Material Type. 574

Table 6-46. Bifacial Lithics at Ichocollo showing Counts of Material Types by Length. 575

Table 6-47. Classifications for Early Agropastoralist components of sites. 579

Table 6-48. Areal features belonging to A03-126 and A03-275 workshop complex. 597

Table 6-49. Material types for Series 5 points in Blocks 4 and 5. 608

Table 6-50. Dimension of structural features at Pausa [A02-39]. 617

Table 6-51. Counts of lithics from surface collection at Pausa [A02-39]. 619

Table 6-52. All ceramics from Taukamayo [A02-26] and vicinity. 628

Table 6-53. Cores from the surface of Taukamayo. 630

Table 6-54. Counts of lithics from surface collection at Taukamayo [A02-26]. 630

Table 6-55. Ob1 and Ob2 obsidian use at A03-738 "Lecceta 1" compared with entire project. 648

Table 6-56. Diagnostic sherds from Huañatira 2 [A03-855]. 653

Table 6-57. Diagnostic ceramics from Callalli Antiguo [A03-662] and vicinity. 670

Table 7-1. Lithics Phase II analysis showing fraction of artifacts with detail measures. 677

Table 7-2. Q02u2 “Quarry Pit” obsidian production system indices. 681

Table 7-3. Q02u3 “Maymeja Workshop” obsidian production system indices. 682

Table 7-4. A02-26u1 “Taukamayo” obsidian production system indices. 683

Table 7-5. A02-39 u3 and u4 “Pausa” obsidian production system indices. 684

Table 7-6. 1x1m test excavation unit proveniences from 2003 fieldwork. 685

Table 7-7. Q02-2u2 Excavation levels from test unit at quarry pit.. 697

Table 7-8. Q02-2u2 lithic technical class by test unit excavation level. 698

Table 7-9. Q02-2u2 quarry pit obsidian artifact material type and color by level. 699

Table 7-10. Q02-2u2 Color in quarry pit artifacts compared with all points from fieldwork. 700

Table 7-11. Q02-2u2 quarry pit: Lengths of flakes and cores with >20% cortex. 701

Table 7-12. Calculated depth of strata in Q02-2u2. 702

Table 7-13. Q02-2u3 Excavation levels from test unit at the A03-126 workshop. 708

Table 7-14. Q02-2u3: c2 showing Technical Class by Level with the Standard Residual. 709

Table 7-15. Attributes of obsidian artifacts from Q02-2u3. 711

Table 7-16. Proportion of Kombewa flakes by level in Q02-2u3. 711

Table 7-17. Cores at Q02-2-u3 by excavation level showing countsof Ob1 and Ob2. 715

Table 7-18. Canonical discriminant function for Q02-2u3 Cores. 718

Table 7-19. Q02-2-u3 characteristics of three clusters from core attributes. 721

Table 7-20. Q02-2u3: ANOVA on complete cores and flakes with at least 50% dorsal cortex. 727

Table 7-21. Q02-2u3: ANOVA on measures from complete flakes. 730

Table 7-22. Canonical Discriminant Function Structure Matrix in order of size of correlation. 733

Table 7-23. Q02-2-u3: characteristics of 3 clusters of flakes based on numerical attributes. 734

Table 7-24. Q02-2u3: c2 showing types of terminations by flake cluster. 736

Table 7-25. Q02-2u3: Obsidian type and color of artifacts by level. 742

Table 7-26. Q02-2u3: Length of complete flakes and cores with ?50% cortex. 742

Table 7-27. A02-39u3 test unit. Excavation levels, overview, and soil description. 751

Table 7-28. A02-39u4 test unit. Excavation levels, overview, and soil description. 751

Table 7-29. A03-39u3/u4 counts of lithics by level. 753

Table 7-30. A02-39u3 and u4: Obsidian flake types by excavation level. 754

Table 7-31. A02-26u1 test unit. Excavation levels, overview, and soil description. 759

Table 7-32. A02-26u1: Artifact counts and averages for select measures. 762

Table 7-33. A02-26u1: Lithic Technical Classes by Level showing counts. 762

Table 7-34. A02-26u1 Excavation and Block 3 surface: Lithic Tech. Classes by Obsidian type. 767

Table 7-35. Counts by length for all complete obsidian flakes with ?20% dorsal cortex. 770

Table 8-1. Visibility index values of high visibility production locations. 791

Ch1 Like Salt or Like Gold?

1.1. Overview

Patterns in Obsidian Circulation in the South-Central Andes

Archaeologists have long recognized the central importance of interregional relationships in much of Andean prehistory, but consistent forms of evidence that can be used to gauge long-distance interaction are relatively limited. Over the past century, systematic studies on a regional scale have provided evidence of interaction based on stylistic attributes, such as the presence of non-local ceramics, design elements, and architectural styles. More recently, geochemical evidence has taken a more prominent role in documenting long-distance interaction and exchange because evidence from a variety of chemically unique materials - such as obsidian - have been accumulating in databases as chemical identification techniques are refined. This project aims to apply geochemical evidence of long-distance interaction from an obsidian source in the south-central Andes to an interpretation of regional developments.

Obsidian is among the least-perishable of a number of materials that were widely transported in the prehispanic Andes. On one level, obsidian should be understood as just another one of the widely circulated items, such as nuts, gourds, wood, shell, metal, coca, sebil, and basalt, in a list that is immeasurably long. As a regionally distributed material, however, obsidian has a number of characteristics that are distinctive from some of other goods that circulated in local and regional trade. As a lithic raw material, obsidian is a reductive technology and therefore distance decay effects are prominent in its regional distribution. Obsidian implements can have extremely sharp edges, and by far the most common formal obsidian tools produced in the south-central Andes were the projectile points. Obsidian artifacts are sometimes found in contexts - such as monuments or burials, that suggest that obsidian had ritual importance - however, throughout the region obsidian artifacts are frequently encountered in domestic contexts and middens with little compelling evidence of ritual or symbolic importance. A central concern of this research project is to avoid the simplistic, dichotomous archaeological construction of a "utilitarian" artifact group versus a "prestige" and/or "ritual" associated artifact group. As a highly visible material that was often made into projectile tips, "utilitarian" obsidian tipped implements themselves may have had a role in social influence or coercion, and were perhaps used in status competition displays. Obsidian made into a sharp weapon simultaneously represents access to non-local goods and, in the case of non-local obsidian, may reflect regional scale alliances. Thus, the circulation of obsidian may have been akin to that of utilitarian products, a such as salt, or along the lines of a precious commodity like gold or turquoise where value is related to production costs and reflects scarcity or exclusivity.


Figure 1-1. Larger study region with modern towns and roads. General obsidian distribution areas are depicted, showing largely a Formative Period extent.

Archaeologists identify these materials as non-local when they are encountered in archaeological sites far from their source areas, and in these contexts archaeologists interpret non-local materials are seen as providing evidence of the long distance movement of goods. The dramatic altitudinal zonation on the flanks of the Andes contributes to the archaeological identification of non-local materials because a number of these products are zone-specific. The major environmental and geological zones with distinctive products include the Amazonian lowlands to the east, the mineral rich western Cordillera, and the pacific littoral.

This project focuses on the source region of an obsidian type that was widely circulated in the prehispanic Andes. This chemical group, first referenced in the 1970s as the "Titicaca Basin Type" (Burger and Asaro 1977), was finally traced to its geological source in the Colca valley of Arequipa, Peru in the 1990s. In the Colca, the geochemical type was documented under two names, the Chivay Source (Burger, et al. 1998) and the Cotallalli Source (Brooks, et al. 1997). The source lies in volcanic terrain that is relatively difficult to access by the modern road network, a situation which has hampered research in this zone but has also resulted in relatively little disturbance to archaeological sites in the obsidian source area. This project sought to comprehensively research this zone in order to address a number of long-standing theoretical questions concerning the human use of obsidian in the larger region. Obsidian from the Chivay source is found in a variety of archaeological contexts in the region ranging from common residential middens to ceremonial structures and burials. Unlike other regions of the world where a variety of artifact forms are produced from obsidian, in the south-central Andes the formal implements made from obsidian are largely restricted to projectile points. Simple obsidian flakes are seen as valuable as well, as the shearing and butchering requirements of dedicated pastoralists are significant, and these sharp obsidian flakes filled a functional category akin to razor blades in modern life.

On a theoretical level, this research can contribute to anthropological models concerning the emergence of social hierarchy, the influence of regularized caravan-based trade, and the role of exchange in non-local goods that were widely-circulated in prehistory. Obsidian escapes easy classification as a "prestige item" or an "ordinary good" because the cultural meaning of the material appears to have been highly variable. Obsidian is typically abundant at archaeological sites close a geological source and scarce in areas more distant from geological sources, and obsidian flakes are plentiful at pastoralist rock shelter sites in the altiplano that appear to be relatively lacking in other material goods. Together with coca leaf and other widely circulated goods of cultural value in the Andes, obsidian appears at times in ritual contexts; for example, obsidian may be used as a knife used in a llama sacrifice, as flakes placed in a burial, or as projectile point tips represented on Tiwanaku and Nasca ceramics. However, obsidian appears to not have been a "preciosity" in the sense that large and recognizable quantities of labor were necessarily invested in the production of the material, and indeed obsidian is often found in commonplace, utilitarian contexts. As a household level enterprise, the acquisition of obsidian may constitute evidence of regional interaction which contributes to an archaeological understanding of the prehistoric emergence of regularized exchange over distance, perhaps scheduled around festivals, seasonal harvests, and other regional scale events that integrated distant communities across the Andean highlands. In sum, while it is difficult to place obsidian on a scale from ordinary to prestige good, it appears that the mere possession obsidian may have had some value as a symbol of participation in the long-distance exchange of a variety of goods of cultural value.

This project investigates the economy of Chivay obsidian in prehistory at three levels. First, fieldwork was conducted at the Chivay source in order to examine production at the obsidian source area itself. This work documented the geological exposures of the material, the archaeological sites in association with the source area, and the physical geography and trail system around the source. Second, the project area included obsidian consumption zones within a day's walk of the source in two adjacent residential areas in order to evaluate the local use of obsidian within the larger archaeological context of the Upper Colca. Third, the regional consumption patterns of Chivay obsidian in the south-central Andes were explored using existing studies from the published sources for a more complete picture of production, distribution, and consumption of the material through prehistory. This dissertation represents the integration of these three perspectives on Chivay obsidian in light of anthropological theories of exchange and Andean culture history.

Archaeological evidence for interpreting obsidian production at Chivay takes the form of differences in diagnostic materials and artifact morphology across space from a surface survey, and from measurable change in stratified deposits from test excavations. Four models of procurement and circulation are proposed here, and the material correlates for these models are evaluated based on expectations and measurable changes in artifacts that primarily consist of flakes and cores. The information potential from surface scatters is limited by the non-diagnostic nature of most artifacts at raw material quarries. Nevertheless, this research has found that a strong correlation exists between obsidian intensification and early pastoralism, and perhaps camelid caravan transport; and the research contributes to answering long-standing questions about the relationship between the Colca valley and prehispanic Titicaca Basin polities. This study explores obsidian production and circulation as a unifying theme that can be studied at a regional scale in order to examine long-term processes that led to social changes manifested in the emergence of chiefdoms and states in the Andes.

1.1.1. Prehispanic Economy

As an investigation of non-western exchange, this research relies on categories and approaches developed in economic anthropology. The analysis and interpretation applied in the Upper Colca project is rooted in the substantivist tradition, but the weaker elements of the original substantivist approach will be avoided. These weaker elements include an attempt at an explicit delimiting of culture areas (where values are assumed to be shared), and implicit evolutionary links between volume of exchange and social complexity. Concepts from economic anthropology based on the work of Polanyi (1957) that include the idea of social distance as developed by Sahlins (1972) will be used to explore diachronic change in the circulation of obsidian in the prehispanic Andes. More recent discussions of the importance of "ordinary goods" (Smith 1999), in contrast to the usual focus on prestige items, will be considered in light of the cultural importance of coca, herbs, obsidian, and other non-exclusive but widely valued items in the Andes.

A principal theoretical question addressed in this dissertation concerns the circumstances of a large shift in obsidian production during the Early Formative, around 2000-1300 BCE It is argued here that the links between prestige goods circulation with models of competitive display are evident, but that influence is also accrued by individuals, such as llama caravan drivers, who generate consistent traffic between areas by circulating both information and non-local goods on a seasonal or an annual basis. The underlying assumption is not that the act of exchange inherently "creates value" (Appadurai 1986), but rather that the circulation of material goods was a pragmatic incentive for maintaining a range of important social and ideological relationships across space (Browman 1981;Nuñez and Dillehay 1995 [1979]). The caravan model described here is not adaptationalist, however, as it is not proposed that greater efficiency or social harmony led to the emergence of social complexity in the Andean highlands. Rather, the successful manipulation by aggrandizing individuals of the institutions that developed around these regional relationships was one early basis for the incipient concentration of power and the establishment of inequalities that are most apparent at early regional centers in the Andean highlands (Hayden 1998;Stanish 2003). The social and economic foundations for the emergence of these early centers clearly predate the ascendancy of these centers during the Middle and Late Formative in the Titicaca Basin, and obsidian is one class of artifact that was circulating consistently since the preceramic period that may reflect the changing socio-political role of exchange over the long term.

A study of obsidian production and circulation among Andean peoples across a time span of nearly 10,000 years demands the consideration of a wide variety of cultural contexts, economic systems, and socio-political structures. A major question about obsidian circulation through time concerns the nature of obsidian exchange and perceptions of value with increased distance from the geological source.

Obsidian was a raw material essential to the daily existence of common people. In one sense, the distribution of obsidian may have had more in common with the exchange of agricultural commodities or salt, than it did with either long distance trade in prestige goods such as lapis lazuli, turquoise, or exotic pottery, or with commodities traded long distances for ritual purposes, such as spondylus and strombus shells. Thus, one should expect the patterns of obsidian exchange to differ from exchange patterns of these other trade goods and reflect different kinds of social and economic interactions (Burger and Asaro 1977: 18).

Expanding on Burger and Asaro's (1977: 18) observation, this dissertation examines the possibility that archeologically perceptible strategies of quarrying and production at the source, considered in tandem with consumption patterns, can help to describe what these "different kinds of social and economic interactions" may have been.

How did the contexts of exchange and value change as obsidian was conveyed away from the source area geographically and with greater social distance? The expression "regimes of value" (Appadurai 1986: 5, 14-15) has been used to describe arenas where shared perceptions of value are used in the construction of worth and equivalencies for circulating goods. What was the role of frontiers, either between cultural or environmental zones, in the transmission and value ascribed to these goods? While there is no simple answer to these questions, obsidian appears to have fallen somewhere between the probable regional circulation patterns of salt on one hand, and gold or turquoise on the other. The production and circulation represented by salt and gold serve as types on a continuum in this discussion, and the given changing patterns in obsidian circulation in the south-central Andes, the pattern falls at different points along this continuum in different prehispanic time periods.

A second major theme in this research is the association between a camelid-focused economy and obsidian circulation. High altitude pastoralists had easiest access to obsidian sources as obsidian was found on geologically young volcanoes high in the puna ecological zone; and additionally, pastoralists had burden-bearing animals capable of carrying heavy loads (Burger and Asaro 1977: 41). The sharp edges of obsidian flakes have great utility for pastoralist functions including shearing, butchery, and castration, although it is important to note that these characteristics are not unique to obsidian. Alternative local materials, such as high quality cherts, are found in a number of regions in the south-central Andes where exotic obsidian was used. Obsidian appears to have been circulated largely by pastoralists following the advent of a pastoral economy; however a simple utilitarian explanation for obsidian use among pastoralists is incomplete given the regional data from consumption sites.

There is key countermanding evidence to a simple utilitarian association between obsidian consumption and the functional needs of pastoralists. First, the obsidian flakes observed at pastoralist sites, even in excavated contexts, are overwhelmingly small and apparently of insufficient size for use as shearing or butchery tools. These flakes appear to be debris from advanced stages of reduction, perhaps from projectile point production and resharpening, and utilized flakes as one would expect among pastoralists are rarely reported in the literature. Second, the trends in obsidian procurement, as will be described in this document, suggest that the focus was on acquiring large nodules; a fact that that is discordant with the predominantly small sizes of the formal obsidian points and tools found both in the vicinity of the Chivay source and in the larger region, even when accounting for distance-decay issues. If obsidian was principally utilitarian because it was used for projectile points for subsistence hunting, why did obsidian use in projectile points expand dramatically only after the food producing economy became well-established (from 3300 BCE onwards)? Finally, pastoralist sites are often found without obsidian. During the Late Intermediate Period and Late Horizon the evidence shows that camelid herds expanded markedly, yet obsidian circulation seems to have relatively declined. If obsidian was primarily for utilitarian pastoralist activities, why did it decline as the herds expanded? It appears that the distribution of Chivay obsidian was heavily influenced by the interaction of complex phenomena that included demand in a network of exchange and caravan-based links between the western margins of the altiplano and the Titicaca Basin, the enduring social network that connected communities in these different regions.

This research at the Chivay source supports earlier observations by Burger et al. (2000: 348) where the strongest correlation with widely distributed obsidian types appears to have been the availability of large nodules of homogeneous obsidian. Given the diminutive size of the formal obsidian tools found in the archaeological record, and despite the issue of resharpening, there is an important display aspect to large obsidian nodules that is lost in utilitarian explanations linking obsidian with pastoralism. The link between camelid pastoralism and increased obsidian consumption in the south-central Andes is not causal, as pastoralists also make wide use of other materials, yet the appearance of pastoralism followed by the traffic of regularized long distance caravans appears to have been largely responsible for disseminating the material.

In light of the patterns of obsidian production and circulation, the anthropological relevance of obsidian appears to go beyond the fact that the material was used to produce sharp-edged tools of use to pastoralists. The analysis in this dissertation, therefore, focuses on the large-scale social and economic changes that occurred in the south-central Andes and on the possible role of obsidian as a material used by early leaders to represent contact with sustained regional networks and to signal social differentiation.

1.1.2. Chemical characterization work in the Andes

This study benefits from the work of the many archaeologists who have documented obsidian in archaeological contexts throughout the south-central Andes and who have conducted chemical characterization studies of Andean obsidian. The basis for research at the Chivay obsidian source returns largely to obsidian sourcing efforts by Richard Burger and by Karen Mohr-Chavez in the 1970s because it was these early studies that established the regional significance of this obsidian type. These studies further served to demonstrate the utility of obsidian for examining regional relationships in the Andes and for catalyzing greater interest in the context and form of obsidian artifacts among archaeologists working in the region. The first chemical analysis of obsidian from the region was Mohr-Chavez and Gordus in 1971 at the University of Michigan (Burger, et al. 2000:271;Chávez 1977). In the 1974-1975 study at Lawrence Berkeley Labs the analysis conducted by Burger and Asaro (1977) analyzed 800 artifacts from 141 archaeological sites in Peru and Bolivia (Burger, et al. 2000:273). They isolated eight major chemical groups from this study, however twelve artifacts did not belong to any of these twelve groups and they were designated "rare" types.

A number of important chemical characterization studies have taken place in the region, though none as sweeping as Burger and Asaro's early work. The most significant work with respect to this study was the sampling conducted by Sarah O. Brooks in the Colca valley during the mid-1990s. Additional samples were collected by Steven Wernke during his 1999 Colca survey work. In the course of the present research, samples were collected throughout the study region. New developments in sourcing technology, in particular the portable XRF machine, promise to expand greatly on the insights provided by the chemical characterization studies of the twentieth century.

1.2. Structure of the Dissertation

1.2.1. Organization of presentation

Following this introduction, chapter 2 provides an overview of theoretical approaches to exchange and to raw material procurement areas from anthropological studies worldwide. Chapter 3 shifts the focus to the Andes with a review of the existing research that informs this investigation. In chapter 3 a discussion of economy, exchange, and long-distance interaction in the region is presented, as well as a summary of the evidence of Chivay obsidian consumption through time. Chapter 4 is a geographically oriented chapter with a discussion of the region in terms of climate, economy, geology, obsidian deposits, and finally the original chemical sourcing work that accompanied this research project. Chapter 5 describes the methods used in this dissertation with a focus on the novel methods employed by this project. Chapter 6 presents the results and analysis of a 33 km2 archaeological surface survey in six blocks throughout the Upper Colca study area. Chapter 7 describes the results of the testing program that included eight 1x1m test units, five of which are analyzed in detail here. Finally, chapter 8 is the summary of significant findings from this research and chapter 8 strives to reconcile the results of this research with the theoretical objects of the study.

1.2.2. Digital data availability

While a major methodological goal of this project was to exploit new technologies for spatial data and the organization of information, the presentation in this Ph.D. dissertation is largely confined by the traditional format of the library monograph. Large segments of data from this project are available online at

where additional maps, photos of features and artifacts, and searchable GIS datasets are available permanently. Links to online materials are provided, as well, in the Appendices at the end of this document.

1.2.3. Dates

Throughout this dissertation chronology will be discussed in calibrated years Before the Common Era (BCE). Where possible, 14C dates will be completely reported using the form that follows. The actual radiocarbon years before present will be presented with a lowercase "bp", the laboratory identification will follow, and then the range of 2? (95.4%) calibrated Before Common Era dates are shown as reported by OxCal v3.9 (Ramsey 1995;Ramsey 2003) using data presented in Stuiver, et al. (1998). Calibrated Before Common Era dates are shown with the uppercase letters "BCE".

1.2.4. Spatial data

Most spatial data from the Andes is in coordinate systems referenced to the Provisional South American Datum of 1956 (La Canoa) based on the International 1924 ellipsoid. In order to be compatible with topographic data, imagery, and the datum native to the GPS system the coordinates have all been converted to the modern WGS1984 datum using the ArcGIS three parameter transformation function "1208: PSAD_1956_To_WGS_1984_8". The 1991 three parameter transformation to WGS1984 for metric UTM data for Peru is as follows ?X = -279 m ± 6 m, ?Y = +175 m ± 8 m, ?Z = -379 m ± 12 m, and was based on 6 collocated points (Mugnier 2006).

A GIS database of chemically-sourced obsidian samples has been compiled from the central and south-central Andes based almost entirely on published materials. In the text that follows travel times are been reported between the source and the consumption locale as calculated using Tobler's (1993) Hiking Function. This function models travel velocity as a function of slope.

Walking velocity (km/hr) = 6 exp (-3.5 * abs (S + 0.05))
Where S = slope in degrees (?Z/?X)

Figure 1-2. Tobler's (1993) Hiking Function models foot travel velocity as a function of slope.

Tobler's function follows Imhof (1950) in deriving travel speeds of 5 km/hr on flat terrain and an optimum travel speed of just over 6 km/hr on a -3.5° downslope. Further elaboration of this hiking speed function, such as on or off path travel and llama caravan versus hiking speed was not attempted as there are too many unknowns to reliably model such differences. While the absolute travel velocity may be unreliable, the relative speeds for comparing one consumption site to the next are informative and correctly factor in the effects of travel over steep terrain versus travel across the gentle slopes of the altiplano. The function has been used elsewhere in archaeological contexts (Gorenflo and Gale 1990;Jennings and Craig 2001; href="/biblio/ref_4270">Kantner 1996;Van Leusen 2002).

Topographic data used in this project derive from GPS and from Digital Elevation Models (DEM) generated from two space-borne remote sensing platforms. Three dimensional GPS data was gathered throughout the project using Trimble GPS units and post-processed using the Arequipa IGS base station. Local topographic relief was acquired from the ASTERDEM dataset (30m), and regional scale topographic data was acquired from the SRTM (90m) dataset. As will be described in more detail in Chapter 5, ASTER imagery and DEM data proved to be extremely useful in designing and executing this work in the mountainous terrain of the Chivay source area.

Quickbird satellite imagery for portions of the study region were made available in 2006 as a part of GoogleEarth v4. These data are distributed by DigitalGlobe and they are 2.4m per pixel multi-spectral imagery pan-sharpened with 0.68m per pixel panchromatic imagery for natural color imagery, and GoogleEarth topographic relief is blended into the imagery from the SRTM 90m layer.

1.2.5. Photographs and scale

Linear units used in this project are exclusively metric. A tape measure was extended to exactly 1m in landscape photographs to provide scale. Despite the tape not being legible, the total length of the exposed tape is always 1m unless otherwise noted. In some instances the tape is not perpendicular to the photographer position, a situation that could lead to foreshortening and in such cases the 1m scale would become invalid. Laboratory photos were taken on a matte grey background with a 1cm grid in the background. Additional notes regarding the methodology are provided in Chapter 5.

1.3. Conclusion

This dissertation combines a theoretical discussion of economy in anthropology and a review of existing knowledge regarding Chivay obsidian in the south-central Andes with new data from the Upper Colca research project. The significance and implications of the data presented here will be need to be reassessed in the future when investigators are provided with additional chemical sourcing studies, new archaeological research in the south-central Andes, and further studies of the Colca area itself.

Ch2 Theory

2.1. Introduction

This study seeks to investigate long-term social and economic change from the perspective of production and exchange in a region with marked developmental changes in prehistory. First, major stages in south-central Andean prehistory have been widely discussed in the literature, including the transitions from foraging to food production, mobility to sedentism, egalitarianism to social ranking, and from low-level exchange to sustained interregional exchange. Should one not expect marked changes in the acquisition and production of obsidian that are correlated with these large scale prehistoric developments? For example, with the domestication of camelids and the emergence of llama caravans the costs in both time and labor in transporting weighty cargo, including obsidian nodules, must have dramatically decreased. Furthermore, with increased specialization one might expect merchant caravans or state-controlled interests to have emerged that benefit from the regional demand for obsidian. Complexities lurk behind each of these assumptions as these changes are largely co-evolutionary. Problems also lie in the very measures that are used to quantify changes in obsidian production, circulation, and consumption in the region.

Theoretical complications rest largely in three major issues; (1) the cross-cultural study of labor and value in economic anthropology, (2) associations between exchange and socio-political complexity, and (3) measures of prehistoric production that are both valid and practicable in the study of quarry areas. Each of these general issues will be addressed in turn below, and in the subsequent chapter (Ch. 3) the focus will be shifted to the south-central Andes.

Investigations using a close adherence to individual high-level theoretical approaches, particularly formal models of exchange and efficiency in production, were explored during the 1970s and 1980s in a number of obsidian quarry studies, most notably in the Mediterranean and in Mesoamerica. In these areas the regional trajectory benefits from distinct evolutionary changes like the development of highly efficient prismatic blades and well-documented mercantile traditions that perhaps conformed more closely to neoclassical economic cost-efficiency. Yet even in these areas, formal models encountered substantial difficulties in linking theoretical expectations with the breadth of regional exchange and in establishing empirical measures of control and efficiency beyond technical attributes of blade production. This study assumes a broader theoretical approach based primarily in finding correlates between evidence for prehispanic obsidian production and circulation and regional developments in economic organization and socio-political ranking. The theoretical issues presented here are linked to (1) evaluating the structure and level of independence of exchange networks and caravans, and (2) the degree of control exerted by household economies, aggrandizing individuals, and population centers prior to, and during, the Formative in the Lake Titicaca Region.

2.2. Anthropological approaches to economy and exchange

It is a critical limitation of archaeology, particularly when not given powerful assistance from historical sources, that it remains tied to the spatial distribution of imperishable objects or non-artifactual materials whose temporal sequence is seldom more than very roughly defined. Questions normally asked by archaeologists about trade and diffusion have been those which they can (or think they might be able to) answer directly and incontrovertibly from their data. But we must ask whether such questions are the important ones for an understanding of human society (Adams 1974: 139-140).

2.2.1. Economies

The study of ancient economies has long been a central theme in archaeological research. Lying at the intersection of human behavior and material goods, economy can often provide a theoretical basis for connecting features and artifacts with larger scale traits in prehistoric societies. Economic approaches to prehistory have the promise of bridging the premodern individuals and institutions of anthropological study with their material remains in a quantifiable way in order to examine processes such as change in subsistence, intensification, and exchange.

A fundamental question in economic anthropology concerns the cross-cultural applicability of economic models developed primarily for explaining behavior in Western, market-oriented societies. In the past 50 years economic anthropology has witnessed protracted debates between advocates of formal and substantive approaches to economy (Plattner 1989: 10-15). The perspective put forward by formalists is that ancient and non-western societies differ from those of modern capitalist societies in degree but not in kind (Wilk 1996). In contrast, substantivists recognize fundamental differences between ancient or non-capitalist societies and modern economies. These debates have been reconciled to a certain extent with the recognition that the two approaches largely complement each other. The distinct assumptions associated with formal and substantive economics, however, are critical to framing research questions about ancient economy, commodities, and exchange.

Formal approaches

Formal approaches explore the outcome of rational decision-making with regard to choices faced by prehistoric populations. By conceiving of economic behavior in terms of universal rationality, these approaches are analytically useful because they allow for the isolation of variables and for cross-cultural comparisons. Formal economic analyses are built on studies of modern markets by human geographers and involve reducing labor, land, and capital to the price as the unit of cost. Anthropologists using formal approaches may apply energy or time as value units to studies of food procurement, raw material provisioning, and settlement choice as in studies by behavioral ecologists (Winterhalder and Smith 2000) and archaeologists (Earle and Christenson 1980;Jochim 1976;Kennett and Winterhalder 2006).

Formal approaches to prehistoric exchange have been used by archaeologists to study the evolution in exchange systems both organizationally and from the perspective of individual behavior (Earle 1982: 2). On the larger scale, regression analysis (Hodder 1978;Renfrew 1977;Renfrew, et al. 1968;Sidrys 1977) and gravity models (Hodder 1974) use assumptions of cost minimization to differentiate between possible exchange systems in prehistory. "The sociopolitical institutions establish constraints in terms of the distribution and value of items. Then, individuals, acting within these institutional constraints, procure and distribute materials in a cost-conscious manner" (Earle 1982: 2). Neoclassical assumptions on the scale of individual behavior have also been used to examine the evolution of market based exchange (Alden 1982) and subsistence goods exchange by territorial groups exploiting high-yielding yet unpredictable resources (Bettinger 1982). A synthesis by Winterhalder (1997) investigates the way in which complex exchange behaviors that have been documented ethnographically can result from models of circumstance such as tolerated theft, trade and risk reduction, and models of mechanism such as kin selection and dual inheritance.

Critiques of formal approaches have been various, with strong methodological criticism coming from influential ex-formalists like Ian Hodder. On the subject of exchange, Hodder (1982: 202) observes that the explanatory power of formal approaches to prehistoric exchange are significantly weakened by the problem of equifinality in the empirical evidence, an issue discussed below. Hodder further argues that middle range links between social contexts, political strategies, and the empirical evidence provided by distributions of archaeological data are insufficiently accounted for with formal approaches such as regression analysis.

Substantive approaches

Substantivism arose in anthropology largely in response to what was perceived as the inapplicability of formal approaches and the assumptions of neoclassical economics to ethnographic case studies (Wilk 1996: 1-26). Substantive approaches emphasize that economy and exchange are fundamentally linked to other aspects of human behavior. To substantivists, economic institutions are effectively cultural traits, therefore techniques designed around "modern" or "western" conceptions of rational individualism are inadequate for application in non-western cultural contexts (Bohannan and Dalton 1962;Dalton 1969). The position was first articulated by Karl Polanyi (1957) that the economy is "embedded" in sociopolitical institutions. This view of economy and non-western exchange has its roots in studies by Malinowski (1922) and Mauss (1925), although the focus in that earlier period of economic anthropology was on social relationships, whereas social change dominated the discourse in economic anthropology during the 1960s (Dalton 1969).

Mid-twentieth century substantivism was based on a functionalist view of society as static and aspiring to the maintenance of equilibrium within the social environment (Schneider 1974). Interaction took place through reciprocity, redistribution, and exchange. Sahlins' (1965) further elaborated on reciprocity based exchange by placing generalized, balanced, and negative reciprocities along a continuum that served to describe the dynamics of interaction within specific social contexts. Another distinction lay between the transfer of inalienable gifts between reciprocally dependent individuals "that establishes a qualitative relationship between the transactors" (Gregory 1982: 101) and alienable commodities as transfer between reciprocally independent people "that establishes a quantitative relationship between the objects exchanged" (Gregory 1982: 100).

A number of critiques have emerged of the substantivist approach. One critique of substantivist approaches to exchange, one that is true of functionalism more generally, is an early version of agency theory. In the substantivist view, society is constructed of consistent rules within agents must act and includes a moral obligation of exchange (Dalton 1969: 77). In this organic perspective, however, "[t]here is little room for individual construction of social strategies and manipulation of rules, and there is little intimation of conflicts and contradictions between interests" (Hodder 1982: 200).

A second critique focuses on the distinction drawn by substantivists between complex and small-scale societies. Among complex societies, formal approaches are supposed to be more relevant, but in the context of small-scale societies substantive approaches are appropriate, and this dichotomy creates a dilemma of where exactly to draw the line. Monica L. Smith (1999: 111) argues that the ethnographic cases to which many substantivist models refer (i.e., small-scale societies) are lower in population density and involve fewer layers of interaction than would have been found in premodern states and empires and therefore formal approaches might have more relevance in premodern complex societies than is presented by substantivists.

A third critique holds that the direct correlation between forms of exchange and level of socio-political complexity lacks empirical support (Hodder 1982: 201). Anthropological studies suggest that unpredictability in food supply correlates with more extensive reciprocal exchange systems. Reciprocity is encountered more frequently among hunters, fishers and farmers than among gatherers and pastoralists who exploit relatively predictable resources (Pryor 1977: 4).

In the current day, Polyani's framework centering on the distinction between reciprocity, redistribution and market forces continues to be widely used in anthropology and archaeology, despite the development of alternative models that are particularly relevant to studies of commercialized pre-modern states. In particular, Michael E. Smith (2004) argues that more refined differentiation of transaction mechanisms can be used to distinguish the degree of internal and external commercialization in state-level societies, such as those presented by Carol A. Smith (1976). However, the subtleties of ancient commercial enterprise are less relevant in places like ancient Egypt and the Andes where historical and archaeological evidence attest to strong state control of uncommericalized economies without market-based economics, money, and independent merchants. In other words, Polyani's coarser distinctions of reciprocity, redistribution combined with non-market trade are sufficient in regions such as the ancient Andes where most scholars believe commerce played an minor role in prehistoric development (LaLone 1982;Stanish 2003). In Mesoamerica, however, where evidence for prehispanic markets and traders is widespread, anthropologists have found that Polyani's distinctions are vague and that models with further refinement in modes of commercialization are applicable (Braswell 2002;Smith 1976;Smith 2004).

Complementary applications of formal and substantive approaches

Formal and substantive approaches can be applied in a complementary manner. For example, in Stone Age Economics, Sahlins (1972) attempts to explain the existence of supply and demand curves among aboriginal exchange networks where social parameters determine the development of exchange. Winterhalder (1997) discusses some of the concepts presented by Sahlins (1972) using a behavioral ecology approach to gifts and exchange among non-market foragers. Winterhalder finds that issues such as social distance can be addressed more explicitly in a formal framework, albeit more narrowly, because economizing, neoclassical assumptions are a starting point for this type of analysis. As Hodder (1982: 200) points out, formal and substantive approaches are targeting different behaviors; formal economics relies on outputs and performance, while substantive analyses relate to social contexts of exchange.

While these differences make the two approaches irreconcilable on some levels, some studies attempt to integrate both avenues of research. Drawing on the advantages of both formalism and substantivism may be worthwhile, but a tendency to apply formal analyses to state-level societies, due to greater commercialism and specialization, and to apply substantive analyses to small-scale societies, should be resisted (Granovetter 1985;Gregory 1982;Smith 1999). The assumption that in premarket social contexts economic behavior is heavily embedded in social relations, but then is increasingly atomized and conforming to neoclassic analyses in modern, market-oriented societies is inconsistent. Granovetter (1985: 482) writes

I assert that the level of embeddedness of economic behavior is lower in nonmarket societies than is claimed by substantivists and development theorists, and it has changed less with 'modernization' than they believe; but I argue also that this level has always been and continues to be more substantial than is allowed for by formalists and economists (Granovetter 1985: 482).

Similarly, Danby (2002) observes that there are serious logical flaws in the false dichotomy, widely applied by archaeologists, where neoclassic, cost-minimizing logic is applied to premodern, complex societies but increasingly assuming an embedded "gift" economy among smaller-scale economies.

This dissertation research project follows on the substantivist tradition in the Andes, but formal approaches have been influential in exchange studies worldwide. The weaker elements of the substantivist approach will be avoided by not assuming a direct correspondence between socio-political complexity and volume or type of exchange.

2.2.2. Transfer of goods and exchange value

In the 1970s much of the debate had shifted from formalism and substantivism, to Marxism and structuralism. Marxists approach economic anthropology comparatively by focusing on production, arguing that society's basic forms of exploitation and inequality are continually being recreated in modes and relations of production (Godelier 1979). Structuralists hold that one had to understand the total system of meaning in order to interpret the relative value of items in a society, but they reject the functionalist assumption that economic institutions serve to integrate society (Graeber 2001: 18).

In the 1980s, the emphasis shifted to consumption. In an influential paper titled "Commodities and the Politics of Value," Arjun Appadurai (1986) states that anthropologists can more effectively examine cross-cultural patterns in economy and exchange by focusing on exchange from the consumption perspective. Appadurai follows Georg Simmel (2004 [1907]) in arguing that the source of an object's value is based in its exchangeability and the desire of the buyer, not the labor that went into production. As with formal approaches, this emphasis on exchange permits broad cross-cultural comparisons because exchange activities are universal characteristics of human behavior. Appadurai's approach shares some of the limitations of formal approaches. Appadurai focuses on the commodified value and the value that goods accrue primarily through transfer, which put constraints on the analytical potential for looking at the social and symbolic significance of human relationships organized around exchange. Rather, Appadurai investigates an object's "life history" as it has passed through multiple hands. The emphasis is shifted from two individuals exchanging goods to the relationship of equivalencies between the two objects being exchanged (Appadurai 1986: 12-17).

A notable challenge to Appadurai's contention that "circulation creates value" is Weiner's (1976: 180-183) observation that among the societies discussed by Mauss (1925) the value of the objects in question is associated with their original owners and their specific histories, and value is not the product of transfer. Thus, the value is not predominantly a result of demand, as asserted by Appadurai, and it is often the case that value reflects the inalienable properties of specific heirlooms some of which, like the royal crown jewels, do not circulate at all. Igor Kopytoff (1986: 74) describes a variety of items that have "singularity" due to restricted commercialization. Examples of this include ritual items or medicines in western society that are destined specifically for the intended patient that have prohibitions against resale. In his critique, David Graeber (2001: 34) suggests that for any society it might be possible to map "a continuum of types of objects ranked by their capacity to accumulate history: from crown jewels at the top, to, at the bottom, such things as a gallon of motor oil, or two eggs over easy". Unlike Appadurai, Graeber's approach places goods in a continuum that does not depend on transfer to establish the relative value of an item. The distinction between alienable and inalienable goods is linked to relative abundance and exclusivity of the products in question and to historic characteristics, like the degree of influence of a price-regulating market (Miller 1995).

Exchange and social distance

The dominant ethnographic approach in twentieth-century substantivist anthropology contrasts non-western gifting and exchange with capitalist, market-oriented trade in alienable commodities. Malinowski (1922) describes forms of exchange ranging from pure gifts to trade with increasing self-interest and "equivalence" as one moves towards the trade relationship. Mauss (1990 [1925]) attacks the notion of the pure gift and instead focuses on the temporal aspects of gift-giving, with the establishment of what are effectively ancient forms of credit. Further, Mauss focuses on the sociality of gift exchange with the idea that gifting can take hostile forms by inflicting obligations that the recipient may fear. Sahlins (1972: 191) subsumes the sociality of exchange relationships into a single trajectory with the concept of social distance. The "social distance" between exchange partners is a way of conceiving of the degree of familiarity and information transfer between producer and consumer of goods.

Figure 2-1. Varieties of reciprocal exchange (after Sahlins 1972: 199).

In Sahlins' classic taxonomy of reciprocity he identifies generalized, balanced, and negative reciprocity as a continuum from gift through exchange to haggling and theft, with explicitly different moral standards applying to transfers occurring in each sector. These modes of reciprocity were mapped directly on to increasing social distance originating in the household, although Sahlins was unspecific about the institutional forms of exchange, allowing for wide cross-cultural variation.

Besides the functionalist underpinnings of Sahlins' approach, one critique of his social distance model rests on his distinction between the role of the state and social relationships based on economic ties.

Because Sahlins is constructing a general theory, he is not obligated to deal with the refractory ethnographic data of one particular group. Hence, Malinowski's multiple heterogeneous categories of gifting never arise. In its place, Sahlins has a single Big Idea developed in the famous chapter on Hobbes and Mauss. The idea is that economics is politics for primitives. The understructure of human society, the default, is the chronic insecurity of Hobbes' "war of all against all." Primitives lack the state that is Hobbes' preferred remedy (Danby 2002: 24).

The result for Sahlins rests on a weak dichotomy between primitive|modern economic contexts that includes (1) primitive contexts where relations are mediated by social and economic relationships, and (2) modern contexts given over to the logic of neoclassic economics because commercialization and the circulation of alienable commodities are qualitatively different. Danby presents instead a "post-colonial critique [that] sees gift|exchange as yet another mapping of primitive|modern, an ultimately tautologous them|us split in which 'they' are various negations of what 'we' think 'we' are" (Danby 2002: 32). Anthropologists have sometimes followed this dichotomy where long-term investment and the establishment of social relationships in transactions is given over to the gift side, while exchange in alienable commodities is expected in complex, state level economies where exchange transactions are analyzed following the neoclassical approach and spot transactions are conducted in open markets.

In contrast, Danby argues that, if anything, one should expect greater temporal complexity and mechanisms for extending forward credit with greater social complexity.

Rather than building theory by pushing off against neoclassical exchange, let us put neoclassical exchange aside... the asocial spot transactions on the right-hand ends of [Sahlins' (Figure 2-1)] gift-exchange continua are also marginal to wealthy capitalist economies. Forward, debt creating transactions, embedded in social relations that often entail power, are central. Moreover, they are likely to be socially embedded in a variety of ways that should lead us to rethink the dichotomies between non-gift and gift (Danby 2002: 28).

Instead of framing the anthropological discussion of exchange in opposition to neoclassical approaches, Danby argues that the discussion should be reframed around institutions that underlie transactions in society and provide the long-term temporal milieu within which transactions occur. Temporality is particularly important when exchange is framed by production calendars based in agricultural cycles and by social calendars centered on gatherings and ceremonial occasions. By this standard, goods lying on the inalienable end of the continuum are fundamental to examining all economies in an anthropological light.

Common goods and luxury goods

The inalienability or uniqueness of an object sometimes parallels the categories of common and luxury items, with common items being mutable and replaceable. Yet, the classification as 'common' or 'luxury' for a given object is not discrete. It has been widely observed that the distinction between common and luxury items is contingent in space and time.

The line between luxury and everyday commodities is not only a historically shifting one, but even at any given point in time what looks like a homogeneous, bulk item of extremely limited semantic range can become very different in the course of distribution and consumption...Demand is thus neither a mechanical response to the structure and level of production nor a bottomless natural appetite. It is a complex social mechanism that mediates between short- and long-term patterns of commodity circulation (Appadurai 1986: 40-41).

Appadurai presents sugar as an example of a product with widely varying significance, and he observes that it is not merely the possession of exotic goods, but also the social significance of the knowledge surrounding the production or consumption of these goods, that carries significance. Goods that are irregularly distributed in space, portable, and that have wide consumption appeal such as sugar, salt and obsidian, were transported great distances in prehistory. For example, in the Mantaro Valley of the central Andes, Bruce Owen (2001: 280) notes that a number of metal items such as copper needles went from being wealth goods found primarily in elite contexts prior to the Inka conquest, to being utilitarian items and found with equal frequency in commoner contexts under Inka rule.

The movement of commodities has the significant effect of reinforcing relationships between social groups in the sense that the Kula ring promotes regular contact (Malinowski 1984 [1922]). However, in circumstances where the elite strive to control consumption of commodities, exchange can threaten the position of the elite. In Appadurai's perspective, both rulers and traders are the critical agents in articulating supply and demand of commodities.

The politics of demand frequently lies at the root of the tension between merchants and political elites; whereas merchants tend to be the social representatives of unfettered equivalence, new commodities, and strange tastes, political elites tend to be the custodians of restricted exchange, fixed commodity systems, and established tastes and sumptuary customs. This antagonism between "foreign" goods and local sumptuary (and therefore political) structures is probably the fundamental reason for the often remarked tendency of primitive societies to restrict trade to a limited set of commodities and to dealings with strangers rather than with kinsmen or friends. The notion that trade violates the spirit of the gift may in complex societies be only a vaguely related by-product of this more fundamental antagonism (Appadurai 1986: 33).

The suggestion that exchange promotes contact that threatens elite control results from either a relatively independent group of traders, or competing elites that utilize external contacts to advance their positions. Appadurai's point, however, is that in the realm of political finance one might expect established elites to strive to exert greater control over the values of production and exchange rather than relegate this important issue to those who traffic in long-distance goods, such as caravan drivers.

Following Polanyi, one might expect to find that in complex, premodern societies without market-based exchange there was dominance of administered trade. However, a number of scholars have observed that Polanyi (1957) appears to have ruled out precapitalist commercialism in ancient states a priori, as he claimed that there were no true markets or prices that reflected supply and demand in the ancient world, but rather equivalences in value that were set by rulers in an administered context. These "dogmatic misconceptions" (Trigger 2003: 59) appear to have caused Polanyi to distort the historical evidence as his views of administered markets in premodern states have been widely refuted (Smith 2004: 75-76;Snell 1997). Historian Philip Curtin (1984: 58) observes that even within contexts of administered control traders may have had more freedom than previously thought.

In some regional contexts, such as the prehispanic Andes, Polanyi's interaction types continue to be viable because there is little evidence of commercialization and open markets. The prevailing interpretation of prehispanic economy, for the Inka period in the central Andes at least, is "supply on command" (LaLone 1982) with elite economic domination primarily through labor mobilization. This issue will be explored in more detail in the Chapter 3, but Andeanists have largely found that Polanyi's non-commercial typology of premodern exchange suffices for analyses of prehispanic Andean economies.

Exchange promotes interaction between communities, yet Appadurai also explains that "such a tendency is always balanced by a countertendency, in all societies, to restrict, control, and channel exchange" (1986: 38). Under the rubric of "politics of value" Appadurai describes a pattern of institutional or elite control of commerce monopolizing or redirecting the flow of commodities. On this issue Graeber characterizes Appadurai's 1986 framework as "neoliberal" and describes it as follows

Appadurai leaves one with an image of commerce (self-interested, acquisitive calculation) as a universal human urge, almost a libidinal, democratic force, always trying to subvert the powers of the state, aristocratic hierarchies, or cultural elites whose role always seems to be to try to inhibit, channel, or control it. It all rather makes one wish one still had Karl Polanyi (1944) around to remind us how much state power has created the very terms of what is now considered normal commercial life (Graeber 2001: 33).

Appadurai cites examples of historic royal monopolies and sumptuary laws restricting trade items in Rhodesia and renaissance Europe support his argument (Appadurai 1986: 38). However, without the state-organized overhead of basic institutions guaranteeing peace, investment, and the protection of property, the movement of goods would involve much more danger of theft by brigands or goods could be simply confiscated by rival authorities.

Appadurai's focus on consumption is a significant contribution in that it releases anthropologists to address activities specifically related to consumption behavior, even though the neglect of production presents an incomplete picture of economy. Consumption loci are often the contexts where the strongest evidence or patterning of interaction is found. Thus, consumption is particularly useful to archaeologists who may only have evidence of changing consumption patterns, or who must infer links between production and consumption indirectly.

Exchange in subsistence, cultural, and prestige technologies

The distinction between the circulation of common and luxury goods observed by anthropologists can be considered in terms of the larger economy and the organization of technology by linking these concepts to those of practical and prestige technologies. As described broadly by Hayden (1998), practical technologies are those that are primarily organized around principles of sufficiency and effectiveness, while the logic of prestige technologies is fundamentally different because it is oriented towards social strategies where greater labor investment in products serves to communicate the wealth, success, and power of the parties involved. A third category, termed cultural goods, will also be used to examine the relationship between material culture and exchange. The concepts of practical, cultural, and prestige goods will be used to link long term changes in the organization of technology and production with consumption patterns and socio-political evolution.

This framework is relevant to this discussion of anthropological approaches to exchange because it places the contrasting behavior that Appadurai terms "common" and "luxury" goods into an empirical and evolutionary framework capable of addressing change over long time periods. Thus while the earlier discussion of "luxury goods" that transcend social and political boundaries referred to particular contexts of circulation, these are specific manifestations of goods Hayden would include in the broad category of prestige technology, as these objects are labor intensive to produce or acquire.

The concepts of practical and prestige technologies parallel in some ways the economic distinction made by Earle (1987;1994) between subsistence and political economies. The household level subsistence economy is based on satisficing logic, while the political economy involves the mobilization of surpluses and competition between political actors, and is subject to the maximizing strategies of elites. The organization of technology is sometimes approached in terms of three groups as Binford (1962) has done with "technomic", "sociotechnic", and "ideotechnic" categories. For Binford, technomic objects correspond to "practical technology", and the sociotechnic and ideotechnic groups would largely, but not exclusively, overlap with "prestige technologies". Hayden (1998: 15) observes that labor inputs are low but sociotechnic significance is high in Australian Aboriginal string headbands that signal adult status, and labor is low but ideotechnic significance is high in a pair of crossed sticks tied together to represent a crucifix. Hayden points out that despite these exceptions, items of ritual and social significance are often made with relatively costly materials (such as gold crucifixes) but he admits that many items fall between his categories such as decorated antler digging-stick handles, and "the analysis of such objects becomes especially complex where the prestige materials such as metals or jade are actually more effective, but far more costly, than more commonly used materials" (Hayden 1998: 44-45). Obsidian presents a similar dilemma where, on the one hand it is rare in many regions, and it is unusual-looking, and yet it is also more effective for many kinds of cutting functions and so the inducement to use the material is not straight-forward. As an object can move between categories it should be said that ultimately the significance of an item is not an inherent property of the object, but rather it is created by contexts of use or consumption and should best be considered in terms of labor, exchangeability or life-history (Appadurai 1986;Graeber 2001).




Subsistence Goods

Practical technology. As a response to stresses item must be effective. Widely available, minimizing costs due to satisficing logic, distributed shorter distances.

Simple tools: axes, saws. Simple baskets and pottery. Bulky, low value foods (i.e., cereals, tubers). Common metals (later Old World prehistory).

Cultural Goods

Widely available goods, but with information content and social and ideological significance. Aesthetic but non-labor intensive and non-exclusive designs.

Some projectile technology. Textiles, shell, herbs and medicines such as coca leaf. Items for commonplace rituals.

Prestige Goods

High labor inputs, investment logic due to political potential of consuming labor, sometimes wide distribution in circumscribed contexts. Competition may mean that these are consumed or destroyed.

Rare metals; serving vessels: ceramics, baskets; jewelry; tailored clothing. Musical instruments, high value foods, rare or costly herbs and medicines.

Table 2-1. Three categories of exchange goods, the boundaries on these categories are contingent on contexts of production, circulation, and consumption. Many items like meat, maize, and obsidian tools can belong in any one of these groups depending on form and availability.

According to Hayden (1998: 44) any material that is transported more than two days should probably be considered a prestige technology due to labor investment, however he mentions that perhaps a useful intermediate category of "cultural goods" could be defined consisting of non-prestige ritual or social artifacts. In his analysis of long distance caravans in the Andes, Nielsen (2000: 66-67) borrows concepts from Hayden's practical and prestige technologies with some modification for a discussion of exchange goods. Nielsen defines "subsistence goods" and "prestige goods", but he also defines a third category as "cultural goods" to include maize used for subsistence but also for ritual, as well as coca and textiles. These goods are often non-exclusive, but their form and consumption carries significance and the circulation of such items does not conform to the satisficing logic of practical technologies and subsistence goods.

Ordinary goods and archaeological theory

The fact that many artifacts cannot easily be classified into one of the above groups in a particular time or cultural context shows the difficulties inherent in developing a generalized framework for addressing production and exchange through prehistory. Obsidian is a prime example of a material that has practical value, but it is visually distinct and it is also a material with cultural significance and prestige associations in some contexts. Thus, while the focus in exchange studies has been on prestige goods linked to status competition, because these activities have evolutionary consequences, subsistence goods may also contain social or cultural information. The association of items luxury or commonplace categories is a function of geography, technology, and socially defined valuation.

Information content and everyday goods

Monica L. Smith (1999) develops an argument based on a dichotomy between "luxury goods" and "ordinary goods", where some of the ordinary items used in household activities and moved through kin-based exchange networks form an important, material component of group identity. This essay could also be used to support an argument for an intermediate category of cultural goods. Smith notes that the circulation and consumption of such ordinary but visually distinct household goods serve to maintain cultural links and symbolize status markers that probably precede, and indeed form the structure for, later social ranking. She observes that archaeological discussions of exchange often falsely imply that exchange links were established by an exclusive elite population and these links eventually become established and expand to ordinary goods.

Brumfiel and Earle (1987: 6) find the distinction between luxury and utilitarian goods setting the stage for the organization of economic activity in early complex societies, citing "... the lack of importance of subsistence goods specialization for political development." The sequence in which different types of goods are incorporated into exchange patterns is explained as an evolutionary sequence paralleling developments in sociopolitical complexity, so that "as trading routes and trading relationships became more firmly established, everyday goods were added to the merchants' repertoire...and came to supply not only valuable items for elites but also food staples and utilitarian wares for people in the society generally" (Berdan 1989: 113; cited in Smith 1999: 113).

Some scholars have historically placed a priority on the influence of status or luxury goods by elites in some centralized political sphere in stimulating and maintaining long distance trade links (Brumfiel and Earle 1987;Smith 1976), arguing that the political objectives of elites and aspiring elites were the impetus for long-distance links. Yet as M. L. Smith (1999) applies the socio-semiotics of Gottdiener (1995), the consumption of particular materials can have social significance and can convey information content at a variety of levels. Thus the capacity for kin-based reciprocal exchange networks to distribute household items over distance, or household level caravans to emphasize relatively mundane products, should not be underestimated.

The intent of many archaeologists focusing on the role of status goods exchange seems to be not necessarily to deny the capacity and symbolism of household-level exchange as much as it is to emphasize the political and economic significance of exchange in status goods controlled by elites. Goods that circulate widely within a particular community may serve to express community participation or corporate affiliation (Blanton, et al. 1996). Hayden's distinction of practical and prestige technologies focuses instead on effectiveness, for the first group, and high labor inputs for the second. Thus, to reconcile this with M. L. Smith's argument, the third group that includes "cultural goods" conveys important social and ideological information beyond the satisficing "effectiveness" stipulation, but simultaneously is widely available and cross-cuts social hierarchy.

Availability and consumption patterns

One reason that subsistence goods, cultural goods, and prestige goods are non-exclusive categories is that consumption patterns associated with these goods have changed as availability changed through time. The availability of a given material changes through time, be it obsidian in the prehispanic Andes or glass drinking vessels in ancient Rome, and availability conditions the importance of its consumption (Appadurai 1986: 38-39;Smith 1999: 113-114). In cases of intensified craft production, availability may be determined by labor specialization, production units, intensity, locus of control, and context of production (Costin 2000). With commodities based on raw materials, the primary determinant of availability in most cases is geographical distance from the source, but economic patterns, socio-political barriers, technology of procurement and transport, and rate of consumption all affect availability.

Along with naturally occurring raw materials that are irregularly distributed across the natural landscape, such as obsidian where sources are rare, one may expect behaviors associated with scarcity to apply to geologically occurring minerals only with diminished availability as one moves away from the source of these goods. Thus the availability of goods such as obsidian over the larger consumption zone for these materials will vary from abundant to scarce depending on geographical relationships and socio-economic links between the source and the consumption zone. The archaeological study of commodity distribution, and in particular the relationship to economy and to socio-political evolution, spawned years of research into regional exchange beginning with the work of Colin Renfrew (1969) and that of his colleagues.

The circulation of flaked stone

Focusing here on the differences in artifact use in regions where raw materials are abundant, versus those places where they are rare, permits several generalizations with regards to raw material consumption. As a material class, flaked stone is durable and often the material is sourcable to geological origin point and due to these features, lithics analysis share some attributes with consumption studies of other artifact classes. One characteristic that differentiates lithics is that they are among the more resilient archaeological materials, and are sometimes used as a proxy for mobility or exchange. However, the lithics material class is comparable with other artifacts of consumption like ceramics, food goods and textiles, in that lithics are used to produce goods that range from mundane or subsistence-level to elaborate forms that imply the goods were inscribed with social or ritual importance.

Lithics have important differences from other artifact classes, however. Principally, the production of stone tools is a reductive technology and flaked stone tools inevitably become smaller with use. This directionality of lithic reduction, which allows for technical analysis and refitting studies, signifies that, unlike metal projectiles, textiles, and other widely exchanged materials, the down-the-line transfer and use of lithics has distinctly circumscribed use-life based on reduction. The second major implication of the reductive nature of stone is in regards to the social distance between the producer and consumer. As stone artifacts become inexorably smaller with production and use, larger starting nodules can take more potential forms and have a generalized utility that is progressively lost as reduction proceeds. In terms of the exchange value of a projectile point as a "cultural good", the roughing out of a lanceolate point, for example, may have determined the cultural value of the preform such that it would have had less potential, and therefore less value, in contexts where triangular points are used. With scarce lithic raw materials the size of the item probably related directly with its reduction potential; therefore larger nodules would probably have had value in a wider range of consumption contexts.

Lithic procurement, distribution, and consumption are in some ways comparable with other classes of portable artifacts, and in some ways quite distinct. Archaeologists have used the spatial relationships between lithic raw material and behavior to study the ways in which the availability of a particular material type affects prehistoric behavior with respect to production, curation and mobility (Bamforth 1986;Luedtke 1984;Shott 1996). Procurement, distance from source, and the embedding of lithic provisioning in subsistence rounds have specific consequences with respect to raw material use in the vicinity of a geological source area (Binford 1979;Gould and Saggers 1985), an issue to be discussed in more detail below. Regardless of mode of transfer and other distributional issues, the use of lithic raw materials, as with other artifact classes, is contingent on variability in a number of dimensions. These dimensions include whether the material is abundant or rare, lightly or intensely procured, laden with cultural or prestigious associations, as well as circulation and demand, although many of these variables can be difficult to isolate archaeologically.

2.2.3. Transfer of goods and socio-political complexity

Prehistoric exchange has long been central in anthropological and economic models of socio-political change. Exchange has been discussed by social theorists, anthropologists and archaeologists the debating the ultimate causes of socio-political inequality that emerged independently during the Holocene in different regions of the world. In the course of the last 12,000 years, archaeological evidence documents a change from a worldwide of hunting and gathering groups with relatively egalitarian, village-level social organization to, in a few places world-wide, in-situ development of state-level society with large settlements, greater concentration of wealth, and hierarchical economic and political organizations. While anthropological models accounting for political and economic change have been refined over the past century, evidence of exchange has consistently served as a material indicator of regional interaction and differential access to exotic goods within a community.

Exchange as a prime mover?

Regional exchange is of particular utility to anthropological archaeology because it is a material phenomenon that occurs at the intersection of social relationships, resource procurement, and individual or community differentiation. However, exchange has been attributed with great causal significance in the past and, while it provides evidence of particular utility to anthropologists and archaeologists, it is important to not ascribe excessive significance to exchange in human affairs. For example, to V. Gordon Childe (1936) long distance trade was a prime mover that had direct social evolutionary consequences. Engaging in trade had derived benefits that can affect the balance of power between competing polities and stimulate larger forms of organization.

In the 1970s, Rathje (1971;1972) proposed that it was the lowland Maya need for "essentials" like obsidian, salt, and grinding stones that propelled them to seek external sources for these scarce goods, initiating long-distance relationships. Those who control access to scarce materials accrued benefits from their position in the trade network, he argued, and hierarchies emerged from the differential access to these required, or desired, trade goods. Virtually all of Rathje's postulates have since been disproved, but some variant of the theme first articulated in Rathje's exchange model underlies many contemporary investigations of ancient trade (Clark 2003: 33-34).

In the multilinear evolutionary framework of Johnson and Earle (1987), trade is part of an array of processes that are relevant to socio-political development, but trade does not become a dominant process until the formation of the Nation-State.

Economic Processes in Political Evolution

Local Group

Regional Chiefdom

Early State






risk management


(risk management)

(risk management)









Table 2-2. Economic processes in political evolution (Johnson and Earle 1987: 304) where capitalized words indicates a dominance of that process, parenthesis indicate important secondary processes.

The approach taken here is that all human groups engage in exchange in some form and it therefore serves as a consistent index of regional relationships through time. Furthermore some trade items, such as obsidian, appear to have been part of a suite of features that, at particular times and socio-political contexts, serve to differentiate individuals or groups from the rest of their community for purposes relating to political strategy. While exchange is not a prime mover for socio-political change, transfer between peoples and groups in different forms provides a vehicle for social differentiation and a materialization of geographical relationships that are closely linked to socio-political organization. Thus, while one may see exchange occurring over the long term, it is not a prime-mover but it is changes in evidence of production, transfer, and consumption of non-local goods that inform theoretical models of prehistoric change.

Exchange concepts reflect the current theory

Forty years ago, advances in technical analysis in archaeology, in particular geochemical sourcing, initiated several decades of research into exchange by archaeologists. Following the dominant theoretical approaches of the time, adaptationalist or managerial models held that socio-political change was stimulated by factors that impacted the entire cultural system. Such changes could be stimulated by a combination of factors, including external pressures and internal pressures. External pressures include resource stress, drought, and warfare. Internal pressures include greater efficiency in organization despite larger population levels, irrigation, and organizational complexity due to circumscription.

More recent theoretical approaches focus on social and political strategies prioritize the active role of individuals in the process of political change. The role of commodity exchange in the onset and dynamics of socio-political complexity connects the exchange behavior documented by anthropologists with the long term changes that underlie the archaeological evidence of exchange in prehistory. Leveling mechanisms among hunter-gatherers promote intergroup sharing and are especially prominent when resource predictability is low. With greater resource predictability, and especially when resources are intensifiable, these leveling mechanisms often begin to break down and social differentiation is observed.

Exchange is thought to have been one of a number of interacting behaviors by "aggrandizing" individuals that had the indirect result of institutionalizing social inequality (Clark and Blake 1994;Flannery 1972;Hayden 1995;Hayden 1998;Upham 1990). What were the socio-political contexts that permitted individuals to differentiate themselves, and how did individuals pursuing their interests result in the long-term changes observed archaeologically?

Early leaders

While early social hierarchy emerged from the so-called egalitarian contexts of small-scale societies, Wiessner (2002: 233) observes that "hierarchy characterizes the societies of our closest non-human primate ancestors and seems to be deeply rooted in human behavior" and social hierarchy can therefore be considered a "reemergence" of hierarchy. Egalitarianism is not the "tabula rasa for human affairs on which aggrandizers impress their designs" (Wiessner 2002: 234); rather, it is the result of coalitions and complex institutions of weaker individuals in society that curbs the ambitions of the strong and results in egalitarian structures that were as complex and varied as hierarchical power structures. Wiessner argues that cooperative behavior, such as elaborate exchange relationships, in egalitarian institutions fostered ideologies of coalition building and cooperation in kinship-based networks that extended far beyond local groups. While the egalitarian ethos constrained the competitive activities of aggrandizers, these institutions also provided powerful tools to the cooperative structure that could be used to the advantage of aggrandizers. Exchange has particular significance in agency models that strive to explain how the activities of aggrandizers can result in institutionalized inequalities because exchange can contribute to an understanding of the relationship between structure and agency at this juncture.

Aggrandizers had to work within powerful institutional boundaries that already existed in egalitarian societies in order to forward their interests. The changing nature of exchange, embedded within shifting institutional contexts, complements the emphasis on alienable commodities (Appadurai 1986, see Section 2.2.2). Aggrandizing individuals operate within institutional contexts, indeed their vehicle to promotion is social recognition, yet the ability to organize resources to obtain lower costs for particular items benefits aggrandizers and their supporters. Exchange is a universal feature of human societies, and studies that document the diversity of forms that exchange takes in contexts of early social ranking can shed light on the specific strategies used by aggrandizers that resulted in institutionalized inequality (Clark and Blake 1994).

Prehistoric institutions

Beginning with settlement pattern studies explored by Julian Steward, processual archaeologists often assume regional perspectives that emphasized social and organizational processes (Willey and Sabloff 1974). Building on Polanyi's observation that fundamental aspects of human institutions are economic, archaeologists understood that documenting the regional distributions of artifacts resulting from prehistoric mobility or exchange could contribute to reconstructing past societies. Archaeologists can use direct evidence from the production and consumption of archaeological materials and inference about the likelihood and form of prehistoric exchange to point to the institutional contexts of the ancient economy.

In Polanyi's (1957) seminal article outlining the substantive approach, "The economy as an instituted process," the economy is characterized in terms of two linked properties. First, there is the material process by which items are produced, circulated, and consumed; second, there is the economic form organized around socio-political relationships that arranges interactions diachronically and spatially. In this context, institutions serve as rules and obligations connecting human organizations around the process of producing and circulating goods. The implication is that an archaeological study of variation in prehistoric economies requires that archaeologists can document differences in the institutions that structured past economies.

Many contemporary theoretical approaches downplay the importance of structural analyses in favor of agent-based models, however anthropologists using New Institutional Economics (North 1990) are emphasizing the interdependence of institutions and economics and the implications for activities of individuals (Acheson 2002;Ensminger 2002). Douglass North (1990: 3) defines institutions as "the rules of the game in a society or, more formally, the humanly devised constraints that shape human interaction". This view considers how institutions developed from the decisions of individuals over long periods of time, affecting transaction costs through the benefits of coordinated activity. "Based on the advantages of lower transaction costs in commodity flows, [these anthropologists] argue that political and social institutions developed to regulate commodity flows, maintain regional peace, and guarantee contracts" (Earle 2002: 82). In this perspective, institutional complexity developed as rules to govern the economizing nature of individuals, and the emergence of such institutions should be correlated with increasing quantities of commodity exchange in prehistory. The perspective that argues that volume of exchange should correlate with increased social complexity has roots in theories of progressive evolution (Childe 1936), adaptionalist views (Steward 1955), and the managerial role of chiefs (Service 1962).

However, a positive correlation between evolving institutional complexity and a uniform increase in volume or variety of exchange in prehistory is not supported empirically (Brumfiel 1980;Hughes 1994;Kirch 1991). That is, contrary to the evolutionary expectations of some theorists, the volume or variety of goods transferred does not necessarily reflect the political or institutional complexity in a given society. Perhaps some of this inconsistency is the result of elite control of the circulation of commodities as proposed by Appadurai (1986: 38). Earle (1994: 420-421) observes that on the whole, the archaeological evidence is characterized by a great deal of variability in the types of goods exchanged, the volume of exchange, and the social contexts of exchange. In order to address this variability in an evolutionary framework, Earle suggests that exchange should be considered in terms of two categories, the subsistence economy and the political economy[1](Johnson and Earle 2000), with exchange taking different forms in each category of the economy. The exchange in these two forms of economy correspond largely to the distinction discussed earlier between ordinary goods and prestige goods where, according to Earle, political strategy is advanced by elites by, among other things, the manipulation of exchange relationships.

In a non-evolutionist application of New Institutional Economics in anthropology, Wiessner (2002: 235) differs from North in arguing that transaction costs for exchange are actually high in small-scale societies due to the close relationship between social and economic transactions; there is less neutral space for "unembedded" economic behavior with their associated overhead costs. In such contexts egalitarian institutions developed to foster trust and make interactions more predictable.

In non-market economies in which kin-based exchange systems play an important role in reducing risk (Wiessner 1982;Wiessner 1996) the goal of exchange is to be covered in times of need. In this context, the social and the economic are closely intertwined (Mauss 1925) and it is undesirable for returns to be stipulated as to time, quantity, or quality (Sahlins 1972). The most valuable information in such exchanges is the details of the partner - what he or she has to offer and will offer over the long run (Wiessner 2002: 235).

Elaborate institutions regulating behavior are not new, argues Wiessner, but if there is a weakening in the egalitarian prohibition against the accumulation of wealth or power the "alchemy of ambition" (Wiessner 2002: 234) drives a few individuals to seek preferential access to resources. In contexts of regional packing and circumscription it is argued that hierarchies emerge (Brown 1985;Johnson 1982) that are perhaps founded on the control of labor and surpluses, and instituted through ritual (Aldenderfer 1993;Hayden 1995)

Materialist perspectives hold that economic gains are used to bring about changes in the social order, and that these changes are then legitimized and perpetuated through ideology. These economic gains are generally achieved through intensification of production and through finance organized by ambitious individuals (Boone 1992;Earle 1997;Hayden 1995). Clark and Blake (1994: 17), building on practice theory (Bourdieu 1977;Giddens 1979), describe a model whereby institutionalized inequality is the unintended outcome of political actors competing for prestige by using strategies that match the self-interests of their supporters. Ultimately, these strategies may develop into redistributive institutions that appear to build on the social-leveling mechanisms described by Wiessner (1996), but operate on a larger scale and result in political advantage for redistributing leaders and their allies. In other words, it is institutions themselves that form the basis for leveling mechanisms, and it is institutions that are transformed into vehicles that serve to, in part, legitimize social inequalities.

An important temporal component to exchange is connected to institutions as well because transactions involve anticipation and scheduling. As Colin Danby (2002) argues, most transactions in neoclassical economic analyses in anthropology are considered as synchronic "spot transactions" of commodities, while long term interpersonal relationships enter the domain of gifts in a gifts: commodities dichotomy. The temporality of reciprocation is a means by which economic transactions are embedded inside of social calendars, but various socially-mediated methods of extending credit are well-established and probably an ancient manner of precipitating exchange in market-based transfers as well.

Pastoralism, caravans, and social inequalities

The domestication of cargo-bearing animals contributed several important elements that transformed the nature of regional exchange relationships. First, there are some cross-cultural commonalities in the structure of contemporary societies practicing pastoralism, and it is probable that these factors had some role in prehistoric pastoral societies as well. Second, cargo-bearing animals transform the costs of transport and, consequently, the nature of long-distance interaction. Finally, the structure of wealth in animal herds conveys particular scalar advantages to powerful kin-groups that possess large herds. These factors condition long-term transformations such as sedentism, the nature of food production, and the institutionalization of social inequalities. These issues will be explored in three sections: household level articulation with agriculturalists, wealth accumulation among herders that is limited by risk and pasture, and caravans and the organization of pastoral labor.

Household level articulation with agriculturalists

An important structuring principle to pastoralist exchange is that herding systems are not economically independent because humans must consume a sufficient diversity of major food groups for nutritional reasons; a condition known as non-autarkic (Khazanov 1984;Nielsen 2000). Depending on available wild plants, herders may acquire a portion of their non-animal products from gathering activities but the more common solutions involve a mixed agro-pastoral strategy or articulation with agriculturalists. Furthermore, herders with animals capable of bearing loads are the natural agents for facilitating this articulation with agriculturalists (Browman 1990;Flores Ochoa 1968). Thus, exchange relations are a basic necessity for dedicated herders. For herders with cargo animals, the transport of exchange goods in some capacity was likely a regular feature of pastoral household economies, and to become more common and less laborious in terms of quantity of goods exchanged due to assistance of cargo animals. As pastoralist households are not autarkic, due to the need for non-animal foods on a regular basis, households usually have cargo animals as part of their herd and relatively brisk exchange networks are likely to develop between households without the need for elites, administrative oversight or investment from super-family organization.

Limitations to accumulation by herders

Inequalities are evident in most pastoral societies as owners with large herds are better able to maximize by grazing all available lands when conditions allow and hiring additional help with herding tasks like shearing and butchering. Furthermore, large herds can reproduce more quickly, are better able to survive hardship, and better maintain a minimum herd size threshold for viability. Nevertheless, pastoral wealth is widely recognized as unstable due to the overall vulnerability to drought, disease, parasites, predators, theft, and accidents that can cause declines of over 50% in a given year (Khazanov 1984: 156;Kuznar 1995;Nielsen 2000: 42;Salzman 1999). Herders mitigate this risk by diversifying production, maintaining extensive exchange networks, holding access to grazing land in common, and utilizing other institutional means of risk reduction, such as the redistribution mechanism of suñay among Andean pastoralists (Flannery, et al. 1989).

A significant pastoral institution that appears to function as a leveling mechanism is the corporate ownership of pasture. While herds are typically held by individuals or kin-groups, herding is spatially extensive rather than intensive, and access to pasture in herding societies almost universally requires community negotiation (Ingold 1980;Khazanov 1984;Nielsen 2000: 46-51). Among pastoral societies that do not store fodder the carrying capacity of the land, and therefore the intensifiability of production, is limited by the season with the lowest productivity (Nielsen 2000: 43). Finally, in some regions of the world, such as the Andes, modern herds are bilaterally inherited which serves to prevent accumulation in specific descent groups (Lambert 1977;Webster 1973: 123). Thus, herding does not organizationally contain the seeds for social inequality, but intensified pastoral wealth in the form of very large herds has been documented as one principal form of investment in ethnohistorically known hierarchical societies in herding regions.

Caravans and the organization of pastoral labor

Pastoralist societies that organize into seasonal trade caravans share structural characteristics; some of these characteristics tend towards promoting social inequalities and others that counter-act the tendency (Browman 1990). The organizing of a trade caravan is often simplified among dedicated pastoralists because pastoralism is relatively efficient in the use of labor, and the herding and caravanning schedules can be prioritized (Nielsen 2000: 44-45). A single herder can monitor hundreds of animals on a typical, uneventful day of pasturing without a great deal of effort, and as physical labor is low as compared with agricultural tasks, children and the elderly often contribute and broaden the herding labor pool. As a consequence, during caravan season the loss of several capable family members (usually adult men) to the caravan journey for weeks or months during a single year may not unduly hamper the productivity of a household of dedicated pastoralists.

As mentioned above, all pastoral households must acquire non-pastoral products through diversification or exchange, however the ability to organize a caravan inherently favors the wealthier herders for several reasons.

(1) Herders with large herds are more likely to have a sufficient number of hearty animals capable of enduring long journeys with cargo.

(2) Caravan animals provide the mechanism of transport. Therefore, for direct exchange consisting of spot transactions, pastoralists must initiate the trade opportunity by traveling to their trade partners with a sufficient surplus of goods to acquire goods in exchange.

(3) The rewards of such trade caravans accrue differentially, allowing those who regularly participate in such ventures to acquire access to non-pastoral resources, a more extensive social network and perhaps fictive kin among distant trade partners, and enhanced prestige among their community.

While some elements of dedicated herding societies favor differential accumulation of wealth in the form of large herds, diversification of the resource base, and extensive trade networks, the realities of high risk to pastoral wealth and low intensity of land use affect all herding households equally to the extent that they dedicate themselves to pastoralism. Thus, while redistributive mechanisms and corporate tenure of pasture serve to stabilize pastoral systems, the structure of herding systems also provides a few opportunities for strategic advantage to more opportunistic and aggrandizing elements of prehistoric society. In addition, despite the risk in herding systems the ownership of a large herd may directly confer prestige on pastoralists (Aldenderfer 2006;Hayden 1998;Kuznar 1995: 45), and inequalities in pastoral wealth can be channeled into more enduring and intensifiable avenues such as increased exchange (ultimately resulting in large trade caravans) or a mixed agro-pastoral strategy. As a pastoral economy with cargo transport capabilities first takes hold in a region, increased social differentiation may reflect a co-evolution between more extensive exchange relationships, greater sedentism, population increases, and larger population centers.

2.2.4. Definitions of exchange

Studies of exchange can contribute to understanding human behavior because, more than any other species, humans have possessions and shift them between individuals. A person can acquire an object of wealth either by producing it or exchanging for it. In the terms used by economic anthropologists, exchange takes a number of forms in social interaction. Wealth is the objective of a person's labor and is therefore culturally determined. Markets refers to a market-based economy where prices reflect supply and demand (LaLone 1982: 300), as is not to be confused with aggregated transfers of various forms occurring in marketplaces. Exchange or trade are commonly used terms for processes referred to more generally by economists as transfer or allocation (Hunt 2002). Commodities, goods, and products are used here synonymously here and do not imply exchangeability or alienability.

Polanyi and the economics of exchange

The principal forms of economic organization outlined by Karl Polanyi (1957) - reciprocity, redistribution, and market forces - have been widely used in anthropological discussions of exchange. Paul Bohannan describes the relevance of these economic modes to exchange:

Reciprocity involves exchange of goods between people who are bound in non-market, non-hierarchical relationships to one another. The exchange does not create the relationship, but rather is part of the behavior that gives it context.
Redistribution is defined by Polanyi as a systematic movement of goods towards an administrative center and their reallotment by the authorities at the center.
Market Exchange is the exchange of goods at prices determined by the law of supply and demand. Its essence is free and casual contract (Bohannan 1965: 232).

In principle, the price-fixing aspect of market based exchange has an integrative effect and entails communication between segments of the exchange sphere. For the trader, market exchange involves risk and potential profit. For the producer, the consumption patterns and fluctuations in demand should be sensed all the way back in the contexts of production. Market articulation in land-locked regions of Asia and North Africa provided the initial contexts for large-scale caravans in the Old World. With caravans perennially under threat of robbery or obligation to pay duties for crossing sovereign land, the "nomadic empires of the Turk, Mongol, Arabic, and Berber peoples were spread out like nets alongside transcontinental caravan routes" (Polanyi 1975: 146-149). In contrast to market mechanisms, Polanyi also described exchange modes of institutional or administered trade, where material needs were satisfied through the movement of goods but the practice was not motivated by "profit" for merchants in the market sense of the term but rather to meet institutional goals (Salomon 1985: 516;Stanish 1992: 14;Valensi 1981: 5-6). In this type of trade, value and equivalencies are established by political authority or by precedent.

Characteristics of a market economy

Markets are of particular interest in this discussion because while market principles, to a certain extent, underlie all formal economic approaches in anthropology, market-based economies are far from universal in the premodern world, even among state-level societies. A basic definition of a market is "the situation or context in which a supply crowd (sellers) and a demand crowd (buyers) meet to exchange goods and services" and where the market principle is operating (Dalton 1961: 1-2). Three characteristics used by Earle to evaluate the evidence for market-based exchange in the Inka state include: (1) The importance of specialized institutions of production and exchange divorced in their operations from other institutional relationships; (2) The development of a medium of exchange to facilitate the systematization of exchange values; (3) The percentage of goods utilized by a household that are obtained by exchange (Earle 1985: 372-373).

The presence of exchange institutions, either as bustling marketplaces or distributed as "site-free" exchange houses, have a characteristic described by Earle (1985: 373) as a context where "non-exchange relationships, such as kinship and political ties, will not unduly constrain choice". The alienability of products, the strong influence of price motive and the detachment from production and other social linkages, underlie many of the features of these proposed institutions. The consensus among Andeanists is that during Inka domination, and probably during the preceding periods, market-based economies were not found in the central and southern Andes with a few exceptions (Earle 2001;LaLone 1982; but see href="/biblio/ref_2005">Salomon 1986;Stanish 2003).

The location of transfer becomes important with respect to price-fixing markets. In aggregations the public nature of the contact and the circulation of information is quite different from isolated exchanges. These issues are linked to the spatial and temporal configuration of exchange in market economies because just as periodic gatherings, central places, and rank-size geographic relationships serve to distribute goods in some settlement systems, these aggregations serve to distribution information about availability of products and changing prices to buyers and sellers (Smith 1976;Smith 1976). When exchange takes place in a private courtyard rather than a public marketplace then there is reduced risk of the neighbor overhearing the barter exchange value offered to another (Blanton 1998;Humphrey and Hugh-Jones 1992). Greater public visibility and monitoring in market contexts might be expected, and greater privacy, interpersonal negotiation, and temporal depth to exchange relationships in private barter exchange configurations.

Variations on Modes of Exchange

Some scholars have built on Polanyi's four original modes of exchange, others have developed entirely new schema (Smith 1976). Earle (1977: 213-216) argues that Polanyi's (1957: 250) definition of redistribution as "appropriational movements towards a center and out of it again..." is vague and Earle observes that this definition is so broad that it could to apply to economic systems ranging from central storage of goods in Babylonia to meat distribution in band-level hunters. Earle advocates separating leveling mechanisms from institutional mechanisms, where institutionalized redistribution involves wealth accumulation and political transmission between elites across broad regions in the mode of peer-polity interaction (Earle 1997).

Andean political economy

Stanish (2003: 21) expands on Polanyi's system by describing political economy in the prehispanic Andes with deferred reciprocity taking the form of competitive feasting and political support (Hayden 1995;Stanish 2003: 21). Stanish observes that while there was an implicit or explicit (Service 1975) evolutionary sequence going from reciprocity to redistribution and finally to markets, more recent evidence suggests that these modes can co-occur and that the relationships are too complex to collapse into a single sequence.

Types of reciprocity

Sahlins (1972: 194-195) further elaborated on aspects of Polanyi's reciprocal mode with generalized, negative, and balanced reciprocity. Generalized and negative reciprocity are opposite ends of a continuum (see Figure 2-1, above). Generalized reciprocity refers to sharing, altruism, and Malinowski's "pure gift", while negative reciprocity is the attempt to maximize personal gain from the transaction through haggling or theft (Sahlins 1972: 195-196). Polanyi's basic modes of exchange have persisted in economic anthropology for almost fifty years. Some argue that Polanyi's modes of exchange are limiting in that they do not provide a means to analyze precapitalist commercial activity (Smith 2004: 84), however the benefit to Polanyi's exchange modes is that they are sufficiently general to be comparable cross-culturally and the three modes are discrete enough to be, in some cases, archaeologically distinguishable. Furthermore, if commercial activity is unlikely in the study region, as in the prehispanic south-central Andes, Polanyi's modes capture the necessarily economic variability.

Geographical characteristics

Renfrew (1975) considers trade as interaction between communities in terms of both energy and information exchange. Renfrew (1975: 8) tabulated Polanyi's schema as follows







No Central Place





Central Place

Central Organization


Table 2-3. Characteristics of reciprocity and redistribution (from Renfrew 1975: 8).

Renfrew follows with an exploration of the greater efficiency implied by central place organization in terms of material and information exchange. The universality of central place organization in the development of complex political organization worldwide has been called into question in pastoral settings. In the south-central Andes, anthropologists have proposed that alterative paths to complex social organization could have been pursued by distributed communities linked by camelid caravans with the anticipated central place hierarchy not occurring until relatively late, as proposed by Dillehay and Nuñez (1988) and in a different form by Browman (1981).

Renfrew has developed a graphical representation of the spatial relationships implied by each mode of exchange.


Figure 2-2. Modes of exchange from Renfrew (1975:520) showing human agents as squares, commodities as circles, exchange as an 'X', and boundaries as a dashed line.

The exchange modes depicted by Renfrew (1975:520), shown in Figure 2-2, efficiently convey the variety in organization represented by exchange relationships. In some regions of the world, such as the prehispanic south-central Andes, market-based economies are not believed to have operated which modifies one's expectations for the activities of traders. The full suite of these ten modes is not expected in any one particular archaeological context worldwide, but the figure serves to underscore the complexity of isolating particular types of exchange based on archaeological evidence. Furthermore, these modes are not necessarily mutually exclusive, as some of these modes may have been operating simultaneously unless restrictions on production, circulation, or consumption of goods were in place.

Exchange network structures

Based on geographer Peter Haggett's (1966) work, network configurations can be used to describe the characteristics of interaction that resulted in the distribution and circulation of goods. In describing Andean caravan transport, Nielsen (2000: 73-74, 91) uses the following terms: (1) distance that goods are transported; (2) segmentary vs. continuous - in segmentary networks a given node is connected to a small number of other nodes, while continuous networks each node is connected to all other nodes; (3) convergent (focalized) vs. divergent (non-focalized) - in convergent networks the individuals participating in exchange, and the goods they transport, tend to concentrate in a small number of central places or exchange locales.

Reciprocal exchange relationships that take the form of down-the-line trade may be described as continuous networks when these mechanisms serve to move goods between ethnic groups and across regions. Centralized political control by elites and true market mechanisms might result in network convergence at central places (Smith 1976). To Nielsen's third set of terms, convergent (focalized), divergent (non-focalized), one may add diffusive to describe the pattern of a single type of item radiating from the center to the surrounding region as occurs with obsidian.

Figure 2-3. Network configurations.

These network configurations serve to draw attention to the limitations of using raw material distributions as a proxy for all exchange behaviors. Obsidian exchange is sometimes used by archaeologists as gauge of the volume, frequency, and structure of prehistoric exchange relationships. As noted by Clark (2003), raw materials from geological sources diffuse continuously from a single point to the region, presumably following trade routes, until the materials are found deposited at archaeological consumption sites. In contrast, much exchange between complementary groups, such as between agriculturalists and pastoralists, is non-focalized and often segmentary, as it links producers and consumers through a variety of localized articulation methods. The structural differences between diffusive exchange networks and regular, household-level interaction are well demonstrated in mountain regions with distinct ecological zonation, such as the Andes.

Archaeological inferences that do not differentiate the expectations of one network configuration from another are problematic. Often, diffusive configurations will have evidence of the transport of goods in the opposite direction, as one might expect in system where obsidian is acquired through reciprocity-based relationships. However, the pattern where goods are reciprocated to the source area is a configuration model to be tested rather than one that can be assumed.

2.2.5. Exchange and social distance

As discussed by Robin Torrence (1986: 5), exchange is not directly observable but requires interpretation of the evidence found in consumption sites and in the initial procurement and production areas for artifacts.

Figure 2-4. Model for inference about prehistoric exchange (from Torrence 1986: 5).

Seldom does the act of exchange leave direct evidence of having occurred in a particular location, and the activity must be inferred from the circumstances surrounding production, exchange, and consumption of the product. Note that "Acquisition" (Figure 2-4), or the source area for a product, links directly to all other modes except discard and re-use. In other words, as observed by Torrence, quarry areas are in a unique position for investigating a complete exchange system because it is only "Acquisition" at the quarry area that articulates in some form with most of the major nodes in the Figure 2-4 conceptual model.

When archaeologists encounter non-local materials in their studies, there are commonly three alternative interpretations for this evidence of contact: (1) migration, (2) trade or exchange, (3) conquest by a non-local group. Differentiating these forms of contact from archaeological consumption data can be difficult, and a larger view of the context of exchange is required.

In the 1970s when exchange studies were being widely discussed by archaeologists, two principal approaches were adopted: (1) The system-level view, presented by Renfrew and his associates (1969;1972;1975), and, (2) the political or social view of trade relations (Adams 1974;Friedman and Rowlands 1978;Kohl 1975;Tourtellot and Sabloff 1972) that became more prominent in the 1980s. The systems-oriented approach integrates data into a comprehensive framework, but it is weakened by gradualistic and adaptationist underpinnings as it is

...assumed to have a smoothly, internal inevitability of its own... however it is absurd to think of this as the path that at least the more complex societies have normally followed. They dominate the weaker neighbors, coalesce, suffer themselves from varying forms and degrees of predation, develop and break off patterns of symbiosis - all in dizzyingly abrupt shifts (Adams 1974: 249).

As more recent, agent-centered analyses argue, exchange is a dimension of society that is particularly susceptible to the ambitions of entrepreneurs or aspiring elites because expressions of non-local association and alliance are one manner in which social differentiation can be achieved (Appadurai 1986:38;Clark and Blake 1994;Wiessner 2002: 233).

Exchange across boundaries

Anthropological accounts relate that acquiring resources through trade with neighbors is sometimes a dangerous undertaking that is maybe not far removed from raiding and warfare. "There is a link, a continuity, between hostile relations and the provision of reciprocal prestations. Exchanges are peacefully resolved wars, and wars are the result of unsuccessful transactions" (Lévi-Strauss 1969: 67). Further, Sahlins (1972) observes that reciprocity in trade between unrelated exchange partners in contexts without overarching political control can be a very delicate affair because sometimes it is only the perception of fairness in the exchange that maintains peace.

When people meet who owe each other nothing yet presume to gain from each other something, peace of trade is the great uncertainty. In the absence of external guarantees, as of a Sovereign Power, peace must be otherwise secured: by extension of sociable relations to foreigners - thus, the trade-friendship or trade-kinship - and, most significantly, by the terms of the exchange itself (Sahlins 1972: 302, emphasis in the original).

Yet exchange and markets across boundary areas are common features in world history. Such exchange could consist of "border mechanisms" with trade-partnerships or fictive-kin ties that permit "interactions across tribal boundaries under conditions of peace and personal security" (Harding 1967: 165). Zones between social groups are particularly likely to have active trade and markets if there is some cultural or ecological variation between the groups, such that the products circulating in each zone complement one another (Hodder and Orton 1976: 76). The No-man's-land described as the "Intertribal sector" by Sahlins (1965;1972: 196-204) is a good locale for such exchange because complementary products are available for exchange, and because the morality of exchange in neutral geographical territories permits balanced or negative reciprocity between traders. In perception of value, as well as geography, these border zones can be conceived as overlapping social "spheres of interaction" as described by Barth, where "entrepreneurs will direct their activity preeminently towards those points in an economic system where the discrepancies of evaluation are greatest, and will attempt to construct bridging transactions which can exploit those discrepancies" (Barth 1967: 171). These discussions of the details and great the variability in entrepreneurial strategy anticipate the challenges of an agent-based approach to ancient trade (Adams 1974: 243;Hodder 1982).

Exchange and social distance

The question of the isolation of producers and traders from consumers also connects theories about exchange with the issue of the physical form of the exchange goods. Archaeologists investigating the role of stone artifacts in prehistoric economies observe that lithic manufacture is a subtractive technology as stone artifacts always get smaller with use and maintenance. The degree of reduction of stone material at a lithic source determines the kinds of forms that subsequent artifacts will take.

Jonathon Ericson (1984) applies Sahlins' concept of a continuum of social distance in exchange relationships, discussed above, to the directional, reductive nature of lithic production systems. He explores the idea that the degree of lithic reduction that may occur could be reduced when social distance increases because the producer would have less information about the consumer and the end forms that material will take (Ericson 1984: 6). For example, if the procurer does not know if the nodule will be formed into bifacial lancoleate knife or a triangular projectile point by the consumer, it would be better to leave the nodule in a larger form.

Exchange partners can be slow to respond to changes in the needs of consumers in a given exchange system (Ericson 1984: 6;Harding 1967;Rappaport 1967;Spence 1982), and the effects of social distance can impact production, exchange, or consumption patterns. The implications of social distance for lithic reduction are that, as a subtractive process, reduction circumscribes the potential artifact forms that a nodule of raw material may take in the future. Countering this tendency, people can reduce risk in tool production by producing blanks closer to the source of a raw material where the value of a material is lessened and the costliness of knapping error or breakage, and inconsistent or poor-quality material, is reduced. Stylistically, producers may wish to impart a local motif to the material; alternately, in order to maximize distributive potential, a good may be left in a minimally reduced form.

The exotic and value creation

Non-local exchange goods are prominent in anthropological models of socio-political change because exchange goods can accrue value directly as a function of scarcity, labor input, or through social and symbolic reference. As discussed above under the subject of practical goods and prestige goods, a given object may move between practical and prestige categories in different places, times, and social contexts. The availability of an item in a given milieu communicates not only the relative scarcity but, for alienable goods, the exchange value of that item; such items may also contain allusions to distant regions, social groups, and esoteric knowledge.

Theoretical models assert that in order for sacredness or exotic power to be conferred through possession of non-local goods, those goods cannot be widely available or mutable in economic circles accessible to just anyone (Clark and Blake 1994;Goldstein 2000). The possession and circulation of these goods have also been considered as part of a network strategy, distinct from a corporate strategy, towards acquiring influence and leadership (Blanton, et al. 1996). These goods may have served as indicators of long-distance association for trade and alliance, and also have served as a means of differentiation during this time of incipient political competition. In another approach, one that focuses on differential reproductive success, Craig and Aldenderfer (in press) use costly-signaling theory in a formal, biological adaptationist framework to model the development of social inequalities through the differential use of obsidian in southern Peru at the Archaic and Formative transition. Exotic materials have been used to demarcate commonplace from supernatural referents, or are at least part of a constellation of behavior and objects that signal status difference.

A pattern noted frequently by archaeologists is that close to the source of a raw material there is no distinction associated with the commodity as the item is abundant, whereas farther from the source, where access is intermittent, the possession of such commodities may acquire greater symbolic importance (Knapp 1990: 161;Renfrew 1986). It follows that if one moves from a place where a product is scarce and found in ritual contexts, towards the source of that product such that it becomes less scarce, one may observe a reduction in ritual or exclusive association for that group of goods. This theme will be considered with obsidian use in the Lake Titicaca region.

Furthermore, what of those who transport exotic goods? Mary Helms argues that "we should consider long-distance travelers or contact agents as political-religious specialists, and include them in the company of shamans, priests, and priestly chiefs and kings as political-ideological experts or 'heroes' who contact cosmically distant realms and obtain politically and ideologically useful materials therefrom" (1992: 159). These agents are in a position to benefit, in an entrepreneurial way from the value difference, between the source and the consumption zone, but in many cultural contexts their participation and social roles appear to be circumscribed. The association of non-local goods with status or prestige, long-distance alliance or esoteric knowledge is contingent on a variety of factors upon which it is difficult to generalize, but one can examine these archaeologically through artifact form and context.

2.2.6. Territoriality and access to raw material sources

Exchange and social distance have particular configurations when they occur in contexts of unusual raw materials. The spatial dependence of procurement, distribution, and consumption on access to particular source locations creates a context where social distance may correspond directly with procurement and consumption patterns. There is broad cross-cultural variation worldwide in territoriality and access to raw material sources. The effort and the benefits associated with territorial circumscription and resource control are frequently considered in the context of specialized production. Resource control has been discussed for circumstances where competition for a resource can lead to an attempt to "monopolize" access (Torrence 1986: 40-42). Following formal economic principles, the value of a good should be a function of its availability where value would escalate as a result of restricted access and limited supply (Brumfiel and Earle 1987: 7), a process that has been proposed for shell bead exchange among the Chumash of California (Arnold 1991). The variety of territorial control strategies documented worldwide invites a broader consideration of the diverse boundary negotiations that may have been occurring at a raw material source through prehistory.

Territoriality and resource procurement among foragers and pastoralists is a topic that has been explored in anthropology in recent years (Cashdan 1983;Casimir and Rao 1992;Kelly 1995). Approaches range from ecological models based in optimal foraging theory and site catchment analysis, to organizational models that include perimeter defense and social boundary demarcation, and finally to process-oriented models that consider the effects of sedentism, circumscription and population pressure. Carolyn Dillian (2002: 95-116) reviews the issue of territoriality and obsidian procurement using more contemporary theory.

California multiethnic access

One possible arrangement is multiethnic procurement at a geological source. Ericson (1982: 136) describes a situation in California's Napa Valley where the Saint Helena obsidian source fell within the territory of the Wappo ethnic group and the Wappo were, in turn, surrounded by the Lake Miwok, Coast Miwok, Wintun and Pomo groups (Kroeber 1925). It is reported that in exchange for obsidian, the Wappo received items such as bows, beads, shells, mats, fish, headbands, and clams (Davis 1961), but it is not known if extraction was conducted by the local Wappo or through direct access by the surrounding groups themselves. Ericson (1984: 7) reviews evidence of multiethnic access to lithic sources. In California there are a number of distinctive obsidian sources that have been studied. At some areas the quarry management was "tribal but related and nearby groups had the right to quarry either freely or on the payment of small gifts. Wars resulted from attempts by some distant tribes to use a quarry without payment. On the other hand, the Clear Lake obsidian quarries were neutral ground" (Bryan 1950: 34).

The Wintun of California practiced round-trip fasting when traveling to obsidian sources (Dubois 1935). A functional interpretation of this behavior from Ericson (1984: 7) is that fasting would avert exhaustion of resources around much-visited obsidian sources. A regionally important quarry could conceivably get so much use, with many groups exploiting the source, that the ecology of the source area catchment would get depleted.

Geochemical studies of obsidian artifacts in California have revealed a concordance between obsidian distributions, as shown by chemical studies, and the geographical boundaries of ethnographically documented cultural groups (Bettinger 1982;Hughes and Bettinger 1984;Luhnow 1997). Dillian (2002: 294-297) found that despite ethnographic accounts of Karok direct procurement at the Glass Mountain source in Modoc territory, the knapping evidence at the source suggests that local Modoc were conducting virtually all of the reduction on site and then exchanging with neighboring tribes where it was widely circulated.

Other Source Areas

The well-known red pipestone quarry in Minnesota "was held and owned in common, and as a neutral group" (Holmes 1919: 262). In the Western Desert of Australia, Gould et al. (1971) report that chert and chalcedony sources themselves are not held through a concept of quarry ownership. All material of usably good knapping quality is equally valued, and knapping is not a skill that is assigned great importance. However, an important totemic affiliation exists between a person and stone from the region in which they were born. Cherts from a person's ancestral region are sometimes visually distinct and therefore materials of a particular region will be sought and transported over long distances as a physical link to those regions.

These preferences appear to be a reflection of the close totemic ties each man has to the particular region in which he was born and from which he claims totemic descent. Thus, a man may have a sense of kinship with some of these localities, and he will value the stone material from them as a part of his own being. Stone materials thus acquired are not sacred in any strict sense but are nevertheless valued highly enough to be transported over long distances by owners (Gould, et al. 1971: 161-162).

Ideological and emotive links to raw material are a consideration in human activities around source areas. Social and symbolic restrictions on quarries have also been documented ethnographically in New Guinea (Burton 1984;McBryde 1984). In terms of exchange behaviors, quarry areas present special problems and opportunities for archaeologists. One of the principal difficulties with examining ethnicity and access to quarries is correlating archaeological evidence with social and symbolic behavior associated with quarries. Some theoretical models are contingent on measuring evidence of maximization and control of production at quarry areas, and these models are often contingent on the detection of boundaries and restricted access based on material evidence (Torrence 1986). Unfortunately, as demonstrated by some of the studies above, many social and symbolic limitations on quarry access leave no direct material correlates and can be extremely difficult to detect archaeologically.

2.2.7. Discussion

While formal and substantivist economics have been used to investigate regional interaction and exchange worldwide, this dissertation research follows on the substantivist tradition in Andean archaeology. The exchange issues explored above can be summarized in following three themes.

Exchange value

Exchange from the perspective of commodity 'exchangeability' and demand by consumers is a cross-culturally comparable (and often archaeologically detectable) means of assessing value, but this approach depends on the goods actually circulating. Others have observed that some inalienable objects (heirlooms) are valued precisely because they do not circulate, and the ability of an object to "accumulate history" is another means of establishing value albeit a measure that is difficult to establish archaeologically.

Social distance

The continuum of social distance is useful in that it captures the role of different behavior and institutions in exchange as one moves away from the household, and it may parallel formal models of kin selection. Further, the concepts of production and social distance for a commodity like obsidian can be empirically linked to expectations about the degree of lithic reduction as one moves from production to consumption contexts. The use of the social distance concept to position a diachronic, "primitive" household (substantive) exchange against a synchronic, "modern" and commercialized (formal) realm is a problematic and false dichotomy. Virtually all exchange contexts contain elements of both social contracts and economical behavior. Social distance, territoriality, and access to products can be manifested in a variety of ways that range from the organization of technology, socially circumscribed access, and symbolic restrictions; a situation that poses difficulties for archaeologists attempting to establish the relative accessibility of a particular product prehistorically.

Social and political consequences of exchange

Exchange is a mechanism that brings goods and people together between ecological zones and across social boundaries, and exchange creates strategic opportunities for individuals and institutions. Exchange is sometimes used to reinforce status differences because the possession of exotic goods, and the necessary surplus to acquire non-local products, differentially favors those with established regional ties. However exchange across boundaries, along with warfare and ritual, can be dangerous and may be the domain of strategic and opportunistic individuals as traffic in the exotic, and contact with foreign elements, is liable to challenge the social order. Conversely, one of the often-noted social consequences of regular exchange is more mundane: exchange serves to reinforce long distance social ties over time, to buffer risk, and to express cohesiveness through common access to distinct resources.

In recent decades, exchange studies in archaeology have acquired new technical rigor with advances in chemical proveniencing. While the cultural, institutional, and theoretical ramifications of exchange remain complex and nuanced, the demonstrable fact of chemical characterization offers refreshing certainty to the otherwise conditional and qualified study of ancient exchange.

2.3. Chemical provenience and exchange

In the early twentieth century archaeologists investigated prehistoric exchange, primarily by using stylistic criteria, in order to demonstrate contact between two culture areas from a diffusionist perspective. Modern studies of exchange through scientific sourcing began with the work of Anna O. Shepard (1956) who found that petrographic analysis of ceramic tempers could be used to differentiate ceramic types. Since Shepard's early work, methods for chemically characterizing artifacts has grown rapidly (Glascock 2002). By the late 1960s, with geochemical evidence for long-distance interaction accumulating worldwide, archaeologists began developing systematic approaches to evaluating exchange.

The fundamental issue for many archaeological studies of exchange is that exchange processes have social evolutionary consequences (see Section 2.2.3 ). In current debates over the role of exchange, theoretical approaches range from those premised on looking at efficiency in production and exchange, and those approaches that focus on the role of social dynamics and agency models for long-distance relationships.

2.3.1. Quantitative approaches to regional exchange

An influential approach to long-distance exchange was developed by Colin Renfrew and his colleagues (Renfrew, et al. 1968) who asserted that the spatial distributions of a raw material like obsidian could be used to infer not only extent of interaction, but mode of exchange. In subsequent investigations these distance decay relationships were further explored and in 1977 Renfrew defined the Law of Monotonic Decrement (LMD):

In circumstances of uniform loss or deposition, and in the absence of highly organized directional (i.e., preferential, nonhomogeneous) exchange, the curve of frequency or abundance of occurrence of an exchanged commodity against effective distance from a localized source will be a monotonic decreasing one (Renfrew 1977: 72).

Here, and in other publications, Renfrew (1975;1977) and Hodder (1974;1978;Hodder and Orton 1976), sought to interpret exchange relationships and the friction of distance from the shape of "fall-off curves" where the abundance of material is plotted against cost, usually distance from the source. The novelty in this approach is that it sought to determine "types" of exchange that substantive anthropologists had placed in evolutionary sequence using explicit, formalist measures of abundance and cost. These distance decay graphs included "Down-the-line" exchange thought to represent reciprocity, "freelance" trade representing barter, and even laissez-faire capitalism. More general and robust characterizations were described as well. For example, low value, often cumbersome goods were shown to have different distance decay profiles than prestige-goods exchange (Hodder 1974;Hodder and Orton 1976: 124).

Other scholars adopted this approach and in places like Mesoamerica the method held promise because obsidian sources were abundant, and material from many sources were, for the most part, visually distinct (Braswell, et al. 2000). Raymond Sidrys developed a "Trade Index" that, he argued, showed that major ceremonial centers acquired obsidian in volume from greater distances.


Figure 2-5. Log-Log fall-off curve of obsidian density (grams of obsidian/m3 of fill). Line 1 is a regression line derived from major ceremonial centers while line 2 is derived from minor centers (Sidrys 1976: 454).

The axes of the fall-off graphs are on linear, power or logarithmic scales. Depending on the data available, interaction can be measured by using absolute or relative measures. Abundance of a non-local good at site A is assessed in absolute terms using weight of material from site A divided by estimated population for site A, or abundance is measured in relative terms using percentage by weight or artifact count from site A in the total raw material class for that site (Earle 1977: 6;Renfrew 1977: 73).

In his work with Neolithic Near Eastern obsidian Renfrew (1969: 157) specifies that areas less than 300 km from the geological source area, at least for Neolithic modes of exchange, are in the "supply zone" because obsidian represents over 80% of the material in lithic assemblages at sites found in that zone. Areas beyond the 300 km band Renfrew refers to as the "contact zone" and it is in that area that he argued the shape and angle of the fall-off curve could provide insights into prehistoric economy. The "interaction zone", or universe of study, for a sourceable material was defined by Renfrew to be the area within which 30% or more of the obsidian was derived from a single obsidian source. A similar approach was taken with Mesoamerican obsidian in the Oaxaca area by Jane Wheeler Pires-Ferreira (1976: 301) but in this case a 20% threshold was used to define the interaction zone. There appears to be no ethnographic basis for the source, contact, or interaction zone threshold values used by Renfrew and others.

In the course of further exploration of the parameters associated with fall-off curves, serious weaknesses were identified that limit their utility for identifying forms of exchange solely on the basis of the shape of the curve. Hodder (1974;1978;Hodder and Orton 1976: 127-154) simulated a large number of simple random walks and found that generally similar fall-off curves could be produced by different combinations of variables, a condition known as equifinality. In other words, a Gaussian artifact distribution that results in a fall-off curve that Renfrew would have described as "down-the-line exchange" could as have been the result of random walks. Hodder and Orton (1976: 142-143) found that the more convex curves have higher a values and that these are the result of a greater number of short steps usually associated with highly portable value goods. Ammerman et al. (1978: 181-184) explored the fall-off curve with simulation studies and found that in the interpretation of down-the-line models one must consider the accumulated effects of time over the long-term. Further, they argue that accumulation rates in archaeological studies of down-the-line systems can modeled by using realistic estimates of "passing" and "dropping" of artifacts in distribution systems. Critical evaluations of the utility of fall-off curves demonstrate that Renfrew's goal of distinguishing reciprocity, redistribution, and market exchange from two-dimensional graphs was overambitious. It has been shown, however, that the relative shape of fall-off curves can aid in differentiating high value commodities from bulky, utilitarian goods. These graphs are valuable for highlighting variations from the LMD, and these deviations can point towards avenues for further data exploration. Hodder and Orton (1976: 155-160) suggest applying trend surface analysis to distribution data, an approach that was implemented in the 1970s raster-based computer mapping package SYMAP. Trend surface analysis and geostatistics have become considerably easier using modern GIS methods.

2.3.2. Other Distance Decay studies

Approaches to the study of attenuation with distance were applied in a variety of other regions worldwide, and these studies have been ably summarized by Torrence (1986: 10-37). A valuable theme in these studies has been the exploration of deviations from Renfrew's Law of Monotonic Decrement, and suggestions for improvement of distance decay models. Gary Wright (1970) proposes that predictions made by the LMD could be improved by controlling for variations in consumption through time and between types of consumer sites, and he emphasizes that investigators consider the influence of alternative materials in the study area, such as the effects of local flint on demand for imported obsidian. Wright (1970;Wright and Grodus 1969: 47-52) also advocates using weight of a non-local good as a measure of abundance, rather than count or percentage of each artifact class. For time periods prior to the domestication of beasts of burden in the Near East, when the weight of artifacts was borne directly by human carriers, weight would more likely influence behavior and discard patterns. Renfrew also considers weight in his analysis, but he observes that weight would be influenced by stylistic and functional factors.

Geographical relationships are considered in more detail in subsequent research. Jonathon Ericson (1977;1977;1981;1982) conducted an exhaustive study of California obsidian distributions and, by focusing on deviations from the LMD, and using regression analysis, he was able to assess the strengths and weaknesses of the approach. Ericson uses trend surface analysis available in the early raster-based mapping system SYMAP to produce maps showing isometric distributions of obsidian types in consumption sites (Ericson 1977: Figures 1-4). In comparing this form of analysis with Renfrew's distance decay graph distributions Ericson states that "in two-dimensional analysis only the magnitude of an observation and its distance from a source is considered, the spatial position of the observation is not considered in its local context; and this simplification masks significant variability in the data" (Ericson 1977: 110). Ericson's density maps demonstrate that, while distance from source is a primary determinant of obsidian type, the distributions of obsidian consumption locations are not symmetrical around the source areas. Ericson explores the spatial relationships by superimposing trail systems, alternative material distributions, and ethnolinguistic group boundaries (Kroeber 1925) on the obsidian consumption density maps and finds that these spatial phenomena influence obsidian distributions.

Similar to Ericson's implementation of trend surface analysis, Findlow and Bolognese (1982: 60-70) perform a SYMAP analysis using the percentage of lithic assemblages represented by obsidian throughout their study region. Their maps are useful in that they show the changing territoriality and direction of obsidian procurement through time in the region. However, as their obsidian percentage isolines only display aggregated obsidian versus not obsidian data, and do not differentiate between source types of obsidian, the maps are difficult to interpret in terms of exchange distances from the sources through time.

Quantitative data available for Ericson's (1977: 121-123;1977: 249-257) analysis include isotropic distance calculations between discard locations and geological source areas, and estimated population by consumption area. A multiple linear regression analysis using the percentage of obsidian from a single source showed that, as stated by the LMD, the distance from the source had the highest predictive power. However, the estimated population in a given consumption area had only slightly less power in predicting the source of the obsidian being used than did distance to the closest source, and the distance to the second closest obsidian source had virtually no predictive power. Ericson interprets these data in terms of the degree of utilitarian use of a commodity and level of necessity for the average person in the community. Thus, in the proximity of the geological source of a raw material widely used in the community, the source types with high population levels in the immediate vicinity of the source would have been more widely used in exchange systems (Ericson 1977: 120).

An improvement in the calculation of effective distance with an incorporation of the influence of topographic relief in the cost accumulation with fall-off curves was first explored by Findlow and Bolognese (1980;1982). In their study in the U.S. Southwest the authors manually develop the linear solutions known today in GIS as "Least-cost paths" or, as Findlow and Bolognese (1980: 239) express it, "the line between the site and the source that at once minimized distance and topographic relief". They find that when these paths are used in the cost function then variability is accounted for more strongly than when using an isotropic distance estimate.

The technology of transport can dramatically alter the effective distance, and changes in mode of transportation have been proposed as explanations for variability in distance decay curves through time (Hodder and Orton 1976: 113,117-118;Renfrew 1977: 73). Torrence (1986: 122-123) reviews the issue and she also discusses boat transport (1986: 135-136). The adoption of river and sea-going vessels in Mesoamerica are discussed by Sidrys (1977: 103-105), and in the Mediterranean by Ammerman (1979;1978). Gary Wright (1970) mentions the importance of considering the weight of transported material for the period before the domestication of cargo animals and the availability of caravan trade networks in the Near East. Similarly, working in the Andes, Richard Burger and his co-authors (2000: 348) consider the impact of camelid domestication and llama caravan networks on obsidian distributions.

The prediction, following the LMD, that artifact size or weight should diminish with distance from the source area (Wright 1969: 47-52;Wright 1970;Wright and Grodus 1969) is not necessarily supported in cases where the form of the artifact, such as projectile points or bladelets, take precedence over effective distance from source. Angela Close (1999) found that at early Neolithic sites in southwestern Egypt flint backed bladelets, even very close to the source, were produced to be very narrow due to hafting requirements, defying the prediction of LMD. Close (1999) also found that among the late Neolithic sites unretouched, unbroken flint debitage did not conform to the LMD either, a situation that she attributes to the probable use of domestic cattle in transporting flint cores (Close 1996). With formal tools, such as projectile points, artifact form may take precedence over size. Hofman (1991) found with Folsom points that did not diminish in size with distance from the raw material source. However, incidence of resharpening can be expect to increase with distance from the source, affecting tool size (Andrefsky 1994).

The gravity model was another approach borrowed from geography and applied to the study of regional artifact distributions (Chappell 1986;Hallam, et al. 1976;Hodder 1974;Renfrew 1977: 87-88). Gravity models are used in situations where artifacts made from raw materials from a number of competing sources are found in a given consumption site. The claim is that the approach quantifies the "attractiveness" of a given material type over other types available to the consumers in a site by comparing actual proportions with expected proportions of the material given the predictions of the LMD. However, as Torrence (1986: 27) observes, "in reality, the gravity model is merely a means for describing and comparing distributions which are already reasonably well documented." While gravity models could be useful in parsing complex temporal patterns in raw material use in production and consumption at a given site, the approach does not provide a means of inferring the character of exchange relationships between consumption sites in prehistory.

Throughout the 1970s, the regional exchange literature shows an increasing awareness of the limitations of a purely spatial approach to inferring modes of exchange through formal geographical regression analysis. Renfrew's ambitious models linking geographical distance-decay with substantive modes of exchange connected to an associated evolutionary socio-political level, were meant to be sufficiently general to operate through time and in different cultural contexts. However, from simulation studies and from archaeological applications, researchers observed that it was necessary to incorporate supplementary information along with the geographical data on long-distance exchange and consumption, in order for distance decay studies to be useful. The view that chemically derived exchange data must be considered in association with information about archaeological context was first expressed by Wright (1969) and was widely echoed in the contributions to Ericson and Earle's (1982) volume "Contexts for Prehistoric Exchange" as well as in later volumes on exchange including Ericson and Baugh (1993) and Baugh and Ericson (1994). Preserving the strengths of geographical analyses from the 1970s period of investigation and combined with a greater consideration of artifact form, regional variation in reduction strategies and site-specific contexts of consumption, offers a fruitful way forward.

2.3.3. Site-oriented studies of exchange

As obsidian moved farther from its source, the size of the pieces traded progressively decreased and the relative value increased. ...Small blades were obtained simply by smashing a large block with a stone, while in time blades were broken into smaller fragments to obtain newly sharp edges (Harding 1967: 42).

Establishing consistent links between the types of social structure and the forms and organization of production in association with exchange was the approach taken by Ericson (1982) and, most explicitly, by Torrence (1986). By investigating the standardization and error rates in blade reduction strategies at obsidian production sites on the Greek island of Melos, Torrence was able to evaluate the degree of specialization involved in the quarrying and production of obsidian at the source.

In her review of archaeological site-oriented studies of exchange, Torrence (1986: 27-37) uses a few major themes to characterize site-oriented investigations. These themes include measures of abundance, source composition percentages, and the variability in archaeological context and artifactual form of import.

Measures of abundance

Investigators have compared changes in the abundance of non-local material with domestic goods that are assumed to represent population. The ratio of the weight of obsidian to volume of excavated dirt (as shown in Figure 2-5) or as a ratio of weight of ceramics (as a proxy measure for population) has been used in a number of studies in Mesoamerica (Sidrys 1976;Zeitlin and Heimbuch 1978: 189). In studies of Near Eastern obsidian exchange Renfrew (1969;1977) develops indices for the presence of obsidian based on counts and weights per phase and per excavated cubic meter, and he also uses the count of obsidian artifacts as a percent of the total lithic assemblage count. Renfrew qualifies his conclusions due to a lack of consistency in the data, but he estimates that the total quantity of to arrive at the site was relatively small and he observes a decrease in obsidian at Deh Luran sites. Working at the Olmec site of San Lorenzo Tenochtitlan, Cobean et al. (1971) develop an index that compared the number of obsidian flakes, blades, and total debitage with grinding stones and slabs for each phase at the site, with the abundance of grinding equipment serving as an estimate of number of households. The authors interpret the increasing index of obsidian to grinding slabs as evidence of a gradual rise in "prosperity" for individuals at the site, though they only briefly explore the implications for changing exchange relationships (Cobean, et al. 1991).

In Torrence's examination of the aforementioned studies she concludes that "the major difficulty with studies of consumption based on measures of resource abundance is that they lack the necessary linking arguments between patterns of consumption and type of exchange" (1986: 28-30). She states that what is needed for site-level studies of abundance and exchange is "a series of arguments describing how resource use will respond to specific types of exchange" akin to the explicit connections that Renfrew developed on the regional scale between distance decay and forms of exchange.

Site level composition from multiple sources

When raw material is available to consumers from several competing sources, investigators have used the relative quantity of materials from the different sources at individual sites to gain insights into ancient exchange. Broad exchange relationships are frequently inferred from the presence of obsidian from a distant source. When material from several rival obsidian sources are represented, either strongly or weakly, in different phases at a site, the cause of the changes in representation is often attributed to shifts in the geography of regional political or economic relationships. Geographical explanations are invoked when there is a clear deviation from the LMD, and yet evidence of significant in-situ political change is not found in investigations among studies in Mesoamerica (Zeitlin 1978: 202;Zeitlin 1982) and the Near East (Renfrew 1977: 308-309).

Changes in the mechanisms of exchange, rather than simply geography, are attributed to changes in the source composition from different sources when socio-political changes are perceived by investigators. Studies have inferred redistribution occurring when sites that act as central places (Christaller 1966 [1933]) at the top of the settlement hierarchy have disproportionate quantities of non-local materials irrespective of their distance from the source. This phenomenon has been discussed for Tikal (Moholy-Nagy 1976: 101-103;Sidrys 1977).

When sourcing studies are conducted at the scale of the household unit, though it is often labor intensive and costly to conduct proveniencing studies to an extent that are statistically meaningful, it is possible to discern convergence patterns that can be connected to exchange mechanisms (Pires-Ferreira and Flannery 1976;Santley 1984;Torrence 1986: 35). The most common application of evidence of variability between households in source composition is to infer redistribution from the presence of low inter-household variability in source composition.

In a reciprocal economy where individual households negotiate for their own obsidian, we would expect a good deal of variation between households, both in the sources used and the proportions of obsidian from various sources. Conversely, in an economy where the flow of obsidian is controlled by an elite or by important community leaders, who pool incoming obsidian for later distribution to their relatives, affines, or fellow villagers, we would expect less variation and more uniformity from one household to another (Winter and Pires-Ferreira 1976: 306).

The concept is been depicted in a graphic from.


Table 2-4. Household composition of raw materials should vary with different types of exchange. On the left, individual households acquire source materials more directly, on the right pooling and redistribution results in greater inter-household consistency in raw material composition (Winter and Pires-Ferreira 1976: 311).

This approach was influential in Mesoamerica (Clark and Lee 1984;Santley 1984;Spence 1981;Spence 1982;Spence 1984 ) where the technology of transport did not change significantly except for perceived changes in the form of boat technology. Transport changes could significantly impact the inter-household diversity of source composition in places, such as the Near East, where caravan exchange networks followed from the domestication of pack animals (Wright 1969). The impacts of caravan exchange on artifact variability has long been discussed in the Andes (Dillehay and Nuñez 1988;Nuñez and Dillehay 1995 [1979]), although household-level sourcing data has not been available to date.

Variation in site-level contexts and artifact form

Intrasite variability in lithic distributions, and the morphology of those artifacts, can shed light on exchange patterns. Several projects in Mesoamerica have inferred redistribution as the mechanism of exchange when concentrations of non-local lithic materials at major sites point to debris associated with specialist workshops and other kinds of intrasite use of space. Blade manufacturing debris has been used in this manner at Tikal (Moholy-Nagy 1975;Moholy-Nagy 1976;Moholy-Nagy 1991;Moholy-Nagy 1999), the Basin of Mexico (Sanders, et al. 1979), and Teotihuacan (Spence 1967;Spence 1984).

The proportion of artifact forms of a non-local material in a single site has been used to characterize mechanisms of exchange. This approach has proved useful with obsidian exchange in places where distinctive reduction strategies are associated with finished artifact form, such as blade technology, and these strategies can be recognized in lithic material found in consumption sites. Winter and Pires-Ferreira (1976: 309-310), working at two sites in Oaxaca, argued that blades of high quality, non-local obsidian were introduced to the sites in finished form and that this constituted evidence of elite pooling for prismatic blade reduction followed by redistribution to local sites. The higher quality material was transformed into more valuable artifact forms in workshops located outside of their study area, and then in the Oaxaca sites the presence of these artifacts that were apparently the result of contact with elite spheres of exchange, was interpreted as evidence of redistribution of the finished artifacts. The sourcing and exchange work of Pires-Ferreira has come under some criticism because in her initial proveniencing study she only used two diagnostic chemical elements, and more recent analyses suggest that many of her sourcing attributions are incorrect (Clark 2003: 32). Further, Sheets (1978: 62) argues that the edges of prismatic blades are too fragile to have been transported in completed form (Clark and Lee 1984: 272).

Systematic studies of the intrasite contexts and artifact form of non-local material have the potential to provide insights into exchange and social structure. Intrasite spatial patterns can be investigated in combination with both quantitative data stemming from geographical distances and artifact abundance, and with qualitative data inferred from artifact form and technological aspects of production (Torrence 1986: 36). These kinds of intrasite patterns are susceptible to distortion by dumping patterns and intrasite studies must be sensitive to the problems of conflating material from workshop refuse, household middens and construction fill (Moholy-Nagy 1997).

In a useful review of Mesoamerican obsidian studies Clark (2003: 32-39) summarizes major advances in provenience and trade models, and highlights important avenues for improvement. One important observation made by Clark is that in archaeologists' efforts to design systematic approaches to exchange, early studies failed to distinguish "power as energy and power as legitimacy" (Clark 2003: 38). It could be argued that these forms of power are largely comparable, as the ability to procure, display, and redistribute non-local goods demonstrates power in both energy and in prestige. However, as was emphasized by substantivists in their discussion on the inalienability of certain products, the value of particular status items rests precisely in their incommensurate nature.

Sanders and Santley's (1983) proposal that special goods were returned to Teotihuacan in exchange for obsidian products, and that these were cashed in for corn from the fringes of Teotihuacan's domain in hard times, fails to appreciate that once symbolic goods circulate like economic goods among the hoi polloi such goods lose any legitimizing powers (Clark 2003: 38).

A framework for investigating prehistoric exchange must negotiate between these features of the archaeological record. Material evidence of long-distance exchange is abundant and easily measured, but interpreting the significance and perceived value of these non-local products in the social and political context of their consumption is the measurement of greatest relevance to comprehending the role of exchange over time.

2.3.4. Discussion

A period of vigorous exploration of spatial models of exchange followed on the initial availability of chemical characterization methods in the 1960s. In addition to responding to new proveniencing methods, these developments also reflected a convergence between formal analytical models borrowed from quantitative geography (Haggett 1966;Harvey 1969) and the spatial models of processual archaeology in the early 1970s. As the potential and the limitations of the formal geographical approaches have became more evident, archaeologists have been able to take stock of the insights from that period (Bradley and Edmonds 1993: 5-11;Clark 2003: 32-42;Hodder 1982;Torrence 1986: 10-37, 115-138).

Exploration of a variety of promising measures and ratios for regional analysis is one of positive products of these regional distance decay studies. For obsidian distance decay, the ideal analytical situation would involve a variety of metric measures for each artifact, and the chemical type would be known for every artifact of sourcable material. While such detailed analyses have not been generally available, analysts have made do with other measures such as visual assessment of obsidian type (although it is unreliable in many regions) and relative measures of abundance. These measures of abundance include: (1) proportion of obsidian in total lithic assemblage by count or weight, (2) proportion of obsidian flakes to other artifact classes like ground stone or ceramics, and (3) density of obsidian by excavated volume of soil. A variety of additional inventive measures were explored during this period. Further details from consumption contexts, such as variability in material type by spatial and temporal provenience, are important for regional studies, although acquiring consistent and comparable measures from various archaeological projects can be difficult.

Technological innovations promise to lend greater support to both chemical characterization and geographical analyses in coming years. Developments in GIS and other spatial technologies have greatly facilitated the management and analysis of spatial data. An associated technological development is portable chemical characterization devices such as portable X-ray fluorescence (XRF) units. It is likely that, in the not-too-distant future, XRF analysis will become a routine part of lab analysis for many archaeological projects (Jeff Speakman, May 2006, pers. comm.), a development that will significantly expand the archaeological significance of quarry studies and geographical relationships between production and consumption locales.

In sum, the confluence of spatial technology with the widespread use of non-destructive chemical characterization methods in archaeology will likely result in the emergence of a rigorous proveniencing and spatial analysis sub-discipline in archaeology in coming years. The foundations of this sub-discipline were explored in the 1970s but with greater refinement in theoretical approaches to exchange and the organization of technology, as well as the issues of artifact variability, and the consumption contexts of exchange, one may expect significant contributions from regional analysis in coming years.

2.4. The View from the Quarry

For over one hundred years archaeologists in the Americas have noted the potential of studying quarries, yet consistency is lacking in the approaches that have been taken, making it difficult to compare quarrying behavior cross-culturally. The principal theoretical objective for quarry studies in many parts of the world is establishing a link between models of social change and the extraction of resources at a give quarry through time. Despite the productive research into quarrying during the early part of the twentieth century, quarries worldwide have received insufficient attention by archaeologists.

Beginning with the pioneering work of William Henry Holmes (1900;1919), archaeologists have recognized that the remains of quarrying and mining have the potential to provide valuable information about the past. Holmes' work was a major contribution as he discussed the broad issues of the geographic locations of known quarries and workshops, artifact manufacture, the quantities of waste material, and the use of fire in quarrying. The sole major component of modern quarry studies that Holmes did not explore was technical flake analysis, which wasn't developed yet, and a discussion of the chronology of the use of the quarries, which was very difficult to achieve prior to radiocarbon dating.

With the development of geochemical sourcing methods during the 1960s that linked materials with certainty to their geological source areas, quarry studies began to be of greater interest. The goal of most quarry studies is, through a combination of evidence from the resource procurement and the consumption contexts, to examine changes through time in the mechanisms of exchange that link the production and consumption together. These mechanisms are believed to reflect the prevailing organization of the stone tool economy.

A central challenge in conducting archaeological research at quarries is that at the typical raw material source, due to the sheer volume of archaeological materials and the variability in procurement contexts in prehistory, it is vital to have a clear research strategy. By targeting research questions and theoretical goals, and then determining the appropriate sampling methods, the abundance of material at a quarry can be approached with a few specific guiding questions to be answered, as well as an eye for new, unanticipated findings. Several researchers have discussed frameworks for studying the ancient quarrying of stone. Torrence describes the rich potential of studying production and exchange from the perspective of the quarry in terms of the unique position of the quarry for the study of a "complete exchange system" (Torrence 1986: 91, see Section 2.2.5).

These frameworks for quarry research fall into three sets of approaches, corresponding roughly to major theoretical groups in archaeology, and most will use a combination of these approaches emphasizing (1) efficiency, (2) social factors, and/or (3) ideology. The theoretical approach taken is often conditioned by the available data. For example, an ideological approach is strongest when demonstrable ethnographic, historical, or archaeological data are available that display a clear ideological basis for behavior concerning the stone tool procurement. Similarly, efficiency and error rates are more measurable from some reduction strategies, like prismatic blade production, and therefore cost minimizing analyses are very fruitful with such data.

2.4.1. The specialization and efficiency framework for quarry studies

The most thoroughly-articulated efficiency model is the one developed by Torrence during her dissertation work on the Greek island of Melos (Torrence 1981;Torrence 1986). Her approach at the obsidian quarries of Melos focuses on lithic reduction sequences in order to detect the changes in morphology of flakes that point to increased specialization due to standardization of reduction strategies. Her goal is to establish "a framework for measuring exchange" by developing a continuum for production efficiency that aims to link particular levels of efficiency with the social correlates that indicate the existence of different forms of exchange. Citing Rathje (1975: 420-430), Torrence approaches the study of long-term changes in the efficiency of extraction and manufacture in terms of sophistication of technology, simplification, standardization, and specialization (Torrence 1986: 42). In her research she is able to establish a continuum of efficiency starting with the irregular, non-specialist production on one end, and high efficiency production characterized by ethnohistorical evidence from modern gunflint knapping, on the other end.

Characteristics of prismatic blade production facilitate the kind of examination for efficiency and specialization used effectively by Torrence. First, core-blade reduction sequences leave relatively visible evidence of the technical stages for archaeological analysis. Secondly, prismatic blades are extremely efficient as measured experimentally using cutting-edge to weight ratios (Sheets and Muto 1972). Finally, archaeologists have developed measures of error rates for obsidian blade production with the aim of establishing the degree of knapping expertise from debitage, and these measures correlate with efficiency gauged in both time and consumption of material (Clark 1997;Sheets 1975). These efficiency measures are relevant in areas with blade production, but they are not applicable to research in areas with exclusively bifacial reduction traditions, such as in the south-central Andes.

The efficiency measures that Torrence develops in her approach linking production with exchange has the advantage of being explicit and comparable across regions. Further, her measures of production efficiency complement the formal assumptions that underlie the fall-off curves used to analyze regional exchange models in an evolutionary approach that projects greater efficiencies in organization through time. That is, specialized blade production is the most efficient means of producing cutting implements, and freelance trade in a market based economy, following Polanyi's schema, is the most efficient means of moving goods to consumers. Efficiency measures have heuristic value, as divergences from expected efficiency models can prompt the pursuit of theoretical inquiries into the cause of the deviance from the anticipated efficiency models.

The most problematic aspect of this framework is the dependence on the theoretical link connecting efficiency measures to prehistoric institutions (Bradley and Edmonds 1993: 10), particularly in light of the substantivist/formalist debate in anthropology specifically referencing gifting and exchange relationships. Bradley and Edmonds (1993: 10) observe that, as with Renfrew's regression approach, the problem of equifinality undermines the system of inference when the principal link between production efficiency and elaborate institutions is reduction evidence from workshops. They question the supposition that larger socio-political structure and evolutionary stages can be reconstructed from the limited perspective of workshop production and regional distribution patterns based largely on consumption sites that often have poor temporal control and few contextual associations.

More recent studies have pursued another direction and have avoided comprehensive formal "frameworks" in favor of more willingness to incorporate case-specific details, social dynamism, and historical factors in developing social models of exchange (Friedman and Rowlands 1978;Hodder 1982).

Formal, cost-minimizing assumptions about human actions and incentives have provided archaeologists with a much-needed analytical structure to the study of prehistoric exchange, but critics note that it cannot provide a complete picture of ancient economies.

As Hodder notes, most studies have been predicated on the idea that progress can be made by assuming that people in the past considered costs and benefits along formal economic lines (Hodder 1982). Torrence's (1986) study is a case in point, for it is only by making this assumption that she is able to apply the same scale of measurement to people as different from one another as hunter-gatherers procuring workable stone for their own use, and the makers of gunflints for sale in the modern world market (Bradley and Edmonds 1993: 10).

A formal approach can serve as one of the layers in a more comprehensive analysis that also integrates finer scale social factors as well as historical particulars into the analysis.

2.4.2. Analysis of a production system

In a multi-tiered approach that examines evidence from workshops, from residential sites in the vicinity of the quarry and finally lithics from more distant, consumption contexts, Ericson (1984) describes a general "Lithic Production System".


Variable (numerator)

Normalizer (denominator)


Exchange Index

Single source

Total material

Count, weight, %

Debitage Index


Total Tools and debitage

Count, weight, size, %

Cortex Index

Primary and secondary reduction flakes

Total debitage

Count, %

Core Index

Spent cores

Total cores and tools

Count, %

Biface Index

Bifacial Thinning Flakes

Total debitage

Count, %

Table 2-5. Measurement indices for procurement system (after Ericson 1984: 4).

The indices presented by Ericson depend upon general artifact type categories and provide a basis for comparing activities between workshops, local sites, and distant consumption locales. Note that Ericson did not separate complete flaked stone artifacts from broken artifacts, as this was the early 1980s, and thus his resulting indices using weight measures were likely skewed.

The measures in Ericson's table emphasize the goal of allowing comparability between archaeological datasets over widely studied areas by principally relying on general metrics that are commonly gathered in laboratory analysis. In contrast, the current Upper Colca study employed technical analyses of complete flakes and cores in order to highlight differential reduction strategies between assemblages, or between bifacial core versus flake-as-core reduction.


Figure 2-6. Stages of production from quarry, local area, and region (after Ericson 1984: 4).

Ericson presents the spatial distribution of lithic production in terms of stages of production and zones of geographic proximity to the source area. A consistent implementation of this approach on a local and regional scale requires the sourcing (visually or chemically) of the lithic material. Despite the use of commonly gathered measures in indices of production, it would be necessary to ensure that relatively consistent practices in excavation and analysis procedures were in place to permit this kind of regional comparability.

Ericson asserts that workshops should be studied on a general level rather than pursing spatial and temporal variation in activities at the source. He writes that if "there are a number of different workshops at the source, the data must be merged to form a composite picture of production" (Ericson 1982: 133). The emphasis is on documenting the predominant production strategy at a given source area, but at the cost of characterizing variability within a production context. If changes in the use of a given material have been differentiated from stratified deposits at consumption sites in the larger region, how are these changes to be linked with production activities? While workshop sites often lack datable materials or temporally diagnostic artifacts, collapsing all excavated workshop data, and perhaps surface evidence as well, into a single composite picture sacrifices the detail that stratified deposits can provide.

Excavation and analysis procedures in Ericson's California dataset appear to be relatively consistent for the past half-century, however the systematic collection and analysis of non-diagnostic lithics has been historically deemphasized by archaeologists in some regions of the world, such as the south-central Andes. Archaeological practices worldwide have not placed equal emphasis on gathering and quantifying lithic artifacts which would lead to problems in applying the kinds of metrics Ericson suggests on a regional scale. Ericson's approach emphasizes the use of quantifiable and comparable measures over a geographically extensive region.

Barbara Purdy (1984) emphasizes chronology in her study of a chert procurement site in Florida. She did not find datable, organic remains in association with stratified excavations that would provide greater evidence of the temporal associations, but the strata in her test units allowed her to differentiate distinct episodes of use of the site. She found that weathered chert artifacts in a sandy-clay soil were separated by a lithologic discontinuity from less-weathered artifacts produced using a different technology. She was also able to use the relative weathering of chert and thermoluminescent dating on fire-altered chert cobbles from the quarry site.

2.4.3. Specialization at a Mexican Obsidian workshop

Obsidian quarry workshops at Zináparo-Prieto in the state of Michocan in western Mexico were investigated comprehensively by Veronique Darras (1991;1999). Darras undertook excavation as well as systematic survey of vicinity of the substantial quarry area that was used most intensively during the Classic to Post-classic transition (A.D. 850-1000). She documented mines and workshops, as well as associated residential structures and public buildings, in an inventory of 45 sites in the region.

Darras describes mine shafts and open-air quarry pits that parallel obsidian procurement methods elsewhere in Mesoamerica. The mine shafts (Darras 1999: 72-80), are over 25m in depth and include support pillars as well as evidence of torch lighting and ventilation shafts, can be compared with Classic and Post-classic period mining methods used at other Mexican quarries that include the Sierra de las Navajas (Pachuca) source in Hidalgo (Pastrana 1998), and the Orizaba mines (Cobean and Stocker 2002). Open-pit quarry depressions are reported by Darras at Zináparo-Prieto that measure between 10-15m in diameter, with a few larger than 15m in diameter. Quarry depressions with an encircling debris mound were described by Healan (1997) at the Ucareo source, also in Michocan, as "dough-nut quarries".

The lithic analysis approach used by Darras, citing the tradition of Tixier (1980), addresses the abundance of material (the typical problem with quarry research) by employing two levels analysis (Darras 1999: 108-115). The first measures abundance using a relatively expedient typological classification based on material quality, technical class, reduction stage, and relative size and form of artifact. The second level of analysis, conducted only on a representative samples, consists of a techno-morphological analysis of flakes including a detailed study of platform characteristics and fracture and termination types.

Darras finds no evidence of pressure flaking or prismatic blade technology, a situation that makes this workshop analysis to be more comparable to reduction sequences elsewhere outside of Mesoamerica where prismatic blade production also did not occur, such as obsidian production the south-central Andes. Darras finds that percussion industries follow two contemporaneous sequences: one that uses "cada plana" (plane-face) cores, and the other that uses conical cores, each sequence meticulously outlined. These twin industries, and the lack of prismatic blade production, are unusual in the region and Darras suggests that this is due to a lack of regional political control during this stage.

Darras succeeds in integrating household archaeology with her workshop studies, and as revealed by comparison with obsidian consumption in an associated village, Darras documents substantially more obsidian production than appears to have been used locally. However, she does not extend her analysis to the immediate region beyond villages adjacent to the quarry areas, leaving a disjunction between reduction evidence from the immediate quarry workshops and the larger pattern of obsidian artifacts radiating into the region. Darras' study is one of the most thorough workshop investigations to date, and her work provided a valuable model for the current Upper Colca project research because in the manner that test excavations were used to connect quarry evidence with initial reduction trajectories at associated workshops.

2.4.4. A contextual approach to Neolithic axe quarries inBritain

A focus on social context is emphasized in more recent studies of exchange. Building on Hodder's (1982) development of a "contextual approach" to exchange, archaeologists following this approach argue that the study of production sites has been incomplete because the social and sometimes historical elements of quarries and production systems have been neglected in processual tradition that focuses overwhelmingly on the organization of technology and production efficiency. In particular, Hodder and others have argued that the perspectives on exchange articulated by Mauss (1925) and even by Sahlins (1972) have been largely neglected (to judge from citation patterns) in the formal approaches previously described.

A study by Richard Bradley and Mark Edmonds (1993) focuses on axe production and circulation in Neolithic Britain through an examination of quarry production at the Great Langdale complex in the Cumbrian mountains in the Lake District of the English uplands. This area was a source for a fine-grained volcanic tuff material used for producing stone axe heads that were flaked, ground, and often polished, and have been found in Neolithic. Through petrographic analysis it has been possible to connect many axes made from tuff and granite found throughout Britain with the raw material source in the Great Langdale region, but subsourcing resolution has not been possible.

Background and methodological approach

Bradley and Edmonds (1993: 5-17) begin with a useful critique of the formal approaches to exchange first articulated in the work of Renfrew and his colleagues. In reviewing prior approaches, Bradley and Edmonds perceive weaknesses and untenable assumptions in the links asserted by prior researchers connecting (1) efficiency in reduction strategies, organization of production, and degree of hierarchy in social organization (Torrence, Ericson, others), and (2) geographical distributions of types of artifacts with the nature of social organization (Renfrew et al.).

Bradley and Edmonds emerge with a strategy that permits them to connect temporal change in quarrying and the organization of reduction in the immediate vicinity of the source with perceived changes in knapping strategies. They also explicitly attempt to incorporate evidence from social and symbolic constraints on quarries, where historical specifics about specialization, quarry access, and socio-political boundaries appear to trump the larger patterns of circulation documented during the 1970s by Renfrew, Hodder, and others (Bradley and Edmonds 1993: 9, 63). Further, the authors observe that while these social and symbolic control variables are extremely difficult to appraise from archaeological evidence, these unknowns were ultimately some of the most important variables informing Torrence's (1986) formal approach to efficiency and socio-economic control of production. On these grounds, Bradley and Edmonds devote more effort to methodological and theoretical goals that they feel are attainable: documenting variability in production, inferring connections with regional consumption evidence through temporal context, and exploring social and symbolic generalities through ethnographic analogy from quarries and with evidence of regional ground axe distributions in Britain.

Technical analysis

Bradley and Edmonds' (1993: 83-104) lithic analysis begins with experimental knapping studies that allow them to identify the character and frequency of different classes of flaked stone generated during production. Establishing reduction stages from flakes of Great Langdale tuff material is particularly difficult because the material does not contain a visible cortex. Bradley and Edmonds pursue strategies that produce a wide range of flake and core morphologies in an effort to capture the range of possible variation in reduction at production sites. Their expressed aims are to move beyond simple measures of efficiency in production, and they also seek to use their experimentally-derived assemblages to inform their analysis such that that they do not merely derive a single, generalized reduction sequence but, instead, shed light on how knappers controlled form, and anticipated and avoided mistakes.

It was just as important to discover which methods couldhave been used to make an artifact as it was to establish which were actually selected. It is clearly important to understand howthe production process was structured in a given context, but we also need to discover whyit took the form it did (Bradley and Edmonds 1993: 88, emphasis in original).

In earlier work Edmonds (1990) cites from, and applied, the chaîne opératoireapproach to his investigation of quarry production. Curiously, in the Bradley and Edmonds (1993) volume the French term does not appear (although they cite the chaîne opératoireliterature), and in their quarry production studies they instead choose terms like "pathways" to describe sequential reduction.

Part of Bradley and Edmonds' aim is to reintegrate symbolic and social perspectives into the study of quarry production and exchange, empirical archaeologists may ask: how do Bradley and Edmonds conduct symbolic and social analysis with lithic attribute data from a quarry workshop? In their data-oriented investigation of reduction strategies, Bradley and Edmonds gather standard lithic attribute data that is largely in common with those that follow the processualist tradition; it is in the interpretations and assumptions about economy that the differences emerge. For example, many of the assemblages are described by Bradley and Edmonds along a production gradient that range from "wasteful and inefficient" to "careful preparation" or ad-hoc versus structured production using evidence from flake dimensions, platform morphology and preparation, flake termination type, and other attributes.

For Bradley and Edmonds, efficiency is investigated primarily in a heuristic manner in association with spatial context in that they suggest expediency, investment, and reduction strategies over the larger quarry area. Their technical analysis is able to conclude, among other things, that axes appeared to be exported from the quarry area in two forms: (1) crude asymmetrical rough-outs with hinges and deep scars, and (2) more processed and nearly finished artifacts that lacked only polishing.

They also note that some processing appeared to occur at some distance from the quarry, and in other cases nearly all of the production sequence occurs at the quarry. The authors used evidence of excessive labor expenditure, such as in axe grinding and polishing, as a contrast to the "rational", cost-minimizing expectations of efficient production expectations. For example, they explain that polishing of axe heads is laborious, but it results in greater longevity in the axe because during use the irregularities can act as platforms for unintentional flake removal. Furthermore, polished axes remain in their hafts more consistently. Polish over the entire surface, however, is not necessary and is labor intensive. Were axes polished over their entire surface to increase their exchange and gift value through greater labor investment? In evolutionary approaches, labor investment in goods is cited as a form of prestige technology (Hayden 1998, see Section 2.2.2). Bradley and Edmonds present an insightful review of existing approaches and bring quarry analysis one step further with their in-depth technical analysis that informs a novel yet cautious interpretation with an incorporation of elements of the symbolic and social theory current in the early 1990s.

Chaînes opératoires at quarry workshops

The chaîne opératoireanalysis of quarry materials described by Edmonds (1990) deserves further mention. The chaîne opératoireapproach conceives of lithic reduction as a comprehensive system or "syntax of action" from the origin of lithic material at the quarry to reduction, reuse, and abandonment within a larger context of human action. Sellet (1993: 106) writes that the chaîne opératoireis a "chronological segmentation of the actions and mental processes required in the manufacture of an artifact and its maintenance in the technical system of a prehistoric group. The initial stage of the chain is raw material procurement, and the final stage is the discard of the artifact."

In the realm of lithic production, Shott (2003) and others question whether the chaîne opératoireapproach is notably different from Holmes' century-old concept of "reduction sequence" (Holmes 1894) and Schiffer's Behavioral chain analysis (Schiffer 1975). Two major differences distinguish chaîne opératoirefrom the lithic reduction sequence analysis in the North American tradition. First, while the American reduction sequence focuses solely on lithic production, chaîne opératoireapplies to apprenticeship and expertise relating to allmaterial culture behavior beginning with lithic production but also including ceramics, textiles, architecture, wine-making, and others (Tostevin 2006: 3). Second, even when confined to lithic production, the chaîne opératoireapproachexplicitly attempts to infer the choices and intentionality of the knapper, as well as to capture greater context and breadth by seeking to address the larger context of activity and action.

French archaeology has a long tradition of attributing archaeological variability to choice, stretching back from Boeda (1991) to Bordes to Breuil. This tendency is balanced by an equally strong American-based resistance to attribute variation to cultural choice until all other factors, particularly reactions to environmental stimuli, have been excluded, forming one of the central tensions in the current debate (Tryon and Potts 2006).

Others argue that while attempts to consider the intentionality of the knapper are commendable, in their weaker form sequential models like chaîne opératoirerun the danger of being overly typological and rigidly unable toincorporate behavior that diverges from a one particular linear progression (Bleed 2001;Hiscock 2004: 72).

From the perspective of quarry procurement and workshop activities, production would have proceeded with a goal in the mind of the knapper based on the quality of the raw material as well as organizational issues like technological requirements and mode of transport. For Edmonds (1990: 68) implementing a chaîne opératoireapproach at a quarry workshop appears to have been complicated by an overwhelming presence of flaked stone belonging to early stages production, and a paucity of evidence concerning consumption. It seems that observing a complete "syntax of action" is hampered by the incomplete view of reduction, as the evidence is overwhelmingly from workshops and not from consumption sites. Thus, while quarry workshops have relatively few classes of artifacts the techniques, such as refitting, are seldom practicable when there is an abundance of early stage material and an under-representation of advanced reduction flakes and complete, or near complete, tool forms.

Variability in initial reduction of cores observed at the Great Langdale quarry workshops reflect the available raw material and the quarrying methods used to procure the material, and these frame the starting context for following a chaîne opératoireanalytical model. However, the chaînesequence is often truncated and largely capable of merely determining early reduction characteristics that are basic to all reduction sequence analyses in the tradition of Holmes (1894). For example, at Great Langdale the researchers determine, mostly from platform characteristics, that there was expedient, ad hoc production in one period and more precise and controlled flaking in another. Sequence models seem to be of limited utility in contexts where only initial workshop production is available, a condition that perhaps explains the avoidance of an explicitly chaîne opératoireapproach in the later Bradley and Edmonds (1993) volume. In sum, chaîne opératoireis roughly synonymous with American 'reduction sequence' and Schiffer's Behavioral Archaeology in terms of low and mid-level theory, but in high level theory it is not generally presented except for the theory of chaîne opératoireitself (Tostevin 2006). Given the difficulties of meaningfully applying chaîne opératoireto data based almost entirely on quarry workshop contexts, the concepts will not be attempted in this research.

Interpreting the axe trade

While earlier approaches focused on efficiency and evolutionary schema in a commodified vision of production, Bradley and Edmonds (1993) attempt to take a middle road where they use measures of efficiency and investment observed in workshop contexts in a heuristic manner, but they principally base their interpretations of production on the changing socio-political context of the larger consumption zone. They consider artifacts in terms of dichotomies that probably existed in some form, dividing the circulation of inalienable gifts from alienable commodity production. They seek to consider the implications of gifts in the theoretical terms that relate gift-giving and status acquisition with the political strategizing of elites in Neolithic Britain. Further, they consider the "regimes of value" where gifts and commodities circulate and are assigned value that is a construction of political contexts and not merely a reflection of measurable costs using the concepts of social distance borrowed from Sahlins. Finally, Bradley and Edmonds consider the circulation of axes as wealth goods and the ways in which elites may influence the specialized production of axes (Brumfiel and Earle 1987) and the circulation and consumption in a peer polity situation (Renfrew and Cherry 1986) or through control of deposition (Kristiansen 1984). Ironically, while neoevolutionist approaches are explicitly rejected, one is left with the question: what is theoretically significant about changes observed in the circulation of axes if not the link to evolutionary changes in socio-political organization? In both Edmonds (1990: 66-67) and Bradley and Edmonds (1993), evolutionary explanations are avoided in the analysis of production, but in regional exchange the changes observed attributed to increasing levels of social ranking as indicated by competition over exchange networks and specialized knowledge during the Later Neolithic.

2.4.5. Discussion

The principal challenges of quarry research in archaeology were articulated in the seminal work of Holmes, one hundred years ago. These difficulties include the sheer quantity of non-diagnostic artifacts and sampling issues, the lack of temporal control, and stratigraphy that is either complex or non-existent, it is no wonder that relatively few projects have targeted quarry areas in the intervening century.

Advances in the last few decades include methodological improvements like rigorous attribute analysis and greater standardization of measures and digital measurement devices that have sped up lab work. Theoretical advances include an exploration of principles of production efficiency and subsequent articulation of the problems and prospects of this formal approach. Principal among these are further incorporation of data from adjacent contexts, the use of other lithic material types close to a major source, and the incorporation of evidence from other material classes like ceramics. More recent advances include the incorporation of additional datasets into quarry and workshop analysis, a wider theoretical scope, and an attempt to understand the wider intentionality and decision sequence of quarry procurement and initial production.

A promising theoretical approach to procurement and exchange would recognize the need for a consistent framework against which to assess changes in production, circulation and the regional demand, but it is also one that responds to local variation and multiple reduction trajectories outside the scope of formal concepts of efficiency. Production and circulation of lithic raw material from quarry sites often span broad time periods and must reconcile with a great variety of cultural and organizational forms. In regions of the world (such as the south-central Andes) where foraging was largely replaced by agro-pastoralism, where residential mobility was replaced by sedentary communities with mobile components, and where egalitarian social structure changed into ranked and ultimately stratified societies, these large scale changes must be reconciled with evidence of technological organization and exchange. While a skeleton of expectations can be built from the general anthropological evidence provided in this chapter, the regional specifics of Andean prehistory flesh out the character of production and exchange of Chivay obsidian through time. Models that are applicable to the case of Chivay obsidian and that assimilate issues from this chapter with the Andean regional trajectory are presented at the end of the following chapter.

Ch3 Archaeology of Region

The role of interaction across long distances has been a persistent theme in archaeological studies in the south-central Andean highlands. The emergence of the Tiwanaku state in the stark and sparsely populated altiplano seems improbable unless one considers the larger geographical setting and the apparent importance of sustained links over long distances that contributed to developments in the Lake Titicaca Basin. Archaeological research that focuses on the Formative Period, the time that preceded Tiwanaku in the Titicaca Basin, has demonstrated that regional centers coalesced out of a multitude of small villages that were articulated through a highly mobile sector of the economy based on camelid caravan transport. This chapter examines the roots of this mobile sector in the economy of the Terminal Archaic and the Formative by focusing on the transport of obsidian and the significance of early evidence of transport in the origins of long distance relationships in the south-central Andes.

This chapter will contextualized the theoretical issues surrounding raw material production and exchange in the south-central Andes. Subsequently, this chapter will explore some models of long distance interaction in the south-central Andes. Next, a review of the regional evidence for the production and circulation of Chivay obsidian from the a few major sites in the highlands will be followed by a temporal survey of Chivay distributions revealed by previous archaeological research. These regional data will be discussed by progressing through time beginning with Chivay obsidian use by early mobile foragers, and continuing in time through to the Inka period. A subsequent section considers broader patterns for use of obsidian in the Andes. This chapter concludes with a discussion of four models for obsidian procurement and circulation from the Chivay source, and the associated material correlates for these models.

A larger set of questions guides this chapter: Do the distributions of Chivay obsidian conform to those of utilitarian products like salt and dried meat, or do the distributions resemble those of prestige goods like spondylous, copper, or gold? Escaping the utilitarian / prestige goods dichotomy, a third group termed "cultural goods" better captures the traditional networks that circulated material like obsidian in the Andes. This chapter will argue that obsidian is an example of a class of material that moves between these simple classifications and that it requires a consideration of context. What pressures were responsible for the rapid change in production and circulation of obsidian during the Terminal Archaic Period? What can obsidian circulation reveal about the regional context and the strategies of early aggrandizers at the beginning of the Formative in the Titicaca Basin?

Developments in obsidian production at the Chivay source intersect with a wide swath of the Andean prehistoric sequence, and these developments will be explored below in three broad cultural periods: the "Archaic Foragers", "Early Agropastoralists", and "Late Prehispanic".


Figure 3-1. Chronologies discussed in the text.

"Archaic Foragers" are characterized as mobile foraging groups responsible for the early human use of the obsidian source, exchange, and consumption of the material. Use and exchange of obsidian by Andean mobile foragers during the Early, Middle, and Late Archaic Periods will be considered in light of ethnographic and archaeological evidence.

Next, with the domestication of camelids and chenopodium occurring sometime during the Late Archaic, food production came to dominate the economy by 3300 cal BCE heralding sweeping changes that included early social ranking that first appeared during the Terminal Archaic and early Formative Periods. During this period, referred to here as the Early Agropastoralist period, the production and exchange of obsidian appears to have changed as the material became more widely circulated in the consumption zone and mobility was facilitated by the gradual inception of regular camelid caravan transport routes. Reciprocity-based exchange systems would have coalesced, while personal procurement activities persisted from the earlier period, resulting in distinctive production evidence in the source area. The evolution of expansive states during the Middle Formative and Late Formative through to the Tiwanaku Horizon shifted the regional political structure, and these developments had repercussions in the Chivay source region. As regional centers in the Titicaca Basin emerged and began to dominate the political landscape, the established exchange patterns may have shifted to reflect the role of redistribution believed to have been occurring at the centers. Political forces became increasingly powerful in the Titicaca Basin, as well as expanding out of the site of Wari far to the north in Ayacucho, and the procurers of Chivay obsidian in the Colca area were increasingly working in a border land between powerful polities. What kinds of evidence will be reflected by the combined impacts of local provisioning, reciprocity, and elaborate redistribution systems in the Chivay area and in the Titicaca Basin consumption zone? With the collapse of Middle Horizon polities, the effects of warfare on exchange are apparent in the limited distributions of obsidian during the Late Intermediate Period. Finally, the Inka period was a time of great regional integration and tremendous redistribution of goods, but it appears that obsidian had reduced significance in favor of competing materials like metals. The changing nature of prehispanic obsidian exchange, and the link between exchange theory and the south-central Andean sequence, will be considered in this chapter.

3.1. Andean Economy and Exchange

The Andes present a valuable opportunity for examining anthropological models of economy and exchange in prehistory. Distinctive historical aspects of Andean development, including the emergence of pristine states at high altitude, the administration of vast empires without a formal system of writing, and the wealth of ethnohistoric data provided by Spanish chroniclers, offer important research problems for economic anthropology. This study investigates the procurement and circulation of obsidian from the Chivay source in the south-central Andes during a broad time period that includes major shifts in the economy and in socio-political organization.

Throughout prehistory, the long, narrow Andean cordillera presented distinctive challenges to human groups that were addressed through a variety of technological and social strategies. Here, a focus on lithic raw material procurement and exchange permits obsidian circulation to serve as an indicator for particular types of regional interaction.

Exchange has a complex role in mediating human relationships over distance (see Section2.2.2), and in the south-central Andes, obsidian appears to have served as both a political tool and relatively ordinary aspect of economic activity. In part, the persistence of exchange in herding regions reflects lack of autarky among the dedicated pastoralists; they require vegetative and agricultural products, and such items are widely available in the sierra and foothills. In reference to the early development of long distance caravan networks in the altiplano, David Browman (1981: 413) notes "[t]he trade in consumables is less spectacular than the trade in luxury items, and more difficult to detect archaeologically, but it was much more important to the average altiplano inhabitant." In other words, due to the environmental contrasts in the Andes, relatively mundane consumables like salt, ajipeppers, and even coca leaf could precipitate a low-level but persistent demand for exchange of goods between adjacent ecological zones and, in some cases, across larger distances.

A central point of the following discussion is that wide-ranging exchange networks, apparently organized at the level of the household and facilitated by caravan transport, are a persistent theme in the south-central Andean highlands. These networks do not integrate easily with the exchange typologies presented above (Section 2.2.4), and this form of articulation is sometimes seen as irrelevant "background" reciprocity in models of early competitive leadership. However, this distributed mode of integration may have served as an early foundation for subsequent political organization in the region. This study focuses on obsidian procurement and distribution and infers that other goods were also being transferred along these networks. While the simple assumption that evidence of obsidian circulation is analogous to prehistoric trade in a multitude of other more perishable goods is problematic, the persistence of obsidian exchange in the south-central Andes is compelling evidence of generalized contact over distance.Andean approaches to regional economy are reviewed here in order to examine the distribution of obsidian and other goods through diverse mechanisms of procurement and exchange.

3.1.1. Economic organization and trade in theAndes

In a cross-cultural perspective, Andean exchange relationships throughout prehistory exhibited characteristic organizational traits of societies dwelling in mountain ecological zones; although in the late prehispanic periods distinctive features of inter-zonal control emerged in the south-central Andes. There is general consensus among Andeanists that the mechanisms of merchantilism and market economies - prices reflecting supply and demand - did not exist in the late prehispanic south-central Andes[2].

Polanyi's substantivist economic typology based on peasant households in non-capitalist settings is still widely used in the Andes with some modification.

Reciprocity, redistribution, and non-market trade are the institutional means by which indigenous Andean economies operate. All evidence points to the overriding fact that true market systems did not operate in the central Andes, as they did in central Mexico and in a number of complex societies of the Old World. Exchange did exist on a massive and pervasive scale, however, and the concept of administered trade is the superior means of understanding this phenomenon in the prehispanic central Andes. Trade existed, but it was not one based on market principles. Virtually all cases of trade were administered by some corporate group, constituted along sociological (kinship) or political lines (Stanish 1992: 15).

The major modes of economic interaction will be reviewed below. Finally, the closing section of chapter 3 will explore more specific material expectations of how each of these economic modes may appear in procurement areas such as the Chivay obsidian source.

Direct Access

Direct household procurement of goods is largely structured by the geographic relationships between consumer residence and a given resource. Access to products from complementary ecological zones by consumers who undertake a personal voyage to acquire and gather those goods is a form of direct access. For resources that are widely distributed in ecological areas or along altitudinal bands, direct access is a recurring theme in Andean prehistory. The procurement of unique or unevenly distributed or resources, such as salt or obsidian, is a different configuration entirely in a vertically organized region like the Andes, because distant consumers are forced to articulate with production areas far beyond their immediate and complementary neighbors in an altitudinally stratified exchange relationship. This kind of multiethnic, direct household procurement for salt occurs in the Andes to this day (Concha Contreras 1975: 74-76;Flores Ochoa 1994: 125-127;Oberem 1985 [1974];Varese 2002).

Direct access by foragers was the first mode of procurement in the Andes, and this acquisition mode probably dominated in Early Holocene prior to the population growth that permitted the development of reciprocity networks. The procurement of raw materials in a manner that is incidental to other subsistence activities is a more efficient means of acquiring these goods, an activity described as "embedded procurement" by Binford (1979: 279). Communities, ethnic groups, and even prehistoric states appear to have maintained direct access to resources in other zones, and (as is stipulated by the definition of the direct accessmode) this kind of articulation is for direct consumption or redistribution on the level of the corporate or state entity. This is part of a much celebrated pattern in Andean research, a feature known as vertical complementary (Murra 1972), a topic that will be returned to in more detail below. Direct access by states to unusual sources of raw material such as metals and minerals are well documented in the Andes. These include as Inka mine works, and distinctive Inka artifacts and architecture are commonly encountered in association with the procurement areas. In access between herding and agricultural sectors, there is also an ethnographically documented direct access mode described as "ethnic economies". In this direct access mode, ethnic groups will control parallel strips of vertical land holdings, and sometimes non-contiguous tracts, ranging from lower agricultural zones to high puna that may lie several thousand vertical meters above. This pattern is documented on the eastern slope of the Andes for the Q'eros of Cuzco (Brush 1976;Flores Ochoa, et al. 1994) and several communities in northern Potosí in Bolivia (Harris 1982;Harris 1985). The important concept of the direct access organization is that entities that were consuming the goods were directly responsible for acquiring them. If there is inter-household barter or transfer of any kind, then the arrangement likely belongs to a type of reciprocity relationship.


The institution of reciprocity is important in all societies, and in the contemporary Andes reciprocal relationships are elaborate and permeate village life. It is an arrangement for the transfer of labor or goods that is organized without coercive authority between entities equal in status, although sometimes disputes are settled by community leaders. Andean labor reciprocity includes agricultural work, roof raising, canal cleaning, terrace building, and other services; as such, reciprocity structures the traditional village economy in the Andes (Alberti and Mayer 1974;Stanish 2003: 67). These kinds of reciprocal arrangements are frequently delayed, although compensation can be accelerated through recompense in products. For example, a herder might bring a caravan down to a farming area in the lower valleys and spend some days contributing labor to the agricultural harvest in exchange for some portion of the yield.

In many premodern economic transactions relationships of balanced reciprocity structured these arrangements. Two forms are likely in the Andes, that include (1) a multitude of small, household exchanges creating "down-the-line" artifact distributions (see Figure 2-2), presumably this type of exchange is responsible for the long distance transmission of small, portable goods for much of the pre-ceramic period. A second form consists of (2) barter exchange relationships with regular long distance caravans that articulated with settlements, and perhaps at periodic fairs, that transported goods over potentially greater distances. This mode would effectively consist of unadministered trade. The institution of reciprocity will be explored in more specific contexts below.


The basis of relations between political elites and non-elites in the prehispanic Andes was shaped by redistribution. In Andes during the later prehispanic period, redistributive mechanisms linked elites to non-elites through the redistribution of consumables like coca, and feasts of food and chichabeer in exchange for labor and political support. "The manipulation of redistributive economic relationships among the elite and their retainers, most notably of exotic goods and commodities, stands at the core of the development of Prehispanic Andean complex societies" (Stanish 2003: 68). Often these surplus goods were produced through efficient mechanisms orchestrated by elites, and the benefits and prestige derived from these surpluses would be accrued disproportionately by political leaders.

The central collection of goods for redistribution or use by the state includes taxation which, in the Inka period, was through mit'alabor. With respect to the exchange of lithic raw materials, Giesso (2003) argues that at Tiwanaku household stone tool production was a form of taxation. Giesso cites ethnohistoric evidence referencing the Inka period and argues that the household knapping of projectile points could have contributed to the provisioning of the state armory. The means by which non-local material arrived in the Tiwanaku homeland is unclear, but in the Inka case the raw material was acquired locally or it was provided by the state (Giesso 2003: 377). Earle (1977: 215) argues that redistributionshould be considered as two major groups with leveling mechanisms on one side and complex institutional mechanisms for wealth accumulation on the other (see Section 2.2.4).

Administered Trade

Non-market administered trade was another major means of transfer in the prehispanic south-central Andes. Building on Polanyi's classic definition of non-capitalist economic types, Charles Stanish (2003: 69) describes a form of elite-administered, non-market trade that was capable of procuring non-local goods and that provided wealth to elites as follows. Garci Diez's Titicaca Basin visita, a 16thcentury Spanish census document (Diez de San Miguel 1964 [1567]), describes how local elites in the Lake Titicaca Basin would have their constituents organize llama caravans for trading expeditions to adjacent regions. In neighboring areas such as the Amazon basin to the east, the Cusco valley, and western slope valleys, agricultural goods such as corn, fruits, and other products were sweeter, faster growing, and more abundant than in the Titicaca Basin. The colonial visitaindicates that, based on the colonial currency, corn in the Titicaca Basin was worth 5.7 to 6.9 times the amount that it was worth in the Sama valley (an area in modern-day southern Bolivia) where it was abundant. Stanish (2003: 69-70) argues that administered trade benefited elites because they were able to appropriate this difference in value, and through feasting and other ceremonial functions, a portion of this wealth was redistributed to commoners. It is important to mention that these same colonial sources indicate that the commoners also organized trading ventures and would take advantage of these elite-organized journeys to conduct private barter exchange on the side. With reference to herders conscripted into elite orchestrated trade caravans "those in Lupaca country 'who had their own cattle [cargo llamas]' (Diez de San Miguel 1964 [1567], f. 13v) went to the coast and to the lomas to barter on their own. …the maize growers on the irrigated coast were eager for the highlander's animals, their wool and meat" (Murra 1965: 201). Thus, elites organized large caravans and apparently possessed the surplus wealth and the camelid caravan animals in advance to initiate the trading expedition, but the herders that they conscripted also engaged in household-level barter. For the elites, their organizational efforts earned them significant wealth and status for a relatively modest outlay of costs. For the herders, it appears that they were able to embed household economic transactions with their mit'alabor service by conducting their side barter activities. Apparently, even the powerful Lupaca elite had to concede some independent trade activity to their caravan drivers. What, then, of the relationship between caravanerosand elites during earlier periods, when elites probably had less consolidated power than during the contact period?

The evidence suggests that "administered trade" was not the first form of long distance caravan exchange. As mentioned above, relationships of balanced reciprocity have long served to articulate herders with those living in complementary ecozones such as sierra agriculturalists, coastal fishers, and residents of the eastern lowlands. But the long distance transport of diffusive goods like obsidian are well-demonstrated and form a continuous network configuration that contrasts with the segmentary, vertically organized exchange between valley and puna (Figure 2-3). Archaeological distributions (Browman 1981;Burger, et al. 2000;Dillehay and Nuñez 1988;Nuñez and Dillehay 1995 [1979]) and contemporary ethnoarchaeological studies (Lecoq 1988;Nielsen 2000;West 1981) attest to the capacity for small scale, household-level organization of multi-week caravan expeditions. This evidence suggests that there were probably at least two major types of long distance caravans operating from the Late Formative onwards. The question then becomes: what was the relationship between household-level caravans and elite-administered trade? Did elites co-opt functions that were previously coordinated on the household level? If elites acquired control of some segment of caravan traffic, what strategies did elites use to wrest control from caravan drivers that, the evidence suggests, were very independently-minded people (Browman 1990: 419-420;Nielsen 2000: 517-520)? These questions concerning the origins and configurations of regional interaction in the south-central Andes are at the center of this discussion of changes in obsidian procurement and the regional circulation of goods in prehistory on the perimeter of the Lake Titicaca Basin.

3.1.2. Economy and exchange in mountain environments

Cross-cultural studies of human adaptation to mountain environments have revealed a number of common features between production strategies employed by people in the Andes, the Himalaya, and the Alps (Funnell and Parish 2001;Guillet 1983;Rhoades and Thompson 1975;Tomka 2001). These commonalities in adaptation to mountain settings include:

  1. Both specialized and mixed procurement systems are found, but there is a predominance of mixed systems.
  2. Specialized procurement systems are interconnected through regional exchange.
  3. Land holdings by a particular social unit can be non-contiguous and distributed across multiple ecological niches.
  4. In extensive procurement zones, such as grazing areas, land tenure is communal, whereas in intensive procurement, such as irrigated farmland, tenure is often family based.

These strategies are responses to characteristics of mountain settings that include altitude-based biotic ecozones, limited productivity in any single zone, and risk to herding and farming in production activities.

These regular features of production in mountain settings provide a comparison against which to evaluate procurement strategies in the Andes. A number of characteristics of production common to mountain environments have been inappropriately conceived as exclusively Andean in an essentialist tradition referred to as lo andino (Starn 1991;Van Buren 1996), while conversely others have sought to impose Andean models on regions where the model does not necessarily apply (Goldstein and Messerschmidt 1980). These models of regional interaction in mountain environments, both in Andean and general geographical models, can be contrasted with regional distributions of raw materials. Obsidian and other raw materials circulated widely in the Andes, and the spatial patterns described by these materials, may be examined in light of other regional patterning like stylistic distributions, as well as economic models of regional interaction.

Vertical complementarity

The contrasting ecological zonation found in low-latitude mountain regions worldwide has resulted in distinctive social configurations that appear to reduce risk, broaden the selection of products available in a given zone, and provide opportunities for strategic advancement by particular individuals or groups. These configurations have been investigated in two broad sets by Andeanists. First, there are a number of scholars who address vertical complementarity as a general process that is comparable with other mountain regions of the world. Second, there is a particular configuration known as "Vertical archipelagos" first described by John Murra (1972) that has been widely discussed in the Andean literature.

Verticality writ large

Vertical complementarity encompasses a variety of strategies for the problems posed by human use of resources at different scales, and by the broad natural diversity across relatively small distances in mountain environments (Aldenderfer 1993;Masuda, et al. 1985). These problems of articulation are addressed through mobility, through direct control of different zones by a single group, by mutualism between residents of different zones, and through a variety of exchange relationships. Vertical organization has been recorded among modern Quechua and Aymara communities (Brush 1976;Brush 1977;Flores Ochoa, et al. 1994;Harris 1982;Harris 1985;Platt 1980).


Figure 3-2. Contemporary types of Andean zonation (Brush 1977: 12 ).

Based on contemporary observations, Brush (1977: 11) describes three systems:

  1. the archipelago type, or vertical complementarity, as conceived by Murra;
  2. a compressed type, where a single village can access and control all resources without resorting to means of long distance control;
  3. the extended type that corresponds to horizontal complementarity.

On a more localized scale it is possible to see vertical control strategies within a particular valley in a mixed agropastoral strategy that has been called "compressed archipelago" (Figure 3-2). In the central part of the Colca valley on the western slope of the Andes in Arequipa, Peru, Guillet describes vertical household relations.

To what extent do households integrate both puna pastoralism and valley farming into their production and exchange strategies? First, most households residing in the puna tend to specialize in herding and do not have agricultural fields that they cultivate directly. Similarly, many village households neither belong to family surname groups with access to puna pastures nor count themselves among those who have gained control of communal pastures ( botaderos) on the slopes behind the village. Households that follow such specialized strategies must perforce use the exchange nexus to obtain complementary products (Guillet 1992: 133).

Additional evidence for micro and macro vertical complementarity is discussed in the context of the Colca valley (Casaverde Rojas 1977: 172;Málaga Medina 1977: 112-113;Pease G. Y 1977;Shea 1987: 71).

Vertical complementarity can viewed as an anthropological principle that describes the propensity for social groups in mountain environments, from foragers to state societies, to geographically broaden their social and economic base and reduce risk by exploiting a variety of environmental settings (Aldenderfer 1993;Guillet 1983;Salomon 1985). Salomon (1985: 520) presents complementarity strategies in prehispanic Ecuador as varying in two dimensions.


Figure 3-3. Diagram of institutions of Andean complementarity (from Salomon 1985: 520). Numbered modes reference the "Modes of Trade" in the figure from Renfrew (1975: 41-43 ) shown in the previous chapter as Figure 2-2.

One is between decentralized systems based on reciprocity, and the other is based on centralized systems of redistribution. There is an underlying neo-evolutionary correspondence implied in many of these models, as chiefdoms and states are believed to have been responsible for the network convergence perceived in redistributive systems, however it is important to observe that due to the variety in products, social relationships, and economic configurations, it is likely that a great many of the institutions presented in Figure 3-3 occurred simultaneously, and in general there is no direct correlation between confluence and evolutionary typologies. Another important dimension involves the number of political units participating in the interaction ranging from direct access, dyadic relations, exchange systems and open barter. The vertical complementarity literature in the Andes is valuable for considering prehispanic exchange relationships in that it has compelled a number of scholars to explore explicitly the relationship between ecological zonation, production, and social organization.

Vertical archipelagos

The work of John Murra on vertical complementarity has been among the most influential ethnohistoric studies in the Andes. The premise of the vertical archipelagos model is that the rapid altitudinal change along the flanks of the Andes produced a pattern where social groups residing in non-contiguous ecological strata formed distinct communities that developed around intensified production in these strata. Polities and ethnic communities sought to control a variety of these resource pockets following the Andean ideal of self-sufficiency.

Murra's (1972) seminal article distilled observations from ethnohistoric sources, in particular the visitacensus Garci Diez (1964 [1567]) of Chucuito in Puno, conducted only 35 years after the Spanish invasion. Murra showed that late prehispanic altiplano societies obtained direct access to products from a variety of ecological niches through this practice, and that the strategy was a guiding model of organization in some Andean polities (Salomon 1985). According to Murra (1985) the principal characteristics of the vertical archipelago model can be summarized as follows:

  1. Ethnic groups aimed to control numerous "floors" or ecological niches in order to exploit the resources that are found there and maintain self-sufficiency. As a consequence, powerful polities residing in the altiplano had considerable wealth in pastoral herds but also controlled land on both the western and eastern flanks of the Andes.
  2. The population was concentrated on the altiplano and those living in the peripheral 'islands' maintained continuous social and economic contact with the core region.
  3. The institutions of reciprocity and redistribution guaranteed that those living in 'islands' maintained the rights to products available in the center. These rights were defined through kinship ties that were ceremonially reaffirmed in the center.
  4. A single 'island' could be shared by different ethnic groups, though the coexistence could be tense.
  5. The relationships and functions of these colonies changed through time and may have become more exploitative. They are perhaps an antecedent to the mitimaqstrategy of the Inka empire.

While a chiefdom level of organization and centralized power is a principal characteristic of the Late Intermediate polities that practiced vertical complementarity in the period examined by Murra, the concept was been explored and expanded by archaeologists in the subsequent decades. In Murra's original description, verticality referred explicitly to direct control of a diverse resource base without engaging in trade with other ethnic groups, thereby preserving what Murra (1972) has described as the ancient Andean ideal of economic self-sufficiency that permeated Andean society and ideology far beyond his Lupaca case study. Stanish argues that Murra was explicit about excluding exchange processes, in part, because he perceived "a structural linkage between exchange and markets" (1992: 15). As prehispanic market mechanisms were absent in the south-central it was believed that barter and exchange were also of minimal importance and other means of articulation, such as direct control, were emphasized. Murra, however, later modified the definition to include specialized exchange centers, a change which Browman (1989: 324) argues confuses the issue because it subsumes a variety of processes into a single model.

On a theoretical level a further limitation of the original 'verticality archipelago' model lies in its adaptationalist orientation (Earle 2001).Adaptationist models of exchange and regional control have their basis in Service's (1962) proposal that these arrangements arise from environmental diversity, and then chiefs emerge to administer and redistribute goods produced regularly by their retainers.As observed by Van Buren (1996: 346), the archaeological origins and perpetuation of the archipelago pattern was founded on the assumption that groups benefit, as a whole, from the control of multiple tiers and the ecological resources that are produced in those archipelagos. As mentioned above in the discussion of administered trade, colonial documents emphasize the independence of commoners and the ability of subjects to practice subsistence barter, a pattern that leads authors to suggest that the vertical archipelagos pattern may have had more of a political basis than a foundation in ecological and subsistence practices. The ultimate roots of such a system may lie instead in the capacity of the rulers of such groups to organize larger scale trade and convert the value differential between products in the different ecological zones into political prestige through feasting and ceremony (Stanish 2003: 69-70).

Horizontal complementarity

Soon after Murra published his 1972 paper about vertical complementarity, researchers began noting that vertical complementarity is only one of a number of strategies employed in the Andes. A different kind of geographical interaction pattern, one that stays within broad regions such as the altiplano or the littoral, has been called horizontal complementarity. Contrasted with Murra's vertical complementarity model, in horizontal complementarity polities would directly control many parcels in a given niche. Here, the adaptationalist argument is somewhat less obvious since a horizontal complementarity strategy seems redundant unless particular resources were only available in one sector of a given horizontal territory. This phenomenon could reflect a cultural mechanism for the continued social integration of communities distributed widely across a vast region. An analogous situation can be observed in the culturally-constructed Yanamamö trade for ceramic wares, maintained to provide a social catalyst for villages to come together for feasting and marriage-making, despite the risk of conflict and treachery (Chagnon 1968).

Relations of horizontal complementarity between coastal valleys have been observed on the littoral (Netherly 1977;Rostworowski 1977;Shimada 1982), and for the higher altitude altiplano area where this kind of organization is referred to as the "altiplano mode of economic integration" (Browman 1977) and, to a lesser extent, the modern "extended type of Andean zonation" (Brush 1977: 12-16;Gade 1975). The extended type takes place in regions of expansive, contiguous production where exchange forms the basis for circulating products from other zones. In the expansive altiplano where there are fewer impediments to travel and, with the domestication of camelids, relatively low transport costs with cargo animals, materials may have been conveyed over substantial distances.


Figure 3-4. Subsistence exchange for products by ecozone versus single-source, diffusive goods.

A general schematic of these contrasting relationships is shown in Figure 3-4. While there is a great deal more complexity and variability to economy and exchange than is communicated in Figure 3-4, for example tubers are grown in the high elevation zones and herding does occur below 3800 masl, the purpose is to highlight the contrasting network characteristics. These two characteristics include

(1) Subsistence exchange.The acquisition of products available by zone contrasts

(2) Diffusive.The network for goods that radiate diffusively, such as obsidian and salt.

While the mixed agropastoral strategies are also very common, particularly in ecotone areas such as the puna rim, the subsistence exchange that articulates dedicated pastoralists with agriculturalists has been documented ethnographically in many studies of which only a selection are shown in Figure 3-4 (Bastien 1978;Brush 1977;Custred 1973;Flores Ochoa 1968;Palacios Ríos 1981;Winterhalder and Thomas 1978).

Ecological complementary has long served to transfer goods between ecozones, however the movement of goods laterally through areas of homogeneous resources, such as horizontal complementarity across the puna, implies other mechanisms of transfer. In addition to strategies described previously, it has been proposed that large periodic markets are a possible solution to the problems of social and economic integration among agropastoralists living in widely distributed settlements.

Browman (1990: 405-411) reviews the evidence for prehispanic and colonial period markets in the Peruvian and Bolivian altiplano. The colonial period aggregations in the Andes are products of historical circumstances, however Browman's evidence suggests that periodic fairs in some form were a feature of the prehispanic economy. While evidence concerning the actual goods exchanged at periodic fairs during the prehispanic period is scarce, these events could have been an effective way for obsidian to have been distributed in highland Peru and these fairs will be discussed in more detail below. Seasonal or annual gatherings are also ethnographically known among foragers in arid regions of the world with low population densities (Birdsell 1970: 120;Steward 1938). Social aggregations in some form can be expected to date back to the foraging period and these aggregations probably included, among other activities, the exchange of goods.

Evidence for barter, markets, and marketplaces in theAndes

"They did not sell… nor did they buy… they thought it was a needless operation" and all universally planted what they needed to support their households and thus did not have to sell foodstuffs, nor raise prices nor did they know what high prices were…(Garcilaso 1960 [1609]: 153; cited in Murra 1980: 142).

Was obsidian exchanged through a market system reflecting supply and demand in the prehispanic Andes? Were exchange specialists present in the prehispanic Andean highlands? If so, when did they appear and what did they transport? While extensive exchange has been documented in the prehispanic Andes, it is widely believed by Andeanists that exchange in the prehispanic central Andes was not based on market institutions. Building on the typology developed in Chapter 2 following Renfrew(1975)and Salomon(1985), differing degrees of specialization and independence can be expected to have existed among actors in exchange relationships (Figure 2-2andFigure 3-3). If specialized traders were present in prehispanic Andes, how did they interface with state authority during the Middle Horizon and Late Horizon, and what was their position during periods of regional conflict like the Late Intermediate Period?

Much of the evidence supporting the alleged lack of market-based exchange is derived from studies of Inka economic organization, where ethnohistorical accounts and archaeological datasets converge. A great deal has been written on the topic of Inka economy, and obsidian exchange was relatively diminished during this period, therefore this discussion will be limited to a few relevant issues regard exchange specialization. Several of the sixteenth century chroniclers are clear that while barter was widespread, the barter values of goods did not reflect fluctuations in supply and demand as in a market economy. However, the lack of consensus on the issue of markets and commerce during the Inka period stems from inconsistency in the cronistasthemselves. As reviewed by Murra(1980: 139-152)and LaLone(1982)the denial of a market exchange by Garcilaso de la Vega (quoted above) can be contrasted with numerous accounts of large and small markets, and a long tradition of barter exchange of various types.

The issue of Late Horizon marketplaces and market exchange is explored here because one of the principal questions that may be considered with changing obsidian distributions through time is the possible role of commercialism in prehispanic Andean exchange. The appearance of exchange specialists, such as freelance caravans moving certain commodities and responding to the changes in barter values that result from surpluses and shortages, would have presented a mode of transport distinctive from that of local reciprocal exchange or regional or state redistribution.

The vertical archipelago model functioned as an alternative to trade for goods from neighboring areas because "regional differences in production were, by preference, handled by means of colonization instead of through barter or trade."(Murra 1965: 201). As mentioned, Stanish argues that Murra initially excluded exchange mechanisms because of a perceived association between exchange and market economies. Further it can be argued, building on Appadurai(1986: 33), that in certain contexts of ranked or stratified societies with elaborate redistribution mechanisms, market systems of exchange represent a threat to the centralized ideological power of redistribution. "There is great advantage to leaders who are able to portray their resource-control strategies as reciprocity, redistribution, and generosity. Non-centralized resource-control strategies are, by definition, not 'control' strategies"(LaLone 1982: 296). If centralization is a principal determinant of state control on market exchange, were the peripheries involved in greater numbers of barter transactions?

Merchants in the Andes: Late Prehistoric and ethnohistoric evidence

Coastal Trade

The strongest evidence for merchant specialists in the Andean region comes from relatively peripheral areas of the Inka Empire, from the Pacific coast of what is now Peru, and from the coast and highlands of prehispanic Ecuador. The coastal Late prehispanic traders of Chincha, near the Paracas peninsula, have been described as consisting of 6000 merchants who traded in Cusco, among the Colla (and presumably the Collagua), and in Ecuador, but little is known about how these expeditions were organized(Patterson 1987;Rostworowski 1976;Rostworowski 1977;Sandweiss 1992). LaLone(1982: 308)notes that Rostworowski was not able to connect Chincha traders with marketplaces despite her assertion that these represented commercial exchange. Sandweiss(1992: 10)believes that Chincha trading expanded a great deal under the Inka following the Inka conquest of the Chimu to the north.

Coastal products such as spondylous and other goods were known to have been transported in large balsa rafts. There are numerous contact-period accounts of merchants plying the Pacific littoral beginning with the renowned loaded boat of balsa logs encountered off the coast of Ecuador by Pizarro on his second trip south, several years before the actual Spanish invasion of the Andes(Hemming 1970;Murra 1980: 140). The boat had a crew of 20 and had a small cabin and cotton sails. Murra is confident the boat was an Inka "registry" because the crew knew Quechua and a few were captured by the Pizarro's army who later used them as interpreters. While the Spanish paid particular attention to precious metals, the contents appear to have contained wealth goods including gold and silver ornaments, bracelets and anklets, headdresses and mirrors, and a great deal of cotton, wool and rich embroidery. A small weighing scale as well as a great deal of shell, probably spondylous and strombus, were found on board(Hemming 1970;Murra 1980: 140). Although the activity of this boat was described as trade (the word rescataris used in the text) by the Spanish observer Sámano-Xerez written in 1527-1528(Porras Barrenechea 1937: 21), it is highly likely that this boat, with its cargo of elite goods, was in fact carry ritual offerings from the Inka to some northern destination. The evidence for long distance exchange between Ecuador and Mesoamerica has been long been a topic of interest in New World archaeology(Coe 1960;Zeidler 1977).

Ecuadorian highlands

The strongest evidence for mercantilism and markets in the prehispanic Andean highlands comes from colonial Ecuador(Hartmann 1971;Salomon 1986). Hartmann argues that the Inka economy had a significant market component based the following evidence: (1) the Spanish saw gatherings that they identified as "markets" from the very earliest reports, although as they were coming from Mexico the Spanish used the Nahuatlword " tianguez"; (2) commodities were plentiful and varied, including both staple and luxury items; (3) both Quechua and Aymara had specialized terms for buying and selling; (4) market activity was not suppressed by the Inka authorities, only regulated to suit their interests(Hartmann 1971).

The existence of markets in prehispanic Ecuador is a particularly interesting question because the Quito area was conquered by the Inka only 30 years prior to the Spanish invasion and therefore the region had only recently been absorbed into the Inka Empire. Salomon(1986)examined the ethnohistoric evidence for precolonial and colonial markets in highland Ecuador and found that the contact period evidence provides insights into pre-Inka customs as well as the Inka response. The Quito valley is in a position to serve as a hub for the transfer of products from the Amazon lowlands, the Pacific coast, and the páramohighlands. In this sense, the Quito valley is in a similar geographical configuration, but on a smaller scale and a different ecological zone than the Lake Titicaca Basin. The strong dependence of early Spanish residents on the markets, and the founding of new markets by Spaniards, leads to some uncertainty as to the precolonial importance of markets. However, there is a variety of evidence for a pre-Inka merchant class in central Ecuador called mindaláesthat gathered in stationary markets and controlled trade in cotton, coca, and salt that they would bring from lower-lying regions(Salomon 1986: 203-204). Barter exchange between non-specialized traders occurred as well, typically of household surplus goods, and Salomon argues that both mindalamerchant organization and non-specialized barter were ancient developments in Ecuador. In contrast, Patterson(1987)argues that the merchantile organization was not a long-established system but rather a Late prehispanic period response to opportunities presented on the northern border of Tiwantinsuyu. Vertical archipelago organization is also found in Ecuador both in agricultural production and in targeted procurement communities such as stable colonies for salt production(Oberem 1981 [1976]: 79). However, Salomon shows that archipelagos were, in most cases, a late phenomenon that was introduced by the Inka. In addition, obsidian distributions in Ecuador can provide insights into Andean exchange in a context with functioning markets(Burger, et al. 1994).

As for trade specialization elsewhere in the Andes, LaLone(1982: 307)sees a latitudinal gradient from north to south where markets and freelance traders may have been more abundant in the northern periphery of the Inka empire, and subdued or non-existent in areas full under Inka control. A notable exception to this gradient are the sea traders from Chincha(Rostworowski 1977;Sandweiss 1992). The evidence is far less secure for the southern periphery of the Inka Empire, but the implication of the Ecuadorian data is that solving zonation problems through the vertical archipelago approach was promoted by the Inka in Ecuador following the Inka conquest, which perhaps calls into question the pervasiveness and pre-Inka antiquity of the vertical archipelago strategy in the south-central Andes as well.

Traveling Peddlers in the south-central Andes

In contemporary contexts, peddlers are found with frequency in areas that form boundaries areas between different commercial spheres of interaction and lacking in consistent distribution of goods. Browman (1990: 422) reviews ethnographic evidence for mobile peddlerswho perform bulk-forming and bulk-breaking services in the south-central Andean highlands. The peddlers will provide manufactured items to rural pastoral communities, though often at a substantial mark-up, and will trade for items like hides and wool in time for purchasing fairs in the regional centers. The puna between Lake Titicaca and the western slopes in Arequipa are particularly active with comerciantes ambulanteswho schedule their travel cycles to correspond with patron saint festivals, as well as distributing goods to communities without regular markets (Flores Ochoa 1977: 148 ;Flores Ochoa and Najar Vizcarra 1976) as well as traveling herbalists and related groups (Bastien 1987). In the small community of Cerrillos in southwestern Bolivia near the Argentinian border, Nielsen (2001: 166) reports that peddlers, referred to as cambalacheros, would pedal bicycles from the city of Oruro bringing clothes and metal pots to sell or to barter for hides.

The Collaguas of the Colca valley were frequently visited by itinerant peddlers from Puno, according to Casaverde (1977: 185). These vendors known as polveñosmaintain established compadrerelationshipswith Colca valley households in order to have reliable hosts and potential buyers in the valley. Although transactions frequently take place through barter and the host and other residents are not obliged to trade, the profit motive of the peddlers is understood. These examples illustrate some of the variety in forms of distribution that may have had some basis in the prehispanic economy.

Limitations of ethnohistory and ethnography

Ancient economy in the Andes has been most fruitfully studied by combining archaeological evidence with ethnohistoric and contemporary sources. General models of great importance to Andean studies such as vertical complementarity, the question of the commerce in the prehispanic economy, and a structure of dual organization in the Inka administration are examples of models emerging from ethnohistoric sources but with empirical support from archaeological research. Ethnohistorically based models also have their limitations in that they can unduly influence interpretation much in the way that the "tyranny of ethnography"(Wobst 1978)should not necessarily define the range of archaeological possibility and inference. Ethnographic sources are valuable for the evidence of pastoral patterns, the priorities of caravan drivers, and the articulation between mixed and specialized pastoral and agricultural practices. However, modern features including the presence of the cash economy, markets, truckers plying the highways, and various modern job opportunities impact the structure of exchange relationships and caravan transport(Browman 1990;Nielsen 2000).

Ethnographic studies of contemporary llama caravan drivers are focusing, by necessity, on relatively marginalized communities that are sufficiently conservative to continue to herd llamas despite a variety of often more lucrative alternatives (Lecoq 1988;Nielsen 2000). Scholars have noted, however, that caravan drivers enjoyed relatively high status and autonomy in the late prehispanic and early colonial period which contrasts strongly with the economically marginalized modern day caravanero(Murra 1965). Furthermore, herding has become relatively low status in many regions of the contemporary Andes because the economic focus has moved downslope since Spanish contact due to the growth of the coastal economy and the importation of various low altitude crops and livestock. In other words, the prominence of caravan drivers as central economic agents has been greatly diminished in recent centuries, and ethnographically documented interactions with agriculturalists is one circumstance in which this relative shift in power may result in distorted perceptions of prehispanic interaction patterns. The relatively high status accorded to truck drivers and other purveyors of goods and information in contemporary Andean villages was likely to have been ascribed, instead, to the relatively cosmopolitan drivers of llama caravans during prehispanic times.

3.2. Long distance trade

The transport of goods over distances that exceed immediate complementarity relationships in mountain environments is well-documented archaeologically and ethnographically in the south-central Andes. Point specific resources like obsidian and salt have a distinctive, radiating distribution pattern as compared with subsistence exchange between ecological zones (Figure 3-4). Mechanisms that include direct acquisition and down-the-line exchange (boundary reciprocity) are likely to have been long-term exchange modes that served to disseminate of goods horizontally through a single ecological setting like the Andean altiplano. However, it is known ethnohistorically that long distance transport with the aid of camelid caravans, either with direct procurement and including very few transfers (long distance trade caravans), was a common method for the lateral distribution of goods. When did long distance caravan transport begin to dominate regional exchange in the south-central Andes, and who initiated this form of transfer between far-flung populations? Principal factors that influence the origin and perpetuation of long distance trade routes by highland pastoralists in the south-central Andes include the following features.

(1) Cargo animals:While not exceptionally strong, llamas are effective cargo animals because they are relatively compliant, they are not water tethered, and they can consume a range of grasses found on the altiplano so that they do not have to transport their own fodder.

(2) Topography:By virtue of the open and predominantly low-angle topography of the altiplano, the movement of loaded cargo animals across the altiplano requires lower effort than travel along the eastern or western sierra that are bisected by deep valleys.

(3) Resources:On an inter-regional scale, the altiplano divides complementary resource areas from the Amazon lowlands to the Pacific Ocean and these converged on the altiplano during particular time periods.

These features created circumstances that allowed for the wide-distribution of materials like obsidian and other products in the south-central Andean highlands. Long distance exchange and spatial relationships have been presented as a primary factors in the appearance of early social complexity during the Middle and Late Formative (Bandy 2005;Stanish 2003: 159-164). However, given the antiquity of camelid domestication in the Andes the long distance caravan pattern probably predated the Middle Formative by a millennium or more. The presence of caravans and the transport of complementary goods around the high, flat altiplano are part of a number of characteristics that created the circumstances within which social inequalities evolved. Competition for social power emerged during the Formative from a context that included these features of long distance exchange both in terms of the capacity for regional interaction, and the social institutions that surrounded the organization and scheduling of exchange in the region. These regional exchange mechanisms had long term consequences based the theory of Clark and Blake (1994: 17) who argue that social ranking was the unintended outcome of early political actors, operating within the institutional constraints of their circumstances, pursuing short term prestige goals for themselves and for their supporters. Following this model, the established circulation mechanisms of non-local products during the Early Formative, and perhaps earlier, are likely to have had a significant influence on the strategies pursued by aggrandizers during subsequent periods such as the Middle Formative.

3.2.1. Household-level caravans

Adaptationalist explanations for the origins and significance of long distance caravan networks are unsatisfactory, yet the more explicit aggrandizer models for the rise of elite-administered caravans that are documented ethnohistorically, refer to a regional-scale phenomenon that occurred relatively late in the prehispanic past (Stanish 1992: 14;Stanish 2003: 69). As was discussed above, the evidence suggests that household level organization of long distance caravans should be considered as a possible hypothesis for the long distance circulation of goods prior to the Middle or Late Formative when archaeological evidence for ranked society appeared in the Titicaca Basin.

In terms of the two configurations described above as subsistence exchangeand the single-source diffusivegoods (Figure 3-4X), the capacity and incentive for long distance caravan activity appear to be much older than the evidence for elite administration. In other words, on a geographical level and in terms of archaeological distributions one should consider that long distance caravan exchange was possibly organized on the household level thousands of years before elites were clearly organizing labor for the construction of monuments and the hosting of feasts.

What were the contexts in which individuals and households to began to organize long distance caravans in prehistory? Many of the "diffusive" goods with radiating distributions were not required for subsistence, strictly speaking. While small quantities of salt are biologically necessary for humans, salt is available in low densities in many parts of the altiplano and hunters and pastoralists can actually acquire salt from the consumption of meat and blood. In terms of the need for obsidian, it is evident that obsidian has flaking characteristics not available in other stone materials, but high quality cherts are available in many regions where the archaeological evidence shows that obsidian was imported from relatively large distances. Semi-precious stones like lapis lazuli are found in archaeological contexts such as burials after 3300 BCE in the south-central Andes. A variety of perishable goods probably circulated along such exchange networks as well, and as wild plants and agricultural products from particular regions are renowned to this day, such as hot peppers, sweet corn, herbs, and potent coca leaf, there was likely to have been variability in products available from particular valleys over others (especially during the earlier stages of plant domestication). If these products were transferred by caravans, perhaps in dried form, they would have radiated along distributions that were closer to diffusive sources like obsidian.

At risk of creating a rigid dichotomy, these items that were relatively common-place in some regions may have been closer to "prestige goods" because they were transported substantial distances. Following agency models for early leadership, those who mobilized resources to acquire such goods may have been striving to differentiate themselves from their communities and their neighbors. However, it is important to note that many of these goods, such as obsidian, do not appear to have been used exclusively by a small, restricted segment of the population in the archaeological contexts where they are found. Nevertheless these goods were widely transported and the mere presence of these goods indicates that some effort was expended to acquire the product (and, according to Hayden (1998: 44), these are automatically a form of prestige technology). In terms of the importance of "ordinary goods" (Smith 1999, see section 2.2.2), such products may have had important cultural and ethnic associations that lent value to the acquisition of these goods and prestige to those who could acquire them while simultaneously having widespread availability in a community. Archaeologists often define "prestige goods" based on spatially-delimited contexts that imply restricted or elite contexts of consumption, and this sometimes leads to claims that these alleged prestige goods were the impetus for long distance exchange. Rather than inferring that prestige goods defined the exchange networks and then "everyday goods" followed suit, Smith (1999) notes that the cultural information and these non-local yet widely-consumed materials like obsidian are perhaps better considered as "cultural goods", linked to ethnicity signaling and practice among herders (Nielsen 2000: 521-526), and complemented by other markers of ethnicity the most visible of which were probably textiles. As regional links became established, obsidian may have become a means for demonstrating participation and cultural affiliation with the subsistence level networks that exchanged goods including obsidian. Even if archaeologists are not able to detect exclusivity in the access to such goods in intra-site spatial distributions, as the materials were widely consumed in the community, it seems probable that the ability to acquire these goods reflected positively on the individual or household capable of procuring non-local products.

3.2.2. Incentives for early caravan formation

Down-the-line exchange can account for many archaeological distributions of cultural goods such as obsidian, both in volume and in temporal persistence. In terms of efficiency, down-the-line reciprocity networks require lower travel distance and less risk than is incurred traveling through unfamiliar territory (Renfrew 1975: 44). Furthermore, as Nielsen (2000: 24, 514) notes, caravan driving is hard work and it only becomes worthwhile when distances are large and/or loads are heavy or bulky. These bulky items tend to be transported medium distances (40-120 km), but smaller and more valued items might be transported long distances (Nielsen 2000). Traveling peddlers also carry small and light items without the added responsibility of managing many camelids in a caravan, therefore in some ways caravans and peddlers are complementary alternatives for the independent circulation of goods. Nielsen further observes that ancient caravans probably carried a variety of products in their cargo to reduce risk, and they likely carried anythingthat was worth transporting and that could be traded within their social and cultural parameters.

Given the evidence for increased sedentism from the Terminal Archaic and onwards (Aldenderfer 1998;Craig 2005), down-the-line exchange may be considered as the null hypothesis, with direct procurement and household level trade caravans as alternative hypotheses. Consistent but low levels of down-the-line exchange embedded in social relations form the background against which to examine other exchange modes such as caravans and possible non-market mercantilism. What kinds of possible incentives existed for the development of household-level caravans in order to acquire such goods, given the wide variety of potential trade and consumption patterns in prehistory? Incentives for household level caravan organization may include

(1) The maintenance of a social network and a demonstration of extra-local alliances that exceed the level of basic regional interaction implicated in subsistence-related exchange.

(2) Demand for a greater quantity of types, and the number or sizes of products than can be acquired from down the line exchange.

(3) A need or desire to exceed the products available to one's neighbors or to avoid dependency on an entire reciprocity network of neighborly relations.

(4) In a biological adaptationist framework, the risk and potential of caravan driving was a form of costly signaling (Aldenderfer 2006;Craig and Aldenderfer In Press).

Secure evidence of caravan activity, beyond simple evidence of pastoralism from faunal remains and evidence of corrals, may be a challenge for archaeologists and the systematic study of spatial distributions of non-local goods may shed light on the differences between subsistence pastoralism and caravan trade activity. One example of caravan activity from consumption contexts may come in the form of distinctive evidence of pooling of certain goods at particular sites along major transportation corridors, but establishing changes in relative density probably requires representative samples from archaeological collections from contexts that were presumably major caravan networks and those that were more peripheral to principal caravan routes. Sourcing from contemporary household-level assemblages may serve to differentiate such pooling (Figure 2-4), as was attempted with Oaxaca obsidian by Pires-Ferreira (1976). In the south-central Andes, stronger evidence for differentiating caravan trade from down-the-line trade may come from differences in artifact form and context from a variety of non-local products. In short, differentiating caravans from down-the-line exchange will require inference from multiple lines of evidence and from domestic contexts, evidence that is extremely scarce in the south-central Andes.

3.2.3. Exchange between herders and farmers

Ethnographic accounts of the interaction between caravan drivers and agriculturalists may be colored by the comparatively low status of caravan drivers in modern circumstances. Among isolated agricultural valleys in prehispanic Andes caravan drivers likely represented an important link to both information and non-local products. As the owners of the means of transport and the initiators of long distance interaction, caravan drivers were in a strong position to influence trade negotiations with dedicated agriculturalists that did not have their own llama herds or did not have schedules that permitted them to undertake long voyages. Strong links and commitments between herders and particular agriculturalists are, then, in some ways against the interests of herders because in a market context they "held the cards" in terms of negotiating favorable conditions of exchange. From the perspective of regional economic interaction two principal groups that, as per Dillehay (1993: 253), largely complement one another and may result in a relatively stable political and social environment:

(1) Mobile herders with an economic focus on hunting, pastoralism, and who are limited in their scheduling by the needs of the herd and annual cycle caravans.

(2) Relatively sedentary agricultural communities, often with an animal husbandry component as well. These households are largely restricted in their scheduling and long distance travel due to the requirements of agriculture.

In this configuration the mobile herders have greater autonomy, but they depend on articulation with dedicated agriculturalists. The relationship between mobile pastoralists and sedentary agriculturalists in Old World contexts has been the subject of comparative study by Khazanov (1984: 198-227). One common pattern is for trade between nomads and sedentary society to be manipulated for political purposes by administrators and elites in the regional centers. In the fifteenth century, the Chinese state sought to regulate exchange with nomadic tribes occurring at trading posts on the frontier. A recurring pattern occurred where the Chinese government would attempt to control the nomads by restricting trade, and nomads would, in turn, "acquire the right to trade by using arms." (Khazanov 1984: 206). That is, if exchange was curtailed the nomads would resort to violent means as described by Levi-Strauss (1969: 67) and Sahlins (1972: 302) who note the link between exchange and warfare (Section 2.2.5).

In the Near East, nomads were in a more profitable position because often they were the essential link between isolated oases. While administrators may have sought to control nomads trading with their communities, the demand by farmers for the milk and meat products from nomads, and the transportation services offered by nomads, often placed the herders in an advantageous position (Khazanov 1984: 208). In the Andes, a similar pattern linked agricultural valleys like the Colca in the high sierra, the mid-altitudes, and perhaps the littoral, by way of camelid caravans, although by the Late Prehispanic it appears many valley communities had their own large herds pastured in the adjacent puna.

3.2.4. Types of products carried by caravans

Ethnographic and ethnohistoric accounts of the types of goods carried by llama caravans provides a context for discussing the volume of trade and spatial relationships in the prehispanic period. Nielsen (2000: 65-67) observed that caravan drivers transport virtually anything that they think they can transport and will be able to trade, and therefore the argument that certain caravans were transporting purely subsistence goods, while others carried purely prestige goods, is probably unfounded. Nielsen notes that as part of their diversification strategy, caravans in prehistory probably variously transported subsistence goods, cultural items, and prestige goods as they articulated with networks at different levels with a diversity of social associations.

Informative ethnohistoric evidence for caravan transport comes from reports describing the provisioning of the infamous silver mines at Potosí that are reviewed by Browman (1990: 408). These reports state that 40,000 llamas were reserved by Potosí for provisioning and another 60,000 were brought as support for indigenous workers fulfilling their tax obligations through labor. Although with these Potosí data it is difficult to extrapolate from the substantial transportation requirements for mining and ore milling, and the demands of the Spaniards overseeing the mining operations, to a conception of goods that may have circulated during prehispanic times, these data are informative on the variety of items mobilized for the mining effort. Goods included manufacturing items such as cloth, wool, wood and dung as fuel, and building supplies. Subsistence goods included potatoes and ch'uno,meat, maize and chicha, and various fruits and vegetables. Goods that could be classified as cultural / prestige items included herbs, medicines, stimulants including quantities of coca leaf, and hallucinogenics like ayahuasca(1990: 408).

In Mesoamerica, where cargo animals were not available, human bearers carried goods for hundreds of kilometers and canoe transport was used extensively. Drennan (1984: 110) observes that textiles may have represented a significant portion of the goods being transported long distance in the prehispanic period. Similarly, woolen textiles in the Andes were an important trade good for highland pastoralists and probably represented a substantial part of the goods offered for barter between pastoralist caravan drivers and agriculturalists from the beginning of the mutualistic relationship between pastoralists and agriculturalists. This demand for textiles would have been especially strong among agriculturalists living outside of the cotton-producing coastal area. The domestication of camelids for cargo transport sometime in the latter part of the Archaic allowed for more efficient transport of goods throughout the remainder of the prehispanic period. However, this is not to suggest that daily staples were transported, as in modern circumstances where fruits are trucked to the altiplano markets where they are bought for daily consumption.

3.2.5. Caravan travel distances and speeds

Ethnoarchaeological studies provide details on the more immediate decision making practices of caravan drivers including the daily routine, the rate of travel, and the scheduling of rest days. References in the ethnographic and ethnohistoric literature concerning to the velocity and capacity of llama caravans provides benchmarks for estimating the rate of travel in prehispanic times. There is some variability in the reported weights, speeds, and distances in the ethnographic and ethnohistoric literature (Bonavia 1996: 501-515).






Caravan distance calculated by coca chews

3 km level, 2 km uphill

Approximately 40 min per "cocada", 6-8 per day.

(Zahm 1911: 180)


Up to 40-45 kg

(Tschopik Jr. 1946: 533)

Llama caravan (ethnohistoric)

75 - 100 lb loads (34.1 - 45.4 kg)

10-12 miles (16-19 km) / day

(Murra 1965: 185)

Llama caravan

25-30 kg

(< 40 kg)

15-20 km / day

Ten hour marches. A long trip can last 30 days.

(Flores Ochoa 1968: 118, 130)

Llama caravan

35 kg

25 km / day (150 km journey)

From 8-9am to nearly 4 pm, or 8 hrs per day

(Custred 1974: 277)

Llama caravan (ethnohistoric)

11 miles (17.7 km) / day

1 day: from daybreak to noon

(Murra 1980: 48)

Llama caravan

25 kg

20-25 km / day

1 day: 7 am to 2-3 pm.

(Lecoq 1987: 8, 20)

Llama caravan

25 kg

15-20 km / day

Journey: 2-3 months

(Flannery, et al. 1989: 106, 115)

Llama caravan (ethnohistoric)

10-20 km / day

1 day: From dawn until early afternoon

(Hyslop 1984: 294-298, 302)

Llama caravan (model for long trips)

30 kg

20 km / day

6 days a week of travel

(D'Altroy 1992: 85)

Llama caravan

15-20 km / day

1 day: From dawn until early afternoon.

(Browman 1990: 403)

Llama: Lighter loads

25-35 kg

300-400 km journey

Journey: 2-3 months

(Browman 1990: 403)

Llama: Heavy loads

50-60 kg

Short distance

Short duration

(Browman 1990: 403)

Llama caravan (salt blocks)

23 kg

15-25 km / day

From 4-5am to 2-4pm, or 6-9 hrs/day, no stops

(Nielsen 2001: 169-176)

Table 3-1. Reported llama caravan loads, distances, and times.

Caravan drivers generally arise at first light and begin preparing for the journey and loading animals for an early departure. Caravans often travel until early afternoon when camp is established and the animals are allowed to graze. As camelids do not pasture at night (whereas Equusdo) ample time must be provided for animals to feed during the afternoon in order to avoid stressing the animals (Nielsen 2000: 446-449). Rest days are taken regularly on caravan routes that exceed six days, with Nielsen (2000: 461) reporting one rest day for every three to five days of travel (Lecoq 1988: 185-186;West 1981: 70). The top priority with respect to nightly campsite selection is the needs of the herd animals. Quality pasture is sought for the animals, the next priority is sufficient water, and additionally the emotional condition of the llamas is considered as reported the llamas can be restless in certain camps. Subsequently the needs of humans are considered including hunting opportunities, trade opportunities, and the comfort of the camp. Thus, while economic and social demands frame the larger scale decisions of caravan routes and products to transport for trade, the needs of the herd animals dominate in short term decision making (Nielsen 2000: 490).

Male llamas are larger and caravan animals are typically castrated males, based on some reports, but "left intact" according to others. Flores Ochoa (1968: 118), reports that castrated llamas produce better meat and wool, but non-castrated llamas make better caravan animals. According to most other reports caravan llamas are castrated because they are stronger and tamer, and this practice allows herders to manage mixed herds (Browman 1990: 398;Nielsen 2001: 168;West 1981: 66).

Estimating travel velocity

A computer model for estimating travel speed based on topography, where velocity over a segment of trail is calculated as function of slope, could be derived using the function presented by Tobler (1993) for hiking and horseback riding (the source of cost paths shown in Figure 3-5). While these topography based calculations have serious limitations (Connolly and Lake 2006: 252-255), for general estimates over larger regions with measurable changes in terrain steepness these estimates are superior to the simple use of slope for estimating velocity. Such models would preferably be derived using original data from fieldwork, perhaps based on a contemporary study that takes into account the size and weight of the cargo animal, the amount and type of cargo, and the performance of the cargo animals based on trail conditions mapped using GPS receivers.

In the course of ethnoarchaeological fieldwork accompanying a llama caravan in Bolivia, Nielsen (2000: 449;2001: 184) notes that there were differences in the going and returning portions of a 2-3 month caravan journey that had implications for the overall travel speed. The out-going trip involved visits with companions and a variety of rituals at propitious locations along the route. During the return journey, in contrast, no rituals were performed but the animals moved more slowly because they were carrying large loads of produce uphill from the eastern lowlands; and they had been traveling for months and reportedly their feet hurt. Thus the lack of ritual performance and visiting and bartering saved time, but the walking speeds were slower, breaks were longer, and layover rest days were more frequent and longer.

Evidence of relative travel distances in the region

In the Mantaro region of the central Andes, Earle (2001: 310) notes that long distance exchange (classified as taking place over a distance of 50 km or greater, to lands beyond those held by the local ethnic group), rarely includes staple items. During the Late Intermediate and Late Horizon, items that were found in the Mantaro area include items of metal and shell described as "wealth items", emphasizing the political nature of long distance exchange in this period.

While Colca valley obsidian was primarily conveyed throughout the Titicaca Basin during prehispanic times by unspecified groups, during the colonial period Colca valley polities are known to have been dominant caravan operators in the region. Residents of the Colca valley owned large camelid herds and they were responsible for extended caravan journeys to the altiplano during the Colonial period. Documentary evidence shows that the Collaguas from the main Colca valley initiated caravans to transport products, such as Arequipa wine to Cusco and Canas, and Colca corn to "wherever they would like to sell it" (Crespo 1977: 56). A testament to the role of the Collaguas, and the Colca valley in Arequipa more generally, comes from Toledo's (1924 [1570-1575]) visita generaldescribing populations of Colca community members resettled as farmers in the new city of Arequipa, Yanahuara, and environs. Toledo states that these Collaguas mitimaesgroups should produce wine in these lower-altitude areas for transport to Cusco and to the mines of Potosí in Bolivia over 700 km to the south-east (Málaga Medina 1977: 114).

(1) Patio exchange

Information exchange concerning the relative barter values for goods are subsumed by the intimacy of the exchange context. In ethnographic accounts of exchange the negations occur predominantly in the privacy of domestic patio areas, and not in a public forum, such as a marketplace, where prices and equivalencies are public knowledge (Humphrey and Hugh-Jones 1992).

(2) Temporality of caravan travel

Caravan travel involves negotiating long distances with unforeseen delays. Due to the difficulty in scheduling encounters between caravans as highly mobile segments of the population, settled villages probably served as important nodes in the fluid regional interaction network (Dillehay and Nuñez 1988). The manner in which these temporal factors intersect with the diachronic nature of reciprocal exchange is counter to the synchronic expectations of establishing immediate, market-based equivalences for goods.

When caravan drivers do not have established trade partners in a settlement they may initiate a new trade relationship with unknown partners upon entering the community. Casaverde Rojas (1977: 177) describes how women in Cabanaconde, in the lower Colca valley, would besiege the first caravans to arrive seasonally at the village entrance with offers. The women would attempt to establish trade partner relationships by negotiating favorable terms for agricultural goods and a place to corral the caravan animals in exchange for labor and pastoral products. Nielsen (2001: 183) states that when market prices fluctuate then contemporary caravanners in Oruro, Bolivia, will sometimes avoid established trade partners in order to better pursue profit opportunities with the greater variety offered by modern market forces; the compadrazgoinstitution seems to be waning. As Danby (2002) and M. E. Smith (2004) both argue, commercialization and alienability of products in all ancient economies should perhaps best be considered in terms of degrees and not in absolutes.

3.2.6. Circuit mobility and role of the periphery

The circuit mobility model of Nuñez and Dillehay (Dillehay and Nuñez 1988;1995 [1979]) conceives of the development of regional interaction in the south-central Andes in terms of decentralized "circuits" traveled by regular camelid caravans. This articulation between far-flung communities is envisioned as beginning in "Late Archaic" (rather, during the "Terminal Archaic" using the terminology in this dissertation) and Formative times.

Herder-caravan societies moved in fixed spiral-like transhumance paths between two or more axis settlements either along a puna-to-puna vector, a puna-to-coastal vector, or a puna-to-selva vector… Continuity and stability was given to the circuit herder-caravan movement by settlements at both ends of its pathway. For this movement to have maintained equilibrium, its pathway must have been balanced by relatively homogeneous, fixed axis settlements which offered multiple resources and services from their particular ecological zone and by ferias (or fairs) where goods were exchanged (Dillehay and Nuñez 1988: 611).

This historical model is of importance here because it highlights the limitations of a core-periphery focus for addressing certain regionally distributed processes like the emergence of control over regional exchange routes during the Formative in the Lake Titicaca Basin (Nielsen 2000: 88-92;Yacobaccio, et al. 2002: 171-172). In models of Late Formative complexity in the Titicaca Basin presented by Browman (1981) and, to a lesser extent, Kolata (1993: 274), caravans and long distance trade play a prominent role. In Browman's formulation, the core areas of regional centers became increasingly powerful due to craft specialization guilds and other institutions that reached their apex at Tiwanaku, albeit many of these expectations for guilds have not been borne out in more recent research (Isbell 2004: 216;Rivera Casanovas 2003). Kolata holds it was the productivity of raised-field agriculture that formed the principal economic mechanism behind Tiwanaku's florescence, with caravan based articulation being a secondary component. Both scholars emphasize the dominant position of the core areas of regional centers, a position that on a regional scale was ultimately attained by the Tiwanaku state. In contrast to these models that emphasize centralization, the Nunez and Dillehay model holds that it is integration by way of caravan trade routes themselves, and that these routes developed into "leading circuits" when they served to connect important centers. Further, as traditions became established, the relationships between circuits and principal settlements along these routes provided temporal continuity and stability to a system that is otherwise fluid and mobile. Their model gives more autonomy and influence to this integrating, caravan-based element in society such that "the sedentary (or axis) settlements of the population were maintained (and often created) and controlled by the mobile sectors." (Dillehay and Nuñez 1988: 621).

This model also contains weaker points, such as an adaptationalist underpinning for the origins and incentives for participation in these caravan exchange networks, and the model ascribes the emerging dominance of Titicaca Basin centers as resulting from their environmental characteristics. The system is described as "harmonious and cohesive" (Dillehay and Nuñez 1988: 620) as caravans are used to efficiently spread patchy resources over a wider region. The adaptationalist interpretation of harmony is based on a lack of archaeological evidence for warfare in the Formative, with the Andean tradition of reciprocity serving as the cohesive force. Further, the system segregates discrete "highland" and "lowland" communities, while archaeological evidence supports a gradient with no clear demarcation.

A more current theoretical orientation for this model would focus on the motivations of traders, the influence of non-local goods in axis communities, and the status acquired by caravanners based on the importance of their role on a regional scale. The emergence of this distinctive, dispersed economic form that appears simulatenously with beginnings of social inequality during Terminal Archaic and Formative may shed light on the underpinnings of social inequalities that developed during the Formative. Further, a greater exploration of the strategic relationship between aggrandizers in Titicaca Basin regional centers and these long distance traders may provide evidence for how the first elites in early polities were able to coordinate labor and garner resources from the region. Despite the historical and adaptationalist focus, the Nuñez and Dillehay model brought a focus on the role decentralized networks in connected emerging settlements that grew to have regional influence. This emphasis on the mobile sector of society highlights the potential influence in both resources and political support that could come from second-tier communities dispersed across the large expanses of the altiplano in the political strategies of early elites in Titicaca Basin centers.

3.2.7. Compadrazgo relationships and commerce

A trade relationship termed compadrazgo,or simply conocidos,is reported between regular trade partners; typically between those who live in complementary ecological areas (Browman 1990: 404-405;Flores Ochoa 1968; but see ;Nielsen 2000: 437-438;Nielsen 2001: 182-183). For example, if a llama caravan driver from a particular area of the puna and a farmer in a mid-altitude valley with a variety of products have regularly exchanged goods over the years, and then they teach their children of the relationship using fictive-kin terms; a tradition of mutualism is established between herder and farmer that can potentially last for generations. The relationship offers stability and predictability to both sides of the exchange in barter rates, types of goods, and quantities to be exchanged.

The nature of this encounter is critical to understanding Andean reciprocity relationships and the degree of alienability of goods being exchanged (Burchard 1974;Mayer 1971). As mentioned above, the caravan driver has mobility and choices in terms of travel routes and communities where to partake in exchange. The maintenance of long term exchange partners through compadrazgois therefore a constraint on caravan autonomy. Two characteristics of compadrazgo relationships appear to underscore the embeddedness of the interaction:

Yapa and over-reciprocation

Barter relationships are often cemented with a yapa:a little bonus given to the buyer that takes the form of an over-reciprocation to assure future transactions (Browman 1990: 421;Sahlins 1972: 303). The magical powers attributed to the yapanotwithstanding, Sahlins' (1972: 308-314) develops a functional economic explanation for over-reciprocation where it serves a similar mechanism to price fluctuations in market-based societies. When an over abundance of product A relative to product B exists in a barter situation based on traditional equivalencies (hence, a lower value for A may result, in a market system), the provider of product A may over-reciprocate and thus, based on the morality of reciprocal arrangements, guarantee future compensation from the trade partner.

The primitive trade partnership is a functional counterpart of the market's price mechanism. A current supply-demand imbalance is resolved by pressure on trade partners rather than exchange rates. Where in the market this equilibrium is effected by a change in price, here the social side of the transaction, the partnership, absorbs the economic pressure. The rate of exchange remains undisturbed - although the temporal rate of certain transactions may be retarded (Sahlins 1972: 311).

The discrepancy that must be resolved synchronically in neoclassical market economics is resolved diachronically in reciprocal arrangements (Danby 2002). Browman (1990: 421) does not believe the over-reciprocation device described by Sahlins is in evidence in the Andes. As Browman observes, there is ethnographic evidence that suggests that barter rates do, in fact, fluctuate in response to supply and demand. The arrangement described above is one possible configuration that occurred in prehistoric circumstances, however, and it is a possible means of assuring the long term persistence of exchange relationships (Burchard 1974;Mayer 1971).

Seasonal fairs and the temporality of caravans

If seasonal fairs and aggregations were a feature of the prehispanic altiplano, as discussed by Browman (1990), interactions may have taken notably different forms in those contexts. Seasonal fairs may have had the significance of religious festivals in the contemporary Andes where the devout sometimes travel for weeks in order to arrive at auspicious times. Fairs and cultural occasions may, then, have been blended with economic transactions.

Scheduled festivals with elaborate dances, music, and costumes are a major cultural contribution in contemporary altiplano communities like Paratía (Flores Ochoa 1968) and despite the lack of simple material correlates for archaeological study, cultural items like song and dance were probably significant features in a variety of prehispanic reciprocal exchange contexts (J. Flores Ochoa 2005, pers. comm. July 2005). Despite the relative marginalization of altiplano cities in the modern economy (or perhaps a reflection of this marginalization), traditional festivals endure as important cultural features in the Titicaca Basin. Citing early twentieth-century sources, Browman (1990: 409) reports that at major shrine at Copacabana, Bolivia, between 40,000 and 50,000 "traders" would converge at times scheduled to coincide with ceremonies at the shrine.

If economic transactions occurred in association with these festivals in prehispanic times, either as a central feature or relegated to the periphery of the cultural events, the transactions may have assumed certain characteristics of marketplace exchange. These characteristics would have included public knowledge of barter equivalences and perhaps more immediate, synchronic exchange due to the short time period of convergence at the festival. As mentioned, however, marketplace concentrations do not necessarily imply true "market economies" with fluctuating prices reflecting supply and demand (LaLone 1982). Assuming that economic transactions that may have occurred at these fairs did not create moral conflict (by debasing sacred ceremonies with lowly economic transactions, in a Euro-American perspective) they would have created an excellent context for the transfer of both cultural goods and prestige items, and for the control of certain exchange practices by administrators or elites. Nevertheless the problem remains that dedicated agriculturalists with harvest goods for exchange would have been absent from these fairs on the altiplano because dedicated agriculturalists would not have the schedule or the herd demographics that would have permitted them to initiate long distance caravans. Therefore a variety of strategies probably developed to allow the transfer of products with the emergence of caravans that traveled, on the large scale, according to schedules dictated by seasonal gatherings, harvest schedules, and other economic and cultural circumstances. These developments imply the emergence of something of a continuum between the more alienable exchange that occurred in seasonal gatherings, and more inalienable barter that occurred in the intimate exchange context of compadrazgorelationships.

3.2.8. Discussion

The environmental and cultural context of the south-central Andean highlands framed the circumstances in which emerged the long distance traffic in various products during the prehispanic period. The domestication of camelids sometime in the past 6000 years allowed for more efficient transport of bulky goods. There remain many unknown aspects to the network that articulated dedicated agriculturalists with pastoralists in the prehispanic past, however archaeological and ethnographic evidence allows for inference regarding the following major points:

(1) Network configuration.Products available by ecozone were transported in numerous, segmentary articulations between ecozones while other products, only available in a few locations, were apparently conveyed diffusively and were between transported ecozones and across homogeneous terrain like the altiplano (XFigure 3-4X).

(2) Motivations for change in mode of interaction.The domestication of animals and plants, changes in sedentism, and the development of social inequalities were some of the factors that contributed to development of long distance caravans. It is evident that the original modes of interaction: Direct acquisition and down-the-line exchange, were supplemented by household organized caravan trade, and finally administered caravan trade, but the timing of these changes is difficult to establish with precision.

(3) Means of trade.Features of the Andean barter economy such as enduring trade relationships between households in complementary zones cemented by institutions like over-reciprocation and fictive kinship, are well demonstrated ethnographically in the region. However, seasonal market-like gatherings are also reported in the region and exchange of goods in those contexts may have been more alienable, and transactions may have been more synchronic in nature. Such gatherings may also have had evolutionary significance because they potentially relate to the development of early leadership in regional centers including ceremony, feasting, the use of monumental architecture, and centralized control of trade in certain goods.

The persistent themes in mountain agropastoral settings of non-autarkic economies and risk reduction through mixed subsistence strategies serve as a reminder that variability was probably the rule in exchange relations as well. Absolutes in exchange patterns were probably rare, and a degree of both self-interested trade and elements of embedded, social and symbolic fraternity likely existed between trade partners since early antiquity. While a number of plausible models have been proposed for the later Prehispanic periods where both household-level and elites-administered trade caravans appear to have transported a variety of goods in the region, the initial contexts for caravan trade remain largely unexplained. These initial contexts are particularly important because this mode of organization contributed to the regional context and institutional base in which early leaders in the Titicaca Basin had to operate in order to begin the process of expanding their influence in access to labor, resources, and ideology of their communities.

If enduring regional interaction had persisted since the early days of the pastoral economy during the preceramic, this may be connected to the factors that lead to an increasing consolidation of power in the Titicaca Basin during the Formative Period. These questions are central to understanding the foundations of regional integration that emerged in the Lake Titicaca Basin during the Middle and Late Formative Periods. The above discussion has sought to elaborate upon a possible context for early household-level caravan organization that is principally based on the "circuit mobility" model of Nunez and Dillehay (Dillehay and Nuñez 1988;1995 [1979]) but without following the adaptationalist approach, and with more specific empirical contexts for early caravans. In the ensuing discussion of obsidian procurement and circulation in the south-central Andes existing evidence from obsidian circulation in the region is presented that provides the context for examining obsidian production in more detail.

3.3. Regional patterns and major sources of obsidian

Sources of tool-quality obsidian occur in discrete locations along the Western Cordillera in the south-central Andes. The spatial pattern of obsidian sources in southern Peru, spaced at an interval of approximately 100-300 km along an arc trending north-west to south-east, is evident from maps of these source distributions. It is volcanic processes at a continental scale that have resulted in emplacement of obsidian at 3000 and 5000 meters above sea level. As will be discussed in more detail in Chapter 4, obsidian sources appear to be regularly spaced paralleling the spine of the Western Cordillera, however geologists believe that "the complex tectonic and lithospheric variables at continental subduction zones preclude the development of any regular pattern of volcano spacing" (Clapperton 1993: 77-80). Nevertheless, volcanoes do occur at major crustal fractures and at fracture intersections.


Figure 3-5. Known south-central Andean obsidian sources used in prehistory with least cost paths (Tobler's function on SRTM DEM data) from the three major sources to sites in the region.

In the prehispanic central and south-central Andes obsidian from three sources dominated the archaeological use of this material, followed by obsidian from a number of minor, local sources that were only used for tool production in the area of geological origin (Figure 3-5). Finally, there were countless sources of geological obsidian that were used negligibly, or not at all, by humans apparently due to the undesirable attributes of the glass. Outside of the study region, additional major Andean obsidian sources have been studied to the south in northern Argentina and surrounding terrain (Yacobaccio, et al. 2004;Yacobaccio, et al. 2002) and to the north in Ecuador (Burger 2003;Burger, et al. 1994).

The three most widely-circulated obsidian sources in Peru and north-west Bolivia are the Quispisisa, Alca, and Chivay types, and these three types will be investigated here in more detail. Burger et al. (2000: 348) argue that it is the large nodule size and the homogeneity of the glass that is the primary determinant of how widely obsidian was circulated in prehistory and by these criteria these three sources are exceptional in the Andes.

In the course of this research a relational GIS database has been assembled using published data on obsidian sourcing in the south-central Andes (Burger, et al. 2000;Craig 2005: 908-916;Frye, et al. 1998;Giesso 2000) and using these data a few summaries have been generated for the purposes of this study. The trends in these data show broad patterns in the use of obsidian by chemical type, but given the early stages of this kind of research, these values also reflect a sampling bias based in the history of research and the collections available for sampling. With further development of sourcing technologies, including the more widespread use of portable XRF units, obsidian from a wider variety of sites will be characterized in coming years.




Ave. Distance
(Km) *

Ave. Time
(Hrs) **

Max. Distance
(Km) *

Max. Time

(Hrs )**






















Table 3-2. Three major Peruvian obsidian sources showing average and maximum distances and times. * based on straight-line distance measure, ** Least-cost path walking time based on Tobler's(1993) hiking function.


Figure 3-6. Cumulative frequency graph showing three major Peruvian obsidian sources by consumption site altitude.

As will be discussed below, the dominant pattern that has emerged for both the Chivay and the Alca obsidian types is one of highland interaction. Throughout the prehispanic period, artifacts of Chivay obsidian have never been found below an altitude of 1280 masl, and that was in the context of a Tiwanaku colony, at the site of Omo in Moquegua (Burger, et al. 2000: 338;Goldstein 1989). Alca type obsidian artifacts are similarly found mostly at high altitude; the coastal Quebrada Jaguay paleoindian finds are anomalous for Alca distributions. The next lowest altitude context for Alca is also at Middle Horizon Omo. Quispisisa, on the other hand, is commonly found at coastal sites from the Nasca area and northward, beginning with a strong representation at the cotton preceramic site of San Nicolas (Burger and Asaro 1978;Vaughn 2005).

One possible explanation for the lack of Chivay and Alca obsidian in low elevation contexts is that camelid caravans were less common in the low elevation coastal areas of southern Peru and northern Chile. While there is strong evidence of camelid corrals in northern coastal Peru (Shimada 1982), the examples of pastoral sites in the coastal areas of the south-central Andes are rare. Stanish (1992: 57) states that there no sites with corrals below 1,500 masl in the Moquegua drainage, and he presents other evidence for the scarcity of pastoral sites in southern coastal Peru. However, Wheeler et al. (1995) examine the remains of 26 camelids, both llamas and alpacas, at the Chiribaya site of El Yaral at 1000 masl in Moquegua. Moving northward in the Pacific coastal drainages, both Christina Conlee (2000) and Kevin Vaughn (2005) report evidence of domesticated camelids at low altitude in the Nasca area. Vaughn (2005: 92) describes a corral-like structure at the site of Marcaya at 1000 masl. As mentioned, the Nasca zone is rich in Quispisisa obsidian. A variety of factors may have contributed to the pattern of Quispisisa obsidian consumption on the coast and Alca and Chivay consumption in the highlands including the greater breadth of the altiplano in the Titicaca Area, the position of geological source relative to the coast, and the severe aridity in the Atacama area of southern Peru and northern Chile.

3.3.1. Obsidian and larger geographical associations

The specific cultural and temporal associations of these distributions are explored in a series of publications by Richard Burger, particularly in Burger et al. (2000). There appear to have been two overarching geographical associations that cut across the various distributions of obsidian in space and time.

Availability, demand, and mobility

Chivay was largely conveyed into the Lake Titicaca area by Titicaca Basin communities and polities that were organized around the lake and the adjacent broad altiplano. For sites in the Lake Titicaca Basin, the database shows that 453 obsidian artifacts have been sampled from all time periods, and 408 (90%) of the samples were of the Chivay obsidian type. Another interesting aspect to this pattern is that the proportion of Chivay obsidian holds true for the north Titicaca Basin as well as the south when the two areas are considered separately. Using Bandy's (2005: 92) division of the Titicaca Basin where the site of Camata and northward are considered the North Basin, and with the South Basin extending as far south as Khonkho Wankane and south-west to Qillqatani, all obsidian samples in the region were aggregated from the entire prehispanic period. The LIP site of Cutimbo, though slightly south of Camata, was joined with the North Basin sites while Incatunahuiri and the Ilave valley sites were counted with the South Basin.



Chivay Type

Alca Type

Other Types

NorthTiticaca Basin









SouthTiticaca Basin


















Table 3-3. Obsidian in the north and south Titicaca Basin by counts and percents.

These patterns reveal that Chivay obsidian accounts for approximately 90% of obsidian from the Titicaca Basin, with a greater presence of Alca obsidian in the North Basin. It seems that the presence of obsidian from alternative sources, whether it was the Tumuku source (still unlocated) or other obsidian sources further to the south, counter balances the access to Alca type in the North Basin. This pattern, with heavy use of Chivay obsidian in the South Basin, reflects overall integration in the Titicaca Basin through prehistory as well as the lack of high quality obsidian sources with large nodules south of the Chivay source, until one arrives at the Zapaleri source close to the frontier of Bolivia, Argentina and Chile, located 700 km to the south of Lake Titicaca (Yacobaccio, et al. 2004: Fig. 1). It has been noted that, in cultural terms, the Chivay distributions are especially linked to areas integrated by the Tiwanaku state, the Pukara polity, and with their predecessors: the Middle Formative communities linked by Yaya-Mama stylistic features (Burger, et al. 2000). However, what is particularly compelling about the evidence from obsidian is that these social and economic links in obsidian exchange appear to predate the cultural-stylistic links and become defined in the Terminal Archaic roughly 1000 years earlier than regional evidence of the Yaya-Mama stylistic attributes that first appear at Chiripa in the southern Lake Titicaca Basin.

The distribution of Alca material is confined by the availability of Quispisisa on one side and Chivay on the other, but nevertheless Alca obsidian was transported the furthest with material during the Middle Horizon being conveyed as far as the Wari sites of Huamachuco and Marca Huamachuco near Cajamarca (Burger, et al. 2000: 336). Alca material was also transported the furthest distance at an early date, as it was conveyed 769 km to Chavín de Huantar during the Early Horizon where it was found in both elite and commoner contexts (Burger, et al. 2000: 313-314). Quispisisa predominates in Wari assemblages but it also circulated widely before the Middle Horizon, such as the 589 km to Chavín de Huantar during the Early Horizon.

Terrain adjacent to source

The altitudes of obsidian consumption sites (XFigure 3-6X) also reflect the terrain adjacent to each of the sources. The Quispisisa source is in an area with deep river valleys and a number of the sites sampled are from lower elevation or coastal contexts. The Alca source is similarly close to deep river valleys, but it also borders the altiplano and the much of the Alca samples came from the highlands of the Department of Cusco. The Chivay source is on the periphery of the broad altiplano and with no major altitude loss it was possible to transport material to a variety of sites that include Pikicallepata to the northeast and Qillqatani to the southeast. Further detail on these obsidian distributions are described below, and are published elsewhere (Burger, et al. 2000; cited in Craig 2005: 908-916;Frye, et al. 1998;Giesso 2000).

3.4. Chivay Obsidian Consumption Contexts

Archaeological research conducted at four sites in the south-central Andean highlands, the consumption zone for Chivay obsidian, provide evidence for the changing use of Chivay obsidian over long time periods. This section will begin by focusing on the four sites because they important here as the sites have provided stratified evidence of obsidian consumption during the Archaic and Formative in well-dated archaeological contexts. This review of the sites will be relevant in subsequent discussions of change in obsidian use over time in the region, and for interpreting activities at the Chivay source.

3.4.1. Asana

The site of Asana is at 3435 masl in the Osmore drainage of Moquegua, and it lies 185 km or 50 hours by the hiking model to the south-east of the Chivay source. Excavations at Asana were directed by Aldenderfer(1998: 76-80)between 1986 and 1991 and in the lower levels of the deeply-stratified site obsidian was found in low concentrations.

Titicaca Chronology (BCE)

14C yr bp

Lab Code (Beta-)


14CSample Provenience

Obsidian Provenience

Obsidian Type



Early Form.





II (sup), TU1


Aconcahua (visual ID)

n=3, 0.4%

Late Archaic





XIVb, I25b


n=1, 2%

Middle Archaic





XVIIa, G24b

KF4a: J29D-8; XVIIA





XVIIa, G24b

KF4b: J29D-8; XVIIA


Early Archaic





PXIX, P38b




PXX, V51a

KF5: F28D-5; PXXi





PXXI, R43c






PXXI, R43c

KF2: W41A-4; PXXII










n=11, 0.36%

KF3: T42B-7;F28D-5


n=1, 0.08%

KF1: U41D-6; PXXV






Table 3-4. Asana obsidian samples, collections, and associated14C samples by level (Aldenderfer 1998: 131, 157, 163, 209, 268;Frye, et al. 1998).

These data from Asana show that low frequencies of Chivay obsidian occurred regularly through Early and Middle Archaic levels at the site despite the presence of high quality cherts nearby, and low quality obsidian from the Aconcahua source located about 50 km away. During the Late Archaic, obsidian disappears from the site and in a level that is transitional between the Terminal Archaic and Early Formative, obsidian returns in the form of lower-quality Aconcahua material. These data will be discussed in more detail below as obsidian use is explored by time period.

3.4.2. Qillqatani rock shelter

The rock shelter of Qillqatani is at 4400 masl along the Río Chila to the southwest of Mazo Cruz, and it is 221 km or 50.3 hours walking time, by the hiking model, across the puna from the Chivay obsidian source. Qillqatani lies on the southwestern perimeter of the Lake Titicaca Basin in the headwaters of the ríoHuenque drainage in the Department of Puno, and it is not far from the headwaters of the Osmore drainage on the western slope in the Department of Moquegua. The shelter has a strong pastoral component, but excavations revealed occupations dating back to the Middle Archaic. The shelter is also known for elaborate rock art panels. Qillqatani was excavated in 1990 by Aldenderfer in two adjacent blocks, labeled East Block and West Block (1999;2005;in prep.). The East Block measured 20 m2in horizontal extent, while the West Block measured 60 m2.

(a) Titicaca chronology,

calibrated dates

(b) Qillqatani periods,

calibrated dates

(c) Uncalibrated14C dates in r.c.y.b.p.

(d) Qillqatani provenience
of obsidian

(e) Obsidian sample

(f) Block

LH AD1476 - 1532

LH 1476 - 1532


Level D11d-507/WIII:



LIP AD1100 - 1476

LIP 1100 - 1476



H12C-3; WV; LIP



Tiwanaku Horizon

AD400 - 1100


AD 500 - 1100


Level D9a/F7/WVIII:


E9A-5; WVII; F-3






Late Formative

400 BCE - AD500



D6C-8; WX; F-5; 1.45



Late Formative

500 BCE - AD 400

D9A-8; WX; F-5; 1.45



G13A-10; WX



G13A-10; WX



D6C-9; WXI; F-6; 1.6



H12D-10; WXI; F-12; .75



H12D-10; WXI; F-12; .75



I12D-8; WXI; F-13; 2.00



Formative C

900 - 400 BCE


East Block dates

Level D25d/F10/EVI:


Level F23a/F4/EXI:


West Block date

Level G7c-513/WXII:


E25B-10; EIX; F-7



Middle Formative

1300 - 500 BCE

D25D-6; EIX; F-7



D26B-10; EXI; F-12



F23B-11; EXI; F-3



F24B-9; EXI; F-13



D6D-11; WXIII; F-8; .75



F11A-3; WXIII; 1.65



D26D-13; EXIV; F-10



D26C-13; EXIV



D26C-13; EXIV



Formative B

1500 - 900 BCE



D8C-12; WXVIII; F-11



D8C-12; WXVIII; F-11



G7D-11; WXVIII; 5.5



Early Formative

2000 - 1300 BCE

Level D6/F7/WXVI:

Level D6a-522/WXIX:2940±70bp(B43925)

E9D-12; WXVIII; F-9



Formative A

2000 - 1500 BCE


E6C-14; WXX; F-7; 1.4




E9A-15; WXX; 2.4



H13A-15; WXX; 4.6



Terminal Archaic

3300 - 2000 BCE

Terminal Archaic

3300 - 2000 BCE


Level F9a-532/WXXIV:3660±60bp(B43926)

F28B-19; WXXIV; HF-1



F28B-19; WXXIV; HF-1



F28B-19; WXXIV; HF-1



G9E-19; WXXIV; 98.70



G9E-19; WXXIV; 98.20



G10B-20; WXXIV; HF-2



Late Archaic

5000 - 3300 BCE

Late Archaic

5000-3300 B.C.E


Level E9d-541/WXXX:

E9D-18; WXXX



Middle Archaic

7000 - 5000 BCE

Middle Archaic

7000-5000 B.C.E

Below level WXXX:7100±130bp(B18926)



Source unknown


Table 3-5. Qillqatani excavation levels, radiocarbon dates, and obsidian samples. (a) Titicaca chronology, (b-c) levels and14C dates from Qillqatani, all dates are on charcoal and were analyzed by Beta Analytic, from Aldenderfer (1999). (d-f) Obsidian from Qillqatani chemically provenienced at MURR from data presented by Frye, Aldenderfer, and Glascock (1998), with "Tumuku" type replacing reference to "possible Alca-2", and Aconcahua type abbreviated to "Aconc."


Figure 3-7. Qillqatani data showing percentage of bifacially flaked tools and percentages of debris made from obsidian per assemblage by count (Aldenderfer 1999;Aldenderfer in prep.).











Qillqatani Period


Ave Wt.


Ave Wt.



Ave Wt.


Ave Wt.


Late Horizon



































Late Formative












Formative C

900-400 BCE











Formative B

1500-900 BCE










Formative A

2000-1500 BCE











Term. Archaic

3300-2000 BCE











Late Archaic

5000-3300 BCE











Middle Archaic

7000-5000 BCE











































Table 3-6. Qillqatani periods by tools (bifacially flaked) and debitage (all other lithics), in obsidian and non-obsidian categories.


Excavations in the two blocks at Qillqatani revealed a switch from hunting and gathering to a predominantly pastoral economy between the occupation in level 25, and that of level 24 that was stratigraphically superior (Aldenderfer 2005: 20;). The occupation in Level 24 appears to have been the first residential occupation, and it is considerably thicker than were previous levels that appeared to have been logistical in nature. The ovoid structures of Level 24 are slightly larger and are free standing, while in the preceding levels the structures are against the rear wall of the rock shelter. Level 24 is the first level with ceramics, but it is being considered Terminal Archaic in this analysis due to economic evidence and the exceptionally early ceramics at Qillqatani.

Qillqatani Period

Weight in Grams

Total Count

0 - 1 g

1 - 2 g

2 - 3 g

3 - 4 g

4 - 5 g

Late Horizon
















Late Formative






Formative C







Formative B







Formative A







Terminal Archaic







Late Archaic






Middle Archaic







Total Count







Table 3-7. Counts of obsidian debitage at Qillqatani by weight (g).

The earliest obsidian at Qillqatani is a Middle Archaic sample from an unknown source. In the Late Archaic, Chivay obsidian begins to appear at the rock shelter. Subsequently, during the Terminal Archaic obsidian from Chivay occurs in relatively large numbers and it persists until the end of the Early Formative (the Qillqatani Formative Cperiod) where chemical testing of ten obsidian samples reveals there is greater heterogeneity in obsidian procurement. In this period, Chivay obsidian makes up 60% of the obsidian tested by the Missouri University Research Reactor (MURR), the Tumuku type makes up 20%, and the Aconcahua type and Alca type make up the remaining 10% (one sample of each).

Three obsidian artifacts from two proveniences at Qillqatani, West level XI and East level XIV, were found to belong to the Tumuku chemical group, a group that derives from an as-yet-unlocated source that probably lies in the southern department of Puno. This chemical group has not yet been well characterized, as no source samples are available and the samples analyzed by NAA in the 1970s at Lawrence Berkeley Labs (Browman 1998: 309-311; Burger and Asaro 1977) are not easily comparable with more recent NAA results from MURR. As a consequence, in initial studies of these three Qillqatani artifacts it was proposed that the artifacts belong to a subgroup from the Alca obsidian source in northern Arequipa, a subgroup that was referred to as "Alca-2" (Frye, et al. 1998) and "Alca-Z" (Jennings and Glascock 2002: 111). More recently, Michael Glascock at MURR has stated that these three Qillqatani samples probably derive from the Tumuku source (M. Glascock, March 2006, pers. comm.).

Flake sizes at Qillqatani indicate that relatively large flakes were being discarded (Table 3-7), particularly during the Terminal Archaic and Early to Middle Formative. A central question concerning obsidian circulation is to what degree were flakes used for shearing and butchering tasks? Were flakes sufficiently large to be hand-held for these activities? Length measurements from these flakes from Qillqatani are not available, but an estimate of the size of those flakes based on their weight is provided by comparing similar obsidian flakes from the Chivay source dataset. An obsidian flake weighing 4g from the Chivay dataset had the following dimensions: 45 x 13 x 7mm. A smaller obsidian flake weighing 3g measured 31 x 14 x 6mm. These flakes appear to have been sufficiently large for shearing and butchering tasks. Further technical studies, as well as use wear and residue analysis, may shed more light on the use of large obsidian flakes in the regional consumption zone.

Small quantities of material and regional models of exchange

The actual volume or weight of obsidian at Qillqatani is relatively small. Obsidian from the Chivay source appears in small quantities in nearly every stratum in the approximately 8000 year sequence. The most informative data on long distance exchange come not from the total mass of obsidian, but from the variability in the relative percentage of obsidian flaked and obsidian tools in each lithic assemblage through the sequence. Nevertheless, the weight of obsidian tools at Qillqatani totals 43 g, and the sum of the weight of obsidian flakes is 382 g out of a total of 11,085 g of lithic material. Thus, the entire quantity of obsidian excavated at Qillqatani is only about 1.5% of the cargo that could be carried by a single llama.

These figures are meant to highlight a central issue with obsidian studies overall: the actual quantities of obsidian encountered and evaluated from the consumption sites throughout the south-central Andes are relatively low. The significance of obsidian circulation over the larger region is not matter of weight or volume, but rather a question of consistency and changes in the proportions of particular sources utilized over time. These data from Qillqatani underscore this issue. A comparison of the persistence of non-local obsidian at Qillqatani as of level 24 (X

Table 3-6XandXTable 3-7X) with the non-local obsidian at Asana (XTable 3-4X) during earlier periods of the Archaic, suggests that on a regional scale mechanisms of exchange, or direct acquisition, were more intermittent during the earlier Asana phases. In contrast, Qillqatani has high quality cherts and low quality Aconcahua type obsidian available in the immediate proximity of the rock shelter. Thus, the question becomes: why transport Chivay obsidian 200 km when alternative materials are locally available? Further, why was the transmission of Chivay obsidian so consistent through time? In this perspective the low but consistent quantities of Chivay obsidian that were conveyed throughout the south-central Andean highlands during the prehispanic period can be seen as a gauge of highland interaction and horizontal complementarity, with wider implications for exchange.

3.4.3. Sumbay

The work of Máximo Neira Avendaño (1990;1968) is distinguished as being the earliest systematic Archaic Period excavations in the Arequipa highlands. In the mid-1960s Neira excavated at seven preceramic sites close to the Sumbay train station, including several rock shelters, known as Sumbay-1 through Sumbay-7. The sites were later renamed "Ccollpa-Sumbay" by Eloy Lináres Malaga in 1984 (1990;1992) as part of the CONCYTEC survey.

The Sumbay area lies 42 km to the south-east of the Chivay source across the pampa and it is 9.3 hours from the obsidian source area by the hiking model. The main cavern, SU-3, is a rock shelter that is concealed in narrow canyon that cuts through the puna. This rock shelter measuring 15m across, 11m deep, and 6m in height at the dripline, was badly looted in the 1930s, and Neira's excavations focused on the remaining intact portions of the cave. The cave is known for exceptional rock art panels featuring both petroglyphs, ochre pictographs, and a variety of animal species including elongated camelids, a hunting puma, humanoid figures, and surithe Andean ostrich.

Seven excavation pits were placed in the main shelter, SU-3. Pit 5 produced obsidian samples in every level along with a large number of tools made from pitchstone, a dull vitreous material formed from weathered obsidian that has absorbed water from the environment. Pitchstone has 4-10% water while obsidian has 0.1 - 0.5% water (M. Glascock, 2006 pers. comm.). The source of this pitchstone is probably somewhere near Sumbay and further study of this pitchstone may provide an interesting complement to the chemical characterization of obsidian.


Figure 3-8. Sumbay pitchstone projectile points.

Titicaca Chronology



14C yr bp

Lab # (Bonn)


Obsidian Type




Many very small flakes.

Late Archaic

5000-3300 BCE





One pitchstone knife, 5 broken pitchstone points, 1 quartzite scraper, 100 flakes.





Two obsidian samples (Chivay)

One pitchstone foliate point with concave base, 4 pitchstone scrapers, 6 broken pitchstone points, 1 broken obsidian point, 62 flakes.

Middle Archaic

7000-5000 BCE



One obsidian sample (Chivay)

Six pitchstone points: 2 incomplete stemmed with concave base, 1 almost complete, 3 foliate points, one with broken tip and broken base. Seven incomplete pitchstone points, 2 incomplete quartzite points. One pitchstone knife, 110 flakes, 1 worked bone, 5 bone concentration.

Table 3-8. Sumbay, SU-3 Pit 5. Strata with obsidian samples and associated artifacts(Neira Avendaño 1990: 32-33).

Twenty-five obsidian samples were analyzed from the Sumbay area. Twenty samples from the surface and terreplain of the rock shelter SU-2, two from the surface of the rock shelter SU-3, and 3 from excavations in SU-3, Pit 5 (Burger, et al. 1998: 209;Burger, et al. 2000: 278). All samples turned out to be of the Chivay type. Since stratum 3 was dated to the Middle - Late Archaic transition, the stratum 4 obsidian sample is stratigraphically below than that, so it is probably Middle Archaic in date.

3.4.4. The Ilave Valley and Jiskairumoko

Located in the Department of Puno in the western Lake Titicaca Basin, the Ilave river valley is an open, terraced river valley and it forms the highest volume river drainage to flow into Lake Titicaca from the west. Jiskairumoko (95-189) is the largest multicomponent Archaic/Formative site identified during survey and it lies 200 km from the Chivay obsidian source or 44 hours travel time across the expansive puna grasslands. Mark Aldenderfer conducted research in the Ilave valley from 1994 to 2002 including a pedestrian survey, a testing program, and excavation work (Aldenderfer 1997;Aldenderfer 1998). In 1997 and 1998 Cindy Klink conducted research in the Huenque drainage, a principal tributary that joins the Ilave from the south (Klink 2005;Klink 2006;Klink and Aldenderfer 1996). Nathan Craig excavated at the site of Jiskairumoko and tested at numerous other sites between 1999 and 2002 (Craig 1999;Craig 2005;Craig and Aldenderfer 2002;Craig and Aldenderfer In Press). Tripcevich conducted a viewshed analysis of sites in the Ilave drainage using data from the 1994 and 1995 Ilave valley survey (Tripcevich 2002).


In the Ilave valley, obsidian was scarce prior to 3300 BCE, but it appears in a number of Terminal Archaic and Formative contexts after that date. Diagnostic projectile points from the Ilave area reveal a dramatic change in material type with the Terminal Archaic.


Figure 3-9. Comparison of projectile point counts in the Ilave Valley and the Upper Colca.

Using only diagnostic projectile points from the Klink and Aldenderfer (2005) point typology, these data reveal a shift to greater use of obsidian with series 5 projectile points. Following the point typology, all series 1 through 4 points are diagnostic to the Terminal Archaic or earlier, except for types 4C and 4E which have been excluded from this analysis. The lower part ofXFigure 3-9Xreveals data classified into the same three groups, but with data from the Chivay obsidian source area resulting from the 2003 Upper Colca project. Comparing these two datasets reveals that with the advent of Series 5 points in the Terminal Archaic after 3300 BCE there is a sudden upswing in obsidian use in the Ilave consumption zone that corresponds perfectly with a dramatic shift in projectile point production at the Chivay source area itself. There appear to have been pan-regional changes occurring with the widespread adoption of series 5 style projectile points, the use of obsidian, and these changes are possibly related to the greater use of the bow and arrow.

Jiskairumoko (95-189)

The site of Jiskairumoko was excavated using a broad, horizontal décapage technique between the years 1999 and 2002, and Craig's (2005) dissertation forms the site archive. Radiocarbon dates fromexcavation work at Jiskairumoko range from the Late-Terminal Archaic transition with a date of 4562±73bp(AA58476; 3520-3020 BCE)through to the end of the Early Formative, with a date of3240±70bp (Beta-97321; 1690-1390 BCE). Note that Craig (2005) has the Terminal Archaic dating to 3000-1500 BCE, while in the Upper Colca Project and in this dissertation the period dates to 3300 - 2000 BCE All references to the "Terminal Archaic" have been adjusted to the latter time range in this document unless otherwise noted.

Craig and Aldenderfer chemically characterized 68 bifacially-flaked obsidian tools from excavated contexts at Jiskairumoko (96% of all obsidian tools from Jiskairumoko) by sending these items to M. Steven Shackley at the UC Berkeley Archaeological XRF Lab(Craig 2005: 513, 908-916). The samples were compared with four geological samples that provided to Shackley from nodules that I had collected from Chivay source in 2003. Three of the nodules provided to Shackley were from the Maymeja area and one was from east of Cerro Hornillo. The XRF results showed that 97% of the Jiskairumoko obsidian tools (n = 66) were from the Chivay source and 3% (n = 2) were from the Alca source. These values diverge slightly from obsidian type distributions in the Titicaca Basin more generally (XTable 3-3X), where 90.5% (n = 466) of all of the obsidian artifacts tested to date from the Titicaca Basin are Chivay type. These data indicate that the people at Jiskairumoko were more intensively using Chivay obsidian in the Terminal Archaic and Early Formative than were Titicaca Basin residents over all, especially in contrast to the variety in obsidian types that emerge from samples dating to later times.

Obsidian artifacts from excavations at Jiskairumoko reveal that obsidian was used in similar proportions for projectile points from excavated contexts as for points from surface contexts in the Ilave valley. Craig(2005: 685)observes that 18% (n = 54) of the excavated Jiskairumoko projectile points are made from obsidian. He also notes that 50% of these have edge modification in the form of serration or denticulation while only 15% of non-obsidian points have edge treatment, a statistically very significant pattern that he attributes to the symbolic importance of this material. Obsidian was found, along with other lithics and a carved camelid effigy, in Burial 1 at Jiskairumoko in the grave of an elderly adult female with cranial deformation, a primary deposit dating to the Terminal Archaic or 3019-2859 cal BCE(Craig 2005: 680-682). Graves excavated in the Ilave valley area also contain a variety of non-local grave goods including obsidian, lapis, and gold, but non-local grave goods only occur in contexts dating to the Terminal Archaic and later.

3.5. Andean Obsidian Distributions through Time

While obsidian spatial distributions are well demonstrated, it is more difficult to ascertain the temporalpatternsbecause many of the obsidian artifacts from the south-central Andes that have been sourced are from an uncertain temporal provenience. The Andean obsidian sourcing literature, organized and published primarily by Richard Burger, has provided a temporal context for obsidian samples wherever possible. Temporal control for obsidian artifacts fall into three categories

(1) excavated contexts with a temporal association.

(2) surface materials spatially associated with temporally diagnostic artifacts.

(3) direct chemical sourcing of temporally-sensitive diagnostic projectile points.

Good temporal control from excavated contexts is preferred, but a number of the samples are from the number 2group, above, because the samples have weak temporal control and are from sites that are assumed to be single component based solely on diagnostic artifacts. Finally, a fourth method, obsidian hydration dating, has been used in the Andes but with mixed results. The high diurnal and seasonal temperature variation in the Andean highlands suggests that hydration dating can only be reliably used as a relative dating method in conjunction with14C dates (Ridings 1996).

Chronologies based on stages on the one hand, and periods or horizons on the other, are used in Andean archaeology. Some pan-Andean research, such as studies of regional obsidian exchange, have used the horizons absolute chronology developed by John H. Rowe (1967) in collaboration with Dorothy Menzel and based on the Ica ceramic "master sequence". This approach allows for a maximum of compatibility between regional datasets and it has been used widely, and even extended into the Titicaca Basin (Burger, et al. 2000). However, archaeological research conducted in the Lake Titicaca Basin has largely followed a chronology based on evolutionary stages that differs from the horizon chronology primarily during the Formative Period and the Tiwanaku Period (Hastorf 1999;Kolata 2003;Stanish 2003: 85-90). As the bulk of Chivay type obsidian artifacts are found in the Lake Titicaca Basin, this dissertation will review Andean obsidian distributions using a Titicaca Basin chronology, and it will follow the Formative Period temporal divisions used by Stanish (2003). These regional data will be presented in terms of three major temporal periods in order to be consistent with the framework adopted with the new data reported for this dissertation.


Titicaca Chronology

Years cal. AD/BCE

Late Prehispanic

Late Horizon

A.D. 1476 - 1532

Late Intermediate

A.D. 1100 - 1476


A.D. 400 - 1100

Early Agropastoralists

Late Formative

500 BCE - AD400

Middle Formative

1300 - 500 BCE

Early Formative

2000 - 1300 BCE

Terminal Archaic

3300 - 2000 BCE

Archaic Foragers

Late Archaic

5000 - 3300 BCE

Middle Archaic

7000 - 5000 BCE

Early Archaic

9000 - 7000 BCE

Table 3-9. Temporal organization of data.

These three major temporal divisions serve to differentiate the major forces that appear to have structured obsidian procurement and distribution. Rather than focusing on the "preceramic" and "ceramic" dichotomy, which is only peripherally related to obsidian use, a decision was made to present data in terms of these three major divisions. Furthermore, socio-political changes that accompanied obsidian circulation in the Terminal Archaic are better considered in conjunction with the later developments in the Formative than together with the earlier periods of the Archaic.

The use of the term "Archaic Foragers" is not meant to suggest that no pastoralism occurred prior to 3300 BCE Many scholars place the initial domestication sometime between the Middle and the Late Archaic in the chronological terms used here (Browman 1989;Kadwell, et al. 2001;Wheeler 1983;Wing 1986). However, initial domestication and adoption of a dedicated pastoralist economy are distinct events, and the evidence from a number of sites in the south-central Andean highlands suggests that the process of shifting to a food producing economy in the Andean highlands featuring camelid pastoralism and agriculture based largely on seed-plant and tuber cultivation was complete by circa 3300 BCE in the Upper Colca region. Likewise, these divisions between "forager" and "pastoralist" are not meant to suggest that foraging activities ceased in the Terminal Archaic, merely that foraging diminished in importance in terms of caloric intake and food security.

The Early, Middle, and Late Archaic Periods have the additional benefit of being differentiated by relatively consistent changes in projectile point morphology through prehistory. A projectile point typology has recently been developed by Klink and Aldenderfer (2005) that synthesizes data from previously developed point typologies with new evidence from stratified Archaic and Formative excavations in the region. This typology has proven to be extremely useful for investigating surface distributions of projectile points, and it will was used throughout this dissertation both for assessing the age of obsidian points on a regional scale, and for assigning temporal categories to surface sites identified in the course of our 2003 survey work.


Figure 3-10. Projectile Point Typology with Titicaca Basin chronology (Klink and Aldenderfer 2005)


The temporal groups shown in Figure 3-10 visibly depict the changes in temporal control provided by projectile point evidence between the "Archaic Foragers" and the "Early Agropastoralists". With the advent of series 5 projectile point styles in the Early Agropastoralists time, the forms of projectile points became a far less effective means of differentiating time periods because projectiles do not change with as much regularity. Series 5 projectile points, except for type 5a, are diagnostic only to the time after the domestication of camelids and the dominance of food producing, pastoral economy. The following review will examine only a few of the most significant aspects of these distributions by time period, with a focus on insights from obsidian exchange on the preceramic and Early Agropastoralist stages of Andean prehistory.

3.5.1.Archaic Foragers in the South-Central Andes

To date, relatively few Archaic archaeological sites have been excavated in the area of the south-central Andean highlands, and therefore the regional knowledge of this important time period is limited. A brief overview of each period will be provided here, with an emphasis on archaeological sites containing Chivay type obsidian, in order to contextualize the regional relationships and interactions that were occurring during these preceramic periods. This review will focus on the western slope of the Andes and the Lake Titicaca Basin, the region where Chivay type obsidian abounds. Much of the Archaic Foragers obsidian in the literature that has been sourced was derived from undated, surface contexts from sites that are assumed to date to the Archaic based on projectile point styles or other inference by archaeologists. Klink and Aldenderfer's (2005) projectile point typology permits the classification of site occupations by time period with greater certainty during the Archaic when projectile point styles change with regularity. In some cases, the projectile point typology can be used to assign a date range to temporally diagnostic obsidian projectile points that have been directly analyzed for chemical provenience.

In Burger et al.'s (2000: 275-288) review, procurement and exchange during the preceramic is evaluated through chemical provenancing data from 87 obsidian samples, 70 of which are from the Chivay source. Unfortunately, of the thirteen sites where the samples were collected in the south-central highlands, only three sites (Chamaqta, Asana, and Sumbay) contain excavated contexts placing the samples in preceramic levels. The other samples are assigned to the preceramic group through inference from projectile points, or because they are from a-ceramic sites. The difficulty in this arises from the fact that obsidian was used with much greater frequency in the beginning of the pastoralist periods, as was mentioned above in the case of the Ilave valley (Section 3.4.4 ).

Obsidian materials from both the Chivay and Alca sources were transported relatively long distances from the Early Archaic onward, indicating that material from these obsidian sources has properties that were prized at a relatively early date. Sites such as Asana and Qillqatani, both over 200 linear km from the Chivay source, contain non-local obsidian through much of the preceramic sequence despite the immediate availability of high quality cherts, as well as the lower-quality Aconcahua obsidian, in the local area. This evidence underscores the nature of obsidian provisioning in the Archaic Period. As compared with later pastoralist use of obsidian, Archaic foragers were relatively selective in their use of raw materials. This may reflect the priority placed on dependable hunting technology as compared with the shearing and butchering needs of pastoralists, or the different social and symbolic priorities attached to obsidian by hunters versus herders.


Figure 3-11. Chivay type obsidian distributions during the "Archaic Foragers" time
(circa 10,000 - 3,300 BCE).

Early Archaic (circa 10,000 - 7,000 BCE)

Evidence from the Early Archaic in the south-central Andes is slim, but it is accumulating gradually. The available data suggest that during this time period human groups went from initial exploration of the highlands to more stable, year around occupation of the sierra and puna. The best evidence from the highlands comes from a handful of stratified sites in the region with Early Archaic components, as well as distributional evidence from archaeological survey work(Aldenderfer 1997;1998;1999;2003;Craig 2005: 452-484;Klink 2005;Nuñez, et al. 2002;Santoro and Núñez 1987;).

A regional study of projectile points from excavated contexts by Klink and Aldenderfer(2005: 31, 53)noted that projectile point styles diagnostic to the Early Archaic (Series 1 and Types 4a, 4b) and the Early-Middle Archaic transition (Type 2a) have a broad geographical distribution. These types are found throughout a region that includes the littoral, western sierra, and puna areas of extreme northern Chile and southern Peru, and as far north as Pachamachay cave in the Junin puna(Rick 1980). Given the low population densities of that time period, this wide regional stylistic distribution suggests that a high degree of mobility was being practiced in the Early Archaic and first part of the Middle Archaic. Further to the north, in contrast, raw material types in the Junin puna of central Peru suggest to Rick(1980)that very low forager mobility was occurring throughout the early preceramic period. Early Archaic obsidian distribution data largely confirm with the model of high mobility and show that humans found good sources of lithic raw material in the sierra early, and then transported the material or exchanged it widely from a relatively early date.

Current paleoclimate evidence indicates that glaciers advanced during the late glacial between 11,280 and 10,99014C yr bp, and then, despite cool temperatures, the glaciers receded rapidly, perhaps as a result of reduced precipitation(Rodbell and Seltzer 2000). Glaciers were in approximately modern positions in southern Peru by 10,90014C yr bp (circa 11,000 cal BCE). The Early Archaic Period, with reference to diagnostic projectile point styles(Klink and Aldenderfer 2005), begins at circa 9000 cal BCE During the Early Holocene and the first part of the Early Archaic conditions were wetter and cooler than modern conditions. Subsequently, during the latter part of the Early Archaic, the climate began an episode of long-term aridity that lasted through the Late Archaic(Abbott, et al. 1997;Argollo and Mourguiart 2000: 43;Baied and Wheeler 1993;Paduano, et al. 2003: 272). Opportunities in the highlands for human foraging groups were created by a number of new ecological niches for plants and animals that opened up during the Early Holocene. These niches represented a resource pull for mobile foragers that countered the increased difficulty of subsistence in the hypoxic, high altitude environment faced by the non-adapted early settlers in the highlands(Aldenderfer 1999). A review of the evidence for forest cover on the altiplano during the Archaic, with deforestation occurring as a result of pastoral intensification, is provided by Gade(1999: 42-74).

During the Late Pleistocene lacustrine period, paleo-Lake Titicaca (Lake Tauca) was much larger but only a few meters higher than is modern Lake Titicaca (Clapperton 1993: 498-501). Radiocarbon dates on shells show that Lake Tauca was probably still present as late as 10,08014C yr bp (or 9900 cal BCE). Sediment cores from Lake Titicaca indicate that there was an increase in sub-puna vegetation and fire from vegetation prior to 9000 BP (Paduano, et al. 2003: 272). Evidence from paleoclimate records and fluvial geomorphology point to a time of increased aridity and salinity in Lake Titicaca, with short, episodic moist spells beginning around 8000 cal BCE and continuing through the Middle Archaic and Late Archaic Periods (Rigsby, et al. 2003;Wirrmann, et al. 1992). The climatic data for this time suggest that with deglaciation in the Early Holocene a resource niche opened up that exerted a pull on plant and animal species towards the high altitude regions, and that early human groups responded to these opportunities by colonizing the high Andes.

Chivay obsidian in the Early Archaic

Evidence of human use of the Chivay Source beginning in the Early Archaic Period comes from survey work adjacent to the source and from excavations at the site of Asana, 200 km away in Moquegua. As will be discussed in Chapter 7, during the course of survey work in the area of the Chivay source in 2003 the Upper Colca team collected several dozen Early Archaic type projectile points, the majority of them made from obsidian. Evidence for the regional consumption of Chivay obsidian begins with the site ofXAsanaXthat were described earlier in SectionX3.4.1XandXTable 3-4X. Aldenderfer(1998: 157, 163;2000: 383-384)encountered small quantities of obsidian in two levels belonging to the Asana II/Khituña Phase, placing them in the Early Archaic Period. In level PXXIX just one obsidian flake from Chivay was identified in an assemblage consisting of 1,152 lithic artifacts weighing 746 g. This level could not be dated directly, but it is assessed at 9400 uncal bp as it lies stratigraphically above a14C sample from level XXXIII dating to 9820±15014C yr bp (Beta-40063; 10,000-8700 BCE).

Obsidian from Chivay appeared in greater quantity in level PXXIV where eleven flakes of Chivay obsidian made up 0.36% of the lithic assemblage by count. These flakes were all small tertariary flakes, chunks, and shatter, suggesting to Aldenderfer(1998: 163)that a bifacial core of Chivay material was reduced on site. A14C sample from level PXXIV dated to 8720±120 bp (Beta-35599; 8250-7550 BCE). The Khituña phase has been interpreted by Aldenderfer(1998: 172-173)as representing a residential base in the high sierra and the beginnings of permanent settlement above 2500m in elevation. This interpretation is based on the presence of high sierra and puna lithic raw materials, but no coastal materials.

In surface contexts in the Ilave valley, surveys directed by Aldenderfer(1997)located three obsidian projectile points in forms that are possibly diagnostic to the Early and Early-Middle Archaic transitional period (types 1A, 3A, 3B). These three points were analyzed using a portable XRF unit in 2005 and all three were found to be of the Chivay type(Craig and Aldenderfer In Press).

Alca obsidian at Quebrada Jaguay

The earliest obsidian identified in the south-central Andes comes from one of the oldest confirmed sites in South America, the Paleo-Indian site of Quebrada Jaguay 280 near the coast north of Camaná, Arequipa. Sandweiss et al.(1998)report that Alca type obsidian was the dominant lithic material at the site in the Terminal Pleistocene and Early Holocene phases that were identified. Twenty-six Alca type obsidian flakes came from the older occupation that occurred during the Terminal Pleistocene and the context is dated by twelve14C dates on charcoal falling between 11,105±260 bp (BGS-1942; 11,700-10,400 BCE) and 9850±170 bp (BGS-1956; 10,100-8700 BCE). A later occupation contained one flake of Alca obsidian and it belonged to the Early Holocene II phase dated by four14C samples falling between 8053±115 (BGS-1944; 7350-6650 BCE) and 7500±130 (BGS-1700; 6600-6050 BCE). Three other obsidian flakes from the Alca source were collected but they could not be assigned to a temporal context.

Sandweiss et al. observe that the Chivay source is "less than 20 km further; the absence of Chivay obsidian at QJ-280 may indicate that this source was covered by glacial readvance during the Younger Dryas (circa 11,000 to 10,00014C yr B.P.)"(Sandweiss, et al. 1998: 1832). The Maymeja volcanic depression of the Chivay obsidian source indeed shows notable evidence of recent glaciation, but obsidian matching the "Chivay Type" is also found in smaller nodules on adjacent slopes outside the Maymeja zone as well as in the streambed below Maymeja having eroded through alluvial erosion (see Section4.5.1for maps and a discussion of the Chivay obsidian source geography). Procurement of obsidian from Chivay did not necessarily require entering the Maymeja area, however better source material can be obtained from the within the area.

Regional scale GIS analyses show Quebrada Jaguay to be 154 linear km from the Alca source, or 35.7 hours by the hiking function. In contrast, the trip to Quebrada Jaguay from the Chivay source is 170 linear km and 41.5 hours by the hiking function. Despite this relatively difference in distance, no Chivay obsidian was found at Quebrada Jaguay. It is worth noting that, despite this early find of Alca obsidian on coast, both the Alca and the Chivayobsidian types were predominantly found at high altitude throughout prehistory based on current evidence from obsidian sourcing studies. In the bigger picture, the coastal Quebrada Jaguay paleoindian finds are anomalous for Alca distributions. The next lowest altitude context for Alca to date is at the site of Omo in Moquegua, a Tiwanaku colony site at 1250 masl dating to the Middle Horizon.

Recent work by Kurt Rademaker and colleagues at the Alca source has shown that bedrock outcrops of obsidian are found as high as 4800 masl, and there are large pieces found at 3,800 masl that appear to have been transported downslope by glacial action and colluviation (K. Rademaker 2005, pers. comm.). Deposits of Alca obsidian at much lower elevations have been reported that are probably the result of pyroclastic flows(Burger, et al. 1998;Jennings and Glascock 2002). These deposits found close to the floor of the Cotahuasi valley at 2500 masl were probably available throughout the Younger Dryas and perhaps further sub-sourcing of Alca obsidian will better answer the question of which Alca deposits were being consumed during the Terminal Pleistocene by the occupants of Quebrada Jaguay.

Quispisisa Obsidian in the Early Archaic

Another early use of obsidian in the Andes is the evidence of Quispisisa obsidian in levels dated to ~9000 uncalBP or 8500-7750 cal BCE (Burger and Asaro 1977;Burger and Asaro 1978;Burger and Glascock 2000;MacNeish, et al. 1980) from Jaywamachay a.k.a. "Pepper cave" (Ac335) located at 3300 masl in Ayacucho in the central Peruvian highlands. As was recently explained by Burger and Glascock (2000;2002), for nearly three decades the Quispisisa obsidian source was mistakenly believed to have been located in the Department of Huancavelica, but in the late 1990s the true Quispisisa source was finally discovered in the Department of Ayacucho. Jaywamachay is the closest of the sites excavated by MacNeish's team to the newly located Quispisisa source. It is 81.3 linear km or 22.0 hours walking calculated using the hiking function, while the bulk of the sites excavated by MacNeish were in the Ayacucho valley approximately 120 km from the relocated Quispisisa source. The Puzolana (Ayacucho) obsidian source is located close to the Ayacucho valley and high-quality, knappable obsidian is available at this source, but the maximum size for these nodules is 3-4 cm, significantly limiting the potential artifact size for pieces made from this source (Burger and Glascock 2001). It is interesting to note that, in several cases in the south-central Andes, the Archaic Period foragers seem to have ignored small sized or low-quality obsidian sources that later pastoralists ended up exploiting. The Aconcahua source near Mazo Cruz, Puno, previously described, was not used at the adjacent Qillqatani rock shelter until the Middle Formative (Frye, et al. 1998), and it was not used at Asana until the Terminal Archaic (Aldenderfer 2000: 383-384).

Burger and Asaro(1977: 22;1978: 64-65 )chemically analyzed a projectile point of Quispisisa obsidian found at La Cumbre in the Moche Valley that was reportedly from a context dated to 8585±280 bp or 8500-6800 cal BCE(Ossa and Moseley 1971). The investigator, Paul Ossa, now doubts the Early Archaic context and instead believes the point may be Middle Horizon in date (Richard Burger, 11 March 2006, personal communication), but it is still a noteworthy case of long distance transport within the Wari Empire, and it perhaps involved the use of coastal maritime transport. Using the relocated Quispisisa source in the Department of Ayacucho the linear distance is 846.7 km and by the hiking function the distance is calculated as 199.8 hours, primarily confined to the coast.


Regional relationships indicated by both stylistic criteria and obsidian distributions point to high mobility during the Early Archaic Period. Recent evidence of projectile point similarities from northern Chile suggest that mobility was high along the Pacific littoral, and between the sierra and puna during the Terminal Pleistocene and Early Holocene around 25° S latitude(Grosjean, et al. 2005;Nuñez, et al. 2002). During the Early Holocene, foraging groups resided around high altitude paleo-lakes on the altiplano of northern Chile between 20° S and 25° S latitude that persisted longer than did the paleo-lakes at the latitude of Lake Titicaca. These lakes subsequently dried up during the Middle Archaic Period and occupations declined until the end of the Late Archaic Period but this distributions point to the mobility and early high altitude occupation of the adjacent puna of Atacama.

On the whole, obsidian proveniencing and analysis has shed light on human activities during the Paleoindian period and Early Archaic at a regional scale. Chemical analysis techniques, non-destructive XRF analysis in particular, are becoming more pervasive because the methods are being refined and the equipment is becoming more portable. Further field research will provide a greater understanding of this time period, but evidence will also likely to emerge from existing collections as obsidian projectile points diagnostic to the Early Archaic are systematically provenienced.

Middle Archaic (7000 - 5000 BCE)

The Middle Archaic Period, as with the preceding period, is still poorly understood in the region as very few highland sites containing stratified deposits have been studied from this period. In the Lake Titicaca Basin paleo-climatic evidence points to significant aridity, a lake level approximately 15m below the current stand, and high salinity in Lake Titicaca that is comparable to that of modern day Lake Poopó(Abbott, et al. 1997;Wirrmann and Mourguiart 1995). Human occupation during the Middle Archaic in the Lake Titicaca Basin was notably higher than in the preceding period, but research shows that settlement continued to occur in the upper reaches of the tributary rivers and not adjacent to the shores of the saline Lake Titicaca(Aldenderfer 1997;Klink 2005).

Obsidian distributions during the Middle Archaic

At the site of Asana on the western slope of the Andes, Aldenderfer(1998: 223)observed architectural features that suggest that a longer residential occupation of the site by entire coresidential groups was occurring in the latter part of the Middle Archaic Muruq'uta phase.One flake of obsidian from the Chivay source was found in the upper levels of theMuruq'uta phase (XTable 3-4X).This occupation dates to the Middle Archaic -Late Archaic transition as the flake was stratigraphically above a14Csample that dated to 6040 ± 90 bp (Beta-24634; 5210-4720 BCE).

At the rock shelter of Sumbay SU-3 obsidian was recovered from excavation levels that date to as far back as the transition between the Middle and Late Archaic (SectionX3.4.3X).


Relatively little is known about the Middle Archaic in the south-central Andean highlands. Elsewhere in the south-central Andes, archaeologists have noted an absence of settlement during the mid-Holocene timeframe corresponding to the Middle and Late Archaic Periods. In the dry and salt puna areas of northern Chile, and along the south coast of Peru, a significant decline or absence of mid-Holocene sites has led investigators to refer to this period as the silencio arqueológico(Nuñez, et al. 2002;Núñez and Santoro 1988;Sandweiss 2003;Sandweiss, et al. 1998). This designation apparently does not apply to the Titicaca Basin or to the sierra areas of the Osmore drainage where no Middle Archaic occupation hiatus has been observed.

With gradual population increases and adaptation to the puna, social networks extending across the altiplano and connecting communities and their resources residing in lower elevations with puna dwellers, were probably beginning to take form. Exchange of resources, including obsidian, between neighboring groups may have been in the context of both maintaining access to resources and risk reduction. From a subsistence perspective, Spielmann (1986: 281) describes these as buffering, a means of alleviating period food shortages by physically accessing them directly in neighboring areas, and mutualism,where complementary foods that are procured or produced are exchanged on a regular basis. Another likely context for obsidian distribution during the Archaic Period is at periodic aggregations. Seasonal aggregations have been well-documented among foragers living in low population densities, where gatherings are the occasion for trade, consumption of surplus food, encountering mates, and the maintenance of social ties and ceremonial obligations (Birdsell 1970: 120;Steward 1938). If analogous gatherings occurred among early foragers in the south-central Andes it have would created an excellent context for the distribution of raw materials, particularly a highly visible material like obsidian that was irregularly available in the landscape.

Late Archaic (5000 - 3300 BCE)

The Late Archaic in the south-central Andes signals the beginning of economic changes in the lead-up to food production, and furthermore the first signs of incipient social and political differentiation are evident in a few archaeological contexts. Some of the obsidian samples reported by Burger et al(2000: 275-288)with possible Late Archaic affiliations are from surface contexts at multicomponent sites and it is therefore difficult to confidently assign these samples to any particular period of the Archaic.

Chivay obsidian during the Late Archaic

Evidence of Late Archaic obsidian use comes from excavations at the previously discussed sites of Asana, Qillqatani, and Sumbay. Several obsidian samples in the Burger et al. (2000) study were of portions of diagnostic projectile points that resemble types in Klink and Aldenderfer's (2005) point chronology. From illustrations and text(Burger, et al. 2000: 279, 281)the Qaqachupa sample appears to belong to the Late Archaic. Burger et al. describe this point as resembling a Type 7 point from Toquepala in Ravines'(1973)classification, and Klink and Aldenderfer(2005: 44)mention that their Type 4D point strongly resembles the Toquepala Type 7, making it diagnostic to the Late Archaic.


In excavated data from Asana, in Moquegua, one sample of Chivay obsidian was found in a Muruq'uta phase occupation in Level XIV-West lying above a palimpsest dated to 6040±90 (Beta-24634; 5210-4766 BCE) (Aldenderfer 1998: 269), as shown above in SectionX3.4.1X. This level lies on the Middle Archaic / Late Archaic transition. Interestingly, Chivay obsidian disappears at Asana subsequent to this time period. Furthermore, evidence of lower raw material diversity in the Late Archaic lithic assemblage point to greater geographical circumscription.


Chivay obsidian at Qillqatani is first found in the Late Archaic level WXXX dated to 5620±120 (Beta-43927; 4800-4200 BCE). The Chivay material is the second oldest obsidian fragment identified from Qillqatani (SectionX3.4.2X), the oldest obsidian is from an as-yet unknown source. Notably, the assemblages from the Qillqatani excavations do not begin to contain obsidian from the Chivay source until considerably later date than did the excavations from Asana.XFigure 3-7Xreveals that both obsidian tools and debris increase as a percentage of the assemblage in the Late Archaic. The counts of obsidian, however, are still relatively low.

Qillqatani is slightly further away than Asana than from the Chivay source, and it is at a higher altitude and further to the east. It has been suggested that perhaps the obsidian was transported via a coastal route at this early date (Frye, et al. 1998). As mentioned, Alca obsidian was also found on the coast in the Terminal Pleistocene levels although, like Chivay, Alca obsidian distributions conform over the long term to a highlands orientation. As no Chivay obsidian as ever been found below the 1250 masl (at Omo), and all Archaic Period Chivay obsidian is found above 3000 masl, the littoral route between Chivay and Asana seems improbable.


Craig and Aldenderfer(In Press)report that two obsidian projectile points from the Ilave valley in a type 3f form, diagnostic to the Late - Terminal Archaic were analyzed in 2005 with a portable XRF unit and were found to be of the Chivay type.


At Sumbay SU-3,three obsidian samples were analyzed from excavated contexts and all three turned out to be of the Chivay type(Burger, et al. 1998: 209;Burger, et al. 2000: 278), as shown in SectionX3.4.3X. Two14C dates were run and returned dates from the early part of the Late Archaic(Ravines 1982: 180-181). One sample was from stratum 3 and it was dated to 6160±120 (BONN-1558; 5400-4750 BCE). Another sample was from above it in stratum 2 and it dated to 5350±90 (BONN-1559; 4350-3980 BCE). One of the three obsidian samples came from Stratum 4 of unit 5, while the other two samples came from higher levels. The Stratum 4 sample probably dates to the Middle Archaic.


Obsidian from securely dated Late Archaic contexts show something of a reduction in regional distribution and a greater focus on locally available lithic material, suggesting a reduction in mobility or exchange in the Late Archaic. Similarly, projectile point styles became increasingly more limited in spatial distribution, with greater local variability during the Late Archaic implying reduced mobility(Klink and Aldenderfer 2005: 53). This is consistent with Aldenderfer's(1998: 260-261)observations about reduced mobility during the Late Archaic Qhuna Phase occupation at Asana when the occupants ceased to use non-local lithic raw materials. During this phase, Aldenderfer also describes increasingly formalized use of space at Asana, evidence of a ceremonial complex, and greater investment in seed grinding. In short, during this time a circumscribed population with reduced mobility was probably living in higher densities and exhibiting signs of ceremonial activity that are consistent with the scalar stress model for the emergence of leadership(Johnson 1982).

It is also worth considering the impact that scarcity may have on valuation. The lack of discarded obsidian signifies that it was not being knapped or resharpened and it was probably not abundant, but that does not mean that obsidian was not known in the larger consumption zone during this period. In the subsequent time period, the Terminal Archaic, obsidian becomes abundant on a regional scale at the same time as a host of other social and economic changes were occurring. This period, and the previously discussed Middle Archaic, correspond with what was referred to as the silencio arqueologico(Núñez and Santoro 1988)due to a dearth of archaeological data observed by investigators working in Northern Chile. The reduced evidence of circulation of obsidian from Chivay appears to correlate with a reduction in archaeological evidence regionally.

Thereis strong representation of Chivay obsidian in Arequipa at Sumbay, and it is likely that Late Archaic projectile point forms are found in the North Titicaca Basin as reported by Burger et al. (2000). However, at Asana there is little obsidian from the Colca. Possibly these reduced distributions of obsidian reflect the reduced mobility and more complex architectural investment in Late Archaic contexts at Asana(Aldenderfer 1998), and prior to the development of extensive, long distance exchange that were potentially initiated by early caravan networks during the Terminal Archaic.These conclusions, however, on not based on particularly robust data, as the sample of sites for this time period is relatively small.

Discussion and review of the Archaic Foragers period

The transport of obsidian from both the Chivay and Alca sources fluctuated during the Archaic Foragers period, and the distribution remained confined to the sierra and altiplano areas of the south-central Andes. The small and irregular quantities of obsidian consumed throughout the region could have been procured and transported by mobile foragers, or conveyed through down-the-line exchange networks. Conceivably exchange could have taken place a seasonal gatherings as well, based on analogy from other foraging groups worldwide, although there is no direct empirical support for such gatherings in the south-central Andean Archaic. At Asana, the most well-stratified highland Archaic site investigated to date, evidence of non-local obsidian was intermittent and consisted of very small quantities of material in a pattern that is perhaps exemplary of the discontinuous nature of regional exchange during this time.

Obsidian from the Quispisisa source (340 linear km to the north-west of the Chivay source), was likewise used primarily in highland contexts during the Archaic Foragers period, particularly in rock shelters excavated by MacNeish and colleagues in the central highlands. While Quispisisa material was circulated at a number of preceramic coastal sites(Burger and Asaro 1978), current evidence suggests that those contexts post-date 3300 cal BCE (i.e., cotton preceramic) and are therefore Terminal Archaic in the terms used here. It appears that while Quispisisa material was used widely on the coasts subsequent to the Archaic Foragers period, material from Chivay and Alca were virtually always confined in the highlands with the exception of the earliest obsidian evidence of all: Alca at Quebrada Jaguay. In the Archaic Foragers period, as in later periods, the evidence from obsidian is contrary to the widely-discussed Andean vertical complementarity models, because obsidian distributions suggest that the movement of people or products between ecologically complementary zones was not widespread.

3.5.2.Early Agropastoralist obsidian distributions

The "Early Agropastoralists" period (3,300 BCE - A.D. 400) begins with what appears to have been a shift to a chiefly pastoral lifeway, greater regional interaction, and a more intensive production and circulation of obsidian from the Chivay source area in the Terminal Archaic. In this discussion, the Early Agropastoralists period continues through the Late Formative and subsequently beginning AD400, with the ascendancy of Tiwanaku, the "Late Prehispanic" time block begins. The changes during the Terminal Archaic that mark the beginning of the Early Agropastoralist time include the growing importance of food production, the expanded production and circulation of obsidian, and socio-political differentiation that began to appear in the Terminal Archaic, phenomena of greater interest to this research than the presence or absence of pottery.


Figure 3-12. Chivay type obsidian distributions during the "Early Agropastoralists" time (3,300 BCE - AD 400).

Why an "Early Agropastoralists" time block?

In lieu of the traditional preceramic / ceramic divide in archaeology, the "Early Agropastoralist" time block was defined for this project because, in many ways, the conditions during the final millennium of the Archaic had more in common with events in the Formative than with the preceding Archaic Periods. Differentiating sites as "Early Agropastoralist" beginning in the Terminal Archaic requires that one delimits a firm boundary on a process that was millennia in the making. Considering the Terminal Archaic as the beginnings of agropastoralism, and thereby linking it to the attendant settlement distribution that includes series 5 projectile points, connects cultural and economic evidence from throughout the region with a generalized estimate of when food production took hold in terms of scheduling, mobility, and prioritizing the needs of agricultural planting and herding. As mentioned, the term "Early Pastoral" is not meant to imply that no food production occurred before this date and hunting and gathering did not persist after this date. Indeed the current evidence suggests that camelids were first domesticated sometime during the Middle or Late Archaic, probably after 5000 BCE (Browman 1989;Kadwell, et al. 2001;Wheeler 1983;Wing 1986). The "Early Agropastoralist" period refers to a period after 3300 BCE that features a predominantly food producing economy in the highlands including intensive pastoralism, as well as seed-plant and tuber cultivation. These changes in food production may have been linked to the accumulating evidence of increased rainfall and shortening of the dry season beginning circa 2500 cal BCE (Baker, et al. 2001;Marchant and Hooghiemstra 2004;Mourguiart 2000). These changes in economy are joined by evidence of increased sedentism, widening stylistic distributions pointing to long distance cultural integration, and the beginnings of social differentiation apparent in architecture and grave goods. The circulation of obsidian was linked to these phenomena because increased regional interaction was perhaps articulated through camelid caravans, although the presence of caravans in the Terminal Archaic is being explored, not assumed, in this study.

In the Titicaca Basin, the changes incurred between the Terminal Archaic and the Late Formative are monumental. However, from the perspective of the Chivay source the intensification revealed archaeologically, both at the source and in the consumption zone, is notable beginning in the Terminal Archaic, and the intensification does not change notably throughout the Formative Period. Defining the ending of the "Early Agropastoralist" period as the end of the Late Formative is similarly problematic. By the Late Formative, the economic influence of Pukara and other regional centers probably impacted the peoples of the Upper Colca, but direct evidence returning from Pukara, only 140 km away, is scarce in the Colca. In contrast, during the Tiwanaku times, regional states with socio-economic impacts over hundreds of kilometers dominated both the circulation of goods like obsidian, and regional settlement organization.

Terminal Archaic (3300 - 2000 BCE)

The Terminal Archaic ushers in a suite of social and economic changes in the south-central Andes and, consistent with these developments, obsidian begins to circulate in significantly greater quantities during this period. Obsidian is increasingly used for projectile points during the Terminal Archaic, a trend that continues into the Formative Period (Burger, et al. 2000: 294).

Regional patterns in Terminal Archaic sites are somewhat difficult to assess from surface finds because the Terminal Archaic is lacking in exclusive diagnostic artifacts in both the lithic or ceramic artifact classes. As is shown inXFigure 3-10X, the most common projectile point styles that belong to the Terminal Archaic, such as Types 5B and 5D, also persist through the ceramic periods, leaving only type 5A and part of type 4F as diagnostic to exclusively the Terminal Archaic (Klink and Aldenderfer 2005: 48). Furthermore, by definition, the Terminal Archaic is a preceramic period, which precludes a ceramic means of assigning chronology. From site organization characteristics, a site might be considered to be a Terminal Archaic site if it has pastoralist attributes but it is aceramic and has series 5 projectile point types represented that belong, at least partially, to the Terminal Archaic.

The beginnings of the Andean agropastoral strategy are apparent during this period at sites like Asana where seasonal residential movement occurred between the high sierra and the puna (Aldenderfer 1998: 261-275;Kuznar 1995). Another attribute of the Terminal Archaic at Asana is a disappearance of ceremonial features. Evidence for a shift from hunting to pastoralism comes primarily from evidence of corrals (Aldenderfer 1998), from changes in the ratio of deer to camelid remains, and from the ratio of camelid neonate to adult remains. At sites where the process has been observed, the transition to full pastoralism at various sites in the Andes is usually perceived as a gradual process dating to sometime between 3300 to 1500 BCE At Qillqatani, however, the transition was relatively abrupt and it occurred in level WXXIV that dates to approximately 2210-1880 BCE (XTable 3-1X).

Chivay obsidian during the Terminal Archaic

While the number of excavated Terminal Archaic sites is relatively small, general processes are apparent from recent work at several sites in the south-central Andean highlands.

At Ch'uxuqullu on the Island of the Sun, Stanish et al.(2002)report three obsidian samples from the Chivay source in Preceramic levels. Eight obsidian flakes were found in aceramic levels and were described as being from a middle stage of manufacture, and three of these were sourced with NAA. Two samples came from levels with a14C date of 3780±100bp (Teledyne-I-18, 314; 2500-1900 BCE) and a third Chivay obsidian sample comes from an aceramic level that immediately predates the first ceramics level which occurred at 3110±45bp (AMS-NSF; 1460-1260 BCE). Citing paleoclimate data, the authors observe that boat travel was required to access the Island of the Sun at this time.

The Ilave Valley and Jiskairumoko

Terminal Archaic projectile points in the Ilave valley demonstrate the dramatic shift in the frequency and use of obsidian that occurred in this time period (SectionX3.4.4X). XRF analysis of 68 obsidian artifacts excavated from Jiskairumoko show that the Chivay source was used with particular intensity in this period. The XRF study found that 97% of obsidian bifaces from Terminal Archaic and Early Formative levels at Jiskairumoko were from the Chivay source, however published obsidian studies from all time periods show that typically 90% of all obsidian analyzed from Titicaca Basin sites are from the Chivay source (XTable 3-3X).


The percentages of obsidian tools and debris remain generally similar to those in the Late Archaic level except that the counts are much higher (XFigure 3-7X). By count, obsidian tools are doubled, and obsidian debris is 4.3 times greater, and all the obsidian has visual characteristics of the Chivay type (Aldenderfer 1999).

Based on the density of obsidian, the relationship with the Chivay source area 221 km away seems to be well-established by this time, and it is a relationship that becomes even more well-developed in the Early Formative. Six samples of obsidian were analyzed at the MURR facility from Terminal Archaic contexts that are associated with a radiocarbon date of 3660±120 (Beta-43926; 2210-1880 BCE). All six obsidian samples were from the Chivay source.


At the site of Asana, obsidian reappears at the end of the Terminal Archaic during the Awati Phase dated to3640±80 (Beta-23364; 2300-1750 BCE)where it makes up 0.4% of lithic materials, but this obsidian was not from the Chivay source. It was judged from distinctive visual characteristics to have come from the Aconcagua obsidian source only 84 km to the east of Asana, near the town of Mazo Cruz(Aldenderfer 2000). Aconcahua type obsidian has characteristics that are less desirable for knapping due to fractures and perlitic veins that cross cut the material (see Appendix B.1), and while it was possible to derive sharp flakes for shearing and butchering functions, the material was probably not used for projectile point production(Frye, et al. 1998).

The shift to Aconcahua obsidian in the Awati phase at Asana is particularly puzzling given the evidence for Chivay obsidian circulation at this time period. It is precisely at the end of the Terminal Archaic that a dramatic spike in the use of Chivay obsidian at Qillqatani (SeeXTable 3-5XandXFigure 3-7X) took place. One may ask: Why is it that when the occupants of Qillqatani are importing Chivay obsidian in unprecedented quantities, the people of Asana are getting only small quantities of low-quality obsidian? In addition, this low-quality obsidian comes from Aconcahua, a source adjacent to Qillqatani?

Given the pattern of early Chivay obsidian at Asana, these Terminal Archaic distributions suggest that the high sierra residents at Asana were not participating in an altiplano-based circulation of goods as the Qillqatani residents. The residents of Asana never again participate in the circulation and consumption of Chivay obsidian, while at Qillqatani the consumption of Chivay material continues strongly for another one thousand years.

Other obsidian types during the Terminal Archaic

Obsidian from alternative sources in the circum-Titicaca region, including the unlocated sources of Tumuku and Chumbivilcas types, are used in greater quantity during the Terminal Archaic judging from associated projectile point evidence provided in Burger et al.(2000: 280-284).

Evidence from close to the Alca obsidian source provides new information about long distance interaction during Terminal Archaic. At the site of Waynuña at 3600 masl(Jennings 2002: 540-546)and less than one day's travel from the Alca obsidian source, recent investigations have uncovered a residential structure with evidence from starch grains resulting from the processing of corn as well as starch from arrowroot, a plant necessarily procured in the Amazon basin(Perry, et al. 2006). Given the long distance transport of arrowroot, it is conceivable that the plant material arrived as a form of reciprocation or direct transport from travelers moving between the Amazon and the Alca obsidian source. The Cotahuasi valley also has major salt source and other minerals that would potentially draw people procuring such materials. The starch samples were found in a structure on a floor dated by two14C samples. One sample was dated to 3431±45 (BGS-2576) 1880-1620 BCE, and another was 3745±65 (BGS-2573) 2350-1950 BCE

Further north, the Quispisisa type obsidian was particularly abundant during the Terminal Archaic at the preceramic coastal shell mound site of San Nicolas, along the Nasca coast, in a context associated with early cotton(Burger and Asaro 1977;Burger and Asaro 1978: 63-65). Quantitative data on the consumption of obsidian at San Nicolas are unavailable and the temporal control is weak because the "cotton preceramic" date is derived from association with cotton and no ceramics, not from direct14C dating.


The distribution of obsidian from all three major Andean obsidian sources: Chivay, Alca, and Quispisisa, expanded considerably during the Terminal Archaic. It is notable that both the Chivay and Alca sources expanded, but the distribution remained confined to the sierra and altiplano areas of the south-central Andes. In comparison, obsidian from the Quispisisa source (340 linear km to the north-west of the Chivay source), has been found in significant quantities in Ica on the coast of Peru. While many of the early coastal obsidian samples have weak chronological control, the quantities of Quispisisa obsidian found in possible Archaic contexts is noteworthy. The fact that Chivay obsidian has never been found in coastal areas, and Alca is not found on the coast after the Paleoindian period, is remarkable considering the extensive evidence of coastal use of Quispisisa obsidian beginning in the Terminal Archaic.

Early Formative Period (2000 - 1300 BCE)

In this time period, new socio-economic patterns became well established in the south-central Andes while the distribution of Chivay obsidian was at its maximum both in geographic extent and in variability of site types. During this time period, societies were characterized by sedentism, demographic growth, increased specialization, and early evidence of social ranking(Aldenderfer 1989;Craig 2005;Stanish 2003: 99-109).

At sites like Jiskairumoko in the Titicaca Basin, architecture from the Terminal Archaic consisted of pithouses with internal storage, ground-stone, and interments with grave goods that included non-local items (like obsidian) that were presumably of value(Craig 2005). During the Early Formative around 1500 BCE this architectural pattern gives way to larger, rectangular, above-ground structures that lacked internal storage.

As reviewed by Burger et al. (2000: 288-296) using the Ica "Master sequence" chronology under the roughly contemporaneous "Initial Period" (though the Early Formative ends 300 calendar years earlier), the distributions of obsidian are notable in their extension both north and south from the Chivay source, and in their concentrations at early centers like Qaluyu. Obsidian from Chivay also persists on the Island of the Sun in the Early Formative(Stanish, et al. 2002).

Chivay obsidian distributions

At Qillqatani, obsidian flakes are available in quantity during this Titicaca time period which roughly overlap with the Formative A and B (XTable 3-5X). Seven obsidian samples were analyzed at MURR that all corresponded with the Chivay source. These samples were found in levels adjacent to a context that dated to2940±70bp (Beta-43925; 1380-970 BCE).

It appears that obsidian was being obtained as nodules or blanks and being knapped down to projectile points. Obsidian flakes represent 15% by count (n = 160) of all the flakes in the period "Formative A", and obsidian projectile points represent 12% by count of all points in this level(Aldenderfer 1999).

Obsidian is first found at Qaluyu dated to ~3250 uncal bp which calibrates to 1640-1420 BCE(Burger, et al. 2000: 291-296), or the end of the Early Formative where Burger et al. report evidence of nodules or blanks arriving for further reduction at the source. Some of the Chivay obsidian found at Qaluyu were medium sized pebbles; one example Burger et al. analyzed had the dimensions 1.9 x 1.4 x 1.2 cm. Qaluyu is 140 linear km from the Chivay source, or 34.6 hours by the hiking model, so it is roughly 70% of the distance of Qillqatani from the source. There are substantial quantities of obsidian at Qaluyu during the subsequent Middle Formative.


Paleoclimatic studies have documented a decline in abundance of arboreal species in the Lake Titicaca Basin and an increase in open-ground weed species reflecting disturbed soils after ca. 1150 cal BCE(Paduano, et al. 2003: 274). This has been interpreted as reflecting intensification in food production and population that occurred during the Early Formative. Some argue that the grasslands of the altiplano are an anthropogenic artifact of a human induced lowering of the treeline due in part to the expansion of camelid pastoralism(Gade 1999: 42-74). The increasing sedentism, cultivated plants, expanding camelid herds and the growing influence of prominent settlements like Qaluyu in regional organization signify that significant socio-economic dynamism was underway by this period. The evidence of hierarchy takes the form of "very moderate social rank" acquired by certain individuals, such as religious specialists(Stanish 2003: 108). These traits are first suggested by jewelry, grave goods, and non-residential structures in the Terminal Archaic, but they become further elaborated during the Early Formative after 2000 BCE.

There is no evidence of political ranking in the structure of settlement distributions between Early Formative sites in the Lake Titicaca Basin, and all the sites are less than one hectare in size(Stanish 2003: 108), but the evidence of obsidian circulation from Qillqatani and other pastoral sites suggests of regional exchange system not yet focused around Lake Titicaca. Evidence acquired from the earliest occupational levels at Titicaca Basin regional centers suggest that regional interaction was low. At the site of Chiripa, in the southern Lake Titicaca Basin, Matthew Bandy describes traces of sodalite beads, obsidian, and sea shell in excavations from Early and Middle Formative contexts.

It is clear, then, that long distance trade in mortuary and prestige items took place as early as the Early Chiripa phase [1500-1000 BCE]. Equally clear, however, is that this early exchange involved very small quantities of the objects in question. This trading would seem to have been very sporadic and infrequent. There is no evidence in the Early Formative Period for the sort of regular caravan trade postulated by Browman (his "altiplano mode"; see Browman 1981: 414-415) (Bandy 2001: 141).

However, the evidence for obsidian distributions at Qillqatani that are discussed here point to routine exchange, perhaps by caravan trade, along on the Western Cordillera in the Early Formative. It appears that future regional centers like Chiripa were not yet participants in these exchange networks.

The dynamic nature of interregional interaction that occurred during the Early Formative is evident from the diversity in obsidian types at sites, and the lack of geographical restrictions that appear to structure exchange during later periods. While settled communities were evident, the lack of hierarchy and centralization suggests that they were not integrated by supra-local organization. Yet the economic basis for long distance relationships appears to have taken form by the Early Formative. The decentralized and variable nature of exchange in this period, which is abundantly evident at Qillqatani and yet not evident at sites like Chiripa, implies that another form of integration was linking communities, like the residents of Qillqatani, with the Chivay source 221 km to the north-west.

Middle Formative Period (1300 - 500 BCE)

During the Middle Formative Period, social ranking became established in the Titicaca Basin. The changes are most evident in the settlement structure as some regional centers grew to become far larger than their neighbors and feature sunken courts, mounds, and specialized stone and ceramic traditions. Stanish(2003: 109-110)interprets these changes in terms of an ability of elites to mobilize labor beyond the household level.

The stylistic evidence suggests that during the Middle Formative the north and south Titicaca Basin were relatively separate spheres, with Qaluyu pottery in the north and fiber-tempered Chiripa ceramics in the south extending only as far north as the Ilave river. However, Chivay obsidian is encountered in both the North and South Basin. Christine Hastorf (2005: 75)suggests that by the end of this period (the Early Upper Formative) evidence of ethnic identity and ritual activity is supported by ritual architectural construction and non-local exchange goods. It is further inferred that "Plants such as coca ( Erythroxylumsp.), Anadenanthera ( A. colubrine, A. peregrine), and tobacco (Nicotiana rustica)surely would have been present in the Basin by this time, perhaps associated with snuff trays…"(Hastorf 2005: 75). An increase in long distance exchange is commonly found as part of a complex of features associated with ideological and social power during the Middle Formative in the region, and it appears that existing exchange routes, such as the one along the Western Cordillera connecting Chivay with Qillqatani, were increasingly routed towards the Titicaca Basin regional centers during this time.

Chivay obsidian distributions

Chivay obsidian occurs in small quantities at a number of Middle Formative sites in the southern Basin including Chiripa and Tumatumani, and it persists at Ch'uxuqullu on the Island of the Sun. At Tumatumani, 3% of the projectile points are made from obsidian(Stanish and Steadman 1994). Bandy(2001: 141)reports that at Chiripa they recovered only small quantities of obsidian in the time spanning 1500-200 BCE and these were in the form of finished bifaces. For the entire time span, after four excavation seasons, they report only 87.1g of obsidian. Tumuku type obsidian was identified in Chiripa in Condori 1B component circa 1500-1000 cal BCE levels(Browman 1998: 310, dates calibrated). At the site of Camata, on the lakeshore south of the modern city of Puno, four obsidian samples were analyzed from contexts that range from circa 1500 - 500 cal BCE and all four were of Chivay type obsidian(Frye, et al. 1998;Steadman 1995).

The evidence from the North Titicaca Basin regional centers is even more intriguing as there is a significant presence of Alca type obsidian during the early part of the Middle Formative, and Chivay obsidian is found in the Cusco Basin during this time. Qaluyu, Pikicallepata, and Marcavalle all contain both Chivay and Alca obsidian between approximately 1100 - 800 cal. BCE, the early part of the Middle Formative(Burger, et al. 2000: 292;Chávez 1980: 249-253). Subsequent to this overlap in obsidian use, there appears to have been significant overlap in other stylistic attributes as well. These similarities include common traits in ceramic vessel forms between Chanapata vessels in the Cusco area and Qaluyu vessels in the North Titicaca Basin(Burger, et al. 2000: 292).

However, during the latter part of the Middle Formative after 800 BCE the Yaya-Mama religious tradition first emerges at the site of Chiripa(Bandy 2004: 330;Chávez 1988), a tradition that eventually unifies the north and south areas of the Titicaca Basin during the Late Formative. As noted by Burger et al.(2000: 311-314), with the appearance of the Yaya-Mama tradition the Alca and Chivay obsidian distributions become more asymmetrical. Alca obsidian makes up 16% (n = 9) of the obsidian in a pre-Pukara context at the site of Taraco on the Titicaca lake edge in the North Basin, however while Chivay obsidian was found at Marcavalle and other Cusco sites previously during the Early Formative, obsidian from the Chivay source is absent during the Middle Formative and it does not re-appear in the Cusco region again until the Inka period. Alca obsidian, on the other hand, expands outward during this period as it is found in the Titicaca Basin to the south-east, and it is also is transported a great distance to Chavín de Huantar. Both of these examples of long distance transport have been attributed to religious pilgrimage(Burger, et al. 2000: 314).


At Qillqatani, the Middle Formative comprises the bulk of Formative B layers and all of the Formative C layers (see Qillqatani data in Section X3.4.2X). In Formative B layers, Chivay obsidian is the only type represented in the four samples that were analyzed and the lithic assemblage suggests that formal tools were not being produced at the site as no evidence of obsidian tools were found in these levels. Obsidian flakes, however, persist as 18% of the lithic assemblage from that level. Subsequently, in the Formative C level that begins around 900 BCE and corresponds approximately with the latter half of the Middle Formative as well as the rise of the Yaya-Mama tradition in the south Titicaca Basin, there is a distinctive shift in the use of obsidian at Qillqatani. Whereas all prior obsidian samples from Qillqatani were Chivay after the initial Middle Archaic sample, the obsidian samples in Qillqatani Formative C levels are only 60% from the Chivay source.

The other samples come from Aconcahua, a source of lower-quality obsidian that is near the Qillqatani shelter, and from Tumuku, an as-yet undiscovered source that may be located close to the three-way border between Peru, Bolivia, and Chile, and finally Alca obsidian occurs for the first and only time at Qillqatani in these levels. Given that Alca obsidian also occurs at Pukara, Incatunahuiri (surface) and at Taraco in quantity (16% of assemblage, n = 9), the presence of Alca obsidian at Qillqatani is consistent with the abundance of Alca material in circulation in that time. A sample of Alca obsidian has also been found on the Island of the Sun (Frye, et al. 1998), though it was from a surface context.

Table 3-6Xshows that obsidian tools in Formative C levels at Qillqatani are abundant (n = 19) and relatively large on average (1.21 g), and the non-obsidian tools were also very abundant (n = 187) for this level.


Middle Formative obsidian distributions appear to demonstrate the emergence of a distinctive Titicaca Basin exchange sphere. One could argue that the emerging elites that mobilized labor to build the initial mounds and courts, and sponsored specialized artistry in stone and ceramics, may have precipitated a demand for greater exotic exchange goods as a source of prestige. Stanish(2003: 162)believes that this is the process that occurs later, during the Late Formative, when he argues that this process is connected to wealth generation for sponsoring feasts and other activities, though he admits the data are sparse. Evidence of long distance exchange from contexts belonging to the Early Formative and first half of the Middle Formative (1500-1000 BCE) at sites like Chiripa are sparse, irregular, and generally involve very small, portable goods; however the evidence from Qillqatani supports other models of more regular interaction along established exchange routes.

Late Formative Period (500 BCE - AD 400)

During the Late Formative, significant social ranking developed and dominated the socio-political landscape in the Lake Titicaca Basin. The complex polities that emerged on either end of Lake Titicaca were distinguished by architecture, stoneworking, and ceramic traditions. A three-tiered site size hierarchy is evident in the Late Formative, and the construction of prominent terraced mounds with sunken courts occurred in a few major sites at this time. The elite ceramics and stoneworking, and the construction of elaborate mounds as a venue for large-scale feasts and human sacrifice at sites like Pukara and early Tiwanaku, can be interpreted as a means of demonstrating the large-scale organization of labor by elites (Stanish 2003: 143).

The patterns of obsidian circulation that emerged at the end of the Middle Formative became very well established in the Late Formative. Excavations at centers in the North Basin have found a reduction in the presence of Alca obsidian in the Titicaca Basin, but the Alca material is still present although it is found in minor quantities at the larger sites as compared with Chivay obsidian. The diversity of obsidian types in the Titicaca Basin samples reported by Burger et al. (2000: 306-308) for this time span is low, as compared with the diversity of types used in the Cusco area, because in the Titicaca Basin it is virtually all Chivay obsidian.

In recent excavation work at Pukara on the central pampa at the base of the Qalasaya, Elizabeth Klarich (2005: 255-256) found that obsidian was generally available and obsidian use was not associated with any intrasite status differences. Probable representations of obsidian points and knives appear in Pukara iconography, and small discs have been found at Pukara and Taraco that were possible ceramic inlays (Burger, et al. 2000: 320-321). The Late Formative Period on the southern end of the Titicaca Basin has small amounts of Chivay obsidian at Chiripa and Kallamarka. InGiesso's (2000: 167-168)review of lithic evidence from several Formative Period sites he found no obsidian use in these collections except for samples from Khonkho Wankane which recent research directed by John Janusek has revealed to be principally a Late Formative center. The furthest confirmed examples of transport of Chivay obsidian are these examples encountered in southern Titicaca Basin sites. The two furthest confirmed contexts for Chivay obsidian transport are represented bytwo samples from a Late Formative context at Kallamarka (Burger, et al. 2000: 308, 319, 323) and six samples found at Khonkho Wankane (Giesso 2000: 346), both about 325 km from the Chivay source, or 72 hours by the walking model.

At Qillqatani, the obsidian returns to being primarily Chivay type during the Late Formative although 1 out of the 9 obsidian artifacts analyzed from Late Formative levels was of the Tumuku type (SectionX3.4.2X). During the Late Formative there is a steady decline in the percentage of debitage made from obsidian at Qillqatani, a trend that continues in the Tiwanaku period.

In the Colca Valley, there are few traces of the Late Formative consumers of Chivay obsidian. As will be further explored in this research project, there is very little Qaluyu or Pukara material diagnostic to the Middle or Late Formative in the Chivay obsidian source region. Steven Wernke found a diagnostic classic Pukara sherd with a post-fire incised zoomorphic motif that resembles a camelid-foot from a Formative site above the town of Yanque in the Colca (Wernke 2003: 137-138).

Pukara materials are known to have circulated, probably through trade links, throughout the south-central Andes. Pukara sherds have been found in the valley of Arequipa in association with the local Formative Socabaya ceramics (Cardona Rosas 2002: 55). In the Moquegua valley, both Chiripa-related and Pukara pottery have been found (Feldman 1989), and a Pukara textile was found in a possible elite grave context in the Ica Valley (Conklin 1983).


The decline in forest cover in the Lake Titicaca Basin beginning during the Early Formative intensified throughout the Formative Period. Between 2500 - 800 cal. BCE a decline in fine particulate charcoal was detected in a lakecore drawn from the southern end of Lake Titicaca's Lago Grande(Paduano, et al. 2003: 274). The pollen and charcoal record indicates thatforest cover finally disappeared around AD0, coincident with population increases and resource pressure associated with Late Formative Period socio-economic intensification.

The obsidian circulation during the Late Formative shows distinct patterns in the Titicaca Basin and in Cusco. In the Titicaca Basin, the diversity of sources is reduced, with virtually all the material coming from Chivay and Alca, and the Alca samples are primarily affiliated with a single period at Taraco. Even at the site of Qillqatani, the diversity is reduced as compared with the previous Middle Formative level (Qillqatani Formative C). Current evidence suggests that the economic circulation was more integrated and that it was probably under some form of control by the dominant regional centers of this time. The furthest confirmed evidence of Chivay obsidian transport is from among Late Formative contexts at Kallamarka and Khonkho Wankane.Yet, diagnostic evidence from Titicaca Basin Late Formative polities in the Colca valley area is scarce. Part of the difficulty in understanding the Formative at the Chivay obsidian source area is that the Formative ceramic sequence in the Arequipa highlands is still being refined.

Discussion and review of "Early Agropastoralists" obsidian

From the Terminal Archaic through the Formative Periods the circulation and consumption of obsidian expanded dramatically. Chivay and Alca obsidian is found in a wide variety of sites from isolated rock shelters to early regional centers, and it appears in consistent quantities that differ considerably from the intermittent nature of regional obsidian supplies in preceding periods. The increased circulation of obsidian is part of a spectrum of changes that began in the Terminal Archaic, but obsidian distribution patterns are particularly useful because they are quantifiable.

In some ways obsidian appears most closely linked to tasks associated with camelid pastoralism, such as shearing and butchering. However, in the central Andean littoral, Quispisisa obsidian is found in density in "cotton preceramic" and Initial Period (i.e., Terminal Archaic and Early Formative) coastal sites where presumably pastoralism, if present, was a very minor part of the economy. Furthermore, in the south-central Andean highlands obsidian is found in pastoralist rock shelters but also in regional centers, and the small discarded obsidian flakes and small projectile points do not appear to have served as adequate shearing implements due to their size.

Social complexity during first part of the "Early Agropastoralists" period is manifested most prominently along the Pacific littoral. In the coastal context of what is now the Department of Lima, in central Peru, monumental architecture dating back to 3000 BCE is well-established at the site of Aspero, and the transition to yet more monumental preceramic construction at sites slightly inland has been documented in recent research(Haas, et al. 2005;Shady Solis, et al. 2001). The shift inland is argued to be related, among other things, to the increased importance of harnessing labor surpluses through the production of cotton for anchovy nets and textiles, and to competitive monument building between elites. Evidence of exchange with highland and Amazonian groups is apparent in the form of tropical feathers and other non-local prestige goods. While these developments occurred over one thousand km to the north of the Chivay area along the Pacific coast, early complex organization is also reported in northern coastal Chile in the Chinchorro II and III traditions spanning the Late Archaic through the Middle Formative(Rivera 1991). A long history of cotton production for nets and textiles, imported wool textiles, elaborate burial traditions, and long distance exchange with the highlands and the Amazon characterize the later Chinchorro tradition. These regional patterns underscore the wide scope of the changes that occurred during the Early Agropastoralists time. It has been argued that the beginning of social inequality in the highland Andes may have been stimulated indirectly by demand for wool from aggrandizers in coastal societies(Aldenderfer 1999).

Returning to evidence from highland obsidian distributions, the widespread circulation of obsidian during the Terminal Archaic and Formative is perhaps best discussed in the terminology used in the historical model of Nuñez and Dillehay(Dillehay and Nuñez 1988;1995 [1979]). While this adaptationalist model has theoretical limitations, it serves as a useful alternative to evolutionary chiefdom models that assign a paramount role to central-places in exchange, despite the decline of central-place models in geography(Smith 1976: 24). Following the Nuñez and Dillehay model, it is possible that what is occurring in the Terminal Archaic and Early Formative is the emergence of regular exchange pattern between Qillqatani and Chivay that took the form of a caravan trade "axis" along the western Cordillera. This exchange pattern is among the earliest systematic and demonstrable cases of the circulation of diffusive items through largely homogenous altiplano terrain (XFigure 3-4X) in the pattern that has also been described as horizontal complementarity or the "Altiplano mode"(Browman 1981). Subsequently, during the Middle and Late Formative in the Titicaca Basin, former "axis settlements" like Qillqatani, and the Western Cordillera axis more generally, became relatively less important in long distance exchange relationships. Regional centers in the Titicaca Basin like Taraco, Pukara, Chiripa, Khonkho Wankane, early Tiwanaku, and other centers, expanded in influence during this period of peer-polity competition. In these circumstances, the acquisition and ceremonial use of exotic goods appears to have been an important part of the competitive strategies of aggrandizers. Other evidence for Titicaca Basin-based exchange dynamics include the possible production of hoes of Incatunahuiri olivine basalt at Camata around 850-650 BCE that were then transported and used in southern Titicaca Basin sites(Bandy 2005: 96;Frye and Steadman 2001). While the nature of the relationship between emerging elites in Titicaca Basin polities and caravan drivers that provided links between settlements throughout the region is difficult to describe with precision, it appears that interregional articulation became considerably more elaborate by the end of the Early Agropastoralist period.

3.5.3.The Late Prehispanic

During the Late Prehispanic period the circulation of Chivay obsidian was subjected to pan-regional forces by expansive Tiwanaku and Wari and again during the Inka period. While obsidian appears to have had fairly consistent use in both the Tiwanaku and Wari domains, the use of obsidian relative to other goods appears to have declined.


Figure 3-13. Chivay type obsidian distributions during the "Late Prehispanic" time (AD 400 - 1532).

Tiwanaku and Wari (AD400 - 1100)

Following the Titicaca Basin chronology the Tiwanaku period begins with the emergence of the Tiwanaku state during the time that has been called "Tiwanaku IV" beginning around AD400 (Stanish 2003). This is centuries earlier than the start of the Middle Horizon, as defined by the presence of Wari in Ica. Wari in Ica spans the period from approximately AD 750 - 1000.

Obsidian consumption at Tiwanaku

Research by Martín Giesso (2000;2003) on the Tiwanaku lithic industry has revealed patterns in the differential use of lithic material in the region. First, it should be noted that obsidian is not abundant in the Tiwanaku core area in terms of the capital of the expansive state. Giesso (2003: 365-366) notes that obsidian artifacts make up only 0.8% of the lithics by count (n=86), of the collections at Tiwanaku. Yet the Tiwanaku core lies 315 km from the Chivay source in Euclidean distance, or 70 hours by the hiking function and therefore while obsidian is relatively scarce, it has been conveyed a considerable distance. The spatial distribution and the artifact form of obsidian at Tiwanaku are more analytically relevant than the total count or weight of obsidian which was quite small. Intrasite data on the contrasts in temporal and spatial use of obsidian at Tiwanaku are critical because, among other things, these data would provide a gauge of the relative persistence of access to the Chivay source as Tiwanaku's regional influence expanded.

Giesso's comprehensive source sampling showed that ten different types of obsidian were in use in the Tiwanaku heartland, although of these ten samples only four samples derive from known source locations. He notes that of the ten obsidian types, only Chivay (Cotallaulli) type is transparent and rest are described as "opaque". Giesso's spatial assessment of the distribution of lithic production activities in the core region showed that obsidian microdebitage was concentrated at certain mounds and in residential sectors as compared with chert, quartzite, and other locally available products. Giesso also organized a sourcing study of basalt artifacts in the Tiwanaku area, and he examined the contexts of production at the Querimita quarry located on the southwestern shores of Lake Poopó in Bolivia (Giesso 2003: 369) just over 300 km south of Tiwanaku. The evidence from Querimita on the production and consumption of basalt provides an interesting regional contrast to the spatial patterning of Chivay obsidian. There are both Wankarani (Formative) and Tiwanaku sites in the vicinity of the Querimita basalt source in the direction of the shores of Lake Poopó, but diagnostic Tiwanaku materials at the quarry itself are not reported (M. Giesso 2006, pers. comm.).

At Tiwanaku, quantities of obsidian and quartz have been found in the construction fill at the ceremonial Mollo Kontu mound, a structure that has served as a local fertility shrine. Nicole Couture argues that "quartz and obsidian fragments also served as mountain icons, in the way that exotic crystals, minerals, and rock candies are used today by Aymara yatiris, or ritual specialists, to represent mountains and thunder in rites to promote agricultural and social fertility" (Couture 2003: 225). Obsidian from Mollo Kontu has been traced to six different sources: Quispisisa, Chivay, Sora Sora, Cerro Zapaleri, and two unlocated sources (Giesso 2000). "The high density of quartz and obsidian artifacts, often five to ten pieces per excavation level, indicates that they were not accidentally included as part of redeposited rubbish, but rather were deliberately added to the clay fill" (Couture 2003: 215). The diversity in the Tiwanaku urban core was such that Chivay material represented a relatively low percentage (76%) of the material, although the ratio of Chivay material in Giesso and Glascock's study on the whole was 90% Chivay, which is typical for the Titicaca Basin. Interestingly, Giesso found that the 19 obsidian samples from the Akapana and Putuni were all transparent samples from the Chivay source, a pattern supported by the ceramics assemblage from those sites that was entirely local. This pattern suggests to Giesso (2003: 368) that a cultural strategy of "ideological purity" was taking place at that location at Tiwanaku due to exclusive use of those particular materials that are perceived as "local" to the Titicaca Basin. Giesso notes that obsidian is found in the form of standardized projectile points, type 4E according to Klink and Aldenderfer (2005), and that evidence of production is found in commoner residential contexts, whereas finished points are often associated with elite ritual contexts. This suggests to Giesso that labor contributions for some segment of the Tiwanaku heartland commoner population may have taken the form of projectile point manufacture.

A comparison of the circulation and use of Chivay obsidian with Querimita basalt strongly suggests to Giesso that the source areas, production, and transport of both exotic materials were controlled by the Tiwanaku state. If so, research at the Chivay source should reveal some evidence of Tiwanaku materials as it did at Querimita, and perhaps research will reveal state mandated standardization of production activities. However, if no evidence of Tiwanaku presence is found at the Chivay source, then this suggests that the relationship between raw material procurement, long distance transport, and state sponsored activities involved a more nuanced relationship between the state and the peripheral economy.

Burger et al. (2000) also analyzed obsidian from Tiwanaku and in their study, from a collection of 18 samples, all from surface contexts, the results were all of the Chivay type. The original Burger and Asaro (1977) study of Bolivian obsidian revealed that three samples purportly from the site of "Sora Sora" were from the Titicaca Basin source, now known as the Chivay source. At 554 km from the Colca Valley this is the furthest reported transport of Chivay type obsidian to date. However, some doubts have arisen as to the spatial origin of these three samples supposedly from the site of Sora Sora (Burger, Dec 2006, pers. comm.), and given the anomalously high transport distance, these three samples are suspect unless additional supporting data become available. The second furthest reported conveyance of Chivay obsidian is 325 km to the Late Formative contexts of Kallamarka and to Khonkho Wankane in Bolivia, as mentioned above.

Obsidian consumption at Wari

At the imperial capital of Wari, obsidian flakes were likewise found in domestic contexts and have been interpreted as "kitchen waste" as no specialized obsidian production sites were found at the site of Wari (Isbell, et al. 1991: 48; contra ;Stone 1983). In the funerary area of Wari known as the Cheqo Wasi sector the obsidian artifacts include a broken knife, 79 flakes (both utilized and unutilized), and two triangular fragments of obsidian with flat surfaces that may have been polished mirrors (Benavides 1991: 64).

At the Wari site of Pikillacta, in the valley of Cusco, evidence from obsidian flakes found in household compounds suggests that obsidian production took place at the household level. In the Haycuchina and Waska Waskan residential clusters, evidence of craft production consists of "waste flakes" of turquoise-colored stone, obsidian, and broken marine shell all collected from the surface (McEwan 1991: 99). Obsidian flakes were found in middens at Pikillacta along with both utilitarian and elite ceramics.


In Tiwanaku levels at Qillqatani there is a notable decline in the use of obsidian, and of the two specimens analyzed one was of the local low-quality Aconcahua type and the other was Chivay.XFigure 3-7Xreveals that there is a reduction in the percentage of obsidian in the assemblage from this level. In the "debris" category, obsidian falls below 10% for the first time since the Middle Archaic Period. Ceramics were clearly Tiwanaku influenced, but they were relatively poorly made (Aldenderfer in prep.). One may speculate that with the establishment of Tiwanaku colonies in Moquegua, the dominant caravan traffic patterns shifted to an east-west pattern in the Qillqatani area. It is also possible that if the Tiwanaku economic sphere was dominating the circulation of Chivay obsidian, then the material would probably have been conveyed more directly into the Basin and the western cordillera exchange would have been relatively diminished.

The Colca area during the Tiwanaku Period

Relatively little is known about the Colca valley in time periods prior to the Late Intermediate period. The political context of the Colca during this time is intriguing and somewhat ambiguous because the Colca valley lay on the frontier between Wari and Tiwanaku. External cultural influences in the Colca valley proper during the Tiwanaku times appear to have been entirely from Wari, yet Chivay obsidian is not found in Wari sites except in sites with a Tiwanaku component in Moquegua. Given the predominance of Chivay obsidian at Tiwanaku one might expect to encounter diagnostic Tiwanaku material somewhere close to the Chivay source. A number of scholars have commented on the surprising lack of Tiwanaku material in the Colca area given the obsidian distributions in the Tiwanaku heartland (Brooks 1998: 311-313, 454-459;Brooks, et al. 1997;Burger, et al. 1998: 211-212;Burger, et al. 2000: 340-342;).

The geographically closest Tiwanaku sites to the Colca in the department of Arequipa are found on the south-west edge of the city of Arequipa (85 km due south), with the best evidence coming from the site of Sonqonata (Cardona Rosas 2002: 78-87). The Arequipa highlands are largely unstudied, and the closest known Tiwanaku site to the Chivay source on the altiplano appears to be 147 km to the east in the department of Puno at the site of Maravillas just north of Juliaca (Stanish 2003: 189). Tiwanaku sites in this region are generally associated with lakeside agriculture or are found at cross-roads along major travel corridors (C. Stanish March 2006, pers. comm.), and to date no Tiwanaku sites are known in the largely pastoral periphery of the northwestern Lake Titicaca Basin in the direction of the Chivay source. Tiwanaku pottery has been found in Cuzco at Batan 'Urqo (Glowacki 1996: 245) and a Tiwanaku snuff tablet was found at La Real in the lower Majes valley of coastal Arequipa (García Márquez and Bustamante Montoro 1990: 28), but these appear to have been examples of trade goods along the Tiwanaku-Wari frontier.

In the main Colca valley close to the Chivay source, Wari-related ceramics have been excavated from a site just 4 km downstream from the town of Chivay and approximately 10 km from the Chivay obsidian source. A trench excavation exposed red-slipped wares from a domestic context with Tiwanaku Horizon dates in archaeological work associated with the William Denevan's Río Colca Abandoned Terrace Project at the site of Chijra (Malpass 1987: 61;Malpass and De la Vera Cruz 1986: 209, 216;Malpass and De la Vera Cruz 1990: 44-46, 57). Radiocarbon dates associated with these red-slipped wares came from a hearth in a house terrace and produced dates of 1140 ±80 (WIS-1713; AD680-1030) and 1290±90 (QL-4015; AD600-900) (Malpass 1987: 61). This ceramic style has recently been investigated in detail by Wernke as part of his elaboration of the Colca valley ceramic sequence (Wernke 2003: 466-477). Other Wari-influenced sites in the Colca Valley include a large, recently located site in on the north side of the Colca river named Charasuta close to the town of Lari (Doutriaux 2004: 212-223) and the site of Achachiwa near Cabanaconde (de la Vera Cruz 1987;de la Vera Cruz 1988;Doutriaux 2004: 202-207). Curiously, Chivay obsidian is not known to have circulated in the Wari sphere at all despite the proximity of these sites to the Chivay source.

In the main Colca Valley, indications of Wari ideological and stylistic influence are the strongest evidence for external links during the Tiwanaku Horizon. In addition to Wari influences at Chijra and Charasuta, the site of Achachiwa provides intriguing evidence of exotic obsidian intruding into the Colca valley. Achachiwa is a large site adjacent to the modern town of Cabanaconde that has a large Middle Horizon component that appears to be Wari influenced, as well as components belonging to a the local LIP occupation and an Inka occupation (de la Vera Cruz 1987;de la Vera Cruz 1988;Doutriaux 2004: 202-207). Brooks reports that she collected seven obsidian flakes for analysis from Achachiwa that were visually distinct from obsidian she had encountered elsewhere in the Colca (Brooks 1998: 447). Of these seven flakes, none were of Chivay type obsidian although the site is only 46 km downstream of the source. Her analysis showed that six of the flakes were from the Alca source (96 linear km away) and one was from the Quispisisa source (300 km away), a strongly asymmetrical pattern that is non-the-less consistent with Middle Horizon obsidian distributions in southern Peru.


Chivay obsidian distributions during the Tiwanaku period are somewhat of an enigma. As Tiwanaku persisted for longer than Wari one might expect Tiwanaku evidence at the Chivay source either preceding or simultaneous with Wari presence in the Colca valley. However, the nature of procurement and distribution of Chivay obsidian during the Tiwanaku period was such that diagnostic materials from the consumption zone do not appear at the Chivay source. There appears to have been considerable nuance in the relationships between state leaders, corporate integration, and ethnic local kin-based groups during the period of Tiwanaku hegemony. Groups living in the Tiwanaku peripheral areas were perhaps consistent with Dillehay's characterization as"…a patchwork of overlapping, geographically disparate, and apparently politically semi-autonomous core valleys, oases, and plateaus or foci of cultural development, each of which primarily exploited its own immediate peer area"(Dillehay 1993: 247). Along these lines,Stanish (2002: 188) notes that early states appear to have "selectively incorporated certain areas around the basin" rather than attempt comprehensive control. John Janusek summarizes a variety of evidence demonstrating that "Tiwanaku was an incorporativemore than it was a transformativestate, simultaneously employing multiple strategies of regional control and influence" (Janusek 2004: 162), along the lines of corporate political strategies described by Blanton et al.(1996). Accordingly, a direct correlation between the predominance of Chivay obsidian in the Tiwanaku economy and clear, material evidence of incorporation in the archaeological remains of communities in the Chivay source area should not be expected. The political affiliation of Colca valley communities was probably made more complex due to the presence of the Wari frontier during the Middle Horizon.

The asymmetrical, export-only exploitation is consistent with a pattern that has been observed at a number of prehispanic obsidian sources in the Andes. Obsidian sources generally have few diagnostic artifacts or architecture in association with quarrying. It was mentioned earlier that Formative distributions of Alca obsidian were strikingly asymmetric with export to long distance consumption sites but no corresponding diagnostics from those consumers back at the Alca source (Burger, et al. 2000: 314, 323). Similarly, the Quispisisa source has little diagnostic evidence the immediate vicinity linking the source to Wari or any other known group, despite the long history of use of that obsidian type in Wari sites (Burger and Glascock 2000;Burger and Glascock 2002). The asymmetrical nature of Chivay obsidian use during the time of Tiwanaku and Wari is, therefore, consistent with a pattern apparent at raw material sources elsewhere in the Andes where obsidian procurement does not involve reciprocation or discard of diagnostic artifacts from the consumption zone.

Late Intermediate Period (AD1100 - 1476)

Current evidence suggests that following the collapse of the Tiwanaku state, a prolonged drought occurred until AD1200 after which time Aymara chiefdoms emerged in the region referred to as Collasuyu by the Inka. A central question of this time period concerns the extent to which obsidian distributions can reveal whether Tiwanaku period interaction patterns persisted into the LIP in the forms assumed by economic organization and long distance exchange. While the LIP is known as the auca runaor the "time of strife" when fortified hilltop refuges " pukaras"were constructed in abundance in the Titicaca Basin and adjacent territories, the weapons used in these conflicts appear to have primarily been percussion weapons like slings and clubs, and not obsidian tipped-projectiles.

Chivay obsidian during the LIP

In a review that considers the LIP and Late Horizon together in one discussion, Burger et al.(2000)note the declining presence of obsidian in most LIP and LH contexts in the region. Many of the chemically provenienced obsidian samples from the Titicaca Basin come from surface contexts with either LH or LIP associations, and the pattern revealed in the Titicaca Basin from these samples is one of almost exclusive use of Chivay type material. One of the reasons for the strong presence of Chivay obsidian over Alca obsidian may have resulted from cultural affinities between Aymara groups(Browman 1994). The Colla in the North Titicaca Basin and the Collagua in the upper Colca area share many traits including the construction of pukaras(Wernke 2003: 262-263), mortuary features including chulpaburial structures and fiber-wrapped mummy encasings(de la Vega, et al. 2005;Wernke 2003: 225-234), and other commonalities.

In the northern Lake Titicaca Basin, Arkush(2005: 247)notes that two-thirds of the occupied pukaras throughout her survey area contained obsidian on the surface, and that chert flakes and blades are common. In one instance, at Calvario de Asillo (AS1) prepared obsidian cores and a concentration of flakes was encountered. Obsidian projectile points were small, triangular base-notched points (series 5), consistent with the Klink and Aldenderfer point typology for the Terminal Archaic and onwards(Arkush 2005: 709-711). Metal objects are common as well, with copper and copper-alloy pendants and tupus(long pins) frequently found at pukaras.

Six obsidian samples from the mesa-top pukaraand chulpaburial tower complex of Cutimbo were analyzed by Frye et al.(1998)and the samples were 100% from the Chivay source. Burger et al.(2000: 343-344)indicate that one obsidian sample from the renowned chulpacomplex of Sillustani was also from the Chivay source. Burger et al. observe that while a single Alca flake was excavated at Taraco from an LIP context, and another was found on the surface at the LIP or LH site of Kolkeparke near Ayaviri(Burger, et al. 2000: 343), the Titicaca Basin is otherwise entirely supplied from the Chivay source during the LIP.

At the rock shelter of Qillqatani (SectionX3.4.2X) a large percentage (15.4%) of the tools from LIP levels were made from obsidian, although the count is very low (n = 2) and the tools were extremely small. The single obsidian sample that was analyzed proved to be from the Chivay source.

In the south-western Lake Titicaca area, Hyslop(1976: 118-119)found that the Lupaqa sites he encountered in the course of his road system survey contained primarily basalt and quartzite flaked stone. More recent survey work in the Ilave and Huenque drainages(Klink and Aldenderfer 1996)found that high-quality cherts are abundant in the region, which suggests that the prominent use of basalt at the sites Hyslop encountered was by choice, not by necessity, and was perhaps a reflection of lakeshore agricultural activities.

Late Horizon (AD1476-1532)

The expansion of the Inka Empire during the Late Horizon resulted in a restructuring of the long distance movement of goods. The ability of the Inka state to transport stone is vividly demonstrated by the transport of hundreds of andesite ashlars weighing up to 700 kg apiece the 1600 km distance from Cusco to Saraguro, Ecuador (Ogburn 2004;Ogburn 2004). The ability to move people and goods over long distances was an integral part of state apparatus; however, it is evident from obsidian distributions that not all goods were more widely distributed during the Late Horizon.

It appears that if the Inka had a particular demand for a substance it could be acquired from over great distances. The Late Horizon was referred to as the "tin horizon" by Lechtman (1976) due to the lengths that the Inka would go to procure tin for copper production. In the Mantaro Valley, tin was not present in Late Intermediate Period coppers but it was found in all seven copper implements from Late Horizon levels where it made up, on average, 5% of the metal composition (Earle 2001: 311;Owen 2001: Tables 11.1-11.3). This tin is thought to have come from mines in Bolivia or southern Peru, and this exchange was facilitated by the Inka state (Lechtman 1976). Thus tin was transported from southern Peru, perhaps from the vicinity of obsidian rich lands of Arequipa, but apparently little obsidian was transported along those same Inka transportation routes.

Ethnohistoric accounts indicate that during the Late Horizon the control of natural resources sometimes occurred through restricted access to raw material sources in the Andes. Access to tunnels leading to particularly rich gold deposits at Inka gold mining operations were restricted at the tunnel mouths (Burger and Glascock 2002: 364;Sancho de La Hoz 1968 [1534]: ch. XVIII: 332). The deposits of other natural resources, such as obsidian, are distributed across the geological landscape such that directly controlling access would have been difficult or impossible.

Burger et al.(2000: 343-347)report very little Chivay obsidian circulating in contexts that are definitely Late Horizon in date. Small quantities of obsidian were found in Cusco, and these turned out to be entirely of the Alca type except for one notable case of Chivay material at Machu Picchu. In one of the only cases of Chivay obsidian use in the department of Cusco since the Late Formative, several unmodified nodules of Chivay obsidian were identified from a collection of small obsidian pebbles excavated by Hiram Bingham in 1912 at the gateway to Machu Picchu. Burger et al. (2000) note the significance of the unmodified state of these small obsidian pebbles, as the small size of these nodules suggest that it was not the tool-making potential, but rather the natural glass itself that was "suitable in its apparently natural state as an offering or sacred object comparable to quartz crystals"(Burger, et al. 2000: 347;Rowe 1946: 297). This view is consistent with observations about the possible significance of the 'essence' of obsidian as natural glass in Andean cosmology that will be explored in more detailed below.

Discussion and review of "Late Prehispanic" obsidian

Obsidian distributions during the Late Prehispanic period reflect the dramatic changes in economy and ideology imposed by powerful states in the region. In the Tiwanaku and Wari spheres of influence obsidian circulation achieved its largest known extent. Particularly intriguing are the examples of non-local obsidian consumption in close proximity to rival obsidian sources, a phenomenon that occurs twice during the Middle Horizon. Three samples of Chivay obsidian were found close to the Sora Sora obsidian source in western Bolivia in a possibly Tiwanaku context (Burger, et al. 2000: 340), and Quispisisa and Alca obsidian were found at the Wari-influenced site of Achachiwa, only 46 km downstream of the Chivay source, in the lower Colca valley (Brooks 1998: 447). During the LIP this pattern of high mobility and sprawling interregional contact is reversed as regional circulation of obsidian is curtailed and the Chivay regional pattern returns to a distribution confined to the Colca region and the North Titicaca Basin. Finally, during the Inka period, obsidian appears to have had a largely diminished significance, particularly in contrast to other materials that were circulated prodigious distances by the Inka. This discussion of the spatial distributions of prehispanic obsidian will be complemented by a review of the uses and forms that obsidian artifacts assume in the archaeological and ethnohistoric record.

3.6. Obsidian Use in the South-Central Andes

Obsidian was knapped into relatively few artifact forms in the south-central Andes. The most common formal stone implement was a bifacially flaked projectile point, but other bifacial tools such as knives and scrapers were also commonly produced throughout the consumption region. The other major technical class for obsidian artifacts were simple flakes. As a sharp, but fragile, cutting implement, a freshly struck obsidian flake was potentially useful for butchery and wool shearing purposes. Dransart (2002: 108-109) reports that in rural communities modern wool shearing is accomplished either with the lid of a tin can that has been folded over so as not to cut the user's hand, or with a broken piece of bottle glass. Contemporary herders in the Colca region report that obsidian flakes, and sometimes broken glass vessels, are used for castrating animals because it was explained that as non-metal tools do not oxidize, they are less likely to introduce infection into the animal (T. Valdevia 2003, pers. comm.).

A distinctive Andean method of camelid slaughter can be accomplished with a small, sharp flake is described by George Miller (1979: 27-36). The ch'illamethod of slaughter consists of laying the animal down, cutting a small incision near the sternum with a small flake of stone, and reaching in and manually breaking the ascending aorta where it leaves the heart. Ethnoarchaeological studies have shown that simple flakes are often used in butchering and shearing, and therefore a prehistoric association between pastoral facilities and lithic flakes, both utilized and unutilized, seems probable.

3.6.1. Variability in Andean obsidian use

In many regions of the world prehistoric artifacts made from obsidian can be generally classified by whether the principal function is for display or for some utility more directly related to subsistence. In Mesoamerica and the ancient Near East, both areas with complex societies and elaborate stone tool production, obsidian was used to make bowls, vases, eccentrics, seal stamps, statuettes, and tables, as well as items of personal decoration such as labrets, ear spools, necklaces, and pendants (Burger, et al. 1994: 246). Craftspeople also developed efficient, high utility obsidian technologies as well, such as prismatic blades, a form that allows archaeologists to quantify cutting edge to edge-length, error rates, and other efficiency measures.

In the south-central Andes the diversity of artifact forms made from obsidian is relatively low, and it is difficult to differentiate display from utilitarian applications. For example, obsidian projectile points are both sharp and highly visible, suggesting that the points had a display function that underscored their utility as a weapon and as a cutting tool.

Hafting in the Andes

Archaeological evidence shows that bifacially flaked obsidian implements were hafted in a variety of ways. The majority of the hafting evidence from the Andes comes from coastal sites due to superior preservation. Projectile points were hafted to spears, spear thrower darts, and arrow shafts. Obsidian bifacial tools were also hafted to wood or bone handles for use as knives. Hafting materials varied regionally, but hafting was often accomplished using gum or resin, and hafts supported with cotton string have been found in some coastal sites (Carmichael, et al. 1998: 79).

Evidence of use of obsidian

The artifact types predominantly manufactured from obsidian include projectile points and tools for cutting and shearing tasks. While simple flakes are wide-spread and were probably used abundantly for butchering, scraping, and shearing purposes, the utility of obsidian flakes is frequently discounted when flakes are relegated to the "debitage" or "debris" class. In Andean studies bifacially-flaked instruments are the most commonly analyzed obsidian artifact class in archaeological reports from sites in the Andes.

Obsidian artifacts are sometimes found in association with iconographic representations of dark colored artifacts that are similar in appearance to the very obsidian artifacts found in that context (Burger and Asaro 1977: 15). Such is the case with black-tipped darts and knives depicted on Ocucaje 8 through Nasca 6 ceramics and textiles, and obsidian artifacts found in tombs from those contexts. Building on the discussion in Burger and Asaro (1977: 13-18), examples of obsidian artifacts from the south-central Andes follow.






Weapon (probable dart point), conflict


Looted tomb at Hacienda Mosojcancha, Huancavelica.

A point made from obsidian was found embedded deeply in a human lumbar. (Figure C-10).

(Ravines 1967)

Weapon, with spear throwers


Grave 16 at Asia,
(central coast of Peru).

Preceramic context.

Found in association with spear throwers

(Engel 1963: 56;Uhle 1909)

Weapon, hafted


Tombs at Hacienda Ocucaje, Epoch 10.
Early Horizon context.

Points hafted with gum and, in one case, cotton thread to wooden foreshafts.

(Burger and Asaro 1977: 14)



Carhua (south coast , Peru)

Point penetrating through arm muscle near humerus (Figure C-1).

(Engel 1966: 212)




Paracas Necropolis

Well-preserved harpoon (Figure C-9).

(Engel 1966: 180c)



Hafted projectile


Nasca Phase B1 and B2 diagnostic attribute

Phase B has "Atlatl darts (arrows) in series as ornaments" (Figure C-6).

(Carmichael, et al. 1998: 151)

Weapon, poison


Eastern Lake Titicaca

Obsidian is among point types dipped in strong poison from herbs, perhaps curare.

(Cobo 1990 [1653]: 216-217)

Weapon, bow and arrow

Point depiction

Tiwanaku pottery

Archers with bow and black

tipped arrows depicted on a Tiwanaku q'ero(Figure C-4).

(Bennett 1934: 426-459;Posnansky 1957: plate XXa)

Weapon, hunting

Point depiction

Nasca B vessel

Depiction of darts sailing towards a group of camelids .

(Burger and Asaro 1977: 16)

Tool, Ritual

Knife depiction

Nasca B pottery

Black knives associated with taking of trophy heads.

(Burger and Asaro 1977: 15)

Tool, Ritual

Knife depiction

Nasca textiles, Epoch 1

of EIP

Black knives associated with taking of trophy heads.

(D'Harcourt 1962: 110, 112)

Tool, Ritual

Knife, hafted

Early Nasca

Bifacial knife hafted to painted dolphin palate (Figure C-5).

(Disselhoff 1972: 277)

Decorative, Ritual

Mirror fragment

Huancayo, Middle

Horizon 2 context

Fragment of obsidian mirror ground and polished to .4 cm thickness.

(Browman 1970: 86)

Decorative, Ritual


Huarmey, Wari

Mirror mounded in carved wooden hand (Figure C-8).

(Lavalle and Perú 1990: 185)


Obsidian knives with blood-stains.

Cerro Colorado, Paracas

Part of medical kit that also contained a chachalote (sperm whale) tooth knife, bandages, balls of cotton, and thread.

(Tello 1929: 55)


Chillisaa kala, Aymara for "black flint"

Titicaca Basin

Speculation about tools used for trephination.

(Bandelier 1904;Marino and Gonzales-Portillo 2000)

Table 3-10. Examples of obsidian use in the south-centralAndes (part 1).







Material used in folk cures

Canchis, Cuzco; and elsewhere

Modern use in folk cures, the stone was believed to have curative powers.

(Burger and Asaro 1977: 17;Cobo 1990 [1653])


"knives of crystalline stone"

Titicaca Basin ( ?)

Abdominal surgery by "sorcerers"

(Cobo 1990 [1653])

Animal castration

Flakes, unmodified or retouched


"We use sharp pieces of obsidian or glass to castrate herd animals it doesn't cause infection like rusted metal knives."

T. Valdevia 2003, Pers. Comm. (my translation).


Flakes, unmodified or retouched


"Aboriginal shearing required special implements, perhaps obsidian knives." Some modern pastoralists use broken glass and tin lids for shearing.

(Dransart 2002: 108-109;Gilmore 1950: 446)

Table 3-11. Examples of obsidian use in the south-centralAndes (part 2).

Many of these examples are shown in Appendix C. of this volume. While the diversity of artifact forms was relatively low, it is evident that the visual and fracture properties of obsidian were relevant to the tools that were used from this material. Further north in the Andes, in Ecuador, a greater percentage of obsidian artifacts seems to have filled a primarily decorative role, including abead, an ear spool, and three polished mirrors(Burger, et al. 1994: 246).

The evidence of use lithics may also take the form of cutting and scraping marks on faunal remains, but the lithic material type can rarely be established by these means. The continued use of glass and obsidian in modern contexts for shearing, butchery, and castration suggests that the prehispanic metals such as copper and bronze did not displace obsidian and other lithic materials for utilitarian tasks. Prehispanic metals were used largely for display, although metals were used in some weaponry, such as mace-heads (Lechtman 1984).

Obsidian in warfare

The evidence for obsidian use in conflict in the Andes comes from a variety of archaeological and ethnohistorical sources, but it is primarily in the form of indirect evidence. The data reveal a great expansion in the production of small obsidian projectile points after the onset of a food producing economy. The majority of Late Horizon weapons that the Spanish faced during their invasion appear to have been bola stones and percussion weapons like maces and slings(Cahlander, et al. 1980;Korfmann 1973), as well as padded armor.

Archaeologists working in the south-central Andes vary historically in their assessment of the use of projectiles in the highland region(Giesso 2000: 43). Bennett(1946: 23)asserts that the use of bow and arrow were not important on the altiplano, Metraux(1946: 244-245)says that spear throwers were in use, while Kidder(1956: 138)indicates that evidence from projectiles show that arrows were widely used at Tiwanaku. Of the projectile points analyzed by Giesso from Tiwanaku, 19% were made from obsidian, with the highest concentration of points coming from excavations in the civic/ceremonial core at the Akapana East, K'karaña, and Mollo Kontu mounds(Giesso 2000: 228-238).

In a dramatic example of obsidian use as a weapon, one archaeological findError! Reference source not found.shows a probable spear or dart point penetrated the victim's abdomen from the left anterior abdomen and lodged on the anterior side of the lumbar vertebra.

Evidence of use of the bow and arrow in warfare

One of the strongest patterns in obsidian distributions in the south-central Andes is the sudden onset of obsidian use with series 5 projectile points in the Terminal Archaic, after 3300 cal BCE In defining the small type 5D projectile points Klink and Aldenderfer(2005: 54)suggest that the widespread adoption of these point types may reflect the use of bow and arrow technology, as these points fallwithin the size range described by Shott (1997: Table 2) as associated with arrow points (although with substantial overlap with the size of the smaller dart points).


Figure 3-14. Type 5d projectile points from a Terminal Archaic level at Asana and Early Formative levels at Qillqatani, from illustration in Klink and Aldenderfer (2005: 49).

Additional metrics for differentiating arrowheads from dart and spear points have been suggested by Thomas (1978: 470) and Patterson (1985). With Mesoamerican projectile points, Aoyama (2005: 297) has found a very significant correlation between evidence from microwear analysis and projectile dimensions in his effort to differentiate arrowheads from dart and spear points. The relationship between the mass of the item and the velocity and distance of the projectile are discussed in detail by Hughes (1998). She notes that the innovation of fletching allowed for balanced projectiles that had smaller projectile tips and lighter shaft materials, which in turn permitted greater distance and velocity in weapons systems (Hughes 1998). On a related point, the notched base in all type 5 projectile styles reflects a change in hafting technology. Greater impact loads can be absorbed by mounting projectiles tips with notched bases into slotted hafts (Hughes 1998: 367;Van Buren 1974).

Obsidian has a number of attributes that make it a particularly effective material for projectile point tips, most notably are the predictableconchoidal fracture qualities of obsidian that allow it to be pressure flaked into very small projectile points. The sharpness of freshly knapped obsidian margins is unrivalled, andethnographicallyobsidian is renowned as a brittle material that can fracture on impact, causing the fragments to produce greater bleeding in the victim of an obsidian projectile (Ellis 1997: 47-53). In materials science, obsidian has extremely low compression strength because it has no crystalline structure (Hughes 1998: 372). Due to the brittleness or lack of compression strength, missed shots can result in many broken obsidian projectile tips, and perhaps for this reason obsidian was not the predominant material for projectile point production until after the domestication of camelids when hunting was decreasingly the primary source of meat (as inferred from the ratio of camelid to deer bone in excavated assemblages).

Increased velocity and distance is possible with the use of bow and arrow, and obsidian has greater penetrating power, although lower durability, than other material types. Do these changes suggest a greater use of projectiles for warfare? Greater use of small, fletched projectiles might be expected in contexts where the individual needs (1) increased distance from the target (2) increased velocity, perhaps due to change in prey or to the use of padded armor. The use of obsidian may increase in contexts where one needs (1) a material that can be knapped into small, light projectiles as discussed above, (2) increased penetration with sharper material, (3) decreased concern for durability because in warfare the weapon will perhaps be retrieved by the opponent.

Although empirical evidence for warfare is scarce in the Terminal Archaic in the highlands, a study of 144 Chinchorro individuals from Terminal Archaic contexts in extreme northern Chile found that one third of the adults suffered from anterior cranial fractures that probably resulted from interpersonal violence, and men were three times more likely than women to have these wounds (Standen and Arriaza 2000). These wounds appear to have been caused by percussion weapons like slings, but the coastal evidence for interpersonal violence in this time period is strong. Additional support for the introduction of the bow and arrow in the Terminal Archaic come from the region among Chinchorro burials in coastal Northern Chile. Bows among the Chinchorro grave goods date to circa 3700 - 1100 cal. BCE, or the Terminal Archaic and Early Formative (Bittmann and Munizaga 1979;Rivera 1991;Standen 2003).

Possible use of poisons on projectile tips

A potential explanation of widespread adoption of the small type 5D projectile points is the greater availability of poisons that reduced the need for heavy, destructive projectiles.South American poison arrows are usually quite small and such arrows are frequently tipped only with a sharpened wooden point. According to Ellis(1997: 55),virtually all ethnographic examples of arrow use include some kind of poison applied to the arrow in order to have either a toxic or a septic effect on the victim. Ellis observes that due to the great variety of substances used to create such toxins, in many regions of study these substances would probably contaminate any chemical attempt to use residue analysis to differentiate the types of poisons used, or even the prey that was hunted, with a particular used projectile point.

A variety ofhighly effective poisons are applied to the tips of projectiles by hunters in the Amazon Basin today(Heath and Chiara 1977).In the prehispanic Andean highlands, trade contacts with the Amazonian lowlands to the east may have made available poison concoctionsfor application to projectiles, most notoriously the fast acting paralysis alkaloid curareprepared from the vine Chondrodendron tomentosum(Casarett, et al. 1996).

Bernabé Cobo (1990 [1653]: 216-217) discusses the use of bow and arrow with poisons by "expert marksmen" in his chapter on warfare. Hedescribes the widespread and expert use of the sling for warfare, which is consistent with reports elsewhere on the use of slings, but then he states that bow and arrow were more significant in warfare.

The most widespread weapon of all the Indies, not only in war but also in the hunt, was the bow and arrow. Their bows were made as tall and even taller than a man, and some of them were eight or ten palms long, of a certain black palm called chontawhose wood is very heavy and tough; the cord was made of animal tendons, cabuya, or some other strong material; the arrows, of a light material such as rushes, reeds, or cane, or other sticks just as light, with the tip and point of chontaor some other tough, barbed wood, bone, or animal tooth, obsidian point, or fish spine.
Many used poisoned arrows, their points anointed with a strong poison; but, among the nations of this realm, only the Chunchosused this poisonous herb on their arrows, and it was not a simple herb, but a mixture of various poisonous herbs and vermin; and it was so effective and deadly that anyone hit by one of those poisoned arrows who shed blood, even though it might be no more than the blood resulting from the prick of a needle, died raving and making frightful grimaces(Cobo 1990 [1653]: 216-217).

The Chunchos ethnic group is described as living in the "forests east of Lake Titicaca on the border of the Inca Empire".It is possible that Amazonian poisons became available in altiplano during the Terminal Archaic due to expanding exchange networks, together with Amazonian hallucinogenics and other lowland products, dramatically altering the efficacy of arrows. As a sharp but lightweight weapon, obsidian tipped arrows would have represented an effective poison delivery system to animal and human victims alike.

While there is no direct evidence for the use of Amazonian poisons on the altiplano, the dramatic change in projectile technology with the type 5D type, co-eval with expanding exchange networks in the region, suggest that a new technology for weapons systems, such as poisons and new shaft and fletching materials, may have influenced the design of projectiles in this time.The transitional economic context of the Terminal Archaic involved many changes, including shifts in both food production and interregional exchange, and the technology of lithic production show significant alterations that correspond to this period.

Multiethnic access to geological sources in the Andes

Salt procurement in the Amazon basin on the eastern flanks of the Andes provides an example of multiethnic access to a raw material source. During annual voyages to the Chanchamayo salt quarry ten days outside of their territory, the Asháninka of the Gran Pajonal (Ucayali, Peru) combine salt procurement with exchange with neighboring groups who share access to the source(Varese 2002: 33-35). Arturo Wertheman, a missionary who traveled through the region in 1876, explains

Throughout the year small bands of Asháninka traders traveled [the Gran Pajonal] paths to obtain salt, carrying with them tunics or ceramics to exchange for other items and for hospitality. With them traveled their traditions, their hopes, and the information of interest to their society. The Pajonal, the vast center of Campa territory not yet invaded by whites, appears to have been the center of culture and tradition through which the Indians journeyed, like a constant flow of life through their very society(Varese 2002: 120)

Other ethnographic examples of multiethnic access to salt quarries are mentioned in the ethnographic literature. Oberem(1985 [1974]: 353-354)describes access by both the Quijo and the Canelo people to a large rock salt quarry located on the Huallaga River, a tributary to the Amazon lying further north on the border of Peru and Ecuador.

In the Cotahuasi valley of Arequipa, to the north of the Colca valley, the rock salt mine of Warwa [Huarhua] has been exploited since Archaic times(Jennings 2002: 217-218, 247-251, 564-566). Access to the Warwa quarry, which lies near the border of the departments of Arequipa, Ayacucho, and Apurímac, is described by Concha Contreras(1975: 74-76)as including caravan drivers from all three of those neighboring departments.

El primer viaje lo hacen, mayormente, en el mes de abril. En esta época cientos de pastores se concentran en esta mina. El camino es estrecho y accidentado hasta llegar a la misma bocamina…Desde el fondo de la mina los pastores cargan a la espalda la cantidad de sal que necesitan llevar, de tal manera que hacen muchos viajes al socavón de la mina. En todo este tramo tardan 4 días, porque después de la mina, siguen cargando en la espalda hasta una distancia de aproximadamente cinco kilómetros, donde quedaron las llamas pastando, puesto que hasta la misma mina no pueden entrar juntamente con sus llamas(Concha Contreras 1975: 74).

There appear to be a number of protocols associated with salt acquisition at Warwa. Notably, the multiethnic visits by caravanners from various departments coincide in April despite the tight working quarters at the salt mine. Furthermore, there seems to be concern for impacts in the much-visited the mine area itself. The cargo animals are actually grazing a distance from the quarry and humans are obliged to carry the loads to these areas instead of attempting to load the animals close to the mine. These cases of multiethnic raw material access provide examples of the social and institutional nature of access to unique geological sources, and the focalized attention that these source receive from surrounding ethnic groups. The analogy with regional acquisition and diffusive transport of the raw material is perhaps the closest modern analogy to the nature of prehispanic obsidian quarrying that remains in the Andes.

3.6.2. Symbolic significance of obsidian

Inferring the symbolic significance of obsidian in the prehispanic Andes involves approaching the topic from a several lines of evidence because there are few direct indicators of valuation and symbolic meaning of obsidian in the Andes. Below, these issues are raised as a series of questions that may, or may not, be answered with archaeological evidence from the region:

(1) How did the symbolic importance of obsidian change through prehistory?

(2) Do changes in use of obsidian reflect local geographic availability?

(3) Did the visual properties of obsidian resonate with Andean aesthetic traditions?

(4) Was there social significance in the visible differences between major obsidian types?

Part of the difficulty in inferring the symbolic properties of obsidian is the result of a relative reduction in the importance of obsidian during the Late Horizon and Colonial period, to judge from ethnohistoric accounts. Ethnohistoric sources are a prime source of information, although through Spanish eyes, of cultural significance and symbolic value in the prehispanic Andean world. Despite the apparent decline of obsidian use in the late Prehispanic and Colonial periods, archaeological distributions suggest that obsidian had substantial cultural and ceremonial associations in particular contexts during in Prehispanic times.

How did the social and symbolic importance of obsidian change through time and space?

Obsidian flakes and bifacial tools were sometimes included as grave goods from the Terminal Archaic (3300 cal BCE) and onwards. Ethnohistoric accounts and archaeological evidence indicate that obsidian "knives" were employed in ritual practice as well as in medical procedures throughout the region, although it is not evident if these were bifacially-flaked tools or simply freshly struck flakes. In addition, concentrations of flakes of quartz as well as obsidian from a variety of non-local sources were found in the Tiwanaku ceremonial mound of Mollo Kontu in a regular pattern that was interpreted by Couture (2003) to indicate deliberate inclusion in clay fill rather than accidental redeposition.

Based on analogy from other goods, it is possible that in areas where obsidian is used in ceremonial contexts a finely-made obsidian implement may have been less likely to have been exchanged as an alienable product traded in barter. In some historical contexts, a ritual item may become a singular item and acquire a distinctive 'genealogy' related to its social history that precludes it from being bartered and circulated in equivalence for something ordinary such as a sack of potatoes (Hodder 1982;Kopytoff 1986).

However, on the whole, obsidian appears to be one of a number of products that escapes easy classification. On one hand, obsidian circulated in established networks controlled by llama caravan drivers who were responsible for the distribution of a variety of goods on the altiplano. These caravans appear to have been organized by households and derived communities of relatively humble means, to judge from other aspects of their material lives excavated at sites like Qillqatani. On the other hand, as a luminescent material with irregular spatial availability, obsidian had properties that qualify it as a "prestige good" in many contexts (Hayden 1998). In terms of rituals organized and performed on a local level, obsidian was probably one of a number of materials that were traditional but non-the-less slightly rare like ochre and shell. Craig (2005: 683-693) explores the evidence for the symbolic importance of discrete color groups in ritual items like ochre and obsidian in the Titicaca Basin Archaic and Formative site of Jiskairumoko by examining the patterning of obsidian and ochre in comparison to what were common-place, functional applications of these goods.

From the perspective of early aggrandizers in the Terminal Archaic and Formative Lake Titicaca Basin, obsidian probably represented a somewhat elusive product for labor investment and therefore was not a principal substance used in elite strategies by the time of the Middle and Late Formative. From the perspective of prestige technologies serving the agenda of aggrandizers by locking up surplus labor (Hayden 1998), the issue might be considered in terms of the following factors:

(1) Obsidian was relatively easy to acquire, perhaps too easily acquired, for the people living in the highlands of Arequipa; and it was not concentrated in specific, controllable points accessed only through mine shafts.

(2) Obsidian has functional properties that assured its continued circulation for the production of cutting and piercing implements, making it insufficiently rare to have served as a "preciosity" (Clark and Blake 1994;Goldstein 2000).

(3) While craft specialization was fostered in more closely-controllable contexts in regional centers during the Formative, obsidian sources were perhaps sufficiently distant to have escaped this kind of specialized craft production. In contrast, consider the various non-utilitarian obsidian products made by craftspeople in Mesoamerica, discussed above.

(4) As obsidian was primarily made into projectile point tips, it was perhaps associated with hunting, or the threat of violence. In some regional traditions, such as in Nasca iconography, obsidian appears to have been associated with trophy head taking. However, in the south-central Andean highlands there is no evidence of a direct association between violence, emerging leadership, and control of surpluses along the lines of Hayden's (1995) despotic leader model.

Obsidian products, particularly finely-made bifacial tools, were perhaps one of a series of items that served to differentiate status-seeking individuals in early transegalitarian contexts, but with later crafts investment obsidian appears to have been assigned a relatively specific role for projectile point production and for cutting implements.

Do changes in use of obsidian reflect local geographic availability?

One way to examine the social and symbolic significance of obsidian is to examine the changing use of the material as availability declines with distance from the geological source. To judge from the distant consumption zone, obsidian was used in both mundane contexts and also in ritual or ceremonial contexts. At Tiwanaku, obsidian was found dispersed into the fill of ceremonial mounds (Couture 2003). It is also sometimes found in Titicaca Basin burials as of the Terminal Archaic. As reported by Craig (2005: 570-574, 679-682), at Jiskairumoko and at other sites in the Ilave valley, obsidian has been found in burials dating to 3300 cal BCE and later along with other non-local goods including lapis lazuli (sodalite) beads, gold discs, and gold beads, as well as ritual items like a camelid effigy made from bone. If the associations of obsidian with ritual power were related to its non-local origins, one might expect this pattern to have been weaker in closer proximity to the Chivay source. Items from ritual contexts, such as grave goods from the Colca Valley, and in close proximity to the Chivay obsidian source, may provide data to test this hypothesis.

Do the unusual visual qualities of obsidian resonate with Andean traditions?

Andean traditions place a priority on visual attributes, and links between the visual purity of a material and its essence have been widely noted in some segments of the Andean literature. As a natural glass, the aqueous properties of obsidian cause the material to reflect light which may display the workmanship of obsidian artifacts, as well as the potential sharpness of obsidian tools. Obsidian used for prehistoric tool production was often a homogeneous glass that was visibly consistent, pure in color, and sometimes transparent or banded. The importance of visual qualities of metals in the Andes has received greater attention:

The social arena in which metallurgy received its greatest stimulus in the Andes was the arena dominated by status and political display. An underlying cultural value system that appears to have strongly influenced the visual manifestation of status and power was a color symbolism oriented around the colors of silver and of gold. The most innovative and interesting aspects of Andean metallurgy arose from attempts by Andean metalsmiths to produce metallic gold and metallic silver surfaces on metal objects that were made of neither metal (Lechtman 1984: 15).

In Andean metallurgy, the appearance of consistency in color, reflectivity and material was prioritized because visual characteristics conveyed information about the inherent essence and animation of the object (Lechtman 1984: 33-36).

There is little direct support in Andean archaeological or ethnohistorical sources for inference regarding how obsidian was perceived, but it would be consistent for obsidian, a stone with the appearance of watery luster, to be associated with ceremonial power and ritual sacrifice given the well-demonstrated importance of stone and water in Andean cosmology.

Were the visual differences between obsidian types in the Andes important?

Visible differences between obsidian chemical types in the region are principally in terms of nodule size, fracture characteristics, glass color, and cortex. While nodule size and fracture characteristics are believed to have been important determining factors in explaining which obsidian types were circulated widely in prehistory (Burger, et al. 2000: 348), glass color is more of visual aesthetic issue. The obsidian sources in the south-central Andes are predominantly black or grey. Obsidian from the Chivay source is often a transparent grey and banded, and Quispisisa obsidian sometimes has a red coloration, although Burger et al. (2000: 314) state that the Quispisisa type is visually indistinguishable from Alca obsidian. As with many goods with discrete places of origin, these visual differences communicate information about the spatial origin of the stone that would have visually linked the material with regions and socio-political groups to knowledgeable viewers. To Brooks (1998: 452;1997) the transparency of Chivay obsidian was a reason for its wide circulation in prehistory. Others have commented on the transparency of the material including Giesso (2000;2003), and Burger et al (2000: 296). Giesso (2003: 368) observes that archaeological and ethnographic evidence from the Andes indicate that "transparent elements were viewed as mediators between different cosmological worlds". Further study may permit evaluation of observed patterns in the contexts of obsidian use that are linked to color.


The social and ritual significance of obsidian in particular prehispanic Andean contexts appears to have varied across time and space, and with further research in the region into evidence of production and consumption these differences will be better understood. Archaeologists have established that the visual attributes of particular materials like metal were important in the late Prehispanic Andes. If this mode of interpretation may be extended, high quality obsidian shares some visual characteristics with metal such as shininess and an appearance of material purity. Furthermore, obsidian from particular areas was often visually distinct and this may have conveyed information in regions such as Moquegua where a variety of obsidian types have been encountered. Obsidian was irregularly available across the landscape, and the mere possession of this highly visible material in obsidian-poor regions had possible social significance because it suggested that the holder participated in long distance exchange networks or had alliances with groups in obsidian-rich areas.

3.7. Models for the Procurement and Circulation of Chivay Obsidian in Prehistory

A number of models have been presented for regional interaction and exchange in the south-central Andes (Bandy 2005;Browman 1974;Browman 1980;Browman 1981;Burger and Asaro 1978: 68-70;Dillehay and Nuñez 1988;Nuñez and Dillehay 1995 [1979];Stanish 2003). This study evaluates a selection of these models at the Chivay obsidian source in the highlands of Arequipa. Renfrew explored various configurations for interaction and his "exchange modes" are reviewed above in Chapter 2 (Figure 2-2). These models will be discussed with respect to activity at the Chivay source area and material expectations for what may result from each model in the vicinity of the obsidian source. It should be noted that due to the extremely thin cortex on many Chivay obsidian nodules, decortication is not a consistently useful measure of reduction level or labor investment, but none-the-less the Upper Colca lab analysis sought to measure percentage of remaining cortex on flaked stone artifacts. When the geological cortex is of the extremely thin variety, it is sometimes left on the face of tools and it does not pose an obstacle to knapping. As cortical flakes from obsidian with a thin cortex are often smooth, and can be equally sharp, one should therefore not assume that nodules will be decorticated in the quarry area.

A number of modes of procurement are explored here, but these acquisition and exchange modes are not mutually exclusive either in time or space. That is, a variety of processes were likely to have been occurring simultaneously. For example, a independent caravan have transported an obsidian nodule to a site in the Ilave river valley, and then obsidian nodule may have been transmitted through down-the-line trade from the Ilave area to the Tiwanaku area. These models, therefore, will focus specifically on procurement and initial transport from the Chivay source because that is where more direct material correlates for these different models can be expected.

3.7.1. Direct acquisition Model


Direct acquisition by the end user entails high mobility and multiethnic access to the Chivay source. In this model, those traveling to the source would procure only sufficient obsidian for their household or community needs and no more. The evidence of procurement at the source would be the direct impacts of communities of consumers where obsidian was perhaps circulated in a context of generalized reciprocity but specifically not exchanged against goods in truequebarter, as that is a different type of procurement. Thus, this category consists of direct, personal visits by the immediate consuming household throughout prehistory.


Exchange Mode


Material Correlates

Direct Access


Personal (household) procurement through visits to the source. No exchange.

High variability in procurement, advanced reduction, low density production. Possible presence of discarded non-local low-value materials, and non-local temporary architecture.


Reciprocity, including barter

Local procurement supplying regional demand through exchange. Barter relationships, delayed reciprocity, and other arrangements between neighbors may have been reciprocated with obsidian.

Low variability in procurement, medium reduction, low to medium density of procurement and reduction. Local debris and architecture at procurement area. Non-local portable objects brought into region possible from reciprocation.

Caravans, Independent

Freelance or non-market central places, reciprocity and barter.

Household organized caravans transporting a variety of goods near the Colca. Procurement and transport of obsidian as one of these goods.

Production associated with pastoral facilities, dense processing activity that is moderately systematic. Small triangular proj. point production, possible evidence from variety of non-local goods and architecture.

Caravans, Administered

Emissary, colonial enclave, entrepôt. Redistribution, barter.

Elite commissioned and delegated caravans transporting goods near the Colca. Obsidian procurement for elite use at regional center.

Production with pastoral facilities. Systematic reduction by part time specialists. Possibly non-local elite-related material. Possible evidence of control of source.

Table 3-12. Models of procurement and exchange for Chivay obsidian. Compare terms with those used in Figure 2-2 and Figure 3-3.

Based on low population densities during the Early Holocene it is assumed that the earliest regional consumers of Chivay obsidian in the south-central Andes, the residents of Asana circa 9400 BP, acquired obsidian directly. Direct household acquisition of resources, such as salt procurement, persists to this day in a few places in the Andean highlands. It is possible that the multi-ethnic nature of access to Andean salt mines may serve as a model for procurement that occurred at Chivay during much of prehistory.

As it is generally the altiplano pastoralists that possess the major means of transport, llama pack animals, and direct household procurement by such groups was perhaps common for many types of goods. For example, herders from the community of Paratía (Flores Ochoa 1968: 87-109), to the north-west of Juliaca in the department of Puno, made regular trips to the Colca valley to acquire agricultural goods. If Paratíans traveled directly, it is likely that they used the Quebrada Escalera route passing to the north of Nevado Huarancante which would pass only a few kilometers from the Maymeja area of the Chivay source. On the return voyage from the Colca, if llamas were not overly burdened, a special stop could have been made to acquire a few nodules of obsidian. Similarly, the household-organized llama caravans described by Nielsen (2001) could well have obtained lithic raw material for household consumption if their travel route passed adjacent to a scarce raw material source on a return journey.

This mode of transport requires that consuming households had the social and physical means to travel to the obsidian source. The ability to partake in regional transport, even in a logistical fashion described by Flores Ochoa and by Nielsen, required strong animals, the food security to undergo a risky journey, peaceful conditions and personal security on the isolated travel routes, probable social relationships or contracts with communities encountered along the way, and knowledge of the extraction source area. Conditions of multiethnic access were likely variable in prehistory and knowledge of travel routes and sources of raw material were best obtained through cordial relations with local groups. In short, this means of obsidian distribution requires relatively cosmopolitan travelers with the resources to personally visit the sources of goods needed by the household. This mode of transport is relatively inefficient because individual households or communities sharing resources, by definition, have to personally acquire and produce obsidian in this model of obsidian procurement.

Material Expectations

In the immediate vicinity of the obsidian source archaeological evidence of the direct acquisition mode would produce high variability in procurement, but relatively small quantities in production because consumption would be limited to the households of those visiting the source. Greater amounts of advanced reduction evidence, including bifacial thinning flakes and tools, broken and discarded during manufacture, will be in evidence at the source area. This type of procurement would have the greatest chance of resulting in diagnostic projectile points in the area as advanced reduction and potential discard in the area of the source is expected. When temporal control is available, primarily from datable organic material in excavated contexts, direct acquisition should result in irregular visits to the source area based on household need. In excavation units cultural material will likely be low-density, as soil and perhaps ash from adjacent active volcanoes will have time to accumulate as visitation rates are low.

There is some chance that undecorated, non-local pottery might be found in association with quarrying or with adjacent rest area bofedal zones. For example, Formative Period ceramics from the southern Lake Titicaca are typically fiber tempered, but in the north basin and in the Colca they are grit tempered. Evidence of fiber tempered plainware may have resulted from discard during direct access procurement by southern Titicaca Basin visitors. Other temper and paste characteristics may serve to identify non-local pottery. Alternately, non-local pottery may result from reciprocation activities (i.e., the Down-the-line model), but the notion here is that non-decorated, utilitarian pottery may have been too crude to have served as barter goods.

3.7.2. Multiple Reciprocal Exchanges (Down the line) Model


The multiple reciprocal exchanges model involves the direct acquisition of obsidian by local people for the express purpose of transferring the obsidian to neighboring communities in exchange for other items. In this model, people residing in Chivay area, perhaps no more than one day's travel from the source, procure material and transport it to a location where the obsidian is then exchanged with neighbors.

This kind of procurement resembles Renfrew's (1975: 520) "Reciprocity" and the "Down-the-line" modes (see Figure 2-2), where goods are exchanged with neighboring groups of roughly equal status through a variety of configurations that are essentially reciprocal in some form. While a synchronic barter of obsidian for, say, a chunk of ochre from the neighboring region is easiest to describe conceptually, reciprocity relationships are a manifestation of a wide range of social mechanisms (Section 2.2.2). In the context of the Chivay area, reciprocal exchange arrangements such as barter for other products, or for grazing rights, for labor, or for social functions such as bridewealth, may be evident. Furthermore, delayed reciprocal arrangements between neighbors are extremely common. Evidence of Down-the-line exchange may be encountered in a wide variety of socio-political contexts from Archaic Foragers to agro-pastoralists living on the periphery of states during the Late Prehispanic. If demand for a product is sufficiently high, archaeological evidence may be encountered of individuals devoting themselves to procurement in order to satisfy regional demand, but substantial quantities of goods would have to be reciprocated because the households that are sponsoring this increased procurement would have fewer provisioners working to bring subsistence to the household. Thus, if down-the-line demand is sufficiently great then the households devoting themselves to elevated rates of procurement would need to barter products for subsistence goods. In the Andes, the pattern observed from items like salt (Concha Contreras 1975: 74-76;Nielsen 2000) suggests that in modern circumstances when demand, and barter values, are sufficiently high, then the down-the-line network may be simply by-passed in favor of procurement through personal or caravan acquisition. Caravans from the consuming zone or adjacent highland areas will make the journey to procure the material and transport it for household use and for barter (a combination of the Direct Access model and the Independent Caravans model).

In a functioning Down-the-line system the flow of information is also important. The changes in the regional demand for a product like obsidian can return to the source area procurers through direct requests, or it might be reflected in increased barter value in a market context. Information exchange may also return specific demands from consumers as to the size, form, or quality of the source material. The temporal regularity of down-the-line reciprocal trade may also be quite variable, as down-the-line networks may dwindle and then be revived during a seasonal gathering or ceremonial occasion. Reciprocal relationships can take the form of mutualism and buffering, they may result from a need to complement the resources on a neighbor's territory, and they often present opportunities for ambitious individuals to advance their interests through differential access to non-local goods.

Material Expectations

Down-the-line procurement involves local people visiting the Chivay source and acquiring goods to supply the reciprocal exchange network, however large that it may be. In the quarry area one should expect local visitors, and therefore local styles in both discarded materials and local architecture. Procurement may take place in the context of embedded economic activities, such as hunting forays into the high country or pasturing of camelids in the rich bofedal adjacent to the source.

If reciprocation for obsidian takes the form of portable objects, such as non-local ceramics, one may encounter diagnostic, non-local goods in the communities adjacent to the obsidian source. These may be in the form of styles belonging to neighboring communities, or more exotic styles may be found on non-local goods that could have arrived through the exchange network from even more distant areas. There is a high likelihood that reciprocation for obsidian would have taken place in other forms as well: goods that are perishable, labor, or other assets that are otherwise less easy to detect.

If locals are involved obsidian procurement at the Chivay source one may also find that the large nodules available at the Chivay source are used in the local economy as well. That is, if nodules in the Chivay source are up to 30cm in length then large flakes, either cortical or non-cortical, may be expected to have been discarded in residential contexts in local communities. If large nodules are available then those that are not exchanged with reciprocal partners are put to use for local needs. Thus when large cores and flakes are procured in the Chivay source area, then appropriately large flakes should be discarded in the middens of communities in the adjacent consumption zone in the upper Colca.

Procurement and initial reduction at the source will have relatively low variability because it is conducted by the same local methods. Local people will have better knowledge of high quality extraction loci and perhaps there is lower variability in procurement locations as a result. As reciprocity networks, and particularly barter arrangements, are contingent on visual attributes of bartered items one should expect medium reduction of material at the source or in the adjacent communities. At the very least, nodules will be partially decorticated and an initial strike that provides entry into the core should be expected, as this serves to expose the quality of material on the interior to barter partners. Furthermore, if transport does not involve camelid cargo animals (because reciprocity is either taking place pre-domestication during the Archaic Forager period, or otherwise does not involve camelids) one might expect a greater concern for the weight of the nodules and therefore further reduction in the vicinity of the obsidian source.

More advanced reduction may also be expected as it minimizes risk and waste by producing blanks, preforms, and prepared cores in the vicinity of the source where obsidian is abundant. However, according to the Down-the-line model producers have the greatest social distance, and therefore the least information, about their consumers. Advanced reduction limits the possible forms that artifacts may take, and therefore producers would need to know what kinds of tools consumers were planning to produce in order to move beyond initial stages of reduction. Thus medium level reduction might be expected, but not an abundance of advanced reduction at the Chivay obsidian source.

3.7.3. Independent Caravans Model


Long distance transport of goods by way of camelid caravans was well established in the prehispanic Andes. The strongest evidence for the importance of caravan transport comes in the form of ethnohistoric and ethnoarchaeological studies described earlier in this chapter, however archaeological evidence of caravan traffic is usually very light and it often requires inference from indirect evidence. The "Independent Caravans Model" described here consists of caravans organized on the household level, although ethnographic studies show that, in practice, the members of several households will often band together for company and for safety while participating in long distance caravans. It is worth pointing out that independent long distance transport does not necessarily involve cargo animals. It is possible that small quantities of obsidian were carried bytraveling peddlers. As a variant to this model, one should consider that peddlers carrying small portable items, mostly cultural goods like herbs, shell, feathers, but potentially small obsidian tools or cores, could have circulated objects widely without the assistance of llamas.

According to this model, a household with a sufficient number of cargo animals, usually castrated male llamas, will initiate a trade caravan by transporting goods that they expect will be in demand, to regions that they anticipate will have complementary goods to offer them. According to some descriptions, caravans are pursuing a directed acquisition of specific goods and then they return directly home, while other models describe entire circuits where herders acquire goods, travel, barter for other goods, travel some more, perhaps re-trade their new goods and so on; finally returning to their place of origin several months later.

The mere presence of products distributed over larger distances is not proof of caravan transport, either household organized or administered, because other modes such as direct acquisition and down-the-line models actually result in widely dispersed goods as well. Furthermore, many of the distinctive objects that archaeologists recognize as non-local are often small enough to have been transported without cargo animals. Establishing the beginnings of caravan transport is not a simple task because there is no one signature for long distance caravan organization that is distinctive from other modes of transport. Furthermore, many of the goods are believed to have been perishable, complicating efforts to interpret prehispanic trade caravan patterns. Finally, studies of contemporary caravans emphasize that diversified strategies characterize caravan driving, whether in making daily decisions while on the trail, or in the larger context of economy and exchange. It is thus difficult to define a consistent indicator for caravan activity.

Portable diagnostic artifacts, whether decorated ceramics or other exotic goods, are often relatively small and therefore the artifact weight and total quantity frequently cannot be used to differentiate between caravan transport, traveling peddlers, and down-the-line exchange. The temporal regularity of exchange, however, is a consistent measure that archaeologists can recover from stratified deposits. When regular caravan transport routes developed then the scheduling of such transport may have been linked to the timing of annual events such as harvests and annual ceremonies, and if so these cyclical patterns would result a steady accumulation of non-local goods through time. In contrast, down-the-line exchange depends upon the articulation of many individual exchanges and it is not linked to the acquisition of scheduled harvest products in the same manner as caravan transport and therefore the presence non-local goods would have been irregular.

In terms of the network configurations discussed earlier (Figure 2-3), the configuration that describes the diffusion of obsidian in the region is distinct from the configuration of the regular articulation between herders and farmers that involved the barter of pastoral products for agricultural products. However the regular conveyance of some agricultural goods adjacent to a raw material source creates a context for conveying larger quantities of obsidian regionally. Ethnographic studies indicate that caravans will opportunistically embed exchange into other activities. For example, Nielsen (2000: 488) explains that caravans primarily organized around salt transport would carry a variety of other trade items, and they would occasionally stop to procure raw materials, such as lithics, when the caravan route travels past a known source. Similarly, there is caravans that visited the Colca valley in prehistory from the Titicaca Basin were passing with 3 km of the Chivay source if they used one of the popular routes into the Colca from the south-east direction. Thus, procurement of obsidian was likely to have been associated with long distance exchange opportunities.

Material Expectations

Caravan procurement would have consisted of pastoralists traveling to the Chivay source, acquiring obsidian that they believed to have exchange value, perhaps processing the nodules to some extent, and then transporting the material to consumers or to other traders in areas far from Colca. As temporality is a significant part of caravan organization, regular and dense procurement activities are expected when a caravan undertakes a detour from the principal travel route. Some preparation or maintenance of the trail from the principal travel route to the obsidian source is expected if the animals are heavily laden. The amount of processing and initial reduction in the source area probably reflected the number of days that a caravan would have been willing stop, and conversely the transport of whole nodules is conceivable with the assistance of cargo animals such that some extra grams of weight were less of a limiting factor. As mentioned, cortex is often very thin on Chivay obsidian, therefore it would not be surprising if some percentage of the material transported away from the source area included unreduced, whole nodules.

Ethnoarchaeological studies report that caravan drivers look for camps that include high quality pasture, water, and corrals if possible. When these features occur close to an extraction site, a relatively dense obsidian processing area may be found nearby because herders can nourish their livestock while simultaneously working stone. Because of the temporality of caravan activity, procurement was probably intensive, though episodic, through time. When caravans or animals were present, processing would occur and then the material would be conveyed away and the quarry area would see little use until the next episode of intense production. Evidence of production may be variable, however, because caravans could also transport whole nodules without very much difficulty.

Many of the artifact types that were produced from obsidian transported by caravans can deduced based on archaeological evidence acquired from consumption contexts dating to the Terminal Archaic and later. The principal artifact form made from obsidian during the time that camelid caravans conceivably operated (subsequent to the Late Archaic), is the small triangular projectile point diagnostic to the Terminal Archaic and later. This is not to say that other tool forms were not being made, as other non-diagnostic biface forms could also date to the Terminal Archaic or later, however the vast majority of bifacially-flaked obsidian artifacts are projectile points. Furthermore, cores were likely transported away from the obsidian source because simple flakes serve as valuable cutting tools. Therefore, the majority of obsidian production would have prioritized the production of cores and flakes that serve as blanks for triangular point production. If advanced reduction occurred in the quarry area it is likely that small, triangular point forms were the objective.

Reduction strategies that target the production of small triangular points were probably relatively flexible because these point forms are not especially long or delicate, and therefore it would have been possible to produce the appropriate blanks from a variety of core forms. None-the-less, the overall variability in formal tools forms produced from obsidian is exceptionally low because series 5 points look quite similar. Therefore in terms of intensified production from the Terminal Archaic and onward, relatively consistent reduction strategies in the procurement workshop zone may be encountered.

Because these caravans were independent, one might expect slightly greater variability than in the other models of obsidian distribution because various individual households were participating in this procurement, and methodological variation by region might be reflected in the reduction methods. Finally, independent caravans may behave in other, relatively variable ways such as in architecture, in divergent ceramic styles, and in the types of obsidian pieces that were being exported.

3.7.4. Elite-Sponsored Caravans Model


Ethnohistoric evidence suggests that elites commissioned long distance trade caravans to procure materials that were used in a variety of elite strategies at the regional center. As discussed above, Stanish argues that Titicaca Basin elite-administered long distance trade involved acquiring goods for favorable barter rates in distant valleys and then acquiring prestige from the redistribution of these goods. Late Prehispanic elites were probably in a good position to initiate large caravans: they had immense camelid herds and their followers owned them tax payments in the form of labor. In addition, elites would have had the surpluses necessary to initiate a large scale trading venture.

Would such caravans have visited the Chivay source and extracted obsidian for elite consumption or redistribution? Stanish (2003: 69) argues that all Prehispanic trade was administered trade, as opposed to market-driven trade, but he specifically excludes trade in obsidian as "small and light" and capable of being transported through down-the-line exchange. Obsidian does not appear to have been a high prestige item along the lines of precious metals in the prehispanic Andes. Under the Inka there is evidence of control of access to tunnels leading to rich gold mines in the Andes (Burger and Glascock 2002: 364). With obsidian, however, there is no evidence of elite control either in obsidian consumption patterns in regional centers, or in the Arequipa obsidian quarry areas that are generally dispersed and it would have proved difficult to limit access to them (Jennings and Glascock 2002: 115-116). Thus extensive elite-administered acquisition or redistribution of obsidian should not be expected. Nevertheless, given the importance of exchange and non-local goods in issues relating to the origins of social complexity, any evidence of elite-organized raw material procurement should be studied closely.

Material Expectations

Elite-administered caravans would be relatively difficult to differentiate from independent caravans in their source activities. Elite related diagnostic pottery may be encountered in the vicinity of the source area. There may have been some degree of greater standardization if these were part time or full-time specialists working for the elites. Elite-sponsored procurement may involve greater intensification than would be expected from independent caravan procurement because these task groups likely have been organized and dedicated to the procurement objective. Finally, the export of large nodules may have increased as elite-sponsored caravan trains were reportedly large and capable, and the weight of nodules would not have overly interfered with the progress of the caravan following this model. In addition, if the object of elites was prestige building, larger nodules would probably have been more impressive in the distant consumption zone. In short, differentiating elite sponsored caravans at the source may be relatively difficult unless pottery or some other diagnostic material is found to have been associated with procurement.

3.8. Summary

This chapter establishes the background for further research into the production and circulation of Chivay obsidian in the prehispanic Andes. The chapter began with a review of different forms of exchange based on Polanyi's framework but discussed in light of Andean prehistory. The specific geographical and cultural conditions that distinguish ancient Andean economies were placed in a context that is comparable with other mountain culture regions.

Theories of culture change might suggest that regional exchange requires the managerial or coercive power of elites to organize caravans and benefit differentially from the trade goods. However, archaeological evidence shows that the persistent obsidian exchange that moved small quantities of obsidian since the Early Archaic saw a significant increase during the Terminal Archaic. This increased circulation of obsidian is an early symptom of the dramatic economic and social changes that would occur in the south-central Andean highlands in Formative times.

The evidence of this circulation was perhaps not apparent at early centers because these centers were not yet dominating regional exchange patterns through the control of labor and large herds of camelids that are believed to have been a feature of Late Formative primate centers. The mechanism responsible for disseminating obsidian to the region beginning in the Terminal Archaic is perhaps related to a number of interrelated phenomena. The principal factors governing this change may include: the lowered cost of interaction and transport of weight facilitated by caravan animals, the expansion of regional-scale social networks, and the impetus provided by the needs of some to differentiate themselves in this time of emergent social ranking through the ownership of non-local goods such as obsidian.

Ch4 Regional Geography

Regional Geography and Geology of the Upper Colca Project Area

The Chivay obsidian source is located on the margins of the Peruvian altiplano above the Colca Valley in southern Peru. The geographical position of the obsidian source area (71.5° S, 15.6° W) at on the western edge of the altiplano above the richly productive Colca valley had bearing on circulation of Chivay obsidian in regional prehistory. This chapter begins with an overview of the geographical relationships in the Upper Colca study area including climate, biotic zones, and resource availability. The chapter will then describe the ecological and geological context in the Colca Valley, and the influences that spatial relationships may have had in prehistory.


Figure 4-1. View of the volcanic Chivay source area above the town ofChivay in the Colca valley. The high point in the center-right of the frame, Cerro Ancachita, is above the Maymeja area. This view is from the Arequipa highway to the south-west of the Chivay source at 4720 masl.

Subsequently, further details on the geological structures in the area of the Chivay obsidian source will be presented, together with the geomorphological process that have influenced obsidian procurement in the area. Finally, this chapter concludes with a discussion the formation of silicic lavas and tool-quality obsidian, and the research that has been conducted to date documenting Tertiary obsidian flows in the Colca Valley.

4.1. The geography of the Colca Valley study area

A brief review the geography of Upper Colca study area better contextualizes the ancient lifeways and relationships that are the subject of this dissertation. The 2003 research project was organized into three distinct blocks of contiguous survey (1, 2, 3) with other adjacent areas numbered 4, 5, and 6. The three major survey blocks correspond with ecological zones in the Upper Colca area and the overall ecological variability in the project area will be reviewed here by a discussion of each block.

Ecological complementarity between highland herders and valley agriculturalists is a widespread feature of Andean economy. Research into the productive potential of "pure" pastoralism in the Andes has shown that the caloric return and efficiency is insufficient to sustain communities without inputs from non-pastoral communities (Thomas 1973;Webster 1973). The Upper Colca study area lies immediately above the altitude of intensive valley agriculture that begins at the village of Tuti (3840 masl), corresponding with the upper portion of the suniand with the punaecological zones, and reflecting the local precipitation and temperature gradient (ONERN 1973: 39;Pulgar Vidal 1946). The town of Tuti is at approximately the altitude of Lake Titicaca, and in some ways the land use practices observable upstream of Tuti in the Callalli area are comparable to those in the Lake Titicaca Basin, although this western slope location does not have the moderating effect of the large lake body itself.

Local elevation



Upper Colca Survey Blocks

4000 - 5000


Rich pasture lands and rugged volcanic terrain in the Chivay obsidian source area, the San Bartolomé puna near Chalhuanca, and below to the town of Pulpera.

1, 2, 4, 5

3600 - 4000


High elevation agricultural lands, lower quality pasture, upper river valley from Chivay to Sibayo / Callalli.

3, 6

3300 - 3600


Lower portion of the main Colca valley.


Table 4-1. Andean ecological zones with approximate local elevation values for each zone.

Annual rainfall ranges from 550 and 750 mm, depending on elevation, and the average annual temperature is 1° C in the puna (ONERN 1973). In this area, precipitation is higher with altitude, but temperature, especially night time temperature, drops with altitude. The result is a balance between the potential for dryland agriculture, altitude, and climate that has periodically allowed farming in the Upper Callalli area in prehispanic and Hispanic times.


Figure 4-2. Survey blocks in the Upper Colca study area are shown with modern production zones described in Table 4-1. These zones reflect changes in altitude and rainfall as the Colca river, descending westward, is surrounded by puna pasture lands.

A relationship of mutual benefit exists between residents and resources of the higher altitude ecological zones and the vegetatively-productive main Colca valley. A principal assumption of this research is that the procurement of resources like obsidian have long been structured by the human use of complementary ecological zones. Some form of regular articulation between the highlands and the valley therefore probably existed through much of prehistory without special recourse for obsidian provisioning. This interaction may have taken the form of regular visits to the Chivay obsidian source area by valley agriculturalists, who may have been hunting or herding in the adjacent highlands, or by way of visits by highland pastoralists traveling from the puna regions in order to visit the Colca valley to barter for agricultural goods.

4.1.1. Climate across the study area

The Colca valley is located in the western cordillera of Andes in a semiarid climate that is cool and unpredictable. As a high altitude region of tropical latitude, diurnal temperature variation is more prominent than annual temperature variation (Denevan, et al. 1987;Troll 1968). Precipitation is highly seasonal, however, and it is changes in rainfall and availability of pasture that strongly influence the scheduling and intensity of pastoralism and dry land agriculture throughout the study area. The south-central Andes is a region south of 15° S latitude, outside of the Intertropical Convergence Zone, and precipitation is relatively unpredictable with high interannual variability because it results largely from the convection of humid air from the Amazon Basin to the east (Johnson 1976). The western slope of the Andes thus lies in a rain-shadow, and warm, dry winds from the east result in low average annual precipitation as one descends the western flanks of the Andes. In the ecotone that is the Upper Colca study area the herding economy thrives, because these communities can seasonally exploit the rich pasture lands and greater rainfall of the high puna while simultaneously interacting with the communities and products of the main Colca valley, communities to which they may have family linkages.

Of the major obsidian sources in the Andes, the Chivay obsidian source is the highest elevation source and the environmental conditions in the area during prehispanic procurement visits to the source are the subject of some speculation. Annual climatic variability under modern conditions at the Chivay source can be inferred from the relationship between altitude and temperature, known as the lapse rate, and this rate may be calculated using records from nearby meteorological stations. A mass of rising air will cool at the dry adiabatic lapse rate that is often estimated as 0.98° C per 100m of ascent for dry conditions, while the saturated adiabatic lapse rate is typically 0.5° - 0.6° C per 100m at dew point when temperatures are around 10°C (Adiabatic lapse rate 1999). This theoretical lapse rate may be compared with empirical temperature data reported by the ONERN environmental investigation, a regional study of the entire Majes-Colca drainage that preceding the construction of the Majes hydroelectric project. These data include mean monthly temperature records from meteorological stations at Pañe near the Colca headwaters, Sibayo on the northern edge of our Upper Colca study area, and the mid-altitude stations of Arequipa and Pampacolca (ONERN 1973;WMO 2006).

/Figs_Ch4/weather_stations_map-2.jpg /Figs_Ch4/weather_stations_graph-2.jpg

Figure 4-3. Temperature by Altitude at mid- and high-altitudeArequipa meteorological stations. All data from ONERN (1973) except Arequipa values from W.M.O. (2006). These data represent a lapse rate of -0.56°C per 100m of ascent.

These empirically-derived mean annual temperature values from the Arequipa sierra show a lapse rate of -0.56°C per 100m, which is in the range of the standard saturated adiabatic lapse rate of -0.5 to -0.6 per 100m, discussed above, from which can be inferred a mean annual temperature of 0.8°C at the Chivay obsidian source. Given the aridity of this region much of the year, the relatively low-slope (saturated) lapse rate is unexpected and the raw tabular data in the ONERN report suggests that a more detailed examination of the data will better allow the temperature variation at the Chivay obsidian source to be inferred.

Models based on local meterological data

Monthly mean temperature data collected from 1952 - 1970 at local meteorological stations show highs, daily means, and lows, and the considerable differences from the lapse rate function that probably result from the dramatic diurnal temperature variations and the relatively thin atmosphere in this region. The seasonal effects on lapse rate can be considered by looking more closely at data from the nearest reliable weather stations, at Sibayo and Pañe.


Figure4-4. Comparison of Temperature highs, means, and lows for August and January from two meteorological stations with records kept between 1952-1970(ONERN 1973).

August (winter, dry)

January (summer, rainy)
































Temp. Change / 100m







Figure 4-5. Mean monthly temperatures (°C) from data in ONERN (1973). Inferred temperatures at the Maymeja area of the Chivay obsidian source shown in italics.

/misc/image014.jpg /misc/image015.jpg

Figure4-6. Precipitation by Altitude (left), precipitation in the study area (lines represent highs, averages, and lows) is highly seasonal as shown in 14 yr precipitation record from Sibayo on right. Data from ONERN (1973).

The local relationship between altitude, temperature, and precipitation on the dry western slopes of Arequipa are evident in these data derived from tables provided by ONERN (1973, Appendix II). Diurnal temperature variation, highest during the dry winter months of July, August, and September results in a steep lapse rate such that low temperatures in Sibayo, near the lowest part of the study area, are only 3° C warmer than those mean winter lows at the Chivay source. Curiously, the high temperature values cross over between summer and winter months (

Figure4-4) resulting in somewhat more of a temperature difference between Sibayo and Pañe during the rainy summer months than the winter months, probably reflecting the higher precipitation at altitude. As most precipitation, including snowfall, comes during the summer months of January and February, greater cloud cover and reduced solar insolation at altitude are expected during the summer months.

At the Chivay obsidian source the inferred temperature values are in keeping with observations made during the course of research at the source in August and September 2003, as will be discussed below. This is not to suggest that the modern climate regime existed during prehispanic times, but these relative contrasts in temperature and precipitation values throughout the Upper Colca study area probably have existed throughout the Holocene although actual precipitation and temperatures varied from those of the modern conditions. Paleoclimatic reconstructions from the Terminal Pleistocene and through the Holocene using data from cores collected in large lakes and glaciers were discussed in Chapter 3.

It is important to stress that adiabatic lapse rate temperature estimates do not account for other important factors affecting local climate such as local insolation, vegetative cover, and wind, as well as temperature contrasts due to solar heating where the atmosphere is thin, and temperature inversions in mountain valleys (uncommon at tropical latitudes). Further, orographic effects such as mountain and valley winds have specific and localized effects in the Upper Colca survey blocks. Seasonality in the upper Colca region includes rain and snowfall in the highest reaches during the wet summer months as well as a high incidence of lightning strikes (a major cause of death in the Peruvian highlands today). Modern herders in the Colca area adapt to these conditions by distributing their impacts and exploiting the high altitude resources primarily during the dry season.

The equatorial bulge in barometric pressure

An issue related to temperature variation is the presence of an equatorial bulge in atmospheric pressure as one ascends to high altitude (Ward, et al. 2000: 26-28;West 1996). In low latitude areas, barometric pressure at a given altitude is correspondingly higher than is pressure at that altitude in mid and high latitude areas of the world.


Figure4-7. Latitude against barometric pressure (West 1996). Lines show altitude in km. Annual temperature seasonality is minimal in the equatorial areas and thus seasonal effects at low latitudes are not shown in this graph.

Models of variance in barometric pressure with latitude at a given altitude based on data derived primarily from weather balloons explain why there is greater available oxygen at 5000m in low latitude areas than in high latitude areas. The differences in pressure are particularly notable during high latitude winters.


15° latitude

60° latitude, summer

60° latitude, winter

Equivalent Altitude
















Table 4-2. Equatorial bulge and effects on barometric pressure (in torr) at 15° and 60° latitude. From data in (West 1996: 1851). Equivalent altitude column shows altitude at 60° latitude with equivalent pressure to the value shown at 15° latitude.

As shown in Figure 4-7 and Table 4-2, the equatorial bulge in pressure means that the available oxygen at the Chivay source at 5000 masl is only close to the pressure that of a location nearly 1000m lower if the source were located at 60° latitude (in winter). In other words, during the winter at 60° latitude, at approximately the latitude of Anchorage Alaska, one would have to be at only 4400 masl altitude to find available oxygen at levels equivalent to that found at 5000 masl at the Chivay source. Barometric conditions similar to those found near the summit of Nevado Ampato at 6000 masl can be found in winter at 5310 masl near Anchorage, AK. The high altitude Chivay source is not as inhospitable as would be this zone at the equivalent altitude in higher latitude areas. Resource patches, such as rich bofedal grazing areas, are sufficient to draw seasonal or permanent residents to these altitude zones in the Andes. Raw material sources, particularly mining operations, are another significant draw to high altitude locations both in past and in modern times (Ward, et al. 2000: 336-344).

4.1.2. Lower elevation biotic zones: Study Area Blocks 3 and 6

Beginning with the lower elevation part of the survey area and moving upstream, the vegetation of the Colca valley above Tuti consists of low grasses dominated by the ichuvariety during the dry season. The Colca River is relatively low gradient in this area, and two or three levels of natural river terraces are evident along the valley margins. The geography of the zone is dominated by the fact that the confluence of two large river systems (Colca and Llapa) occurs here. This area holds the largest settlements in our study area, but both agriculture and pasture appear marginal in immediate area of these towns. Rather, the settlement of the upper river valley seems to reflect the importance of the articulation between the main Colca valley and the broad altiplano. The village of Sibayo rests precisely at the main confluence of the Colca River and the Llapa River, and as is evident in the Shippee-Johnson expedition, Sibayo has long served as a principal modern ingress to the Colca Valley until the Chivay-Arequipa highway was completed (Shippee 1932).


The bulk of the contemporary populations in the study area reside in towns in block 3 established during the colonial period that are distributed along the Colca river. The largest settlement inside the study area is Callalli, a town with a 1993 population of 1295 persons, and across the Colca River the population of the town of Sibayo is 508, and upstream in Block 5 the cooperative of Pulpera numbers 85 residents (I.N.E.I. 1993). These towns are primarily service centers and district seats for widely distributed populations with an economy based largely on pastoral products, and on extensive interaction through trade with their wealth in camelid herds, and have long resided in rural hamlets and herding outposts. Callalli and Sibayo first formed as part of the sixteenth century reducción of the Yanque Collagua, and in the colonial period the province of Collaguas (Caylloma) held three-quarters of the livestock in all of Arequipa (Cook 1982;Manrique 1985: 95-96). The dominance of the herding economy in these upper valley towns is evident in the 1961 census where both Callalli and Sibayo populations are reported as 93% "rural", while the average percentage for all nineteen towns in the Colca census, including large and dense agricultural communities downstream, was only 52% "rural" (Cook 1982: 44).

These early villages also formed an important source of labor for colonial mining ventures in the Cailloma region (Guillet 1992: 25-27). Mercury and copper were mined between Callalli and Tisco (Echeverría 1952 [1804]) and Lechtman (1976) reports a structure in Callalli known as " La Fundición" that is described as "stone metal smelter, probably colonial, said to be for copper smelting: mineral, scoria on surface". A location known as "Ccena" or "Qqena" is described as having "metal smelters near a pre-Spanish occupation site: mineral, scoria, surface sherds" (Lechtman 1976: 11). The toponym "Ccena" can be found close to the Llapa and Pulpera stream confluence upstream of Callalli. These historical smelters were not encountered during the course of our survey in this particular area.

Pyrotechnical installation

One structure ([A03-842], Figure 4-8) was identified in the course of the 2003 survey that appears to be of colonial period construction and it appears to have some kind of pyrotechnical function (B. Owen 2006, pers. comm.). The structure has two doors in the lower area, apparently providing access to the lower furnace. The internal lower construction is built of thermally altered stones and has a cracked lintel. A variety of pyrotechnical structures are known in the south-central Andes (Van Buren and Mills 2005) that were used to heat lead, silver, or copper, and other oven types (e.g., pottery, bread) are known in the region as well. Slag or other evidence of smelting was not encountered in the soils adjacent to the structure, however, although the building is immediately adjacent to a stream channel and such materials may be eroded or difficult to detect.



Figure 4-8. Exterior and interior of probable colonial pyrotechnical structure at Achacota near Pulpera, upstream and south of Callalli [A03-842]. One meter of exposed tape is visible in each image.

Vegetation and dryland agriculture in the study area

The Callalli area consists primarily of high elevation grassland punaecology (1973). Dry bunch grasses are available much of the year in this area, but during the wet season (austral summer) a greater variety of grasses become available and herds are brought to the valley to exploit the pasture of chilliwuaand llapagrasses.

Callalli and Sibayo lie in the upper reaches of agriculture at this latitude and evidence of abandoned fields are visible in the upper valleys. Plants like tubers, oca, and chenopodium were historically viable at this altitude (Echeverría 1952 [1804]), although microclimatic variability is influential in these conditions of marginal dryland agricultural production. Guillet (1992: 24) cites evidence for climate change from historical sources that describe the cultivation of maize above the current limits for this crop, and coca, membrillo, and peppers on terraces that lie at altitudes where these crops are not feasible today. Wernke (2003: 51-52) considers the significance of climate change evidence from ice cores in mountain ranges to the east for Colca valley culture history and agriculture.

In Markowitz's (1992) ethnographic study at Canaceta, a village at 4000 masl and approximately 12 km upstream of Callalli, villagers explained that they had formerly engaged in agriculture at this altitude, but that they had recently abandoned the practice due to lack of rainfall. She observes that practicing a mixed subsistence system of agriculture and pastoralism was an important cultural ideal in the Canaceta, but that in recent years due to changes in the climate and in economic circumstances they had increasingly become specialized on pastoral production complemented by exchange (Markowitz 1992: 48). Under modern circumstances there is probably not a very high return on labor invested in agriculture in this area as the increased economic integration, and an improved transportation infrastructure with intensive farming areas at lower elevations, has further induced residents towards specialized economic practices. The current distribution system emphasizes pastoral production in the highlands and higher yield agriculture on lands at lower elevations.


Figure 4-9. Tuber cultivation at 4200 masl surrounded by large tuff outcrops.

In the course of survey in 2003 cultivation was observed in a few high altitude locations, such as the upper reaches of Quebrada Taukamayo. The plots were at 4200 masl in a north-north-east facing (10°) aspect and a mild slope (8°). The area is relatively sheltered by the presence of lava tuff flows (see Figure 4-9) that may have had a temperature moderating effect acting like large terrace stones that are known to reduce diurnal variation by absorbing heat during the day and releasing heat at night in the immediate valley microclimate (Schreiber 1992: 131).

Flora and fauna

Important flora and fauna to residents of the Upper Colca region include the following (Gomez Rodríguez 1985;Guillet 1992: 130;Markowitz 1992: 42-44;Romaña 1987;Tapia Nuñez and Flores Ochoa 1984). Major flora comprise grasses such as Chilliwa( Calamagrostis rigescens), a frost-resistant perennial grass that thrives during the rainy season and in bofedales, grazed by a wide range of animals and also used for roof thatching. Other important puna pasture grasses include llapa, malva, sillo, and paco,though these are principally consumable by herbivores only during the rainy season. Ichu/ Paja( Stipa ichu) is a common grass used for thatching. In the higher elevation bofedales one can encounter parru, a grass preferred by alpacas. Wild fruits are gathered seasonally by locals including locoti(cactus fruit), q'ita uba(wild grapes), and sanquayo(a plant related to chirimoya) (Markowitz 1992: 43). In the high elevation area of the obsidian source yareta( Azorella compacta), a green, flowering cushion plant is one of the few flora that grow in the unirrigated areas of this harsh volcanic terrain. In addition to animal dung, dried yaretais the only dense, combustible fuel widely available above 4500 masl. As a local herder, T. Valdevia demonstrated, the cushion plant will burn when it is kicked over and allowed to dry out for several weeks. Drought and cold-resistant shrubs, including tolaand cangi, are valuable sources of firewood in the punatoday, though the shrubs are over harvested in many areas.

Fauna species include a number of birds that are hunted for their meat including the Grey Breasted Seedsnipe known as puko elquio( Thinocorus orbygnianus), partridges ( pishaq), and the guallata,the large white Andean Goose (Chloëphaga melanoptera)(Hughes 1987;Markowitz 1992: 43) .Several Andeancondors ( Vultur gryphus), for which the Colca is renowned, flew repeatedly near our work at the Chivay obsidian source in 2003. Wild mammals observed in the study area include viscacha( Lagidium peruana), tarucadeer ( Hipocamelus aticensis), and the wild camelid vicuña (Vicugna vicugnaor Lama vicugna).Thetrout found in the streams represent an important food source, but these fishes were introduced in the nineteenth century.

4.1.3. High Puna: Block 2 survey and adjacent Blocks 4 and 5

This high altitude zone of the study area is dominated by the Pliocene lavas and Quaternary moraines, as will be described in more detail below in the geology section. Much of these survey blocks consisted of porous lava rock and sandy soils covered by a mantle of ash, lapilli (rock fragments between 2 - 64 mm across) and the occasional volcanic bomb (molten rock between 64 - 1000 mm). During the dry season surface water was available only sporadically across the area and, unsurprisingly, reliable water sources frequently have archaeological sites nearby. The renowned "Ventanas del Colca" tuff formations occur on the edge of the San Bartolomé survey area (block 2) at the point where the highway connecting Callalli with Arequipa climbs out of the Pulpera drainage and straightens out for its run across the open puna grassland.

This area is also remarkably wet, the environment is productive, and faunal density is relatively high. Even during the dry season the Block 2 area has reliable water.


Figure 4-10. ASTER scenes with the terminus of volcanic breccia outcrop in green, photosynthesizing plant areas (bofedales) shown in red pixels, and ash in white. The black diagonal line is a seam between the two scenes.

Block 2 geography

Block 2 can be characterized as containing rich bofedales, reliable hunting opportunities, and access to lower Colca Valley resources only two-day's travel away. This area, known as San Bartolomé, can be considered a "puna rim" ecological area because animal and plant species are affected by the presence of lower elevation Colca Valley environments immediately downhill to the north and west, and warmer air rises from the lower valley affecting the local climate. The Pliocene eruption of the Barroso group vents of Huaracante, Hornillo, and Ancachita resulted in the predominantly silicic coulee flows of lava described as "Centro Huarancante" by the INGEMMET study (Palacios, et al. 1993: 139). On the eastern edge of the Huarancante Group the viscous lava flows terminate in breccia outcrops where they overlie crystalline ignimbrites (TBa-c) of older Pliocene age, also belonging to the Barroso formation. Below the toe of the lava the crystalline ignimbrites rich in dacite appear as light colored, sandy soils, where perennial surface water has created good grazing opportunities in this flat expanse to the east of the lava flows. Bofedales are found below each quebrada descending from the lava formations, a pattern that is perhaps the result of subsurface water moving through the lava flows and emerging at the contact zone with the ignimbrite formation.

A number of small rock shelters occur along the base of the upper lava flow. The rock shelters are typically dry, but the floors are sloping out onto talus slope below and so they offer little in the form of shelter inside the drip line. There are some notable exceptions, as will be reported in Chapter 7, and relatively dense occupations were found at few of the best rock shelters. On the whole, Block 2 follows along the eastern periphery of a lava flow where there is a concentration of resources. Water emerges from the lava onto the open pampa, grazing opportunities for wild and domestic herbivores on the bofedales and adjacent grasslands are good, and Andean geese and other puna bird species are seen in the greatest numbers. The lava flow provides topographic variability on the perimeter of a wide and often windy plain, providing shelter and occasional, small rock shelters.


Figure 4-11. View of San Bartolomé (Block 2) area during the dry season from atop a toe of Barroso lava looking northeastward. For scale, our white pickup truck is visible below. A rich bofedal is visible 3 km to the northeast.

Current environmental conditions suggest that the edge of the lava flow served as a rich ecotone with water, shelter, and numerous hunting opportunities for bird species, vicacha, tarucadeer, mammals, and probably the wild camelids guanacoand vicuñaat certain times in the past. We often observed a number of wildlife species in the course of research in this area including viscachaand geese, and perhaps the density of wildlife in this locale was due in part to lack of surface water elsewhere, on the adjacent lava beds and ashy soils. Even well into the dry season, parts of Block 2 area have reliable water and soft grasses.

Regional geography

Block 2 is a natural bottleneck for economic traffic moving between the Colca and the Titicaca Basin along what may be interpreted as the prehispanic trail system. Due to the steep descent to the Pulpera on the north side, and the glaciated volcano Nevado Huarancante on the south-west side, many travelers would probably have traveled through Block 2. The presence of a very large bofedal and rich hunting opportunities in the adjacent lava flows to the west probably made the Block 2 area an even greater attraction for travelers with caravan animals. This area lies on the periphery of the Colca valley and it is approximately one long day from the rich grazing areas of the Escalera access to the Colca valley, and about two days from the town of Yanque.

4.1.4. The Chivay Source: Block 1

The geological source of Chivay obsidian is above 4800 masl among the lava flows from two Barroso (Pliocene) volcanic vents named Cerro Ancachita and Cerro Hornillo. The discussion here focuses on the geographical context of Survey Block 1, and the geology of these volcanic features will be explored in more detail in Section 4.3.3, below.

In the high altitude portions of the Upper Colca survey, most notably in the Chivay obsidian source area above 4800 masl, the local temperature and climatic exposure is strongly affected by the lack of vegetative cover and the katabatic (mountain breeze) and anabatic (valley breeze) winds. The winds were a daily feature during the Block 1 fieldwork in the months of August and September 2003, and the winds are most notable in the mornings and evenings when the temperature differential between the high altitude areas and the warmer Colca Valley are greatest. Extrapolating the local lapse rate from Colca meteorological stations suggests that the obsidian source, at around 5000 masl, would have mean temperatures of 0.5 to 1.0° C year-round (

Figure4-4), not accounting for the effects of wind in this exposed area. In the course of fieldwork daytime high temperatures were approximately 5° C, and nighttime temperatures were commonly -9° C with the coldest night measured at -12° C. These data largely corroborate the estimates derived from the local adiabatic lapse rate discussed above.

4.2. Tectonic geology

The Colca Valley lies on the western margins of the broadest region of the Andean cordillera at 15° south. The Andes consist of several parallel mountain chains and in this region that includes areas of southern Peru, western Bolivia, and northern Chile the parallel chains separate and landforms between these mountain chains comprise the broad, high-altitude valleys and plateaus known collectively as the altiplano. The altiplano is a treeless region at an average altitude of 3700 masl extending from the Lake Titicaca Basin to Bolivia's Lake Poopó thatis second only to the Tibetan Plateau as the largest plateau in the world (Clapperton 1993: 45-47;Pearson 1951). The Quechua term punafor "elevated area" applies to the ecological band of the altiplano, but punacan also refer to a broader regions and life zones. The steep Andean cordillera includes volcanism over some of the thickest crusts in the world (Aramaki, et al. 1984: 217), and has resulted in the steep mountains and compressed ecological zonation that strongly influences human organization and subsistence strategies in the region (Troll 1968;Winterhalder and Thomas 1978).


Figure 4-12. Select raw material sources in the central Andes.




Other names





Titicaca Basin, Cotallalli

Burger, et al. 1998, Brooks, et al. 1997





Frye, Aldenderfer, Glascock 1998





Burger et al., 2000, 2001, 2002.





Yacobaccio et al, 2004

Sora Sora




Giesso, 2002, pers. comm.





Burger,et al. 1998.





Burger, et al. 2005




Umasca, Cusco Source

Burger 1998, Jennings, et al. 2002

Querimita Basalt




Giesso 2000, 2003

Uyo Uyo




Brooks 1998, Wernke 2003




Mistaken for Quispisisa

Glascock, et. al., in press.




Andahuaylas A

Glascock, et. al., in press.




Andahuaylas B

Glascock, et. al., in press.

Table 4-3. Coordinates and names of select raw material sources in the central Andes. Coordinates in WGS1984 datum.

It is volcanic processes at a continental scale that have resulted in the discrete pattern of obsidian sources in the central Andes in an arching line between 3000 and 5000 meters above sea level. As discussed by Clapperton (1993)and Thorpe et al. (1982;1981), the volcanic origins of these features involves a consideration of plate tectonics in western South America. Off of the coast of Peru the oceanic Nazca plate is subducting beneath the South American plate, resulting in the magmatism, seismicity, and tectonism characteristic of the central Andes. Geologists date the initial uplift and deformation of the Andean geosyncline to the Laramine Orogeny during the middle Cretaceous period, a period of worldwide mountain building. The Andean batholith, a massive igneous intrusion that underlies the western range of the central Andes (Cordillera Occidental), formed during this orogenic period. The eastern range of the central Andes (Cordillera Oriental) arose from laterally compressed geosynclinal rocks that emerged as folded stratigraphy. The trench between the western batholith and the folded eastern stratigraphy, known as the Titicaca trough, continued to accumulate sediments through the Miocene. A quiescent period followed when the central Andean region matured through erosive processes into a relatively level surface of the altiplano, moderating the terrain that lay above the volcanic, sedimentary, and metamorphic layers that make up the structure of the Andes. This surface undergoing erosion could not properly be called the altiplanoduring the Late Miocene because, interestingly, the surface was still under 1800 masl by 10 Ma (Gregory-Wodzicki 2000). Estimates of uplift rates on the order of 0.2 - 0.3 mm/yr have created the raised plateau known as the Altiplano from the Miocene to the present. Recent estimates by Thouret et al. based on incision dates in major canyons of Ocona below the Cotahuasi valley suggest that "downcutting may have taken place before 9 Ma but most likely before 3.8 Ma and again before 2.7 Ma, based on dated valley infillings" (Thouret, et al. 2005). They suggest that accelerated valley incision was due to increased runoff that resulted from glaciation of the high Andean peaks, and by implication the Chivay obsidian source was at already at high altitude, probably covered by glaciers, by some time in the Late Miocene and Pliocene. Given the

4.3. Formations in the Upper Colca Valley

The volcanic terrain and the temperate lower valley in the Chivay area have largely conditioned human activity patterns in the region for the past 10,000 years. The local geological sequence will be summarized from text and maps from Palacios et al. (1993) the primary INGEMMET source for the region, as well as the ONERN (1973) geological study associated with the Majes Irrigation project, and the discussion will be supplemented with evidence from more recent specific studies in the region.


Figure 4-13 Rock groups in the Colca region (based on Palacios et al., 1993).

The Colca region is dominated by late Cretaceous sedimentary rocks overlain by Tertiary flows and tuffs of basaltic to rhyolitic composition. Some Jurassic sedimentary strata are exposed consisting of sedimentary and metasedimentary rocks. The region has continued to be volcanically active during the Pleistocene and Holocene epochs in the form of stratovolcanoes that ring the Colca valley as well as exogeneous andesitic domes and flows occurring in the Colca River valley itself.

4.3.1. A geological descent of the upper Colca drainage

If one measures the river from its birthplace in Imata near the department of Puno, among a fan of tributaries that flow into Lake Jayuchaca at around 4500 masl, to its final exit into the Pacific Ocean at Camaná, the course of the Colca River is approximately 450 km (Parodi 1987: 31) with an average gradient of 1º. The river that begins in the low gradient marshlands of the Peruvian altiplano becomes increasingly steep as it descends into one of the world's deepest canyons, finally flattening out as it emerges on the littoral batholith only 48 km from the sea.

For the first 70 km from its origin the Colca River trends northwest at a low gradient, open channel descending towards the community of Huinco (3950 masl). This lower portion of this low gradient section of the Colca drainage is under the 250 million m3Condoroma reservoir, a product of the 1970s Majes Hydroengineering Project that, combined with water contributed by tunnel from the much larger Angostura reservoir in the Apurimac drainage, has resulted in a sustained year-around flow in the Colca between here and the Tuti diversion dam (Gelles 2000;Maos 1985). At Huinco, the river turns abruptly to the south-west where it continues to descend as a gentle channel, cutting through the Miocene lava formation known as Tacaza as well as wind-scoured river terraces. Immediately before entering the northern edge of Block 3 of our study area, the river cuts through strata of uplifted Cretaceous limestone striking west-north-west and forming outcrops and rock shelters on the edges of the upper river terraces (Figure 4-15). Downstream at the confluence with the Llapa River the Colca River returns again to the Tacaza formation for five more kilometers before entering the Inca formation, an andesitic exogenous dome dating to the Middle Pleistocene. This Quaternary formation fills the upper Colca valley from this point downstream to near the town of Coporaque and Yanque where fluvial conglomerates overlie it as a result of natural damming of the river by mudslides.

In its descent, 5 km below Sibayo, the incising of the Colca River begins immediately upon entering the recent exogenous dome of the Inca formation, an the river remains incised until it exits the Colca canyon approximately 100 km downstream. The Llapa River, which joins the Colca from river-left at Sibayo, emerges from the sandy tuff layers known as the Castillo de Callallituff formations (Noble, et al. 2003) and the Chalhuanca rhyolite dome fields to the south-east of Callalli. The entrenched Colca River descends more rapidly upon entering the Inca formation approximately 10 km downstream with perennial tributary streams entering primarily from glacierized stratovolcanoes that ring the Colca Valley. The first of these tributary streams are the creeks entering from the north upstream of Tuti draining the southern and eastern flanks of the Nevado Mismi volcano, while streams that form just on the other side of Mismi have recently been confirmed by a National Geographic expedition in 2000 to be the source of the Amazon. Immediately upstream of the town of Tuti a diversion dam was constructed across the Colca where the river is entrenched in a ravine 50m deep. At this point water collected by the Majes Project is diverted into a system of canals and tunnels along the south bank of the Colca that finally crosses into the fertile Pampa de Majes near the Pacific coast.

Below Tuti, the Colca river channel is deeply incised into Quaternary exogenous domes. Continuing downstream, the river turns west, again becomes entrenched, and begins dropping more steeply in the vicinity of Chivay. The geology, geomorphology and soils of the main Colca valley have been well-studied in the past several decades as a result of the "Colca Valley Terrace Project" organized by William Denevan (1986;1988;Denevan, et al. 1987) and more recent reviews of research in the main Colca valley can be found in dissertations by Wernke (2003: 34-66) and Brooks (1998: 57-84). In its steep westward drop to the Pacific littoral the river subsequently enters the 3,270m deep Colca canyon, the third deepest canyon in the world after the Yarlung Tsangpo (Tibet) and the Cotahuasi (Perú). Here the river cuts through folded sedimentary and metamorphosed layers predominantly belonging to the Yura group (Jurassic and Cretaceous) until it emerges as the Majes River on to the large alluvial plain that leads to the sea.

Figure 4-14. Legend showing geological map units in maps that follow.


Figure 4-15. Geological map units and 2003 Survey Block boundaries (in gray) for the Upper Colca project study area. Legend shown in preceding figure.


Figure 4-16. Geological map units shown on ASTER scene from 28 Sept 2000; legend is shown in Figure 4-14. In general, red pixels show areas of photosynthesizing vegetation (bofedales).

4.3.2. Yura and associated sedimentary strata

Starting with the oldest rock groups in the study region, the Yura group, the geological history of the Upper Colca region will be summarized through to the Holocene. The Jurassic and Cretaceous (> 66.4 Ma) sedimentary strata in the region include quartzite, shale to sandstone, dolomite, and limestone. The Yura sedimentary formation is exposed on both sides of the Chivay-Arequipa road around the Sumbay junction and to the 10 km to the east of the Llapa and Pulpera confluence. Quartzite outcroppings of the Yura formation appear to have provided material for the abundant artifacts made of quartzite observed in archaeological sites in the Callalli area. These quartzite Yura strata do not appear as cartographic units on the 1985 INGEMMET map (Ellison and Cruz 1985), but Parodi (1987: 47) notes that fine-grained quartzite outcrops occur west of Callalli and these features were encountered in recent fieldwork (JKYu west of Callalli on Figure 4-15 ). Quartzite outcrops form in metamorphic sandstone and occur in the oldest metamorphic strata in the Colca region. Similarly, chert, chalcedony, and quartz precipitate from diatoms in sedimentary contexts (Andrefsky 1998: 51-56;Luedtke 1994), and consequently exposures of these materials are found in the Mesozoic strata in the region and in cobbles form in many stream beds. Interestingly, cherts were noted in the Ichocollo creek in Block 6 of the survey, but an examination of the headwaters on the Cailloma (31-s) geology map (Davila M. 1988) reveals no layers older than the Tertiary in that watershed.

Steeply uplifted Cretaceous formations appear in two portions of the study area (Palacios, et al. 1993: 28-30, 36) and these formations, in addition to nodules found in riverbeds, may have provided the local sources of non-obsidian toolstone in the area. The thick calcareous Yura sedimentary exposures north of Sibayo are dramatic examples of uplift of these Cretaceous strata. Additionally, the slopes just east of Chivay below 4000m between the town of Canocota, close to Calera hot springs, and as far south as the Quebrada de los Molinos consist of Murco and Hualhuani formation sedimentary rocks that include siltstone, quartz-rich sandstone, and limestone layers. These areas, in addition to cobbles encountered in riverbeds, may have represented the local non-igneous sources of material for stone tool production that include chert and chalcedony.

4.3.3. Oligocene and Miocene lavas

Oligocene and Miocene lavas

Tacaza Group

Flows belonging to the Tacaza group are found throughout the south-central Andes, however the only portion that appears in the Colca region belong to the older Tacaza with dates in the range of 30.21±0.73 Ma and 26.51±0.6 Ma. At 1900m thickness these Colca lava flows are the thickest Tacaza layers in the larger region (Palacios, et al. 1993: 86).


Figure 4-17. Andesitic Tacaza deposits with breccias and tuff outcrops in the Quebrada de los Molinos drainage. The Chivay obsidian source in later Barroso deposits is found high above, on the right side of the photo.

In the Colca area, these deep deposits of lavas and breccias consist of andesites and trachybasalts (containing higher feldspar content) intercalated with tuff bands. The Tacaza layer appears predominantly on the western half of our survey zone.

Pichu Formation ash flows and the Castillo de Callalli

On the eastern side of the study area, the Tacaza formation is overlain with the Pichu formation consisting of sandy tuffs and white ignimbrites. Among these Miocene ignimbrites layers is the Castillo de Callalli formation, an ash-flow tuff that rises dramatically from Llapa river just upstream of Callalli and is a principal landmark in the Upper Colca region. In the INGEMMET study (Ellison and Cruz 1985) the Castillo de Callalli was assigned to the Pichu formation. This landmark is an approximately 400m hill of silicic ash-flow tuff ranging from densely welded to non-welded tuff. The layering in this formation has recently been subject to a more detailed study involving isotopic dating and phenocryst mineralogy (Noble, et al. 2003). The recent work shows that this formation is not a single stratigraphic unit, as presented in the INGEMMET study, but rather it consists of two layers separated by 16 Ma.


Figure 4-18. The lower section of the Castillo de Callalli is known as "Cabeza de León". Evidence of an LIP pukara was encountered on the summit [A03-935].

The lower part of Castillo de Callalli is adjacent to the main road to Callalli and follows the Llapa River. The Cueva de Quelkata, a rockshelter with a predominantly Terminal Archaic component (Chávez 1978) that was dynamited by the Majes Project road construction, is at the base of this formation. This lower section has well developed columnar jointing and is a densely welded, devitrified ash-flow tuff, and K-Ar dating on phenocrystic hornblende indicates that this lower flow is 20.7±0.6 Ma (Noble, et al. 2003: 33). The upper section is described as "partly welded vapor phase crystallized tuff with the physical characteristics of the distal part of an outflow sheet". Phenocrystic sanidine from this section yielded an40Ar-39Ar age determination of 4.72±0.02 Ma (Noble, et al. 2003: 35). The study suggests that the upper part of the Castillo de Callalli formation is associated with the Cailloma caldera to the north which erupted three times during the Pliocene. Further discussion of Miocene volcanism in the Orcopampa area of the Chila cordillera, between the Chivay and Alca obsidian sources, can be found in Swanson's (1998) geology dissertation.

4.3.4. Pliocene lavas - Barroso group

Barroso lavas in the Colca area include predominantly andesitic and trachyandesitic flows covering an area of approximately 320 km2. In the case of the Barroso group, the emplacements are contiguous lavas that occur as transversal flows, as andesitic domes, and occasionally as rhyolitic domes. At the Chivay source, two vents dating to the Barroso formation occur at Cerro Ancachita and Cerro Hornillo, and at the highest peak in the Centro Huarancante formation named Nevado Huarancante, to the south. Transversal flows and crests emanating from these flows created adjacent peaks such as Cerro Saylluta and Cerro Llallahue. Lavas, silicic coulee flows, and viscous volcanic breccias flowed from these vents and traveled up to 15km, into the Block 3 study area to the east. Curiously, in the Cailloma quadrangle study, immediately north of Chivay, Barroso group volcanism in the Cailloma caldera was dated to the Pleistocene rather than the Pliocene epoch (Davila M. 1988). The important distinction is that recent dating of Cailloma caldera deposits (Noble, et al. 2003: 35) appear to pre-date the Barroso group flows that are responsible for tool-quality obsidian formation in the western Cordillera.

At the Chivay obsidian source, Barroso group flows are superimposed on Tacaza levels, and both groups have been eroded and incised by later fluvial and glacial erosion. It is proposed by Burger et al. (1998: 205) that obsidian occurs at the Chivay obsidian source where silica rich magma from Barroso eruptions cooled rapidly when the flows contacted the older Tacaza group deposits. The emplacement of Barroso group obsidian flows will be discussed in more detail below.


Figure 4-19. Detail of Chivay source, Maymeja area with INGEMMET geological map units shown on ASTER scene from Sept 2000. Contact between TTa and TBa on the west appears offset and likely conforms to the horseshoe shaped valley. Yellow arrow shows direction of glacial striations.

Origins of the Maymeja volcanic depression

The Maymeja area is a depression surrounded by transversal flows and domes that appear to have been heavily glaciated in the Pleistocene epoch. The Maymeja depression contains certain features that resemble those of a volcanic caldera resulting from eruption-induced subsidence and collapse. However, further consultation with volcanologists indicates that the Maymeja depression is likely nota caldera. The characteristics that do suggest that Maymeja is a caldera include: a circular, steep-walled perimeter, occasional ignimbritic deposits, the Anchachita and Hornillo vents located along the margins, and remnant vent-like features in the center of the Maymeja area (Fisher and Schmincke 1984: 360;Karátson, et al. 1999;Szakács and Ort 2001). However, the small size (2 km diameter), an irregular southern and breached western margin, and overall paucity of ignimbritic materials in the region suggest, rather, that the margins of this area were defined by highly viscous rhyolitic lava flows from Ancachita and Hornillo that were subsequently eroded into the circular form of a cirque, particularly on the south-facing (heavily glaciated) slopes, as a result of abundant Pleistocene glaciation. An example of a large Pliocene caldera is the Cailloma caldera that dominates the Cailloma quadrangle immediately to the northwest of the Chivay area (Davila M. 1988). Rather, the Maymeja area can be more generally described as a volcanic depression that underwent significant glaciation during periods subsequent its Pliocene formation.

4.3.5. Pleistocene - Ampato group

Pleistocene - Ampato group

Pleistocene Inca Formation

Named for the site of the Inka bridge over the Colca adjacent to Chivay, the Inca formation consists of Andesite and Trachyandesites that occur in exogenous domes and flows that appear to emanate from the north side of the Colca river, just north of the town of Chivay. These domes are composed of andesites and trachyandesites marked by a high percentages of alkali feldspars (Palacios, et al. 1993). Two Potassium Argon dates from these flows by Sandor (1992: 232-235) indicate that they formed during the Middle to Late Pleistocene (64,000 ± 14,000 bp and 172,000 ± 14,000 bp).

4.3.6. Holocene stratovolcanoes

Arequipa is a volcanically active region with a number of stratovolcanoes that have erupted repeated over the past 10,000 years. These peaks include Huayna Putina, Misti, Sabancaya, and Ubinas (Gerbe and Thouret 2004;Thouret, et al. 2002;Thouret, et al. 2002;Thouret, et al. 2001;Thouret, et al. 2005). The regular deposition of tephra from these peaks provides consistent strata that may aid in archaeological excavation work. Archaeologists working in the western cordillera will benefit from the tephrachronology sequences currently being developed by volcanologists in the region.

4.3.7. Glaciation

In modern times glaciers generally occur above 5,000 - 5,200 masl in Arequipa (Clapperton 1993;Dornbusch 1998;Fox and Bloom 1994), with differences in precipitation being the single largest contributor to variation in snowline altitude between the eastern and western cordillera. During the Last Local Glacial Maximum (LLGM) remote sensing studies of glaciated landforms suggest that there was a regional snowline depression of 600 - 800m in the western cordillera region of Arequipa during the Late Pleistocene (Clapperton 1993;Klein, et al. 1999). However glaciological studies show that the response of snowline to aridity is not uniform across the region, and that "as snowline rises in response to increasing aridity, it becomes less sensitive to temperature perturbations" (Klein, et al. 1999: 81). Recent evidence from ice and lake core studies in central and southern Peru (Smith, et al. 2005) have shown that the LLGM occurred in the tropical Andes around 21,000 cal years ago or over 10,000 years before uncontested evidence of human presence in South America.

The extent of glaciation during the Terminal Pleistocene and Early Holocene in the Colca region is of direct interest to this study of the Chivay source because the Maymeja source area itself was potentially glaciated into the Holocene epoch, and glacial geomorphology appears to have eroded high altitude obsidian deposits like Chivay. Lake and glacial core studies, as well as radiocarbon dates on vegetative material in deglaciated areas, indicate that despite the evidence for glacial advance during the Late Glacial, aka the "Younger Dryas" (9550 - 10,850 cal BCE) in the northern hemisphere, the glaciers of the tropical Andes appear to have retreated during this period (Rodbell and Seltzer 2000: 335;Seltzer, et al. 2002). The evidence suggests that the cooler temperatures were associated with a decline in precipitation, and that this precipitation decline resulted in glacial retreat.

Glaciation of the Chivay source area

These regional data on snowlines are corroborated by evidence of terminal glacial moraines in the Quebrada de los Molinos at 4400 masl and in the adjacent Quebrada Escalera at 4300 masl. On the east side of our study area glacial moraines were observed just west of the Ventanas del Colca feature on the south and east end of a dramatic U-shaped valley at 4350 masl on Quebrada Porhuayo Mayo. The INGEMMET map series (Palacios, et al. 1993) indicates that morainal deposits of silt, sand, and gravel are evident elsewhere in the study area (Figure 4-15), typically at or above 4300 masl, corroborating the evidence from the regional model with a local snowline during the LLGM at 4300 - 4400 masl, or approximately 700 meters lower than conditions evident in 2003.

A team from the University of Maine including professors Daniel Sandweiss and Harold Borns explored the question of the Early Holocene deglaciation of the Chivay Source area in the late 1990s. As shown on Figure 4-15, a10Be date of circa 10,00014C BP (9450 cal BCE) was acquired from a quartzite erratic on a moraine at 4650 masl to the east of the Chivay source area (data courtesy of Daniel Sandweiss, 2006). This sample suggests that the terrain surrounding the source was glaciated at this elevation and higher as late as the Early Holocene. Establishing the rate of deglaciation and the exposed areas at a given time period will require further glaciological study.

The evidence of glaciation in the Maymeja area of the Chivay source area is pronounced. Glacial erosion is evident on the south-facing slopes of the northern part of the Maymeja depression, as is expected in the southern hemisphere. The south and south-western slopes of Ancachita peak are steep and unstable, and at the base of this slope is a recessed glacial tarn that appears to retain water during the wet season. A moraine blocks the exit of this tarn feature, but on the slopes below lateral moraines parallel the path of the glacial tongue descending from below Ancachita. In the most deeply eroded part of this northern area contains the only continuous surface flow of obsidian encountered in the entire study area: the Q02-1 source which contained vertical, subparallel fractures and was unsuitable for tool production. Other effects of glaciation on obsidian distributions include the presence of transported obsidian nodules in parallel and terminal moraines in the Quebrada de los Molinos.


Figure 4-20. Glacial polish and striations (aligned towards camera) on lava flows adjacent to Maymeja workshop on the southern end of the Maymeja area.

In the southern portion of Maymeja a lateral moraine is similarly visible, and the striated and polished benches of lava dramatically attest to the extent of glaciation in the area (see yellow arrow on Figure 4-19). The direction of striation on the these lava flows is consistently south-west or dropping towards the Molinos drainage, and striations persist on high exposed benches suggesting that the glaciers were large as they were striating rocks over twenty meters above the base of the Maymeja area.

During the Middle and Late Holocene glaciers in the south-central Andes, on the whole, have retreated. Glaciological studies conducted in southern Peru and western Bolivia show that retreat is most notable during the time range from circa 10,000 to 3,000 BP (circa 9,000 - 1,000 cal BCE) (Clapperton 1993: 464-466), as well as in last decades of the twentieth century. There is evidence for small advances in glacial extent since 1000 cal BCE, most notably the Little Ice Age circa AD1500-1850. Currently Ancachita is slightly below an altitude permitting glaciation to flow downslope into the Maymeja area, though 5100 masl elevation is glaciated in drainages on peaks like Nevado Sara Sara with large glaciated expanses in the high altitude accumulation zone, and areas of peaks on the eastern side of the Andes like Carabaya (northern Puno) are currently glaciated as low as 4900 masl (Dornbusch 1998). The possible effects of this glaciation on obsidian exposure and weathering in the Chivay area will be discussed in greater detail below.

4.4. How obsidian is formed

Even in volcanically active regions of the world, the geological formation of high quality obsidian is a relatively rare event in nature because a number of features must co-occur for volcanic magma to become tool-quality natural glass. The following description was developed largely from Shackley (2005: 10-15, 189) and Fink and Manley (1987). Obsidian can form when rhyolitic magmas are extruded and quenched in the course of a volcanic eruption. Rhyolitic magmas are silica-rich, acidic melts that are capable of flowing as viscous lavas. As rhyolitic magmas approach the earth's surface, the high water content (up to 10% H2O) begins to escape as vapor, changing the viscosity and the cooling rate of the flow, and resulting in a very low presence of water in obsidian. When water remains trapped in obsidian it sometimes forms bubbles of water vapor, reducing the homogeneity and fracture quality of obsidian for tool production. Fink (1983) found that obsidian emplacement tends to proceed along the following sequence associated with the eruption: (1) tephra fall-out from the initial explosive eruption, (2) basal lava breccia, (3) coarsely vesicular pumice, (4) the principal obsidian flow, (5) finely vesicular pumice, and (6) surface breccia. The best quality obsidian for tool-production often occurs not on the ground surface but slightly underground in subsurface emplacements around a volcanic vent where degassed magma squeezes into rock fractures free of dirt and ash particles. Obsidian over 20 million years old is rarely useable for tool production because, as a geologically unstable material, obsidian gradually devitrifies from a glass into a rock (Francis and Oppenheimer 2004: 163).



Compare with


Rhyolite (Felsic)

As an intrusive rock it is granite

Silica content

Rhyolite is usually >70% wt SiO2

Basalt: <52%, Andesite: 53-63% wt SiO2

Water content

Obsidian: 0.1 - 0.5 H20

Perlite: 3-4%, Pitchstone: 4-10% H20


Quality obsidian is usually <20 Ma

Obsidian >66.4 Ma (KT boundary) is devitrified.


Obsidian: 5.0 - 5.5

Quartz: 7.0

Specific gravity

Obsidian: 2.6 (2600 kg/m3)

Pumice: 0.64, Water: 1.0, Basalt (solid): 3.0

Compressive strength

Obsidian: 0.15

Chert: 100 - 300

Table 4-4. Characteristics of Obsidian.

During the extrusion of rhyolitic lavas it is the supercooling (instantaneous quenching) of the lava that creates obsidian, an atomically disordered natural glass with the structural properties of non-flowing liquid. This lack of crystalline structure in aphyric obsidian results in an isotropic lithic material with excellent flaking properties and the potential for extremely sharp edges because it has no prevailing fracture direction and it fractures at the molecular level. Obsidian has a low specific gravity as it is acidic, it also lacks crystalline structure, and it has relatively low hardness. Obsidian has high tensile strength but it has extremely low compressive strength and, combined with the non-crystalline structure, the result is implements with relatively brittle characteristics and fragile working edges (Hughes 1998: 367;Luedtke 1994: 93;Obsidian 2006;Speth 1972: 52). The cortex ofobsidian from primary deposits can visually vary widely depending on the context ofemplacement and weathering processes. Obsidian flows that cool where tephra is present can melt a thin layer of the adjacent ash and the fused material appears as a thicker cortex (Figure 4-21).

4.4.1. Chemical differentiation

An important attribute of obsidian for archaeological investigation is that obsidian flows are chemically distinctive allowing artifacts to be chemically linked to their geological source areas. These chemical differences in obsidian are the result of certain elements crystallizing to solids and being removed from the magma as per the Bowen reaction series, resulting in a distinctive geochemistry for lava from most magma chambers and sometimes for each extruded lava flow. Prior to, and during, a volcanic eruption, magma evolves as changes in temperature and pressure causes chemical differentiation and leads certain minerals to crystallize and settle out of the melt.

As magma evolves, further melting and crystallization change the nature of the solids, and crystals that accept the incompatible elements may form in the liquid. Some feldspars, for instance, are good hosts for strontium, as is mica for rubidium. "Evolved" obsidian magmas may contain these crystal "hosts," and the ratio of a given element between the liquid and solid phases will change dramatically. Changes of this kind issue a particular chemical character to a given obsidian…The result of these processes is that the incompatible-element mix of a given obsidian source varies from any other and becomes a sensitive indicator of origin (Shackley 2005: 10-11).

This process creates detectable chemical differences in obsidian that permits methods such as X-ray florescence (XRF), Instrumental Neutron Activation analysis (INAA), and various types of inductively coupled plasma mass spectrometry (ICP-MS), and Proton-induced X-ray emission-proton-induced gamma ray emission (PIXE-PIGME) to chemically characterize the material (Neff and Glascock 1995;Shackley 1998).

4.4.2. Obsidian color

The color of obsidian is most often black but it also occurs in red, brown, bronze, purple, blue, green, gray, silver, clear, as well as with banding that includes some of the colors listed above. Obsidian coloration results from the oxidation state of tiny crystals that occur in the melt (Volcano Hazards Program 2000). The black color that is common in obsidian is the result of tiny (< .005 mm) magnetite (iron oxide) crystals, red is usually from hematite present in highly oxidized obsidian, and green results from variations in iron oxidation. Microscopic crystals of various types of feldspars may yield the unique blue, purple, green or bronze colors associated with "rainbow obsidian". Banding results from the folding-in of an oxidized flow surface as the lava continues to move, with each colored streak perhaps reflecting the individual pulses in the obsidian eruption. Gold and silver sheen obsidian is argued to be caused by bubbles of water vapor trapped in the glass that are stretched nearly flat along flow layers (Obsidian 2006). Given the unusual visual qualities of obsidian, the color and banding in a particular nodule are characteristics likely to have influenced human use of the material.

4.5. Pliocene (Barroso group) obsidian in the Colca valley

The obsidian that was widely used in the prehispanic central Andes occurs in Tertiary flows along the western cordillera. In the main part of the Colca valley, obsidian is found in the lower levels of Late Tertiary lava flows on both the north and south sides of the Colca river. East of the town of Chivay, the Chivay obsidian type has been observed to occur where Barroso flows contact the older Tacaza deposits (Burger, et al. 1998: 205). Barroso flows (TBa), clearly evident in photos shown in Figure 4-11 and Figure 4-15, extend atop Tacaza flows to the east of Chivay, and obsidian has been observed where layers described as "Tertiary intrusive, porphorytic" (T-po) extend atop Tacaza flows to the northwest of Coporaque (Figure 4-23). A fission track date on an obsidian sample collected east of Chivay confirms the Pliocene origins of these flows with a result of 3.52 ± 0.15 Ma (Poupeau and Labrin, 10 Oct 2006, pers. comm.). To date, research has shown that high quality obsidian flows in the Colca valley that occurred during the late Tertiary fall into two chemical groups: (1) the Chivay type found to the east of the town of the Chivay, and (2) the "Uyo Uyo" type that lies across the Colca River, west of the town of Coporaque. In the course of fieldwork in 2003 the Upper Colca project sampled obsidian from both locations and had results analyzed by the Missouri University Research Reactor, as will be described below.

4.5.1. Chivay obsidian source observations

Obsidian at the Chivay source was observed obsidian in natural contexts eroding from the base of what is possibly a collapsed rhyolitic dome (Cerro Hornillo), and from the south-west flank of Cerro Ancachita between the elevations of 4900 and 5000 masl. In the majority of locales, obsidian appears as concentrations of cobbles in a pumaceous rhyolite soil matrix where unconsolidated outcrops seem to occur as jointed and weathered flow bands strike the surface. A similar context is described by Healan (1997: 84) at Ucareo, a central Mexican obsidian source, where he notes that unconsolidated outcrops are not in-situ, but such outcrops are best considered as "primary features" because they have not undergone lateral movement. Cobbles from these outcrops often have a very thin cortex at the Chivay source.

The only consolidated obsidian flow to strike the surface in the Maymeja area is finely jointed in the vertical direction, offering fragmented primary material that is poorly suited for obsidian tool making. It is notable that this flow is exposed in a gully in the northern portion of the Maymeja depression where glacial erosion is most pronounced and the bed of a small glacial tarn, forming only during the wet season, is located nearby.



Figure 4-21(a). Small box in lower-right gully shows Q02-1, an obsidian flow eroding out of ashy-pumaceous soils below western arm of Cerro Ancachita. (b). This obsidian is of limited use for tool making because it contains vertical, subparallel fractures.

On the southern half of Maymeja, a heavily exploited unconsolidated outcrop strikes the surface at the principal quarry pit (Q02-2), but only small nodules (5-10 cm long nodules, and a rare piece up to 15 cm long) remain. In these areas several quarry pit features, one large pit measuring 4 x 5m and 1.5 m in depth (Section 7.4.1), and two shallow pits measuring approximately 1-2 m in diameter that are possibly modern were observed further down the ridge. The larger quarry pit is located on a slope and on the downslope side lies a debris pile made up of primarily small, non-cultural (unmodified) nodules of obsidian, but with the occasional flake or retouched flake. Quarry pits surrounded by discard piles have been termed "doughnut quarries" by Healan (1997: 86-87) describing the Ucareo source in central Mexico where such pits have been found in great abundance. Variously sized quarry pits have also been described by researchers at other obsidian sources in central Mexico (Darras 1999: 80-84;Pastrana 1998).

Nodules at the Q02-2 quarry pit are found in two principal forms: a long, narrow form and a spherical nodule form. It is possible that the nodule forms reflects differences in emplacement, with the long, narrow nodules resulting from relatively thin flows while spherical nodules are unconsolidated outcrop forms. As will be discussed in Ch. 7 (Section 7.4.1), these nodule forms appear to have influenced knapping strategies as narrow nodules offer more angles and a different flaking geometry as compared with spherical nodules. Pastrana and Hirth (2003) describe reduction strategies for biface production that exploit long, narrow nodules at the Sierra de las Nevajas (Pachuca) source in central Mexico (Figure 7-2).

Chivay type obsidian outside of the Maymeja area

Elsewhere around the base of the dome Cerro Hornillo obsidian was encountered eroding from the ground in smaller nodules (2-5 cm long). These obsidian exposures are pronounced on the eastern and south-eastern slopes of Cerro Hornillo around 4900 masl where scatters of subangular pebbles and cobbles, or angular shattered felsenmeer carpets of these small obsidian pieces, were encountered. In glacially eroded areas and along wind-scoured ridges these obsidian surfaces occur as lag gravels where finer soil has been transported away by Aeolian processes, leaving only obsidian nodules. These nodules appear to be weathering from rhyolitic flows and the tool-making quality of the raw material seems to be compromised of three characteristics: (1) Size- remaining nodules were typically quite small; (2) fracture quality - heterogeneities in the material caused the material to fracture unpredictably; (3) visual quality - the nodules were often occluded with bubbles and ash particles.


Figure 4-22. Obsidian gravels exposure in tephra soils east of Cerro Hornillo.

In this study, obsidian containing heterogeneities due to the presence of bubbles or ash particles is termed "Ob2" obsidian, while homogeneous obsidian that was probably preferred for tool production in antiquity is referred to as "Ob1" obsidian.

Pulpera / Condorquiña flow

In the course of the Upper Colca Project research an obsidian source was encountered on the eastern toe of the Barroso lava flow where flows emanating from Cerro Ancachita and Cerro Hornillo terminate near the community of La Pulpera. According to the INGEMMET map (Ellison and Cruz 1985), this obsidian appears to have formed where silicic lava flows belonging to the Barroso group contacted Late Miocene ignimbrites from the Pichu formation and cooled rapidly leaving obsidian exposed by erosion in this area. This obsidian does not appear to be of tool-quality as the nodule size is small (<5cm) and it contains many heterogeneities that interfere with the fracture characteristics of the stone. Samples of this obsidian were sent to M. Glascock at the Missouri University Research Reactor in 2002 and the samples were determined to be of the same chemical group as Chivay (Glascock, pers. comm. 2002).

Obsidian Cortex

The cortex of nodules at the Chivay source is often remarkably thin. The spherical nodules, described above, seemed to be more closely affiliated with a very thin cortex that is under a millimeter in thickness where hydration appears only as a slight discoloration on an otherwise smooth external surface. In other cases, particularly on long and narrow nodules, a textured and raised, but sometimes rough and ropey, cortex is evident that was referred to as "tabular cortex".

This geologically derived variation in cortex is meaningful to archaeologists because when the cortex was thin it appears that it was sometimes left undisturbed on the faces of many preforms, but when the cortex was the rough or tabular type, it seems to have been a central obstacle to knapping. Cores partially covered with rough cortex were discarded after flakes were removed from the non-cortical face. One possible explanation for the intensified quarrying observed at the Q02-2 quarry pit is that nodules recovered in this area contained a high frequency of the thin type of cortex, a cortex type which would have represented less of an obstacle to knapping.

The cortex form can influence reduction strategies in important ways. First, if cortex is extremely thin then it does not pose a structural obstacle to knapping and the priority on reducing an item's weight by decorticating it close to the raw material source may be lessened. Second, the thick tabular cortex on one side of long, narrow nodules greatly limits the potential of these nodules unless the cortex can be removed effectively. We speculate that the origin of the very thin cortex is related to the glacial history of the obsidian source. These unconsolidated outcrops were likely to have been compressed and eroded by the presence of glaciers and the effect of this glaciation on obsidian outcrops may have served to further fragment and introduce water into the obsidian flow, which appears in the oldest specimens as a layer of perlite. As a consequence, the extremely thin cortex may have resulted from glacial erosion and moisture introduced during the Pleistocene, rather than from characteristics of the original quenching environment of the obsidian flow. Obsidian hydration dating may allow direct dating of the fracturing of the obsidian, however the unreliability of hydration dating in contexts of high temperature variation and unknown moisture levels (Ridings 1996) suggests that results from hydration rinds alone are probably of limited value.

Rare 6 Type distribution

In her dissertation, Brooks (1998: 443) notes that Glascock identified the Uyo Uyo samples as matching the "Rare 6 Type" obsidian that had been previously encountered in Burger's earlier work with Lawrence Berkeley Lab (Burger and Asaro 1977: 56). However, Glascock cautions that the calibrations are not perfect between the LBL and the MURR results, particularly with small sample sizes. In recent communications Glascock (2006, pers. comm.) is not confident that the Uyo Uyo and Rare 6 Type are the same type and he believes a re-analysis would be required to confirm it.

In Burger's earlier study he identified Rare 6 type from two projectile points in Cuzco and Puno (Burger, et al. 2000: 312-313). One sample was from the surface of the site of Chinchirmoqo that lies 2 km from Pomacanchi in the department of Cuzco. The other Rare 6 Type sample was a projectile point found on the surface of the site of Taraco on the north side of Lake Titicaca. These surface samples are difficult to assign to a specific time, though Burger et al. placed both samples in the "latter part of the Early Horizon and Early Intermediate Period", a period that roughly corresponds to the Middle to Late Formative using the chronology of the present project. Excavation work currently underway by Charles Stanish and colleagues at the site of Taraco may reveal additional Rare 6 Type obsidian artifacts.

4.6. Other Tertiary obsidian in the Colca valley

Other Barroso group deposits in the region include Pampa Finaya on the north side of the Colca River. In the course of his archaeological survey Steven Wernke (2003: 36, 39) surveyed Pampa Finaya and, unlike in the Chivay source area, obsidian was not found on the perimeter of this Barroso flow where it contacts Tacaza layers. However, Wernke identified an obsidian source to the west on Cerro Caracachi, a distinctive knob-shaped peak at the head of Quebrada Huancallpucy that is at a contact zone between a layer described by INGEMMET as an porphyritic intrusive material (T-po). This source was exposed in a distinct stratigraphic context and that the obsidian appeared to be of low quality for knapping as compared with the Chivay type.


Figure 4-23. Geological map units with Uyo Uyo sampling locations. See Figure 4-14 for legend. Selected archaeological sites in the main Colca Valley shown in blue.

On 30 Nov. 2003, I visited the area of Cerro Masita and Cerro Caracachi on the suggestion by Wernke (2003, pers. comm.) in order to collect geological samples, to determine whether higher quality obsidian is encountered on this dome, and to compare the deposits with what had been observed at the Chivay source (see Appendix B4). I approached Cerro Masita was approached from the road in the drainage north-west of Ichupampa and only observed small pieces of low quality obsidian near the summit of Caracachi (documented by Wernke) and on the eastern flanks of Masita. Analysis by at Missouri University Research Reactor (Glascock, 2005 pers. comm.) established that this material belongs to the "Uyo Uyo" chemical group first recognized in samples provided by Sarah Brooks (1998: 443-445). In 1993 Brooks collected modified and unmodified nodules from the site of "Uyo Uyo" near Coporaque that were analyzed at MURR in 1995. Brooks (1998: 443-445) writes that the geological origin of these nodules was not determined, but the presumption was that the source lay uphill from the site of Uyo Uyo, as there were many unmodified nodules at the site, an observation that was confirmed by these hikes. No large nodules were encountered, however, and it appears that the Chivay source east of Chivay remains the sole source of high quality obsidian in large nodule form in the Colca valley.

4.6.1. Tripcevich source sampling work

In the course of Upper Colca 2003 survey work geological samples of unmodified obsidian were collected from a variety of natural contexts throughout the Upper Colca study area as well as elsewhere in the Colca valley. In order to best characterize the elemental variability within a chemical type, Shackley (1998: 100-101) recommends collecting samples from throughout the primary and secondary deposition area in sufficiently high numbers to recognize sub-source variability and reduce the chances of mischaracterization.


Figure 4-24. Bivariate plot showing Dysprosium against Manganese for Tripcevich 2005 samples.


Figure 4-25. Map showing locations of Colca valley obsidian source samples analyzed by MURR in 2005.

On Figure 4-24 only samples NT0005, NT0008, and NT0015 fall within the Chivay ellipse giving the impression that there is chemical variability at the Chivay source. Bivariate plots across other elements, however, show that all of the samples fall squarely within the Chivay ellipse.













Condorquiña source near Pulpera









Condorquiña source near Pulpera









Condorquiña source near Pulpera









East of Cerro Hornillo









Maymeja SE rim along trail departing area at 140°









Quarry pit Q02-2, A03-219


















Chivay, A03-570. Cortex battered, poss. from glaciers.









Chivay, northern Maymeja









Chivay, northern Maymeja









On western shoulder of Ancachita with ash sample

















Quarry pit Q02-2, A03-219























Uyo Uyo


W. side of Cerro Caracachi







Uyo Uyo


East slope of Cerro Masita









Artifact from Yana Salla, Llalli









Near small rock shelter by top of landslide









From across valley near old canal









From across valley near old canal

















Above Ayawasi






Table 4-5. Peruvian obsidian source samples submitted to MURR by Tripcevich in 2002 and 2005. Coordinate datum is WGS84.

Three samples were provided from the low-quality Condorquiña source in the Pulpera area that were collected in 2002 in order to evaluate the chemical variability within the Chivay source. While the quality was poor, this characterization demonstrates further association between that the Chivay obsidian chemical type and the Pliocene Barroso lava flows evident in the Colca valley. One artifactual sample that turned out to be Chivay was from the town of Llalli close to Ayaviri, in Puno.

Additional samples were submitted from the Alca source collected during fieldwork in 2001 from an exposure described by Justin Jennings (Pers comm. 2001) to the west and south of Cerro Aycano. Kurt Rademaker and colleagues (2004) have since encountered much larger obsidian exposures to the east of this area, on moraines located on the northern slopes of Nevado Firura at approximately 4700 masl.

4.7. Conclusion

The geography and geology of Chivay obsidian deposits conditioned the human use of this source in prehistory. Spatial relationships around the source were reviewed here in terms of three primary contrasts. First, on a regional scale, the location of the Chivay source above the rich and productive Colca valley meant that the source was only a few hours from communities residing in the main valley, but that it was also accessible to puna residents, and herders and caravan drivers as the economy based on camelids expanded. Given the inefficiency of pure pastoralism, obsidian exchange was likely part of a larger pattern of sustained contact between herders and agriculturalists. Second, exploitation of particular sources of obsidian over others was probably limited by local conditions, as water is much more available in the Maymeja area than elsewhere around Cerro Hornillo. Finally, the quality of obsidian varies due to the formation and erosion contexts of Chivay obsidian in prehistory. The Q02-2 quarry pit appears to have been the source of obsidian that predominantly consisted of large nodules of homogeneous glass with a relatively thin and inobtrusive cortex, permitting efficient and predictable quarrying and production in prehistory. As will be discussed in the chapters that follow, these geographical factors were influential in the archaeological use of the Chivay obsidian source area as was documented in the course of the 2003 research project.

Ch5 Methods

Research Methods and Data Recording Strategies

5.1. Introduction

This chapter reviews methods and equipment used in the course of the Upper Colca Archaeological Research Project, and it describes the creation of indices and measures that are used in analyses in subsequent chapters. Research at the Chivay obsidian quarry presented two principal challenges. First, the obsidian source itself is situated in a rugged, high altitude landscape that required working out of backpacker campsites and involved careful decisions regarding time budgeting during survey and excavation at the source. Second, raw material source areas present a challenge in the sheer volume of archaeological materials that are typically found in these areas. Sources require a modification of the established "site-oriented" survey paradigm for mapping and sampling during fieldwork, and for analysis during lab work. Anadditional goal of this project was to implement mobile GIS to record archaeological distributions in a digital format that integrates easily with a GIS database, and at a finer scale of resolution than is possible using traditional archaeological survey methods.

This chapter will review the research methods used both in the field during survey and excavation work, and in the subsequent lab analysis. Research design required an explicit selection of survey methods and these will be discussed below. Laboratory analysis of artifact collections was broad such that, for example, simple flakes were analyzed with nearly the detail of projectile points. This expansive analysis strategy required explicit sampling methods so that detailed analysis took place on representative samples from across geographical space and across artifact types, as sampling reduced the total count of artifacts requiring detailed analysis. Collections from survey and excavation were analyzed in two stages, (1) basic sorting and weighing of all collections, and (2) detailed analysis of selected artifacts from the larger population. The integration of field and laboratory digital data permitted the production of detailed summaries promptly for a Peruvian government report, and it allowed for the integration of GIS spatial analysis tools with the detailed data of lab analyses.

5.1.1. Locus rather than Site-oriented survey methods

Raw material sources are archaeologically complex features because, as foci for ancient procurement, there are frequently a large number of overlapping palimpsest occupations. A siteless survey approach is theoretically compelling (Binford 1992;Dunnell 1992;Dunnell and Dancey 1983;Ebert 1992;Foley 1981;Thomas 1975), however in practice most projects must balance detailed mapping against expediency and recording speed, as will be described below. The theoretical aims of this research were to detect and record meaningful variability in prehispanic artifacts and features throughout the study area, but also to focus on concentrations of lithic reduction activity and the variable material types that were evident in lithic scatters. The Upper Colca survey was not a "siteless survey" if that technique is taken to mean that the position of every artifact is recorded individually. Rather, it involved recording concentrations of non-diagnostic artifacts as loci in a mobile GIS system using a GPS polygon to delimit the loci (Tripcevich 2004;Tripcevich 2004). The result is a regional survey approach that is approximately as fast as traditional survey recording methods, but with much finer resolution and with greater detail on data that are more relevant to the theoretical goals of the research as determined by the field researchers.

5.1.2. Data recording in both categorical forms and field journals

The approach to data recording taken with this research project was essentially two-pronged where the mobile GIS archaeological recording system was complemented by more subjective notebook records written in a narrative form. These two complementary recording methods were employed, to some extent, in every data recording situation.

Method 1 - Digital forms with a spatial reference:The field crew aimed to record comparable and relatively objective data categories used digital forms linked to GPS-based spatial provenience in the mobile GIS data recording system. This system shares many of the advantages, and the limitations, of traditional, paper-based fieldwork forms.

Method 2 - Personal notebooks:Complementing Method 1, all team members had field notebooks where they were meant to record both data and more abstract observations on a daily basis at a variety of scales including the local, site-level scale, observations about regional patterns, and the relevance to larger goals of the project. Field notebooks may also take the form of digital audio recorded as one walks over a site.

While a two-pronged approach using both objective forms (traditionally, as paper forms on a clipboard) and field journals is not unusual, the distinction between the two systems is made more explicit by the use of mobile GIS. The digital recording system takes care of two aspects of archaeological feature and artifact recording: spatial positioning and data logging into comparable, form-based attributes. Conversely, the field notebooks fill the important role of capturing a range of other insights from fieldwork that are difficult and inefficient to capture in a digital form. In a notebook, observations and reflections on the contexts and patterns under study are more naturally recorded in narrative form, along with schematics, flow charts, or casual site sketches. The importance of reflection and observation has long been known to geographers and other field scientists.

There is at present enthusiasm for field mapping and their techniques… But map what and to what purpose? Is not this possibly another horn of the dilemma? ...Routine may bring the euphoria of daily accomplishment as filling in blank areas; the more energy goes into recording, the less is left for the interplay of observation and reflection (Sauer 1956).

Field notebooks provide an important contrast to the regular, systematized data notation of the mobile GIS system. Furthermore, many of the fieldworkers did not have access to the mobile GIS system, and as there was only one system in the 2003 Upper Colca team then only one person could be logging data into that system at a given moment. Months after the fieldwork was over, in the course of the analysis process, the various field notebooks provided daily observations, detailed notes on sites, and an overview that corroborated the geographically detailed and categorical organization of the mobile GIS database. Another valuable form of complementarity between the two systems is that one is geographical while the other is temporal in structure. Mobile GIS is inherently spatially focused, as it is a cumulative process of recording and improving spatial features with attributes and mapping detail, and when the GPS is activated the user is automatically viewing data layers in close proximity to the current location. Field notebooks, however, are temporally organized as they are logged chronologically as the field research progresses. Field notebook pages are easily photocopied and scanned into a file such as PDF, allowing the written logs can join the digital database albeit in an unsearchable form. Ultimately these two forms of data acquisition complement one another, and as mobile GIS increases in capability and in popularity, the dirt-smudged field journal will most likely continue to serve an important, material function.

5.2. Geographical context

As this research involves GIS analysis at regional, local, and intrasite levels of analysis, methodological issues related to spatial data organization will be described below.

5.2.1. Geographical datum and regional data sets

At the regional scale, data on obsidian consumption patterns in the central and south-central Andes were gathered from the original sources cited in summaries principally by Burger, et al.(2002;2000), as well as some more recent contributions from other sources. Regional studies of obsidian distributions require incorporating spatial data from a variety of maps and records. The bulk of these data consist of site locations and other archaeological phenomena, and these data derive from studies between 1960 and 2000 where spatial data in the Andean countries was referenced to an older coordinate system based on the Provisional South American Datum of 1956 - La Canoa, (PSAD56). Currently, the governments of these Andean countries are shifting their cartographic divisions to a modern datum based on World Geodetic System 1984 (WGS84) or the related Sistema de Referencia Geocéntrico para las Américas 2000(SIRGAS 2000) (Fortes, et al. 2006), a new continental reference system for South America. The Upper Colca Project functioned entirely in WGS84 in order to be consistent with global data sets such as satellite imagery and other forms of spatial data that have recently become available. The datum change from PSAD56 to WGS84 results in an incompatibility between historic datasets and recent work, when obsidian artifacts from older collections are chemically provenienced and the precise spatial origin of these artifacts can be difficult to ascertain. Ultimately all of these sites should be revisited and mapped with GPS using a modern system such as WGS84, but in the meantime a relocation of these site positions and transformations of the historic geographical data are required.

Spatial data in Peru and Bolivia are almost universally rely on the 1956 La Canoa (Venezuela) datum based on the International 1924 ellipsoid known as Provisional South American Datum (PSAD56). If data from PSAD56 and WGS84 data are inadvertently combined, the resulting misalignment of map features in the Arequipa area is approximately 300 to 700 m.



Semi-Major Axis (meters)


International 1924

Prov South American 1956



WGS 1984

WGS 1984



Table 5-1.The two reference ellipsoids used in Peruvian and Bolivian cartography(NIMA 1977).

Three parameter metric transformations that can be applied to UTM coordinates in the central Andes are shown in Table 5-2.


NIMA/NGA 1991 TR8350.2 report







(no error published)


PSAD56 ? WGS84

PSAD56 ? WGS84

PSAD56 ? WGS84

PSAD56 ? WGS84

?X (Eastings)

-279 m ± 6 m

-270 m ± 5 m

-288 m ± 17 m


?Y (Northings)

+175 m ± 8 m

+188 m ± 11 m

+175 m ± 27 m



-379 m ± 12 m

-388 m ± 14 m

-376 m ± 27 m


No. Satellites





ArcGIS Transformation

1208: PSAD_1956_


1202: PSAD_1956_


1201: PSAD_1956_



Table 5-2. Three parameter cartographic transformations for UTM coordinates from PSAD 1956 (La Canoa) to WGS 1984 (Dana 1998;Mugnier 2006: 496;NIMA 1977).

By using the WGS 1984 datum, the geographical data collected in the course of the Upper Colca Project research registered properly with newer spatial data from a variety of institutions such as international, US government, and private remote sensing sources. These datasets include global topographic data like SRTM, satellite imagery and DEM sources like ASTER, and these data are also consistent with web products like Google Earth. Furthermore, WGS84 is the native coordinate system of the Global Positioning System and therefore the dGPS data acquired during fieldwork in the Upper Colca did not require an additional geographical transformation. The majority of maps in the region will be released in WGS84 or SIRGAS 2000 in the coming years.

5.2.2. Regional datasets

Sources of spatial data were used in the regional analysis component of this project consist of digital raster and vector sources, and scanned paper-based maps. The first group include topographic datasets acquired from ASTER (Abrams, et al. 2002) and hole-filled SRTM (Jarvis, et al. 2006), and the second group are vector datasets acquired from Vector Smartmap and derivative datasets (NIMA 1995). Peruvian and Bolivian government maps were scanned from paper, georeferenced, and transformed (IGM 1986;Klinck and Palacios M. 1985;Palacios, et al. 1993) so that all now coincide with the WGS84 datum.

5.3. Data recording approach

5.3.1. Introduction

This research project encompassed three major zones: a river valley zone, a high punazone, and the obsidian source itself. A principal challenge in the archaeological evaluation of contrasting survey regions is the construction of meaningful categories that allow for comparison between these varied zones. Consequently, this project sought to strike a balance between the in-field assessment of sites and features based on the experience of archaeologists, and the significance and categorization provided by ancillary lab results and spatial analysis.

The 2003 field recording approach resulted in the presentation of data in this chapter that is a combination of two major forms of information. (1) Data that were recorded during survey work that described features and artifacts assessed by fieldworkers and quantified both in the field and in the lab, in the course of subsequent lab analysis. These categories are scalable, but the larger structure of the database is rigid in order to allow for comparison between different contexts. (2) Data were derived through inference and subjective assessment in the course of fieldwork by the project director (Tripcevich) and by five other experienced archaeologists who participated in segments of the fieldwork. The insights and notes of other project participants were also included into this subjective data category.

The result is a project that is integrated by using comparable quantitative measures, but that is informed by the subjective experience and interpretation of the archaeologists that conducted the fieldwork. The following data presentation is therefore based on quantitative measures, but the comments and assessments are informed by the insights from daily observations and personal notebooks.

5.3.2. Organization, sampling, and inference

Regional archaeological surveys must devise classification schemes that allow for meaningful comparison between features mapped throughout the survey area. In the Upper Colca Project, some artifact classes had radiating distributions that created special problems when devising comparable categories. This situation is best explained through an example that illustrates the challenges of consistent data recording.

The radial attenuation problem in artifact densities can be observed most dramatically in obsidian concentrations throughout the survey region as one departs from the Chivay source. In such situations, the human eye is easily misled by large concentrations of artifacts because the eye is attuned to the presence of contrasting or unusual materials, an issue that makes consistent sampling methods all the more important. In this example, the field crew would readily observe single flakes of obsidian far from the obsidian source, while at the source area itself a given obsidian flake was common place and obsidian densities were relatively deprioritized. Yet in the obsidian source area a flake of chertwas notable as it provided contrast. While these contrasts are meaningful: it is important that one might find a flake of chert at the obsidian quarry, far from the river where chert is usually found, such features should be mapped and observed separately with observations indicating that this is a-typical for the area. Ample effort should also be given to mapping features that aretypical to the area for faithful representation of general distributions.

This example illustrates a major methodological challenge for archaeological survey that exists in both conventional and digital recording systems. A systematic sampling strategy is the most cost effective way to describe the common features found in a region, because random samples can be extrapolated to the larger population. These distinctions in artifact identification correspond to two types of survey differentiated by Banning (2002: 27-38) as "statistical survey" and "prospection survey", as will be discussed in more detail below. Attributing the finds in the database based on the type of survey strategy that was employed permits a more consistent depiction of broader features of the landscape in later analysis.

5.3.3. Aggregating by Sites versus Loci

Spatially delimited "sites" were deprioritized in this project in an effort to capture changes in the continuous field of obsidian artifacts and related sites as one approaches the obsidian quarry workshop area. Rather than relying on the ill-defined concept of siteas the basic unit of analysis (Dunnell 1992), the Upper Colca Project survey team focused on recording lociof different artifact classes. Site boundaries were in fact recorded, however, because it was impractical to record isolates in the same detail as one recorded spatial structure in the concentrations conventionally thought of as a site.

For example, it was difficult to reconcile the density of the category High density lithic scatterbetween a workshop at the obsidian quarry, on one hand, and a residential base in the valley bottom; a problem that was resolved with sampling. Furthermore, pottery was almost non-existent in the lithic quarry area, even though on a regional scale there is ample evidence of consumption of obsidian by groups that possessed ceramic technology. In order to examine data in an integrated framework, broad categories such as "Site Type" were given minimal priority in favor of explicit and comparable categories based on features and on artifact concentrations described as loci. Broad site typecategories were assessed primarily for purposes of cartographic representation and to facilitate communication in the course of research, however analysis and interpretation focused on basic and comparable categories of data by artifact class.

5.3.4. Site and Loci recording structure

Archaeological sites and features were recorded primarily in terms of categories built around artifact classes while in the field. The site - isolate dichotomy is useful for expediency while doing fieldwork, but "Sites" and "Site Types" did not form the basic unit of analysis. Rather, in the course of survey, if a given group of artifacts was isolated or belonged to a "site", and if it was sufficiently large to be considered a site, then the recording strategy shifted to a more detailed analysis method.

When it came time for a site to be documented, attributes were recorded on numerous levels: on the level of the site, on the level of loci of three feature classes (lithics, ceramics, and structures), and finally on individual artifacts found in spatial association with that site. With these categories, the data and the lab results could be used to categorize the sites after-the-fact based on actual data and not only based on in-field intuition. In this manner, in-field impressions of archaeological features contributed to, but did not structure, the framework in which data was recorded. The structure provided by digital forms and GPS based mapping technology was complemented by interpretive notes and impressions that were written on field journals in daily narratives during fieldwork. The details of our Arcpad mobile GIS recording system are provided later in this chapter.

5.3.5. The Primary Key: ArchID

A stable, primary key ID number was assigned to all phenomena that were individually mapped. Every feature or artifact, including sites, loci, and individual artifacts that were mapped separately, were assigned their own "ArchID" number in a single number series regardless of the archaeological feature type. The ArchID primary key is a unique identifier integer that was value-free, as no further feature information was embedded in the number. For example, some archaeologists may classify sites by number range. In that system, rock shelter sites might be numbered between 100 and 200, and administrative structures might fall between 400 and 500, for example. However, that type of encoding of meaning into ID numbers is problematic for database design. The primary key approach adopted here is consistent with database normalization methods and the First Normal Form (Codd 1970) where one ensures that each table has a primary key that serves as minimal set of attributes that can uniquely identify a record. The First Normal Form further specifies that repeated fields be eliminated, and that each attribute must contain a single value and not a set of values. This kind of tabular organization is intuitive for those who have worked with computer databases, but the database normalization literature makes these features explicit.

The ArchID approach to numbering sites, loci, and artifacts in a single series is consistent with the low-interpretation field documentation system. The approach to survey provenience used in the Upper Colca survey is low-interpretation because an interpreted hierarchy is not encoded in the proveniencing system. For example, in some systems the Site ID# is primary, and structures and artifacts encountered inside that site are numerically subsumed by the site numbering, unless they are isolates. In other words, sites receive the principal numbering system, and any artifacts and features found "inside" sites receive index numbers from a secondary range that force the site assignment into the proveniencing of every artifact in that area. The weakness and spatial dependence of this system become more evident when features from different temporal occupations are recorded in a single, multicomponent site. In contrast, the upper Colca survey used a single number series so that the "site" assignment number did not intrude into the proveniencing of every feature inside the site, as features were mapped individually and thus were independent spatial entities.

The advantage to this approach, and to categorizing sites in later analysis rather than in the field, is that documentation and interpretation are distinct steps and data can be reinterpreted and individual loci reassigned to other time periods independently of the site context and spatial provenience in which they originally belonged. In other words, the GIS does the work of spatial provenience, proximity, and overlay, while numbering systems are dedicated only to the task of serving as a key for referencing records and tables in the database. Categories and types were used in this document for analysis, data presentation and summary, however, in the course of original data acquisition during fieldwork there was an explicit effort to document features based on simple artifact and feature characteristics rather than by a generalized typology or classification. There is no single file with all ArchID numbers represented, as they are distributed across the nine file types shown above in Figure 5-7a. However, in a post-fieldwork GIS processing step an "All_ArchID_Centroids" point file is created that serves as a single reference point for all ArchID numbers used throughout the season (see Section 5.10.1).

5.3.6. Site classification

Meaningful construction of site categories requires a combination of quantitative measures and qualitative assessment based taphonomy and other formation process issues. In the loose volcanic soils of the Upper Colca, relatively high erosion rates result from downslope movement and stream-channel migration, combined with wind deflation and seasonally-intense precipitation. For example, high density lithic loci are commonly found at the base of slopes due to the disturbance and erosion that are part of site formation processes (Rick 1976;Schiffer 1983). These artifact aggregations at the bottom of hill-slopes may appear to qualify as high-density loci when assessed quantitatively and through sampling, however a qualitative interpretation of the context reveals the formation processes at work.

In the Upper Colca project area anthropogenic effects, due to a high incidence of site reoccupation, include the direct and indirect effects of the later reuse of space. These include palimpsests, site maintenance and disturbance, interments, and reuse of construction materials for residential or mortuary structures. The indirect effects are primarily the result of intensive pastoral production that occurred in the Upper Colca in the last few millennia. These include erosion and trampling (although camelids have two digit pads instead of hoofs), and landscape modification for pastoral production such as corral construction and the modification of water distribution to enhance grazing opportunities at bofedales.

5.3.7. Linking Field and Lab data: an example

Every feature and artifact that was mapped individually received a unique identified key (ArchID number) that was used after the fieldwork was over to connect the GPS derived geographical location to associated attribute tables for recording numerical and text characteristics, and other field observations. The ArchID number links Arcpad GPS derived data to these other tabular data in a GIS through One-To-One or a One-To-Many relates. Collections were conducted during the course of fieldwork and in many of these spatial proveniences, a number of individual artifacts were collected and examined creating a One-To-Many relate situation.

An example would best illustrate this situation. In this example, a concentration of lithics is identified and it is mapped and described during fieldwork as lithic locus ArchID: 100. All field acquired data are linked through the number 100, including the polygon delimiting the concentration, environmental and cultural observations at the location, photograph numbers, the date and time of the mapping that is automatically logged. Bags of artifacts collected from that provenience, including ceramics, would receive the ArchID 100 spatial identifier. Furthermore the #100would be noted, or the range of numbers in that area, in the verbal field journal descriptions providing an explicit link between digital tables and interpretive description.

During a later phase of research, when bags of artifacts are opened and analyzed, an artifact-specific level of proveniencing occurs. A collection of artifacts that are spatially provenienced to a polygon: ArchID 100 results, however when it comes to time acquire detailed measurements on those artifacts, some of these artifacts should now be numbered individually. The solution in the Upper Colca Project was to create a catalog ID number, or "rotulo" in Spanish (RotID), such that each spatial provenience has lab numbers inside it numbered 1 to n.In practice, the bags would be tagged with a number and a decimal as inArchID.RotIDor100.15for the fifteenth artifact analyzed from spatial context 100, although digitally the two number series remain integers stored in separate fields. Lithics, ceramics, bone, and any other artifact class were stored together in a single second-level number series, yet in practice there was an attempt during lab work to keep all artifacts of a single class within a contiguous numbering range.

5.3.8. Theoretical relevance of the provenience system

These data recording issues are methodologically specific, but the issues have theoretical importance because later analyses are circumscribed by the units of analysis used in proveniencing. The above descriptions highlight an apparent contrast in the methodology:

Situation 1:Individually mapped and located artifacts and loci that are found at larger sites receive their own ArchID numbers. These are independent entities that may, or may not, belong to the larger site upon further interpretation. Importantly, the GPS derived coordinates store this spatial relationship and the numbering system is independent of spatial location.

Situation 2:Asin the example described previously, lithic locus ArchID 100 contains fifteen artifacts rotulonumbered RotID 1 through 15 that are analyzed and retained as 100.1 through 100.15. In this case, the artifacts are locked into their spatial container that is the ArchID number. While it would be ideal to have geographical positions for every artifact analyzed, it is not practical to spend so much time in the field and therefore, by necessity, the RotID is inside the ArchID spatial provenience. In other words, spatial provenience is the first level, while artifact provenience is the second level.

In sum, the object here is to allow the GIS to manage spatial information and use the ArchID number system not as a geographical hierarchy, but rather as a linking system to other forms of data be they attribute tables, artifact collections, or digital photos. On a theoretical level, when particular loci are subsumed within particular sites by the numbering strategy (as per the hierarchical system where loci are inferred to "belong" to sites), it is impossible to later extract the loci from within the site in the database because their numberings are inextricably linked. If later analysis reveals that the locus is likely a later occupation and not related to the site itself, there is no easy way to reverse the hierarchy in proveniencing. In the non-hierarchical proveniencing system used by the Upper Colca project, the location of that locus inside the site boundary is already conveyed in the GIS and that liberates the ArchID numbering system to serve as an effective primary key for the database.

5.4. Survey Strategy

5.4.1. Goals of Survey and Testing

The goals of the Upper Colca Archaeological Survey were to document the prehispanic use of the Chivay obsidian source and to record changes in obsidian processing evident at the source. Quarry research presents special challenges to archaeologists because prehistoric patterns are obscured by the sheer quantity of non-diagnostic materials from early reduction stages, and the compounded reuse of space over time (Ericson 1984;Torrence 1986). Working at the Chivay obsidian source involved additional challenges in its remote location at high altitude where roads and electrical sources were unavailable. The research design therefore had to maximize the time spent camping at the high altitude source using field methods that could be used effectively to detect variability in obsidian production at the source.

Very few archaeologists had visited the source area prior to this work, and therefore the research team had to accomplish basic documentation of the source area. However, detecting change in obsidian production required a relatively in-depth investigation, such as analyzing lithic production loci and excavating test units to acquire temporal control. Preliminary visits in 2001 and 2002 indicated that the rugged, high altitude terrain around the Chivay source precluded a systematic and extensive survey of contiguous lands near the source. While total coverage surveys are preferable in theory, it simply was not worthwhile to survey many square kilometers of jumbled rhyolite boulders and skree fields, terrain that were barely passable on foot, when the vast majority of all the locations for sizable sites could be targeted by the general criteria evident on maps and imagery. Furthermore, a comprehensive study of activities related to the source area demanded that time was budgeted for an investigation of the highly productive lands approximately one day's travel away from the obsidian source, in order to place obsidian procurement in the context of the local economy. A research strategy that approached the entire region in terms of survey and testing in three major contiguous blocks was deemed the most effective approach to documenting the source region.

5.4.2. Surveys types: Prospection, Statistical, and Spatial Structure

The principal goal of the survey work was to document archaeological distributions in source area and adjacent terrain. Banning (2002) describes the goals of archaeological survey in terms of three principal types of survey that emphasize different research goals.


Prospection / Purposive


Spatial Structure


For findingarchaeological sites.

For estimating population parameters, evaluating probabilistic hypotheses and constructing locational models.

For detecting spatial patterns such as settlement lattices, travel routes. Also good for documenting continuous phenomena.


Prioritize the locating of sites by incorporating background information, predictive models, and remote sensing.

Sampling strategies for documenting artifact density, diversity, and site types within stratified samples or numerical samples.

"Total coverage" or nonsite survey to identify spatial interrelationships that might be missed by sampling approaches.

Table 5-3. Types of archaeological survey described by Banning(2002: 27-38).

Banning makes the point that while sampling is a common approach to archaeological survey, sampling is actually a poor method for prospecting for sites or for characterizing settlement lattices because major clues can fall outside of the sampling window. Each method has specific strengths and weaknesses, and often archaeological surveys are a mixture of several types.

5.4.3. Surveyor interval and sampling

Following Banning's (2002) terms, the Upper Colca survey work was a combination of all three survey types. The survey was prospective because it attempted to document a little known region and find the majority of the large sites associated with the obsidian source, but it was also statistical because survey zones were deliberately stratified so as to permit predictive statements about the use of space throughout the study region. Finally, the Upper Colca research also involved survey for spatial structure because it consisted of three large blocks within which the land was thoroughly surveyed so as to document intersite relationships and travel routes.

Many contemporary archaeological surveys will claim to have conducted "100% survey" of large regions, but then they will have had a survey interval of 30m or more between surveyors. Surveys focused on documenting complex societies with standing architecture are particularly likely to refer to their widely-spaced surveys as "100% surveys". A wide surveyor interval is actually a non-explicit kind of sampling that de-prioritizes smaller sites and those lacking standing architecture, resulting in an often unstated bias in the results. Subsequently, the region is considered "surveyed" though many smaller sites falling between transects were surely missed. While smaller sites are found in these widely spaced surveys, it is only if the site happens to fall across one of the surveyor lines. More realistically, such a survey method is somewhat successful because the surveyors cover a lot of ground but then they will veer off their route to visit high likelihood locations for sites such as rock shelters and lake shores; a technique belonging to the realm of prospection survey.

5.4.4. Survey design

The Upper Colca Project survey goals emphasized investigating the Chivay source area, the geological contexts for obsidian formation, and the principal areas of human settlement within one day's walk from the source. In the implementation of the survey, high-likelihood areas in the obsidian source zone were evaluated using a prospection survey. This included a careful survey of the entire Maymeja area itself and large portions of the southern rim using a surveyor interval of 15m.



Scale / Res.

Comments / Application

ASTER imagery

NASA, JPL(Abrams, et al. 2002)


Visual and NDVI analysis


NASA, JPL(Abrams, et al. 2002)


Representation, slope calculation


NASA, USGS, CGIAR (Jarvis, et al. 2006)


Regional relief mapping

Aerial photos

Servicio Aerofotográfica Nacional, Perú


Historic aerial photos.

Topographic maps

Instituto Geográfica Nacional, Perú


Features, toponyms (PSAD56), scanned

Geology maps



Geology (PSAD56), scanned


NIMA (1995)


Regional map


NIMA (1995)

1:1 million

Continental map

Table 5-4. Digital data sources used in developing the survey strategy.

Selection of survey regions involved the use of a number of spatial data sources (Table 5-4), as well as interviews with local residents, personal visits, and consultation of previously published reports. Preliminary field visits with a Trimble Geoexplorer GPS in 2001 and 2002 involved collecting ground control points and GPS lines on major roads and other features. After post-processing using the AREQ base station (International GPS Service), these data permitted the georeferencing of aerial photos and scanned maps directly to the GPS acquired data.


Figure 5-1. Criteria in designing regional survey from three stage research proposal including obsidian source survey, testing program, and concluding with the river valley survey.

Survey in the area of the obsidian source, outside of the Maymeja depression itself, was selective as it focused on high likelihood areas. Survey coverage in the Blocks 2 and 3 zones was contiguous with a 15m surveyor interval although here the survey region was delimited by other criteria. First, in the Block 2 (San Bartolomé area) a particular strip of land was targeted that paralleled the terminus of a Barroso lava flow. The survey block was surveyed 100% at 15m intervals along this densely occupied region. In Block 3, a maximum steepness and distance to river criteria was used to concentrate survey efforts to the river corridor region. Thus, in Block 3, all lands were surveyed within 500m of the high river terrace above the principal drainage (Colca, Llapa, and Pulpera drainages), and terrain over 15° slope (33% slope) were not surveyed. This maximum steepness limitation excluded many eroded regions where preservation is poor, but it also excluded a number of areas that were perhaps occupied. In order to evaluate the survey criteria in Block 3, a swath of land 1 km wide by 3 km long was surveyed at truly 100% coverage at 15m interval, and these areas could then evaluated to gauge the effects of the survey criteria used elsewhere in Block 3 that excluded the high slope and non-riverside areas. This 100% survey test swath will be described in more detail below.


Figure 5-2. GPS tracks from edges of most survey routes showing emphasis on Blocks 1, 2, 3 and 6.

With GPS units the survey coverage and spatial sampling is made relatively explicit, permitting future researchers in the region to focus on areas that were under-investigated in the 2001, 2002 and 2003 survey efforts. With GPS track loggers becoming easily available, explicit coverage reporting will likely be more widely adopted in the future.

5.4.5. Testing the effectiveness of the B3 survey strategy

While survey criteria for coverage in Block 3 were relatively restrictive, a swath in the vicinity of Callalli with a diversity of topographic and ecological conditions was selected for conducting a "100% survey". The goal of covering ground at a 100% was to evaluate the effectiveness of the survey strategy that was being applied throughout the rest of Block 3. The survey coverage in Block 3 included only areas within 500m of the highest river terrace, and slopes under 15° (33.3%) incline.

The 100% survey revealed seven small sites, some lithic isolates, a lone broken vessel and a wall on a hilltop location that was undiagnostic but is probably a Late Intermediate Period pukara construction. The area of the 100% survey swath is 0.5 km wide by 3.5 km long (area = 1.7 km2) and if onlythe area outside of the regular survey model is included, the area is 1.1 km2. The sites located in the areas outside of the survey model fall into two major groups: pukaras on hilltops and small, eroding lithic scatters with no reliable temporal assignment found on steep open slopes. Other regional evidence points to a pattern of intensified pastoral production during the LIP and Late Horizon, and these dispersed sites may result from herders working while they monitor their flock during the wet season when the hillslopes of Callalli contain rich graze.




Feature type














Visib is +33





























Pukara. Visib is +3.12









Possible Pukara wall.

Table 5-5. Sites and isolates from 100% survey strip that would not have been encountered using the regular Block 3 survey strategy.

Given the high effort expended in completing the 100% survey, and the eroded condition of most sites on steep slopes, a slight modification of the survey strategy would have resulted in the group encountering virtually all the informative sites in the region. The improved survey model would be like the one that was employed (500m from the high terrace and < 15° slope), and furthermore it would include a visit to all the major hilltops in the region searching for pukaras. With prospection survey, it is often true that pukara walls can be identified with binoculars or on imagery (Arkush 2005), allowing for targeted climbs of only those hills with visible walls.

5.5. Mobile GIS for archeological survey

Mobile GIS can be incorporated into archaeological survey methods with varying degrees of change to traditional archaeological survey techniques. This section describes the methods implemented in 2003 where a predominantly digital recording method was used.

5.5.1. Standard survey practice

The standard survey practice in the 2003 season consisted of a single team of four to six surveyors spaced 15m apart. The team swept across hillslopes following contours, and at the end of each survey transect the team would sweep around and return towards the opposite direction in the adjacent transect following a boustrophedon configuration. The survey team would assemble to investigate sites when they were encountered, although the team would not necessarily assemble for isolated finds.


Figure5-3. An example of a pedestrian survey line following a river terrace at a 15 meter interval. In this survey, only one mobile GIS unit is used. GPS units carried by the surveyors at either end of the survey line mapped the extent of all surveyed areas.

The portable equipment carried on the survey by each team member included basic day hiking equipment such as personal gear, lunch, copies of maps and a compass, and everyone had a small FM walkie-talkie. In order to map the extent of daily survey coverage, and to quantify the distances between surveyors, those hiking on either end of the survey line carried a Trimble GeoExplorer GPS logging a polyline attributed with the side of the survey that they were walking (right or left) as well as the number of surveyors hiking that day. GPS datalogger tracking devices are more widespread today (in 2006) and pair of small USB based dataloggers such as the Sony GPS-CS1 for geotagging photos are also suitable for continuous data stream mapping on either end of a survey transect. A separate mobile GIS system consisting of Dell Axim PDA with a Trimble Pathfinder Pocket GPS receiver was carried for mapping archaeological sites into Arcpad as will be described in more detail below.

5.5.2. The contribution of mobile GIS

Current mobile GIS technology contributes to traditional archaeological survey methods in several ways. First, mobile GIS aids surveyors with navigation because the anticipated survey transects, and some other relevant guidance information, can be clearly indicated in conjunction with the current GPS location. Second, mobile GIS allows researchers to record new vector data along with attribute forms that are more flexible than those provided by GPS or by data dictionary approaches in the past. Finally, mobile GIS allows researchers to transport digital datasets into the field so that they can do error checking immediately, review the work of other research teams, and perform queries on large existing volumes of data in digital form.

When a surveyor encountered an archaeological feature the surveyor would first determine if the feature exceeded the specification for isolates and then, should the feature be a site, the surveyor would call a halt to the survey line. Site boundaries were established for two reasons in the 2003 fieldwork. First, in GIS it is generally required that a geographical feature be delimited and that a database record is created before it can be attributed. Thus, one cannot describe a site that has not yet mapped, unless some kind of more complicated work-around is employed such as the creation of a temporary attribute record. A second beneficial effect of delimiting sites as an initial step, however, is that the team is forced to travel over the site completely and assess the extent and variability before an attempt was made to describe it.

5.5.3. Hardware configuration

Mobile GIS systems permit surveyors to attribute spatial locations with a variety of data types. Currently the GPS unit is the primary digital input into the mobile GIS and this permits the mapping of point, lines, and polygons delimiting archaeological features. In the GIS the spatial data is attributed and once post-processing is complete the new data joins the larger GIS database.


Figure5-4. Mobile GIS implementation with ESRI Arcpad 6. New data sources from external instruments are shown in the top row. Where post-processing is needed new data is not integrated with other data until later. New and existing data can be summarized and displayed together.

Existing inputs to the mobile GIS system in 2003 are relatively limited. Important attributes that currently must be entered manually include the digital photo numbers associated with each archaeological feature, and relative measurements collected at the site. Additional instruments in the future might include wireless tapes that can transmit precise lengths back to the mobile GIS of features like the dimensions of a doorway. Alternately, for in-field lab analysis on non-collection survey, wireless calipers or scales could be used to transmit the size of an artifact to the mobile GIS linked to the spatial provenience.

5.5.4. Defining loci and sites

When an archaeological site is encountered during survey in this Arcpad system the site must be delimited first, using the Site-A polygon, and then loci located within the site are delimited and described as related in two accounts (Tripcevich 2004;Tripcevich 2004). The locations of individual artifacts of interest are mapped and bagged separately using a Lithic_P or Ceramic_P geometry type. These include diagnostic artifacts or other materials of specific interest.

Locus / Site

Min. Density Artifacts

High Density

10+ artifacts per m2

Medium Density

5-10 artifacts per m2

Low Density

1-3 artifacts per 2m2


2 artifacts per 10 m2

Table 5-6. Locus and Site artifact density definitions.

Pin flags were used to delimit these features of interest, and generally in the case of most medium and high density loci, the result is a "fried-egg" model of artifact density polygons. In recording these polygon features, one generally went from the geographically largest to smallest entity because, as is also true in desktop GIS, features that are created later appear "on top" of features created earlier and thus larger, later features would visually obscure earlier features. This condition has to be corrected back in the laboratory and thus it was simply easiest to map largest to smallest. When a feature is mapped in Arcpad with the GPS then, subsequently, an attribute form appears that allows for explicit description of the feature.

5.5.5. Attribute Forms

Aside from the site datum points and site boundary polygons, three dominant feature types characterized the archaeological data set in the mobile GIS. Each archaeological data type had an attribute form associated with it that recorded information appropriate for a given feature. Page One of the digital forms comprised a unique ID number generated from a script and a range of numbers for digital photos (JPEG files) documenting a feature.





Figure 5-5. (a) Arcpad screen showing a large site with loci and points. (b) Example of page two of a lithic locus form in Arcpad showing Category 1 and Category 2 columns; in the background, two sites and contour lines are displayed on top of a 15m resolution ASTER satellite image.

Page Two of the attribute forms (Figure 5-5b) contained specific information about the feature type, such as Site, Locus, or Point information. The third page contained eight pull-down menus with environmental attributes for geology, exposure, and other local variables. These values were usually the same within a given site so that the values were "sticky"; they were stored in temporary memory between recording events, and the editable form was repopulated automatically unless a new site feature was being recorded. The final page contained a "Comments" field that accepted up to 255 characters and included a button that would open Pocket Word application with a text file named for the unique ID #, allowing the entry of additional notes if necessary. A link to a separate application that permitted MP3 compression of voice-based comments was available as well, but because the processor demands of sound encoding overly hampered the functionality of the Pocket PC for the GIS application, the feature went unused.

5.5.6. Variability within a Locus

A basic complexity of archaeological survey is that artifact concentrations frequently contain a variety of artifact types, perhaps dating to completely different occupations. This variability presents a particular challenge for a fast, mobile GIS based recording system because in lieu of sampling, all that the archaeologist has time to do is to document his or her rapid assessment of the artifacts that are found within individual loci geometry mapped into the GIS. Additionally, despite of the variability present within the locus, the archaeologist must attempt to generate data over the course of the field season that are consistent and comparable. During the Upper Colca Survey this difficulty was addressed by estimating the characteristics of a primary and secondary attribute category, dubbed Category 1 and Category 2 (see Figure 5-5b), that best characterizes the locus using the custom interface developed for the project.

The problem: How does one evaluate and map a scatter of, say, 5,000 stone flakes in less than one hour, as well as estimate the percentage of obsidian to another material type, such as chert?

In order to achieve statistical rigor and reliability, a sampling strategy was needed. Sampling and collecting artifacts is time consuming, and sampling at every concentration of lithics near a quarry is also unrealistic because there are so many lithic artifacts in such areas. Sampling was therefore carried out at "High Density Loci" with artifact concentrations deemed most worthwhile given the research goals, while a less rigorous approach was applied for artifact distributions of lesser importance. A solution was devised that is geared for conducting cursory inventory, not an in-depth assessment. This solution captured variability by estimating the proportions of the two dominant attribute categories within a given polygon.


(a) (b)

Figure5-6. Maps for two different hypothetical sites recorded in less than one hour. (a) A conventional, low precision sketch map showing only major site features and perhaps subdivided into site sectors (b) Mobile GIS site map with 1-2m dGPS error. Internal distributions, such as the fried-egg density gradient model shown here, can be assessed and rapidly mapped.


(a) (b)

Figure 5-7. (a) Structure of the archaeological Shapefiles with names and descriptions. Each of the Shapefiles had a form associated with it that prompted the user with fields appropriate to that data type. (b) An example of a part of the ID # system that prioritizes spatial provenience in the field.

A hypothetical site description makes the site recording strategy more clear. This example takes place at a site with a large, low-density lithic locus (Figure 5-6b), where the concentration of stone artifacts was mostly obsidian material but also included artifacts made from chert, chalcedony, and quartzite. The mobile GIS user walks around the locus with the GPS running, and the area was recorded into the "Lithics-A" ShapeFile (Figure 5-7a). Lithic concentrations of medium and high density are found inside the locus, creating a 'fried-egg' density map. Subsequent to delineating the locus with a GPS, the custom form (Figure 5-5b) appears. Several steps are followed in filling out the form.

(1) The primary "axis" of variability is determined. In this case, it is stone material type.

(2) Using this variable, the largest group is characterized. This attribute category 1 (C1) was described as "Material: Obsidian," and other attributes of interest to lithic analysis such as amount of cortex, size of debitage, and artifact density in the attribute category, were rapidly estimated. In our case, the density was "Low."

(3) The second most represented group, attribute category 2 (C2), is characterized and its attributes are evaluated, again as quickly as possible. Any subsequent groups were disregarded for expediency and because of the error in estimation and low reliability of the method.

(4) The proportion of stone artifacts in the polygon estimated to meet the description of C1 is entered in the field labeled "C1% of Locus," and an estimate is also generated for category 2.

The method works for a rapid inventory, and it provides a general estimation of materials along with the characteristics and densities within loci. Using this system, archaeologists are encouraged to describe the variability between category 1 and 2 in terms of only one variable at a time. For example, if there were notable differences in both Material Type and Debitage Size in a particular locus, then a second polygon was created. Alternatively, the first polygon was copied, and the different "axes" of variability were distinguished independently. Instant types (i.e., attribute categories) were generated for each polygon by emphasizing the greatest variability within the locus, and this was considerably more spatially explicit than rapid archaeological survey had been in the past despite a relatively small investment in time. Time efficiency was a major objective of the Colca Survey with recording all but the very highest density lithic concentrations, and this approach allowed for rapid feature mapping. A variety of new possibilities for custom field applications are becoming available now that modern digital equipment, such as the mobile GIS used in this archaeological survey, can be modified and streamlined by the archaeologists to suit the needs of research without recourse to professional programmers.

5.5.7. Sampling High-density Loci

For the purposes of the Upper Colca project High density lociwere defined as areas where the density of the artifact scatter appeared to exceed 10 artifacts per m2. As with all loci, these concentrations were mapped using the mobile GIS interface, but then High-density loci were further characterized by collecting all artifacts within two or more 1x1 m sample squares for later analysis back in the lab. The Arcpad SampleDesignscript was used to pseudo-randomly place, using an unaligned-grid method, a sufficient a number of square sample units to cover at least 0.01 of the Shape Area (m2)of the locus as reported in Arcpad. This works out to a 1 m2collection area for every 100 m2of polygon area. The GPS indicator was used to navigate to the randomly generated point locations. When documenting each sample an overhead photo was taken of the 1x1m area from near-nadir for later georeferencing, and then artifacts were completely collected. One or more units were randomly placed somewhere within the polygon, and one unit was always placed right on the location of estimated highest density. During the 2003 season, such collections resulted in an average sampling fraction of 0.014 among the twenty-two samples that were collected during the course of the field season in this process of sampling high density loci.

5.5.8. Collection during survey

Traditionally, it has been impractical for archaeologists to retain precise spatial provenance for surface artifacts that are not particularly interesting or rare. Collected artifacts are aggregated by site, sector, or by locus. However, artifact collection is increasingly seen as a destructive practice. The collection strategy used in the Upper Colca Survey consisted of assigning a unique ID number (ArchID) from a single number series to all spatial proveniences, point locations, loci, or entire sites-very much like postal zip codes for street addresses. After four months of fieldwork, 1100 spatial provenance numbers had been assigned from the series. As described previously, individual artifacts collected from a given provenience were assigned key ID#s after a decimal point. An interesting alternative to handwriting the unique ID# on labels for sample bags collected in the field is to bring a sheet of pre-printed barcode stickers. As the sticker is placed on the sample container, a serial barcode scanning wand can scan the barcode value directly into the GIS record. The barcode scanner approach is somewhat restrictive, however, because the mobile GIS unit must to be available to scan every collection ba