Human Adaptation. Food Production. Amd Cultural Interaction During the Formative Period in Highland Ecuador

University of Calgary PRISM: University of Calgary's Digital Repository

Arts Arts Research & Publications

2012-07-16T07:02:45Z Human Adaptation. Food Production. amd Cultural Interaction during the Formative Period in Highland Ecuador

Zarrillo, Sonia http://hdl.handle.net/1880/49108 Thesis http://creativecommons.org/licenses/by-nc-nd/3.0/ Attribution Non-Commercial No Derivatives 3.0 Unported Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Human Adaptation, Food Production, and Cultural Interaction during the Formative

Period in Highland Ecuador

Sonia Zarrillo

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF ARCHAEOLOGY

CALGARY, ALBERTA

JUNE, 2012

This dissertation focuses on plant food production, human adaptation, and cultural

interaction in the highlands of Ecuador during the Formative Period. I conducted starch

granule analysis of ceramic charred cooking and stone tool residues from some of the

earliest Formative Period sites in the highlands, and one site from the eastern Andean

slopes, to develop a regional synthesis of the timing and nature of highland plant food

production. The main hypothesis tested is whether the stimulus to a Formative lifeway in

the highlands diffused from coastal Ecuador. Sites investigated include La Chimba,

Tajamar, Cerro Narrío, Chaullabamba, La Vega, Trapichillo, and Santa Ana-La Florida

(SALF). I also integrate data from previous botanical analyses at other sites, especially

Cotocollao.

The analyses show that Andean domesticated crops, such as oca, potato, lupines

(chocho/tarwi) and quinoa, as well as maize and beans, are associated with the highland

sites located at, and to the north of, Chaullabamba (La Chimba, Tajamar, and Cerro

Narrío). By employing site catchment analysis, I contend that a highland, “vertical

compact”, agricultural system was being practiced at these sites by at least the terminal

Early Formative period, and perhaps much earlier based on proxy (pollen) evidence and

aspects of the plants’ nutritional profiles, which suggest long-standing familiarity with

the crops. By integrating the latest information available on the crops’ origins, I argue

that cultural interaction was perhaps principally occurring through the Inter-Andean corridor along a north-south axis with other highland groups.

ii In contrast, the far southern highland sites (La Vega and Trapichillo), as well as

SALF, show crops suited to a lowland tropical agricultural system, including manioc, sweet potato, Dioscorea (yam), as well as maize and beans and, possibly, cacao. The results from SALF date to almost the beginning of the Early Formative Period, showing that Coastal Ecuador was not the only “hotspot” for Early Formative Period societies.

Cultural interaction in the far southern highlands shows an east-west axis of orientation.

Groups in the highlands, eastern lowlands, and coast were involved in multiple interaction spheres. Highland agriculture and socio-political complexity developed in- situ, and not from unidirectional diffusion from coastal Ecuador.

iii Acknowledgements

The field work and research for this dissertation would not have occurred without the support of many people and sources of funding. During my studies I was honoured to

receive the Dean’s Research Excellence Award, a SSHRCC CGS Scholarship, the

Friend’s of Head-Smashed-In Scholarship (on two occasions), the Dean’s Entrance

Scholarship, a SSHRCC Doctoral Fellowship, an Honorary Izaak Walton Killam

Memorial Award, the Martha Biggar Anders Memorial Award (on three occasions), a

University of Calgary Silver Anniversary Graduate Fellowship, a University of Calgary

Graduate Research Scholarship, a University of Calgary Graduate Studies Scholarship, an

International Fee Differential Travel Grant, a Faculty of Graduate Studies Travel Award,

a Graduate Students Association Professional Development Grant, and a Student

Activities Fund Travel Grant. In addition, I was particularly honoured to receive the J.B.

Hyne Research Innovation Award. Without the generous financial support of these

internal and external funding agencies and privately-supported awards, the many trips to

Ecuador to conduct my research would not have been possible.

My co-supervisors, Dr. J. Scott Raymond and Dr. Brian P. Kooyman, provided

the inspiration, education, and drive, while I was still an undergraduate student, to

continue with graduate studies. My journey into paleoethnobotany began with a course

(that I took at the last minute, two weeks after the course began) with Brian. For a class

project I undertook an analysis two stone tools, and instead of finding phytoliths, I found

starch granules instead. This led to my honours thesis on ceramic residues from the Loma

Alta site, supervised by Scott. I must admit that I wanted to study the Loma Alta stone

iv tools, thinking that starch granules would not preserve through the original cooking events, but Scott wisely “insisted.” As such, both Scott and Brian have played more than a major role in the focus of my graduate research – they have both been my primary role models, mentors, and sources of endless support and patient encouragement. I particularly want to thank Scott for undertaking a field project in Ecuador at Cerro

Narrío, at a time in his career when he very rightly did not have to, and Brian for always being there to listen to, and provide sage advice, on all aspects of my research. I would also like to thank all the members of my committee: Dr. Dale Walde and Dr. Richard

Callaghan from the Department of Archaeology; Dr. Chendanda “CC” Chinnappa from the Department of Biological Sciences; and Dr. James Zeidler for acting as the external examiner. I very much enjoyed the discussions we had during my defense and benefitted greatly from the experience.

The faculty and support staff of the Department of Archaeology are also deserving of praise. I can honestly say that all of my courses were valuable and inspiring.

Over the years I also benefitted from many conversations with faculty I did not have the opportunity to take courses with. Thank you all for your help and support.

I am also tremendously grateful to all of my friends and colleagues in Ecuador who facilitated and supported my research. Through the support of Diego Quiroga, the

Universidad San Francisco de Quito (USFQ) provided much needed use of laboratory facilities in Riobamba. The excavations and work at Cerro Narrío would not have occurred without the co-direction of Dr. Florencio Delgado (USFQ). We were assisted

v (and could not have managed without them) with excavations and survey by the Delgado brothers: Hernan, Danilo, Nivardo, and Alejandro. Fernando and Daniela Astudillo, as well as many others, also helped out with our excavations. Our 2009 field and lab work would not have happened without U of C students Lance Armstrong, Whitney Mosher, and William Swan, who kindly volunteered with excavations, cataloging and analysis.

The city of Cañar also supported our excavations at Cerro Narrío, as well as, most importantly, the Cañari community of Quilloac. We were very fortunate and honoured to have their support, as well as assistance with excavations provided by Mariana and

Santiago. The Museos Banco Central del Ecuador in Quito and Cuenca also provided access to artifacts, as well as the Instituto National de Patrimonio Cultural offices in

Quito, Cuenca, and Loja, and I would like to especially thank Diego Castro, Danilo

Delgado, Fernando Mejia, Joaquin Moscoso, Félix Alvear, and the late Antonio Carrillo.

Without Antonio’s expert guidance and knowledge, it would have taken far more than one morning to sort through and choose samples from the Challuabamba collection.

Victoria Domínguez suggested I sample some of the Tajamar artifacts. She also provided laboratory space at the Rumipamba Archaeological Park during sampling. I am very grateful for all her assistance. Dr. J. Stephen Athens entrusted me with numerous samples from La Chimba – I promise to sample more of them – thank you Stephen! I also wish to thank Dr. Michael Blake of the University of British Columbia. I have greatly benefitted from many conversations with him, and he also suggested that I investigate Theobroma as a source of some of my unknown starch granules. Finally, I am tremendously indebted to Francisco Valdez. He provided access to the artifacts and substantial help with logistics for sampling and travel to and from SALF. Francisco also located the long-

vi sought-for Catamayo sites’ artifacts and, not only facilitated my access to them, but also helped me while I sampled them! In Calgary, I am very grateful to the students who volunteered in the paleobotany lab, especially Debbie Onos, Mary Lynn Tobias, and

Adam Brousseau.

I also wish to express my tremendous appreciation to Dr. Deborah Pearsall. From my first analysis with the Loma Alta ceramics, through to the completion of my dissertation, she has offered moral support, advice, comparative plant specimens, and hosted my visit to the University of Missouri. I have always been uplifted by her kind words. She is a true role model and friend, and embodies all that is good in academia.

Finally, I wish to thank my family. My mother, sisters and brother have cheered me on; patiently waiting for the day I would be done. Thank you for your love and support. My husband Dugane Quon had to endure many untold hours of “discussion” about my coursework, field work, and lab work, and proofread far too many papers over the past 12 years – I only asked him to read one paragraph of this dissertation. Last, but most importantly and dearest to my heart, I want to thank my children, Jennifer,

Nicholas, Matthew, and Emily. They sacrificed much over the past 12 years and I could not have done this without all of you. Whenever I was down and wanted to quit, cold, wet, and homesick in Ecuador, or just generally tired, I would think of all of you (or get a funny message or email) and know that I had to finish. I hope I can now make up for the many times we missed spending together.

vii Dedication

This dissertation is dedicated to my children, Jennifer, Nicholas, Matthew, and Emily,

and to my mother, Maria Zarrillo.

My children sacrificed a lot of “mom” time over the past years, through no choice of their own. They have always supported and encouraged me, are a continual source of joy and

pride.

Although my mother never had the privilege of a formal education, she is one of the most intelligent people I know. She instilled in me a curiosity of world cultures, as well as an

admiration for traditional gardening methods and the oral transmission of cultural

knowledge. Her garden was, and still is, the envy of the neighborhood.

I only wish I had inherited my mom’s green thumb. Those who can, do. Those who

cannot, write a dissertation about it.

viii Table of Contents

Abstract ...... ii Acknowledgements ...... ii Dedication ...... viii Table of Contents ...... ix List of Tables ...... xii List of Figures and Illustrations ...... xiii Epigraph ...... xvii

CHAPTER ONE: INTRODUCTION ...... 1 1.1 Introduction ...... 1 1.2 The Diversity of Andean Environments ...... 2 1.3 Statement of the Research Problem ...... 3 1.4 Theoretical Approach to the Research ...... 6 1.5 Key Concepts Related to Plant Domestication and Agricultural Systems ...... 9 1.6 Organization of the Dissertation ...... 13

CHAPTER TWO: OVERVIEW OF ECUADOR’S NATURAL SETTING ...... 18 2.1 Introduction ...... 18 2.2 Physiographic Setting ...... 18 2.3 Climate ...... 23 2.4 Volcanism ...... 29 2.5 Vegetation ...... 31 2.6 Chapter Summary ...... 40

CHAPTER THREE: OVERVIEW OF ECUADORIAN PRECERAMIC AND FORMATIVE PERIODS AND PLANT RESOURCES ...... 41 3.1 Introduction ...... 41 3.2 Coastal, Highland and Eastern Lowlands Preceramic Precedents ...... 43 3.2.1 The Coastal Lowlands Preceramic Period ...... 43 3.2.2 The Highlands and Eastern Lowlands Preceramic Period ...... 45 3.3 The Coast, Highlands and Eastern Lowlands Formative Periods ...... 48 3.3.1 The Coastal Lowlands Formative Period and Record of Plant Use ...... 48 3.3.2 The Highland and Eastern Lowland Formative Periods and Record of Plant Use ...... 57 3.4 Chapter Summary ...... 75

CHAPTER FOUR: ARCHAEOLOGICAL BACKGROUND AND SITE ENVIRONMENTAL SETTINGS ...... 82 4.1 Introduction ...... 82 4.2 Santa Ana-La Florida ...... 84 4.3 Challuabamba/Chaullabamba ...... 99 4.4 Catamayo Sites – Trapichillo and La Vega ...... 106 4.5 Cerro Narrío ...... 113 4.6 La Chimba ...... 129 4.7 Tajamar ...... 134

ix 4.8 Chapter Summary ...... 138

CHAPTER FIVE: ARCHAEOBOTANICAL SAMPLING AND METHODOLOGY ..140 5.1 Introduction ...... 140 5.2 Starch Analysis: The Biology of Starch and Archaeological Analysis ...... 143 5.3 Modern Comparative Plant Specimens ...... 147 5.3.1 Starch Granule Comparative Specimens ...... 148 5.4 Modern Comparative Starch Granule Sample Processing ...... 153 5.5 Methods for Recovering Starch Granules from Archaeological Samples ...... 155 5.5.1 Stone Tools ...... 155 5.5.2 Ceramics ...... 157 5.5.3 Heavy Density Liquid Separation for the Recovery of Starch Granules from Stone Tool and Ceramic Residues ...... 162 5.5.4 Microscope Slide Preparation and Microscopic Examination ...... 168 5.6 Methods to Control Contamination of Archaeological Samples with Modern Starch ...... 168 5.6.1 Controlling for Starch Contamination during Excavation ...... 171 5.6.2 Controlling for Starch Contamination during Sampling and Storage ...... 173 5.6.3 Controlling for Starch Contamination during Laboratory Processing and Analysis...... 176 5.7 Methods of Starch Granule Identification ...... 180 5.8 Chapter Summary ...... 183

CHAPTER SIX: STARCH ANALYSIS AND AMS RADIOCARBON ASSAY RESULTS ...... 187 6.1 Introduction ...... 187 6.2 Santa Ana–La Florida ...... 187 6.3 Chaullabamba ...... 214 6.4 Tajamar ...... 225 6.5 La Vega ...... 230 6.6 Trapichillo ...... 235 6.7 Cerro Narrío ...... 240 6.8 La Chimba ...... 246 6.9 Accelerator Mass Spectrometry Radiocarbon and 13C/12C Assays ...... 249 6.10 Chapter Summary ...... 257

CHAPTER SEVEN: DISCUSSION ...... 261 7.1 Introduction ...... 261 7.2 Methodological Issues ...... 261 7.2.1 Are the Starch Granules Recovered “Real”? ...... 261 7.2.2 Assessing the Methods for Sampling ...... 266 7.2.3 Assessing Starch Granule Identification Methods ...... 269 7.2.4 Methods of Quantification ...... 274 7.2.5 Quantification and the Issue of 13C/12C Stable Isotope Assays ...... 281 7.3 Accuracy and Precision in Radiocarbon Dating ...... 285 7.3.1 Assessing the Ceramic Charred Residue AMS Radiocarbon Dates: Chaullabamba ...... 286

x 7.3.2 Assessing the Ceramic Charred Residue AMS Radiocarbon Dates: Trapichillo and La Vega – the Catamayo Sites...... 290 7.3.3 Assessing Highland verses Coastal Formative Period Dates: Potential Sources of Error...... 292 7.4 Paleoethnobotany of the Highlands and Eastern Lowlands during the Ecuadorian Formative Period ...... 297 7.4.1 Fine-tuning the Starch Granule Taxonomic Classifications ...... 300 7.4.2 The Geographical Origins, Ecological Requirements, and Nutritional Aspects of the Identified Plants...... 307 7.4.2.1 Chili Peppers: Capsicum spp. (Solanaceae) ...... 312 7.4.2.2 Quinoa: Chenopodium quinoa (Chenopodiaceae) ...... 316 7.4.2.3 Yams: Dioscorea spp. (Dioscoreaceae) ...... 320 7.4.2.4 Sweet Potato: Ipomoea batatas (Convolvulaceae) ...... 322 7.4.2.5 Chocho/Tarwi: Lupinus mutabilis (Fabaceae) ...... 324 7.4.2.6 Manioc: Manihot esculenta (Euphorbiaceae) ...... 329 7.4.2.7 Arrowroot: Maranta arundinacea (Marantaceae) ...... 334 7.4.2.8 Oca: Oxalis tuberosa (Oxalidaceae) ...... 336 7.4.2.9 Beans: Phaseolus spp. (Fabaceae) ...... 338 7.4.2.10 Potato: Solanum tuberosum (Solanaceae) ...... 342 7.4.2.11 Cacao/Chocolate: Theobroma spp. / Herrania spp. (Malvaceae) .....346 7.4.2.12 Maize: Zea mays (Poaceae) ...... 351 7.4.3 Cultural Interaction ...... 354 7.4.4 Why Nutrition Matters ...... 361 7.5 The Use of Ecological Zones and the Nature of Agricultural Production Systems ...... 366 7.6 Chapter Summary ...... 378

CHAPTER EIGHT: CONCLUSIONS ...... 385 8.1 Introduction ...... 385 8.2 Starch Granule Analysis Conclusions ...... 385 8.3 Conclusions Regarding Highland Formative Period Plant Use and Timing ...... 387 8.4 Conclusions Regarding the Stimulus to Formative Period Complexity in Highland Ecuador ...... 392 8.5 Conclusions Regarding the Nature of Highland Plant Food Production ...... 396 8.6 Final Thoughts and Future Research ...... 398

REFERENCES CITED ...... 403

APPENDIX A: FIVE KILOMETER CATCHMENT AREA TOPOGRAPHIC AND 3-D PROJECTION MAPS FOR LA CHIMBA, TAJAMAR, CERRO NARRÍO, CHAULLABAMBA, THE CATAMAYO VALLEY, AND SANTA ANA-LA FLORIDA ARCHAEOLOGICAL SITES ...... 464

xi List of Tables

Table 2.1. Names and Altitudes of Climate Zones in Ecuador ...... 28

Table 3.1. Plant Remains Recovered from Preceramic and Formative Period Coastal Lowland Sites in Ecuador ...... 56

Table 3.2. Plant Remains Recovered from Highland and Eastern Lowland Locations in Ecuador ...... 71 Table 4.1. Contexts and Radiocarbon Age Determinations from Santa Ana-La Florida.1 ...... 86

Table 5.1. Modern Plant Species Tested for Starch Granules...... 149

Table 5.2. Archaeological Sites and Samples Tested for Starch Granules...... 164

Table 6.1. Santa Ana-La Florida Starch Analysis Results...... 190

Table 6.2. Chaullabamba Starch Analysis Results...... 215

Table 6.3. Tajamar Starch Analysis Results...... 226

Table 6.4. La Vega Starch Analysis Results...... 231

Table 6.5. Trapichillo Starch Analysis Results...... 237

Table 6.6. Cerro Narrío Starch Analysis Results...... 242

Table 6.7. La Chimba Starch Analysis Results...... 247

Table 6.8. Accelerator Mass Spectrometry Radiocarbon and 13C/12C Assays on Ceramic Charred Residue Samples...... 250

Table 7.1. Radiocarbon Dates for the Chaullabamba Site...... 287

Table 7.2. Earliest Reported Radiocarbon Dates for the Cotocollao Site: Contexts and Recalibration...... 295

Table 7.3. Starch Granule Analysis Results from all Archaeological Sites Investigated...... 299

Table 7.4. Plant Taxa Identified from the Ecuadorian Highlands and Eastern and Western Slopes ...... 310

Table 7.5. Elevation Range and Region of Domestication of Plant Species ...... 311

Table 7.5. Starch Granule Analysis Results for Sites Located Above 2000 masl...... 368

Table 7.6. Starch Granule Analysis Results for Sites Located Below 1500 masl...... 368 xii

List of Figures and Illustrations

Figure 2.1. Political map of South America showing the location of Ecuador...... 19

Figure 2.2. Relief map of Ecuador with physiographic features...... 20

Figure 2.3. Chimborazo, the highest mountain in Ecuador...... 21

Figure 2.4. Shaded relief map of Ecuador...... 23

Figure 2.5. Map of annual average precipitation for Ecuador...... 25

Figure 2.6. Map of annual average temperature for Ecuador...... 27

Figure 2.7. Sangay Volcano located in the Cordillera Oriental in Chimborazo Province...... 29

Figure 2.8. Tungurahua erupting, Cordillera Oriental in Tungurahua Province, central highlands...... 30 Figure 2.9. Vegetation map of Ecuador...... 32

Figure 2.10. Dry scrub vegetation near Manta, coastal Ecuador...... 33

Figure 2.11 Entering the cloud forest, western slopes of the Cordillera Occidental...... 34

Figure 2.12. Humid Páramo, Laguna Culebrillas...... 35

Figure 2.13. Ecuadorian Amazon rainforest...... 37

Figure 2.14. Anthropogenic landscape of highland Ecuador...... 38

Figure 3.1. Map of Ecuador showing archaeological and lake pollen core sites with direct evidence for plant resources...... 42

Figure 4.1. Map of Ecuador showing the location of Formative Period Archaeological sites from which samples were obtained...... 83

Figure 4.2. Relief map of Ecuador showing the location of archaeological sites investigated...... 84

Figure 4.3. View of Santa Ana-La Florida and the Valladolid River...... 85

Figure 4.4. Modern slash-and-burn cultivation near SALF...... 87

Figure 4.5. Map of the SALF site showing major stone constructions and topography... 88

Figure 4.6. SALF stone retaining walls...... 90 xiii Figure 4.7 SALF ceremonial hearth with spiral stone constructions...... 91

Figure 4.8. Schematic of the stone-lined shaft tomb showing some of the burial offerings...... 93

Figure 4.9. The 1996-2000 Excavations at Chaullabamba...... 101

Figure 4.10. Map showing the location of one of the Formative Period site clusters in the Catamayo Valley...... 107

Figure 4.11. View from the north of Trapichillo in the Catamayo Valley...... 108

Figure 4.12. A portion of the La Vega site, Catamayo valley...... 109

Figure 4.13. Canal-irrigated manioc field in the Catamayo Valley...... 110

Figure 4.14. View of the Catamayo Valley from Trapichillo...... 112

Figure 4.15. View of Cerro Narrío from the southeast...... 114

Figure 4.16. View of the northeast hill projection of Cerro Narrío...... 115

Figure 4.17. Southwest wall of Unit 3 Cerro Narrío...... 120

Figure 4.18. Opening Unit 3A at Cerro Narrío...... 122

Figure 4.19. Cerro Narrío Unit 4, bottom of Level 4...... 124

Figure 4.20. Opening Unit 7, Cerro Narrío, in 2009...... 126

Figure 4.21. Map of the La Chimba site...... 132

Figure 4.22. Tajamar site excavation, structure 1...... 136

Figure 5.1. Example of starch granule, Tropaeolum tuberosum (mashua/mashwa)...... 146

Figure 5.2. Arracacha (Arracacia xanthorriza, zanahoria blanca) roots...... 154

Figure 5.3. Example of sonication procedure to remove residues from a stone tool surface...... 156

Figure 5.4. Ceramic sherd from Trapichillo (Catamayo Phase B, CR-70) being sampled for charred interior residues...... 158

Figure 5.5. Stirrup-spout bottle from Santa Ana-La Florida during and after sonication procedure to remove interior residues...... 159

Figure 5.6. Air vents in the South American Research Laboratory covered with 3M Filtrete© High Performance Electrostatic Filters...... 177

xiv Figure 6.1. Stone Bowl #1 from SALF exterior bowl sonication (Sample STR-47)...... 193

Figure 6.2. Unknown starch granule types from SALF...... 196

Figure 6.3. Example of Capsicum spp. starch granule from STR-52(A-B)...... 198

Figure 6.4. Example of Dioscorea spp. starch granule from SALF...... 199

Figure 6.5. Example of Fabaceae (cf. Phaseolus spp.) starch granule from SALF...... 201

Figure 6.6 Examples of Ipomoea spp. starch granules from SALF...... 203

Figure 6.7. Example of typical Manihot esculenta (manioc) starch granule from SALF residues...... 205

Figure 6.8. Example of Maranta spp. starch granules from SALF...... 207

Figure 6.9. Example of Theobroma spp. starch granule from SALF...... 210

Figure 6.10. Examples of comparative Zea mays starch granules...... 212

Figure 6.11. Examples of Zea mays starch granules from SALF artifact residues...... 213

Figure 6.12. Example of Fabaceae (cf. Phaseolus spp.) starch from Chaullabamba. .... 217

Figure 6.13. Comparative starch granules of Oxalis tuberosa, Ullucus tuberosus and Solanum tuberosum...... 219

Figure 6.14. Oxalis spp. starch granules from Chaullabamba...... 222

Figure 6.15. Example of Solanum spp. starch granules from Chaullabamba...... 223

Figure 6.16. Examples of Zea mays starch granules from Chaullabamba...... 224

Figure 6.17. Examples of Fabaceae (cf. Phaseolus spp.) starch from Tajamar residues.227

Figure 6.18. Examples of Oxalis spp. starch from Tajamar residues...... 228

Figure 6.19. Examples of Solanum spp. starch from Tajamar residues...... 229

Figure 6.20. Examples of Zea mays starch granules from Tajamar artifact residues. .... 230

Figure 6.21. Examples of Capsicum spp. starch as compared to Type 1 lenticular starch granules from La Vega...... 232

Figure 6.22. Examples of Fabaceae (cf. Phaseolus spp.) and Manihot esculenta starches from La Vega...... 234

Figure 6.23. Examples of Zea mays starch from La Vega...... 235

xv Figure 6.24. Examples of Fabaceae (cf. Phaseolus spp.) and Manihot esculenta starch granules from Trapichillo, CR-70 residue...... 239

Figure 6.25. Examples of Zea mays starch granules from Trapichillo, CR-70 residue. . 240

Figure 6.26. Examples of Fabaceae family (cf. Phaseolus spp.) starches from Cerro Narrío...... 243

Figure 6.27. Examples of Oxalis spp. starches from Cerro Narrío...... 244

Figure 6.28. Examples of Solanum spp. starch granules from Cerro Narrío...... 245

Figure 6.29. Examples of Zea mays starches from Cerro Narrío...... 246

Figure 6.30. Examples of Solanum spp. starch granules from La Chimba...... 248

Figure 6.31. Examples of Zea mays starch granules from La Chimba...... 249

Figure 6.32. Cerro Narrío Unit 3A showing ashy cultural layer...... 254

Figure 6.33. Cerro Narrío Units 3 and 3A showing backhoe trench bisecting the two units...... 256

Figure 7.1. Cumulative Stages of Patterning for Starch Granule Data...... 277

Figure 7.2. Proposed regions of domestication for plant species discussed in the text. . 312

Figure 7.3. Quinoa (Chenopodium quinoa) growing in highland Ecuador...... 319

Figure 7.4. Lupinus mutabilis growing near Quito, Ecuador...... 327

Figure 7.5. Meal of freshly harvested boiled sweet manioc, fresh papaya, wild tomato salad, and, the ever present, introduced rice...... 334

Figure 7.6. Cañari women harvesting potatoes with a foot plow...... 344

Figure 7.7. Theobroma cacao tree with pod growing in the Santa Ana-La Florida site garden...... 347

Figure 7.8. Theobroma cacao pod split open to show seeds and pulp...... 348

Figure 7.9. Maize drying near Saraguro, Loja. Photo taken July, 2008...... 353

Figure 7.10. Major cultural interaction spheres based on botanical data...... 356

Figure 7.11. Ubiquity and Relative Frequency of Plants Identified by Starch Granule Analysis...... 369

Figure 7.11. Modern “Compact Vertical” agricultural system near Cerro Narrío...... 375

xvi

Epigraph

“There are many warm valleys where fruit trees and pulses are cultivated all the year round... Of provisions, besides maize, there are two other products which form the principle food of these Indians. One is called potatoe [sic], and is a kind of earth nut, which, after it has been boiled, is as tender as a cooked chestnut, but has no more skin than a truffle, and it grows under the earth in the same way... The other food is very good, and is called quinoa... of these seeds they make a drink, and also eat them cooked, as we do rice... [and] there are many other seeds and roots.. and these Indians are industrious... ”

Pedro de Cieza de León

Description of the valleys near Quito, ca. 1540

xvii

Chapter One: Introduction

1.1 Introduction

The Formative Period in the highlands of Ecuador, when ceramic-making people

first settled in villages and were reliant on food production, is poorly understood.

Because Formative Period sites appear earlier on the coast, sedentism, food production

and cultural complexity in the highlands is thought to have been stimulated by interaction

with coastal Formative Period groups. Previous botanical research from highland

Formative sites shows that highland domesticated plants and maize formed the basis of plant subsistence. The presence of already-domesticated high-elevation Andean plants at the outset of the highland Formative Period suggests not only that food production has a greater antiquity in the highlands than the current record attests to, but that food production may have developed independently in the highlands; that is, without stimulus from the coastal lowlands. Further research also suggests that maize (corn) may have been grown in the highlands much earlier, and that the route of its dispersal, both to the highlands and to the coast, may have been from the Amazon basin, raising further doubts that the coastal region cultures were a catalyst to both the adoption of food production and cultural complexity in the highlands. This dissertation involves botanical investigations, primarily starch analysis, from highland sites to create a regional synthesis of Formative Period plant utilization. This basic research will allow larger issues to be explored, such as human adaptation in the Ecuadorian highlands during the Formative

Period, the nature of plant food production and agricultural strategies, interaction spheres,

and the stimuli to cultural complexity.

1.2 The Diversity of Andean Environments

From a physical perspective, South America is one of the most diverse regions in the world (Caviedes and Knapp 1995:21-96), and Ecuador typifies this. Generalized maps of the major biomes found in Ecuador – the coastal lowlands, the highlands (and the cloud forest regions that straddle it), and the eastern lowlands – seem relatively uniform, but there is considerable microdiversity and regional variation within each. The diversity in the physical environment creates biological diversity as well, both vertically as climates transition from tropical to cold, and horizontally, as landforms and precipitation can fluctuate, often within short distances (Denevan 2001:9). These factors create a patchwork of local habitats in the Andes, where broadly tolerant plants species and crops overlap (Zimmerer 1999). The indigenous people adapted by domesticating native plants and developing agricultural strategies suited to the environmental vagaries of the Andean highlands. One of the most important characteristics of this adaptation is the complementary use of different zones to produce a variety of crops. Even varieties of the same domesticated plant species have been adapted to different microzones. A single household may have more than twenty small fields to take advantage of different temperature, soils, and moisture, thus reducing risk, enhancing diet breadth, and maximizing output while minimizing effort (Denevan 2001:9; Hastorf 1993:28-30). The

Ecuadorian highland Formative Period, then, must be understood within this unique and challenging environment.

1.3 Statement of the Research Problem

The goal of this research is to gain an understanding of the nature of human adaptation and agricultural systems in the Ecuadorian highlands during the Formative

Period. In comparison to our knowledge about the lowland coastal region, the Formative

Period in the highlands of Ecuador is poorly understood. The period of time in SW

coastal Ecuador when settled village life began (ca. 3500 BC), accompanied by food

production and ceramics, is known as the Early Formative. In the highland regions,

permanent villages concurrent with ceramics and food production do not seem to appear

until the time period corresponding to the coastal Late Formative, ca. 1400-1200 BC

(Bruhns 2003:131-132) - a gap of about 2000 years. Present evidence indicates that the

coastal lowlands were more densely populated and exhibited greater social complexity

than the highlands during the Formative Period. Because of this, it is assumed that

sedentism and socio-political complexity in the highlands, supported by an agricultural

system where maize was preeminent, was stimulated by contact with coastal Ecuador

(Bruhns 2003:131-132). Zeidler (2008:471) notes that this state of knowledge may be due

more to the disparity between archaeological investigations in the two regions than

reality.

With respect to domesticated plants, from the limited botanical data available

maize is present immediately in highland Late Formative village sites along with a group

of Andean domesticates, such as potato, oca, quinoa and chocho/tarwi (Pearsall 2003a).

As noted, this sudden appearance of sedentism and a fully-developed agricultural system

was seemingly from interaction with coastal populations. However, the abrupt

appearance of crops, whose origins of domestication are in the highlands and which are

not present in coastal Early Formative sites, strongly suggests that Andean domesticates were being grown earlier than the Late Formative in the highlands. Further, the Andean system of agriculture, based strongly on carbohydrate-rich root crops, is a unique system that utilizes ecological zones from different elevations (e.g., Hastorf 1993:26-27) and must be studied and understood in its own right. These are the gaps in knowledge that will be the focus of my research. That is, were Andean domesticated crops present earlier than the Late Formative in highland Ecuador and when was maize introduced? If highland domesticated crops (with or without maize) are present at sites coeval with

Formative coastal sites, then it cannot be assumed that the transition to sedentism and an agricultural lifeway was stimulated by contact with coastal Ecuador. In fact, food production may have developed independently in the highlands or may have been stimulated from the Amazon lowlands.

Importantly, many of the food plants utilized at coastal Early Formative sites have their domestication origin to the east, in the Amazon Basin and seasonally dry, low to mid-elevation tropics (Piperno and Pearsall 1998; Pearsall 2008), including manioc, chili peppers, llerén, arrowroot, and achira (Pearsall 2003a). While these crops, with the exception of achira, are not suitable for highland cultivation, they must have been transported over the highlands to the coast after their initial domestication in the Amazon basin and its margins, indicating early and widespread interaction.

To investigate these questions, my research focuses on a number of previously excavated sites, as well as new excavations at Cerro Narrío in the highlands.

Archaeological sites from both the northern (La Chimba, Tajamar) and southern highlands (Cerro Narrío, Chaullabamba, La Vega, and Trapichillo), as well as one site

from the eastern slopes (Santa Ana-La Florida), are included to develop a regional

synthesis of highland Formative Period human adaptation and food production. As highland economies and diet today, as in the past, are primarily focused on starchy roots,

tubers, maize and beans, the analytical technique I employed is starch granule analysis.

The lowland coastal Early Formative site botanical record forms the basis of comparison

to the new information I obtained from the highland sites. The sites investigated in the far

southern highlands, considered a crucial area for trade routes due to its lower elevation

than any other area of the highlands (Braun 1982), provide key information on

directionality and cultural interaction between the Amazon lowlands, the highlands, and

coastal Ecuador. These sites include Trapichillo and La Vega in the Catamayo Valley,

and Santa Ana-La Florida on the southeastern Andean slopes. The Santa Ana-La Florida

site has components that have been dated to the Early Formative (Valdez 2008). This site is particularly important as it is located on the eastern slopes of the upper Amazon/Andes at an elevation of 1040 masl, in an ecological transition zone. As such, botanical research here may identify crops that are coeval with the coastal Early Formative sites, indicating a possible route of dispersal both to the highlands and the coast. To summarize, the specific questions I will be seeking answers to are:

1. What domesticated plants were utilized during the Formative Period in the highlands?

2. When do the sites where domesticated plants are present date to?

3. What ecological zones were being utilized at each site?

4. Are there any regional variations in the types of domesticated plants present?

5. Is there any regional variation in when domesticated plants are present?

6. Do highland domesticates appear at the same time or earlier than maize?

7. What do these data tell us about human adaptation in a high-elevation environment and

cultural interaction spheres?

This research will shed light on, and synthesize data on, plants in the highlands

during the Formative Period, as well as test the assumption that the adoption of food

production in the highlands was stimulated from outside the region, in particular coastal

Ecuador. An ecological approach will be used to understand human adaptation and food

production in a diverse and rugged environment, which by its nature cannot be

understood as an offshoot of a lowland tropical agricultural system (Piperno and Pearsall

1998:5). The results of this research contribute to understanding the nature of the

Formative Period in the Ecuadorian highlands where a paucity of information currently

exists. Importantly, this research establishes a baseline from which additional problems

can be addressed by future research.

1.4 Theoretical Approach to the Research

Paleoethnobotany is the study of the interrelationships between people and plants as manifest in the archaeological record. It is much more than just identifying plants by their preserved remains, but seeks to understand how people affected, and were affected by, plant resources, how these interactions shaped cultural practices, and how people

adapted to and shaped their landscapes (for example, see Ford 1979; Hastorf and Popper

1988; Pearsall 2000, 2004; Piperno 2006a). Even within the disciplines of anthropology

and archaeology, paleoethnobotanical research is often seen as atheoretical and

environmentally deterministic – a critique that is not applied to the study of the early human (and hominid) history, which is also rooted in environmental adaptation and

evolutionary theory (Crumley 1994:2-4). As Crumley (1994:3) notes “archaeologists and

physical anthropologists rail against their sociocultural colleagues, who put issues in

human adaptation, genetics, and population aside in favor of textured analysis of

contemporary human thought and action.” Paleoethnobotany, then, seeks to understand

the results of human choices and activities in an ecological system that encompasses both

human and nonhuman components.

The ecological approach I will be employing in this research is primarily driven

by evolutionary theories as applied to human behavior, termed evolutionary or behavioral

ecology (Smith and Winterhalder 1992), as well as historical ecology (Crumley 1994).

Behavioral ecology was developed by biologists seeking to understand how the behavior of organisms contributes to their reproduction and survival in relation to their ecology

(Hawkes et al. 1997; Krebs and Davies 1993, see also Piperno and Pearsall 1998). As applied to humans, behavioral ecology emphasizes decision-making to adjust to varying ecological circumstances, and how the outcomes of these decisions, filtered through reproductive fitness, are replicated in subsequent generations (Piperno and Pearsall 1998;

Smith and Winterhalder 1992). Thus, behavioral ecology seeks to understand the core processes – the selective pressures – that were working to favor the establishment of food production and other behavioral changes. As Piperno and Pearsall (1998:16) point out, the issue of intentionality, which has long been a source of debate in the use of evolutionary theory as applied to humans (e.g., Dunnell 1989; Rindos 1984; Sahlins

1972), is not denied in behavioral ecology where human behavior and motivation are accepted as strong components. While an ecological approach will be used in this dissertation to understand the nature of human adaptation leading to food production

strategies in the highlands of Ecuador, I believe that these strategies developed within an

historical context of human selection and shared cultural knowledge.

Historical ecology provides a complementary perspective for understanding

human-environmental interactions. Historical ecology as a discipline has been described

as tracing “the ongoing dialectical relations between human acts and acts of nature, made manifest in the landscape” (Crumley 1994:14, emphasis in original). While landscapes

can be altered and created by unintentional human activities (see below), historical

ecology focuses on intentional decisions by people within the milieu of indigenous

environmental knowledge, especially an understanding of resource management and

creation (Erickson 2008:158). Whereas I acknowledge that unintentional human activities

can result in landscape changes, unintentional human activities are not always “bad” or

destructive, but may result in new opportunities, and purposeful decisions are not always

“good” or sustainable, and may lead to reduced biodiversity and/or environmental

degradation to which humans must then adjust. I believe, therefore, that humans are

neither, innately, agents of conservation, nor agents of destruction, but that humans make

decisions on how to best manage resources, and changes to resources, environments, and

landscapes, based on cultural knowledge. As such, rather than simply relying on a site-

by-site analysis of the botanical information gained through my analysis and that of

others for highland Ecuador, I interpret the entire botanical record within, and between,

the landscapes represented based on behavioral and historical ecological perspectives.

Other theoretical models are embedded throughout this dissertation, where topical.

1.5 Key Concepts Related to Plant Domestication and Agricultural Systems

Conceptually, two types of domestication can be distinguished: plant population

domestication; and landscape domestication. Plant population domestication is a co-

evolutionary process whereby human selection on the phenotypes (observable

characteristics or traits) and genotypes (inherited genetic code) of managed or cultivated

plants results in changes to the phenotypes and genotypes of subsequent populations,

which make them more useful and better adapted to human management of the landscape

(Clement et al. 2010:72-73). Landscape domestication involves intentional and unintentional human activities that transform the environment into a landscape more productive for humans and other economic species (Erickson 2006, 2008). These human activities may include the (intentional and unintentional) transplantation of plants and animals, encouragement of economic species, culling of non-economic species, creation of settlements, burning, and farming. Human activities also vary in degree (from subtle landscape changes to highly engineered anthropogenic landscapes), in time and across space, making them historical processes (Erickson 2008:158), as discussed previously.

Plant population domestication and landscape domestication are intertwined because domesticated plant populations usually require landscape management, especially cultivation (ground preparation and improvement, planting, tending, and harvesting)

(Clement et al. 2010:73).

Domestication is considered the end-point of a continuum resulting in the

“domestication syndrome”, which generally makes a species more dependent on humans for its growth and reproduction and less capable of survival in the wild (Pickersgill

2007:925). Domestication syndrome changes that occur in plant species under

domestication may include loss of dispersal mechanisms, increase in size (especially the

plant part selected for), increased morphological variability (size, colour, shape, etc.),

changes in plant habit (lateral branching, synchronization in maturity, etc.), loss of seed

dormancy (rapid germination), loss of mechanical or chemical protection (toxic

compounds, spines), and photoperiodism (adaptation to different day length) (Pickersgill

2007:930-932). However, the degree of changes can vary within and between species

cultivated (Clement et al. 2010:73-74; Pickersgill 2007:925). Wild species are those

where there are no human-mediated changes. Incipiently domesticated species have

undergone a founder/bottleneck event, whereby humans selected a small sample from the

wild population and have propagated descendents from this sample that show reduced genotypic diversity, but limited phenotypic diversity. Semi-domesticated species have undergone several founder events that have further reduced their genotypic diversity, and

their phenotypic diversity has now been enhanced through selection by humans. While

these distinctions are useful in genetic studies to reconstruct phylogenies and the origins

of domestication for different species, they do not have much applicability to this

dissertation, other than in Chapter 7, where I discuss the domestication origins for the

plant species identified. It is not always possible to identify a domesticated species from a

wild species based on starch granule morphology (in Chapters 6 and 7 I especially

discuss this in relation to Dioscorea). This issue – wild verses domesticated plant use – is

discussed further in Chapters 6 and 7.

I use the terms cultivation, horticulture, and agriculture interchangeably to

describe plant food production. Cultivation refers to activities, such as sowing, planting,

weeding, watering, fertilizing, etc., to grow wild and/or domesticated plants. Horticulture

is often used to describe cultivation of small gardens and fields by hand, often with multiple crops. Agriculture describes a larger field where often one type of crop is grown, and may involve the use of animal traction or machinery (Denevan 2001:11). I use the term agriculture for any type of field where crops are cultivated, and so my use of the term does not imply monoculture or the use of animal traction (and certainly not machinery).

The classification of agricultural systems relates to human management of the natural environment, and includes the crops grown, tools used, and the processes and forms of environmental modification (Denevan 2001:13). Denevan (2001:13-15) uses the term “ordinary fields” to characterize cultivated plots where there is no obvious modification of the natural landscape by earth/stone works or by excavation. These fields are usually found in the humid midlatitudes and wet, montane or lowland tropics, with the latter two environments represented by the highland and eastern Andean slopes under consideration in this dissertation. Because there is little or no modification to the natural environment, ordinary fields are difficult to identify archaeologically, although they may be distinguished based on palynology, phytolith analysis, associated vegetation patterns, soil chemistry analysis, and other soil/sediment characteristics (e.g., anthropogenic dark earths) (Denevan 2001:13). Ordinary fields can also be divided between rainfed fields

(temporales), where precipitation is adequate for crop growth, and water concentration fields, where surface or ground water is concentrated, such as through seepage or springs, high water tables, and/or fields naturally irrigated on floodplains (Denevan 2001:15-16).

Different agricultural systems associated with rainfed agriculture include annual cropping, dry farming, shifting cultivation and agroforesty. Annual cropping is when fields are permanently cropped without the need for major resource management, and it requires fertile soil and rainfall (Denevan 2001:15). Dry farming is non-irrigated

agriculture and is practiced in regions with marginal rainfall. Shifting cultivation, also

called swidden agriculture, is when a field is cultivated for a few years followed by a

short or long period of bush or forest fallow, and is usually practiced in the tropical

lowlands. The fallow period is generally determined based on soil fertility. Agroforesty is

usually rainfed and refers to the combination of annual crop cultivation with perennial

tree crops and/or managed natural vegetation (Denevan 2001:15-16).

Landform modification associated with water deficit (i.e., not ordinary fields)

may involve canal irrigation and field ditches. Canal irrigation can be as simple as

unlined channels of earth that divert water from high points and direct it to fields at lower

points on the landscape, but canals can also be much more elaborate. Field ditches, or

furrows, are distinct from the actual canals and are the furrows within fields that are fed

by water from the canals (Denevan 2001:18-22).

There are many other types of classifications for agricultural systems, but the

types I define above are what I have observed in practice in the regions immediately

close to many of the archaeological sites under consideration. These agricultural systems

and techniques will be referred to throughout the following chapters, where appropriate.

1.6 Organization of the Dissertation

Chapter 2 presents Ecuador’s natural setting, and discusses the diversity of its physiography, climate, including precipitation and temperature variations, volcanism, and vegetation. The major physiographic divisions, the coastal plain (or lowlands), the highlands, (or sierra) and eastern lowlands (or selva) are followed throughout the chapter in discussing the variations in climate and vegetation (especially) within and between these broad regions. I also discuss the long-standing human modification of the highland valleys as well as the types of crops grown in the highlands today.

The next chapter, Chapter 3, presents an overview of the Preceramic and

Formative Periods in Ecuador, with particular focus on the plant resources identified through archaeological research for these cultural periods. This chapter provides the cultural and plant economic background from which the highland Formative Period is understood. I first discuss the preceramic precedents for the coast, highlands and eastern lowlands and their records of plant use. This is followed by the Formative Period for the coastal lowlands, which forms the chronological framework from which the highlands and eastern lowlands are understood. The record of coastal Formative Period plant use is also discussed. In the final section, I present the Formative Period for the highlands and eastern lowlands and their records of plant use. Collectively, the costal, highland and eastern lowland cultural chronology and record of plant use form the basis for comparison for the new data realized by starch analysis.

A detailed background of the archaeological sites investigated by starch analysis is provided in Chapter 4. The sites investigated include La Chimba and Tajamar, located in the northern highlands, Cerro Narrío and Chaullabamba, located in the southern

highlands, Trapichillo and La Vega, located in the far southern highlands, and Santa Ana-

La Florida, located on the southeastern slopes of the Andes. The background information

is based on past excavations, which are reviewed, as well as past and recent excavations

for Cerro Narrío. General site descriptions and locations, details on the settlement

patterns, if known, and known absolute dates are provided. The local environmental

settings for the sites are also discussed, as well as the types of crops that can be grown in

each site location. Finally, paleoevenivonmental reconstructions are provided for each of

the sites for the time period (Early to Late Formative) under consideration.

Chapter 5 is a detailed description of the starch analysis sampling, laboratory

procedures, and identification methods used. The beginning of the chapter provides the

theoretical background for archaeobotanical sampling. This is followed by a review of

the nature and production of starch granules, as well as validation for the use of starch

analysis to identify plant use from archaeological contexts. As starch granules recovered

from archaeological contexts are identified by direct comparison to modern specimens, as

well as published descriptions, the botanical species that comprise my comparative

collection are presented, as well as the methods used to prepare them. The next section

reviews the methods used to recover starch granules from the archaeological samples,

including stone tools and ceramic residues, as well as the methods used to isolate starch granules from the sampled artifact residues. A full list, sorted by archeological site, of all

of the samples tested for starch granules is provided. One of the major sections of this

chapter is the methods I employed to control for contamination of the archaeological

samples by starch granules from modern sources. The final section details the methods

used to identify and classify the archaeological, unknown starch granules to taxa.

The results of the starch analyses are presented, by site, in Chapter 6. Accelerator

Mass Spectrometry (AMS) radiocarbon assays were also obtained on at least one ceramic

charred residue sample from each of the sites. These are presented by site and are also

discussed, both collectively and in relation to previous absolute dates for the sites investigated, at the end of the chapter. For each site, I review the starch granule types identified and make identifications to taxa based on the reference collection presented in

Chapter 5, as well as previous published descriptions. The identifications are highly

conservative as they are based solely on the starch granule comparative collection,

published descriptions, and a consideration of related species in the region.

Chapter 7 considers the results realized from the starch analysis and AMS

radiocarbon determinations within the context of previous archaeological research, that

is, the complete, as is known, archaeobotanical record for the Formative Period in the

highlands and eastern lowlands. Because ancient starch analysis is a relatively new field,

many archaeologists and people from related disciplines are not familiar with, and may

be uncertain about, its application. I begin the chapter, then, with a discussion of the

methodological issues related to starch analysis as associated to this dissertation. Topics

discussed include the validity of the starch granules recovered from the analysis, as well

as assessing the methods used for sampling and identification. I follow this with a discussion of issues related to quantifying starch granules recovered from archaeological

samples, and how cultural, non-cultural, and archaeological methods skew the starch

granule assemblages recovered. I then discuss the relationship and problems associated

with starch granule assemblages recovered from ceramic charred residues with respect to

quantification (and relative abundance) based on stable isotope analysis of ceramic food

residues. In the following section I begin to discuss the problems related to the highland

Formative Period chronology, as a whole, and with respect to the ceramic charred residue

AMS radiocarbon dates presented in Chapter 6 for the archaeological sites investigated, as well as in comparison to coastal Formative Period absolute dates. Based on the review and reassessment of highland Formative Period absolute dates, in particular for the

Cotocollao site, the next section discusses the paleoethnobotany of the highlands and eastern lowlands. The first part of this discussion provides corroborating evidence to show that the taxa identified (very conservatively) in Chapter 6 are, in most cases,

representative of domesticated species. Then, based on the entire highland and eastern

lowland Formative Period botanical record, I discuss the geographical origins, ecological

requirements, and nutritional aspects of the plants identified. These details are then used

to discuss cultural interaction spheres, as evidenced by the geographical origins of the

plants identified. The nutritional aspects of the plants identified are used to discuss the

complementary use of plants in various combinations (and elevations) in providing a

balanced diet, again, as evidenced by the plant species identified. The final section

assesses the use of ecological zones, and the types and nature of agricultural systems that

were likely employed at the archaeological sites. These aspects are considered both

within and between broad regions, based on the plant species identified, their ecological

requirements and the local landscape of the archaeological sites.

In the final concluding chapter, Chapter 8, I summarize the results of the

dissertation and use the interpretations from previous chapters to draw some final

inferences about the timing and nature of highland Formative Period human adaptation,

food production, and cultural interaction in Ecuador. Future avenues of research are also detailed.

Chapter Two: Overview of Ecuador’s Natural Setting

2.1 Introduction

The following is a general overview of the natural environments of Ecuador.

Further, more specific, information on the environment settings of the archaeological

sites will be detailed in Chapter 4. Also included in Chapter 4 are paleoenvironmental reconstructions specific for each archaeological site investigated. Unless otherwise noted, the following description on the physiography of Ecuador follows Balsev (1988).

2.2 Physiographic Setting

The Republic of Ecuador (Figure 2.1) is located along the northwestern coast of

South America and spans the equator from 1° 30’N to 5°S. From its most western shores along the Pacific Ocean at 81°W it extends over the Andes and into the Amazon Basin to

75°W (Figure 2.1). Ecuador is politically bounded by Colombia to the north and Peru to the south and east. Lying along the equator, one would think that the country has a uniform tropical climate but the striking topography of the country results in remarkable variations in climate, plant and animal life. Undoubtedly the most prominent physical feature of Ecuador is the Andes. The Andes are relatively young and have been rising for the last 20 million years due to the subduction of several tectonic plates (principally the

Nazca Plate) under the South American Plate. This has transformed what was a rather flat continent into one with a strong physical separation between the mountain environments and the lowlands (Bush et al. 2007:33-34). The Andes massif is so vast that it influences hydrological regimes and air circulation across South America (Young 2011:128). In

Ecuador, The Andes separate the country into three broad (but not homogenous) zones:

The Coastal Plain (la Costa), the Highlands (or Sierra), and the Eastern Lowlands (Selva).

Figure 2.1. Political map of South America showing the location of Ecuador.

Although the 100-200 km wide Coastal Plain appears flat in comparison to the

Andes, it does have a low row of hills (the Colonche Hills) approximately 20-40 km from

the coastline. Basins in the southern half of the plain include the Daule, Vinces, and

Babahoyo rivers that all drain into the Guayas River and ultimately into the Gulf of

Guayaquil. Several smaller rivers drain the northern Coastal Plain, with the Santiago and

Esmeraldas rivers being the most important. Figure 2.2 shows the major physiographic

features discussed.

Figure 2.2. Relief map of Ecuador with physiographic features. Viewed from the east with major physiographic features: 1) Coastal Plain; 2) Colonche Hills; 3) Guayas River; 4) Highlands (Sierra); 5) Inter-Andean Valley; 6) Cordillera Occidental; 7) Cordillera Oriental; 8) Eastern Lowlands (Selva); 9) Serranías. Adapted from Souris (2011a).

The Sierra is 150 km wide and here, in contrast to the Peruvian Andes to the south and the Columbian Andes to the north, the belt is narrower and is arranged in two roughly parallel rows of ridges and peaks, separated by the Inter-Andean Valley (Balsev

1988:568; Buitrón 1952:129; Sauer 1963:328). The highest peaks are over 6000 meters above sea level (masl) in elevation and the ridges connecting them are usually about 4000 masl. Figure 2.3 shows the highest peak in Ecuador, Chimborazo, once thought to be the

highest mountain in the world (its summit is still the furthest point from the Earth’s

center, owing to the equatorial bulge) (Parsons 1982:255).

Figure 2.3. Chimborazo, the highest mountain in Ecuador. It is located in the Cordillera Occidental and is an inactive stratovolcano. Seen from Riobamba, at 2757 masl, Chimborazo loses more than half its true height of 6,310 masl (Parsons 1982:255). The best perspective to view its immense height is from the Pacific lowlands. Photo taken July, 2009.

The eastern chain (Cordillera Oriental or Cordillera Real) is older than the western chain (Cordillera Occidental) and is also wider, higher, and more uniform (less jagged) (Buitrón 1952:128-129; Vuille et al. 2000:2520). The Inter-Andean Valley, a 300 km-long structural depression, is about 20-30 km wide (Hall et al. 2008:1), varies from

2600-2700 masl, and is divided by transverse ranges (nudos or “knots”) that connect the two cordilleras (this can be envisaged as a ladder, with the cordilleras forming the sides of the ladder, the rungs representing the nudos, and the spaces between the rungs as the

valleys). Both the cordilleras and the transverse ridges between them delimit Inter-

Andean basins (hoyas or cuencas), such as those of Quito, Riobamba, and Cuenca, causing them to either drain to the east to the Amazon or to the west to the Pacific through steep-sided valleys (Balsev 1988:568; Mörner 1985:2-3; Parsons 1982:255;

Sauer 1963:329). South of latitude 2°30’ S (at approximately the modern town of Cañar), a single dominant central valley is not present and the Andes form a broad mountain range (Hall et al. 2008:1). Further south towards the Peruvian border, the Andes are of lower elevation than in northern Ecuador, rarely above 4000 masl, and form a complex pattern of ridges of different orientations. The lower overall elevation of the southern

Highlands is clearly visible on the landscape map Figure 2.2 and the topographic map

Figure 2.4.

The Eastern Lowlands of Ecuador comprise the western edge of the Amazon

Basin. Numerous rivers that drain many of the highland basins and the Andean slopes plunge down steep-sided valleys to traverse the flat plain of the Selva, flowing east to the

Amazon River. The Aguarico, Napo, Curaray, Pastaza and Santiago are prominent rivers

of this region. Low mountain ranges, called serranías, (Figure 2.2) lie parallel to the base

of the Andes here (Hall et al. 2008:1-2).

Figure 2.4. Shaded relief map of Ecuador. Note the low elevation of the Southern Highlands in comparison the Central and Northern highlands.

2.3 Climate

Ecuador’s climate varies according to topography, exposure to interacting air masses and tradewinds, and adjacent sea currents (Young 2011:129-130). The Coastal

Plains are intermediate between the desert conditions on the coast of Peru and the tropical rain forest of Columbia’s coast. The southwestern coast on the Santa Elena Peninsula receives only an average of 100 mm of precipitation per year, while the northwestern

Esmereldas Province receives over 2500 mm per year. The more arid southern conditions are created by the cold Peru (Humboldt) sea current that runs north along the west coast of South America, while the northern part of Ecuador’s coast is influenced by the warm counter-current flowing south along Columbia’s coast creating the El Niño-Southern

Oscillation phenomenon (ENSO) (Vuille et al. 2000:2520). Where the two currents meet

off the coast of Ecuador, they turn west into the Pacific Ocean. Thus, a gradient of 100

mm to 2000 mm of precipitation per year occurs over a distance of only 150 km along the

Coastal Plains. Precipitation on the coast also increases from west to east. Due to its

equatorial location, the average monthly temperature is a fairly constant 24°C at sea

level. Exceptions to precipitation patterns on the southwestern Ecuadorian coast (and

extending well down the Peruvian coast) occur during El Niño years, whereby torrential

rains, high river run-off and floods are evident due to a southward expansion and intensification of the intertropical convergence zone (ITCZ) (Vuille et al. 2000:2520).

Figure 2.5 shows a map of the annual average precipitation for Ecuador, and Figure 2.6

shows a map of the annual average temperatures.

The Sierra’s climate depends on the complex interplay of elevation, exposure, and

air mass patterns. Precipitation on the outer slopes of the highlands (the montañas)

approximates 2000 mm per year, while the Inter-Andean Valley is drier due to the

rainshadow/Froehn effect of air masses passing over the cordilleras and down into the

lower-lying valley (Figure 2.5).

Figure 2.5. Map of annual average precipitation for Ecuador. Precipitation in the Coastal Lowlands increases from south to north as well as from west to east. Precipitation in the Eastern Lowlands is mostly uniformly high, while precipitation in the Highlands varies considerable. Map adapted from World Trade Press (1998- 2011).

In general, the Inter-Andean Valley receives less than 1500 mm of precipitation per year, with the driest part being the deep valleys of the Guallabamba, Chota and Léon rivers receiving less than 500 mm per year. The average yearly temperature is 10°C at

2800 masl and 0°C at 4800 masl (Figure 2.6). The snowline, at 4800 masl, is evident on

only a few high peaks.

Figure 2.5 shows the annual precipitation for Ecuador, but observing yearly averages masks the seasonal pattern of rainfall. By using principal component analysis of tropical Pacific and Atlantic sea surface temperature anomalies (SSTA) and mean monthly precipitation and temperatures between 1963-1992 from a network of stations covering the entire Andean range of Ecuador, Vuille at al. (2000) were able to characterize precipitation and temperature variations in the Highlands. Unlike the Coastal

Plains and the Eastern Lowlands that are affected by air masses originating over the

Pacific and Atlantic Ocean/Amazon Basin, respectively, the Inter-Andean valleys are influenced by both oceanic and continental air masses producing two distinct rainy seasons (February-May and October-November). The June-September dry period is more marked than the second dry period beginning around December. The ENSO is most pronounced in the northwestern Andes during December-February (but is also apparent during June-August and September-November) and the eastern cordillera from June-

August, causing below average precipitation during El Niño years (reversed signal from the coastal lowlands) and above average precipitation during La Niña years (the dominant pattern).

Figure 2.6. Map of annual average temperature for Ecuador. While the Costal and Eastern Lowlands have high annual average temperatures, the annual average temperature in the Highlands is highly variable. Map adapted from World Trade Press (1998-2011).

The western Andean slopes between 1° and 3°S is the only region to have markedly increased precipitation during the ENSO peak phase of December-February

(thus similar to the coast), although even here the signal is weak and not true for all

ENSO events. Rollenbeck et al. (2006) used a combination of climate station measurements and remote sensing (weather radar, satellite imagery) to examine rain distribution in the southern highland provinces of Loja and Zamora-Chinchipe

(4°S/79°W) during a “normal” (La Niña, non El Niño) year. The main rainy season is

from April-May with a second, less-pronounced peak in December. A pronounced dry

season occurs from July-September. Table 2.1 shows the climate zones for highland

Ecuador.

Table 2.1. Names and Altitudes of Climate Zones in Ecuador Latin American Name1 English Name2 Elevation (masl)2 Tierra Caliente Tropical 0 - 800 Tierra Templada Subtropical 800 - 2000 Tierra Fría Temperate 2000 - 3400 Páramo Páramo 3400 - 4600 Tierra Helada Perpetual snow > 4600 1From Clawson (2000:82-83). 2 From Holdridge et al. (1947:8).

The Eastern Lowlands, which are dominated by the moisture-rich easterly trade

winds that originate over the tropical Atlantic Ocean and Amazon Basin (Vuille et al.

2000:2520), have a uniform climate with greater than 2000 mm of precipitation per year

and constant temperatures greater than 24°C. It should be noted that humid tropical lowlands flank both sides of the Andes in Columbia and Ecuador, in contrast to Peru

where tropical forest conditions are only found on the eastern slopes (the ceja de la

montaña, or montaña) of the Andes and Amazonian Lowland regions (Sauer 1963:333).

2.4 Volcanism

The Nazca and South American plates have been colliding since the Eocene, 20

million years ago, causing uplifting of the edge of the South American Plate, formation of

the Andes Mountains running parallel to the Pacific coast, and volcanic activity ever

since. Another effect of the tectonic forces still at work is the long record of earthquakes

in the region. Earthquakes have repeatedly damaged parts of cities such as Ambato,

Ibarra, Riobamba, and Quito (Basile 1974:6).

Figure 2.7. Sangay Volcano located in the Cordillera Oriental in Chimborazo Province. Sangay is the southernmost active volcano in Ecuador. Photo taken May 2010.

Volcanic activity in the Andes Mountains is most pronounced in Ecuador where a chain of stratovolcanoes lie along the two cordilleras and in the Inter-Andean Valley from the Columbian border to 2°30’S. Within the northern Highlands of Ecuador, in an

area 300 km long and 100 km wide, a total of 35 volcanoes occur, eight of which are active (Sangay is the most active of all, Figure 2.7).

Figure 2.8. Tungurahua erupting, Cordillera Oriental in Tungurahua Province, central highlands. Photo taken in May 2010.

The pattern of volcanism in the southern Ecuadorian Andes from 2°30’S to the

Peruvian border is different in that there are no active volcanoes. These mountains date as recently as the Pleistocene (25,000 to 35,000 years ago), they rarely reach over 4000 masl, and they are not arranged along fracture zones. The soils in the southern Highlands are very fertile, though, owing to their volcanic origin, despite the absence of recent volcanic activity. This difference in volcanism between the north and south is a result of the passing of the Nazca plate over the Galapagos Hot Spot, causing increased volcanic activity in the northern Ecuadorian Andes (Zeidler and Isaacson 2003:73).

Archaeological sites are often deeply buried beneath thick mantles of volcanic deposits.

Radiocarbon dating has confirmed that in the last 3000 years Cerro Negro, Cuicocha,

Ninahuilca, Pululahua, Guagua Pichincha, Quilotoa, and possibly Rasuyaca have erupted, resulting in substantial ecosystem disruptions in northern Ecuador oriented on an east- west axis as well as major impacts on past human cultures (Zeidler and Isaacson 2003:73-

74). Thus, past volcanic activity has not only created high soil fertility in both the northern and southern highlands (and fallout zones), but cannot be ignored when considering past cultures of the region.

2.5 Vegetation

Ecuador has one of the richest floras of the world with an estimated 20,000 species.

Of these, more than 4500 or a remarkable 23 percent are endemic. Barthlott and colleagues (2005) claim that the Ecuadorian Andes harbour the most species-rich ecosystems on earth. Figure 2.9 shows a map of Ecuador with its vegetation zones noted.

It must be remembered, however, that a great deal of Ecuador’s landscape, especially the

Inter-Andean Valleys and the Coastal Lowlands, has been converted (since ancient times as well as during modern) to agriculture. Therefore Figure 2.9 shows the potential natural vegetation zones.

Figure 2.9. Vegetation map of Ecuador. Note that this map shows potential vegetation. In fact, most of the coast and highlands have been converted to agriculture. Perry-Castañeda Library Map Collection (1973).

Vegetation patterns are primarily influenced by the combined effects of elevation and precipitation. In fact, owing to its position on the equator and the presence of the

Andes Mountains, most, if not all, of the earth’s ecological zones are present in Ecuador from tropical to alpine (Young 2011:128). On the Coastal Plains, a gradient exists from south to north ranging from deserts, where annual herbs dependent on occasional rains predominate (Figure 2.10), to tropical rainforests, with large trees reliant on continuous high year-round precipitation, prevail.

Figure 2.10. Dry scrub vegetation near Manta, coastal Ecuador. View to the west, Pacific Ocean. Photo taken November, 2008.

The multitude of microenvironments present in the Andes cannot be overstated as

conditions and vegetation can vary greatly even within small areas depending on soil

type, altitude, aspect, and local precipitation, creating a mosaic of different ecological

settings (Bush et al. 2007; Gentry 1988; Sandweiss and Richardson 2008; Young 2011).

Mountain slopes exhibit gradients from lowland deserts to cold highland deserts and from

tropical rainforests to wet alpine vegetation where herbs predominate. The mountain

forests form a continuous belt that covers the eastern slopes of the Ecuadorian Andes,

whereas on the western slopes, especially in southern Ecuador, the mountain forests are

interrupted occasionally by arid valleys with xeric vegetation.

Figure 2.11 Entering the cloud forest, western slopes of the Cordillera Occidental. Note the dense vegetation, ferns, and ever-present mist. Photo taken July 2007.

Humid mountain forests of the Andean slopes (Figure 2.11) from around 1300-

1800 masl are, as mentioned earlier, referred to as the selva de ceja de montaña or ceja de la montaña (“mountain-eyebrow forest” or cloud forest) (Caviedes and Knapp 1995:66).

Cooler mountain forests occur on both the eastern and western slopes of both Cordilleras as well, between about 1800 to 3300 masl. Therefore, caution is required in grouping all mountain forests together because, just as for other zones in the Andes, species composition varies tremendously depending on local conditions, and humans have occupied and modified these landscapes for millennia (Bush et al. 2007:33; Young

2011:128). Above 3500 masl in the highlands, trees are absent and this humid herbaceous vegetation zone is called the páramo (Figure 2.12).

Figure 2.12. Humid Páramo, Laguna Culebrillas. Laguna Culebrillas is located in the Cordillera Oriental at 4173 masl, above the tree-line. Photo taken April 2009.

Grazing and burning has lowered the forest line in many areas and so instead of a gradual shift from forest to shrubby vegetation to open herbaceous vegetation, the transition from trees to open grassland may be abrupt in many areas. Although the species diversity in the páramos is not as rich as the montane or lowland forests, it is not terribly poor either. In 400 m2 of shrubby páramo at 3300 masl, 102 species of vascular

plants and 26 species of non-vascular plants were observed, and at 4200 masl 94 and 82

species of vascular and non-vascular plants, respectively, were observed (Vuille et al.

1988:570). Although the severity of the páramos has resulted in it being compared to

temperate and arctic ecosystems, the páramos have a greater number of degree hours per

year, conferring higher productivity and a higher resilience to agricultural disturbance

(Ramsay and Oxley 2001:161).

Vegetation of the Eastern Lowlands (Figure 2.13) is species-rich, as is typical of

humid tropical forests. In a single hectare plot, trees alone (with stem diameters greater

than 10 cm) usually number over 200-300 species (Valencia et al. 2004:214). Most trees

are evergreen, growing and shedding their leaves simultaneously and continuously,

although some may shed all of their leaves for a short period at irregular intervals

(Cochrane et al. 1985:55). Undisturbed “climax” forests are considered to have a three-

tiered structure – an upper canopy of trees usually over 30 m in height, a second layer of

smaller trees, and a lower story – with trees in these layers often specifically adapted to

their particular niche (Cochrane et al. 1985:55). In addition to the great diversity of trees, lianas and epiphytes are common, while the shrub layer is often underdeveloped (often just ferns and seedlings) under a closed canopy (Cochrane et al. 1985:55). The lack of sunlight results in the forest floor being rather open and easy to walk through to enjoy the

diversity of flora while keeping an eye out for anacondas, jaguars, multitudes of

venomous and non-venomous reptiles, jumping spiders, biting insects of all sorts, and

other strange creatures that await to feast on tender human flesh.

Figure 2.13. Ecuadorian Amazon rainforest. Huaorani village on a tributary of the Napo River, northeastern Ecuador, in the Yasuni National Forest. Photo taken March 2009.

Isolated Inter-Andean valleys may contain many variants of climate and vegetation due to elevational differences, rainshadow effect, slope, aspect, and soils leading to diversity of microhabitats favouring speciation (Josse et al. 20011:152). The tropical Andes are at the top of the list of worldwide hotspots for the number of species/area and endemism (Myers et al. 2000). Jørgensen and colleagues (2011:195) compared the number of species in 500 m elevational bands from Ecuador, Bolivia and

Peru, and Ecuador has the most number of species between 500 and 3500 masl. The tropical Andes though are also one of the most severely threatened regions of the tropics

(Jetz et al. 2007; Mittermeier et al. 2004). As mentioned previously, the landscapes of the

Inter-Andean valleys have been highly modified and reshaped by humans over the past

millennia (Figure 2.14).

Figure 2.14. Anthropogenic landscape of highland Ecuador. The town in the center is Guasuntos, Chimborazo Province, in the Cordillera Oriental at ca. 2800 masl. Photo taken in April 2008.

Forest and other natural ground covers in the highlands have been cleared for growing crops and animal grazing, causing loss to its biological richness (Wassenaar et al. 2007). In the late 1950’s Miller (1959:202-205) described the current and future potential of agriculture in Ecuador, and his observations for the types of crops grown are still applicable to today. Potatoes (Solanum tuberosum) and introduced cereals, such as corn (maize, Zea mays, introduced in pre-Columbian times), barley (Hordeum vulgare) and wheat (Triticum spp.) (in the northern highlands), are grown below the páramo.

Native and introduced crops are also important in the higher elevations of the Inter-

Andean valleys, including white carrot/zanahoria blanca/arracacha (Arracacia xanthorriza), quinoa (Chenopodium quinoa), oca (Oxalis tuberosa), ulluco (Ullucus tuberosus), mashwa/mashua (Tropaeolum spp.), and introduced broad beans (Vicia faba).

In the warmer intermontane valleys, corn is especially important and grown mostly without the need for irrigation. Squashes (Cucurbita spp.) are a companion crop to corn, and rye (Secale cereale, introduced), lentils (Lens culinaris, introduced), kidney beans

(Phaseolus vulgaris), and cabbages (Brassica oleracea, introduced) are also grown.

Introduced forage crops, such as alfalfa (Medicago sativa), ryegrass (Lolium spp.), clover

(Trifolium spp.), and orchard grass (Dactylus spp.), grown in pastures and for harvest, are also important as they are the base of Ecuador’s dairy industry. Eucalyptus (Eucalyptus globules) trees, originally from Australia, are grown in plantations and anywhere that is too steep, stony, or marginal for corn and pasture, and are an important source of domestic fuel. The lower-elevation valleys of the southern highlands are warm and fertile, allowing for sugar cane (Saccharum spp., introduced) and subtropical crops such as avocados (Persea americana), coffee (Coffea spp., introduced), tomatoes (Solanum lycopersicum), and others to be grown.

Archaeological evidence for plants from highland sites is reviewed in Chapter 3, and some further information on potential crops for the sites included in this dissertation are discussed in Chapter 4. As mentioned earlier, information on paleoclimatic/environmental reconstruction is also provided for the archaeological sites investigated in Chapter 4.

2.6 Chapter Summary

Ecuador is a land of diversity. Whether one considers the physical environment,

temperature and rainfall patterns, or vegetation, great variation exists between and within

Ecuador’s three major regions – the Coastal Plain (coastal lowlands or costa), the

Highlands (sierra) and the Eastern Lowlands (selva). The Andes massif affects air

circulation and rainfall patterns not only in Ecuador but across South America. Coastal

Ecuador’s rainfall pattern is also strongly influenced by the cold Humboldt current and

the warm counter-current, with this being reflected in disparate patterns of vegetation

from the xerophytic vegetation seen on the southwestern coast to tropical rainforests in

the northern coastal region. The multitude of microenvironments present in the Highlands

cannot be overstated as conditions and vegetation can vary greatly even within small

areas depending on soil type, altitude, aspect and local precipitation, creating a mosaic of

different ecological settings. The Eastern Lowlands, which are dominated by the

moisture-rich easterly trade winds that originate over the tropical Atlantic Ocean and

Amazon Basin, have a uniform climate and the vegetation is species-rich. Two factors

that have shaped, and continue to re-shape, Ecuador’s landscape are volcanism (and

earthquakes) and humans. Volcanism is most pronounced in the central and northern

highlands (but affects lowland regions as well), and a great deal of Ecuador’s landscape,

especially the Inter-Andean Valleys and the Coastal Lowlands, has been converted (since

ancient times as well as during modern) to agriculture.

Chapter Three: Overview of Ecuadorian Preceramic and Formative Periods and Plant Resources

3.1 Introduction

The Formative Period in Ecuador is based on the chronology developed for the coastal lowland region and is characterized by the appearance of settled villages, food production, ceramics, and established trade networks (Estrada 1956, 1958; Evans and

Meggers 1954, 1957; Hill 1975; Lathrap 1970; Lathrap et al. 1975; Lathrap et al. 1977;

Marcos 2003; Marcos and Michzynski 1996; Meggers 1966; Meggers at al. 1965;

Raymond 1998, 2003a; Raymond et al. 1980; Zeidler 2003, 2008). The four-thousand year Formative chronology is divided into three Periods: the Early Formative (Valdivia), which spans 4400–1450 Cal B.C.; the Middle Formative (Machalilla) from 1450-800 Cal

B.C.; and the Late Formative (Chorrera) from 1300-300 Cal B.C. (Zeidler 2008:460).

The Formative Period is followed by the Regional Development Period, which starts at about 500 BC and lasts until about 500 AD. In comparison to the coastal lowlands of

Ecuador, the Sierra is poorly known due primarily to a lack of archaeological investigations (Bruhns 2003; Raymond 2003b; Zeidler 2003). In the highlands, as on the coast, the Formative Period is preceded by an aceramic Archaic generalized hunting/foraging period. However, in the lowlands the Preceramic Period has evidence for plant domestication and cultivation, whereas in the highlands the period of hunting/foraging extends to the Middle/Late Formative (Bruhns 2003). This chapter reviews the Preceramic and Formative Period archaeology for Ecuador and the evidence for plant cultivation/use during these broad time periods. Sites in the montaña (cloud forest) regions are included with information for the highlands and eastern lowlands.

Figure 3.1 shows both archaeological and lake core sites that have recovered evidence

(carbonized plant remains, phytoliths, pollen grains, and/or starch granules) for plant

resources, as discussed in this chapter.

Figure 3.1. Map of Ecuador showing archaeological and lake pollen core sites with direct evidence for plant resources. Note that the location of the Dos Caminos, Capa Perros, Finca Cueva, and San Isidro archaeological sites is marked by a black oval. Lake Core sites are those that recovered maize (Zea mays) pollen, as discussed in text and shown in Table 3.2. Map adapted from World Trade Press (1998-2011).

3.2 Coastal, Highland and Eastern Lowlands Preceramic Precedents

Unlike the arid coast of Peru where macrobotanical remains are well-preserved, the

early botanical record from Ecuadorian sites, as with most humid Neotropical locations, is based primarily on microbotanical remains (phytoliths, pollen grains, and starch granules) as well as carbonized macrobotanical remains recovered either in situ or by flotation of sediments. The archaeological record of preceramic human occupation, as mostly mirrored by the location and intensity of archaeological investigations, is

incomplete with gaps in information for chronology, settlement patterns and adaptations

(Bruhns 2003:126-127; Idrovo Urigüen 1999:115-116; Lippi 2003:533; Raymond

2003b:547; Rostoker 2003:541; Zeidler 2003:487, 2008:460, 471). The archaeology of

the Preceramic Period is presented first followed by evidence for plant use, for both the coastal lowlands (section 3.1.1) and the highlands and eastern lowlands (section 3.1.2).

3.2.1 The Coastal Lowlands Preceramic Period

The best documented record of early human occupation in Ecuador comes from the Santa Elena region in the western lowlands. It is in this driest part of the Ecuadorian coast (but not so dry as to preserve non-carbonized macrobotanical remains or animal/human tissue) where the Las Vegas people were foraging, hunting, fishing, cultivating plants, and burying their dead between ca. 10,000 and 6600 Cal BP (Stothert

1983, 1985; Stothert et al. 2003).

The Las Vegas Period is divided into Early (ca. 10,000 - 8000 Cal BP) and Late

(ca. 8000 - 6600 Cal BP) Las Vegas (Stothert 1983, 1985; Stothert et al. 2003). While the

Santa Elena Peninsula may have been wetter during the Las Vegas occupation, with a

more vegetated thorn scrub and wooded savannah landscape rather than the xerophytic vegetation now present, it is not likely that the rainforests found today to the north and east extended into the Santa Elena Peninsula (Piperno and Pearsall 1998:184; Stothert

1983, 1985; Stothert et al. 2003). More than 30 small Las Vegas sites have been located, but by far the largest (ca. 2000 m2, 2 ha) and most important is the Las Vegas type-site

(OGSE-80, Figure 3.1). Based on the plant and animal remains recovered from

excavations people travelled over a large territory and exploited a broad range of

terrestrial, littoral and marine habitats – hunting mammals, amphibians and reptiles,

fishing for near shore and marine species (indicating watercraft), and collecting mangrove molluscs and terrestrial plant resources (Raymond 2008; Stothert 1983, 1985;

Stothert et al. 2003). Analysis of human skeletal remains shows that the Las Vegas people were free of anemia, healthy, and relatively long-lived (Stothert 1985, 1988;

Ubelaker 1980). With respect to this dissertation, the most important information from the Las Vegas site is the record of plant cultivation.

Cucurbita (squash) phytoliths are ubiquitous in the Las Vegas site midden. The size of the squash phytoliths increases over time showing progressive so that by ca.

10,100 to 9,300 Cal BP phytoliths consistent with domesticated C. moschata are present

(Piperno 2006a:143; Piperno et al. 2000; Piperno and Stothert 2003:1055; Stothert et al.

2003:35-37). Other domesticates that span the entire Las Vegas occupation include bottle gourd (Lagenaria siceraria) and the root crop llerén (Calathea allouia). Maize (Zea mays) phytoliths occur in deposits dated to before 7000 Cal BP (Piperno and Pearsall

1998:196-198; Stothert et al. 2003:35). Maize phytoliths were recovered from sediments, and starch granules and phytoliths of maize were found in human tooth plaque dating to

ca. 7000 Cal BP (Pearsall and Piperno 1990; Piperno and Pearsall 1998:187-190; Stothert

et al. 2003:35-38). These data are included with those for the Coastal Formative Period

and are shown in Table 3.1.

Following the Vegas Period an apparent 1000-year hiatus in human occupation

occurs in the archaeological record of southwestern Ecuador prior to the appearance of

Valdivia (Early Formative Period) ceramics and settlements. As Raymond (2003a:41)

points out, “it is not believable that prior to 3500 B.C. occupation of western Ecuador

was confined to one little arid corner [the Santa Elena Peninsula], or that the Vegas

community was a population isolate”. Instead, it is probable that several factors

contribute to the inability to find what would be ephemeral sites (as most Vegas sites are)

in locations other than the Santa Elena Peninsula, where sites are exposed to

destructive/obscuring factors such as erosion, deep burial under river alluvium, or

drowning by rising sea levels prior to 5000 BP (Raymond 2003a:41-42)1. Regardless of

the reasons for the absence of sites intermediate between Las Vegas and Valdivia, the

paleoethnobotanical record for coastal Ecuador continues in the Formative Period, as

discussed below in section 3.3.

3.2.2 The Highlands and Eastern Lowlands Preceramic Period

Despite the fact that the record for human presence in the highlands of Ecuador extends back to the Early Holocene from sites such as El Inga, Chobshi Cave, and

Cubilán (Bell 1965, Lynch 1983, 1988; Mayer-Oakes 1986; Temme 1982) there is no direct evidence to date for the types of plants exploited, despite that these early

inhabitants were surely utilizing these resources. However, vegetation disturbance

wrought by increased fire frequency in the southeastern Ecuadorian highlands are argued to be anthropogenic, not only supporting an Early Holocene human presence but also landscape modification. Lake sediment and soil cores in and around Laguna Cocha

Caranga, located in Podocarpus National Park at 2200 masl on the western slope of the

Cordillera Oriental 8 km southeast of Loja, document an abrupt cessation of mountain forest expansion, increased fire intensity, and open grassy vegetation from ca. 9700 to

1300 Cal BP (Niemann and Behling 2009). Niemann and Behling (2009:10-13) argue that these changes were a result of human impact (and not fires of natural origin) due to the limited deviation of the charcoal record through both dry and wet phases, as inferred through pollen, spore, and algae analyses of the cores. They hypothesize that the fires were a result of hunting activities and that the reduction in human activity after ca. 1300

Cal BP was due to “civil conflicts during pre Inca times”, followed by the “resettling politics of the Inca, and the arrival of the Spanish Conquest” (Niemann and Behling

2009:10). Similar results are also documented by Niemann and Behling (2008) for El

Tiro pass. A pollen core from a bog in El Tiro pass, located just north of Podocarpus

National Park at ca. 2400 masl and about 10 km east of Loja, documents that fires attributed to anthropogenic activities became frequent after 8000 Cal BP and lasted until

500 Cal BP, when the fire frequency dropped likely due to a decrease in human population after the Spanish Conquest (Niemann and Behling 2007:206-210). Although the studies by Niemann and Behling do not document specific plant use, they do indirectly document Early Holocene and later landscape changes likely caused by human agency in the southeastern Ecuadorian Andes.

Archaeological evidence for preceramic occupations on the eastern slopes of the

Cordillera Oriental and in Eastern Lowlands is non-existent (Bruhns 2003:157; Rostoker

2003:541). As with the highlands in general and even more so for regions to the east,

archaeological investigations have lagged far behind those on the coast. As noted earlier, the skewed nature of our knowledge is mostly driven by differences in the pace of archaeological investigations, as well as site visibility (Bruhns 2003:128, 157; Rostoker

2003:541; Zeidler 2008:471) and the relative difficulties in conducting surveys and excavations. As Bruhns (2003:125) notes for the Highlands and Eastern Lowlands, “these two major regions have been traditionally seen as hard to work in because transportation is poor and the climate is considered unfavorable (too cold, hot, or damp)”. In the highlands, as with Coastal Ecuador, there is a gap in the record for human occupation between the early Holocene and the Formative Period. But in the highlands, the disparity is striking – no archaeological sites have been located that are intermediate between the

Paleoindian occupations and the apparent sudden appearance of settlements in the

Formative Period, and so nothing is known regarding the adaptations of this extended

“Archaic” period. Based on the archaeological record alone the eastern lowlands were devoid of humans prior to the Formative Period, although this is probably not the reality.

Therefore, despite the archaeological and proxy evidence for early Holocene human presence and landscape modification in the highlands, we have no direct evidence for the use of plants by people in the Ecuadorian Highlands or Eastern Lowlands prior to the

Formative Period, save for the ephemeral presence of a single maize pollen grain at

Laguna Chorreras (as discussed below in section 3.3).

3.3 The Coast, Highlands and Eastern Lowlands Formative Periods

The Formative Period in Ecuador was first identified in the southwestern coastal lowlands and the chronology developed for that area generally serves as a reference point for all of Ecuador. The four-thousand year Formative Period chronology is divided into three cultural complexes based on absolute dating (radiocarbon and thermoluminescence assays, obsidian hydration) and relative dating (tephrostratigraphy and ceramic cross- dating): the Early Formative culture of southwestern Ecuador is known as Valdivia

(named for the type-site and ceramic complex where it was first defined) and dates to

4400–1450 Cal B.C.; the Middle Formative (Machalilla) 1450-800 Cal B.C.; and the Late

Formative (Chorrera) with a start date that overlaps with terminal Machalilla depending on the region but likely ends about 300 Cal B.C. (Zeidler 2003, 2008). As noted, the

Early, Middle, and Late Formative Period terminology and chronology is used for other regions of Ecuador, along with regionally-specific ceramic complexes used in place of

“Valdivia”, “Machalilla”, and “Chorrera”. For the sake of simplicity and to avoid unnecessary confusion, I will refrain from introducing ceramic complex names where possible. The Formative Period cultural chronology is presented first followed by evidence for plant use, for both the coastal lowlands (section 3.3.1) and the highlands and eastern lowlands (section 3.3.2).

3.3.1 The Coastal Lowlands Formative Period and Record of Plant Use

The Early Formative Period in Ecuador is first seen archaeologically in southwestern Ecuador. After the apparent pause following the Vegas occupation of the

Santa Elena Peninsula, there is a striking shift in settlement patterns into the wetter and

more fertile river valleys of southwestern Ecuador (Raymond 2003a:41). Most Valdivia

(Early Formative) sites are located inland about 10 km apart and/or in separate river valleys, although a few are located close to the sea such as the Valdivia type-site and

Real Alto (Figure 3.1) (Raymond 2003a:42). In comparison to Vegas sites, Early

Formative sites are four to ten times larger with a very distinct organization – a vacant central public (ceremonial) area surrounded by a domestic/residential area (Raymond

1993; 2003a:42). The location of most Early Formative sites on river terraces above fertile bottomlands is suggestive of an economy rooted to agricultural activities

(Raymond 2003:42). Although the size, deep middens with abundant ceramics, and

settlement pattern of Early Formative sites argue for permanent year-round settlement

supported by agricultural surpluses, Raymond (2003a:42-44), Pearsall (2003a:233-234)

and Zarrillo et al. (2008:5009-5010) caution that the Early Formative people were still

accessing resources outside of the local village environment, as shown by the diverse

subsistence economy documented by human stable isotope determinations and the variety

of faunal and floral remains recovered. Stable isotope measurements of human skeletal remains from Loma Alta, although inconclusive because of poor preservation, also indicate a broad-based diet with a slight inclusion of maize, with the stable isotope

signature for maize consumption becoming stronger by the terminal Early, Middle and

Late Formative Periods (Raymond 2003a; Tykot and Staller 2002; van der Merwe et al.

1993). This does not mean that the Early Formative people were not fully-sedentary, but

that extra-local resources were obtained and brought to the villages likely though a

combination of task-groups travelling to hunting, sea-fishing, and gathering locations as

well the as the sharing of resources between communities (Raymond 2003a:44; Zeidler

2008:462-463). Similarly, the spacing and location of Early Formative sites on the

landscape also implies some degree of economically and politically independent

communities, but it is more likely that the nucleated settlements were economically, socially, culturally, and ritually inter-dependent (Raymond 2003a:44-49; Zeidler

2008:464). Increasing social complexity from the onset of the Early Formative to the

Middle Formative Period is evidenced by a greater dependence on agriculture, larger and probably socially-grouped house clusters within some settlements, greater evidence for communal social and ritual activities at ceremonial mound sites, increased status inequality, long-distance maritime and terrestrial trade, and geographical expansion to the south, east and north (Raymond 2003a:49-59; Zeidler 2008:464-465).

Based on radiocarbon dates, stratigraphy, and stylistic analysis of ceramics, the

Middle Formative (Machalilla) culture of coastal Ecuador evolved directly from terminal

(Valdivia) Early Formative antecedents (Estrada 1958; Lathrap et al. 1975:33; Staller

2001; Zeidler 2008:466). Continuity from the Early to the Middle Formative is seen in some ceramic forms, decorative techniques and motifs, but new forms, especially the appearance of the stirrup-spout and double spout and bridge bottles, as well as fine-lined

engraved highly burnished blackwares are diagnostic for Machalilla (Lathrap et al.

1975:33-34; Zeidler 2008:466-467). The mixed farming, fishing and hunting/gathering

economy of the Early Formative continues into the Middle Formative, but there is a shift

in site distribution, increased social hierarchy, and an absence of large civic ceremonial

centers and mound building (Zeidler 2008:467). Small homesteads and hamlets are

dispersed in river valleys, such as the Rio Verde Valley, and more and larger coastal sites

are present, these being located on hilltops or higher terraces overlooking the sea. With

respect to spatial distribution, Machalilla sites are discontinuous with the majority of sites

located along the coast between the Chone River (central coast of Ecuador) to the Punta

Arenas Peninsula in the south. Some sites are located in the far north and the Arenillas

River area of the far south (location of La Emerenciana, Figure 3.1); an absence of sites

in the north central coast (the Jama River Valley) is probably a result of the aftermath of

the volcanic eruption that terminated the Early Formative occupation in that area (Zeidler

2008:466-467). Evidence for increased long-distance trade relationships are found in

stylistic similarities between coastal and highland ceramic complexes as well as the

movement of Spondylus and other marine shells to the highlands and obsidian to the coast

(Zeidler 2008:467). Lathrap and colleagues (1975:34) also suggest maritime trade with

northern Peru and western Mexico.

As noted previously, the Middle and Late Formative Periods (Machalilla and

Chorrera) overlap, this being a product of regional differences for the onset/appearance of the Late Formative Period. Geographically, Chorrera represents the most widespread of the Prehispanic cultures of Ecuador, encompassing all of the coastal lowlands with influence extending into the highlands (Zeidler 2008:468). However, whether or not

Chorrera marks a “true ‘cultural horizon’ characterized by stylistic unity in ceramic manufacture... or whether the numerous late Formative regional variants are better understood as independent regional archaeological cultures” is debated (Zeidler

2008:468). Regional ceramic variation certainly exists, both in the highlands and on the coast, but further study and syntheses are required to determine other inter-regional cultural and temporal similarities and differences beyond ceramic traits (Zeidler

2003:469-469). With respect to ceramics, as Lathrap and colleagues (1975:34) state,

“Chorrera represents the artistic climax of the Ecuadorian Formative” with roots extending back to the Early and Middle Formative Periods. Distinctive and elaborate ceramics include phytomorphic and zoomorphic effigy vessels with whistling spout-and- strap handles, as well as (some very) large anthropomorphic, mold-made figurines

(Zeidler 2008:468), and the culmination of 1000 years of experimentation that resulted in the iridescent painting found on Chorrera ceramics (Lathrap et al. 1975:54-55). Studies by Zeidler and associates (Zeidler 1994, 1995; Zeidler and Pearsall 1994; Pearsall 2004;

Zeidler and Isaacson 2003) in the Jama River Valley provide the best syntheses of

Chorrera regional settlement and land use data. After a pause in occupation of the valley during the Middle Formative Period (due to a volcanic eruption in the northern highlands that spewed debris over the north central coast), the Jama Valley population rebounded in the Late Formative Chorrera Period and 33 sites or site components were identified in a regional survey. The mid-valley San Isidro site (Figure 3.1), that served as a civic- ceremonial center during the Early Formative, was reoccupied and the central platform mound was renewed and enlarged in the Late Formative (Zeidler 2008:470). The settlement pattern is similar to the Early Formative Period in the Jama Valley and elsewhere in that sites are primarily located mid-valley, but in contrast some Late

Formative sites in the valley are now located in forested upland locations where long- fallow swiddening (as opposed to floodplain farming alone) supported greater population densities in the Late Formative (Pearsall 2004; Pearsall and Zeidler 1994). This pattern is also distinct from the preceding Middle Formative Period where sites were principally located along the coast, as discussed above. Finally, evidence from the Jama Valley and

elsewhere support the contention that localized stratified chiefdoms were present during the Late Formative Period (Zeidler 2008:470; Zeidler and Isaacson 2003:110).

Table 3.1 shows the plant remains (with the type indicated – phytolith, starch granules, or carbonized macroremains) recovered from coastal archaeological sites for the Preceramic and Formative Periods. With the discovery of a carbonized maize kernel found embedded in a ceramic sherd, modeled adornos of maize cobs on Early Formative ceramic vessels, and apparent punctate impressions of maize kernels on the same pottery from the San Pablo site (Figure 3.1 and Table 3.1), Zevallos and colleagues (1977) were the first to suggest that maize was a staple in the agricultural base of Early Formative sites in southwestern Ecuador. However, as discussed above, it is now thought that the

Early Formative Period communities had a mixed economy based on floodplain cultivation, fishing, hunting, and the gathering of shellfish and wild plants (Pearsall

2003a:233-234; 2003b:42-44; Zarrillo et al. 2008:5009-5010; Zeidler 2003:462).

Loma Alta and Real Alto, located in SW Ecuador (Figure 3.1), are two Valdivia

(Early Formative Period) sites where extensive botanical investigations (flotation for

macrobotanical remains, phytolith and starch granule analyses) have occurred. At Loma

Alta, maize (Zea mays), manioc (Manihot esculenta), arrowroot (Maranta arundinacea),

Canavalia (Canavalia spp., cf. C. maritima), and chili pepper (Capsicum spp.) starch granules have been recovered from ceramic cooking-pot residues directly dated to 3350-

3000 Cal B.C. (Perry et al. 2007; Zarrillo et al. 2008). Maize, manioc and chili pepper starch granules were also recovered from grinding stones and from sediments dated by association to the same time period (Zarrillo et al. 2008). Previous in situ and flotation analysis of Early Formative (Valdivia 1 and 2) sediments from Loma Alta also recovered

arrowroot, maize, and squash (Cucurbita) phytoliths, and jack bean cotyledon fragments

(Pearsall 2003a). Results from extensive botanical investigations at Real Alto, also dating to the Early Formative, are comparable to the results from Loma Alta. At Real Alto, maize was ubiquitous in domestic contexts by 2800-2400 Cal B.C. in association with roots and tubers, such as Canna edulis (canna/achira), Calathea allouia (llerén), arrowroot, and manioc, legumes (Phaseolus spp. and jack bean), and fruits and tree fruits

(e.g., chili peppers, Annona) (Perry et al. 2007; Chandler-Ezell et al. 2006; Pearsall

2003a, 2003b; Pearsall et al. 2004). Other economically important species include cotton

(Gossypium barbadense), palms (Aracaceae) and sedges (Cyperus or Scirpus). Further evidence for (terminal) Early Formative plant use comes from San Isidro and Capa Perro

(located in the Jama River Valley, Figure 3.1) where phytoliths and carbonized plant remains document arrowroot, canna/achira, squash, maize, and tree fruits, as well as bottle gourd (Lagenaria siceraria), palms and sedges probably used for containers, net floats, thatching, and crafts (Pearsall 2003a; 2004).

Middle Formative Period (Machalilla) evidence for plant use comes from the La

Ponga and Rio Perdido sites (Figure 3.1) where phytoliths of canna/achira and maize and carbonized remains of tree fruits and maize were recovered (Pearsall 2003a). The record of plant use for the Late Formative Period (Chorrera) is more extensive than the preceding Middle Formative. Phytoliths and carbonized macroremains show that arrowroot, canna/achira, other roots and tubers (unknown), (Phaseolus) beans, maize, and tree fruits were utilized at El Mocoral, Dos Caminos, Finca Cueva, and the Late

Formative component of San Isidro, all in the Jama River Valley (Figure 3.1) (Pearsall

2003a; 2004). Evidence from the Anllulla shell mound, located in the Guayas Estuary

(Figure 3.1), shows carbonized remains of maize, unknown roots and tubers, and tree

fruits (Pearsall 2003a). Finally, at the Late Formative site of La Emerenciana, located in

El Oro Province (Figure 3.1) on the eastern coast of the Gulf of Guayaquil, maize phytoliths were recovered from ceramic charred residues dating to the Late Formative

Period (Staller and Thompson 2002). These data, as discussed above and presented in

Table 3.1, illustrate that domesticated plants were present and cultivated throughout the

Holocene in southwestern coastal Ecuador.

Table 3.1. Plant Remains Recovered from Preceramic and Formative Period Coastal Lowland Sites in Ecuador

) ) ) a ia a ce r ) nt a ra se le in e .) n u ) nts d ic pp de sc s e un s s a .) .) a) u m r ia rb p p i ot e rp g a ar ) um a ) sp p u h ci ra r n is c b s s llo i S F e nta e l si i us ia an ) r it ub a g du p um elt l al a a ) e o T r a a i C eo v e ys (M ea s d d/P t/ a (L a e C yp f. s a th a a c e u r n o ( M d n r ( s c a n c a a er u i o t ur n e os ( h a ala a m u ac ce p o R R o o a pp G ta ry (P (C C e / y r ta y G it n a ro (C e ( bi er e e ( (Z c A o C / ru w Archaeological n w G a n r b m m n e o ( p e ( sh o o o le n i p to u k u u ré iz ni a g a e F kn nn rr ott an hil ot uc ac eg eg le a a lm . S d u re n Site Cultural Period A A B C C C C H L L L M M Pa cf Se Sq T U Las Vegas 1 Early Vegas Ph Ph Ph Late Vegas Ph Ph Ph Ph, S Anllulla Late Formative C C C Capa Perro Early Formative C Dos Caminos Late Formative C C C El Mocoral Late Formative Ph Ph C, Ph Ph Ph C Finca Cueva Late Formative Ph Ph C C Ph Ph Ph C C La Emerenciana Late Formative Ph La Ponga Middle Formative C 56 Loma Alta Early Formative C Ph, S S Ph C Ph C, Ph, S Ph Ph, S S C C C C, Ph C, Ph Real Alto Early Formative S Ph, S C C Ph C, Ph, S Ph, S C C Río Perdido Middle Formative Ph Ph C San Isidro Early Formative Ph C Ph C Ph Ph Ph Ph Late Formative Ph Ph C, Ph Ph Ph Ph San Pablo Early Formative C C = Carbonized plant remains, Ph = Phytoliths, S = Starch granules Data compiled from Chandler-Ezell et al. 2006; Pearsall 2003a:217-232; Pearsall 2003b:196-197; Pearsall 2004:91-101; Pearsall et al. 2004; Perry et al. 2007; Piperno 2006:116-118,140-148; Piperno and Pearsall 1990; Piperno and Pearsall 1998:183-199; Piperno and Stothert 2003; Staller and Thompson 2002; Stothert et al. 2003; Zevallos et al. 1977; Zarrillo et al. 2008. 1The Precermaic Las Vegas cultural tradition is divided between Early (ca. 10,000 - 8000 cal. BP) and Late (ca. 8000 - 6600 cal. BP) Las Vegas (Stothert 1985; Piperno and Stothert 2003). Results from OGSE-80, the Las Vegas type-site.

3.3.2 The Highland and Eastern Lowland Formative Periods and Record of Plant Use

As compared to the coastal lowlands, current evidence suggests that the highlands

and eastern lowlands were less populated and socio-politically complex during the

Formative Period, but how much this state of knowledge reflects reality or the differences in archaeological investigations and site visibility between the regions is unknown

(Bruhns 2003:126-127; Idrovo Urigüen 1999:115-116; Lippi 2003:533; Raymond

2003b:547; Rostoker 2003:541; Zeidler 2003:487, 2008:460, 471). In addition, and as

Zeidler (2008:471) highlights, unlike the coastal lowlands where geographically expansive archaeological cultures, such as Valdivia, Machalilla, and Chorrera are evident, currently, most cultural entities in the highlands are discussed in terms of single sites. The current situation is rather ironic as such was not the situation in the past. Max

Uhle recognized the southern highlands in particular as an important region and identified the Chaullabamba2 Culture (Civilization) based on ceramics from several sites, including

Cerro Narrío, and regarded it as the first Formative Culture (Oyuela-Caycedo et al. 2010).

Prior to the 1940’s, when more rigorous work was undertaken by several researchers,

including Collier and Murra (1943), Bennett (1946), Ferdon (e.g., 1940), and others, the

cultural framework for Ecuador was based almost exclusively “on the extensive work of

Jijón y Caamaño [1927] in the central highlands unit of Chimborazo” (Bennett 1946:76).

The archaeological past (Archaic and Formative Periods especially) of the central highlands (including Chimborazo), as will be seen below, is currently one of the least known areas of the highlands. Both Collier (1946) and Bennett (1946) synthesized archaeological evidence for Ecuador as whole, and discussed different regions of the

highlands (north, central, southern) as separate from each other and from other regions.

Meggers (1966) places highland entities, such as the Chaullabamba and Tuncahuán

Phases, within the Regional Development Period in her volume on the archaeology of

Ecuador. However, Zeidler’s (2008:471) comment that highland Formative Period archaeology is mostly discussed at the site-level, as mentioned previously, is currently

largely the case, although Cotocollao (see below) is probably an exception. Bruhns

(2003:133, 137-139) sees ceramic similarities uniting sites in the northern highlands and

the central and southern highlands into two separate “cultural spheres,” although this is

debated as discussed below, and Idrovo Urigüen (1999) also provides an overview of the

highland Formative Period. Zeidler (2008) presents a thorough overview of the Formative

Period for all of Ecuador and serves as a valuable reference on the topic, as does the

Archaeology of Formative Ecuador (Raymond and Burger 2003).

Whereas extensive archaeological research has occurred in the western lowlands

of Ecuador to elucidate the timing and nature of domesticated plant use, very little is

known archeologically from the highland and eastern lowland regions. The lack of

knowledge regarding the adoption of agriculture in the highlands and eastern lowlands is

primarily due to an overall lack of past archaeological research in these regions, as

discussed previously, but is also due to natural phenomena that make it difficult to locate

early sites, such as deep burial under river alluvium (Bruhns 2003:128) and volcanic ash

(especially in the central and northern highlands and adjacent regions) (Zeidler

2008:471), as well as the material remains of earlier societies being obscured by the

activities and settlements of later groups (Rostoker 2003:541). The following synthesis of

the archaeology of the highlands and eastern lowlands begins in the highlands, from the

earliest sites onward, and ends with sites from the western and eastern montañas. This is

then followed by evidence for plant use in these regions.

Salvage excavations at the Cotocollao site (Figure 3.1) were conducted prior to a

large urbanization project in the northern suburbs of Quito, and provide some of the best

information on the highland Formative Period. The Cotocollao village occupation spans

over 1000 years (Villalba 1988:40), from about 1800 to 400 Cal B.C. (Lippi 2003:530)

and therefore overlaps the terminal Early Formative, Middle Formative, and Late

Formative Periods on the coast (Zeidler 2008:472). Cotocollao is not an isolated site as

Villalba has identified more than 70 related sites in the Quito Valley and nearby Los

Chillos and Tumbaco valleys to the east – all of these occupations though abruptly ended

in the Late Formative as a result of the Pululahua volcanic eruption (ca. 400 Cal B.C.)

(Zeidler 2008:472; Zeidler and Isaacson 2003:80-86). The earliest domestic architecture

at Cotocollao shows groups of rectangular wattle-and-daub houses (delineated by post-

holes) and Villalba (1988:64-71) suggests that these “household clusters” may relate to

kin groups. Although no ceremonial or public structures were identified at the site, a

centrally-located cemetery area may have been a focus for rituals (Villalba 1988:64).

Ceramics from Cotocollao are made up of a broad variety of classes, including single

stirrup-spout bottles, a “Cotocollao-type” with a spout-and-loop handle, and a Chorrera

(coastal Late Formative) type with spout-and-strap handle (Zeidler 2008:472). Although

it has been suggested that the Cotocollao vessel forms and stylistic attributes share some

similarity to the (Late Formative) La Chimba site (located ca. 55 km north), Figure 3.1,

and that the two sites are united in a northern “sphere” of interaction (Bruhns 2003:135-

136), Athens (1995:25-29) argues that the two assemblages are very different and only share superficial similarities. Similarly, Ronald Lippi (2003:532) contends that there are no morphological or stylistic connections between Cotocollao, La Chimba and other

ceramic complexes of Imbabura and northeastern Pichincha. Rectangular and round

ground-stone bowls with exterior notched ridges and incised geometric decorations are

also characteristic of Cotocollao material culture and are distinct from stone bowl forms

from the coastal lowlands (San Isidro) (Bruhns 2003:134; Zeidler 2008:472); although I

sampled residues from a selection of surface-collected Cotocollao stone bowls these were

found to be contaminated with modern starch (see Chapter 5). Unlike some other

highland sites, Cotocollao does not show clear evidence for inter-regional exchange

(Zeidler 2008:474). Other highland sites with evidence for extended occupations that

began in the Early Formative Period are Chaullabamba and the Catamayo sites, all

located in the southern highlands.

Unlike sites from the northern highlands, where it is debated as to whether there

are similarities in the ceramic assemblages, some sites in the southern highlands (such as

Chaullabamba, Cerro Narrío, Pirincay, and Putishío) and central highlands (Loma Pucara

and El Tingo) do show similarities in their ceramics, especially in thin “eggshell” bowls

and ollas, as well as “heavier plates and bowls of various forms” and burnished

blackwares (Bruhns 2003:139-140). Zeidler (2008:478) feels that the thicker plates and

bowls and burnished blackwares derive from the Red-on-Cream and Burnished

Blackware traditions from Chaullabamba. He has also indicated that, despite that similar

ceramics are shared between these sites, there are also clear differences with respect to craft specialization and architecture (Zeidler 2008:476). Similarities in material culture may be a result of the southern highlands strategic location for trans-Andean trade and interaction, as first suggested by Braun (1982). Chaullabamba is located along the

Tomebamba River in an expanding suburb of Cuenca, capital of Azuay province. The site is very large, with an area of ca. 70 ha, and is situated along the banks of the Tomebamba

River at an elevation of 2300 masl (Grieder et al. 2009:1-4). Based on a four-phase sequence developed for the ceramic assemblage and five radiocarbon dates, four phases

(Periods) for Chaullabamba were defined, as follows: Period I (ca. 2000-1800 BC);

Period II (ca. 1800-1600 BC); Period III (ca. 1600-1400 BC); and Period IV (ca. 1400-

1200 BC) (Grieder et al. 2002:159-163,Figure 6.3; 2009:21,27-50,61-83). These dates situate Chaullabamba in the Early Formative Period, basically contemporaneous with

Valdivia Phases 7-8b (the end of the Valdivia Phase, Early Formative Period) (Zeidler

2003:519; 2008:460,477). Chaullabamba also provides evidence for long-distance trade.

Based on faunal analysis alone, the people of Chaullabamba were able to obtain taxa exotic to the southern highlands – assuredly from the western coastal lowlands and also possibly from the eastern lowlands (Stahl 2005:323). Further, Stahl (2009) uses several

lines of evidence to convincingly argue that nutritionally-select portions of deer were

being specifically imported from the western lowlands. His results not only support the

contention that Chaullabamba was a locus for interregional contact and exchange

(Grieder et al. 2009:209-217), but may also support the idea that high-status individual(s)

resided at Chaullabamba. More detailed information of the archaeological investigations

of Chaullabamba is provided in Chapter 4, section 4.3. For the purposes of this chapter, it is sufficient to note that the occupation of Chaullabamba dates to the terminal Early

Formative to Middle Formative Periods and that no botanical information is available.

The Catamayo sites are located in the Catamayo Valley in Loja Province. The

valley is located in the low intermediate area between the Central Andes to the South, and

the Northern Andes. Guffroy (2004:17) located 30 sites during a survey of the Catamayo

Valley, seven of which he designates as Formative Period sites. The Formative Period

sites are dispersed in three principle groups in the north part of the valley, within 3-5 km

of each other. Based on the stylistic analysis of more than 20,000 pots, mostly from the

La Vega site, as well as stratigraphy and radiocarbon dates, Guffroy (2004:31-32,59-

78,85-103) developed a four-phase sequence for the Catamayo sites: Phase A (2000 to

1400 BC); Phase B (1200 to 900 BC); Phase C (900 to 500 BC); and Phase D (500 to 300

BC). Therefore, Catamayo A dates to the later parts of the Early Formative Period on the

coast, Catamayo B overlaps with the coastal Middle Formative to Late Formative

Periods, and Catamayo C and D are concurrent with the coastal Late Formative Period.

Extra-regional exchange is evidenced by the presence of marine shell (including

Spondylus) by Catamayo B (Guffroy 2004:91-92; Zeidler 2008:479). The Catamayo

pottery shares some stylistic attributes with Formative Period assemblages from other

regions, particularly in Azuay and Cañar highland provinces in the north, the northern

Peruvian coast and highlands, and coastal Ecuador (Guffroy 2004:85-103), and Valdez

(2008:880) notes some similarities with the Mayo-Chinchipe pottery from Santa Ana-La

Florida in the southeastern lowlands and Catamayo Phase A. Despite these similarities,

which may be derived from inter-regional contact, Zeidler (2008:478) sees the Catamayo area as a “separate locus” of Formative Period occupations in the southern highlands.

More detailed information regarding the Catamayo sites, specifically Trapichillo and La

Vega, is provided in Chapter 4, section 4.4., but no botanical information for the

Catamayo sites have been previously reported.

The final highland site that may have an Early Formative Period component is

Cerro Narrío (Figure 3.1), located in the province of Cañar and north of Chaullabamba located in Azuay province. Detailed information on past and more recent investigations for Cerro Narrío, including a discussion of the debate over the possible Early Formative

Period occupation of the site, is provided in Chapter 4, section 4.5. In brief, the evidence for an Early Formative Period occupation is weak if one solely considers the absence of absolute dates derived from stratigraphic contexts, but stronger if based on the ceramic

chronology developed by Collier and Murra (1943) and seriation and comparison to

coastal Early Formative Period ceramics (Braun 1982). The site most certainly has a Late

Formative Period component, as discussed further in Chapter 4, section 4.5., and shares

Late Formative Period eggshell ceramics and similarities in other vessel forms/decoration with other sites, as derived from Chaullabamba. Based on the stratigraphic position of the ceramics and other artifact classes, Collier and Murra (1943:79-80) developed a relative chronology for the site naming two periods: “Early Cerro Narrío” and “Late Cerro

Narrío”. The Cerro Narrío site is minimally located on and around a hill (historic and modern construction prevents determining the definitive extent of the site). The earlier occupations at the site are located on the lower portions of the hill and also showed

considerable amounts of Spondylus shells worked into figurines and beads, with this evidence for lowland coastal contact diminishing over time. The deposits on the top of the hill were shallower, with this part of the site only being inhabited during the later period of occupancy (Collier and Murra 1943:69, 80). As noted below, Collier and Murra

(1943:38, 81) report finding charred maize in Early Narrío deposits.

Highland Middle Formative period sites, aside from some of the Catamayo sites and site components, are rare. Alausí, located in the city that shares its name in

Chimborazo Province, may have a Middle Formative Period (designated Phase A) occupation based on similarity to late Machalilla (Middle Formative coastal) decorative techniques, but Zeidler (2008:475) notes that “classic” Machalilla decorative techniques

(notching, especially on carinated vessels), and vessel forms (stirrup-spout bottles and

carinated bowls) are absent from the Alausí assemblage. Other documented sites in the

central highlands (and Chimborazo Province in particular) include Loma Pucara and El

Tingo and date to the Late Formative Period. The Loma Pucara occupation begins by

about 670 BC (uncalibrated) and although no botanical evidence is available, later levels

of the site included agricultural implements, such as hoes, as well as food-processing

mortars and pestles (Arellano 1994:118-120; Zeidler 2008:475). Arellano (1994:118-120)

sees similarity in the Loma Pucara ceramic assemblage to Cotocollao (to the north,

supported by abundant obsidian sourced to the Quito basin), Chorrera (coastal lowlands),

and the Upano Valley in the eastern lowlands, but the strongest similarities are to

ceramics from the southern highlands – Cerro Narrío and Pirincay – based on the

presence of eggshell pottery. Eggshell pottery is also present at El Tingo, as is evidence for contact with the eastern lowlands and the Quito basin (Zeidler 2008:476).

The Pirincay site (Figure 3.1) is located in the southern highlands in the Paute

River Valley, ca. 25 km northeast of the city of Cuenca (and the Chaullabamba site), and

may have occupations that begin in the Middle Formative. Bruhns (2003:125-133) feels it

unlikely that any highland archaeological sites date earlier than the Late Formative

Period, and situates Pirincay in the Late Formative Period (Bruhns 2003:132), despite

that the earliest dates for the site, ca. 1500 to 1400 Cal B.C. (Bruhns 2003:132; Bruhns et

al. 1990:224), span the Early to Middle Formative Periods (although it is likely that the

major activities at the site occurred during the Late Formative to Early Integration

Periods). As noted, the ceramics from Pirincay and the Late Formative site of Putishío

share similarities to other sites in the central highlands and (northern) southern highlands,

and both sites have good evidence for craft specialization and exchange. At Pirincay there

is evidence for rock crystal bead manufacture specialization as well as exotic items such

as Spondylus and other marine shells from the coast and obsidian from the Quito region

(Bruhns 2003:152; Bruhns et al. 1990:231). Putishío, located in the northern part of Loja

Province, has evidence for gold-smelting and gold objects made with ceramic molds

dating to the Late Formative Period (3420 ± 225 B.P., corrected for δ13C but not

calibrated). Copper-silver-gold alloying (tumbaga) was added beginning in the Regional

Development Period and continued to the Inca Conquest (435 ± 135 B.P.) (Rehren and

Temme 1994:268-280).

Formative Period sites from the western (montaña) slopes the Cordillera

Occidental include Nambillo and Tulipe (the Nueva Era site, Figure 3.1). While these sites are outside of the Inter-Andean corridor, they are included here as they are located in western Pichincha Province, about 35-40 km to the west of Cotocollao (Quito).

Nambillo has a complex stratigraphy of paleosols separated by a series of three volcanic ash and pyroclastic deposits, which have all been dated (Lippi 2003:530-531; Zeidler

2003:474; Zeidler and Isaacson 2003:84-86). The earliest culture-bearing paleosol has ceramics (Early Nambillo phase) that are similar to those of Cotocollao and the Tulipe

Nueva Era phase, as well as, though less so, to coastal terminal Early, Middle and Late

Formative Period ceramics (Lippi 2003:530-531, Table B2; Zeidler 2003:474; Zeidler and Isaacson 2003:84-86). The Nueva Era site is located just above the modern town of

Tulipe at 1500 masl and 35 km WNW of Cotocollao (Quito). Excavations revealed two occupations – the Integration Period Tulipe phase and the Middle to Late Formative

Period Nueva Era phase (ca. 1500 to 400 Cal B.C.), the latter contemporaneous with

Cotocollao (Zeidler and Isaacson 2003:82-83). The two occupations are separated by a two meter-thick pyroclastic deposit from the Pululahua volcanic eruption (ca. 400 Cal

B.C.) (Zeidler 2008:472; Zeidler and Isaacson 2003:80-86), which also ended the Early

Nambillo phase occupation at the Nambillo site, and, as mentioned earlier, the Cotocollao occupations in the Quito and nearby valleys. Despite Nueva Era’s proximity to the contemporaneous sites of Nambillo (18 km south) and Cotocollao (35 km ESE), the early

Nueva Era phase ceramics “appear unrelated to them” (Lippi 2003:532). Zeidler

(2008:475) feels that the Nueva Era phase ceramics are most similar to Late Formative

Period ceramics from the Jama River Valley and that the Nueva Era site may have played a special role as a western montaña node for Late Formative Period exchange between the northern highlands and coast. Lippi (2003:532), based on survey in western Pichincha

Province, also points out that other sites with Cotocollao-related ceramics are present in

the western montaña, but sherds found further west show closer similarities to the coastal

Late Formative ceramics. These differences in ceramic assemblages between sites in

close proximity to one another highlight the “general pattern found throughout the highlands during the Formative Period” (Zeidler 2008:475), as noted previously.

Therefore, although only a few sites in the western montaña contribute to our understanding of the Ecuadorian Formative Period, they are important in understanding regional cultural variability and interaction; botanical information for the Nueva Era phase of the Nueva Era site are presented below and in Table 3.2.

Formative Period occupations in the eastern montaña and eastern lowlands, as for the preceding preceramic period (discussed above), are poorly understood, although this situation is changing. The Pastaza Phase of the Morona-Santiago Province was once thought to be a Formative Period manifestation, but Athens (1990a) has shown that the

Pastaza Phase may actually date to much later. Other sites from the Morona-Santiago

Province eastern lowlands with purported Formative Period occupations are argued to be

later as well, after review of ceramic assemblages (Bruhns 2003:157-159) and

radiocarbon dates (Rostoker 2003:539-541). Similarly, the Upano River Valley mound

complexes are shown to date to the Regional Development Period (500 BC-AD 500)

(Salazar 2008:283). Zeidler (2008:479) conveys that more recent investigations

conducted by Paulina Ledergerber at two sites in Morona-Santiago have Formative

Period dates, but details of these investigations have not been formally reported. Support for Formative Period human activity and maize cultivation, as discussed below, in

Morona-Santiago comes from the Lake Ayauchi (Figure 3.1) pollen core. The confusing and often contradictory archaeological record for the eastern montaña and lowlands, as mentioned earlier, is possibly a result of the material remains of earlier societies being obscured by the activities and settlements of later groups, as there was a “veritable explosion of human settlement that apparently took place” just after the Late Formative

Period (Rostoker 2003:541). This situation is perhaps illustrative of a long record of human activity in the montaña “frontier” regions on both sides of the Andes in facilitating trade between major ecosystem zones – the eastern lowlands, the highlands, and the coast

(Salazar 2008:265; see also Valdez 2008). Despite this state of affairs, excavations in recent years in Zamora-Chinchipe Province now provide the best evidence for Formative

Period occupation in the southeastern slopes/lowlands.

The site of Santa Ana–La Florida (SALF) is located near the modern town of

Palanda on the eastern slopes of the Andes in the province of Mayo-Chinchipe, southeastern Ecuador. Excavations from 2003 until the present reveal a clear stratigraphic sequence showing the presence of at least three occupations of the site. The first corresponds to the Early Formative Period and is named the Mayo-Chinchipe culture

(Valdez 2008:878) and the final occupation was during the Bracomoro Epoch, associated with the Integration Period ca. A.D. 700-800 (Zarrillo and Valdez 2010). At an elevation of ca. 1040 masl the site is located in the upper Amazonian lowlands and the lower

montaña – a transition zone between the Eastern Lowlands and the Highlands. The site has major structural and architectural stone-work, and, although the site has only been partially excavated and mapped, Valdez (2008:878) considers the site to be a local ceremonial site based on several lines of evidence. Valdez (2008:880) notes some similarities with the Mayo-Chinchipe pottery to Catamayo Phase A (from the Catamayo

Valley, discussed above) and with the Phase 3-8 Valdivia Early Formative Period ceramics from coastal Ecuador, all of which are contemporaneous. Further evidence for inter- and extra-regional interaction is documented by the similarity between SALF stone llipta and small mortars to similar stone artifacts documented over a broad area of the southern Loja highlands to the eastern Bagua (province of northeastern Peru) lowlands, with these being interpreted as representative of the use and trade of hallucinogens

(Valdez (2008:884). Importantly, the SALF ceramic assemblage has stirrup-spout bottles that pre-date the appearance of stirrup-spout bottles from the coastal Middle Formative

(Machalilla) Period, the Late Formative Cotocollao stirrup-spout bottles from the northern highlands, and all stirrup-spout bottles from Peru (Valdez 2008:884) including

Cupisnique and Chavín (Pozorski and Pozorski 2008:625-626). Further, much more detailed, information on the Santa Ana-La Florida site is presented in Chapter 4, section

4.2. Prior to the analyses conducted for this dissertation, no paleobotanical evidence was known for SALF. I will now turn to the record for plant use documented for the

Formative Period for the highlands, western montaña and eastern lowlands (no information is available for the eastern montaña).

Pearsall (2003a:218-221) presents the most recent and complete data set for plant food resources during the Formative Period in the highlands. Those data, with further evidence from a few other studies, are presented in Table 3.2 and discussed below. Plant data from three northern Ecuadorian archaeological sites are presented, including

Cotocollao, Nueva Era, and La Chimba, as well as lake cores from Lake San Pablo (not to be confused with the San Pablo site located in southwestern Ecuador, as discussed above) and Lake Yambo (Figure 3.1, Table 3.2). The record for domesticated plant use for the southern Ecuadorian highlands and eastern lowlands comes from two archaeological sites – Cerro Narrío and Pirincay – as well as lake core pollen results from

Laguna Chorreras and Lake Ayauchi.

Investigations at Cotocollao (Villalba 1998) have revealed some of the earliest and most comprehensive data for plant use in the highlands. As discussed above, the occupation of Cotocollao (1800 to 400 Cal BC) extends from the terminal Early

Formative, through the Middle Formative, and ends in the Late Formative, probably as a result of the eruption of the Pululahua volcano (Lippi 2003:529-531, 534; Villalba 63-7;

Zeidler 2008:471-472). The Cotocollao Early Formative deposits contained evidence for the cultivation of two types of beans (common and lima beans, or two types of common bean) (Phaseolus spp.), chocho/tarwi (Lupinus spp.), maize, oca (Oxalis tuberosa), potato

(Solanum tuberosum), quinoa (Chenopodium quinoa), (probably) achira (Canna edulis), and tree fruits, as identified by carbonized remains, phytoliths, and pollen grains. The record is similar for the Middle/Late Formative although evidence for oca, potato, and quinoa is lacking (Pearsall 2003a:217-232; Villalba 1988:330-340). These results and

their associated dates are also discussed further in Chapter 7, sections 7.3.3. Overall, the suite of crops evidenced from Cotocollao ca. 3750 to 2350 BP is very similar to what is grown today in the highlands (see Chapters 2 and 4).

Table 3.2. Plant Remains Recovered from Highland and Eastern Lowland Locations in Ecuador

) ) ) ) um p.) pus p ) ) os ir s a Sc pp. ) os dium quinoa it Fragments s s tuber or P ium barbadenseolus ber /Tuber lis p ay u u sy m t nopo anum perus Phase a Root a ed ( Lupinus e Che (Gos e ( Z (Sol Cy Oxalis ( ann ne ( ze ( to C ton gum pi i a a e c edge nknown Site Cultural Period cf. Cot L Lu Ma O Pot Quinoa S( Tree FruitU Rind/ Cerro Narrío1 Early Formative? C Cotocollao Early Formative C C, P C, P C, Ph, P P P P C C Middle/Late C C C C, Ph C C La Chimba Late Formative C C C C C C C Integration Period C C C C C Laguna Chorreras2 Preceramic P Late Formative P Lake Ayauchi3 Early Formative Ph, P Lake San Pablo Early Formative Ph, P Lake Yambo Late Formative P Pirincay Late Formative Ph C, Ph Nueva Era3 Late Formative C C C C C = Carbonized plant remains, P = Pollen grains, Ph = Phytoliths Data compiled from Athens 1998:161-164; Bruhns 2003:156; Bush et al. 1989:303-305; Colinvaux et al. 1988:95; Pearsall 2003a:217-232; Piperno and Pearsall 1998:243-261; Villalba 1988:330-340. 1Pearsall and Piperno (1990:332) report "cob fragments recovered from the Cerro Narrio site, dated to ca. 2000 B.C.".

2Hansen et al. (2003:102, 106) report "a single grain of Zea mays was found at the 7000 cal. yr BP level" but becomes more frequent after 4000 cal. yr. BP. 3Nueva Era is located on the western slopes of the Cordillera Occidental in the montaña, ca. 1500 masl and 35 km WNW of Quito (Zeidler and Isaacson 2003:82).

Similar results to those from Cotocollao were revealed at La Chimba (Figure 3.1), a Late Formative to Integration Period site located ca. 55 km northwest of Quito (more information on La Chimba is provided above and in section 4.6, Chapter 4). Evidence for cultivated plants at La Chimba is documented by carbonized plant remains of common or lima beans, chocho/tarwi, maize, oca, and potato; the presence of cotton seeds2 (Pearsall 71

2003a:217-232) indicates that La Chimba people were able to obtain products from the western, and possibly eastern, lowlands.

Corroboration of the Early through Late Formative and beyond cultivation of maize in the northern highlands comes from two lakes: Lake Yambo and Lake San Pablo

(Figure 3.1). At Lake Yambo, 2600 masl and ca. 80 km south of Cotocollao, a pollen core that dates to the Late Formative (2540 ± 170, ca. 1108 to 205 Cal B.C.) shows Zea mays pollen throughout the core (Colinvaux et al. 1988:94-95). Lake San Pablo, located

about 50 km northeast of Cotocollao and 20 km east of La Chimba at an elevation of ca.

2600 masl, reveals that maize pollen was continuously present from the bottom of the

core (ca. 4000 BP, Early Formative) onward (Athens 1998:161-164; Pearsall 2003a:231).

Plant subsistence and use is not as well documented for other sites in the

highlands. Nueva Era is located in the montaña of the eastern slopes of the Cordillera

Occidental at 1500 masl (Zeidler and Isaacson 2003:82), and so the site is in the northern

region of the Ecuadorian highlands (although outside of the InterAndean region) and

contemporaneous with the beginnings of the La Chimba occupation and the end of the

Cotocollao occupation. Carbonized macroremains of maize, sedges, tree fruits, and

unknown roots/tubers date to the Late Formative Period at Nueva Era (Pearsall

2003a:231-232). Currently, no central highlands sites can add to our knowledge of past

plant cultivation and use, but, although meager, some information is available for the

southern highlands and eastern lowlands.

The earliest record of domesticated plant use in the southern highlands comes

from Laguna Chorreras (not to be confused with the Late Formative – Chorrera – Period).

Laguna Chorrera (Figure 3.1) is located on the eastern (Inter-Andean) slopes of the

Cordillera Occidental at 3700 masl in a small valley off the Tomebamba River Valley, about 20 km west of the city of Cuenca (Hansen et al. 2003:81) and the Chaullabamba site (detailed in section 4.3, Chapter 4). Hansen and colleagues (2003:106) report one pollen grain of Zea mays and occasional grains of Ambrosia and

Chenopodiaceae/Amaranthaceae in the 7500 to 4000 Cal BP core zone, suggesting the beginnings of anthropogenic disturbance. After 4000 Cal BP Zea mays, Ambrosia and

Chenopodiaceae pollen become more frequent suggesting forest clearance and “the

possible beginnings of regional agriculture”. At 3700 masl Laguna Chorreras is above the modern limit for maize cultivation (ca. 3100 masl), which is interesting given the large pollen grains of Zea mays are wind-dispersed and do not travel over large distances

(Colinvaux et al. 1988:95). The upslope dispersal of maize pollen to Laguna Chorreras

may be explained by the vertical circulation of atmospheric systems in the Andes

(Colinvaux et al. 1988:95), which likely brought Zea pollen from lower elevations to

Laguna Chorreras.

The Pirincay site (Figure 3.1), also in the southern highlands, is located in the

Paute River Valley, ca. 25 km northeast of the city of Cuenca (and the Chaullabamba

site). Bruhns (2003:132), as discussed above, situates Pirincay in the Late Formative

Period. Because neither dates nor context are provided for the paleobotanical (phytoliths and carbonized macroremains) evidence for maize and beans at Pirincay (Bruhns

2003:156), I have assigned these data to the Late Formative Period (Table 3.2).

Questionable evidence for maize cultivation dating to the Early Formative Period comes

from Cerro Narrío (Figure 3.1), located in the Cañar Valley about 35 km north of Cuenca

(and the Chaullabamba site) and 30 km northwest of Pirincay. Piperno and Pearsall

(1990:322) report that maize cob fragments, dated to “ca. 2000 BC”, were recovered from the Cerro Narrío site, but the context of this find is not described nor is a reference provided in order to evaluate this evidence. Lathrap et al. (1975:33) also refer to a carbonized corncob that was found “in a very deep part of the deposit at the Cerro Narrío site”, but, again, no reference or further information is provided. These references to carbonized maize from Cerro Narrío may refer to Collier and Murra’s (1943:38, 81) findings of charred maize in the Early Narrío deposits of Trench 1, but how a date of

2000 BC was determined is unknown. Further information on the Cerro Narrío site is provided in section 4.5, Chapter 4.

Although not located in the highlands, Lake Ayauchi, located in the Ecuadorian

Amazon (Figure 3.1) was cored to document landscape modification in the eastern

Andean foothills (Bush et al. 1989). Maize pollen and phytoliths, together with abundant charcoal particles and disturbance taxa, occurred in the Lake Ayauchi core between 7,010

± 130 and 4,570 ± 70 radiocarbon years BP, yielding a calendar age of “~5,970-6,100 years BP” (Bush et al. 1989:304). The authors suggest that these data indicate that forest clearing (swiddening) and maize cultivation was occurring in the lake watershed at this time.

While the evidence for plant subsistence from highland, montaña, and eastern lowland sites is meager, and most of the data are from the Late Formative period, there are reasonable indications from Cotocollao, Lake San Pablo and Lake Ayauchi that

domesticated plants, including highland crops such as oca, potato, quinoa, as well as legumes and maize, were grown during the Early and Late Formative Periods.

3.4 Chapter Summary

This chapter reviewed the Preceramic and Formative Period archaeology for

Ecuador and the evidence for plant cultivation/use during these broad time periods. On the coast and in the highlands, the Formative Period is preceded by an aceramic Archaic generalized hunting/foraging period. However, the coastal lowlands preceramic (Las

Vegas) period has evidence for plant domestication and cultivation, including domesticated squash, llerén, and bottle gourd by 10,100 to 9,300 Cal BP and maize by

7000 Cal BP. The four-thousand year Formative Period chronology, first developed for coastal Ecuador, is divided into: 1) the Early Formative Period, spanning 4400–1450 Cal

B.C.; 2) the Middle Formative Period from 1450-800 Cal B.C.; and 3) the Late Formative

Period from 1300-300 Cal B.C., and serves as a framework for understanding Formative

Period manifestations for other regions of Ecuador. The Early Formative Period in

Ecuador is first seen archaeologically in southwestern Ecuador, becoming more geographically expansive on the coast through the Early Formative, Middle (albeit with a discontinuous distribution), and Late Formative Periods, culminating in more complex sociopolitical formations that set the stage for the Regional Development Period that follows. A diverse farming-foraging-hunting/fishing economy is evidenced from coastal

Formative Period sites. Abundant evidence for the use of domesticated plants comes from

a number of sites and includes arrowroot, bottle gourd, canna/achira, chili peppers, cotton, squash, beans (common and jack), llerén, manioc, and maize.

Compared to the coastal lowlands of Ecuador, and due primarily to a relative lack of archaeological investigations, the Ecuadorian montaña, highlands, and eastern lowland regions are poorly known. Although previous and current syntheses of the archaeology of the highlands have identified various cultural manifestations and there have been attempts at developing regional chronologies, much is still debated and unknown. Only a few sites, from fairly disparate locations, have been studied intensively (such as La

Chimba, Cotocollao, Pirincay, Chaullabamba, and the Catamayo sites), and not all of these sites have the time-depth necessary to allow for the development of chronological, cultural developmental, and regional sequences that would be equivalent in quality and information as those for coastal Ecuador.

In contrast to the coast, the archaeological record for Archaic Period plant use in the highlands is silent. There is evidence from pollen cores from lakes located in the southern highlands that suggest anthropogenic fires for forest clearance (probably related to hunting activities) dating back to the early Holocene, but no preceramic plant remains have been recovered as yet from the highlands for this early stage of human occupation.

Apparently the period of hunting/foraging in the highlands endures until the Early to Late

Formative Period when a broad range of highland domesticates and maize make a sudden appearance. The Cotocollao site, located in the Quito basin, has evidence for the use of beans, chocho/tarwi, maize, oca, potato, and quinoa, as well as tree fruits and other unknown roots and tubers, beginning in the (terminal) Early Formative. An Early

Formative presence for maize in the northern highlands is supported by the phytolith, pollen and charcoal record from Lake San Pablo. Archaeobotanical remains recovered from La Chimba, located to the northeast of Quito, document the continued use of beans, chocho/tarwi, maize, and potatoes in the Late Formative Period, and add cotton (probably obtained via trade) to the list. Carbonized macroremains of maize, sedges, tree fruits, and unknown roots/tubers date to the Late Formative Period at Nueva Era, located in the western montaña to the west of Cotocollao. A pollen record from Lake Yambo, located ca. 80 km south of Cotocollao, also supports Late Formative Period maize cultivation in the northern highland region. Unfortunately, there is currently no archaeological evidence for Formative Period plant use from the central highlands.

In the southern highlands, the Laguna Chorrera pollen record documents the

earliest appearance of maize (7500 to 4000 Cal BP) and anthropogenic disturbance, but

the record becomes much stronger after 4000 Cal BP. Laguna Chorrera is located just to

the west of Chaullabamba, and even though there are radiocarbon dates suggesting that the site was first occupied by the (terminal) Early Formative Period, botanical remains for

Chaullabamba are not reported. There have been reports for Early Formative maize at

Cerro Narrío, located to north of Chaullabamba, but how the charred remains were dated

is unknown making the data unreliable. Another pollen core, from Lake Ayauchi, located

at eastern the base of the Andes to the NE of Chaullabamba, shows pollen and phytoliths

of maize that date to almost the beginning of the Early Formative Period, ca. 6000 Cal

BP. Consequently, the only southern highland archaeological site with direct evidence for

plant use is Pirincay, located to the northeast of Chaullabamba. While Pirincay may have

a terminal Early to Middle Formative start date, it is difficult to determine the date for the botanical remains (maize and bean phytoliths and carbonized macroremains) because they have not been systematically reported. No botanical information is available for sites in the southern highlands, nor the eastern montaña or lowlands, except the aforementioned information from Lake Ayauchi. In conclusion, the botanical record for the montaña, highlands, and eastern lowland regions of Ecuador is very limited, and the bulk of the archaeological information comes from the northern region. While the sudden appearance of maize and Andean domesticates in the highlands, as discussed in Chapter

1, has been suggested to be a result of stimulus from the more populated and culturally sophisticated coastal lowlands, the sudden appearance of Andean domesticates at highland village sites may also suggest a lengthier, in-situ developmental period of food production in the highlands.

Notes

1. I have often wondered if a tsunami occurred in southwestern Ecuador sometime just prior to 5500 BP, when the Valdivia sub-Period is first archaeologically visible. This thought was only reinforced by the 2004 Indian Ocean and 2011 Japan earthquakes and tsunamis, where the incredible videos captured allowed me to imagine what a similar catastrophe may have looked like in southwestern Ecuador. In such a scenario, it is possible that tsunami waters surging up the river valleys and back out to the ocean may have scoured away all vestiges of Vegas and pre-5500 BP Valdivia settlements in these locations. On the flatter Santa Elena peninsula, however, a tsunami inundation may have

gushed over the landscape and removed or redeposited artifacts from (already ephemeral)

Vegas occupations down to ca. the 6600 BP level, thus accounting for a hiatus. As the

Santa Elena Peninsula environment was never highly conducive to agriculture, especially once the peninsula became drier, it is unlikely that large permanent Valdivia village sites were ever located there, especially once a shift to more intensive cultivation occurred.

There is some scant evidence for Valdivia occupation and activity on the Santa Elena

Peninsula, such as at Punta Concepción, which is located on a cliff above the sea

(Raymond 2003a:41-42) and its location may have aided in the survival of the site if a tsunami had occurred. So it may be that the Valdivia people made more use of the peninsula than the record attests to, such as fishing, hunting, and collecting estuary resources, but that such small special-purpose sites were erased from the landscape by a tsunami. In this regard it is interesting that: 1) “no later preceramic sediments have been recovered from Site 80 [the Las Vegas site]” (Stothert et al. 2003:35); yet 2) a “post-Las

Vegas” secondary burial (dating to ca. 6710 to 6430 Cal BP) is present at the Las Vegas

site and appears in the Late Las Vegas levels (Stothert et al. 2003:26), suggesting that “

...‘post-Vegas’ peoples may sometimes have buried their dead at the site” (Piperno

2006a:147). Therefore it may be that the now more permanently-settled horticultural

post-Vegas people occasionally returned to the honoured location of their ancestors (the

Las Vegas site) for ceremonies and burial rites, as well as other economic activities on

the peninsula in general. When a tsunami occurred sometime between ca. 6430 and 5500

Cal BP, all traces of the “post-Vegas” to Valdivia activities at the Las Vegas site and on

the Santa Elena Peninsula became flotsam, save one burial at the Las Vegas site and

Punta Concepción (as it was located on higher ground). If so, the post-tsunami Valdivia people learned their lesson and from then on located their settlements on higher ground, above the river valley bottom-lands, and ceased (or had already stopped) returning to the

Las Vegas site. All of this is highly speculative (and probably incorrect), yet merits some consideration as vestiges of such an event would surely be recorded in the geological record for the region. Overland flooding, El Niño marine storm surges, and tsunamis have recently been investigated by diatom, pollen, macrophytic and archaeological analysis at

Pachacamac located on the central coast of Peru (Winsborough et al. 2012), and similar studies should be considered for Ecuador.

2. Two independently domesticated species of cotton are present in the Americas.

Gossypium hirsutum was probably first domesticated in the Yucatan peninsula (Brubaker and Wendel 1994), and was dispersed and further differentiated (local varieties or landraces) throughout the circum-Caribbean area, eastern Brazil, and the southwestern

United States (Piperno and Pearsall 1998:149), as confirmed by genetic evidence (e.g.,

Iqbal et al. 2001). Piperno and Pearsall (1998:149-150), based on archaeological and

genetic evidence (Percy and Wendel 1990), suggest that G. barbadense was first

domesticated in the coastal area of NW Peru and SW Ecuador, from there being

dispersed over the Andes. Recent genetic diversity and geographic patterning supports

that G. barbadense has its domestication center in NW Peru/SW Ecuador and was spread

from this core area across the Andes, “south into Bolivia, east into Brazil, and north into

Colombia, Venezuela and the Caribbean”, and, eventually, the Pacific Islands

(Westengen et al. 2005:400). If cotton had spread over the Andes prior to the Late

Formative Period occupation of La Chimba, then it is possible that the cotton seeds from the site may have originated from the eastern lowlands. A fairly definitive way to discern this is to re-examine the seeds to see if they are of the “kidney-seed” type, which are only known from east of the Andes (Turcotte and Percy 1990). Directly-dated cotton (G.

barbadense) fibres from cotton bolls are present in the Ñanchoc Valley (500 masl),

located on the western Andean slopes of northern Peru, by 6278 to 5948 Cal yr. BP, and

Dillehay and colleagues (2007:1891) therefore suggest that cotton must have been

domesticated earlier than ca. 5500 BP and was already being dispersed. It is also possible

that cotton was being grown in some of the warmer and drier Inter-Andean valleys, as it is today, and was present as an exchange item at La Chimba not from outside of the Inter-

Andean corridor, but from within. If so though, to date, La Chimba is the only highland

site where cotton has been identified, but it is also one of the very few highland sites

where flotation for botanical remains has occurred. Cieza de León (translation of

1864:143), in describing the valleys near Quito, during his travels to the region from

1532 to 1550, states, “they have great store of cotton, which they make into cloth for their

dresses, and also use it for paying tribute.” Unfortunately he does not say whether it was

grown locally or imported. Further research is required to flesh out the answers.

Chapter Four: Archaeological Background and Site Environmental Settings

4.1 Introduction

As discussed in Chapter 3, few Early Formative sites have been identified in the

Ecuadorian Highlands and Eastern Lowlands. Samples from six highland sites and one

site from the eastern slopes of the Andes that date to the Formative Period (based on

previous investigations and absolute dating) are included in this investigation. Figure 4.1

shows a map of Ecuador with the location and names of the archaeological sites from

which samples were obtained, and Figure 4.2 is a topographic representation of Ecuador also showing the locations of the archaeological sites. Although elevations are not shown,

Figure 4.2 illustrates the general relief of Ecuador. The following is information, based on previous excavations, on the archaeological sites investigated. This chapter is not intended to be an exhaustive review of the archaeological sites, but rather to provide background information on the location of the sites, previous investigations, general site descriptions, known absolute dates, and the local environment and potential crops that can be grown. The sites are presented in chronological order based on known dates, and

paleoenvironmental information is also provided for each site.

Figure 4.1. Map of Ecuador showing the location of Formative Period Archaeological sites from which samples were obtained. 83

Figure 4.2. Relief map of Ecuador showing the location of archaeological sites investigated. A, Santa Ana-La Florida (SALF). B, La Vega and Trapichillo. C, Chaullabamba. D, Cerro Narrío. E, Tajamar. F, La Chimba. Adapted from Souris (2011a).

4.2 Santa Ana-La Florida

The site of Santa Ana–La Florida (SALF) is located near the modern town of

Palanda on the eastern slopes of the Andes in the province of Zamora-Chinchipe, southeastern Ecuador (Figures 4.1 and 4.2). It was during a regional survey of the Mayo-

Chinchipe river system that the site first came to the attention of Francisco Valdez and colleagues when a local inhabitant mentioned that heavy equipment used during a road construction project near Palanda had cut through an archaeological site (Valdez

2008:878-880; Valdez et al. 2005, Valdez personal communication 2009). Subsequent excavations (2003 until the present) reveal a clear stratigraphic sequence showing the presence of at least three occupations of the site. The first corresponds to the Early

Formative Period and is named the Mayo-Chinchipe culture (Valdez 2008:878). The middle occupation is named Tacana (Regional Development Period) and the final

occupation was during the Bracomoro Epoch, associated with the Integration Period ca.

A.D. 700-800 (Zarrillo and Valdez 2010). The Early Formative occupation is supported by more than 20 radiocarbon assays that have been obtained from different contexts of the site, as shown in Table 4.1. Samples analyzed for this dissertation originate from an undated portion of the X4(17) midden, XIV burial pit (2200 to 1970 Cal BC, 4150 to

3920 Cal BP, Beta-261402), Extola 1 yellow fill layer (3010 to 2301 Cal BC, Beta-

197175 and GX#30044), and the stone-lined XII shaft tomb (2270 to 2260 Cal B.C.,

4220 to 4210 Cal BP, Beta-197176). A complete list of samples analyzed for starch granules is found in Table 5.2, Chapter 5 and the results are shown in Chapter 6, Section

6.2 and Table 6.1.

Figure 4.3. View of Santa Ana-La Florida and the Valladolid River. The large constructions cover a portion of the site, the larger of the two over the artificial terrace and major stone constructions. The modern town of Palanda is seen in the top left of the image. View towards the southwest, May 2010. 85

Table 4.1. Contexts and Radiocarbon Age Determinations from Santa Ana-La Florida.1 Measured Radiocarbon Lab Number Age (BP) 2 Sigma Calibration Provenience 3010 to 2880 BC Beta-197175 4300 ± 40 (4960 to 4830 BP) Occupation Level -150 cm 2857 to 2301 BC GX#30044 4000 ± 71 (4807 to 4449 BP) Artificial Terrace, burnt floor -40 cm 2841 to 2294 BC GX#30043 3990 ± 70 (4791 to 4422 BP) Ceremonial hearth -90 cm 2460 to 2300 BC Beta-172587 3860 ± 40 (4410 to 4250 BP) Ceremonial hearth -90 cm 2470 to 2040 BC Beta-188265 3830 ± 70 (4420 to 3990 BP) Artificial terrace, burnt floor -50 cm 2395 to 2375 BC Beta-188263 3820 ± 40 (4345 to 4325 BP) Artificial terrace, burnt floor -90 cm 2450 to 2140 BC Beta-261400 3820 ± 40 (4440 to 4090 BP) XIII-10 hearth contents. Prov: 45-47 cm 2450 to 2140 BC Beta-261413 3810 ± 40 (4400 to 4090 BP) Base of levels IX, X 6 and 7, Prov: 95-100 cm 2620 to 1750 BC Beta-210219 3790 ±160 (4570 to 3700 BP) West Terrace 22-23 cm 2450 to 2040 BC Beta-214742 3700 ± 60 (4400 to 3990 BP) Grave pit, entry seal -60 cm 2200 to 1970 BC Beta-261402 3710 ± 40 (4150 to 3920 BP) Burial pit fill XIV- 4 (8) Prov: 192 cm 2200 to 1970 BC Beta-261403 3710 ± 40 (4150 to 3920 BP) Midden III - Prof: 50/60 cm 2270 to 2260 BC Beta-197176 3700 ± 40 (4220 to 4210 BP) XII Shaft tomb, offerings context 1220 cm 2140 to 1930 BC Beta-261408 3700 ± 40 (4090 to 3880 BP) IX-8 redish beige layer Prov: 30-45 cm 2190 to 2170 BC Beta-188266 3690 ± 40 (4140 to 4120 BP) Ceremonial hearth -75 cm 2205 to 1735 BC Beta-188264 3660 ± 90 (4155 to 3685 BP) Artificial terrace, burnt floor -50cm 2120 to 1880 BC Beta-261412 3630 ± 40 (4070 to 3830 BP) Midden III- Prov: 85/90 cm 2010 to 1760 BC Beta-261409 3620 ± 40 (3960 to 3710 BP) VI and VII-8 beige layer. Prov: 35-45 cm 2030 to 1780 BC Beta-261410 3600 ± 40 (3980 to 3730 BP) XIV – 6 Occupation level Prov: 50-80 cm 2010 to 1760 BC Beta-261411 3530 ± 40 (3960 to 3710 BP) X-5 Early occupation midden Prov: -60/ 75 cm 1520 to 1200 BC Beta-210218 3140 ± 70 (3460 to 3150 BP) West terrace hearth -20-30 cm 1485 to 800 BC Beta-181459 2930 ± 150 (3435 to 2750 BP) Exposed road profile -145 cm 395 to 200 BC Beta-188267 2280 ± 40 (2345 to 2150 BP) West terrace, Tacana ceramic context -35/55

1. Dates as reported in Valdez (2008:880) and Zarrillo and Valdez (2010). 86

At an elevation of ca. 1040 masl the site is located in the upper Amazonian lowlands and the lower cloud forest (ceja de la montaña) – a transition zone between the

Eastern Lowlands and the Highlands. SALF sits on a fluvial terrace overlooking the

Valladolid River (Figure 4.3). The Valladolid River is a tributary of the upper Chinchipe-

Marañon river basin, which descends from 3000 to 400 masl (Valdez 2008:878), with the

Marañon River being a major tributary of, and eventually flowing into, the Amazon

River.

Figure 4.4. Modern slash-and-burn cultivation near SALF. These plots are across the Valladolid River and slightly downstream from SALF. The left black arrow indicates a cleared and recently planted plot, and the right arrow indicates the smoke from a new area being cleared with fire in preparation of planting. Picture taken May, 2010.

As characteristic for the eastern slopes in general, the rivers that originate in the

highlands initially plunge though deep, steeply-graded valleys before becoming the

meandering rivers characteristic of the lowlands. Although there is little floodplain land 87

in the immediate region of the site (see Figure 4.3), the river here does not have a very steep gradient, the terrace on which the site is located is relatively flat, and many of the surrounding slopes are not so steep as to preclude mixed horticulture using slash-and- burn techniques to clear plots with little if any modification to the natural slope (Figure

4.4).

Figure 4.5. Map of the SALF site showing major stone constructions and topography. Just above the A is the location artificially modified terrace with stone retaining walls. Note the spiral orientation of the retaining walls and how they converge in the centre. B is the centre of the large double-rowed stone structure, with three symmetrically opposed rectangular stone structures in the interior of the double-rowed stone precinct. Image courtesy of Francisco Valdez, with modifications.

The terrace itself, as mentioned, is relatively flat and has been artificially modified with stone retaining walls on its southeastern end to extend its area, resulting in

the site covering just over 1 hectare (ha) overall (Valdez 2008:878) (Figures 4.3, 4.5, and

4.6). While major stone constructions that altered the natural landscape and river bank are

present, the vast majority of the stone constructions appear to be architectural. Indeed, the

extensive stone constructions (Figures 4.5, 4.6 and 4.7) are striking in that they are unlike

Early Formative Period sites from the coast. Although the site has only been partially

excavated and mapped, Valdez (2008:878) considers the site to be a local ceremonial site

based on several lines of evidence.

A large double-rowed stone structure with a diameter of 40 m (center Figure 4.5,

marked by letter B) is surrounded by several other round stone structures. Valdez

(2008:878) considers that the double-rowed stone structure divides the precinct into two

components – the exterior section marked by the circular stone structures, and “an

interior space where three... rectangular structures are symmetrically opposed in tiers.”

Although the function of these structures is unknown, Valdez (2008:878) notes that there

are few signs of middens to suggest domestic refuse or extended residential occupation

from the explored areas of this portion of the site. The artificially modified terrace

(Figure 4.5, marked by letter A) provides further evidence for ceremonial function.

Found within the interior spaces of some of the retention walls were burial offerings of

turquoise beads and polished stone bowls. These simple pits are considered to be possible

elite burials, based on the location and the offerings, but bone preservation is poor and

looters had also disturbed the southern tip of the terrace where these were found (Valdez

2008:878).

Figure 4.6. SALF stone retaining walls. The top image shows some of the stone retaining walls, and the bottom image is a view (to the southwest) from river level showing the steepness of the bank. The corner of the structure visible in the right image that covers a portion of the site is where the stone retaining walls are located. These stone retaining walls stabilized and extended the surface of the terrace towards the Valladolid River. The modern retaining walls along the bank of the river, seen in the bottom image, were put in place to prevent further erosion and destabilization of the river bank.

The stone retaining walls however are part of elaborate spiral structures that converge at an undisturbed central point where a hearth was found and is interpreted to have a ceremonial function as an offering cache was found under the stones (Valdez

2008:878) (Figures 4.5 and 4.7). It is also here that the best mortuary evidence was discovered in the form of a stone-lined XII shaft tomb (2270 to 2260 Cal B.C., 4220 to

4210 Cal BP, Beta-197176) (Figure 4.7).

Figure 4.7 SALF ceremonial hearth with spiral stone constructions. Only a portion of the stone spiral is shown where it converges on the ceremonial hearth, labelled A. The stone slabs in the bottom left of the image (B) cover the entrance to the stone- lined XII shaft tomb.

Within the shaft-tomb the poorly-preserved long bones of at least two individuals

were discovered along with an abundance of funeral offerings. The chamber had an

overall oval shape, with one of the bodies lying on the west end with ornaments and other 91

paraphernalia around it. The remains of another possible individual were located at the

northeastern end with most of the offerings spread out in a semicircle (Figure 4.8). The grave goods included six ceramic vessels (including four stirrup-spout bottles with different body forms), three polished stone bowls, an anthropomorphic four-legged lime container (llipta, bottom left lower image Figure 4.8) containing traces of calcium carbonate (Valdez 2008:882), a polished red stone mortar and pestle in the shape of a bird

(bottom left upper image Figure 4.8), hundreds of turquoise and malachite beads, ornaments and pendants. One of the most interesting finds was the presence of at least three fragments of Strombus sea shells that had been placed next to the individuals

(Valdez 2008:878). Some of these grave goods are shown in Figure 4.8. In addition, three other pit tombs were found in the eastern side of the constructed terrace, suggesting that the whole artificial terrace portion of the site was used as an elite cemetery.

Figure 4.8. Schematic of the stone-lined shaft tomb showing some of the burial offerings. Four of the artifacts pictured were sampled for residues – asymmetrical stone bowl (bottom centre lower image), red incised stone bowl (bottom centre upper image), donut shaped stirrup-spout bottle (bottom right image), and face stirrup-spout bottle (right centre image). Not to scale. Courtesy of Francisco Valdez.

While the radiocarbon dates for the Mayo-Chinchipe occupation date to the Early

Formative Period, comparison of the pottery assemblage with contemporaneous assemblages from the southeastern lowlands of Ecuador or northeastern Peru is not possible as no other equally-early assemblages have been reported. Valdez (2008:880) notes the Mayo-Chinchipe pottery shows some similarity to Catamayo Phase A (from the

Catamayo Valley, southern highlands) and with the Phase 3-8 Valdivia Early Formative

Period ceramics from coastal Ecuador, all of which are contemporaneous. Valdez

(2008:884) also finds similarity between anthropomorphic and zoomorphic style SALF stone lliptas and small mortars to similar stone artifacts documented over a broad area of the southern Loja highlands to the eastern Bagua (province of northeastern Peru) lowlands, interpreting these as representative of the use and trade of hallucinogens.

Importantly, the SALF stirrup-spout bottles pre-date the appearance of stirrup-spout bottles from the coastal Middle Formative (Machalilla) Period, the Late Formative

Cotocollao stirrup-spout bottles from the northern highlands, and all stirrup-spout bottles from Peru (Valdez 2008:884) including Cupisnique and Chavín (Pozorski and Pozorski

2008:625-626). All-in-all, the lack of household middens, the presence of elaborate stone architecture and elite burials, the stylistic similarities found in the ceramics, polished stone motars, and lime boxes to other regions, and the abundance of exotic items

(including Pacific marine Strombus fragments) and turquoise pendants, beads, and ornaments, argues that the Mayo-Chinchipe people were involved in pan-regional interaction spheres and ceremonial activities at SALF (Valdez 2008). In addition to other

reasons, perhaps the site was situated as a midpoint between the Eastern Lowlands and the Sierra to facilitate trade, travel, and communication between regions.

As discussed in Chapter 2, species diversity and endemism in Ecuador is high with 238 plant families and 16,006 species (Jørgensen et al. 2011:192-194). In the western Amazonian lowlands, species diversity is highest at the equatorial latitude and declines towards higher latitudes (Kessler et al. 2011:210). Further, species richness remains constantly high on the eastern Andean slopes, only declining markedly south of

18° (Kessler 2011:210). Plant diversity in the Andean tropics is also affected by elevation, with the highest diversity in humid regions between 1000-1500 masl (Kessler

2011:210). These findings highlight the rich environmental surrounding of SALF, showing that it is positioned in one of the areas of highest plant diversity in the Andean tropical area. Palanda/SALF is today located in the evergreen lower montane forest vegetation zone, with a mean temperature of ca. 22°C year round and temperature

“extremes” that vary between a minimum average low of 13°C in July and maximum average high of 29.5°C in November (Van den Eynden 2004a:131). Average annual precipitation of the Mayo-Chinchipe River Valley is 3354 mm (Buckalew et al. 1988:16).

The antiquity of samples from the site, and the fact that the site is located at an ecological transition zone, make it worthwhile exploring what the environmental conditions were during the time period for the samples analyzed, ca. 4000 Cal BP / 2050 Cal BC.

In a study seeking to understand the climate and vegetation development in the region of Laguna Rabadilla de Vaca, located in the middle of Podocarpus National Park at 3312 masl (the lake is located ca. 42 km north of SALF), Niemann et al. (2009)

analyzed a lake sediment core for multiple proxies (pollen, spores and charcoal analysis, as well as x-ray fluorescence and magnetic susceptibility scanning). The core spanned

10,380 Cal BP to present (95% probability for all dates) (Niemann et al. 2009:309). In their analysis, the middle-Holocene period (ca. 8990-3680 Cal BP) was characterized by a shift of vegetation zones and the treeline to higher elevations, indicating a warmer climate than at present (Niemann et al. 2009:313). The early middle Holocene (ca. 8990-

6380 Cal BP) was likely drier, while the late middle Holocene (ca. 6380-3680 Cal BP) was wetter (Niemann et al. 2009:313-314). The late-Holocene period (ca. 3680 Cal BP to present) when modern conditions developed was cooler and drier relative to the preceding middle-Holocene period (Niemann et al. 2009:314). Further support comes from Laguna Cocha Caranga, located on the western slope of the eastern Cordillera

(Oriental), 8 km south of Loja at ca. 2000 masl (Niemann and Behling 2009:2). Similar to the Laguna Rabadilla de Vaca results noted above, pollen analysis of two Laguna Cocha

Caranga sediment cores and a soil core show that the period between 6900 and 4200 Cal

BP was warmer and wetter than today (Niemann and Behling 2009:7). Importantly, they also attribute the cessation (at ca. 9700 cal BP) of the Late Pleistocene to early Holocene mountain forest expansion at Laguna Cocha Caranga to anthropogenic fires. From ca.

9700 Cal BP to 1300 Cal BP there is a substantial increase in charcoal in the core, and the amount of charcoal present does not fluctuate between wet and dry episodes during this period. Niemann and Behling (2009:13) infer from these data that humans were modifying the local environment since the early Holocene. These paleoenvironmental reconstructions, showing a warmer and wetter period during the earliest occupation of

SALF, are supported by previous research by Niemann and Behling (2008) at El Tiro

Pass (located just east of the modern city of Loja), as well as other studies in Ecuador and

the central and northern Andes (Niemann and Behling 2008:209-210; Niemann et al.

2009:313-314).

Paleoenvironmental data are not entirely consistent as pollen studies from the

Ecuadorian Eastern Lowlands in Yasuni National Park (Weng et al. 2002:86-87) indicate relatively wet conditions ca. 3700-1000 Cal BP and drier conditions from ca. 1000 Cal

BP to the present. However, as noted by Bush et al. (2007:34,47-48) with respect to

paleoenvironmental reconstructions, precipitation and temperature patterns can vary

substantially with latitude in the tropical Andes, especially when additional factors such

as altitude and aspect are considered, so that regions showing synchronous change in one

period can be asynchronous in another. Finally, two other points need to be mentioned

specifically with respect to montane (ceja) cloud forests: 1) although the cloud forest may

have moved up and down the mountain slopes (different altitudes) over time, this niche

has “been a continual feature of the environment since Andean orogeny created uplands

high enough to induce cloud formation” (Bush et al. 2007:47); and 2) utilizing upslope

systems as proxies for montane forest paleoenvironmental reconstructions is valid (Bush

et al. 2007:42). Thus, based on the paleoenvironmental data closest to SALF (Laguna

Rabadilla de Vaca, Laguna Cocha Caranga, and El Tiro pass), it appears that the local

environment at SALF for the period of ca. 6000 to 4000 Cal BP was warmer and wetter

than today. This would have resulted in the local environment being, if not the same as

today, where the site is situated just below the cloud forest (ceja) and within the Upper

Amazon Lowlands, then more strongly associated with that of the lowlands.

Having established that during the time that SALF was occupied the environmental condition were similar to or even warmer and wetter than today, we can now look at the types of domesticated plants that are suited to this zone. Pearsall (1992,

2008) provides information on South American cultivated plants as grouped into complexes based on elevation, as originally defined by Harlan (1975). Lowland Complex roots and tubers include achira (Canna edulis) llerén (Calathea allouia), arrowroot

(Maranta arundinacea), manioc (Manihot esculenta), and sweet potato (Ipomoea batatas) (Pearsall 1992:193-195, 2008:105-106). To this list we can add true yam/mapuey/cush-cush/ñame (Dioscorea spp.) and yautía/cocoyam (Xanthosoma spp.)

(Denevan 2001:311,320; Piperno and Pearsall 1998:59). A host of domesticated, semi- domesticated and cultivated trees that bear edible fruit include cacao (Theobroma spp.), brazil nut/castaña (Bertholletia excelsa), chambira (Astrocaryum spp.), passion fruit/granadilla (Passiflora spp.), mamey (Mammea americana), peach palm/chonta duro

(Bactris gasipaes) (Denevan 2001:309-316), ciruela de fraile/green plum (Bunchosia armeniaca), cherimoya (Annona cherimola), guanabana (Annona muricata), pineapple

(Ananas comosus), papaya (Carica papaya), pacae (Inga feuillei, the pulp found in the seed pods is consumed), avocado (Persea americana), and lúcuma/eggfruit (Pouteria lucuma), (Pearsall 1992:193-195, 2008:105-109). Other plants used for food, fibre, spices/flavourings, dyes, containers, and stimulants include achiote (Bixa orellana), jack bean (Canavalia plagiosperma), peppers/ají (Capsicum spp.), squash (Cucurbita

moschata), cagua/stuffing cucumber (Cyclanthera pedata), cotton (Gossypium barbadense), mate (Ilex paraguariensis), bottle gourd (Lagenaria siceraria), jaboncillo/soapberry (Sapindus saponaria), tobacco (Nicotiana spp.), and maize/corn

(Zea mays) (Pearsall 1992:193-195, 2008:105-109). Many of these species can also be grown at higher elevations.

Important plants of the Mid-elevation Complex (1500-2500/300 masl) include the tuber jícama (Pachyrrhizus ahipa), the pulse peanut (Arachis hypogaea), both of which

are grown at lower elevations as well. Other plants include tarwi/chocho (Lupinus mutabilis, grown at elevations >2000 masl), lima bean (Phaseolus lunatus), common bean (Phaseolus vulgaris), the pseudo-cereal amaranth (Amaranthus caudatus, grown at higher elevations), and, importantly, the stimulant/anti-fatigue/medicinal coca

(Erythroxylum coca) (Pearsall 1992:193-195, 2008:105-109). The above lists are not meant to be exhaustive or represent plant species that originate from or were necessarily grown in the region of SALF, but show the potential for cultivation of these species and that SALF’s location allows for useful plants that occupy a certain niche (e.g., coca) to be grown. Further explication of the regions of origin for the species identified through starch analysis will be discussed in Chapter 7.

4.3 Challuabamba/Chaullabamba1

Chaullabamba is located (Figures 4.1 and 4.2) in the province of Azuay in a

developing suburb of Cuenca. The site is very large, with an area of ca. 70 ha, and is

situated in the grassy meadows along the banks of the Tomebamba River at an elevation

of 2300 masl (Grieder et al. 2009:1-4). Grieder et al. (2009:3-4) point out that the

Tomebamba Valley is linked to the northern highlands, more southern highlands, coastal

Ecuador, and the Amazonian lowlands by a series of natural river valleys and low passes, that, today, connect Cuenca to the north, south, east and west by a network of roads.

These roads may have been preceded by ancient trails. Investigations at the site were begun in the 1920’s by Max Uhle (Grieder et al. 2009:4), and although Donald Collier and John Murra did not excavate at Chaullabamba, they did examine collections from the

Cuenca area and noted the similarity between Chaullabamba and the “Early Cerro

Narrío” pottery from Cerro Narrío (Collier and Murra 1943:83). Similarly, in 1972

Elizabeth Carmichael obtained radiocarbon dates from a small excavation “500 m downstream from the road to Chaullabamba” (Bruhns et al. 1990:224) as she and colleagues had found material similar to that of Chaullabamba in their investigations in the Jubones Valley (Grieder et al. 2009:5). Carmichael’s radiocarbon dates range from

2964 ± 50 to 2784 ± 50 Radiocarbon Years BP (Burleigh and Hewson 1979:347;

Burleigh et al. 1977:149). Dominique Gomis has (1999) studied the Chaullabamba ceramic collections, but by far the most intensive research has been carried out by

Grieder and colleagues (2002, 2009) who conducted excavation at Chaullabamba

between 1995 and 2000.

Grieder and associates excavated a test unit in 1995 along the south bank of the

Tomebamba River to obtain a stratigraphic profile down to sterile sediment. The 1996-

2000 excavations were conducted ca. 100 m upstream from the 1995 test pit (Grieder

2009:8-9), and from these excavations most of what is known about Challuabamba was

100

revealed. Five areas were excavated (designated “Cuts”) with the location of the cuts based on materials eroding from the river bank as well as magnetometer surveys (Figure

4.9).

Figure 4.9. The 1996-2000 Excavations at Chaullabamba. Cut 1 was excavated in 1995 and is not on this map. Map adapted from Grieder et al. (2009:6).

From the ceramics excavated, a four-phase sequence of “wares” was defined

based on surface colour of the pottery. These are Red-on-Cream, Red-on-Black,

Burnished Black/Gray, and Matte Orange (Grieder et al. 2009:27-50). Similarly, based on

the wares, vessel forms, and five radiocarbon dates, four phases (Periods) for

Chaullabamba were defined, as follows: Period I (ca. 2000-1800 BC); Period II (ca.

101

1800-1600 BC); Period III (ca. 1600-1400 BC); Period IV (ca. 1400-1200 BC) (Grieder et al. 2002:159-163,Figure 6.3; 2009:21,27-50,61-83). These dates situate Chaullabamba in the Early Formative Period, basically contemporaneous with Valdivia Phases 7-8b (the end of the Valdivia Phase, Early Formative Period) (Zeidler 2003:519; 2008:460,477). As noted above, Elizabeth Carmichael obtained radiocarbon dates from a small excavation near to Grieder’s Chaullabamba excavations, and these will be presented and discussed in

Chapter 7 in relation to all dates from the site, including those I obtained from ceramic charred residues. Based on the dates and contexts reported by Grieder (2009:21), and

with the invaluable assistance of Antonio Carrillo who worked with Grieder on the

excavations, I chose samples for analysis from Cut III-H, Levels 5 (4 sherds) and 6 (2 sherds), and Cut 1, Level 3 (4 sherds). Chapter 5 Table 5.2 lists samples and Chapter 6, sections 6.3 and 6.9 details results.

Based on Grieder and colleagues (2009:8-20) excavations, Chaullabamba, or at least the explored areas of the site, does not show any evidence for a ceremonial precinct or public architecture. However, James Farmer (chapter 9 in Grieder et al. 2009:141-

157), in his analysis of the Chaullabamba burials and their offerings, argues that several individuals (Burials 1-4 and 6-7), especially Burial 6, were likely of higher status. This is suggested based on the graves’ locations (within a house floor for Burials 1-4) in relation to other burials, the positioning of the bodies (Burial 6 was in a relaxed position as opposed to flexed), and particularly the grave offerings found with Burial 6 that include

single-spout bottles, numerous serving vessels, a bone wand (for coca lime), turquoise

beads, and 25 potters tools among other items. Domestic structures were substantial, as

102

some of the wattle-and-daub (bahareque) walls measured more than 10 cm thick, and the wooden wall posts were set into stone oval and rectangular foundations, with the floors being packed clay (Grieder et al. 2009:18-20). Over time, house floors were raised above ground level by 20-30 cm, the high earthen platforms surrounded and supported by a double or triple layer of cobbles (Grieder et al. 2009:19-20).

Chaullabamba also provides abundant evidence for long-distance trade. Based on faunal analysis alone, the people of Chaullabamba were able to obtain taxa exotic to the southern highlands – assuredly from the western coastal lowlands and also possibly from the eastern lowlands (Stahl 2005:323). Moreover, Stahl (2009) uses several lines of

evidence to convincingly argue that nutritionally-select portions of deer (Cervidae,

dominated by white-tailed deer) – the meatiest upper limb elements – were being

specifically imported from the western lowlands. His results not only support the

contention that Chaullabamba was a locus for interregional contact and exchange

(Grieder et al. 2009:209-217), but may also support the idea that high-status individual(s)

resided at Chaullabamba.

As Grieder at al. (2009:3) remark, Chaullabamba is located in the “cool, rainy

highlands” at 2300 masl. The year-round average temperature range is actually rather

spring-like – 10°C for a low and 21.6°C for a high – and Cuenca receives about 1000

mm/yr of precipitation, which is lower than expected for a region that is exposed to air

masses uplifted from the Amazon lowlands (Hansen et al. 2003:81). This latter point

draws attention to the fact that a multitude of microenvironments are present in the Andes

and these can vary greatly even within small areas, depending on soil type, altitude,

103

aspect and local precipitation (Bush et al. 2007; Gentry 1985; Sandweiss and Richardson

2008). The closest proxy information for reconstructing paleoenvironmental conditions at

Chaullabamba come from two sites in the western Cordillera. Laguna Chorerras (3700 masl) is located in a narrow tributary valley of the Tomebamba River Valley, and

Pallacocha (4060 masl) is located in a cirque at the head of the Tomebamba River Valley

– both are located in Cajas National Park, to the west of Cuenca/Chaullabamba, and current annual precipitation in the park is double that of Cuenca (Hansen et al. 2003:2).

Results of the pollen cores show that between 10,000 and 7500 Cal BP moister conditions with warmer temperatures than today prevailed. Between 7500 and 4000 Cal

BP conditions become dryer until they stabilized around 4000 Cal BP, reflecting modern

conditions. As noted in Chapter 3, Hansen and colleagues (2003:102,106) report one

pollen grain of Zea mays and occasional grains of Ambrosia and

Chenopodiaceae/Amaranthaceae in the 7500-4000 Cal BP zone, suggesting the

beginnings of anthropogenic disturbance. After 4000 Cal BP Zea mays, Ambrosia and

Chenopodiaceae pollen become more frequent suggesting forest clearance and “the

possible beginnings of regional agriculture” (Hansen et al. 2003:106). Thus,

Chaullabamba Periods I through IV (ca. 3950 BP to 3150 BP) not only had climatic

conditions similar to today, but are also correlated to agricultural indicators through

pollen analyses.

At 2300 masl Chaullabamba falls within the upper limit of the Andean Mid-

Elevation and lower ranges of the Andean High-Elevation Agricultural Complexes

(Harlan 1975; Pearsall 1992, 2008). Mid- and high-elevation roots and tubers include

104

potato/papa (Solanum tuberosum), oca (Oxalis tuberosa), mashwa/mashua (Tropaeolum tuberosum), ulluco/papa lisa (Ullucus tuberosus), maca (Lepidium meyenii) (Pearsall

1992:53,55; 2008:107), arracacha/zanahoria blanca (Arracacia xanthorrhiza), and yacón

(Polymnia sonchifolia) (Denevan 2001:307-320). Pseudocereals are important in the higher elevations and include quinoa (Chenopodium quinoa), cañihua (Chenopodium

pallicaule) and amaranth (Amaranthus caudatus) (Pearsall 1992:53; 2003:107). Pulses that can be cultivated include tarwi/chocho (Lupinus mutabilis), lima bean (Phaseolus lunatus), and common bean (Phaseolus vulgaris). Squash (Cucurbita ficifolia), tree tomato/tamarillo (Cyphomandra betacea), cagua/stuffing cucumber (Cyclanthera pedata), banana passion fruit/caruba (Passiflora mollissima), purple maracuyá/galupa

(Passiflora pinnatistipula), and cape gooseberry/uvilla (Physalis peruviana) can also be grown (Denevan 2001:307-320; Pearsall 1992:53,55; 2008:108-109). As noted for Santa

Ana-La Florida, the flora listed are not meant to be exhaustive or represent plant species that originate from or were necessarily grown in the region of Chaullabamba, but show the potential for cultivation of these species. As noted above, floristic compositions in the

Andes can vary greatly even within small areas so that some plant species may be cultivated within niches at elevations outside of their normal range. As such, and as all of the highland sites fall within the growing ranges of the species listed for Chaullabamba, this species list will serve for all of the highland sites. Further explication of the regions of origin for the species identified through starch analysis will be discussed in Chapter 7.

105

4.4 Catamayo Sites – Trapichillo and La Vega

Both Trapichillo (site 1) and La Vega (site 11) are located in the Catamayo Valley

(Figures 4.1, 4.2 and 4.10) and are two of numerous sites surveyed and explored by

Guffroy and colleagues in the Catamayo and other valleys in Loja province, southern highlands. The valley is located in the low intermediate area between the Central Andes to the South, and the Northern Andes. Although previous archaeological investigations were carried out in the northern parts of Loja province, it was not until 1979-1982 that a systematic archaeological survey and excavations were conducted in the Catamayo/La

Toma valley by Guffroy and colleagues (Guffroy 2004:12-13). In the Catamayo Valley alone, Guffroy (2004:17) located 30 sites, seven of which he designates as Formative

Period sites – these are dispersed in three principle areas of the north part of the valley within 3-5 km of each other.

106

Figure 4.10. Map showing the location of one of the Formative Period site clusters in the Catamayo Valley, including Trapichillo (1) and La Vega (11). Map adapted from Guffroy (2004:30).

The Trapichillo site is located ca. 4.5 km northwest of the La Vega site and 1 km north of the modern town of Catamayo (Figures 4.10 and 4.11). The top of the hill, where the site is located, is 1350 masl and the hill is ca. 100 metres high (Guffroy 2004:31). The

La Vega site (Figures 4.10 and 4.12) is located ca. 3 km southwest of the town of

Catamayo at an elevation of 1234 masl (Guffroy 2004:37). Therefore, La Vega and

107

Trapichillo are in close proximity to each other and at about the same elevation – both are situated at a relatively low elevation for the highlands, as mentioned above, the lowest part of the Andes, and only ca. 200 masl higher than the Santa Ana-La Florida site located to the southeast in the upper Amazon. In 2010, after several years trying to locate the Catamayo artifacts, Francisco Valdez facilitated my access and sampling of the ceramics from the Trapichillo and La Vega sites stored at the Museo del Banco Central in

Cuenca. Figure 4.10 shows the location of Trapichillo and La Vega in relation to other

Formative Period sites in the vicinity.

Figure 4.11. View from the north of Trapichillo in the Catamayo Valley. There is a gully separating the modern cemetery from the hill behind it, which is the Trapichillo site. Photo taken May 2010.

108

Based on stylistic analysis of more than 20,000 pots, mostly from the La Vega site, as well as stratigraphy and radiocarbon dates, Guffroy (2004:31-32,59-78,85-103) developed a four-phase sequence for the Catamayo sites – Phase A (2000 to 1400 BC),

Phase B (1200 to 900 BC), Phase C (900 to 500 BC), and Phase D (500 to 300 BC). The

Catamayo pottery shares some stylistic attributes with Formative Period assemblages from other regions, particularly in Azuay and Cañar highland provinces in the north, the northern Peruvian coast and highlands, and coastal Ecuador (Guffroy 2004:85-103). As mentioned previously, Valdez (2008:880) does note some similarities of the Mayo-

Chinchipe pottery from SALF to Catamayo Phase A.

Figure 4.12. A portion of the La Vega site, Catamayo valley. The mound that encompasses the site is of natural origin, ca. 12 m high, 170 m long, and 140 m wide (Guffroy 2004:37-38). View from the south, May 2010.

109

Unfortunately, provenience information for the Trapichillo sherds was limited to site and catalogue number. Without documentation of further provenience, and, as I was particularly interested in the Formative Period occupation, I selected sherds for sampling that were labelled to Phase. These (few) sherds were in separate plastic bags labelled

Phase A, Phase B, Phase C, and Phase D, and some bore interior charred residues suitable for analysis. For the La Vega site, many more sherds with charred interior residues were available. As the Catamayo four-phase sequence was developed based mainly on samples from La Vega, I chose sherds from various contexts of the site for sampling and analysis.

The samples for both Trapichillo and La Vega are presented in Chapter 5, Table 5.2.

Figure 4.13. Canal-irrigated manioc field in the Catamayo Valley. The irrigation field ditches are separated by mud dams that are opened and closed as needed and the bases of the plants are covered with mulch to retain moisture. Photo taken in May, 2010.

110

The broad, flat Catamayo Valley bottom is currently a major agricultural area for

Ecuador, producing sugar cane, tomatoes, and manioc (Figure 4.13) as well as other crops that are usually grown in the lowlands (Van den Eynden 2004b:18). The valley is hot with an annual average temperature of ca. 23°C, which varies little from month-to- month (Van den Eynden 2004b:4-6), and the average annual precipitation is ca. 1000 mm

(Buckalew et al. 1998:16). In contrast to Figures 4.11 and 4.12, that show why the valley environment has been described as a “semi-desert” with dry shrub vegetation (Van den

Eynden 2004b:4-6), Figures 4.13 and 4.14 show the potential for lush vegetation with simple canal irrigation. Guffroy (2004:11) notes, however, that the current vegetation cover may be due to long-standing anthropogenic impact resulting in environmental degradation and erosion. Because of the elevation and climate of the Catamayo sites, many of the plant species listed as potential cultivars for SALF are also potential crops for the Catamayo Valley, with this also supported by current agricultural practices.

111

Figure 4.14. View of the Catamayo Valley from the summit of Trapichillo. The valley is a rich agricultural area and the warm, consistent temperatures allow crops usually limited to the lowlands to be grown here. The view is to the south and the town of Catamayo is in the top left of the image. Photo taken May, 2010.

Due to the proximity of the Catamayo Valley to SALF, the same locations used as proxies for paleoenvironmental reconstruction (see Section 4.2), will be used. Based on the paleoenvironmental data closest to Catamayo (Laguna Rabadilla de Vaca and El Tiro pass), it appears that the local environment for the period of ca. 4000 Cal BP was warmer and wetter than today (Niemann and Behling 2008:7; Niemann and Holger 2008:210;

Niemann et al. 2009:314). These conditions (wetter and warmer) in the Catamayo Valley may have facilitated the Formative Period occupation. Indeed, both the Niemann and

Holger (2008:209-210) and Niemann and colleagues (2009:312-314) studies attribute

112

high charcoal influx peaks at 3300 cal BP and 3680 cal BP, respectively, as being of anthropogenic origin. As noted for SALF, Niemann and Behling (2009:13) attribute the cessation of the Late Pleistocene to early Holocene mountain forest expansion at Laguna

Cocha Caranga from ca. 9700 Cal BP to 1300 Cal BP to anthropogenic fires, indicating that humans were present and modifying the local environment in the vicinity of the

Catamayo valley since the early Holocene.

4.5 Cerro Narrío

Cerro Narrío is located in the province of Cañar in the Cañar Valley (4.1 and 4.2).

The site is located at 3100 masl on a steep hill that is about 100 m high. The hill overlooks the Rio Quillohuac, which joins the Cañar River not far downstream, and is about 700 m west and across the river from the small city of Cañar (Figure 4.15). Urban

expansion in recent years is engulfing the base of the hill where the earliest occupations

of the site are located (Figure 4.16), and, coupled with almost 100 years of looting, the

site is in peril of destruction (Raymond and Delgado 2008-2009).

113

Figure 4.15. View of Cerro Narrío from the southeast. The site covers the hill and area around the hill. The hill has been called “la Tortuga”, because its shape is similar to a turtle (with its head cut off). The northeast projection of the hill is also seen in Figure 4.17. The small city of Cañar is in the foreground. Photo taken July 2008.

In the early 1920’s Max Uhle visited the site, and although he did not conduct

excavations he described the broad variety of finely-made ceramics that he observed

being looted from the site. Although Uhle’s (1922) hypothesis that the Narrío ceramics

were influenced by the Maya has not withstood additional analysis, he was instrumental in shining international attention on the importance of the site. Since Uhle’s first-hand descriptions of the looting, the site has remained under almost constant attack. During our excavations from 2007-2010, we observed active digging and even sub-surface prospection with a metal detector by people from coastal Guayaquil who travelled to the 114

site specifically to treasure-hunt. The extended years of pillaging has left the surface of the site littered with artifacts, mainly ceramic sherds, and looters pits make the surface topography of the hill appear as though it has been used for artillery practice (Figure

4.16).

Figure 4.16. View of the northeast hill projection of Cerro Narrío. Note the cratered surface from looting and the partially-complete modern house foundations at the flanks of the hill, showing the destruction of the site. Numbers show the approximate location of some of the 2007-2009 excavation units, with units 3 and 3A flanking a backhoe trench. Also note the maize field, left center of photo. Picture taken July 2007.

After Uhle’s visit, and despite that the site appeared dismal even in the early

1940’s, Donald Collier and John Murra, after completing surveys throughout the Sierra, decided that Cerro Narrío was the most favourable place to excavate due to the variety of

115

ceramics found on the surface (and in museum and private collections). It must be remembered that their excavations preceded radiocarbon dating and thus in order to develop a regional chronology for the highlands, a site with stratified deposits allowing for ceramic seriation was needed. In addition, even as early as 1941 when Collier and

Murra were surveying the highlands in search of potential sites, Cerro Narrío was already known to be one of the most important sites of the Cuenca-Cañar region based on Uhle’s observations (Collier and Murra 1943:35). They noted that the deposits on the top of the hill were shallower, with later analysis (based on ceramic seriation) showing that this part

of the site was only inhabited during the later period of occupancy. The earlier

occupations at the site, located on the lower portions of the hill, also showed considerable

Spondylus shell worked into figurines and beads, with this evidence for lowland coastal contact diminishing over time (Collier and Murra 1943:69, 80). Based on the stratigraphic position of the ceramics and other artifact classes, Collier and Murra

(1943:79-80) were able to develop a relative chronology for the site naming two periods:

“Early Cerro Narrío” and “Late Cerro Narrío”. They also observed that although the

ceramics showed new forms and decorative motifs in Late Cerro Narrío, the basic

ceramic complex remained consistent indicating cultural continuity (Collier and Murra

1943:79-80).

The absolute age of the Cerro Narrío occupation has been a source of controversy

among archaeologists (Braun 1982; Bruhns 1989, 2003; Bruhns et al. 1990; Lathrap et al.

1975; see also Raymond and Delgado 2008-2009). Braun (1982) determined, based on ceramic seriation and parallels to coastal Valdivia ceramics, that Cerro Narrío had an

116

Early Formative component. Karen Olsen Bruhns (2003:130) discounts this assessment based on comparison of Cerro Narrío ceramics to those of Pirincay, and she contends that neither site date earlier than the Late Formative (1400 BC at the earliest, but also see discussion in Chapter 3, section 3.2.2). She also argues that an Early Formative radiocarbon date from Cerro Narrío (3928 ± 60 BP, corrected for isotopic fractionation but uncalibrated, BM-896) (Burleigh et al. 1977:149) does not have a firm association with diagnostic artifacts (Bruhns 2003:130). The date in question was derived on charcoal, so while the association between this date and the ceramics is rightly questionable the fact remains that wood is dated to the Early Formative Period, 2580 to

2200 Cal BC (95% probability), as calibrated with OxCal v.4.1 (Bronk Ramsey 2009) using IntCal09 (Reimer et al. 2009). As such, the date may signal the beginning of the occupation at Cerro Narrío. The evidence for an Early Formative Period occupation is weak if one solely considers the absence of absolute dates derived from stratigraphic contexts, but stronger if based on the ceramic chronology developed by Collier and

Murra (1943) and seriation and comparison to coastal Early Formative Period ceramics

(Braun 1982).

From 2007 to 2009 I assisted with excavations at Cerro Narrío under the direction of Scott Raymond (2007-2009) from the University of Calgary and Florencio Delgado

(2007-2008) from the Universidad San Francisco de Quito. The goals of the excavation were to:

1. Map and survey the site to determine its areal extent;

117

2. Conduct excavations to recover diagnostic pottery in order to evaluate the chronology

developed for Narrío by Collier and Murra;

3. Obtain samples for radiocarbon assay from stratified deposits to resolve debates on the

age of the site;

4. Attempt to answer questions about the function of the site as a locus for trans-Andean

exchange by examining the artifact assemblages;

4. Identify if any undisturbed areas of the site are present that would be suitable for future

larger-scale excavations;

5. Enlist the support of the INPC, the municipality, and Cañaris to protect the site (see

also Raymond and Delgado 2008-2009).

In addition to assisting with these endeavours, I obtained samples for botanical

analyses from all of the units and levels from our excavations. For consistency with

samples available from the other sites, to enable AMS radiocarbon dating, and in order to

capture the most inclusive botanical information (especially the use of roots and tubers), I

restricted the analysis of samples from Cerro Narrío for this dissertation to ceramic

charred residues from Units 3, 3A, 4 and 7. See Methods, Chapter 5, as well as Chapter

6, sections 6.7 and 6.9, for results of the starch analysis and AMS radiocarbon and δ13C

assays for Cerro Narrio.

Unit 3 was excavated in 2007 and 2008. The approximate locations of the units

are seen in Figure 4.16. Unit 3 (1m x 2m) was located along the south side of a backhoe

trench that had exposed a profile showing undisturbed strata with a clear cultural

118

component extending from ca. 1-1.5 m below the surface (Figure 4.17). The unit was excavated in 10 cm arbitrary levels (Figure 4.17).

Levels 9 to 12 revealed a layer that appeared, by the abundance of charcoal, ash, pottery fragments, flaked stones, grinding stones, complete and fractured animal bones, and carbonized maize seeds, to be a portion of a food processing and cooking area. Unit 3 was not excavated to sterile sediment as the following year (2008) we completed Level

12 and then turned our attention to the other side of the backhoe trench in an effort to expose a larger area.

119

Figure 4.17. Southwest wall of Unit 3 Cerro Narrío. The profile extends to the bottom of Level 11, and shows the ashy cultural layer beginning at Level 9. The “disturbed” area just below the north arrow is where I obtained a bulk sediment sample for Level 12.

120

In 2008 and 2009 Unit 3A (2m x 4m) was excavated to further explore the ashy cultural layer. With the help of the stratigraphy revealed from Unit 3, an attempt was

made to excavate Unit 3A by natural levels (Figure 4.18). Therefore Layers 3-4 of Unit

3A represent that same ashy cultural deposit as Levels 9-12 from Unit 3. Layer 1

represents a mixed deposit of backhoe overburden. Layer 2 also appeared to be a mixed

deposit, the top part likely some overburden from the time that the road(s) adjacent to the

units were created combined, imperceptivity and with increasing depth, with rocks, pottery fragments, sediment, bone fragments and other materials that were redeposited from the hillside above. Unit 3A Layer 3 (Figure 4.18) is a thick ashy deposit that, like

Unit 3, contained abundant sherds, whole and fractured animal bones, flaked stones

(primary and secondary flakes with minimal, if any shaping), cobbles and carbonized

seeds of maize and legume (species unknown). Layer 3 transitioned to a very compacted,

darker-coloured Layer 4 that, based on the compacted floor and some cobble alignments,

may have delineated a cooking and food processing activity area.

121

Figure 4.18. Opening Unit 3A at Cerro Narrío. Note the ashy layer clearly seen in the backhoe cut profile, and quite distinct from the overlying sediment. I am in the bottom right of the photo cleaning off the backfill from plastic covering Unit 3. Courtesy of Scott Raymond, photo taken July 2008.

The southeast ¼ of Unit 3A was excavated to sterile sediment. There were no stratigraphic indications to suggest that Layer 3 and Layer 4 from Unit 3A or Levels 9-12 from Unit 3 represent different occupations or time periods, and all of the ceramics from both units were consistent with Late Cerro Narrío, although these have not been formally analyzed yet. The ashy layer strata of Units 3 and 3A are both clearly separate from the overlying sediment (see Figures 4.17, 4.18, 6.32 and 6.33) and there were no signs during the excavation, or in the wall profiles, to suggest intrusion into the ashy cultural deposit from above. An AMS radiocarbon assay obtained on a charcoal sample from Unit 3A,

Layer 4, Level 4 returned a date of Cal BC 780 to 410 (Cal BP 2740 to 2260, Beta-

122

274050, 2γ), which correlates to the Late Formative Period as defined from coastal

Ecuador (Zeidler 2003:519; 2008:460).

Unit 4 (1m x 2m) was excavated in 2007 and is located on the northeast trending promontory of the hill (Figure 4.16). As expected in this highly looted area of the site, there was some disturbance noted during excavation. However, there were undisturbed portions especially in the south part of the unit. The deposits in this part of the hill were shallow, and sterile sediment was reached in Level 7, ca. 70-80 cm BS. Level 3 in particular had abundant wood charcoal fragments, some indistinct areas of charcoal- stained sediment, ceramic sherds, carbonized maize seeds, and some post holes that extended to the bottom of Level 6, in the intact southern end of the unit (Figure 4.19). As with Units 3 and 3A, and although they have not been formally analyzed yet, all of the decorated ceramic sherds observed were consistent with Collier and Murra’s (1943)

“Late Cerro Narrío”.

123

Figure 4.19. Cerro Narrío Unit 4, bottom of Level 4. The north end of the unit showed signs of disturbance, which can be seen by the mottled appearance of the sediment in the NW quadrant. The area that appears disturbed in the south end (outlined in red) is where I obtained a bulk sediment sample for Level 5. The area directly above the area outlined in red is where the majority of the wood charcoal was found, the charcoal-stained sediment was located, and carbonized maize seeds were found in situ in Level 3. Photo courtesy of Scott Raymond, with modifications added.

124

Unit 7 was also located on the northeast promontory of the hill, but further to the south of and at a higher position up-slope, of Unit 4 (Figure 4.16). In 2009, and armed with a hand-held GPS, we attempted to relocate Collier and Murra’s “Trench 11”. As

Trench 11 had yielded only an “Early Cerro Narrío” context (Collier and Murra 1943:44), we were hoping to position an excavation unit in the vicinity in order to finally unearth some pottery consistent with the earliest ceramic occupation of the site. Relocating

Trench 11 with any precision was rather impossible for two reasons. First, although

Collier and Murra (1943:44 and Map 3) provide a map of the locations they excavated at

Cerro Narrío, the map is without coordinates nor are any coordinates provided in the description for the Trench 11 excavation. Undeterred, I made a map overlaying Collier and Murra’s Map 3 onto a Google Map Image of the site with coordinate grids to aid in relocating Trench 11. The second reason is that, despite being able to home-in on the general area of Collier and Murra’s Trench 11 with the aid of my overlay map, the spot is so cratered with recent and older looting pits, many superimposed on each other, that it was impossible to delineate a locale that we felt might represent Trench 11.

125

Figure 4.20. Opening Unit 7, Cerro Narrío, in 2009. Note the relatively flat, apparently undisturbed appearance of the ground surface. Photo courtesy of Lance Evans.

Disappointed, we chose a flat, seemingly undisturbed, area in the general locale of

Trench 11 and set about excavating a 1m x 2m area – Unit 7 – as shown in Figure 4.20.

Despite the “undisturbed” appearance of the surface where Unit 7 was located, as excavations advanced it was clear that we were proceeding through a palimpsest of overburden from previous looting activities, filled in with sediment transported by

(probably) both overland water and aeolian actions, to give the appearance of a relatively flat ground surface. The unit was eventually excavated to ca. 2.7 m BS. Lance Evans, one of the undergraduate students that assisted with the 2009 excavations, prepared a report

(2010) on the excavation and, based on his level floor plans, I determined that Layer 2,

Levels 14-17 represented undisturbed contexts and selected ceramic charred residue 126

samples from these levels (excluding Level 17, as no suitable ceramic samples were recovered). Chapter 5, Table 2 indicates all of the samples analyzed for starch and submitted for AMS radiocarbon and δ13C assay from Cerro Narrío. In 2010 a wood charcoal sample from Unit 7, Layer 2, Level 17, Feature 3 (an ash/charcoal deposit) was sent for absolute dating (as well as a sample from Unit 3, as mentioned above), returning a date of ca. Cal BC 810 to 760 (Cal BP 2760 to 2710, Beta-274052, 2γ) was obtianed, which correlates to the Late Formative Period as defined from coastal Ecuador (Zeidler

2003:519; 2008:460).

The weather station in the city of Cañar reports an average annual rainfall of 476 mm and the month-to-month rainfall pattern shows a peak in March (ca. 110 mm), with the lowest average in August (ca. 5 mm) (Cisneros et al. 2001:46). The average temperature (11°C) is remarkably consistent and only varies ca. 1°C from month-to- month, but it is cool (ranges between ca. 6.5°C and ca. 16.5°C) and humid, with an annual average humidity range from 70-85%. Studies allowing for paleoenvironmental reconstruction of the Cañar Valley are meager. The best comes from Laguna Pallcacocha located on the eastern slopes of the western Cordillera (2°46’S 79°14’W) at 4060 masl, about 40 km southwest of Cerro Narrío (Rodbell et al. 1999:518). In their study, Rodbell and colleagues (1999) reconstructed the periodicity of El Niño-driven storms based on a lake sediment core that spans ca. 15,000 calendar years. From 15,000 to 7000 cal BP temperatures were warmer and El Niño events were weak, only occurring about every 15 years or more, but beginning ca. 7000 cal BP the periodicity increased reaching a maximum of every 2-8.5 years apart from ca. 5000 cal BP to the present (Rodbell et al.

127

1999:519. Peak in intensity was ca. 2000-1000 cal BP (Moy et al. 2002:164). Similar results for the onset of increased El Niño frequency (mirroring modern conditions) beginning ca. 5000 BP have also been reported by others (e.g., Rollins et al. 1986;

Sandweiss et al. 1996), although a period of decreased frequency along the central and northern coasts of Peru between ca. 5800 and 3200-2800 cal BP has been associated to a period of cultural florescence and monumental temple construction there (Sandweiss et al. 2001). To these results we can reiterate the paleoenvironmental record reviewed for

Chaullabamba, as those proxies are also located in the western Cordillera, ca. 10 km south of Laguna Pallcacocha. As mentioned previously, results of the pollen cores show modern conditions beginning ca. 4000 Cal BP (Hansen et al. 2003:102-106).

Consequently we can concede that by ca. 5000 to 4000 cal BP local conditions at Cerro

Narrío were most likely similar to today. Even if the site was first occupied during the

Early Formative Period (4530 to 4150 cal BP) (Burleigh et al. 1977) the environmental conditions were probably similar to today, or perhaps a little drier and warmer, as suggested by Moy and colleagues (Moy et al. 2002:164). The Late Formative Period occupants of Cerro Narrío though, based on the radiocarbon dates noted previously, were most likely exposed to climatic conditions much like the present. Although Cerro Narrío is located about 700 masl higher than Chaullabamba, the species list for potential cultivars is similar although more strongly associated with the Andean High-Elevation

Agricultural Complex (Harlan 1975; Pearsall 1992, 2008). At ca. 3000 masl, Cerro

Narrío and environs are located at the limit for maize cultivation, but it is certainly grown there today (see Figure 4.17), and other crops include potatoes (most important), barley

128

(Hordeum vulgare), legumes, various cool temperature tolerant vegetables (like carrots),

oca (Oxalis tuberosa), mashua (Tropaeolum tuberosum) and melloco (Ullucus tuberosus)

(e.g., Jokisch 2002:532-534. Plant species identified through starch analysis are shown in

Chapter 6 and will be discussed in Chapter 7.

4.6 La Chimba

Located about 55 km northeast of Quito, near the border of Pichincha and

Imbabura provinces, La Chimba (Figures 4.1 and 4.21) is a large (ca. 12 ha) deeply

stratified site first excavated by Stephen Athens and Alan Osborn in the early 1970’s

(Athens 1978; Athens and Osborn 1974). The site is located at an elevation of 3180 masl

in the middle-upper reaches of a narrow valley cut by the Río de la Chimba (Stahl and

Athens 2001:161). The terrain of the site is generally flat and at least three “mound-like”

features with profuse habitation refuse are present (Stahl and Athens 2001:161). Three

phases revealed during excavation were named Early, Middle and Late La Chimba, with

the Early and Middle phases separated by a volcanic ash layer (Athens 1978). Thomas

Meyers (1976, 1978) argued that ceramics collected from the surface and mixed deposits

in the Lake San Pablo area (close to La Chimba) (separated into Early and Late Espejo

phases) dated to the Early Formative Period based on stylistic similarities to coastal ceramics, and put forth that the La Chimba ceramics were also Early Formative and came from mixed deposits (resulting in an erroneous Late Formative radiocarbon date) (Meyers

1976:355-356). Meyer’s contention was compellingly dismissed by Athens (1978) and

Athens re-excavated the La Chimba site in 1989, in part to resolve chronological issues.

129

A series of 15 radiocarbon dates that range from 700 Cal BC to Cal AD 200 show that

Early La Chimba dates to the Late Formative Period (Athens 1978, 1990; Stahl and

Athens 2001).

Although it has been suggested that the La Chimba vessel forms and stylistic attributes share some similarity to Cotocollao (located in Quito, ca. 55 km to the south)

(Bruhns 2003:135-136), Athens (1995:25-29) argues that the two assemblages are very different and only share superficial similarities. Similarly, Ronald Lippi (2003:532) contends that there are no morphological or stylistic connections between Cotocollao, La

Chimba and other ceramic complexes of Imbabura and northeastern Pichincha.

Based on a thorough analysis of the animal bone assemblage (almost 40,000 specimens) from La Chimba, Peter Stahl convincingly argued that the presence of entire rabbit and deer skeletons (and a lack of similar complete skeletal representation for other taxa) indicates that rabbit and deer were processed on site and may have been used to prepare jerked meat (ch’arki) for trade; while rabbit and deer dominate the faunal assemblage, a host of other high elevation taxa such as mountain tapir, mountain paca, and mountain lion are also present (Stahl and Athens 2001). Direct evidence for extra- local exchange includes marine shell (including Spondylus and Strombus) from the

Pacific coast as well as Chorrera (Late Formative) pottery from the western lowlands, ceramics (Cosanga) from the eastern lowlands, and Mullumica obsidian from the Quito basin to the south (Athens 1995).

In 2010 I obtained numerous ceramic sherds bearing interior charred residues from

Stephen Athens. All of the sherds were rinsed with water to remove adhering sediment at

130

the time of initial processing, dried, and wrapped in individual foil containers (Athens, personal communication 2010). The sherds originate from the 1989 excavation, test pits

(TP) 5 and 7, and are shown in Table 5.2, Chapter 5. Results of the starch analysis are detailed in Chapter 6, section 6.8. Figure 4.21 is a map of the La Chimba site showing the habitation/refuse mounds identified by Athens and the location of the 1972, 1974 and

1989 excavations.

At 3180 masl La Chimba is (just slightly) the highest site investigated for this dissertation, but only ca. 80 m above the maximum elevation for Cerro Narrío and ca. 180 m above the base of the hill that Cerro Narrío encompasses (and, as noted earlier, the La

Chimba site is rather flat). Although maize is grown in the vicinity, it is not one of the major crops – those being more cold-adapted species like barley, chocho/tarwi, quinoa,

melloco, oca, potatoes, legumes, and mashua/mashwa (Athens 1990b; Stahl and Athens

2001:162-163). Average annual precipitation in the páramo (high elevation grassland),

which is less than 2 km from La Chimba as one ascends the valley, varies between 2500

and 3000 mm (Athens and Stahl 2001:163; Cerruto Torrico 2010:3). Cool rains occur

several times a week between October and May, which are the main growing-season months, as June to September are windy and somewhat dry (Athens and Stahl 2001:163).

131

Figure 4.21. Map of the La Chimba site. Charred residues from ceramics analyzed for starch originated from Test Pits 5 and 7, 1989 excavations, as indicated. Map adapted from Stahl and Athens (2001:162).

Although Olmedo (the closest town to La Chimba) has a weather station, I could only obtain yearly temperature averages for Cayambe (ca. 13 km southwest and in the same valley as Olmedo). For Cayambe, the annual average high temperature is ca. 19°C and the low average is 9.7°C (Weatherforcaster 2012). Cayambe though has an elevation

132

of 2770 masl (410 m below La Chimba) and receives only 1207 mm of rain on a yearly average, and thus La Chimba is wetter and probably somewhat cooler than the temperature data for Cayambe.

Proxies for paleoenvironmental reconstruction include one study conducted at the

Guandera Biological Reserve, located about 65 km northeast of La Chimba, and

Yaguarcocha, ca. 22 km north of La Chimba. A small mire at 3400 masl in Guandera

Biological Reserve was cored to reveal a ca. 7000 Cal BP to present pollen sequence, with several shifts in climatic conditions (Bakker et al. 2008). The time period that encompasses the Late Formative Period occupation of La Chimba (2650 Cal BP) is the

5320 to 2160 Cal BP zone of the core, showing conditions comparable to today, followed by a cooler period between 2160 to 910 Cal BP. Therefore, the Guandera core shows a climate transitioning from modern conditions to relatively cooler from the earliest (Late

Formative Period) occupation at La Chimba to the end of the La Chimba occupation

(1450 Cal BP) during the Integration Period. The Yaguarcocha lake core pollen zone spanning 5000 to 1700 RCYBP shows a warmer and drier climate (Colinvaux et al.

1988:88-89). From these data we can conclude that conditions during the Late Formative

Period at La Chimba were, if not somewhat warmer than today, then probably similar.

The high elevation of La Chimba, coupled with the types of cultigens that are grown today, show that potential crops during the Late Formative Period would be similar to those from Cerro Narrío and the Andean High-Elevation Agricultural Complex (Harlan

1975; Pearsall 1992, 2008).

133

4.7 Tajamar

The site of Tajamar (Figure 4.1) was initially identified in 1996 by Villalba

(Domíguez et al. 2003:228; Jara and Santamaria 2009:55-56). The site, or at least components of the site, have been assigned to the Integration Period (AD 500-1500) based on cross-dating of ceramic assemblages (Domínguez et al. 2003:228; Jara and

Santamaria 2009:55-56). The majority of what is known about the site comes from

survey and excavations conducted by Victoria Domínguez as part of a major project

undertaken by the municipality of Quito to identify archaeological sites within its

(expanding) urban area. The information that follows, unless otherwise noted, comes

from the original project report (Domínguez et al. 2003) to the Instituto Metropolitano de

Patrimonio Cultural de Quito (Institute of Cultural Heritage for Metropolitan Quito). The

report was subsequently edited and published in several volumes. The Tajamar

information is found in Atlas Arqueológico: Distrito Metropolitano de Quito, Volume 1

(Archaeological Atlas: Metropolitan District of Quito) (Jara and Santamaria 2009), but,

as noted, I will refer to the original report (Domínguez et al. 2003:228-229).

The areal extent of the (partially destroyed) Tajamar site is approximately

120,000 m2 (12 ha) and it lies at 2520 masl in the northward expanding suburbs of Quito.

The site is bounded to the west by the Pusuqui River, to the east by the gorge of the

Curinquigue stream, to the north by the town of Pomasqui, and to the south by a cement-

block property wall. The site is located on relatively flat terrain with two steep hills to the

northeast. While it was noted that ceramics were visible on the surface of the crest of the hills above as well as on the raised open area by the river Pusuqui, these areas were

134

outside of the survey block. Fifty-four tests were conducted towards the south side in the raised open area along the gorge of the river Pusuqui. Thirty tests turned out to be positive and were excavated to between 80 cm and 1 m BS. Based on the tests, the stratigraphy was distinct and four strata (two cultural) were observed. The uppermost layer (D1) was a 50cm-thick dark brown sandy sediment that contained ceramic fragments. Below this layer was a 10-15cm-thick layer of volcanic ash/debris (D2), followed by a 40-50cm-thick layer (D3) of light brown silty sediment that again contained ceramics. The final layer, D4, is described as “cangahua” (hardpan) and did not contain cultural remains (Domínguez et al. 2003:228). Therefore it appears as though

Tajamar had at least two occupations, stratigraphically separated by a volcanic event.

Domínguez and colleagues (2003:228) state that some of the ceramic forms are very similar to one of the forms characteristics for the Integration Period, but I do not know if the ceramics that are described came from surface collection and/or from the tests. The brief report concludes with a recommendation to consider the site as a priority for investigation.

In 2010 I became aware of the site and arranged with Victoria Domínguez to sample some of the ceramics and stone tools for food residues with the objective of expanding the number of site-samples for my dissertation research. Sherds bearing interior charred residues were few in the Tajamar ceramic assemblage and I was only able to obtained samples from three. I also obtained residue samples from two mano and two metate (grinding/milling stone) fragments. The Tajamar samples are listed in Chapter

5, Table 5.2 and the results of starch analysis and radiocarbon assays in Chapter 6,

135

sections 6.4 and 6.9. At the time of sampling (2010) I was under the impression that

Formative Period radiocarbon dates had been obtained from Tajamar – it now seems as though no radiocarbon assays were performed prior to the dates I obtained on split samples of charred residues from two different sherds (Chapter 5, Table 5.2 Chapter 6,

sections 6.4 and 6.9). Although images of the site provided to me by Victoria Domínguez

seem to show excavations dating to 2009 (see Figure 4.22), I do not have any information

on those investigations.

Figure 4.22. Tajamar site excavation, structure 1. The volcanic ash/debris layer D2 is clearly seen in the profile. Note that the postholes indicated by the red arrow does not intrude into the overlying ash layer, indicating that the structure predates the volcanic eruption. Based on the estimated 1m scale (shown in red) structure 1 is ca. 3 m in diameter. Photo courtesy of Victoria Domínguez, with modifications as noted.

Quito (and by extension Tajamar) receives 1250 mm of rainfall on average annually, with variation between 890 mm and 1366 mm (Sarmiento 1986:31). The

136

pattern is equatorial with the main dry season in July and August (southern hemisphere winter). Relative humidity remains above 70% all year and the mean annual temperature is about 14°C, with a month-to-month variation of only 0.3°C. Absolute temperature extremes are 28°C and 2°C and frost is “totally unknown” (Holdridge et al. 1947:8-9;

Sarmiento 1986:31). Basile (1974:26) though states that the average annual precipitation for Pomasqui (the town immediately to the north of Tajamar) is only 353 mm (17 inches) with an average annual temperature of 16°C (6° Fahrenheit), and, from research in the late 1930’s, Ferdon (1950:75) reports the same average annual temperature as Basile, but

an average annual precipitation of 432 mm. So it seems that Pomasqui (and Tajamar) is

warmer with considerably less precipitation than Quito. This latter point, as also

mentioned earlier, draws attention to the varied microenvironments that may be present

based on localized conditions. While Tajamar receives ca. 600 mm less rainfall on

average per year and is at a slightly higher elevation than Chaullabamba (in Cuenca), the

climates of the two sites are quite similar. Therefore, Tajamar shares the same potential

for domesticated, semi-domesticated and cultivated plants as Chaullabamba, coincident

with the upper limit of the Andean Mid-Elevation and lower ranges of the Andean High-

Elevation Agricultural Complexes (Harlan 1975; Pearsall 1992, 2008). Surprisingly,

seeing as Tajamar is located so near to Ecuador’s capital city, paleoenvironmental

evidence is limited to the studies discussed in section 4.6 above for La Chimba, with the

addition of a pollen core from Lake Yambo, ca. 100 km south of Tajamar. The Lake

Yambo pollen core dates to the Late Formative Period (2540 ± 170 RCYBP, ca. 3058 to

2055 Cal BP) and shows increased precipitation and the strong presence of Zea mays

137

pollen throughout the core (Colinvaux et al. 1988:94-95). The increased precipitation may coincide with the cooling trend noted by Bakker et al. (2008) between 2160 to 910

Cal BP, and, if so, we may assume that their findings showing conditions comparable to today between 5320 to 2160 Cal BP are also applicable for Tajamar.

4.8 Chapter Summary

Background information for the archaeological sites included in this dissertation was presented in this chapter. The locations of the sites, previous investigations, general site descriptions, known absolute dates, and the local environments and potential crops that can be grown were detailed. Paleoenvironmental information was also provided for each site.

Based on the paleoenvironmental data closest to Santa Ana-La Florida the local environment for the period of ca. 6000 to 4000 Cal BP was warmer and wetter than today. Chaullabamba Periods I through IV (ca. 3950 BP to 3150 BP) not only had climatic conditions similar to today, but also agricultural indicators through pollen analyses of lake cores. In the Catamayo Valley, the local environment for the period of ca. 4000 Cal BP was warmer and wetter than today (which is warm, but quite dry). This may have facilitated the Formative Period occupations of Trapichillo, La Vega, and other sites in the valley, as studies of regional lake pollen cores have suggested high charcoal influx peaks at 3300 cal BP and 3680 cal BP as being of anthropogenic origin. At the

Cerro Narrío site, local conditions were most likely similar to today, especially during the

Late Formative Period, as they were for the Tajamar site. Similar results were

138

reconstructed for the La Chimba site for the Late Formative Period occupations as well, although a transition from modern conditions to relatively cooler may have occurred from the Late Formative Period to the Integration Period occupations. Therefore, paleoenvironmental data from most of the archaeological sites (Chaullabamba, Cerro

Narrío, La Chimba, and Tajamar) show climates similar to modern conditions for the time periods relevant to the samples obtained for starch analysis. At Santa Ana-La

Florida and in the Catamayo Valley (for Trapichillo and La Vega) the local environment was probably warmer and wetter than today.

Notes

1. Grieder et al. (2009:4) state that according to the Ecuadorian Instituto Geógraphico

Militar the correct spelling of the site is “Challuabamba, and that spelling is also used by

Stahl (2005) and Zeidler (2008). I have chosen to retain “Chaullabamba” because I

spelled it that way in all of my catalogues, sample lists, lab sample tubes, microscope

slides, starch analysis forms, pictures, etc. and changing all of my information would

have been time consuming and may have led to errors and/or information being deleted

from computer files. As well, the site name is more commonly spelled (and known by)

“Chaullabamba” in many publications, both in English and Spanish (e.g., Bruhns et al.

1990; Gomis 1999; Idrovo Urigüen 1999; Meggers 1966; Raymond and Burger 2003). In

the future, however, I will use the “Challuabamba” spelling, following Grieder et al.

2009.

139

Chapter Five: Archaeobotanical Sampling and Methodology

5.1 Introduction

Due to abundant rainfall and fairly uniform high temperatures, organic materials

rapidly decompose in humid tropical soils (Baillie 1996:256). The highland regions of

Ecuador also present challenges to the successful recovery of preserved organic remains

depending on local conditions of temperature, humidity, and soil pH, as well as other

taphonomic processes. Regardless of the ecological vagaries that may hamper organic

preservation within and between the sites investigated, macrobotanical remains are highly

unlikely to preserve due to the antiquity of the sites and ample moisture regimes across

the region unless they are carbonized, become mineralized, or are preserved in coprolites,

waterlogged (anoxic), permanently frozen, or extremely arid contexts (e.g., see Bryant

1989; Minnis 1989; Pearsall 2000).

In addition to data from macrobotanical analysis1, microbotanical2 remains

(pollen, spores, starch granules, phytoliths, etc.) may also be preserved and recovered to

elucidate past human-plant interactions and paleoenvironmental information, depending

on the context. Therefore, the “ideal” sampling strategy for recovering data that is the

most complete representation of botanical assemblages present at archaeological sites

includes macrobotanical, palynological (pollen and spores), phytoliths, starch granule,

chemical and biochemical, and, most recently, ancient DNA analyses of sediments (both on- and off-site), features (hearths, middens, etc,), human and animal bones, teeth, gut contents, coprolites, and artifact residues (e.g., see Pearsall 1995, 2000; Piperno 1998,

2006a; Wright 2010). The information realized from each of these analyses may result in

140

more robust taxonomic identifications. For example, stable carbon isotope analysis of

3 human bones from an archaeological site may show a signature for C4 plants in the diet

and maize phytoliths, starch granules, macroremains and pollen may be recovered from

different contexts at the same site, not only making the identification of maize more secure, but also elucidating information on processing and storage methods, importance

of maize in the diet, activity areas, artifact use, etc. In addition, employing varied

methods of botanical investigation may identify taxa that are under or not represented by

one type of analysis, but may be captured by another. Plant underground storage organs4

(USOs) are well-known for leaving few traces of their past use because many of the

economically important species do not produce (or are suppressed from producing)

pollen, do not produce or produce few phytoliths in the plant parts utilized, and require

limited processing prior to eating (they may simply be washed before consumption,

whether raw, boiled, or roasted) (e.g., see Pearsall 2000:178; Piperno 2006a:142;

Torrence 2006:30, 37; Wright 2010:40), so much so that they were (even recently)

thought to be beyond archaeological grasp (Smith 1998:143).

In reality, comprehensive archaeobotanical investigations are infrequently

performed due to various reasons – differential preservation or absence of material or

artifacts suitable for analysis, archaeobotanical sampling may not have been considered at

excavation and some types of samples may not be available (e.g., bulk sediment samples

for flotation analysis), samples may have been obtained at excavation but subsequently

“lost” in the intervening years, and many other reasons. These examples are some of the

hindrances encountered for the majority of archaeological sites that either had the

141

potential to be included or were investigated for archaeobotanical study for this dissertation. Specifically, I was unable to relocate artifacts and samples from many previously excavated highland archaeological sites because artifacts had been discarded some years after original excavation or they could not be located. For archaeological sites where I was able to locate and access their curated artifacts and samples, the types of samples and artifacts available varied. Only a few site collections had small sediment samples available for analysis (none had bulk sediment samples suitable for flotation analysis) and, even when present, did not correlate to contexts I targeted for analysis (i.e., were not from Formative Period contexts). Some site collections had few or no stone tools from Formative Period contexts, either because they were never present at the site or they were incorporated in various museum displays and thus I was unable to locate them, let alone sample them for adhering residues. Other sites with Formative Period contexts had few or no ceramic sherds with adhering interior cooking residues, again, either because they were not present at the site (although this is unlikely given the abundance of sherds at ceramic-bearing sites in Ecuador), were not collected during excavation, or were subsequently discarded. In contrast, a comprehensive sampling strategy was planned prior to excavations at Cerro Narrío, resulting in samples suitable for almost all types of archaeobotanical analyses. Coprolites and gut contents from animal or human remains were not found during excavation (and none are likely to be preserved in the humid highland environment), and only a few human remains were recovered from a disturbed context. All of the faunal and human bones are being analyzed by Peter Stahl (Binghampton University) and were not available for

142

archaeobotanical analysis (from dental calculus, for example); isotopic and other analyses may be conducted in the future.

Therefore, the types of samples available for analysis varied greatly between the archaeological sites. For the most part, only ceramic sherds with adhering interior cooking residues and some few stone tools were available (and not all of the sites had stone tools). As such, in order to expedite processing, analysis, and identification, only starch analysis was performed on the artifacts and samples. Phytolith analysis may have identified additional plants and/or complemented some of the identifications made through starch analysis, but many plant identifications may have been missed had phytolith analysis been performed to the exclusion of starch analysis. As noted above,

USOs are not abundant producers of phytoliths yet were, and still are, very important in

South American cuisine. The Andean system of agriculture is based on carbohydrate-rich

root crops (e.g., Hastorf 1993:27-30,110-120). As such, due to time constraints, the

decision was made to prioritize starch analysis over phytolith analysis due to the potential

for capturing the most complete archaeobotanical record. Despite this shortcoming, all

samples processed for starch analysis have been retained for future phytolith analysis, as

have the bulk sediment flotation samples (for macrobotanical remains) and small

sediment samples (for phytoliths, pollen, etc. analysis) from Cerro Narrío.

5.2 Starch Analysis: The Biology of Starch and Archaeological Analysis

Starch granules (often also called “starch grains”) are formed and can be found in

the organs and tissues of most higher plants, including pollen, roots, tubers, rhizomes,

143

bulbs, corms, stems, woody tissue, leaves, fruits, flowers, and seeds (Shannon et al.

2009:24-25). Plants synthesize two types of starch granules: transitory starch and storage starch. Transitory starches are formed in the chloroplasts of leaves and green stems but are utilized by the plant on a daily basis to support ongoing metabolic requirements

(formed during the light period and utilized in the absence of light) – these starch granules have not been shown to be diagnostic to different plant species. In contrast, storage starch granules are synthesized in amyloplasts of non-green storage tissue during one lifecycle phase (for long-term energy storage), to be utilized in another phase – for example, to support annual re-growth or seed germination (Banks and Greenwood

1975:1; Haslam 2004:1716; James et al. 1985:162; Shannon et al. 2009:24-25; Thomas and Atwell 1997:2-3).

Starch granules are primarily composed of amylose and amylopectin; both are complex organic polymers (large molecules of repeating units) where glucose is the repeating “building block”; although a large body of literature exists on starch biosynthesis, this is a continuing area of research and is not fully understood (Gott et al.

2006:42; Shannon et al. 2009:24-25; Thomas and Atwell 1997:2-3). The structural differences between amylopectin (a very large, branched molecule) and amylose (a smaller, mostly linear molecule), and the helical geometry of both, contribute to the semi- crystalline properties of starch (Thomas and Atwell 1997:2-7). When viewed with a microscope fitted with polarizing lenses, (native/unaltered) starch granules produce a visible and distinctive extinction cross with cross-polarization light – a dark “X” (or

“Maltese cross”) against the bright remaining area of the granule – because some areas of

144

the granule are birefringent (they double-refract, or split, the light into two rays) (Banks and Greenwood 1975:247; Calvert 1997:338; Cortella and Pochettino 1994:177; Guilbot and Mercier 1985:241; Loy 1994:89; Moss 1976:5; Pérez et al. 2009:150; Radley

1976:118). The presence of an extinction cross that rotates in tandem when one of the polarizing filters is rotated allows for the positive identification of starch in samples obtained from archaeological contexts (Cortello and Pochettino 1994:177; Loy 1994:89-

90; Moss 1976). Figure 5.1 shows an example of a starch granule as viewed under plane-

polarized light and cross-polarized light, as well as many of the characteristics used in morphological characterization.

Although other plant and non-plant material can produce or mimic an extinction cross, including bordered pits of plant vessel elements, fecal spherulites produced in the guts of some animals, and Haversian systems in bone, an experienced starch analyst is unlikely to confuse these “starch mimickers” with actual starch granules. Hardy and colleagues (2009:253-254) suggest that the only way to identify starch granules from

archaeological contexts with “complete confidence” is to hydrolyze the starch with a

starch-specific enzyme (such as alpha-amylase). Regardless of whether or not this

statement is technically correct, I am hesitant to partially or completely destroy the

archaeological evidence, in this case starch granules, upon which my results and

interpretations are based.

145

Figure 5.1. Example of starch granule, Tropaeolum tuberosum (mashua/mashwa). Photographed with plane polarized (A) and cross-polarized (B) light. “F” arrow points to a facet, “L” to lamella, “E” to one of the extinction cross arms, and “H” points to the hilum. Notice that the hilum is eccentric (not in the middle of the granule), and that the arms of the extinction cross meet at the hilum.

It is well-documented that distinctive features of storage starch granules are genetically controlled and when carefully observed can be used to identify plants to different taxa (Banks and Greenwood 1975:242; Cortella and Pochettino 1994:172;

Guilbot and Mercier 1985:240; Loy 1994:87-91; MacMaster 1964:233; Pérez et al.

2009:150; Preiss 2009:85; Reichert 1913:165; Tester and Karkalas 2001:513; Thomas and Atwell 1997:7). This is true to the extent that Pérez and colleagues (2009:150, emphasis mine) state that “differences in external (starch) granule morphology are generally sufficient to provide unambiguous characterization of the botanical source, via optical microscopy.” In general it can be said that USO starch granules, from species such as Solanum tuberosum, Canna, Maranta, and Dioscorea, are usually large and ellipsoidal in shape with an eccentric hilum, and some, like Manihot, are spherical or 146

truncated hemispherical (Guilbot and Mercier 1985). Cereal starches are very polymorphous; maize starch granules are polyhedral (especially in flint and pop varieties)

with relatively round angles (Guilbot and Mercier 1985). Pulse (legume) starch grains are

reniform in shape, often with a central elongated or starred hilum due to internal cracks

parallel to breaking planes (Guilbot and Mercier 1985). Thus, starch granules recovered

from archaeological contexts can illuminate humans’ past use of different plants. For

instance, previous research has resulted in the recovery of ancient starches from whole

and fragmented macrobotanical remains, desiccated food, sediments, stone tool surfaces,

the interior of ceramic vessels, charred cooking residues from the interior of ceramic

vessels, tooth calculi, coprolites, resins, etc. (e.g., Balme and Beck 2002; Barton et al.

2009; Hardy et al. 2009; Henry and Piperno 2008; Horrocks and Best 2004; Horrocks et

al. 2007; Lentfer et al. 2002; Loy et al. 1992; Mercader 2009; Parr 2002; Perry et al.

2007; Pearsall et al. 2004; Piperno and Holst 1998; Samuel 1996; Ugent et al. 1987;

Yang et al. 2012; Zarrillo et al. 2008). While the reasons for starch preservation from

these diverse contexts are not entirely understood, it is clear that starch granules can

preserve under certain conditions for extended periods of time. Recovery of ancient

starch from archaeological contexts is only the first step, however, with the identification

of the granules to different taxa as often the major consumer of time and effort.

5.3 Modern Comparative Plant Specimens

The identification of botanical remains, regardless of the type, recovered from

archaeological contexts requires an extensive collection of modern comparative plant

147

species growing in the local region as well as important economic (wild or domesticated) plant species. Without a collection of botanical specimens specific to the analysis (e.g., phytoliths, starch granules, macro plant portion) one must rely on published descriptions and photographs and identifications are less secure.

5.3.1 Starch Granule Comparative Specimens

The work of characterizing starch granules distinctive to different plants in a region is an immense undertaking. For this analysis, modern comparative specimens were obtained from various sources – some were collected in Ecuador, some from grocery stores/markets in Calgary, and others from commercial seed vendors. Every effort was made to ensure that the plant species were correctly identified. For example, every sample of “Arrowroot” flour/starch I purchased was found to not have starch granules characteristic of Maranta arundinacea. Thus, the comparative plant species listed in

Table 5.1 were cross-checked with published descriptions and/or photographs to ensure that species designations were accurate and it should be noted that not all samples resulted in starch granule recovery (e.g., some seeds are oily and do not contain starch or abundant starch).

148

Table 5.1. Modern Plant Species Tested for Starch Granules.

Scientific Name Family Common Name Sample Type Location of Collection

Amaranthus Market, Riobamba, caudatus Amaranthaceae amaranth Organic flour Ecuador cherimoya, custard Cut and dried Market, Riobamba, Annona cherimola Annonaceae apple exocarp, seeds Ecuador sweetsop, sugar Annona squamosa Annonaceae apple, Dried seeds TropiLab Inc.

Arachis hypogaea Fabaceae peanuts Raw seeds Columbia, MO1 zanahoria blanca, Arracacia arracacha, apio Cut and dried Market, Riobamba, xanthorrhiza Apiaceae criollo, virraca root Ecuador

Dried seeds - previously Bactris gasipaes Arecaceae peach palm canned Calgary Market

prickly palm, Bactris major Arecaceae biscoyol, hiuscoyol Dried seeds TropiLab Inc.

Bertholletia excelsa Lecythidaceae brazil nut Dried fresh seed Calgary Market Bixa orellana Bixaceae achiote, annatto Dried seeds Ecuador Market

Dried exocarps, SALF site garden, Bixa orellana Bixaceae achiote, annatto seeds Palanda Ecuador

Bixa orellana Bixaceae achiote, annatto Dried seeds Calgary Market

149

Scie ntific Name Family Common Name Sample Type Location of Colle ction llerén, sweet corn Calathea allouia Marantaceae root Root starch Venezuela Canavalia ensiformis Fabaceae jack bean Dried seeds El Salvador Canavalia ensiformis Fabaceae jack bean Dried seeds St Kitts & Nevis Canavalia ensiformis Fabaceae jack bean Dried seeds Peru Canavalia ensiformis Fabaceae jack bean Dried seeds Columbia Canavalia rosea (C. maritima) Fabaceae wild jack bean Dried seeds TropiLab Inc. Canavalia spp. Fabaceae jack bean Seed starch Peru1

Canna edulis Cannaceae achira root starch Venezuela Capsicum Dried pod and frutescens Solanaceae red chili, ají seeds TropiLab Inc. Capsicum spp. cf. Dried pod and annuum Solanaceae chili pepper, ají seeds Calgary Market Capsicum spp. cf. Dried pod and annuum Solanaceae chili pepper, ají seeds Calgary Market Capsicum spp. cf. Dried pod and annuum Solanaceae chili pepper, ají seeds Calgary Market

Cenchrus brownii Poaceae bur-grass Seed starch Columbia, MO1 Chenopodium quinoa Amaranthaceae quinoa Dried seeds Calgary Market

Chenopodium Market, Riobamba, quinoa Amaranthaceae quinoa Dried seeds Ecuador Chrysophyllum cainito Sapotaceae caimito, star apple Dried seeds Trade Winds Fruit calabash, guacal, Crescentia cujete Bignoniaceae morro Dried seeds Trade Winds Fruit

Cucurbita butternut squash, moschata Cucurbitaceae chicamita, lacayote Dried seeds Mexican origin 150

Scientific Name Family Common Name Sample Type Location of Collection

ñame, yampi, cush cush yam, sacha Dioscorea spp. Dioscoreacea papa Root starch Florida1 ñame, yampi, cush cush yam, sacha Cut and dried SALF site garden, Dioscorea spp. Dioscoreacea papa root Palanda Ecuador ñame, yampi, cush cush yam, sacha Cut and dried Dioscorea trifida Dioscoreacea papa root TropiLab Inc.

pacay, ice cream Dried pod and Market, Riobamba, Inga feuillei Fabaceae bean seeds Ecuador

boniato, sweet Ipomoea batatas Convolvulaceae potato Root starch Florida1

Cut and dried SALF site garden, Ipomoea batatas Convolvulaceae sweet potato roots Palanda Ecuador maca, maca-maca, Lepidium meyenii Brassicaceae ayak willku Root flour Calgary Market Lagenaria bottle gourd, siceraria Cucurbitaceae acocote Dried seeds Calgary Market Market, Riobamba, Lupinus mutabilis Fabaceae chocho, tarwi Dried beans Ecuador manioc, yuca, Manihot esculenta Euphorbiaceae cassava Dried root Surinam manioc, yuca, SALF site garden, Manihot esculenta Euphorbiaceae cassava Cut and dried Palanda Ecuador

manioc, yuca, Manihot esculenta Euphorbiaceae cassava Cuta and dried Ecuador market Maranta commercial arundinacea Marantaceae arrowroot Root starch Columbia, MO1 Maranta arrowroot , Cut and dried arundinacea Marantaceae maranta, sagu rhizomes, starch SALF garden

Cut and dried Market, Riobamba, Oxalis tuberosa Oxalidaceae oca, ocacita root Ecuador 151

Scientific Name Family Common Name Sample Type Location of Collection

Cut and dried Market, Riobamba, Oxalis tuberosa Oxalidaceae oca, ocho root Ecuador Cut and dried Market, Riobamba, Oxalis tuberosa Oxalidaceae oca, ocho root Ecuador Pachyrrhizus erosus Fabaceae jícama Root starch Columbia, MO1

Phaseolus lunatus Fabaceae lima bean Dried seeds Calgary market Phaseolus vulgaris Fabaceae pinto bean Dried seeds Calgary market Phaseolus vulgaris Fabaceae kidney bean Dried seeds Peru1 Phaseolus vulgaris Fabaceae kidney bean Dried seeds Calgary market Sterculiaceae Theobroma cacao (Malvaceae) cacao, chocolate Dried seeds TropiLab Inc.

Tropaeolum Cut and dried Market, Riobamba, tuberosum Tropaeolaceae mashua, mashwa root Ecuador

ulluco, melloco, Cut and dried Market, Riobamba, Ullucus tuberosus Basellaceae papa lisa root Ecuador ulluco, melloco, Cut and dried Market, Riobamba, Ullucus tuberosus Basellaceae papa lisa root Ecuador

ulluco, melloco, Cut and dried Market, Riobamba, Ullucus tuberosus Basellaceae papa lisa root Ecuador

Xanthosma spp. Araceae malanga Root starch St Louis Market1 Xanthosoma Cut and dried atrovirens Araceae yautia amarilla root TropiLab Inc. Xanthosoma tahitian spinach, Cut and dried brasiliense Araceae calalu root TropiLab Inc.

Xanthosoma mangarito, golden Cut and dried mafaffa Araceae elepahnt's ear root TropiLab Inc.

Xanthosoma yautia, malanga, Cut and dried sagittifolium Araceae arrowleaf root TropiLab Inc. 152

Scientific Name Family Common Name Sample Type Location of Collection La Chimba region, Zea mays Poaceae corn, maize Red Flour Ecuador1 Zea mays Poaceae corn, maize Bicolor Flint Highland Ecuador

Red & Yellow Zea mays Poaceae cuzco gigante Striped Flour Huancayo1 Yellow Pointed Zea mays Poaceae conguil Popcorn Otavalo1 North American dent corn, "Indian Flint with Flour Zea mays Poaceae Corn" cap Calgary Market

Zea mays Poaceae nal- tel Orange Flint Rollins Springs Bottom1

Zea mays Poaceae reventador Popcorn Rollins Springs Bottom1

Zea mays Poaceae pod corn Popcorn Seed Savers Flint with Flour Catamayo, Ecuador Zea mays Poaceae maiz morado cap (from Piura, Peru)

Botanical names throughout this dissertation follow the Catalogue of the Vascular Plants of Ecuador (Jørgensen and León-Yánez 1999). 1 These samples were generously provided by Deborah Pearsall of the University of Missouri from her modern starch reference collection. If the origin of the sample is known, it is indicated; otherwise “Columbia, MO” is listed as the origin.

5.4 Modern Comparative Starch Granule Sample Processing

All specimens were prepared in the same manner in a method similar to that

described by Ugent et al. (1984). Exteriors of the USOs were washed with water and

allowed to dry. The USOs were then sliced into thin sections with a clean knife and

allowed to dry. A small portion of each cortex was scraped off with a clean probe and

153

placed onto separate labelled and clean microscope slides. Figure 5.2 shows samples collected in Ecuador being prepared for drying and transport back to Calgary.

Figure 5.2. Arracacha (Arracacia xanthorriza, zanahoria blanca) roots. Washed (left image), cut (center image), and sun-drying with other samples (right image).

Dried seeds were gently cracked open with a clean hemostat and a small amount of the endosperm (for monocots) or cotyledon (for dicots) starch was transferred to clean

microscope slides. A few drops of 50:50 glycerine:deionized5 water was added and

stirred to mix the starch with the glycerine:water. A clean cover slip was applied and sealed with nail polish – the nail polish was applied with separate paper strips to avoid cross-contamination (a paper strip was discarded after each time it was dipped into the nail polish and used). All of the probes, knives and any other instruments used were washed and then boiled in 5% acetic acid (vinegar) in a pressure cooker to gelatinize/destroy starch granules (see section 5.6 below for further comments on the potential for modern starch contamination) and were not used across samples.

Microscope slides were viewed under 400X magnification with plane and cross-polarized

light to identify and confirm diagnostic forms and attributes of the starch grains as

154

compared to the written descriptions and photomicrographs available of the different species. Comparative collection plant preparation and slide preparation was performed in a separate room to that of the archaeological samples to avoid the possibility of contaminating the archaeological samples with modern starches.

5.5 Methods for Recovering Starch Granules from Archaeological Samples

In general, two different types of residue samples were obtained from artifacts – from stone tools and from ceramic charred residues. Both “standard” and unique sampling methods were used depending on the nature of the artifacts and residues. The following describes these methods.

5.5.1 Stone Tools

Laboratory procedures used to remove residues from the stone tools and to recover starch granules from the residues employed previously published methods, with modifications as required due to the nature of the samples (Chandler-Ezell and Pearsall

2003; Pearsall et al. 2004; Zarrillo and Kooyman 2006). For most of the stone tools two

types of samples were obtained – a washed sample and one or two ultrasonic wash

samples. These sample types are:

1. Whether a stone tool had been washed previously or not, it was washed prior to

analysis to remove sediment and incidental starch from the surface of the tool;

2. Ultrasonic washing with distilled water to remove sediment and residue contained

in cracks and crevices of the stone tool representing residues related to tool use.

155

For the wash samples, while holding the tool over a clean container, distilled water was gently sprayed onto the surface of the tool and, if not previously washed, then lightly brushed with new, clean toothbrushes. After the wash samples were obtained, the tools were placed into new, clean containers and then placed onto a rack in an ultrasonic bath

(without heat) and sonicated for 30 minutes. Where possible, the above sample types were obtained from both the “utilized” and “unutilized” surfaces of the stone tools for comparison. Because of the small size of the ultrasonic bath that I had in Ecuador, it was usually impossible to obtain separate samples from “utilized” and “unutilized” surfaces of a stone tool or stone bowl and in such cases “whole tool” samples were obtained.

Figure 5.3 shows a metate fragment from Tájamar being sonicated.

Figure 5.3. Example of sonication procedure to remove residues from a stone tool surface. Metate (milling stone) fragment (STR-80) from Tájamar.

156

The wash and sonicated sample types were then rinsed into new, cleaned and labelled centrifuge tubes. These wet samples were centrifuged at 3000 revolutions per minute (rpm) for 5 minutes and the supernatant decanted, leaving a “residue” pellet. The centrifuge tubes were set in a rack, covered loosely with aluminum foil (so that any starch granules present in the air were less likely to enter the centrifuge tubes) and allowed to dry before sealing with the centrifuge tube caps, sealed in labelled plastic zip-lock bags, and transported to Calgary. They were then further processed to remove clays and organics. Five to 10 millilitres (ml) of 1% NaEDTA (disodium ethylene diamine tetra acetate) was added to each sample, vortexed well to mix, and then placed on an orbital shaker for two hours. Each tube was then filled with deionized water, vortexed, and centrifuged at 3000 rpm for 5 minutes, after which the supernatant was pipetted off (with separate, clean disposable pipettes) and discarded. The rinsing procedure was repeated 3-

5 times until the supernatant was clear, effectively removing the clays. Five ml of 6%

H2O2 (hydrogen peroxide) was then added to each sample to gently oxidize organics and

the tubes were placed on an orbital shaker for 30 minutes after which the samples were rinsed as for the NaEDTA treatment. Samples were dried in a drying oven set to 40°C (so as to not gelatinize starch granules) and further processed for starch recovery as detailed below in section 5.5.3.

5.5.2 Ceramics

All ceramic sherds sampled for adhering charred residues were washed with distilled water and allowed to dry (loosely covered with aluminum foil) prior to sampling.

157

This step was performed to remove adhering sediment and decrease the chance that starch granules present in the sediment might remain on the charred residue. A separate

clean stainless steel dental pick was used for each sherd to gently scrape the charred

residues onto aluminum foil squares (as illustrated in Figure 5.4), which were then

transferred to new, cleaned and labelled sample containers, placed in new, labelled zip-

lock bags, and transported to Calgary for further processing.

Figure 5.4. Ceramic sherd from Trapichillo (Catamayo Phase B, CR-70) being sampled for charred interior residues.

Not all samples from ceramics consisted of charred residues. Several intact

ceramic bottles were also sampled (Zarrillo and Valdez 2010, 2011). Figure 5.5 shows

158

one of the stirrup-spout bottles in the ultrasonic bath. For these, two samples for each bottle were obtained.

Figure 5.5. Stirrup-spout bottle from Santa Ana-La Florida during and after sonication procedure to remove interior residues. The bottle was constantly monitored during the procedure to ensure that the sonication treatment was not causing any damage to the bottle (A and B). After sonication, red paint was visible outlining the mouth of one of the faces (image C), which was not noted previously as the bottles had not been washed since excavation. The bottle has two faces on opposite sides and image D shows the other face.

159

The first sample consisted of adding 150 ml of distilled water to each bottle, swirling and vigorously shaking the water around in the bottle, and the water decanted into new, cleaned and labelled, centrifuge tubes. Another 150 ml of distilled water was then added to each, shaken vigorously, and then left ~12 hours (overnight). The entire bottle was then placed into an ultrasonic bath and sonicated for 30 minutes to dislodge residues from interior cracks and crevices, after which the water was decanted into new,

cleaned and labelled centrifuge tubes. All of the samples were centrifuged at 3000 rpm

for 5 minutes and the supernatant was pipetted off and discarded. Samples were then

loosely covered with foil and left to dry, capped, placed into new, labelled zip-lock bags,

and transported to Calgary for further processing.

Another ceramic sherd was also uniquely sampled for residues. A ceramic pot

from SALF had been removed in situ encased in its original matrix (wrapped in

aluminum foil), and was excavated in the Riobamba lab by Francisco Valdez and I.

Because of the controlled conditions I was able to obtain several samples that I usually would not have access to. A sample of what appeared to be charred residue/sediment

(sample CR-54) and an additional sediment sample from the matrix was obtained (both tested negative for starch, i.e., <3 starch granules recovered). One of the sherds was sampled in a manner similar to sampling for stone tools (Chandler-Ezell and Pearsall

2003; Pearsall et al. 2004) – dry brush, wet brush and ultrasonic washing. The reason the sherd was sampled in this manner was because it had been noted on a different sherd from the site that charred interior residues came off of the sherd along with the sediment when washed – even if the sherd with adhering sediment was allowed to dry before

160

attempting to wash off the sediment (this was possible, however, with sherd CR-63, where some of the residue was sent for AMS C14 determination). These sample types are:

1. Dry brush of sediment adhering to the ceramic sherd representing sediment in

close proximity to the tool (CR-55);

2. Wet brush with distilled water of sediment adhering to the ceramic sherd

representing sediment in immediate contact with the tool (CR-56);

3. Ultrasonic washing with distilled water representing sediment and residue

contained in cracks and crevices of the ceramic sherd representing residues

related to use (CR-57).

A new, cleaned (boiled in acetic acid in a pressure cooker) soft-bristled tooth

brush was used to brush sediment from the sherd into a new, similarly-cleaned container.

This sample (CR-55) was then transferred to a new, clean and labelled sample container.

For the wet brush sample (CR-56), while holding the sherd over a clean container, distilled water was gently sprayed onto the interior surface of the sherd and then lightly

brushed with new, clean toothbrushes. Finally, the sherd was placed interior side down

into a new, clean container with some distilled water and the container was placed onto a

rack in an ultrasonic bath (without heat) and sonicated for 30 minutes (CR-57). The wet

samples were centrifuged at 3000 revolutions per minute (rpm) for 5 minutes and the

supernatant decanted, leaving a “residue” pellet. The centrifuge tubes were set in a rack,

covered loosely with aluminum foil (so that any starch granules present in the air were

less likely to enter the centrifuge tubes) and allowed to dry before sealing with the

centrifuge tube caps, sealed in labelled plastic zip-lock bags, and transported to Calgary.

161

They were then further processed to remove clays and organics, as detailed in Section

5.5.1, after which they were subjected to heavy density liquid flotation/separation in the same manner as other samples (Section 5.5.3).

Once charred residue samples were chosen for analysis, and where sample size was deemed “adequate”, individual samples were split so that some sample would remain for AMS radiocarbon (or other) analysis. Samples to be processed for starch recovery were processed as previously described (Zarrillo et al. 2008), with some modifications.

Samples were placed into new, cleaned and labelled centrifuge tubes, 3-5ml of 6% H2O2

was added to each, and the samples homogenized. Separate, cleaned low-shear

homogenizer rotors were used for each sample. Tubes were then placed on an orbital

shaker for 30 minutes. The centrifuge tubes were then filled with deionized water and

rinsed as described above for NaEDTA treatment of sediment samples (Section 5.5.1).

The H2O2 treatment and homogenization was performed to gently oxidize the charred

matrix and release starch granules. Samples were dried in a drying oven set to 40°

Celsius and further processed for starch recovery as detailed below in section 5.5.3.

5.5.3 Heavy Density Liquid Separation for the Recovery of Starch Granules from Stone Tool and Ceramic Residues

Heavy density liquid flotation/separation for starch recovery was performed as

previously described (Zarrillo et al. 2008), with some modifications. Five ml of 1.8 g/cm3

sodium polytungstate (SPT) was added to each sample. The samples were vortexed well

and then homogenized (as described previously in section 5.5.2), after which they were centrifuged at 3000 rpm for 5 minutes. The supernatant was decanted into new, cleaned 162

and labelled “starch extract” tubes (<1.8 g/cm3 fraction, containing starch granules that

have a density of ~1.5 g/cm3). The remaining residue in the original centrifuge tubes is

the >1.8 g/cm3 fraction, which contains any phytoliths present; these fractions were rinsed, dried, and have been retained for future analysis. The “starch extract” centrifuge

tubes were then topped up with deionized water, centrifuged for 5 minutes at 3000 rpm,

and the top ~1/3 of the supernatant was pipetted off. The tubes were refilled with

deionized water, recentrifuged, and the top ~1/2 of the supernatant was pipette off. This

was repeated a third time with the top ~3/4 of the supernatant pipette off. The first three

rinses are done in this fashion in order to sequentially decrease the density of the

water:SPT solution and avoid pipetting off any small starch granules that may be lighter

than 1.5 g/cm3. Three subsequent deionized water rinses were then performed, pipetting off the supernatant to as close to the residue pellet as possible, in order to rinse out the remaining SPT. Table 5.2 shows the archaeological stone tool and ceramic charred residue samples and their contexts that were processed for starch recovery, as well as samples submitted for AMS radiocarbon age determinations.

163

Table 5.2. Archaeological Sites and Samples Tested for Starch Granules. Cerro Narrio Sample Sample analyzed retained UofC Accelerator Mass (g) +/- (g) +/- Lab Spectrometry C14 Context Sample Type .0001 .0001 Number Sample Code # Cerro Narrio Unit 4 Level 3, FS1, SP30 ceramic charred residue 0.0922 0.0687 CR-101 Level 3, FS1, SP31 ceramic charred residue 0.1115 0.1065 CR-102 CN-U4-CR103, Level 3, FS1, SP60 ceramic charred residue 0.1845 0.2792 CR-103 Beta-312074

Cerro Narrio Unit 3 Level 9 (80-90BS), FS2 ceramic charred residue 0.3126 0.3573 CR-108 CN-U3-CR109, Level 11, FS3, sample 52 ceramic charred residue 0.184 0.1914 CR-109 Beta-312073 CN-U3-CR110, Level 12, FS37, sample 79 ceramic charred residue 0.1756 0.2585 CR-110 Beta-315534

Cerro Narrio Unit 3A S1/2 Layer 3, FS-09-01, sample 60 ceramic charred residue 0.1658 0.0937 CR-104 Layer 4, Level 1, FS-09-02, sample 62 ceramic charred residue 0.2461 0.4435 CR-105 Layer 4, Level 3, FS-0904, sample 66 ceramic charred residue 0.1885 CR-106 Layer 4, Level 4, FS-09-05, sample 71 ceramic charred residue 0.147 0.2016 CR-107

Cerro Narrio Unit 7 Layer 2, Level 14, FS-09-26, sample 23 ceramic charred residue 0.5879 1.8111 CR-111 Layer 2, Level 15, FS-09-27, sample 52 ceramic charred residue 0.6628 1.2774 CR-112 , Layer2, Level 15, FS-09-27, sample 56 ceramic charred residue 0.4563 1.0785 CR-113 Beta-312075 Layer 2, Level 16, FS-09-28, sample59 ceramic charred residue 0.0892 CR-114

164

Chaullabamba Sample analyzed Sample UofC Accelerator Mass (g) +/- retained Lab Spectrometry C14 Context Sample Type .0001 (g) Number Sample Code # CHB-III-H Level 5, Sample 17 ceramic charred residue 0.2055 0.886 CR-86 CHB-III-H Level 5, Sample 20 ceramic charred residue 0.1412 0.222 CR-87 CHB-III-H Level 5, Sample 21 ceramic charred residue 0.2024 1.2796 CR-88 CHB-IIIHL5-CR89, CHB-III-H Level 5, Sample 22 ceramic charred residue 0.2112 0.2648 CR-89 Beta-312072 CHB-III-H Level 6, Sample 14 ceramic charred residue 0.213 0.2609 CR-83 CHB-III-H Level 6, Sample 15 ceramic charred residue 0.21 CR-84 CHB-III-H Level 6, Sample 16 ceramic charred residue 0.1525 0.1683 CR-85 CHB Cut 1 Level 3, Sample 1 ceramic charred residue 0.123 CR-79 CHB-C1L3-CR80, CHB Cut 1 Level 3, Sample 2 ceramic charred residue 0.1095 0.2557 CR-80 Beta-312071 CHB Cut 1 Level 3, Sample 3 ceramic charred residue 0.1921 CR-81 CHB Cut 1 Level 3, Sample 10 ceramic charred residue 0.0989 0.0616 CR-82

La Chimba Sample analyzed Sample UofC Accelerator Mass (g) +/- retained Lab Spectrometry C14 Context Sample Type .0001 (g) Number Sample Code # TP-5 cat#58-1, L13 ceramic charred residue 0.68 0.6429 CR-98 TP-5 cat#58-2, L13 ceramic charred residue 0.159 0.1409 CR-99 TP-7 cat#211-1, L23 ceramic charred residue 0.2293 0.1717 CR-94 TP-7 cat#211-2, L23 ceramic charred residue 0.2576 0.5235 CR-95 TP-7 cat#211-3, L23 ceramic charred residue 0.1842 0.1174 CR-96 TP-7 cat#211-4, L23 ceramic charred residue 0.2834 0.2309 CR-97 TP-7 cat#236 -1, L28 ceramic charred residue 0.3557 0.9252 CR-90 TP-7 cat#236-2, L28 ceramic charred residue 0.2525 0.2006 CR-91 TP-7 cat#236-3, L28 ceramic charred residue 0.3977 0.6725 CR-92 LC-TP7-CR93, TP-7 cat#236-4, L28 ceramic charred residue 0.5588 0.6685 CR-93 Beta-312076

165

Santa Ana - La Florida Sample analyzed Sample UofC Accelerator Mass (g) +/- retained Lab Spectrometry C14 Context Sample Type .0001 (g) Number Sample Code # Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.1045 0.1086 CR-40 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.109 CR-41 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.1061 0.108 CR-42 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.1249 0.1006 CR-43 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.0373 CR-44 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.0944 CR-45 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.0946 0.1799 CR-46 Extola 1, Yellow Fill layer, 95 cm ceramic charred residue 0.0335 CR-47 XIV- 4 TOMB 1, S2W, lowest fill level, ceramic charred residue, 220 cm, ceramic pot w impressions interior 0.0084 CR-48 ceramic charred residue, exterior 0.011 CR-49 possible charred res of sherd CR-54 dry-brush interior of sherd CR-55 wet-brush interior of sherd CR-56 sonicated interior of sherd CR-57 possible charred res 0.533 0.2464 CR-58 XII-5 Shaft Tomb, stirrup-spout face bottle interior rinse CR-50 interior Sonicated CR- 51 XII-5 Shaft Tomb, stirrup-spout donut bottle interior rinse CR-52 interior Sonicated CR- 53 SALF-X4(17)-CR63, X4(17) Sherd 6 ceramic charred residue 0.0874 0.0529 CR-63 Beta-312078 X4(17) Sherd 7 ceramic charred residue 0.0549 CR-64 X4(17) Sherd 8 ceramic charred residue 0.0465 CR-65 XIV-4 (3-9) Tomb, stone bowl 1 interior sonicated STR-47 exterior sonicated STR-48 XIV-4 Tomb, stone bowl 2, polka-dot bowl whole bowl sonicated STR-49 XIV-5 Tomb, stone bowl 3, symmetrical bowl whole bowl sonicated STR-50 XII-5 Shaft Tomb, stone bowl 4, asymmetrical bowl whole bowl sonicated STR-51 XII-5 Shaft Tomb, stone bowl 5, red incised bowl whole bowl sonicated STR-52

166

Tajamar Sample analyzed Sample UofC Accelerator Mass (g) +/- retained Lab Spectrometry C14 Context Sample Type .0001 (g) Number Sample Code # W of Unit 137, Feature R14 cat# Z3 B1- 017, sample 2 ceramic charred residue 0.015 CR-66 Unit 151 Level D9 cat# Z3 B1-017, TA-U151-CR67, sample 7 ceramic charred residue 0.15 <0.15 CR-67 Beta-315535

Unit 137 Level D9 cat# Z3 B1-017- TA-U137-CR68, Maz6 ceramic charred residue 0.3237 0.688 CR-68 Beta-312079

Unit 63 Level D8 cat #Z3 B1-017, sample 3 mano sonicated residue STR-87

W of Unit 137, Feature R14 cat# Z3 B1- 017, sample 4 metate sonicated residue STR-88

W of Unit 137, Feature R14 cat# Z3 B1- 017, sample 5 mano sonicated residue STR-89

Unit 137 Level D3 cat# Z3 B1 0/7 M36, sample 6 metate sonicated residue STR-90

Trapichillo and La Vega #11 Sample analyzed Sample UofC Accelerator Mass (g) +/- retained Lab Spectrometry C14 Context Sample Type .0001 (g) Number Sample Code # CA-PhA-CR69, Catamayo Phase A cat#1102, sample 1 ceramic charred residue 0.7438 1.3254 CR-69 Beta-312070 CA-PhB-CR70, Catamayo Phase B cat#1103, sample 2 ceramic charred residue 0.1004 0.1507 CR-70 Beta-315533 La Vega Zone III, Level 40-45, cat#0070, sample 17 ceramic charred residue 0.6311 0.8568 CR-71 La Vega Zone CV, Level 6-7, cat#0679, LV-0679-CR72, sample 19 ceramic charred residue 0.2308 0.3171 CR-72 Beta-312077 La Vega Zone III, Level 5, cat#0094, sample 13 ceramic charred residue 0.2936 0.3369 CR-73 La Vega Zone XVIII, cat#0133, sample 6 ceramic charred residue 0.4284 0.3841 CR-74

167

5.5.4 Microscope Slide Preparation and Microscopic Examination

A single-channel volume pipettor was used to place an aliquot of “starch” extract on a clean, labelled slide. New and cleaned single-use pipette tips were used for each sample and the aliquot was aspirated from the bottom of the starch extract centrifuge tubes immediately after centrifuging the samples prior to slide preparation. The aliquot was allowed to dry on the slide before resuspending the residue with 50:50 glycerine:water (as a mounting medium). A cover slip was applied and sealed with nail polish – separate applicators were used to avoid cross-contamination. Because the

“starch” extracts are not free from charred (especially in samples from ceramic charred residues) and other organic and non-organic particles, a standard volume of “starch extract” is not always used when preparing microscope slides. The volume is determined by the clarity of the extract, with smaller volumes being employed for extracts that contained abundant charred particles. Volumes range from 10 μl to 160 μl, thus making quantification difficult in most cases, although volumes applied to each microscope slide were recorded (see Chapter 6, section 6.2 and Chapter 7, section 7.2.4 regarding quantification). Microscope slides were scanned in their entirety under cross-polarized light with a research-grade Zeiss transmitted light microscope and photomicrographs were taken with a CCD camera that enabled live-viewing capture to a computer.

5.6 Methods to Control Contamination of Archaeological Samples with Modern Starch

Starches and modified starches are everywhere – their use is common in adhesives, paper, textile products, and food (used in gelling, thickening, antistaling, texturizing, 168

moisture-retention, adhesion, and stabilizing), as binding agents and fillers, in soaps, and

as “powder” for rubber, surgical and disposable gloves, and can become liberated and

airborne by both human activities and natural phenomena, potentially contaminating

artifacts and laboratory samples (Dave et al. 1999; Laurence et al. 2011; Lauriere et al.

2008; Newsom and Shaw 1997a, Newsom and Shaw 1997b, Schwartz and Whistler

2009:1-10; Thomas and Atwell 1997:1). For example, stone bowls and mano and metate

fragments from Cotocollao were to be included in this study, but, upon testing, their

residues were all found to be contaminated with wheat, rye, and/or barley starch granules.

The stone bowls were stored in open boxes in a multi-purpose office/kitchen in the INPC

offices in Quito, Ecuador. Although great care was taken in cleaning all of the

implements and containers used in removing residues from the stone bowls (as per my

usual procedures, see below Section 5.6.2), and I thoroughly cleaned the counter in the

office where the artifacts had been stored and where I performed the sampling, upon

analysis numerous (>20 per slide transect, equalling hundreds per slide) starch granules

consistent with wheat, rye, and/or barley were found in all of the wash samples; starch

granules consistent with Zea mays (maize) were also present. As the control sample was

negative, I must conclude that the artifacts had been contaminated with modern starch(s)

prior to my sampling. This may have occurred in any number of ways, with a few

possibilities being: 1) as all of the artifacts were collected from the surface of the site,

they may have been contaminated with modern starch while exposed; 2) the artifacts may

have been placed in reused bags that contained modern starches (although none of the

stone bowls were in bags when I accessed them); 3) powdered gloves may have been

169

used during washing of the artifacts; 4) the boxes that the artifacts were stored in may

have been contaminated with modern starches; and/or 5) since the room was also used as

a kitchen/dining room, modern airborne starches may have been introduced onto the

surfaces of the stone bowls. Whatever the source of the contamination, and although I

briefly scanned the sonicated sample microscope slides, I did not feel sufficiently

confident to include any results from these samples; there was no way of knowing

whether the starches (in particular maize starch as they were found in the wash samples)

present in the sonicated samples were a result of modern contamination or original tool

use.

Laurence et al. (2011) provide a thorough overview of how pervasive airborne starch

granules are, including how starch inside pollen grains can be liberated into the air during

thunderstorms. Importantly, they also show how some liberated pollen grain starches are

highly similar to seed starch granules from the same species. Specifically, maize pollen

starch granules are morphologically comparable to maize seed starch granules, although

52% of the seed starches were 15 micrometers (μm) or less in size as opposed to 83% for

the pollen starches (i.e., almost 50% of a representative population of starches from

maize seeds should be >15 μm in size vs. 17% for maize pollen) (Laurence et al.

2011:223). These findings may help to discriminate between airborne maize pollen starch

and maize seed starch found on artifact residues if excavation, handling and storage

procedures are in question. Because of the ubiquity of starch in the environment from

multiple sources, great care needs to be taken to guard against incorporating modern

starch granules in archaeological samples. Hart (2011) also recently tested survey-

170

collected artifacts for phytoliths and starch granules. What he found was that dry brush samples (as opposed to wet brush and sonicated samples) from pottery, in some cases, may contain phytoliths and starch granule assemblages consistent with surrounding sediments, but that modern phytolith and starch granules were not found deposited “in the pores and crevices of the artifacts where primary residues are most likely to be hidden”

(Hart 2011:3252). Therefore, if overlying sediment/residue samples (dry brush and wet brush samples) are systematically removed, then the sonicated residue samples are unlikely to be contaminated with modern starch and/or phytoliths. As such, numerous precautions, similar to those used and/or recommended by others (e.g. Hart 2011;

Laurence et al. 2011; Loy and Barton 2006; Mercader 2009; Mercader et al. 2008;

Zarrillo and Kooyman 2006) were taken during excavation, sampling, storage, laboratory processing, and analysis to ensure that the starch granules recovered in this analysis are

derived from the archaeological contexts.

5.6.1 Controlling for Starch Contamination during Excavation

During excavations at Cerro Narrío all suspected stone tools were immediately

placed in new plastic bags and/or wrapped in new aluminum foil upon excavation.

Similarly, ceramic sherds bearing interior charred residues that were found in situ or in

the screen were wrapped in new aluminum foil. Occasionally some sherds with charred

residues were missed and placed in the level bag with other sherds, but were then

recovered either before or after being washed (with water) and immediately wrapped in

new aluminum foil; this was noted on the foil wrapping, and for added assurance none of

171

these sherds were selected for analysis. Sediment samples from all excavation levels were collected for microbotanical analysis and to test for the possibility of starch granules being present in the sediments from which artifacts were recovered. Of the sediment samples tested that were from the from the same context as artifacts sampled for this analysis, none were found to contain more than 1-3 starch granules per 0.5 g sediment, similar to the results obtained by others for the presence of starch in sediments (e.g.,

Barton et al. 1998; Mercader et al 2008:293-298; Zarrillo and Kooyman 2006:485).

Artifacts sampled from other archaeological sites from which samples for this analysis were obtained had varied means of artifact storage, although none were stored in open boxes (like the stone bowls from Cotocollao) and all were in plastic bags. For some of the sites, ceramic sherds with charred residues were in separate bags (either individually bagged/wrapped in foil or in one bag together) from other sherds and artifacts, while for

other sites sherds with residues were in the same level bag as sherds without charred

residues. Sherds selected for sampling were individually wrapped in new aluminum foil

with the site and provenience information recorded and set aside for sampling. Stephen

Athens (2010, personal communication) states that La Chimba sherds that were noted to

posses food residues were set aside from the other sherds at excavation, washed with

water, wrapped in aluminum foil when dry, and placed in individual plastic bags for

storage. All stone tools sampled for this analysis (excluding Cotocollao) had been stored

in individual bags, separate from other artifacts.

172

5.6.2 Controlling for Starch Contamination during Sampling and Storage

Sampling of artifacts primarily occurred in the USFQ Riobamba laboratory set aside for our use. Prior to using the USFQ lab, I cleaned the walls, surfaces, sinks, drawers, cupboards and floor with a ~.5% solution of NaClO (sodium hypochlorite/bleach), followed by vinegar. The door and windows were always kept closed. As for the Calgary lab, no paper products (paper towels, etc.) were used in the

Riobamba lab, food was not allowed in the lab, and any fabric cloths were boiled in vinegar in a pressure cooker prior to use. The only other locations in Ecuador used to sample artifacts included a house we rented for the 2007 field season, the INPC offices in

Quito (as noted previously), a storage room at the Banco Central Museum in Cuenca, and the laboratory at the Rumipamba archaeological park, Quito. At the rental house (which was newly built and not previously used) a room separate from our living space was used as a laboratory. I cleaned this room prior to, and between, any sampling in a similar method as detailed for the USFQ Riobamba lab; some of the Cerro Narrío artifacts were sampled there (Units 3 and 4). As previously noted, despite my attempt at “starch hygiene”, the stone bowls from Cotocollao that were sampled in the INPC offices in

Quito were found to be contaminated (likely prior to my sampling) with modern starch.

The Banco Central storage room was cleaned prior to sampling sherds with charred residues (no stone tools were sampled) from Trapichillo and La Vega, and all sampling occurred in one day within a three hour period. The Rumipamba archaeological park laboratory had a “proper” laboratory, where food was never allowed, that was separate from other areas of the archaeological facility (offices and cataloguing lab) (Victoria

173

Dominguez, personal communication). Despite this assurance, I thoroughly cleaned the

laboratory area prior to sampling the Tájamar artifacts, which occurred over a four hour

period in one day. Finally, the ceramic sherds from La Chimba were sampled in the

South American Research Laboratory in the Department of Archaeology at the

University of Calgary as they had been sent directly from Hawaii where they had been

stored since excavation; all of the sherds were individually bagged and labelled, as noted

above, and were rewashed with distilled water prior to sampling, as detailed below.

All sherds with charred cooking residues were rinsed with distilled or deionized

water (sprayed from a new, cleaned spray bottle), set on aluminum foil and loosely

covered with another sheet of aluminum foil, and allowed to dry prior to sampling. Stone

tools were also rinsed with water in the same manner as the ceramic sherds. Although the

ceramic sherd wash samples were not retained for analysis, the stone tool wash samples

were – all tested negative for starch; that is, no sample had more than 3 starch granules

present (except the above noted wash samples for the Cotocollao stone tools). At no time

were artifacts left uncovered when not wrapped in aluminum foil or stored in plastic

bags. Disposable plastic gloves (previously tested for starch and found to be starch-free)

were worn when handling artifacts (see Figure 5.4). Other precautions that were taken

included retaining control samples of the distilled water that was used to wash off

artifacts and to sonicate tools, and control samples of the cleaned dental picks, aluminum

foil, storage vials, plastic containers, and any other instruments or materials used during

sampling (see section 5.6.3 below for comments on “blank control samples” during

processing). In 2007 I used distilled water sold in sealed bottles (Cerro Narrío units 3 and

174

4 artifacts) and thereafter I used distilled water freshly processed specifically for me by a

pharmacy in Riobamba; none of the distilled water control samples were found to contain

starch.

The dental picks used to scrape charred residues off of the sherds, as well as the

new centrifuge tubes, new plastic containers used during sonication, and the new sample

containers that the charred ceramic residues were placed in were washed with bleach,

rinsed with distilled water, and then boiled in 5% acetic acid (vinegar) in a pressure

cooker prior to use. This method has been shown effective to gelatinize/destroy starch

granules (Zarrillo and Dickau 2008, unpublished results). The acidity of the vinegar (~pH

2-3), coupled with a boiling temperature of >100°C made possible with the use of a

pressure cooker, is more effective than boiling instruments and containers in regular

distilled or deionized water (in a conventional pot) because acid hydrolysis and an

ensured boiling temperature above the gelatinization temperature of native (unaltered)

starch granules are both working to destroying starches.

All residue samples (both charred ceramic and stone tool) were double-bagged in new

zip-lock bags (labelled with provenience) and then stored in cleaned plastic bins for

transport to the University of Calgary. The samples remained in the zip-lock bags and

storage bins and were stored in the South American Research Laboratory in the

Department of Archaeology at the University of Calgary, and were only removed for

sample processing.

175

5.6.3 Controlling for Starch Contamination during Laboratory Processing and Analysis

All of the final processing of samples was performed in the South American

Research Laboratory in the Department of Archaeology at the University of Calgary. In

2007 the laboratory was thoroughly cleaned and reorganized and a section of the lab (wet lab area) was set-up for starch analysis. As noted earlier, no modern plant specimens are stored or processed in this lab. Since the starch laboratory was set up, I have always left one microscope slide out on the lab bench where processing occurs and one slide in the fume hood. Prior to starting an analysis, I check these slides for any airborne starch that may have “landed” on the slides, put out new slides that are checked once the analysis is complete, and then another set of slides are placed until the next analysis. From 2007 until the spring of 2010 all of these control slides were either negative for starch or had,

at most, no more than 3 starch granules on a slide. From the summer to the fall of 2010

construction work was being done in the Earth Sciences building, including work on the

air ducts. Control slides left out during this time period resulted in >20 starch granules

per slide. I suspected that the gross starch contamination was a result of the construction

and duct work retrofitting that had occurred, and the lab was again thoroughly cleaned.

As a further precaution, being concerned that starch from construction materials and

debris (sheetrock, etc.) was airborne and being distributed through the building’s

ductwork, and that any starch adhering to the interior of the ductwork might also be

dislodged and distributed throughout the building, I secured 3M Filtrete© High

Performance Electrostatic Filters over the air vents in the laboratory (see Figure 5.6).

176

These filters are capable of capturing particles between .3 and 1 μm6, which in

theory should capture airborne starch and pollen that would be of consequence (starches

less than even 4-5 microns would be very difficult to identify without the use of electron

microscopy). Control slides were then placed on the lab bench and in the fume hood and

remained out from the spring of 2011 until the fall of 2011; that is, the same length of time and over the same season as the control slides noted above that had >20 starch granules per slide in 2010. These new control slides tested negative for starch granules, indicating that the re-cleaning of the lab and addition of filters over the ductwork vents was effective in eliminating airborne starch (at least to the extent that no starch was found on the control slides).

Figure 5.6. Air vents in the South American Research Laboratory covered with 3M Filtrete© High Performance Electrostatic Filters. Image A shows two of the three vents and image B is a close-up one vent. These images were taken ~6 months after the filters were installed – they were white in colour when new.

As well as the control slides that were left on the lab bench and in the fume hood throughout the period of time that processing of samples occurred for this analysis,

177

additional slides were left out for each batch of samples processed. For example, I usually process samples in batches of three, seven, 11 or 15. The number of samples processed is determined by the number of samples from a context (samples from one site, or more than one site at a time) and the types of samples (I do not run ceramic residue samples at the same time as stone tool or sediment samples). Because the centrifuge has four holders for 50ml centrifuge tubes, I can only run samples in multiples of four. With each batch of samples, a blank control is included. Therefore, batches will consist of three archaeological samples and one control sample, or seven archaeological samples and one control sample, or 11 archaeological samples and one control sample, or 15 archaeological samples and one blank control. Because more than 15 samples is difficult to manage and this may lead to mistakes being made during processing, I never run more than 15 samples in one processing batch. Therefore, in addition to the control slides left out from the time I started processing samples until the time I finished processing samples for this analysis, I also left out control slides for the time period that individual batches of samples were processed. All of the control slides left out during the individual batch runs as well as the control slides left out from spring 2011 until fall 2011 were negative for starch (at least to the extent that no starch was found on the control slides).

In addition to the control slides to test for airborne starch, as noted above, blank

“processing control samples” were incorporated with every batch of samples processed.

These controls are used to test for modern starch contamination and/or cross

contamination of starch from one sample to another. The blank control sample undergoes

all of the same procedures as the archaeological samples from the start of sampling

178

through to final microscope slide preparation. When removing charred residues from pottery sherds, a dental pick cleaned at the same time that dental picks used to scrape residues off of the sherds, was swirled into a similarly cleaned sample vial (labelled as

“sampling blank control” with the sample numbers indicated) with some of the distilled water that was also used to wash off the sherds prior to sampling; an additional unused aluminum foil square, from the same aluminum foil roll as the foil used when scraping off charred residue, is also rinsed with distilled water into the “sampling blank control.”

A similar procedure to test all of the utensils and containers used in sampling stone tools for residues was also performed. These control samples were tested for starch prior to the start of additional processing in Calgary (all were found to be negative for starch) and also incorporated as the “processing blank controls”. I feel that incorporating blank control samples during processing is one of the best ways to ensure that modern starch has not contaminated archaeological samples (Zarrillo and Kooyman 2006:485), and also allows the analyst to isolate and address issues if control samples do test positive for starch (see also Chapter 7, section 7.2). None of the “processing blank controls” resulted in the recovery of more than three starch granules on an individual slide.

Other routine precautions that are taken to prevent modern starch contamination include, frequent cleaning of the lab benches and equipment with a wet cloth (not a dry cloth that may simply make dust and starch granules airborne) before, during and after processing, frequent (at least once-per-week) wet-mopping of the floor, food is not allowed into the lab, and all instruments, containers, centrifuge tubes, microscope slides, disposable pipettes, cover slips, etc. used in sampling and processing are boiled in

179

vinegar in a pressure cooker for a minimum of 30 minutes and then stored in cleaned containers in cupboards. In addition, all cleaning products, reagents, solutions, lab wipes, etc. have been tested and found to be starch-free. Although the heavy density liquid

(SPT) used to separate starch from the charred cooking and stone tool residues was

collected, filtered (with a 25μ filter initially to remove larger particles and then three

more times with a 2-5μ filter), and then boiled for reuse (for phytoliths analysis only), only new SPT was used for all of the starch extractions for this analysis. Also, as

mentioned previously, samples and artifacts were never left uncovered, exposed to the air.

5.7 Methods of Starch Granule Identification

As noted in section 5.5.4, the microscope slides were all scanned in their entirety.

Despite all of the precautions taken to decrease the chances that modern starch might be

introduced into archaeological samples, and as it is always possible that a few modern

starch granules may be present, microscope slides that had less than five starch granules

were considered “negative” results. In such cases, a second (and often 3rd and 4th) microscope slide was prepared to confirm the “negative” results. Identifications of starch granules are based on comparing multiple three-dimensional characteristics of the archaeological starches to modern comparative starches. When starch granules were found on the microscope slide, they were catalogued on a data sheet with a unique identification number, described on the data sheet, initially measured using a reticule fitted in one of the microscope ocular lenses, and photographed from multiple views

180

(rotating the starch granule by applying gentle pressure on the coverslip) under plane- and cross-polarized light. Final, more accurate, measurements are made with the CCD camera image processing software, which was specifically calibrated for the microscope used. When describing granules, the following characteristics were noted:

1. Shape of the granule (spherical, ovoid, lenticular, reniform, polygonal, etc.);

2. Size of the granule (maximum length and width in microns);

3. Single granule vs. compound granule (presence or absence of articulation facets);

4. Angularity of edges (rounded, sharp, etc.);

5. Location and nature of the hilum (centric, slightly eccentric, very eccentric; at broad or

narrow end; closed or open hilum);

6. Presence and nature of fissures (single, multiple, radiating, Y-shaped, stellate, ragged,

and location – transverse or longitudinal);

7. Number and nature of facets – pressure facets (rounded and uniform edges) or

articulation facets (sharp and irregular edges);

8. Presence and nature of lamellae (highly visible, faint, concentric, etc.);

9. Extinction cross features (shape of the cross – right angled arms, acute/obtuse arms,

arms that broaden or constrict towards edges of the granule; strong cross or weak cross,

etc.);

10. Presence and nature of surface features (rough or smooth surface; equatorial groove;

projections, indentations, etc.).

Published descriptions and photographs of starch granules were also used to aid

identifications (e.g., Chandler-Ezell et al. 2006; Cortella and Pochettino 1994; Czaja

181

1978; Guilbot and Mercier 1985; Kostanje and Babot 2007; Lisińska and Leszczyński

1989; Moss 1976; Pearsall et al. 2004; Perry 2001, 2002; Piperno and Holst 1998;

Reichert 1913; Seidemann 1966; Young 1953) and these will be noted where necessary.

Table 5.1 shows the modern plant species obtained for starch granule comparison.

When archaeological starch granules possessed diagnostic characteristics –

including size, shape, pattern of the extinction cross, presence or absence of fissures,

lamellae, etc. – and could be confirmed by comparison to known modern starch samples,

they are identified to family, genus, or species as possible based on morphology. In some

cases where known modern starch samples were not available for comparison, and

identifications were made based on published descriptions and photomicrographs,

archaeological starches are identified as possible; that is, not confirmed (cf.). Where

archaeological starches had one or more of the diagnostic characteristics absent, they

were also identified as cf. Identifications and criteria used will be discussed further with

the results. Starches described as “damaged/unknown” could not be identified for several

reasons: they had enlarged cracks and/or were portions of a broken starch granule; and/or

they were obscured by adhering organics or charred material making morphological

identification impossible. Starch granules that are unknown but appear distinctive are

grouped by similar characteristics into “types” (e.g., “Type 1”, Type 2”). “Starch

clusters” are composed of several starch grains clumped together – they are not

considered to be compound grains nor are the starches necessarily from one species as

several different starch morphotypes are usually present (as opposed to “monodominant

clusters” [Mercader et al. 2008]). Finally, the “starch cluster dam/gel” category

182

encompasses starch clusters that show changes associated with boiling starches in excess water – swelling, loss of birefringence on some of the aggregated granules, and/or disruption of granule integrity and exudation of granule contents (Donald 2004:169-170;

Zarrillo et al. 2008:5007).

5.8 Chapter Summary

The types of samples available for paleobotanical analyses varied between the archaeological sites I could access for this dissertation. This, along with time constraints, led to the prioritization of starch analysis of the samples. Starch granule analysis was chosen as the method of paleobotanical analysis as it offered the best possibility for recovering the most complete record of past plant use from the artifacts available for analysis – stone tools and ceramic charred residues. In particular, had phytolith analysis been performed to the exclusion of starch granule analysis, many plant identification, specifically USOs, may have been missed. As the use of starchy roots and tubers (USOs) are important today in the Andes, as was the case in the past, it was important to use the best method of analysis that would capture this information. Samples were obtained from six highland archaeological sites – La Chimba, Tajamar, Cerro Narrío, Chaullabamba, La

Vega, and Trapichillo – and one site from the eastern slopes of the Andes, Santa Ana-La

Florida. All samples processed for starch recovery have been retained for future phytolith analysis. In order to identify starch granules recovered from the artifact residues, a collection of comparative modern plant species was developed. These samples allow for more secure identifications of starch granules than relying on published descriptions and

183

illustrations alone, although these sources of information are also important and are used in identifications. One of the most important aspects of the analyses performed was to

control for modern starch granule contamination throughout all steps, as possible – from

excavation, cataloging, sampling of artifacts, storage of samples, initial and final sample

processing, and slide preparation. The results of the control samples tested throughout all steps show that modern starch contamination was limited. For extra assurance, archaeological samples that had less than five starch granules recovered are considered

“negative” results, while those that resulted in five or more starch granules are considered to “positive” for starch. The starch granule analysis results are presented in Chapter 6.

Notes

1. Macrobotanical remains consist of plant parts that are visible and identifiable to the naked eye or with low-powered microscopy, including wood, fruits, seeds, nutshells, fibres, etc. (Ford 1979; also see Wright 2010 for recent overview).

2. Microbotanical remains are identified with high-powered microscopy (Ford 1979; also see Wright 2010 for recent overview), including pollen, spores, phytoliths, starch granules (also often called grains). Phytoliths are produced by some plants when soluble silica (also sometimes calcium oxalate), present in the ground water, is absorbed and then

solidifies in intra- and extra-cellular spaces (Esau 1965; Pearsall 1995, 2000; Piperno

1995, 1998, 2006a; Rovner 1983; Wright 2010). Phytoliths vary in size and shape

depending on the plant taxon and plant part where they are formed and many have been

identified that are specific to different taxa, allowing for identification based on

184

morphology coupled with, in some cases, statistical analysis (Pearsall 1995, 2000;

Piperno 1988, 2006a).

3. The C4 plants use an alternate mechanism of carbon fixation (the reduction of carbon dioxide to organic compounds by living organisms) during photosynthesis than the C3 plants. In C3 plants, the initial product produced by carbon fixation is a 3-carbon compound, whereas in C4 plants the carbon is first fixed as a 4-carbon compound

(Campbell et al. 1999:182-183). Carbon isotopes are present in the atmosphere as 13C and

12C with a constant ratio of 1:100 of 13C:12C (13C is slightly higher in ocean water).

13 Because C3 plants fix less C into their tissues than C4 plants, and these differences are

passed along the food chain to eventually be fixed in human (and animal) bone tissue, the

relative importance of C3 and C4 plants in human diet can be assessed by analyzing

human bones (see for e.g., DeNiro and Epstein 1978; Katzenberg et al. 1995; Schwarcz et

al. 1985; Van der Merwe and Vogel 1978). Maize is an important C4 plant for the region

of study.

4. Although definitions vary, botanists use many specific terms to differentiate

subterranean plant vegetative organs that are swollen with food reserves (often as starch

granules, but also as sugars) (Hickey and King 2000:40). These include: “bulb” (e.g., lily,

onion); “corm” (e.g., crocus, taro); “rhizome” or “rhizome tuber” (e.g., achira/Canna, iris); “root tuber” (e.g., cassava/manioc, sweet potato); “stem tuber” (e.g., potato) and

“taproot” (e.g., carrot) (Hickey and King 2000:56-59; Torrence 2006:36). To avoid the use of distracting and confusing technical terminology, and following Torrence

(2006:36), I am using the term “underground storage organ” (USO) throughout this

185

dissertation – although it should be noted that not all of these plant organs are found

exclusively or entirely underground (e.g., rhizomes).

5. Depending on location (Ecuador vs. Calgary) I had differential access to deionized and distilled water used for sampling (washing and sonicating artifacts), laboratory processing of samples, and slide preparation. In Ecuador, I used distilled water as it was processed specifically for me (and was thus more secure as far as modern starch contamination as I could obtain samples for control testing) and the laboratory and locations where I worked did not have piped-in distilled or deionized water. In Calgary I use (and have regularly tested) the piped-in deionized water available in the Earth

Sciences building. Basically, deionized water is made with water that has passed through an ion exchange column to remove charged particles, while distilled water is made from water that is boiled and the steam that rises from the boiling water is condensed. The condensate is largely free of any impurities that were present in the water before it was distilled. The water from the pharmacy in Riobamba was in actual fact deionized and then distilled making it pure enough for use in pharmaceutical preparations. For the purposes of my research, both are adequate. The important point is that all sources of water used, whether distilled or deionized were tested for modern starch contamination.

6. See: http://solutions.3mcanada.ca/wps/portal/3M/en_CA/AirQualityProducts/Filtrete/ProductI

nformation/Products/HighPerf/

186

Chapter Six: Starch Analysis and AMS Radiocarbon Assay Results

6.1 Introduction

As discussed in Chapter 3, few Formative Period sites have been identified in the

Ecuadorian highlands. Samples from six highland sites and one site from the eastern

slopes that date to the Formative Period (based on absolute dating from previous

investigations) were examined. The following are the results of starch analyses performed

on these sites and they are presented in chronological order based on radiocarbon assays obtained as part of my analyses. Only samples that tested positive for starch granules, i.e., five or more starch granules on a microscope slide, are included in the results tables

unless needed to aid in the interpretation of other results (generally from the same artifact

or context).

6.2 Santa Ana–La Florida

The site of Santa Ana-La Florida (SALF) is located near the modern town of Palanda on the eastern slopes of the Andes in the province of Zamora-Chinchipe, southeastern

Ecuador at an elevation of ca.1040 masl, as noted in Chapter 4. The Early Formative

occupation is supported by more than 20 radiocarbon assays (Valdez 2008:880; Valdez et

al. 2005; Zarrillo and Valdez 2010) that have been obtained from different contexts of the

site (see Chapter 4). Samples from SALF represent the earliest contexts analyzed for this

dissertation.

187

Nineteen separate artifacts were tested with 27 different samples analyzed (excluding wash control and standard control samples – all of which tested negative for starch, see

Section 5.5). Of the 27 samples (this total includes multiple samples from some artifacts), starch granules were recovered from 15 (5 or more starch granules in a sample); 11

(58%) of the 19 artifacts produced positive results for starch (Table 6.1). From the

samples where >4 starch granules were observed on a slide (positive result), a total of 794

starch granules were observed on the sample microscope slides. Although the results

from CR-52 and CR-54 were not “positive” they are included in Table 6.1 as they relate

to other samples from the same artifact or context, as discussed below, but they were not

used in calculating totals for different taxa, ubiquity, relative frequency, or the total

number of starch granules recovered from all the samples. Relative frequency/percentage

calculates the percentage the starch granules of a type (or taxon) present within the total

assemblage (Miller 1988:72; Renfrew 1973, in Hastorf 1993:166). For this calculation,

794 is used as the total starch granules (including individual damaged/unknown granules

and starch clusters/aggregates), and was calculated by:

Total starch granules by type (or taxon) Total starch granules recovered x 100

Ubiquity (or presence analysis) calculations describe the number of contexts that a type or taxon of starch granules is recovered from, regardless of the number of starch granules present for each type or taxon (Hastorf 1993:166; Pearsall 2000:212-216;

Popper 1988:60-64; Wright 2010:51-52). For this analysis, the total number of individual

188

contexts any type or taxon of starch granule could be recovered from was 11 (all the artifacts that tested positive for starch, excluding CR-52 and CR-54). For each type or taxon, a score of 1 was added each time it was present from one of the 11 contexts. Note that the following samples are grouped together as originating from the same context:

CR-51/-52; CR-55/-56/-57; and STR-47/-48; hence if a type or taxon was present in either CR-51 or CR-52 (and the same for the other groupings) it was only counted once.

The following is an example ubiquity calculation for Zea mays:

Total presence of type (or taxon) by context Total number of contexts x 100

Zea mays starch granules are present in CR-49 = 1 Zea mays starch granules are present in both CR-5/-51, same context = 1 Zea mays starch granules are present in CR-53 = 1 Zea mays starch granules are present in both CR-55/-56/-57, same context = 1 Zea mays starch granules are present in CR-63 = 1 Zea mays starch granules are present in CR-65 = 1 Zea mays starch granules are present in both STR-47/-48, same context = 1 Zea mays starch granules are present in STR-49 = 1 Zea mays starch granules are present in STR-50 = 1 Zea mays starch granules are present in STR-52 = 1 Total contexts Zea mays present 10

10 11 x 100 = 90.9%

Therefore, as shown in Table 6.1, Zea mays starch granules are present in the residues of

90.9% of the artifacts that tested positive for starch granules from SALF.

189

Table 6.1. Santa Ana-La Florida Starch Analysis Results.

d l pe ere r /ge r h te la ta am la sp a . . n p. l . . p known r d ticu pul spp p ule s ase-sha UC n u riangu spp p c a ys v ste ster s es spp a s s /un u , le 1, c 2, t ea a a y y d clu cl 1 um r e ot a ge Sample co o ih brom a h e 6, smal e s o Zea m m m c rch p yp io om a Number Sample Type 1 Ty Type 1 T Capsic D FabaceaeIp Man Marant The cf. Zea ma Zea D Star Sta Total per sample Ceramic Charred Residue - rim/exterior CR-49 dimples 4 1 5 Stirrup-spout Face CR-50 Bottle - Rinse 25 7 4 2 5 6 3 19 2 73 Stirrup-spout Face CR-51 Bottle - Sonicated 29 9 3 1 5 1 5 1 26 6 63 4 1 154

Stirrup-spout Donut CR-52 Bottle - Rinse 1 3 4

Stirrup-spout Donut CR-53 Bottle - Sonicated 2 8 1 1 8 3 3 6 22 2 48 5 109 Charred "residue"/sediment2 from matrix of CR-54 1 cermaic pot 1 1 2 Ceramic sherd - dry CR-55 brush interior 8 1 4 13 Ceramic sherd - wet CR-56 brush interior 2 3 1 1 1 1 9 1 26 1 46 Ceramic sherd - CR-57 sonicated interior 13 8 2 6 4 2 6 23 2 72 1 1 140 Interior charred CR-63 residue3 13 11315 1622 35 Interior charred CR-65 residue 1 5 6 Stone Bowl #1, STR-47 Interior sonicated4 51 1 11 18 241 43 Stone Bowl #1, STR-48 Exterior sonicated 8 1 1 2 3 7 1 1 24

Stone Bowl #2, STR-49 whole bowl sonicated 26 4 1 1 1 1 9 2 3 1 4 20 4 77 Stone Bowl #3, recontructed5, whole STR-50 bowl sonicated 5 2 7

Stone Bowl #4, STR-51 whole bowl sonicated 4 1 16

Red Incised bowl, STR-52 whole bowl sonicated 13 4 2 3 1 4 7 2 18 1 1 56 Totals 141 49 1 2 8 6 3 34 15 22 25 118 16 325 22 4 3 794 Ubiquity6 81.8 72.7 9.1 18.2 45.4 36.4 18.2 54.5 63.6 63.6 63.6 81.8 36.4 90.9 RF7 17.8 6.2 0.1 0.2 1.0 0.7 0.4 4.3 1.9 2.8 3.1 14.9 2.0 40.9 2.8 0.5 0.4

1CR-54 to CR-57 all related to same sherd from XIV-4 Tomb vessel. 2Encased in original matrix, excavated in the lab. Sample of charred residue/sediment and an additional sediment sample tested negative for starch. 3AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312078, 4450 ± 30 BP (measured), Cal BC 3500 to 3350 (95%) , 13C/12C -14.4 4ALL stone bowls were washed with distilled water prior to sampling to test for incidental starch - all were negative for starch. 5UHU Glue used in resonstruction tested negative for starch 6Ubiquity = Total number of contexts starch type was present/11 (number of separate artifacts tested) x 100. Results rounded up to one decimal place. 7Relative Frequency = total starch by type/total starch recovered from all samples (excluding CR-52 and CR-54) x 100. Results rounded up to one decimal place.

190

Results from CR-50 (rinse) and CR-51 (sonicated) are from the same context – the stirrup-spout face bottle. The sonicated sample, as expected, contained 2.1 times (over

200%) more starch than the rinse sample (154 vs. 73, respectively). A more dramatic increased recovery of starches (>2700%) was noted for samples CR-52 (stirrup-spout donut bottle rinse) and CR-53 (sonicated sample), wherein the sonicated sample had over

27 times the number of starch granules that the rinse sample contained (109 vs. 4, respectively). These results would seem to indicate that the sonication procedure was effective in recovering a greater number of starch granules than simply rinsing the bottles with distilled water and collecting the effluent (see Chapter 5, section 5.4.2 for method).

Although the same volume of distilled water (150 ml) was used in both procedures (rinse and sonicated) and the same amount of residue was mounted on the microscope slides (80

μl), the distilled water used for the sonication procedure had been left to sit in both bottles prior to sonication for ~12 hours and this alone may have resulted in greater starch granule recovery.

Samples CR-54, -55, -56 and -57 are also from the same context – a sherd from a ceramic pot that was excavated in the laboratory where it was possible to obtain several control samples (see Chapter 5, section 5.4.2 for method). In addition to the charred

“residue”/“sediment” sample CR-54 (that was thought to possibly be residue from the interior surface of the pot that had become dislodged and incorporated in the matrix) a sediment sample was tested from the general matrix sediment; both of these samples tested negative for starch – the CR-54 sample had 2 starches (negative results) and no starches were observed in the matrix sediment sample. Starch granule recovery increased

191

incrementally from the dry brush sample (CR-55) to the wet brush sample (CR-56) to the sonicated sample (CR-57), with 13, 46, and 140 granules respectively recovered in 40 μl of residue per slide. With the adhering sediment and matrix samples being negative for starch, the starch granules recovered from samples CR-55, -56, and -57 are thus representative of use-related residues and not transference of starch from the sediment to the surface of the ceramic sherd. Starch granule recovery with the sonicated sample was over 300% greater than the wet brush sample and over 1000% greater than the dry brush sample, similar to what would be expected with stone tool residues – that is, if a tool was used to process starchy foods, the greatest amount of starch should be present in the sonicated sample (Pearsall et al. 2004).

The final matched samples are STR-47 (interior sonicated) and STR-48 (exterior sonicated) from stone bowl #1. This was the only stone bowl where I was able to sonicate

the interior and exterior separately because it was not complete (the bowl had not been

reconstructed and I used the largest piece). The other bowls were either intact or had been

reconstructed and I was not able to separately sample the interior and exterior portions as

the bowls were too large for the ultrasonic bath I used in Ecuador. Note though that I

sampled the actual UHU glue used to reconstruct stone bowl #5 as well as wash samples

of all of the bowls, and all of these samples were negative for starch. The results for STR-

47 and STR-48 are interesting in that although no starch was found in the wash sample

from stone bowl #1, starch granules were recovered from the exterior of the bowl. The

portion of the bowl encompassed mostly the side (about 1/4 - 1/3 represented) with a

small portion of the bottom (see Figure 6.1). Therefore it is possible that contents of the

192

bowl had spilled onto the exterior surface during use, and/or the exterior surface came into contact with starchy plants (from surfaces or during handling) during food/drink preparation and the results are representative of original use. There were 179% more starch granules (n=43) in the STR-47 (interior) sample than the STR-44 (exterior sample, n=24) for the same volume of distilled water used during sonication (500 ml) and residue per slide (80 μl of residue).

Figure 6.1. Stone Bowl #1 from SALF exterior bowl sonication (Sample STR-47).

Twelve different morphotypes were originally defined, and of these eight were

identified to different taxa. Of the unidentified morphotypes, Type 1 starch granules

(Figure 6.2 A-B, see also Figure 6.21 E-F) are lenticular in shape with a diffuse cross (the arms meet at the hilum) and a sharper cross when viewed in profile (elongate oval), with

193

faint to prominent lamellae on the flat view (appear spherical). Very similar starch granules were observed in samples from Loma Alta (Zarrillo 2004) and although they are numerous (n=141) representing 17.8% of the starch granule assemblage, and are present in residues from 81.8% of the artifacts (second only to Zea mays in relative frequency

and ubiquity, Table 6.1), they are not well represented in samples from highland sites

(but also found in samples from Chaullabamba, La Vega and Trapichillo) (see sections

6.2, 6.5, and 6.6). Type 6 starch granules (Figure 6.2 B-C) are spherical, <8 μm, with

centric to slightly eccentric hila (sometimes slightly open) and sharp extinction crosses.

Spherical starch granules such as these are found in many species, even when other more

diagnostic starch granules are produced (e.g., Solanum tuberosum, potato), therefore

these types cannot be assigned with confidence to any taxa in the absence of diagnostic

characteristics. Type 11 (Figure 6.2 E-F) was present in only one sample and represented

by only one starch granule. Although it is a unique form, being cupulate, it may be a

starch granule altered by heat or partial hydrolysis and may thus be representative of any

number of species. This type of starch granule may be from Dioscorea spp. as cupulate

forms have been reported for species native to South America (Chu and Figueiredo-

Ribeiro 1991:473), although this form is not the diagnostic type (see below). Finally,

Type 12 starch granules (triangular) have a pyramidal three-dimensional shape, centric

hila, with some faint fissures emanating from the hilum in a Y shape following the edges

of the granule (Figure 6.2 G-H). As with the cupulate Type 11 starch granule, the Type

12 triangular starch granules may be from Dioscorea spp. or Maranta spp. as triangular

starches have been reported, although, again, this form is not the diagnostic type (Chu

194

and Figueiredo-Ribeiro 1991:473; Reichert 1913:813). In any event, Type 11 (n=1) and

Type 12 (n=2) starch granules only represent 0.3% of the total 794 starch assemblage.

Starch clusters/aggregates (Figure 6.2 I-J), some showing signs of gelatinization with enlargement and/or distortion of granules and loss of birefringence, were also noted. I, as well as others, have argued that these are representative of starchy foods being cooked in excess water and/or toasting and fermentation (low-temperature heat) (Chandler-Ezell et al. 2006:110; Zarrillo et al. 2008:5009-5010). Figure 6.2 shows examples of the unknown starch granule types

195

196

Figure 6.2. Unknown starch granule types from SALF. A-B, Type 1 Lenticular, shown in flat view (CR-51), scale 35 μm. C-D, Type 6, Spherical (CR-51). E-F, Type 11, Cupulate (STR-49), scale 20 μm. G-H, Triangular, as viewed looking “down” on the pyramidal shape (STR-49), scale 25 μm. I-J, Example of a starch cluster from STR-52; note that there are at least three different starch granules aggregated together with partial loss of the extinction crosses, but still birefringence (as seen in J), scale is 30 μm

196

Identified taxa include starches consistent with Capsicum spp., Dioscorea spp., Fabaceae family, Manihot esculenta, Theobroma spp., and Zea mays. Diagnostic domesticated chili

pepper (Capsicum spp.) starch granules are produced in the pericarp of the fruit, are

simple granules that are flattened-lenticular in shape, have a distinctive central depression

when viewed “flat” (similar in appearance to a red blood cell), a central longitudinal line

when viewed on its side, and are much larger (13-45 μm) than wild Capsicum starches

(maximum size 6.3 μm) (Korstanje and Babot 2007:70; Perry et al. 2007:986, Supporting

Information). Although these starch granules are flattened lenticular in shape, they are distinct from the Type 1 lenticular starch granules in that they are not as flattened and have a much sharper extinction cross with arms that do not flare towards the edges of the granule, are more elongate than spherical in flat view, and none of the Capsicum starch

granules identified had lamellae (compare to Figure 6.2 A-B). In addition, the arms of the

extinction cross follow the depression at the hilum before separating (Figure 6.3, B and

D, also see Figure 6.21 A-B). All of the Capsicum spp. starch granules observed in the

SALF artifact residues (Figure 6.3 A-B) were consistent with domesticated varieties of

chili peppers (Figure 6.3 C-D), based on size and characteristics, but a species-level

identification was not made due to a lack of comparative samples for all species of

domesticated Capsicum. Although the total number of starch granules only represented

1% of the entire assemblage from SALF, their ubiquity was ~45% (Table 6.1).

197

Figure 6.3. Example of Capsicum spp. starch granule from STR-52 (A-B). Note the depression in the centre of the granule (lighter central area of granule in A). C-D, comparative starch granule of Capsicum frutescens. Scale bars are 15 μm.

Starch granules identified as Dioscorea spp. were recovered from ~36% (Table 6.2,

Figure 6.4) of the artifacts yielding positive results from SALF. Dioscorea trifida

(domesticated yam, ñame, yampi, cush-cush yam) starch granules are described as simple ovoid/ellipsoidal/elliptical granules with a distinctive truncated distal end, sometimes described as “mitriform”, visible lamellae, a closed or slightly open hilum with two faint

198

fissures/lines radiating from the hilum towards the distal end (partially outlining a short cuneiform depression that extends from the hilum to the distal end), and range in size from 20-84 μm (Bou et al. 2006:378-379; Chu and Figueiredo-Ribeiro 1991:473; Perry

2001:129-130; Piperno 2006b:62; Piperno and Holst 1998:775; Piperno et al. 2000:895-

896; Radley 1953:332; Young 1953:plates 33,34). While all of the starch granules recovered from SALF artifacts were consistent with descriptions and comparative samples of Dioscorea trifida, the small number of starch grains recovered (n=6) and the high number of native Dioscorea species (33, with four endemic) present in Ecuador

(Jørgensen and León-Yánez 1999:438-439) that I did not have comparative samples of, meant that I only felt confident with a genus-level identification.

Figure 6.4. Example of Dioscorea spp. starch granule from SALF. A-B, from CR-57, scale 30 μm. Comparative Dioscorea spp. starch granule (C), scale 35 μm. Note the truncated distal end, presence of lamellae and two short faint fissures/lines radiating from the hilum towards the distal end in both the starch granule recovered from the CR-57 residues (A) and the comparative sample (C).

Starch granules consistent with the Fabaceae (legume/bean) family were recovered

from two artifacts from SALF. Although the starches recovered were consistent with

Phaseolus spp. (common or lima bean) starches in size and other characteristics (Figure

199

6.5 A-B), a family-level identification (cf. Phaseolus spp.) was made as only a total of three starch granules were recovered – in Ecuador the Fabaceae family is represented by

77 genera and 299 species (32 endemic) (Jørgensen and León-Yánez 1999:468). The diagnostic features of Phaseolus starch include: a simple grain, oval or reniform in shape; a long shallow indentation or ragged longitudinal fissure visible when the grain is rotated onto its side; small fissures may sometimes radiate from the indentation/fissure; the arms of the extinction cross separate when the polarizing filter is rotated; lamellae that follow the outline of the granule; size range is 11 μm - 50 μm (Korstanje and Babot 2007:65;

Perry 2001:135; Radley 1953:331; Reichert 1913:387). Phaseolus spp. starch granules

are quite distinct from starches of other Fabaceae genera found in Ecuador, such as

Canavalia spp. seed starch (Zarrillo et al. 2008), Arachis hypogaea (peanut) seed starch

(which is not abundant, nor diagnostic), Geoffroea decorticans fruit pulp starch, Lupinus

mutabilis (tarwi/chocho) seed starch, and Pachyrhizus ahipa root starch (Korstanje and

Babot 2007:54-72). The seed starch granules of Anadenathera colubrina, Acacia visco,

and Prosopis spp. are also quite different in size and morphology from Phaseolus spp.

starch (Duncan et al. 2009:13204; Korstanje and Babot 2007:54-72); these species are

classified as Fabaceae by Korstanje and Babot (2007:54-72), but as Mimosaceae by

Jørgensen and León-Yánez (1999:591-600). Therefore, the published descriptions of

starch granules for other Fabaceae family genera are dissimilar from Phaseolus spp.

making genus-level identification more secure, but the few starch granules recovered

merits caution.

200

Figure 6.5. Example of Fabaceae (cf. Phaseolus spp.) starch granule from SALF. A- B is a starch granule from STR-47 residues, scale 40 μm. C-D is a comparative starch granule of Phaseolus vulgaris, scale 30 μm. Note the same irregular elongate shape, longitudinal ragged fissure, lamellae, and diffuse extinction cross with multiple arms present in both the archaeological and comparative starch granules.

Starch granules consistent with Ipomoea spp. (sweet potato, boniato) were recovered from 54.5% of the artifacts from SALF. With a total assemblage of 34 starch granules

(4.3% relative frequency), a secure identification could be made. Sweet potato (Ipomoea

201

batatas) diagnostic starch grains are similar to those of manioc (Manihot esculenta).

Sweet potato starch occurs as single granules, aggregates of granules, and compound

granules but the distinctive hemispherical/domed shapes have clearly marked pressure

facets that rarely number more than three (Reichert 1913:884). They have centric or

eccentric hila, may or may not have lamellae, and range in size from 3 to 34 μm. The

diagnostic sweet potato hemispherical starch grains do not have the distinctive “inflated”

bell-shaped appearance of manioc starch (in Ipomoea starch granules the broadest part of the granule is the faceted end), nor is the hilum marked with the stellate fissures that manioc starch granules possess (Korstanje and Babot 2007:57; Perry 2001:131-132,

2002; Piperno and Holst 2000:897; Radley 1953:332; Reichert 1913:884). Although the

Ipomoea starch granules (Figure 6.6 A-B) that were recovered from the SALF artifacts were consistent with Ipomoea batatas comparative samples (Figure 6.6 C-D), there are

48 (3 endemic) species of Ipomoea in Ecuador, several of which are reported as being cultivated, including I. batatas, I. purpurea, and I. quamoclit (Jørgensen and León-Yánez

1999:409-411). No other genera in the Convovulaceae family are reported as being cultivated (Jørgensen and León-Yánez 1999:408-412; Rios et al. 2007:182-183).

Therefore, in the absence of more comparative samples (or published descriptions), a genus-level identification was assigned to the SALF artifacts’ residue starch granules.

202

Figure 6.6 Examples of Ipomoea spp. starch granules from SALF. A is from STR-47 with the surface somewhat degraded and/or with adhering organics, scale 20 μm. B is from STR-49, scale 20 μm. Note the two clearly demarcated flat edges (facets) and open hila with slight fissuring present in both A and B starch granules, as compared to C, Ipomoea batatas comparative starch granule, scale 25 μm. Also note that the broadest portion of the granule is from facet edge to facet edge. Compare to Figure 6.7, Manihot esculenta.

Fifteen starch granules (1.9% of the starch assemblage) were identified as Manihot esculenta and they were identified to the species level owing to the distinctiveness of the starch granules and the fact that only three species are known from Ecuador. M. brachyloba and M. leptophylla have not been mentioned as cultivated, although the leaves, but not tuber, of M. leptophylla are eaten, as noted by Rios et al. (2007:188).

Further, both grow at an elevation ~500 m below the location of SALF, while Manihot esculenta (manioc, yuca, cassava), which grows between 0 and 2000 masl (Jørgensen and

León-Yánez 1999:465), is the most likely source of the starch granules. Although the

Euphorbiaceae family has 50 genera and 243 species (45 endemic) in Ecuador, the vast majority are either not utilized, or other plant parts are used as poisons, in construction, as a source of latex, or for purposes other than consumption (Jørgensen and León-Yánez

1999:455-468; Rios et al. 2007:130-131, 132, 152, 157, 162, 170, 175, 179, 184, 206- 203

207, 217, 219). Six species are reported to be used for food, although no USOs are reported as being consumed. Of the species that are used as food, one is introduced

(Jatropha curcas) and the seeds are consumed (Jørgensen and León-Yánez 1999:464;

Rios et al. 2007:184). Three native species are also reported as having seeds used for food, with two found in Amazonia (Caryodendron orinocense and Plukentia volubis) and one in the Sierra (Croton alnifolius) (Rios et al. 2007:149, 162, 210). Two other

Euphorbiaceae family species are reported to have fruits used for food, Alchornea glandulosa and Tetrochidium macrophyllum, with both found in Amazonia (Rios et al.

2007:131, 228). Therefore, all other genera in the Euphorbiaceae family can be eliminated as possible sources of the tuber starch granules from SALF.

The diagnostic features of manioc starch include: compound grains (but often found isolated), bell-shaped with one to five basal facets and an “inflated” proximal end (the faceted end is not the broadest part of the granule); hilum is centric with diagnostic stellate fissures radiating from the hilum; no lamellae; size range is 5 μm – 35 μm (Moss

1976; Perry 2001:132-133, 2002; Piperno 2006b:51-58; Piperno et al. 2000:897; Radley

1953:325-326; Reichert 1913:876; Young 1953:plates 11-14). Despite the small number

of manioc starch granules recovered (n=15), their ubiquity was ~64% (Table 6.1).

204

Figure 6.7. Example of typical Manihot esculenta (manioc) starch granule from SALF residues. A, from CR-51, note the distinctive stellate fissure at the hilum, facets, and that the broadest portion of the granule is not the faceted end, making it bell-shaped. B, Manihot esculenta comparative starch. Scales are 20 μm. Compare to Figure 6.6.

Maranta spp. (arrowroot) starch granules were identified in ~64% of the artifacts that tested positive for starch from SALF and represent ca. 3% of the total starch granules observed (n=22). Arrowroot (Maranta arundinacea) starch grains are described as being usually oval or ovoid, range in size from 10 to 54 µm, with fine lamellae being sometimes stronger and complete around the hilum (which is often open and visible as a white dot, or sometimes marked by a single transverse “V” or “Y” shaped fissure) (Moss

1976:19; Perry 2001:131; Piperno and Holst 1998:775; Reichert 1913:813-815). While

Maranta spp. starch granules are similar to those of Dioscorea spp., the latter have a very distinct blunted/truncated distal end allowing for discrimination between the two genera

(compare Figures 6.4 and 6.8); this highlights the importance of rotating starch granules to view the granule from all aspects when making identifications.

205

Comparison of the SALF Maranta starches to the modern comparative samples and to published photomicrographs of Maranta arundinacea demonstrate they were compatible (Moss 1976:19; Reichert 1913:plate 88; Seidemann 1966:13). However, there are four species of Maranta in Ecuador in addition M. arundinacea (which is grown between 0-1000 masl). M. amazonica, although not noted as being cultivated, does grow at similar elevations (0-1000 masl), and the leaves, but not the USO, are reported to be used for medicine (Rios et al. 2007:188). M. gibba (0-500 masl) and M. ruiziana (0-500 masl) are found at lower elevations (but M. ruiziana is noted as being cultivated and used as food on the coast) (Jørgensen and León-Yánez 1999:558-559; Rios et al. 2007:188).

Finally, the only other species in the Marantaceae family that are reported as being used for food are Calathea spp. (llerén) (Rios et al. 2007:147, 297). Calathea starch granules are described as similar to Maranta arundinacea (and other USO starches), in that they are elongated and have eccentric hila and their size range (22 to 42 μm) is encompassed by the size range for Maranta arundinacea (Perry 2001:128-129).

However, distinctive Calathea starch granules have highly eccentric hila, have a more

“squared off” shape than the distinctive oval/ovoid M. arundinacea granules, and fissures at the hilum are very rare – highly elongate, rod-like shapes have also been reported

(Dickau et al. 2007:3655; Perry 2001:128-129). None of the starch granules forms identified as Maranta spp., or any of the other granules, from SALF, are consistent with

Calathea spp. and consequently it is eliminated as a possible source of starch in the

SALF assemblage. As a result, while it is likely that the SALF starches are from M.

206

arundinacea (domesticated arrowroot), they are only identified to the genus level because of the possibility that they may come from wild species of Maranta, as discussed above.

Figure 6.8. Example of Maranta spp. starch granules from SALF. A is from CR-63, scale 40 μm. B is from CR-57, measured size 13.2, scale 15 μm. C is a comparative starch granule of Maranta arundinacea, scale 25 μm. Note the ovoid shapes, open hila (A and B marked by fissures), presence of faint lamellae, and absence of truncated distal ends (compare to Figure 6.4).

Theobroma cacao (cacao) seed starch samples were prepared and characterized specifically for this analysis. During scanning of SALF slides distinctive small elliptical starch granules were noted, and because T. cacao is thought to have been domesticated in the region where SALF is located (tropical lowlands of NE Peru/SE Ecuador)

(Motamayor et al. 2002), a suggestion was made by Michael Blake of the University of

British Columbia to investigate the possibility that T. cacao starch was present in the

SALF artifact residues. There are few previous studies characterizing T. cacao seed starch (I could only locate one), and none to my knowledge that describe seed starches of the wild species of Theobroma. Although not abundant, starch composes 4.5-7% (avg.

5.3%) of cocoa seeds (Schmieder and Keeney 1980:557). While Schmieder and Keeney

(1980:556, and Figure 1) highlight the compound form of cacao starch, their Figure 1

207

also shows elliptical and spherical forms; they also found that granules range in size from

2 to 12.5 μm. I prepared comparative starch slides from samples of dried (and not roasted) T. cacao seeds that TropiLab Inc. labelled as originating from Suriname. I did not prepare comparative samples from commercial cocoa as I could not locate products

that were not subjected to dutching (alkali processing) (Snyder et al. 2009:615), which

may alter the morphology of the starch granules. I measured and characterized 100 starch

granules and three specific forms were noted:

1. Spherical or sub-spherical (somewhat polygonal) with closed or slightly open centric

or slightly eccentric hila that range in size from 3.8 to 6.3 μm and comprised 30% of the

starches;

2. Small compound granules composed of two or three truncated-spherical granula, with

closed or slightly open centric or slightly eccentric hila that range in size (total size of

compound granules) from 3.8 to 10 μm and comprised 48% of the starches; and

3. Ovoid (teardrop-shaped) starches with eccentric hila located at the broad end of the

granule (Figure 6.9 C-D), distal end often pointed or very narrow in comparison to

proximal end, closed or slightly open hila sometimes marked with one or more short

fissures radiating laterally and/or towards the distal end, with a sharp extinction cross

where the arms may bend or curve, a size range from 6.3 to 12.5 μm, and which

comprised 22% of the granules measured and characterized.

I consider the ovoid/teardrop shape to be diagnostic for Theobroma as I have not

noted this form in other comparative starch samples. Jørgensen and León-Yánez

(1999:921) list six species of Theobroma in Ecuador; all are found in the eastern

208

lowlands with the exception of T. gileri, which is native to the coast. T. bicolor and T. subincanum are found from 0 – 1000 masl (and are thus possible sources, as is T. cacao), while T. glaucum and T. grandiflorum are found at 0 – 500 masl, with the latter noted as being cultivated, as is T. bicolor. Because of a lack of other T. cacao comparative samples, and especially a lack of wild Theobroma samples, I have limited the identification to the genus-level. Furthermore, these identifications must remain tentative as comparative samples of Herrania spp. seeds were not available for characterization of their starch granules. Of the 10 genera and 50 species (9 endemic) in the Sterculiaceae1

family found in Ecuador, Herrania spp. (cacao de monte/monkey cacao) is the only

species reported to be consumed other than Theobroma, and the pulp and seeds of the

Herrania pods are consumed much like Theobroma are (Rios et al. 2007:146, 178, 228-

229; Jørgensen and León-Yánez 1999:918-922). Therefore, the identification of

Theobroma spp. starch must remain tentative until further samples of Theobroma and

Herrania species seeds are obtained and characterized.

That said, starch granules identified as Theobroma spp. comprised ~3.1% of the

SALF starch assemblage and were recovered from ~64% of the artifacts (Table 6.1).

Figure 6.9 shows an example of a Theobroma spp. starch granule from SALF as

compared to the T. cacao diagnostic starch form I identify above (ovoid/teardrop shape).

209

Figure 6.9. Example of Theobroma spp. starch granule from SALF. A-B is from CR- 57, C-D is the diagnostic starch granule form identified for Theobroma cacao. The tapered distal end of C is slightly out of view due to the position of the granule. Scales are 10 μm.

The diagnostic features of Zea mays maize include: simple (single) grain in a blocky polygon/polyhedral shape (due to pressure facets) with slightly smoothed edges in hard endosperm varieties (flint/pop) and spherical forms in floury endosperm varieties; hemispherical and “vase” shapes are also possible; a central or slightly eccentric hilum, often open and with a single fissure or fissures (commonly three in the shape of a “Y”);

210

no lamellae; a rough, cratered or grooved surface, especially in hard endosperm varieties; the arms of the extinction cross seem to bend and follow the edges of the polygon in three dimensions; size range 4 μm - 25 μm (Eckhoff and Watson 2009:378-379; Korstanje and

Babot 2007:63; Pearsall et al. 2004; Perry 2001:136; Piperno and Holst 1998:775;

Reichert 1913:344-353; Young 1953:plate 46). Comparative examples of Zea mays are shown in Figure 6.10.

For this analysis I employed an extremely conservative approach to identifying maize starch granules from SALF and all of the other sites. The reason for this is that although a comparative sample of Cenchrus brownii (that produces starch granules in its seeds that are similar to, but can be distinguished from, Zea mays starch) (Zarrillo et al. 2008

Supporting Information:1-2) was available for comparison, there are five other species of

Cenchrus in Ecuador that range from 0-3000 masl, with one species (C. myosuroides) found in coastal, Andean and Amazonian environments from 0-3000 m (Jørgensen and

León-Yánez 1999:813). No species of Cenchrus, however, are reported to be consumed

(Rios et al. 2007). Still, only starch granules that possessed the characteristics noted above, were >10 μm in size, and that had clear hilum fissure(s) were identified as Zea mays (see Figure 6.11 A-B, C, D, and F, and compare to Figure 6.10). Starch granules that met other morphological characteristics but were between 8 and 9 μm and that possessed clear fissures at the hilum were identified as cf. Zea mays. Starch granules >8

μm that met other morphological characteristics but did not have clear hilum fissure(s), regardless of the presence of an open hilum, were also identified as cf. Zea mays (see

Figure 6.11 E for an example, and compare to Figure 6.10). Based on these criteria, Zea

211

mays starch granules were recovered from ~90% of the SALF artifacts that tested

positive for starch. Zea mays starch conservatively represents 42.9% (including Zea mays

and Zea mays vase-shaped categories), the highest of all types, in the starch assemblage,

and may represent 57.8% of all starches recovered if the cf. Zea mays granules are

included (see Table 6.1).

Figure 6.10. Examples of comparative Zea mays starch granules. A-B Hard endosperm Zea mays starch granule, note the polygonal shape with pressure facets, distinctive “Y” fissures at the hilum in image A (scale 25 μm) and the way the arms of the extinction cross in image B follow the contours of the granule. Image C (scale 20 μm) is another hard endosperm maize starch granule, but the hilum is open and distinct fissures are not seen – this granule would be identified as cf. Zea mays if seen in the archaeological residue samples. Images D (scale 20 μm) and E (scale 12 μm) are both soft endosperm Zea mays starch granules with a more spherical form. Note though that the granule in image D has clear fissures, whereas E has an open hilum – D would be identified as Zea mays, and E as cf. Zea mays. Image F is an example of a “vase-shaped” Zea mays starch granule with a fissure at the hilum, scale 15 μm.

212

Figure 6.11. Examples of Zea mays starch granules from SALF artifact residues. A- B is the same starch granule from CR-67. Note the central fissures, polygonal shape, and the way the arms of the extinction cross follow the contours of the granule in image B, scale 25 μm. Image C is a Zea mays starch granule from CR-53, scale 20 μm. Images D-E are two starch granules from CR-63 (both scales are 15 μm). D is identified as Zea mays as the hilum fissures are clearly seen, whereas E is identified as cf. Zea mays as the fissures are not prominent, although all other characteristics are consistent with maize. F is an example of vase-shaped Zea mays from CR-50, scale 20 μm. Compare to Figure 6.10.

As a final point, an accelerator mass spectrometry (AMS) radiocarbon assay was obtained on a charred residue sample CR-63. The charred residues from this sherd was selected for dating for two reasons: 1) it was one of the few samples that both, a) tested positive for starch, and b) had residue remaining for analysis; and 2) unlike the other

SALF samples that came from well-dated contexts, the CR-63 sherd was recovered from a midden. As the site has more than one component (Valdez 2008:879-880, Table 43.1), I 213

wanted to ensure that the sample dated to the Early Formative Period. The results of this assay (Beta-312078) are: 4450 ± 30 BP (measured radiocarbon age), 13C/12C (δ13C) ratio

-14.4 ‰, Cal BC 3380 to 3350 (2 sigma, 95% probability), which falls well within the

Early Formative Period (4400-1450 Cal BC) (Hill 1975; Marcos 2003; Meggers 1966:24-

28; Raymond 2003a; Zeidler 2003:519, 2008:460); this result will be discussed later with respect to Table 6.8, section 6.9 where all the C14 dates are presented. The reader will

13 note, moreover, that the δ C shows a strong signature (-14.4 ‰) for C4 plants; 16 of the

35 starch granules (~46%) observed in the CR-63 residue slide were identified as maize

(Table 6.1).

6.3 Chaullabamba

The site of Chaullabamba (ca. 2300 masl), as noted in Chapter 3, is situated along the

Tomebamba River in the southern highlands of Ecuador in what is now a subdivision of

Cuenca, the capital city of Azuay province (Grieder et al. 2009:3). Radiocarbon assays situate the earliest occupation of the site to the terminal Early Formative to Middle

Formative (2300 -1700 Cal BC) (Bruhns et al. 1990:224; Grieder et al. 2002, 2009:8, 21-

22; Stahl 2005:317) (also see Chapter 4, section 4.3). Two AMS radiocarbon dates obtained from ceramic charred residues analyzed for starch (Tables 6.2 and 6.8) date to

the Middle (1490-1320 Cal BC) to Late Formative (1290-1060 Cal BC) (Zeidler

2008:460) (Table 6.2, as well as Table 6.9, section 6.9). As shown in Table 5.2 (Chapter

5), 11 ceramic charred residue samples were processed for starch granule recovery and

214

positive results were achieved from 7 (~64%). A total of 121 starch granules were observed on the sample microscope slides (Figure 6.2).

Table 6.2. Chaullabamba Starch Analysis Results.

spp.) ed p n /gel l sha m ca a ular Phaseolus se- know r d c eri s e . spp. va st lenti sph e (cf. pp ay /un r sample , s m ed 6, g clu pe UC Sample ea a h Z m . a tarc Number Sample Type 1 Type FabaceaOxalis Solanumcf Zea maysZea maysD StarchS clusterTotal Ceramic charred CR-79 residue 1 20 1 1 23 1 47 Ceramic charred CR-801 residue 23132192 23 Ceramic charred CR-81 residue 1 1 4 6 Ceramic charred CR-85 residue 3 3 3 1 1 1 12 Ceramic charred CR-86 residue 1 4 3 4 12 Ceramic charred CR-87 residue 3 1 1 5 Ceramic charred CR-892 residue 33 1 61 216 Totals 2 9 10 27 6 10 2 48 4 1 2 121 Ubiquity (%) 28.6 57.1 42.9 57.1 28.6 100 28.6 100 42.9 RF (%) 1.6 7.4 8.3 22.3 4.9 8.3 1.6 39.7 3.3 0.8 1.6 1AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312071, 2930 ± 30 BP (measured), Cal BC 1290 to 1060 (95%) , 13C/12C -23.4 ‰ 2AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312072, 3090 ± 30 BP (measured), Cal BC 1490 to 1320 (95%) , 13C/12C -22.0 ‰

215

Type 1 lenticular and Type 6 spherical granules encompass 1.6% and 7.4%,

respectively, of the total starch granules recovered, while four (3.3%) starch granules

were too damaged or obscured to be identified. The total number of discrete types

identified is eight, and identified taxa include Fabaceae, Oxalis spp., Solanum spp., and

Zea mays. Fabaceae (cf. Phaseolus spp.) starch granules represent 8.3% (n=10) of the

Chaullabamba starch assemblage and were found on ca. 43% of the ceramic charred

residue samples that tested positive for starch. Again, for the reasons stated earlier,

although all of the starches from the archaeological samples were consistent with

Phaseolus spp. starch granules, they are conservatively identified only as Fabaceae (cf.

Phaseolus spp.), although the assemblage of granules is larger for Chaullabamba than for

SALF. Figure 6.12 shows examples of Fabaceae starch granules from the Chaullabamba

residues.

Starch granules consistent with Oxalis spp. (oca) were observed in the residues of 4/7

(~57%) of the artifacts that tested positive for starch from Chaullabamba (Table 6.2), and represent 22.3% of the total starch assemblage (second only to Zea mays). It is important here to detail the characteristics that distinguish Oxalis tuberosa starch granules from those of Ullucus tuberosus (melloco/ulluco) and Solanum tuberosa (potato). These are

described below and shown in Figure 6.13.

216

Figure 6.12. Example of Fabaceae (cf. Phaseolus spp.) starch from Chaullabamba. A-B, from CR-89 and C-D is from CR-80, scales 40 μm. Compare to Figure 6.5 C-D.

Oxalis tuberosa (oca) starch granules (Figure 6.13 A-D) can be compound (2-5 granules of unequal size), but the diagnostic shapes are described as “irregular” with asymmetrical forms – oval, elliptical, pear-shaped, and prismatic – sometimes with a truncated-rounded distal end, range in size from 22-60 μm, with distinct highly-eccentric hila as an open circle or line, distinct lamellae, and, importantly, lines radiating from the hilum to the border of the grain (sometimes a centric fissure), fibrous surface appearance,

and extinction cross arms that are irregular, broken and/or intersect at more than one

point (Cortella and Pochettino 1995:459; Hernández-Lauzardo et al. 2004:359; Korstanje

and Babot 2007:54; Torres et al. 2011:384).

217

Oxalis tuberosa (oca) starch granules can be distinguished from Ullucus tuberosus

(melloco/ulluco) starch in that melloco starch granules are smaller (10-30 μm), although there is some size overlap with Oxalis and I have noted larger sizes. Ullucus tuberosa starch granules also have bowl- and bell-shaped granules (in addition to irregular granules similar to oca) and sometimes have one or more rounded projections. Similar to oca, melloco starch granules have eccentric hila with a circle or small fissure, but the lamellae are more regularly spaced than those of Oxalis, and there is a more diffuse extinction cross with arms that meet at one point (Cortella and Pochettino 1994:459;

Korstanje and Babot 2007:56). No starch granules were observed in the starch granule assemblage from Chaullabamba (or any of the other sites) that met these collective criteria, but see a comparative sample Figure 6.13 F.

Another high-elevation USO that produces ovoid or elliptical starch granules is

Solanum tuberosum (potato). As shown in Figure 6.13 G-H, the conspicuous starch granule forms are ovoid, oval and ellipsoidal, they range in size from 15-120 μm

(although mean size is ~40 μm), the surface of the granules are smooth, the hilum is usually a distinct, small, round spot at the broad end of the granule and may occasionally be marked by a small fissure, the lamellae are fine and tend to follow the outline of the granule, and the extinction cross arms are distinct and are usually broader at the edges of the granule (Korstanje and Babot 2007:57; Reichert 1913:882; Weber et al. 2002:190).

218

219

Figure 6.13. Comparative starch granules of Oxalis tuberosa, Ullucus tuberosus and Solanum tuberosum. A-D, two different starch granules typical of Oxalis tuberosa, scales are 50 μm. Note the irregular shapes, fibrous appearance of the surfaces, the centric fissures and lines, irregular lamellae, and arms of the extinction cross that are irregular in shape and width. E-F, Ullucus tuberosus comparative starch granule, scale 40 μm. Note the bell-shape with projections at the distal end of the granule, the hilum is at the narrow end of the granule (as it is for the Oxalis in C), more regular-shaped lamellae, and although the arms of the extinction cross are somewhat diffuse, they are more even in width than Oxalis. G-H, Solanum tuberosum comparative starch granule, scale 45 μm. Note the ovoid shape with the eccentric hilum at the broad end, smooth, non-fibrous granule surface, distinct and more regularly-spaced lamellae that follow the outline of the granule, and the evenly prominent extinction cross arms that broaden towards the edges of the granule.

219

Despite that all of the starch granules observed in the Chaullabamba residues were consistent in all characteristics with Oxalis tuberosa, the starches were only identified to the genus-level as there are 3 genera and 53 species (5 endemic) in the Oxalidaceae family in Ecuador, although only Oxalis spp. are reported to be consumed (Rios et al.

2007:200-201; Jørgensen and León-Yánez 1999:775-778). Of the other Oxalidaceae taxa, there is only one species of Averrhoa (A. carambola), which is an introduced tree limited to the coast (0-500 masl) (Jørgensen and León-Yánez 1999:775), and so it can be eliminated as a possible source of the starches from Chaullabamba. There are six species

(one endemic) of Biophytum reported (all are herbs): 1) B. boussingaultii apparently ranges from 1000-1500 masl, which is almost 700 m lower than Chaullabamba (ca. 2300 masl), and, more importantly, its presence in Ecuador is in question: 2) B. columbianum is native to the coast from 500-1000 masl, and so it is found 1300 m lower than

Chaullabamba’s elevation; 3) B. dendroides, coastal and Amazonian (0-500 masl), and is reported to be used as a poison (Rios et al. 2007:141); 4) B. heinrichsae, Amazonian

(500-1000 masl); 5) B. somnians, Amazonian (0-500 masl); and 6) B. soukupii,

Amazonian (0-500 masl), which is only tentatively identified (Jørgensen and León-Yánez

1999:775-776). While I cannot say with absolute certainty that none of the Biophytum species contributed to the Chaullabamba starch assemblage, it seems unlikely as they are found well outside the elevation of Chaullabamba outside of the Inter-Andean Valley, and, moreover, none are reported to be cultivated or consumed (Rios et al. 2007) (I could not find any description of starch granules, however). Therefore, with the other two

Oxalidaceae genera almost certainly eliminated as possibilities, the Oxalis species are the

220

most likely source. Because there are 43 species (four endemic) of Oxalis in Ecuador

(excluding species only found in Galapagos and introduced species), and various species range from 0-4500 masl (even though only O. tuberosa is reported to be cultivated), the

Chaullabamba starches are identified only to genus. Figure 6.14 shows examples of starch granules from the Chaullabamba ceramic charred residues identified as Oxalis spp.

The Solanaceae family is extremely diverse and contributes several important species of food plants, including potatoes, tomatoes, chili peppers, naranjilla, and tamarillos, as well as others used for drugs and medicine, e.g., tobacco, datura, tree datura, Brunfelsia

(used in ayahuasca/ yajé), and Latua (an hallucinogenic used on the coast of Chile)

(Knapp 2002, 2005). There are 35 genera and 351 species in the Solanaceae family in

Ecuador (Jørgensen and León-Yánez 1999:900), yet despite this diversity, all genera other than Solanum can be eliminated as possible sources of starch similar to Solanum

tuberosum because they are limited to elevation <500 masl or are endemic to the

Galapagos, are introduced, are not reported as being used or cultivated, are cultivated as an ornamental flower, shrub, or tree, only their seeds, leaves and/or bark are used in folk- medicine medicine or as an hallucinogenic, or their fruits (but not USO’s) are consumed

(Coe 1994; Frisch and Frisch 2004:350; Gallegos 1988; Jørgensen and León-Yánez

1999:900-918; Knapp 2002, 2005; Rios et al. 2007, 2008; Schultes and Hoffman

1992:46).

221

Figure 6.14. Oxalis spp. starch granules from Chaullabamba. A-B, from CR-85, scale 40 μm. C-D, from CR-79, scale 35 μm. Note the irregular shapes, eccentric hila marked by small fissures, fibrous surfaces, irregular lamellae and patterns of the extinction crosses. Compare to Figure 6.13, especially A-D.

While all starch granules identified as Solanum spp. are consistent in all respects with those of Solanum tuberosum, they are identified only to genus as Solanum species other than S. tuberosum are reported to be cultivated in Ecuador (Jørgensen and León-Yánez

1999:909-918). In particular, S. phureja cannot be eliminated as a possible source of the starch granules observed in the Chaullabamba residues (and residues from the other sites) as it is a potato cultivated between 3000-5000 masl (Jørgensen and León-Yánez

1999:916) and in modern times it has been crossed with S. tuberosum in order to investigate and develop disease-resistant potato hybrids (e.g., Maine et al. 1993). Note

222

though that the natural range of S. phureja is at a higher elevation than the site of

Chaullabamba (2300 masl). That said, 4.9% (n=6) of the starch granule assemblage from

Chaullabamba were identified as Solanum spp. from ca. 29% of the artifacts that tested positive for starch; examples are shown in Figure 6.15.

Figure 6.15. Example of Solanum spp. starch granules from Chaullabamba. A-B, from CR-80 residue, scale 35 μm. C-D, from CR-85 residue, scale 40 μm. Note the smooth granule surfaces, that the hila are distinct, small, round spots at the broad end of the granules, that the lamellae follow the outline of the granules, and that the extinction cross arms are prominent, continuous and broader at the edges of the granules. Compare to Figure 6.13 G-H.

223

Zea mays polygonal and vase-shaped starch granules represent 41.3% (n=50) of the

starch granule assemblage from Chaullabamba, and 49.6% if the cf. Zea mays granules

are included. Zea mays starch was recovered from 100% of the artifacts that tested

positive for starch from Chaullabamba. Figure 6.16 shows examples of Zea mays starches

from the Chaullabamba ceramic charred residues.

Figure 6.16. Examples of Zea mays starch granules from Chaullabamba. A, from CR-80 residue, scale 25 μm. B, from CR-79 residue, scale 25 μm. C, from CR-82 residue, scale 20 μm. D, from CR-85 residue, scale 20 μm. Compare to Figure 6.10.

224

6.4 Tajamar

The Tajamar site is located in the Cayambe valley, in the far northern suburbs of

Quito, at an elevation of 2520 masl (Domínguez et al. 2003:228; Jara and Santamaria

2009:55-56). An assemblage of 311 starch granules were recovered from seven artifacts, including charred ceramic residues and stone tools (Table 6.3). No previous radiocarbon

assays were available, but based on the ceramic assemblage, the site is thought to date to

the Integration Period (AD 500-1500) (Domínguez et al. 2003:228; Jara and Santamaria

2009:55-56). All of the artifacts that were sampled for residues produced positive starch

results and two AMS radiocarbon dates were obtained from the CR-67 and CR-68

charred ceramic residues (Table 6.3, Table 6.8 and see section 6.9), situating the Tajamar sherds that were sampled to the Middle (1430-830 Cal BC) to Late Formative (1300-300

Cal BC) (Zeidler 2008:460).

Type 6 spherical starch granules (n=43) were observed in 100% of the artifact

residues and represent 13.9% of the starch granule assemblage, as shown in Table 6.3).

Taxa identified by starch analysis include Fabaceae (2.6% of the starch assemblage,

Figure 6.17), Oxalis spp. (6.4%, Figure 6.18), Solanum spp. (2.2%, Figure 6.19), and Zea

mays (Figure 6.20). The Zea mays (49%, n=153) and cf. Zea mays (19%, n=59), if

combined, represent 68% of the entire assemblage, and were both found (100% ubiquity)

in all of the artifact residues from Tajamar.

225

Table 6.3. Tajamar Starch Analysis Results.

.) p sp lus seo ha rical P r e f. s h c . pp. y ste p s a sp sp m lu , eae ( m c UC Sample 6 c lis nu ea mays h l per sample e a Z c ol . tar Number Sample Typ Faba Oxa S cf Zea Damaged/unknownS Tota Ceramic charred CR-66 residue 1 7 12 5 25 Ceramic charred CR-671 residue 2 6 9 3 4 9 3 1 37 Ceramic charred CR-682 residue 2 2 6 3 2 7 2 1 25 Mano STR-873 sonicated 1 1 4 32 1 39 Metate STR-883 sonicated 31 4 21 28 4 2 90 Mano STR-893 sonicated 1 1 6 30 38 Metate STR-903 sonicated 51535257 Totals 43 8 20 7 59 153 16 5 311 Ubiquity (%) 100 28.6 57.1 42.9 100 100 RF (%) 13.8 2.6 6.4 2.2 19 49 5.1 1.6 1AMS C14 assay on split sample of interior charred residue tested for starch: Beta-315535, 2870 ± 30 BP (measured), Cal BC 1210 to 1010 (95%) , 13C/12C -22.7 ‰ 2AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312079, 2660 ± 30 BP (measured), Cal BC 920 to 810 (95%) , 13C/12C -20.5 ‰ 3ALL stone tools were washed with distilled water prior to sampling to test for incidental starch - all wash samples were negative for starch.

226

Figure 6.17. Examples of Fabaceae (cf. Phaseolus spp.) starch from Tajamar residues. A-B, from CR-67, scale 40 μm. C-D, from CR-68, scale 30 μm. Compare to Figure 6.5 C-D.

227

Figure 6.18. Examples of Oxalis spp. starch from Tajamar residues. A-B, from CR- 67, scale 40 μm. C-D, from CR-68, scale 50 μm. Compare to Figure 6.13, especially A-D.

228

Figure 6.19. Examples of Solanum spp. starch from Tajamar residues. A-B, from CR-68 residues. C-D from CR-67 residues. Scales are 25 μm. Compare to Figure 6.13 G-H.

229

Figure 6.20. Examples of Zea mays starch granules from Tajamar artifact residues. A, from CR-67 residues, scale 15 μm. B, from CR-68 residues, scale 20 μm. A-B from ceramic charred residues, AMS C14 dated, see Table 6.3. C, from STR-87 mano residue, scale 25 μm. D, from STR-88 metate residues, scale 20 μm. E, from STR-89 mano residue, scale 20 μm. F, from STR-90 metate residues, scale 15 μm. Compare to Figure 6.10.

6.5 La Vega

The La Vega site is located in the Catamayo Valley ca. 3 km southwest of the modern town of Catamayo, province of Loja, in the southern Ecuadorian highlands. Despite its highland location, the elevation of the site is 1234 masl (Guffroy 2004:37), only ca. 200 masl higher than the Santa Ana-La Florida site located to the southeast of La Vega in the upper Amazon. The interior charred residues from eight different sherds were sampled and processed for starch granule analysis, and all tested positive for starch, as shown in

230

Table 6.4. A split-sample of the charred residues from CR-72 was sent for AMS radiocarbon dating (Tables 6.4 and 6.8, and see section 6.9 below), and the results show that the sherd dates to the Middle (1430-830 Cal BC) to Late Formative (1300-300 Cal

BC) (Zeidler 2008:460). Note also that the δ13C shows a strong signature (-14.4 ‰) for

C4 plants. A total of 190 starches were observed; 10 (5.3%) could not be identified because they were either too damaged or were obscured by adhering residue.

Table 6.4. La Vega Starch Analysis Results.

spp.) olus n r ase ta w cal h n . P kno eri spp. f s n mple c sa lenticula s sph ae ( may cluster 6, ged/u ce ea may pe 1, e Z yp aba ea otal per UC Sample Number Ty T CapsicumF Manihotcf. escule Z Dama StarchT CR-71 24 8 3 6 5 13 1 60 CR-721 1 1 8 10 4 25 2 51 CR-73 2 2 3 5 12 CR-74 4 1 4 1 10 CR-75 2 1 1 2 2 8 CR-76 2 5 6 13 CR-77 6 6 2 14 CR-78 2 3 4 9 4 22 Totals 35 5 12 11 16 29 70 10 2 190 Ubiquity (%) 75.0 37.5 37.5 25.0 25.0 100 100 RF (%) 18.4 2.6 6.3 5.8 8.4 15.3 36.8 5.3 1.0 1AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312077, 2720 ± 30 BP (measured), Cal BC 1190 to 1000 (95%), 13C/12C -14.4 ‰

231

Figure 6.21. Examples of Capsicum spp. starch as compared to Type 1 lenticular starch granules from La Vega. A-B, Capsicum from CR-71, scale 25 μm. C, Capsicum from CR-73, scale 15 μm. D, Capsicum from CR-76, scale 15 μm. E-F, Type 1 lenticular starch from CR-71, scale 35 μm. Note that the Type 1 lenticular starch granule (E-F) has prominent lamellae, a more spherical shape, and a more diffuse extinction cross (although this is partially obscured by adhering residue in image F), as compared to images A-D, the Capsicum starch granules that have a more elliptical shape, linear depressions at the hila, and a sharper extinction cross (image B) – also note that the extinction cross arms meet at the center of the hilum for Type 1 lenticular (image F), while the arms of the extinction cross for Capsicum follow the linear depression at the hilum before separating (image B). Compare to Figure 6.2 A-B and Figure 6.3.

Type 1 lenticular granules represent 18.4% (n=35) of the assemblage and were

recovered from 75% of the samples. Five type 6 spherical granules were also observed, representing 2.6% of all starches. Identified taxa include Capsicum spp. (6.3%, n=12) observed in the residues of three of the samples (ca. 37% ubiquity), as shown in Figure

232

6.21. As discussed for SALF (see section 6.2), the lenticular Capsicum starch morphological characteristics are distinct from the Type 1 lenticular starches.

Eleven starch granules were identified to the Fabaceae family (cf. Phaseolus spp.) and represent 5.8% of the starches from La Vega; they were observed in 25% (2/8) of the artifact residues. Manihot esculenta starches were also observed in 25% of the artifact residues, and represent 8.4% (n=16) of the total starches recovered. Figure 6.22 shows examples of the Fabaceae and Manihot esculenta starches from La Vega.

Starch granules identified as cf. Zea mays and Zea mays represent 15.3% (n=29) and

36.8% (n=70), respectively, of the starch granules from La Vega. Both were recovered from all of the artifacts tested (100% ubiquity). If combined, their total relative frequency is 52.1%, over half of all starches observed in the La Vega residues.

233

Figure 6.22. Examples of Fabaceae (cf. Phaseolus spp.) and Manihot esculenta starches from La Vega. A-D, Fabaceae starch from CR-71 (A-B) and CR-72 (C-D), scales 30 μm. Figure 6.5 C-D. E-H, Manihot esculenta starch granules from CR-71 (E-F) and CR-72 (G-H), scales 20 μm. Compare to Figure 6.6.

234

Figure 6.23. Examples of Zea mays starch from La Vega. A, from CR-71, scale 20 μm. B, from CR-72, scale 20 μm. C, from CR-73, scale 30 μm. D, from CR-75, scale 20 μm. E, from CR-77, scale 20 μm. F, from CR-78, scale 15 μm. Compare to Figure 6.10.

6.6 Trapichillo

The Trapichillo site is also located in the Catamayo Valley ca. 4.5 km northwest of the La Vega site and 1 km north of the modern town of Catamayo, province of Loja, in the southern Ecuadorian highlands. The top of the hill, where the site is located, is 1350 masl and the hill is ca. 100 metres high (Guffroy 2004:31). Therefore, La Vega and

Trapichillo are in close proximity to each other and at about the same elevation – both are situated at a relatively low elevation for the highlands. Only two artifacts were analyzed from Trapichillo and both are interior charred residues from ceramic sherds. When

235

examining the collection from Trapichillo, only a few sherds with adhering interior charred residues were found assigned to the four-phase Catamayo ceramic series. Phase

A has been previously identified as dating to the terminal Early Formative Period (2000 –

1400 BC) and Phase B to the Middle to Late Formative Periods (1200 – 900 BC)

(Guffroy 2004:85-86, 91). Therefore, as I was especially interested in the earliest occupation of Trapichillo (as for the other sites), and owing to the dearth of sherds with residues that were assigned to one of the Catamayo Phases, only four samples were available – one for Phase A, one for Phase B, one for Phase C, and two (very small) samples for Phase D. The two residue samples (CR-69 and CR-70), representing the

Phase A and Phase B samples, respectively, were both split, with one part processed and analyzed for starch and the other part sent for AMS radiocarbon analysis. The results are shown in Table 6.5 (also see Table 6.9 and section 6.9 below).

Inexplicably, the CR-69 Phase A residue returned a date a minimum of 1500 calendar years younger than the CR-70 Phase B sherd. As the exterior of both of the sherds bore no diagnostic decoration (as is common for cooking pots in general) it may be that the sherds were incorrectly assigned during original excavation and/or cataloguing or I made an error in cataloguing (although see Figure 5.4, Chapter 5 showing the actual Phase B sherd being sampled with catalogue information on the sampling aluminum foil).

236

Table 6.5. Trapichillo Starch Analysis Results.

p.) sp s d

ar le cal Phaseolu e-shape p eri . s h (cf ys va ster lenticulsp e u , , ma ed/unknown cl UC Sample 1 6 ag Zea ch tal per sam o Number Sample Type Type FabaceaManihotcf. esculenta Zea maysZea maysDam Star T Ceramic charred CR-691 residue 9 10 3 18 2 42 Ceramic charred CR-702 residue 4 5 4103 1211251 Totals 4 5 13 20 6 1 39 3 2 93 Ubiquity (%) 50 50 100 100 100 50 100 RF (%) 4.3 5.4 14.0 21.5 6.5 1.0 41.9 3.2 2.1 1AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312070, 1100 ± 30 BP (measured), Cal AD 680 to 870 (95%), 13C/12C -15.9 ‰ 2AMS C14 assay on split sample of interior charred residue tested for starch: Beta-315533, 2870 ± 30 BP (measured), Cal BC 1190 to 980 (95%), 13C/12C -24.3 ‰

Irrespective of where the error lies, neither of the radiocarbon results I obtained date to the Early Formative Period (Table 6.5). The Phase A sherd (which also shows a δ13C signature for C4 plants) dates to the Integration Period (AD 500-1500), while the Phase B sherd date corresponds to the Middle (1430-830 Cal BC) to Late Formative (1300-300

Cal BC) (Masucci 2008:489; McEwan and Delgado-Espinoza 2008:505; Meggers

237

1966:24-28; Salazar 2008:263; Zeidler 2008:460). Section 6.9 details these results in relation to all of the AMS radiocarbon dates obtained.

Of the 93 starch granules observed in the Trapichillo ceramic charred residues, 4.3%

(n=4) are Type 1 lenticular, 5.4% (n=5) are Type 6 spherical, 13% (n=13) are assigned to

Fabaceae, 21.5% (n=20) are identified as Manihot esculenta, 6.5% (n=6) are cf. Zea mays, 42.8% (n=40) are Zea mays, and 5.3% (n=5) were either damaged or a starch aggregate (Table 6.5). Figure 6.24 shows examples of Fabaceae (cf. Phaseolus spp.) and

Manihot esculenta starch granules, while Figure 6.25 shows examples of the Zea mays assemblage – all are from the CR-70 residues.

238

Figure 6.24. Examples of Fabaceae (cf. Phaseolus spp.) and Manihot esculenta starch granules from Trapichillo, CR-70 residue. A-B and C-D, two Fabaceae starch granules. E, F, and G, three Manihot esculenta starches. Scales: (A) 30 μm, (B) 50 μm, (E) and (F) 15 μm, (G) 25 μm.

239

Figure 6.25. Examples of Zea mays starch granules from Trapichillo, CR-70 residue. Scales: A and C-F 20 μm, B is 15 μm. Compare to Figure 6.10.

6.7 Cerro Narrío

As discussed in Chapter 3, Cerro Narrío is located on the edge of the modern town of

Cañar, in Cañar province, in the (northernmost) southern highlands of Ecuador. The site is located in the Cañar Valley and encompasses a large hill ca. 3000-3100 masl (Collier and Murra 1943:35). Survey and excavations were conducted by J. Scott Raymond and

Florencio Delgado-Espinoza from 2007-2009; I assisted with the survey and excavations and selected and processed samples for botanical analysis (Table 5.2). Three contexts were selected for starch analysis: Units 3 (Levels 9-12) and 3A (Layers 3-4) represent the same context and four samples from each were analyzed; Unit 4, Level 3 (four samples

240

analyzed); and Unit 7 (Layer 2, Levels 14-16, four samples total). Only ceramic charred residues, which can be AMS radiocarbon dated, were analyzed. Of the sixteen samples analyzed, 11 (69%) produced positive results for starch (≥5 starch granules per sample).

In addition, sediment samples from the contexts that the sherds originated from were also tested for starch, and none produced positive results (0-3 starch granules per sample).

Although one previous radiocarbon date of unknown provenience for Cerro Narrío is reported as Early Formative, 2580-2200 Cal BC (95% probability) (Burleigh et al. 1977), the AMS radiocarbon dates from the ceramic charred residue samples CR-103, CR-110, and CR-113 consistently date to the Late Formative Period (900 to 550 Cal BC), as shown in Table 6.6. Two wood charcoal samples were sent for radiocarbon analysis in

2010. Unit 3A, Layer 4, Level 4 returned a date of Cal BC 780 to 410 (Cal BP 2740 to

2260, Beta-274050, 2γ), and Unit 7, Layer 2, Level 17 returned a date of Cal BC 810 to

670 (Cal BP 2760 to 2620, Beta-274052, 2γ). The AMS C14 date for the CR-109 ceramic

13 charred residue sample, which also shows a δ C signature for C4 plants, is rejected as it is not in line with other dates from the same context (Tables 6.6 and 6.8) – this will be further elaborated in section 6.9. Although the starch analysis results for the CR-109 residue are shown in Table 6.6, they were not used in calculating totals for different taxa, ubiquity, relative frequency, or the total number of starches recovered from all the samples.

241

Table 6.6. Cerro Narrío Starch Analysis Results.

pp.) s s olu hase e P . s nknown spp. ay u sampl spp s r 6, spherical pe UC Sample ea m may aged/ anum Z m abaceae (cf. l . ea tarch clusterotal Number Sample Type F Oxalis So cf Z Da S T Unit 4 Level 3 Ceramic charred CR-100 residue 5 27 4 3 39 Ceramic charred CR-1031 residue 7 4 7 6 12 1 37 Unit 3A S1/2 Layers 3 and 4 Ceramic charred CR-104 residue 3 1 5 4 7 4 1 25 Ceramic charred CR-106 residue 9 2 2 4 22 7 46 Ceramic charred CR-107 residue 6 5 11 6 12 8 2 50 Unit 3 Level 9, 11, 12 Ceramic charred CR-108 residue 4 15 15 6 23 3 4 70 Ceramic charred CR-1092 residue 12 13 1 1 25 2 54 Ceramic charred CR-1103 residue 10 8 9 17 1 3 48 Unit 7 Layer 2, Level 14, 15, 16 Ceramic charred CR-111 residue 2 7 6 5 27 4 1 52 Ceramic charred CR-1134 residue 4 8 9 24 3 2 50 Ceramic charred CR-114 residue 141 72112 37 Totals 5 45 43 54 61 192 36 18 454 Ubiquity (%) 20 80 80 70 100 100 RF (%) 1.1 9.9 9.5 11.9 13.4 42.2 7.9 4.0

1AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312074, 2540 ± 30 BP (measured), Cal BC 800 to 570 (95%) , 13C/12C -24.1 ‰ 2AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312073, 960 ± 30 BP (measured), Cal AD 780 to 970 (95%) , 13C/12C -12.9 ‰ 3AMS C14 assay on split sample of interior charred residue tested for starch: Beta-315534, 2530 ± 30 BP (measured), Cal BC 800 to 550 (95%) , 13C/12C -24.1 ‰ 242

In total, 454 starch granules were observed in the ceramic charred residues from all contexts (10, as the CR-109 results are excluded) at Cerro Narrío. Despite the large starch granule assemblage, only five different types were identified, with four identified to taxa, including Fabaceae (cf. Phaseolus spp.), Oxalis spp., Solanum spp., cf. Zea mays and Zea mays. Type 6 spherical granules represent 1.1% of the assemblage and were recovered from ca. 20% of the samples. Starch granules identified to the Fabaceae family were recovered from 8/10 contexts (80% ubiquity) and represent 9.9% (n=45) of all starches

(Figure 6.26).

Figure 6.26. Examples of Fabaceae family (cf. Phaseolus spp.) starches from Cerro Narrío. A-B, from CR-103 (Unit 4). C-D, from CR-110 (Unit 3). E-F, from CR-114 (Unit 7). Scales 30 μm. Compare to Figure 6.5 C-D.

243

Similar results to those of Fabaceae were realized for Oxalis spp. starch granules in that they represent 9.5% (n=43) of all starches and were recovered from 80% of the samples that tested positive for starch (Figure 6.27).

Figure 6.27. Examples of Oxalis spp. starches from Cerro Narrío. A-B, from CR-103 (Unit 4), note the multiple extinction cross arms in B that meet at more than one point, scale 30 μm. C-D, from CR-110 (Unit 3), note that this is a rare compound granule, scale 35 μm. E-F, from CR-111 (unit 7), scale 45 μm. Compare to Figure 6.13, especially A-D.

Solanum spp. starch granules were identified from ca. 70% of the samples and represent 11.9% (n=54) of the total starch assemblage. Figure 6.28 shows examples of potato starches from Unit 4, Unit 3, and Unit 7.

244

Figure 6.28. Examples of Solanum spp. starch granules from Cerro Narrío. A-B, from CR-103 (Unit 4), scale 25 μm. C-D, from CR-107 (Unit 3), scale 35 μm. E-F, from CR-113 (Unit 7), scale 20 μm. Compare to Figure 6.13 G-H.

Starch granules identified as cf. Zea mays and Zea mays were observed in all of the samples that tested positive for starch (100% ubiquity). They were also the most numerous of all starch types identified, with cf. Zea mays representing 13.4% (n=61) and

Zea mays 42.2% (n=42.2), and, if combined, represent more than half (55.6%) of the starches from Cerro Narrío. Figure 6.29 shows examples of Zea mays starches from Unit

4, Unit 3 and 3A, and Unit 7.

245

Figure 6.29. Examples of Zea mays starches from Cerro Narrío. A, from CR-103 (Unit 4). B, from CR-110 (Unit 3). C, from CR-107 (Unit 3A). D-F, from CR-111, CR-113, and CR-114 (Unit 7). All scales 20 μm, except A 15 μm. Compare to Figure 6.10.

6.8 La Chimba

The La Chimba site is located in the northern highlands of Ecuador, in northeastern

Pichincha province. At 3180 masl (Stahl and Athens 2001:161), it is at a similar elevation as Cerro Narrío. Based on stylistic analysis of ceramics and radiocarbon determinations, three periods have been defined for the site: Early La Chimba (690-440 Cal BC); Middle

La Chimba (440-44 Cal BC); and Late La Chimba (44 Cal BC – AD 250) (Stahl and

Athens 2001:165). Eight ceramic charred residue samples were processed and four (50%) returned positive results for starch. Sample CR-93 was split and sent for AMS C14

246

dating, and the result, 830 to 790 Cal BC (95%) (see Table 6.7), is consistent with that reported for the same context (TP-7, Level 28) (Stahl and Athens 2001:165), placing the

TP-7 samples (CR-90, CR-93, and CR-95) in the Late Formative Period. According to

Stahl and Athens (2001:165), TP-5 Level 13 (the CR-99 sample) also dates to Early La

Chimba (ca. 630 BC) and the Late Formative Period.

Table 6.7. La Chimba Starch Analysis Results.

own e n l p. p ys nk s uster samp ys d/u l r um ma UC Sample n a ma pe Ze mage . ea a otal Number Sample Type 6,Sola sphericalcf Z D Starch cT Ceramic charred CR-90 residue 6 4 29 5 4 48 Ceramic charred CR-931 residue 2 13 3 18 3 2 41 Ceramic charred CR-95 residue 4 17 1 2 24 Ceramic charred CR-99 residue 13 7 24 3 47 Totals 12 26 14 88 9 11 160 Ubiquity (%) 75 50 75 100 RF (%) 7.5 16.2 8.7 55 5.6 6.9 1AMS C14 assay on split sample of interior charred residue tested for starch: Beta-312076, 2590 ± 30 BP (measured), Cal BC 830 to 790 (95%) , 13C/12C -22.0 ‰

247

Only three different starch granule types were identified, the least number as compared to the other sites, and, interestingly, oca (Oxalis spp.) starch granules are absent, despite the elevation of the site. Type 6 spherical granules comprised 7.5% of the total assemblage (n=160) and were recovered from 75% of the ceramic charred residue samples, as shown in Table 6.7. Starch granules identified as potato (Solanum spp.)

(Figure 6.30) represent 16.2% of the assemblage and were recovered from two of the four positive samples (50% ubiquity).

Figure 6.30. Examples of Solanum spp. starch granules from La Chimba. A-B, from CR-93, scale 35 μm. C-D, from CR-99, scale 40 μm. Compare to Figure 6.13 G-H.

248

Zea mays starches represent 55% (n=88) of the starch granules assemblage from La

Chimba and are present on all (100% ubiquity) of the samples that were positive for starch. If the cf. Zea mays starch granules (n=14) are added, then the relative frequency of maize is 63.7%.

Figure 6.31. Examples of Zea mays starch granules from La Chimba. A, from CR- 90. B-C, from CR-93. D, from CR-95, E-F, from CR-99. Scales A, C, D and F are 20 μm, B and E are 15 μm.

6.9 Accelerator Mass Spectrometry Radiocarbon and 13C/12C Assays

Table 6.8 presents all of the accelerator mass spectrometry (AMS) radiocarbon and

13C/12C (δ13C) assays. The results are presented by site, from oldest to youngest based on

249

the 2 sigma (σ) 95% probability calibrated dates, with the oldest site Santa Ana-La

Florida (SALF) located on the eastern Andean slopes in the upper Amazon region.

Table 6.8. Accelerator Mass Spectrometry Radiocarbon and 13C/12C Assays on Ceramic Charred Residue Samples.

13C/12C Conventional Site Context Beta # Submitter #1 Measured Age (‰) Age 1 Sigma Calibration2 2 Sigma Calibration2

X4(17) ca. Cal BC 3490 to 3360 circa Cal BC 3500 to 3330 SALF midden 312078 SALF-X4(17)-CR63 4450 ± 30 BP -14.4 4620 ± 30 BP (Cal BP 5440 to 5310) (Cal BP 5450 to 5300)

Cal BC 1430 to 1410 circa Cal BC 1490 to 1320 Chaullabamba III-H-5 312072 CHB-IIIHL5-CR89 3090 ± 30 BP -22.0 3140 ± 30 BP (Cal BP 3380 to 3360) (Cal BP 3440 to 3280) Cut 1, ca. Cal BC 1260 to 1130 circa Cal BC 1290 to 1060 Chaullabamba Level 3 312071 CHB-C1L3-CR80 2930 ± 30 BP -23.4 2960 ± 30 BP (Cal BP 3210 to 3080) (Cal BP 3240 to 3000) Unit 151, ca. Cal BC 1190 to 1050 Cal BC 1210 to 1010 Tajamar Level D9 315535 TA-U151-CR67 2870 ± 30 BP -22.7 2910 ± 30 BP (Cal BP 3140 to 3000) (Cal BP 3160 to 2960)

Unit 137, Cal BC 900 to 830 (Cal Cal BC 920 to 810 Tajamar Level D9 312079 TA-U137-CR68 2660 ± 30 BP -20.5 2730 ± 30 BP BP 2850 to 2780) (Cal BP 2870 to 2760) Zone CV Mz BCD-8, Cal BC 1120 to 1010 circa Cal BC 1190 to 1000 La Vega Level 6-7s 312077 LV-0679-CR72 2720 ± 30 BP -14.4 2890 ± 30 BP (Cal BP 3070 to 2960) (Cal BP 3140 to 2950) Catalogue# Cal BC 1120 to 1010 circa Cal BC 1190 to 980 Trapichillo 1103 315533 CA-PhB-CR70 2870 ± 30 BP -24.3 2880 ± 30 BP (Cal BP 3070 to 2960) (Cal BP 3140 to 2920)

Catalogue# ca. Cal AD 690 to 780 circa Cal AD 680 to 870 Trapichillo 1102 312070 CA-PhA-CR69 1100 ± 30 BP -15.9 1250 ± 30 BP (Cal BP 1260 to 1170) (Cal BP 1270 to 1080) Unit 7, Layer 2, ca. Cal BC 890 to 810 Cal BC 900 to 800 Cerro Narrío Level 15 312075 CN-U7-CR113 2690 ± 30 BP -24.2 2700 ± 30 BP (Cal BP 2840 to 2760) (Cal BP 2850 to 2750)

Unit 4, ca. Cal BC 790 to 760 circa Cal BC 800 to 570 Cerro Narrío Level 3 312074 CN-U4-CR103 2540 ± 30 BP -24.1 2550 ± 30 BP (Cal BP 2740 to 2620) (Cal BP 2750 to 2520)

Unit 3, ca. Cal BC 790 to 670 circa Cal BC 800 to 550 Cerro Narrío Level 12 315534 CN-U3-CR110 2530 ± 30 BP -24.1 2540 ± 30 BP (Cal BP 2740 to 2620) (Cal BP 2740 to 2500)

Unit 3, ca. Cal AD 830 to 940 Cal AD 780 to 970 Cerro Narrío Level 11 312073 CN-U3-CR109 960 ± 30 BP -12.9 1160 ± 30 BP (Cal BP 1120 to 1010) (Cal BP 1170 to 980) TP-7, Cal BC 810 to 800 (Cal Cal BC 830 to 790 La Chimba Level 28 312076 LC-TP7-CR93 2590 ± 30 BP -22.0 2640 ± 30 BP BP 2760 to 2750) (Cal BP 2780 to 2740) 1The final letters and numbers of the submitter number correspond to the UC lab number. E.g., -CR63 = sample CR-63 2INTCAL09 calibration. In some cases multiple calibration ranges were reported. In those cases the end-limits of the oldest and youngest ranges are given as a single range, preceeded by "circa".

The Beta-312078 date from SALF, regardless of whether the 1σ or 2σ calibrated date

is considered, is contemporaneous with Valdivia Phase 1b (Zeidler 2003:519). None of

the highland site samples that were dated are earlier than the Middle to Late Formative

Period, and, in fact, can be argued to all date to the Late Formative Period. The Middle and Late Formative Periods, as defined from coastal Ecuador, have dates that overlap 250

between 1300 and 830 Cal BC (Zeidler 2008:460-470), and only the CR-89

Chaullabamba sherd date (Beta-312072) barely breaches this overlap and thus may be considered as dating to the Middle Formative Period.

Based on previous excavations and radiocarbon dates, Chaullabamba is reported to span the terminal Early Formative to Middle Formative Periods (2300 -1700 Cal BC)

(Bruhns et al. 1990:224; Grieder et al. 2002:162, 2009:8, 21-22; Stahl 2005:317) (also see Chapter 4, section 4.3). However, two AMS radiocarbon dates obtained from ceramic charred residues analyzed for starch (Tables 6.2 and 6.8) date to the Middle (1490-1320

Cal BC) to Late Formative (1290-1060 Cal BC). The sherd with the AMS date of 1490-

1320 Cal BC (CR-83, Beta-312072) originates from Cut III (3) H, Level 5. Grieder at al.

(2009:21) report a radiocarbon date obtained on wood charcoal from Cut 3 H, Level 4 as

2334-1744 Cal BC (1σ, 68%) (3530 ± 72 BP, TX-9241). Therefore the CR-83 ceramic charred residue AMS date is ca. 424 to 844 years younger than the date for the wood charcoal date even though the CR-83 sherd was stratigraphically below the wood charcoal. The possibility that the dated wood charcoal derives from old wood is further discussed in Chapter 7. The Late Formative Period date comes from sample CR-80 charred residue (1290-1060 Cal BC, Beta-312071) and the sherd originates from Cut 1,

Level 3. A radiocarbon date obtained on wood charcoal from Cut 1, Level 4 is reported as

2334-1709 Cal BC (1σ, 68%) (3950±200, TX-8439) (Grieder 2009:21). In this case, although the CR-83 sherd was recovered one level above the wood charcoal (and so the dates are stratigraphically consistent), Level 3 of Cut 1 appears to be ca. 649 to 1044 years younger than Cut 1 Level 4. These findings will be further elaborated in Chapter 7.

251

In contrast to the results from the charred residue sample AMS radiocarbon dates that either confirms the Late Formative Period contexts for some sites or questions dates that are earlier than Late Formative Period from other sites, the Tajamar site did not have any previous absolute dates. As noted earlier, the site, or at least components of the site, have been assigned to the Integration Period (AD 500-1500) based on the ceramic assemblages

(Domínguez et al. 2003:228; Jara and Santamaria 2009:55-56). However, the AMS C14 dates (Table 6.8) obtained from two separate charred residue samples (CR-67, Beta-

315535 and CR-68, Beta-312079) returned dates for the Late Formative Period, ca. 500 years prior to the onset of the Integration Period. Therefore, dates from two separate contexts (Units 151 and 137, both Levels 9), show that the site has a Late Formative component. I do not know the depths below surface that the two sherds were recovered from, but if they were 10 cm levels, that would put Level 9 ca. 80-90 cm BS, within the

D3 cultural layer and below the D2 volcanic ash layer (Domínguez et al. 2003:228), see

Chapter 4 (section 4.7 and Figure 4.22). Therefore, if the Integration Period characterization was based on surface collection and/or layer D1 (0-50 cm BS) ceramics, then the site has an Integration Period component and a Late Formative Period component, separated by a volcanic event. The Pululahua (Pululagua) volcano, which is very close to the Tajamar site, erupted in ca. 2358 BP (Athens 1998:169), 2305 ± 65

(uncalibrated, Hall 1977 in Zeidler and Isaacson 2003:75-76) corresponding to 752 to

182 Cal BC (Isaacson and Zeidler 1998:45-46; Lippi 2003:530; Zeidler and Isaacson

2003:75-76), and is a likely candidate for the D2 volcanic ash layer. The eruption may

252

have ended the Tajamar Late Formative Period occupation; whether the site was occupied prior to the Late Formative Period, and for how long, is unknown.

Another date inconsistent with previous findings comes from Trapichillo. As noted in section 6.6 above and in Chapter 4 (Section 4.4), Catamayo Phase A has been previously dated to the terminal Early Formative Period (2000 – 1400 BC) and Phase B to the

Middle to Late Formative Periods (1200 – 900 BC) (Guffroy 2004:85-86, 91). The two ceramic charred residue samples AMS dated (CR-69, Beta-312070 and CR-70, Beta-

315533), representing Catamayo Phase A and Phase B, respectively, returned puzzling dates (Tables 6.5 and 6.8). While the CR-70 date (Beta-315533, Cal BC 1190 to 980) is consistent with Catamayo Phase B and the Middle to Late Formative Periods, the CR-69 date (Beta-312070, Cal AD 680 to 870) is a minimum of 1500 calendar years younger than the CR-70 Catamayo Phase B sherd. As the exterior of both of the sherds bear no diagnostic decoration, as is common for cooking pots in general, it may be that the sherds were incorrectly assigned during original excavation and/or cataloguing or I made an error in cataloguing. As previously mentioned, Figure 5.4 Chapter 5 shows the actual

Phase B sherd being sampled with catalogue information on the sampling aluminum foil, and so it does not seem likely that I made an error in cataloguing the samples. Again, regardless of where the error lies, neither of the radiocarbon results I obtained date to the

Early Formative Period. The anomalous “Phase A” sherd dates to the Integration Period

(AD 500-1500), while the Phase B sherd date corresponds to the Middle (1430-830 Cal

BC) to Late Formative (1300-300 Cal BC) as it should (although, as discussed above, it does not exceed the Middle to Late Formative Period overlap and so is probably Late

253

Formative). The date of the sherd is not in question, and does show that Trapichillo was occupied and in use during the Integration Period (AD 500-1500), but it does not confirm delegating Catamayo Phase A to the Early Formative Period. As with the results from

Chaullabamba, this situation is disappointing as the Trapichillo site is one of the few highland sites that is purported to have an Early Formative Period component (Guffroy

2004:85-86, 91), yet I could not confirm this. The La Vega site is also reported to have

Catamayo A (Early Formative) deposits, but a single AMS radiocarbon assay on a split ceramic charred residue sample (CR-72, Beta-312077) shows that the sherd dates to the

Middle (1430-830 Cal BC) to Late Formative (1300-300 Cal BC) (Zeidler 2008:460).

Figure 6.32. Cerro Narrío Unit 3A showing ashy cultural layer. The ashy layer is clearly distinguishable from the overlying sediment and appears to represent a single, continuous occupation.

254

As highlighted in Tables 6.6 and 6.8 (in dark gray), the Unit 3, Level 11 (CR-109,

Beta-312073) date for Cerro Narrío is rejected as it is not consistent with other dates for that context. Another ceramic charred residue sample from Unit 3 (Level 12, CR-110,

Beta-315534) was sent for analysis, with a date returned that is a minimum of 1330 years older than the CR-109, Level 11 sherd charred residue date. Based on stratigraphy, Units

3 (Levels 9-12) and 3A (Layers 3-4) represent the same context. The cultural deposit is marked by an ashy layer that is distinct from the sediment above and appears to represent a single, continuous occupation as there are no visible changes in the layer indicating abandonment and reuse (Figure 6.32).

Units 3 and 3A are separated by an approximately 1 m wide backhoe trench, which was present prior to excavation (see Chapter 4, section 4.5). The ashy sediment layer appears continuous between the two units, only interrupted by the backhoe trench (Figure

6.33). An AMS radiocarbon assay (from 2011) obtained on a charcoal sample from Unit

3A, Layer 4, Level 4 resulted in a date of Cal BC 780 to 410 (Cal BP 2740 to 2260), which is consistent with the Unit 3, Level 12 date (Beta-315534) as well as the dates obtained from Unit 4, Level 3 (Beta-312074), Unit 7, Layer 2, Level 15 (Beta-312075), and the wood charcoal sample date from Unit 7, Layer 2, Level 17 – all of which are Late

Formative. Despite the fact that the Unit 3 CR-109 sherd (Beta-312073) originated from

Level 11 (approximately 10 cm or less) stratigraphically above the Level 12 sherd (CR-

110, Beta-315534), the three other dates from the Unit 3 ashy cultural layer are consistent with each other and therefore the CR-109 Unit 3 Level 11 sherd date seems unlikely to originate from the same context.

255

Figure 6.33. Cerro Narrío Units 3 and 3A showing backhoe trench bisecting the two units. Top image shows Unit 3A just being opened in 2008, with Unit 3 already excavated to the ashy layer from the 2007 excavations (photo courtesy of Scott Raymond). The bottom image shows Units 3 and 3A excavated down to approximately the same depth in 2009 and the top of the ashy layer is highlighted in red. Between 2008 and 2009, developers installed the cement sewer pipe shown in the bottom image.

256

Although there were no signs of disturbance in the Unit 3 profile the surface of the site is littered with sherds. Therefore it is possible that the CR-109 Unit 3 Level 11 sherd originated either from the surface or from the wall of the unit, above the ashy layer, and fell into the unit while Level 11 was being excavated. The date of the sherd is not in question, and does show that the Cerro Narrío was occupied and in use until at least Cal

AD 970, but it does not provide a date for the Unit 3, Level 11 context for the reasons stated above.

6.10 Chapter Summary

The starch analysis results of stone tool and charred ceramic residue samples from seven archaeological sites are presented in this chapter. Six sites are located in the highlands and one site in the eastern slopes/Upper Amazon region. All of the sites produced positive results for starch granules allowing identifications to be made.

Results from Santa Ana-La Florida, the lone site located in the Upper Amazon, show that starch granules consistent with Capsicum spp., Dioscorea spp., Fabaceae family (cf.

Phaseolus spp.), Ipomoea spp., Manihot esculenta, Theobroma spp., and Zea mays were recovered from artifact residue samples (stone bowls, intact stirrup-spout bottles, and ceramic sherds) that tested positive for starch. An AMS radiocarbon assay was obtained on one of the ceramic charred residue samples (obtained from a midden) and the date

(Cal BC 3380 to 3350, 2γ) is consistent with other previously reported Early Formative dates for the site.

257

The Chaullabamba site (southern highlands) samples (all ceramic charred residues) resulted in the recovery of starch granules identified as Fabaceae (cf. Phaseolus spp.),

Oxalis spp., Solanum spp. and Zea mays. Based on previous excavations and radiocarbon dates, Chaullabamba is thought to span the terminal Early Formative to Middle

Formative Periods (2300 -1700 Cal BC) (also see Chapter 4, section 4.3). However, two

AMS radiocarbon dates obtained from ceramic charred residues analyzed for starch date to the Middle (1490-1320 Cal BC) to Late Formative (1290-1060 Cal BC).

Starch granule analyses of stone tool and charred ceramic residue samples from the

Tajamar site revealed Fabaceae (cf. Phaseolus spp.), Oxalis spp., Solanum spp., and Zea mays were utilized. No previous radiocarbon dates are reported for Tajamar, but it has been assigned to the Integration Period (AD 500-1500) based on the ceramic assemblages. However, AMS C14 dates obtained from two separate charred residue samples returned dates for the Late Formative Period, ca. 500 years prior to the onset of the Integration Period. Therefore, dates from two separate contexts (Units 151 and 137, both Levels 9), show that the site has a Late Formative component.

The Trapichillo and La Vega sites are both located in the Catamayo Valley in the southern Ecuador highlands, although at elevations almost comparable to SALF. Starch analysis results from La Vega identified Capsicum spp., Fabaceae (cf. Phaseolus spp.),

Manihot esculenta, and Zea mays; results from Trapichillo were the same except that

Capsicum starch granules were not recovered. Catamayo Phase A has been previously identified as dating to the terminal Early Formative Period (2000 – 1400 BC) and Phase

B to the Middle to Late Formative Periods (1200 – 900 BC), yet AMS C14 assays on

258

split ceramic charred residue samples resulted in confusing dates in that the “Phase A” sherd dates to the Integration Period (AD 500-1500), while the Phase B sherd date corresponds to the Middle (1430-830 Cal BC) to Late Formative (1300-300 Cal BC) (as it should). As discussed above, it is likely that the lone Phase A sherd that I was able to sample was probably miscataloged and not truly representative of Catamayo Phase A.

The La Vega site is also reported to have Catamayo A (Early Formative) deposits, but a single AMS radiocarbon assay on a split ceramic charred residue sample (CR-72, Beta-

312077) shows that the sherd dates to the Middle (1430-830 Cal BC) to Late Formative

(1300-300 Cal BC) Periods.

Starch analysis results from three different excavation units from Cerro Narrío are internally consistent in that Fabaceae (cf. Phaseolus spp.), Oxalis spp., Solanum spp. and

Zea mays starch granules were identified from all three contexts at this southern highlands site. One AMS radiocarbon date obtained on a ceramic charred residue sample is rejected as it is not consistent with other dates from the same context or with diagnostic artifacts from that context. All of the other C14 dates are consistent with the Late

Formative Period, as expected based on the ceramics recovered during the excavations.

Finally, the La Chimba site, located in the northern highlands, charred ceramic residue samples analyzed showed the least number of identified taxa – Solanum spp. and

Zea mays. An AMS C14 date obtained from one of the split ceramic charred residue samples returned a Late Formative Period date, which is consistent with previously reported dates for the site. Further elaboration of the importance and meaning of the

259

starch analysis and AMS radiocarbon dates realized from these analyses will be discussed in Chapter 7.

Notes

1. Jørgensen and León-Yánez (1999) classify Theobroma and Herrania as Sterculiaceae in the Catalogue of the Vascular Plants of Ecuador. However, also in 1999, Theobroma and Herrania were reclassified in the family Malvaceae (Bayer et al. 1999). The

Catalogue of the Vascular Plants of Ecuador remains the most thorough and recent reference for Ecuadorian plants, and so, for comparison and consistency, I followed the classifications found therein. Despite the reclassification, Herrania remains the main possible alternate genera, in addition to other Theobroma species, that may have starch granules most similar to Theobroma cacao. In Chapter 7, section 7.3.1.11, I use the current Malvaceae classification.

260

Chapter Seven: Discussion

7.1 Introduction

The starch analysis conducted from the highland Ecuadorian sites of Cerro Narrío,

Chaullabamba, La Chimba, La Vega, Trapichillo, and Tajamar, and the one site located on eastern slopes, Santa Ana-La Florida, provide new information on plant resource use.

Other species identified by the starch granule analysis support previous data for domesticated plant use in the highlands. In this chapter I first review the methodological and quantification issues of starch granule analysis. This is followed by an appraisal and discussion of chronological issues related to the highland Formative Period, showing that domesticated plant species and a highland agricultural system were present when

Formative village sites first appeared in the highlands. With the addition of previous botanical analysis from the Cotocollao and La Chimba sites, I then use the entire paleobotanical record for the Formative Period in the highlands and eastern lowlands, together with site catchment analysis, to discuss what the botanical record tells us about cultural interaction in the highlands and eastern lowlands during the Formative Period.

Finally, I discus what the botanical record reveals about the nature of regional agricultural systems, with particular emphasis on the highlands.

7.2 Methodological Issues

7.2.1 Are the Starch Granules Recovered “Real”?

In Chapter 5 I discussed the considerable potential for contaminating archaeological samples with modern starch. Are the starch granules recovered from the

261

archaeological samples analyzed for this dissertation research representative of the contexts from which they originate? Extensive measures were taken to control and test for modern starch granule contamination. As discussed in Chapter 5, section 5.6, the wash samples from stone bowls and mano and metate fragments from the Cotocollao site were found to be conspicuously contaminated with hundreds of wheat/rye and/or barley starch granules per sample. A blank control that was initiated at the time of sampling, through to final lab processing in Calgary, was found to be negative for starch. A control sample is considered “negative” when three or less starch granules are observed on a microscope slide (always entirely scanned). A blank control undergoes all of the same steps as an archaeological sample. This means that I sonicate a sample of the distilled water that is used to remove residues from the artifacts in a plastic container (cleaned in the same manner and at the same time as the containers, sampling instruments, centrifuge tubes, and microscope slides used for the archaeological samples), I swirl and brush a clean toothbrush and/or any other instruments used into the water sample, the control water sample is placed into a clean centrifuge tube, it is dehydrated and dried together with the archaeological samples, stored in the same new zip-lock bag as the archaeological samples, transported to Calgary and undergoes all of the same processing steps together with the archaeological samples. These precautions, coupled with the unusually high number of starch granules of wheat/rye and/or barley (that cannot be representative of pre-Columbian use) recovered per sampled (in the hundreds), show that the Cotocollao artifacts were already contaminated with modern starches at some time before I sampled their residues.

262

Considerable measures were taken during excavation, sampling, storage, laboratory processing, and microscope slide preparation to ensure that the starch granules recovered were derived from the archaeological contexts, as detailed in Chapter 5, sections 5.5.1 – 5.5.3. For all of the archaeological samples obtained and analyzed, the blank control samples were negative for starch granules. Blank controls that are initiated at the time of sampling are not the only control samples analyzed, however. Numerous control slides left on laboratory benches and in the fume hood in the South American

Research Lab at the University of Calgary where I performed final sample processing were also negative for starch, showing that airborne or other incidental modern starch granules did not contribute to the archaeological samples’ starch granule assemblages.

Despite these precautions it is almost unavoidable that a few modern starch granules may become incorporated into archaeological samples at some point during excavation, cataloguing, sampling for residues, storage and/or processing. Therefore, I only considered an archaeological sample “positive” when at least five starch granules were observed when scanning a microscope slide; any less was considered “negative.” Is this number (≥ 5 starch granules = positive result) valid? This number (≥ 5) is based on my almost 10 years of experience in starch granule analysis whereby there has only been one occasion (see Chapter 5, section 5.5.3) where I have had a control sample show more than three starch granules (the majority of control slides, especially the blank controls, are usually free of starch). When starch granules are observed on a control sample slide, another slide is made from the same control sample. In all cases the second slide had either less than three or no starch granules. Consequently, I feel that a positive detection

263

limit of a minimum of five starch granules is valid and most archaeological samples analyzed for this dissertation had far more than that. Of the 79 archaeological samples analyzed, 59 (ca. 75%) produced positive results. Of the positive results, only ca. 19%

(n=11) of the samples had between 5 and 10 starch granules, while ca. 81% (n=48) resulted in >10 starch granules (ranging from 12 to 154) (39 samples, ca. 66%, had more than 20 starch granules).

As discussed above and in Chapter 5, numerous lines of evidence demonstrate that the starch granules recovered from the artifact residues are a result of tool use and not from any form of contamination. Another line of evidence supporting this contention is that similar, but not identical (as one might expect if systematic contamination from unknown sources was occurring) assemblages of starch granules were recovered from different contexts (samples). Furthermore, starch clusters/aggregates (Chapter 6, Figure

6.2 I-J), some showing signs of gelatinization with enlargement and/or distortion of granules and loss of birefringence, as well as isolated starches showing similar changes associated with cooking in excess water and/or toasting and fermentation (low- temperature heat) (Chandler-Ezell et al. 2006:110; Zarrillo et al. 2008:5009-5010), were observed in almost all of the samples (see Tables 6.1 through to Table 6.7, Chapter 6), including stone tool residue samples. None of the few starches observed in the control slides that had starch granules (all ≤ 3), had changes to morphology or other characteristics (loss of birefringence) to indicate exposure to heat, and no starch clusters were observed. These multiple lines of evidence also support the conclusion that the starch granules recovered from the ceramic charred and stone tool residues are a result of

264

use. Although further research is required to assess how diagenesis affects starch granules

(Collins and Copeland 2011) and whether natural factors may result in starch granule gelatinization over time (rather than use-related cooking/processing), I suspect that starch clusters/aggregates embedded in and recovered from ceramic charred cooking residues are a good indication of actual cooking events. Indeed, one may argue that all starch granules recovered from ceramic charred residues, where they were trapped inside a protective matrix since discard, not only provide direct evidence for food preparation and consumption (Zarrillo et al. 2008), but are also protected from diagenic change, as similarly argued for starch recovered from dental calculi (Hardy et al. 2009:254).

I feel that the precautions taken to reduce the risks for modern starch contamination coupled with the negative results for starch realized from the blank controls and other control samples, as well as the points made above, not only validate my results but also show that these precautions are necessary. I have used blank controls since I started performing starch granule analyses in 2003 (e.g., Quon 2003; Zarrillo 2004; Zarrillo and

Kooyman 2006; Zarrillo et al. 2008), and I know of no other published ancient starch analysis study that documents the use of blank controls. I strongly feel that testing laboratory surfaces, sample containers, sampling instruments and tools, centrifuge tubes, laboratory glassware, distilled/deionized water sources, lab reagents, etc. is not adequate to ensure that incidental modern starch contamination, from all potential sources (not just airborne starch), has not occurred and that such testing must be done together with blank control samples. To do otherwise goes against standard analytical practices and undermines the quality and integrity of the results obtained (Reeuwijk 1998:sections 7.3

265

and 8.4). These comments are not meant to question the results of previous analysis by other researchers, but to suggest that more can be done to further demonstrate the integrity of starch analysis results. The methods used to prevent incidental starch contamination should be explicitly stated when results are published.

7.2.2 Assessing the Methods for Sampling

Methodological issues with respect to the different sample types (stone bowl and stone tool residues, sediments, residues from intact ceramic bottles, and charred ceramic cooking residues) mainly revolve around mechanisms for preservation and sampling procedures. As noted in Chapter 5, the mechanisms that allow starch granules to preserve and be recovered from stone tool residues is unknown, especially in light of the prevalence of enzymes and microbes present in sediments that normally degrade starch granules. As Haslam (2004) has thoroughly reviewed, and as suggested by others (e.g.,

Loy 1994:110-111; Barton et al. 1998; Pearsall et al 2004:437; Perry 2001:186; Piperno and Holst 1998:768, 772; Zarrillo and Kooyman 2006:484-485), starch granules likely survive in stone tool residues because they are protected in cracks and crevices present on stone tool surfaces. More recently, Weiner (2010:222) has questioned this potential mechanism for starch granule preservation on stone tool surfaces, stating “bearing in mind the sizes of starch grains (from 1 to 100 or so microns in diameter), it is difficult to conceive of a widespread niche of that size, such as a crack, that would actually protect them.” I do not agree with this statement. Examining the surfaces of even obsidians

(volcanic glass) and cryptocrystalline cherts, both of which have extremely smooth

266

surfaces, with a simple 10x magnification hand lens will reveal many suitably large

“niches”, such as along fracture planes, in naturally occurring imperfections/inclusions, and/or where microflakes have been removed (either from use or by intentional shaping), where starch granules may be embedded and protected from degrading enzymes and microbes. Further, I suspect that using stone tools to grind or scrape organic matter produces a biofilm coating, perhaps similar to dental plaque, that results in starch granules that are deeply embedded in cracks and crevices on the stone tool surface, being sealed beneath (and within) the “tool plaque”, aiding in their preservation.

Sampling methods for stone bowl and tool residues and sediments follow standard procedures developed previously that have been shown to be effective in recovering starch granules (e.g., Pearsall et al. 2004; Zarrillo and Kooyman 2006). Unfortunately the size of the ultrasonic bath I used in Ecuador did not allow me to isolate utilized and unutilized surfaces for some of the larger artifacts, as I would have preferred. Doing so may have allowed for comparison of starch granule recovery from the used and unused surfaces to further support tool-use. As well, many of the artifacts had been previously washed prior to my sampling procedures. While ideally unwashed lithic tools are preferred for residue analysis, washing does not preclude the successful recovery of starch grains from stone tool residues (Loy 1994:96; Piperno and Holst 1998:772;

Zarrillo and Kooyman 2006:484, 492-493). In those cases my first priority was to wash the artifacts, collect the effluent, and test the wash samples for incidental starch contamination. As such, I did not obtain dry brush and wet brush samples from most artifacts, but only a wash and ultrasonic wash sample. In all cases (see Chapter 6) the

267

wash samples were negative for starch, confirming that the starch granules recovered by ultrasonic wash were representative of artifact use.

The method for removing residues from one ceramic sherd sampled (sample CR-54 from SALF) was similar to that employed for a stone tool (Pearsall et al. 2004) in that dry brush, wet brush, and sonicated samples were obtained. This method allowed me to show that the sonicated sample had 10 times more starch granules/volume than the dry brush sample, indicating that the starch granules recovered were representative of use-related cooking residues (see Chapter 6, section 6.2). This method, then, has applicability for other studies where charred cooking residues are either not present or difficult to isolate from adhering sediment, as was the case for many of the sherds from SALF.

Sampling residues from the interior of the intact ceramic bottles from SALF presented a challenge in that I did not have direct access to the interior surfaces of the vessels allowing for dry brush and wet brush samples to be acquired. Faced with this I decided to obtain an interior rinse sample and an interior sonicated sample, utilizing the same volume of distilled water for both sample types. The results (see Chapter 6, section

6.1) show that the sonicated sample (CR-51) from the stirrup-spout face bottle had 2.1 times (over 200%) more starch granules than the CR-50 rinse sample (154 vs. 73, respectively). A more dramatic increased recovery of starches (>2700%) was noted for samples CR-52 (stirrup-spout donut bottle rinse) and CR-53 (sonicated sample), wherein the sonicated sample had over 27 times the number of starch granules than the rinse sample (109 vs. 4, respectively). While these results seem to indicate that the sonication procedure recovered a greater number of starches than simply rinsing the bottles with

268

distilled water and collecting the effluent, the distilled water had been left to sit in both bottles prior to sonication for about 12 hours, and this alone may have resulted in greater starch granule recovery. In any case, the results show that this method was successful in recovering starch granules from intact ceramic bottles. Further testing, for example, by collecting the water left to sit in bottles overnight and analyzing that separately from a rinse sample and a sonicated sample, will allow for further conclusions to be drawn on the efficacy of the different sampling techniques.

Finally, the procedure used to recover starch granules from charred cooking residues adhering to the interior of ceramic pots has previously been shown to be effective and has the benefit of allowing for AMS radiocarbon assays to be obtained on the same sample (Zarrillo et al. 2008). Issues regarding the accuracy and precision of the

AMS radiocarbon results are discussed below in section 7.4. All of the sherds, whether previously rinsed of adhering sediment or not, were rinsed with distilled or deionized water to reduce the possibility that starch granules from sediments, or from previous handling, may have been transferred to the interior surfaces of the sherds. I did not, however, collect the rinse effluent for testing. One improvement to this procedure would be to conduct a formal study and test whether or not starch granules are present in ceramic pre-sampling rinse effluent.

7.2.3 Assessing Starch Granule Identification Methods

The method of identifying starch granules to taxa is based on comparing multiple three-dimensional characteristics of the archaeological starches to modern comparative

269

starch granules. The underlying assumption, of course, is that starch granules from modern plant species have not changed morphologically over time, allowing us to identify starch granules from archaeological contexts. It is important here to mention that

Donald Ugent, the pioneer of archaeological starch analysis studies, used starch analysis to confirm the identification of desiccated tubers and rhizomes recovered from middens at numerous sites (ranging in age from 2250 to 295 B.C.) from the Casma Valley of coastal Peru (Ugent et al. 1981; Ugent et al. 1982; Ugent et al. 1984; Ugent et al. 1986).

Consequently, Ugent was not relying on comparing the archaeological starch granules to modern comparative samples alone, but also fine and gross morphology of the desiccated plant remains, scanning electron microscopy, and iodine-staining, to identify sweet potato

(Ipomoea batatas), potato (Solanum tuberosum), achira/canna (Canna edulis), and manioc/yuca (Manihot esculenta). Later, Ugent (1994) also used paper and thin-layer chromatography and ultraviolet light spectrophotometry to confirm the differentiation of ancient starch types. With respect to even greater time-depth, Ugent (1997; Ugent et al.

1987) identified wild potato (Solanum maglia) macroremains based on starch granule analysis at the Monte Verde site (Chile), dating to about 12,500 BP. The pioneering work of Donald Ugent not only illustrates the use of starch granule analysis in documenting past plant use, but also confirms that modern comparative starch granules from different species can be used to identify starch granules from archaeological contexts.

Having reaffirmed that starch granule morphologies of modern species can be used to identify ancient starch granules, as discussed in Chapter 5, an extensive comparative collection of known modern plant species is required from the region of study, inclusive

270

of as many different species as possible, in order to make secure identifications. It is also particularly important to have an extensive comparative collection in regions where native (wild) plant species may have been utilized, as differences in starch granule morphology between closely related native species may not be as well-defined as they may be between a domesticated plant species, its wild progenitor, and related species. In almost all cases, published description and micrographs of starch granules, as reviewed in

Chapter 6, indicate that my identifications (to genera and/or family) are probably overly conservative and that most of the starch granules can be identified as domesticated species. The lack of a more diverse comparative collection, in some cases, precluded more definitive identifications. In other cases, the specificity of starch granules morphology is limited to the genus level, as for Dioscorea (see sections 7.4.1 and

7.4.2.3). For example, I identified Fabaceae family starches as cf. Phaseolus spp. because of the high number of Fabaceae family species in Ecuador and my lack of comparative samples. The other reason for applying caution and not “over-identifying” is related to the number of starch granules of the same morphotype that are recovered from the same context.

In addition to requiring a broad and inclusive starch comparative collection, the identification of an archaeological starch can be restricted by a low recovery rate of similar starch granules. That is, the archaeological starch in question may be an isolated example – the more starch granules recovered of the same type from an archaeological context, the greater the confidence in identification. This is an assemblage-based approach to identification, whereby a group of archaeological starches that are similar in

271

morphology is assessed as a whole in comparison to different reference taxa assemblages, similar to that used in phytolith analysis (e.g., Pearsall 2000:444-454; Piperno 2006a:103-

105). This is the most conservative method for identifying unknown starch granule assemblages as it addresses inter- and intra-species starch granule variation (e.g., Cortella and Pochettino 1994; Ugent et al. 1986; Piperno et al. 2000). There are instances where a starch granule is so unique in morphology for a given region that it can be confidently identified with only a few archaeological starch granules present. An example would be

Canna edulis (achira), which has extremely large starch granules (up to 180 μm) and they are unlikely to be confused with anything else. This latter approach to identifying archaeological starches is a diagnostic-based approach, whereby a particular form of starch granule is exclusively unique to a particular taxon. Another method for identifying individual starch granules from an assemblage of unknown starches is by using images to obtain measurements (size and shape) and other observed characteristics (e.g., extinction cross style and angles) and then apply multivariate statistics (discriminant function analysis) to classify the unknown starch granule to known assemblages (Torrence et al.

2004). In a somewhat similar method, Wilson and colleagues (2010) also used image analysis techniques, but with pattern recognition software, to assign probability identifications of individual starch granules (and assemblages) to taxa. These techniques would be useful for initial classification of unknown starch granule assemblages, but they are limited by utilizing two-dimensional characteristics to identify three-dimensional objects (starch granules). Until further refinements to image analysis methods are made, I feel that final identification should be made by direct observation and characterization of

272

starch granules to reference collections. Because there is natural variation in starch granule morphological characteristics, even within a species (not every starch granule from a species will have all diagnostic characteristics), the most prudent method for identifying starch granules, then, is an assemblage-based approach. This is the method that I employed. As such, the fewer archaeological starches of a similar type that are recovered, the lower the certainty is for identification. Returning to the example of

Phaseolus, only three starch granules consistent with Phaseolus were recovered from

SALF. For this reason, and in the absence of a more inclusive comparative collection, I identified these starch granules to the lowest taxonomic level possible (Fabaceae) with good confidence. Even though 45 starch granules consistent with Phaseolus spp. starch were recovered from the Cerro Narrío samples, I cannot say with certainty that other

Fabaceae family species did not contribute to, or are represented by, the cf. Phaseolus spp. starch granules assemblage.

For this analysis I also employed an extremely conservative approach to identifying maize starch granules. As discussed in Chapter 6, section 6.2, the reason for this is that although Cenchrus brownii (that produces starch granules in its seeds that are similar to, but can be distinguished from, Zea mays starch) (Zarrillo et al. 2008 Supporting

Information:1-2) was available for comparison, there are five other species of Cenchrus in Ecuador that I did not have comparative samples for. As such, only starch granules that possessed very restricted morphological characteristics (see Chapter 6, section 6.2) were identified as Zea mays.

273

The identification of cacao (Theobroma spp.) seed starch granules is based on identifying and characterizing a diagnostic starch granule morphology based on modern comparative Theobroma cacao seeds. There were three different starch granule morphologies in the comparative seed specimens: 1) spherical or sub-spherical, which comprised 30% of the assemblage and ranged in size from 3.8-6.3 μm; 2) small truncated-spherical compound granules, which comprised 48% of the starch granules and varied in size from 3.8-10 μm; and 3) ovoid (teardrop-shaped) starch granules that represent 22% of the starch granules and ranged in size from 6.3-12.5 μm. I consider the ovoid/teardrop form to be diagnostic due to its distinctive shape and other attributes, as discussed in Chapter 6, section 6.2. Because of a lack of other T. cacao comparative samples, and especially a lack of wild Theobroma samples, I have limited the identification to the genus-level. The only other Sterculiaceae family species reported to be consumed in Ecuador is the monkey cacao/cacao de monte (Herrania spp.) tree, wherein the seeds and seed pod-pulp are consumed much like Theobroma (Rios et al.

2007:146, 178, 228-229; Jørgensen and León-Yánez 1999:918-922). Therefore, as I cautioned in Chapter 6, the Theobroma spp. identifications are tentative until I can expand my Theobroma spp. comparative collection and obtain and characterize the seed starch granules for Herrania spp.

7.2.4 Methods of Quantification

In Chapter 6 I applied three main quantification techniques: 1) absolute counts of the starch granules; 2) presence/absence (ubiquity) analysis; and 3) relative frequency

274

percentage ratios. I also pointed out that some of the 13C/12C assays from the ceramic charred residue samples submitted for AMS dating, may be seen as another (independent from starch analysis) form of quantification. This issue will be discussed separately, in section 7.2.5, below. The most basic technique of documenting the occurrence of plant remains is to cite their occurrence in samples. Even a basic laundry list of species identified can provide important qualitative information about domestication, trade, subsistence, seasonality of occupation, etc. (Wright 2010:51). Ubiquity, or presence analysis, documents the number of proveniences in which a plant species is recovered

(Hastorf 1993:166; Pearsall 2000:212-216; Popper 1988:60-64; Wright 2010:51-52).

Relative frequency/percentage calculates the percentage of starch granules of a type (or taxon) represented within the total assemblage (Miller 1998:72; Renfrew 1973, in Hastorf

1993:166). Both cultural and non-cultural factors bias the numbers and types of plants that are recovered from archaeological contexts. Therefore, taxa identifications and abundances do not directly reflect the full range of past human-plant interactions.

There are many sources of patterning that must be indentified in order to make interpretations that are a reflection of past activities based on an assemblage of plant remains (Popper 1988:53-54). Different uses of plants and related activities can leave different patterns of plant remains. For example, potatoes (and other USOs) may be cultivated and harvested in a plot removed from the habitation area, with only the potatoes (and not the rest of the plant) brought back to be washed, cooked without peeling, and eaten whole. Therefore the only plant remains recovered may be starch granules (or preserved coprolites) and only if the potatoes were cooked in a pot and not

275

roasted/baked (where the only vestiges then may be some charred remains from a cooking accident). This pattern would be in contrast with a hypothetical scenario of maize use. Maize may be ground on stone tools, cooked in pots, and/or toasted and popped (where the chance for more durable charred macroremains entering the archaeological record is greater), and these uses and activities will leave different patterns than those for potato. Popper (1988:55) provides a useful flow chart illustrating the cumulative factors affecting patterning of archaeobotanical data, beginning with people’s beliefs prescribing what plants are used and how they are used, to the recorded types and numbers of plant remains recovered from archaeobotanical analysis. I have modified

Popper’s table to illustrate the same for starch granules (in particular), as shown in Figure

7.1.

As can be seen in Figure 7.1, a group’s worldview will determine the types of plants utilized. The first source of bias is that not all plants produce storage starch granules (as discussed in Chapter 5), and, of those that do, storage starch granule production is not the same across species. As well, the plant part(s) used may not produce storage starch. Where plants are cultivated and used, how and where they are processed, stored and disposed of are additional factors that will affect patterning. The types of processing implements, such as milling stones (but also baskets, sacks, bowls and platters made from perishable materials) and the types of cooking and serving vessels, and how they are handled, will also vary. The result of this is that not all plants will be exposed to all, or possibly any, implements. These factors will bias the types, quantity and nature of the botanical remains entering the archaeological record.

276

Figure 7.1. Cumulative Stages of Patterning for Starch Granule Data. Modified from Popper (1988:55).

There are also several factors that will skew the preservation of botanical remains.

With respect to storage starch granules, as noted in Chapter 5 (section 5.2), they are mainly composed of amylose (20-30%) and amylopectin (70-80%) (from now on I will revert to using the term “starch granules” or simply “starch” when referring to storage 277

starch granules, as transient starch granules are not pertinent to this or the following discussions). Although the chemical composition of starch is quite simple, the structure is not. For example, while proteins are easily characterized by the sequence of the 20 different types of amino acids that they are built from, and nucleic acids can be analyzed to determine their nucleotide sequences, such is not the case with starch components as it is practically impossible to organize the carbohydrate components into a meaningful sequence (Bertoft 2004:57-59). Differences between the amylose and amylopectin structures, as well and the amounts and composition of other minor constituents (such as lipids), vary and impart diversity to the characteristics of starch granules from different botanical sources (Thomas and Atwell 1997:7-11). While amylose is a “minor” component in most granules, it has a strong influence on starch properties. When starch granules are exposed to iodine a complex is formed between iodine and the amylose fraction, imparting a deep blue colour. The colour and the intensity are dependent on the chain length of the amylose, which varies between species. Cereal starches in particular do not stain dark blue with iodine, not because the granules have less amylose than other starches, but because a considerable part of the amylose is complexed with lipids, and the lipid-amylose complexes have no affinity for iodine (Bertoft 2004:61). This is only one example of the tremendous variety exhibited in the physiochemical composition and characteristics of starch granules. The consequence of these and other variations for differential preservation of starch granules from different species is unknown. Starch granules from different species definitely display different gelatinization temperatures, and this characteristic alone has been used in the food and manufacturing/commercial

278

industries to differentiate and characterize starches from different species (Kesler and

Bechtel 1953:427). When exposed to heat in excess water, starch granules exhibit irreversible changes such as swelling, loss of birefringence, native crystalline melting, and eventually total solublization (Thomas and Atwell 1997:25-27). For instance, the temperatures for the onset of gelatinization to completion are 64-73°C for Zea mays and

64-84°C for Ipomoea batatas (sweet potato) (Moorthy 2004:327; White and Tziotis

2004:303). However, a quick look back at the results from SALF (Chapter 6, Table 6.1) shows that maize starch granules are far more prevalent than sweet potato starch in the

SALF cooking pots. Based on gelatinization temperature alone, and all other factors being equal, sweet potato starch granules should have been preserved at a higher rate. But all else is not equal. I have argued elsewhere (Zarrillo et al. 2008:5009) that several factors, including differences in physiochemical and structural properties of starch granules from different species, as well as food processing and cooking techniques, likely bias the types and number of starch granules recovered from ceramic charred residues.

Here I extend that contention for starch granules preserved (or not preserved) on artifacts and in sediments. As noted in Figure 7.1, many other factors, such as the vagaries in depositional environments (temperature, moisture regimes, soil pH, microbes and enzymes, exposure, burial) within and between sites, affect (and probably interact with each other in ways we do not as yet understand) the preservation of botanical remains

(including starch granules). Finally, the types, numbers, and characteristics of the samples collected from archaeological contexts, the ways these samples are processed, as well as our ability to make identifications, also bias the final starch analysis results.

279

Returning to the validity of quantification techniques then, we can see how even simple measures such as absolute counts, ubiquity, and relative frequency may not be

(and are almost assuredly not) a true reflection of the full range of plant use, nor are they even likely to be an accurate reflection of the relative abundance of the species identified.

In this regard, I agree with Kadane’s (1988:211) comment that “forming a ratio does not in itself make preservation issues go away” and that absolute counts, ubiquity, and relative frequency measures share similar problems (biases). Therefore, while I provided absolute counts, ubiquity, and relative frequency measures in Chapter 6 for all of the starch analysis results, I limited any discussion on comparison of these measures to restricted contexts, such as the CR-50 and CR-51 stirrup-spout face bottle from SALF

(section 6.2). In these instances cultural, taphonomic, and processing factors are most equal in that the results of different samples from the same artifact are being compared.

Conversely, the collective results of starch analysis from all sites, as shown in Chapter 6, show that maize starch granules are overwhelmingly in the majority as compared to all other identified and unidentified taxa, but it would be incorrect to conclude from these data that people were eating twice as much maize as other starchy plants. The reason for these differences in data comparison is straight forward. In the first case I am making a reasonably fair assumption that the variables that affected the data being compared are the same. In the second case it would be inappropriate to make such a simple inference since there are numerous variables impacting the sites, and that these variables have most likely skewed the data (and not in consistent ways). In this regard, applying more advanced statistical tests (e.g., using inferential statistics) is even less appropriate because

280

doing so would assume that my sample data is an accurate representative (normally distributed) sample (Sanders1990:7-9). What we are left with then, and as I will discuss below, is a list of starchy plants that are present at the sites investigated. This list is probably incomplete, even with respect to plants that produce abundant starch granules, and is almost certainly incomplete as far as the full range of plants that were used at these sites.

7.2.5 Quantification and the Issue of 13C/12C Stable Isotope Assays

Table 6.8 in Chapter 6 showed the AMS radiocarbon and 13C/12C (δ13C) assays obtained from the ceramic charred residue samples submitted for analysis. As mentioned in Chapter 6, and as can be seen in Table 6.8, some results show enriched δ13C values consistent with an increased incorporation of C4 plants (with maize being the most likely

C4 plant) in those cooking residues (see also Chapter 5, note 3, for a discussion on the

13 difference between the C3 and C4 photosynthetic pathways). Samples with enriched δ C values are present from only 4/13 (ca. 31%) of the samples analyzed, and include one from SALF (Beta-312078, -14.4 ‰), one from La Vega (Beta-312079, -14.4 ‰), one from Trapichillo (Beta-312070, -15.9 ‰), and one from Cerro Narrío (Beta-312073). It is noteworthy that these particular samples from Trapichillo and Cerro Narrío do not date to the Formative Period. Therefore, despite maize starch granules being recovered from all samples in greater numbers than all other starch granules, the δ13C determinations do not support that maize was the most abundant plant food cooked in the pots. In fact, if we were to assess the presence of maize solely on the δ13C assay, we would conclude that

281

maize was absent from 9/13 (ca. 69%) of the samples that were analyzed based on the depleted δ13C values (≤ -20.5‰). Can these differences in quantification be resolved?

I have argued elsewhere (Zarrillo et al. 2008:5009 and Supporting Information), that cooking pot residues embody a veritable “stew” of the different foods (both floral and faunal) that were cooked in a pot over its use-life, and therefore δ13C determination may not reflect the relative proportions of C3 and C4 sources that contributed to the cooking residues accurately. Indeed, Hart and colleagues have shown through analyzing and quantifying experimental cooking residues that: 1) maize is systematically underrepresented, even when it contributed the bulk of the cooking matter, because a non-linear relationship exists between the percent of maize cooked, in comparison to other foods, and the resultant δ13C (Hart et al. 2007); and 2) the mobilization of carbon from C3 resources and maize in cooking residues is determined by the form of maize cooked (Hart et al. 2009). As explained by Hart and colleagues (2009:2210-2211), when whole-kernel or hominy/mote (alkali-treated) maize is used, it is masked by the C3 food sources cooked in the same pot. This is further complicated by the experiments conducted by Lovis and colleagues (2011), which show that even if maize is cooked alone with hardwood ash (as in the production of hominy/mote), the δ13C value of carbonized residues will be depleted, masking the presence of maize. Conversely, when corn meal

13 (or flour) is used it masks the C3 resources and enriched δ C values are realized (Hart et al. 2009:2210-2211). Of course, when maize is cooked alone (and I would extend this to very low-temperature cooking to begin fermentation for maize chicha – beer – production), in any form (except when C3 wood ash is used), in a pot dedicated to such

282

use, enriched δ13C values will be obtained from the cooking residues. Consequently, it is important to know what resources may have been cooked in a pot and what their relative proportions were (a very difficult task) in interpreting δ13C values from cooking pot residues (Hart et al. 2007, 2009). It is suggested, however, that δ13C values may be used to “monitor maize preparation practices” (Hart et al. 2009:2211), even if they cannot be used to estimate the percentage of maize cooked, or, in the case of depleted δ13C values, even if maize was ever cooked in a pot (false negatives). The research by Hart and colleagues (2007, 2009) and Lovis and colleagues (2011) is interesting in that we might conclude that the four results with enriched δ13C values from SALF, La Vega,

Trapichillo, and Cerro Narrío, may indicate that: 1) maize corn meal or flour was cooked in combination with C3 resources; or 2) only maize was cooked in those four pots (and not with C3 wood ash). Raviele (2011:271) also showed that abundant maize starch granules are found in experimental cooking residues when dried (mature) maize (whether whole or coarsely/finely ground) is cooked, but not with green (milk stage) maize. These combined findings then, may indicate that mature dried maize was cooked, either as corn meal or flour, in the four pots with enriched δ13C values (or these pots were dedicated to maize cooking/processing/fermentation), whereas mature dried whole maize or hominy/mote may have been cooked in all of the other pots resulting in positive starch results, but depleted δ13C signatures.

Returning to the question of whether the differences between “quantification” based on starch granule analysis and δ13C assays can be resolved, the answer is no. This is because neither approach can quantify the relative abundance of starchy foods cooked,

283

in the case of starch analysis, or the relative abundance of C4 verses C3 resources based on δ13C assay of cooking pot residues. However, although I cannot quantify the relative abundance of maize by either starch granule analysis or by the δ13C assays, if I were

13 relying on δ C alone, and did not consider the attenuating effects of C3 resources, I would have concluded that 9/13 (ca. 69%) of the samples showed no evidence for maize use. The starch granule evidence, however, shows that maize was cooked in all of the pots that produced positive starch results, including the nine pots with depleted δ13C signatures. These findings may indicate that a range of maize processing and cooking practices were employed, such as the of use of corn meal/flour verses dried, whole-kernel maize, the production of mote, and chicha brewing, with these different culinary practices reflected in the varied stable isotope signatures from different pots (and the preservation of starch granules). Moreover, then, these data show that sophisticated cooking techniques were used, a diverse variety of foods were cooked in (most of) the pots, and consequently a broad-based subsistence economy is indicated for SALF, Trapichillo, La

Vega, Chaullabamba, Cerro Narrío, Tajamar and La Chimba. These results mirror those shown for coastal groups (as discussed in Chapter 3). Further, these discrepancies between the results obtained by different types of analysis highlight the importance of not relying on one method of analysis alone, and in using multiple lines of evidence in interpreting botanical (and other) data. Further research is needed to better understand the mechanisms involved in starch granule differential preservation, especially for cooking pot residues (Collins and Copeland 2001; Henry et al. 2009; Raviele 2011; Yang and

Jiang 2010:1155; Zarrillo et al. 2008:5009), before any meaningful headway, if any, can

284

be made with respect to quantification. The analysis and interpretation of δ13C determinations of experimental and archaeological cooking residues, especially if combined with compound-specific δ13C measurements on lipids from the cooking residues or absorbed by the pots (e.g., Reber and Evershed 2004; Reber et al. 2003; and see Malainey 2011 for a recent overview of different techniques), may be a promising line of enquiry to complement starch granule analysis.

7.3 Accuracy and Precision in Radiocarbon Dating

The meanings of the terms "accuracy" and "precision" in radiocarbon dating are quite different from each other. Accuracy refers to the date obtained being a “true” estimate of the sample age within the range of the statistical limits (or ± value of the date). Precision, on the other hand, is the reproducibility of repeated measurement, regardless of whether the resulting measurements are accurate. Several radiocarbon dates might be obtained that are similar (reproducible and precise), but not accurate for the context being dated (in other words, precisely wrong). Random errors affect the precision of a measurement, whereas systematic errors determine accuracy (Bowman 1990:39-40).

If a systematic error is present and persists through a series of measurements, then the results, although highly replicate, will produce a bias relative to the true (accurate) date.

As presented in Chapter 6, section 6.9, AMS radiocarbon assays were obtained on ceramic charred residue samples from all of the sites investigated. Two sites,

Chaullabamba and the Catamayo sites (Trapichillo and La Vega) stand out in that the ceramic charred residue radiocarbon dates are younger than previously reported dates for

285

these sites. Therefore, it is important to assess the ceramic charred residue AMS dates within the context of other absolute dates for the Chaullabamba and the Catamayo sites.

7.3.1 Assessing the Ceramic Charred Residue AMS Radiocarbon Dates: Chaullabamba

The results of all known dates for Chaullabamba are shown below in Table 7.1, and include dates obtained on wood charcoal, animal bone, and ceramic charred residues. It is important to note that all material submitted for radiocarbon assay may be contaminated with older and/or younger carbon, making the context of the material being dated especially important in interpreting the results (Bowman 1990:23-28). Six of the fifteen known radiocarbon dates for Chaullabamba date to before 1450 Cal BC (terminal Early

Formative Period) with 2 sigma (95%) probability, and, of these six, only two (TX-8439 and TX-9241) have ranges that fall entirely within the Early Formative Period. Therefore, are the two Early Formative Period wood charcoal dates older than the actual age of the contexts, or are they accurate and reflect the earliest occupation of Chaullabamba? It is always possible that the charred wood samples that were dated (in both cases) are from long-lived trees (or dead fall that was collected), and are therefore older than the actual date that the wood was used. Unless the wood samples were identified prior to radiocarbon dating to assess whether or not they were long-lived species, there is no way of knowing for certain.

286

Table 7.1. Radiocarbon Dates for the Chaullabamba Site. Measured Assay Age Calibrated1 Age Calibrated1,2 Age Lab # Sample Type Method Provenience (RCYBP) BC (1σ) BC (2σ) Wood TX-8439 Charcoal Conventional Cut 1, Level 4 3950 ± 200 2334 to 1709 circa 3010 to 1880* Ceramic Charred Beta-312071 Residue AMS Cut 1, Level 3 2930 ± 30 circa 1260 to 1130 circa 1290 to 1060 Wood TX-9025 Charcoal Conventional Cut 2, Level 3 3130 ± 60 1460 to 1340 circa 1530 to 1260* Wood TX-9241 Charcoal Conventional Cut 3H Level 4 3530 ± 72 2334 to 1744 circa 2120 to 1680* Ceramic Charred Beta-312072 Residue AMS Cut 3H Level 5 3090 ± 30 1430 to 1410 circa 1490 to 1320 Wood TX-9027 Charcoal Conventional Cut 3B Levels 1-3 3200 ± 60 1510 to 1390 circa 1630 to 1320* Wood TX-9026 Charcoal Conventional Cut 3A Levels 1-3 3160 ± 50 1470 to 1370 circa 1530 to 1310* AA55501 Animal Bone AMS Cut 3G, Level 5 2947 ± 49 1260 to 1230 circa 1370 to 1000* AA55500 Animal Bone AMS Cut 3G, Level 4 2972 ± 47 1300 to 1110 circa 1380 to 1040* AA55499 Animal Bone AMS Cut 3G, Level 3 2768 ± 49 980 to 950 circa 1030 to 810* AA55498 Animal Bone AMS Cut 3G, Level 2 2734 ± 49 920 to 825 circa 1000 to 800* AA55497 Animal Bone AMS Cut 3G, Level 1 2704 ± 49 900 to 815 circa 980 to 790* Wood BM-897 Charcoal Conventional ? 2909 ± 55 circa 1210 to 1010* circa 1290 to 930* Wood BM-906 Charcoal Conventional ? 2800 ± 48 circa 1020 to 890* circa 1120 to 830* Wood BM-907 Charcoal Conventional ? 2964 ± 50 circa 1300 to 1110* circa 1380 to 1020* Wood BM-909 Charcoal Conventional ? 2784 ± 50 circa 1010 to 850* circa 1060 to 810* TX, AA, and BM dates as reported in Grieder et al. (2009:21), Burleigh et al. (1977:148-149), and Burleigh and Hewson (1979:346-347). Italicized 1σ dates are as reported and were not recalibrated. The contexts of the BM dates are unknown. 1In some cases multiple calibration ranges were reported. In those cases the end-limits of the oldest and youngest ranges are given as a single range, preceeded by "circa". 2All Beta and BM (British Museum) dates are corrected for measured isotopic fractionation. It is assumed that the TX and AA dates were corrected as well, as they are reported as RCYBP (Radiocarbon Years Before Present), but it is unknown whether the fractionation correction was based on measurement or an assumed value. *Calibrated dates determined using OxCal 4.1 (Bronk Ramsey 2009), using IntCal09 calibration database (Reimer et al. 2009), rounded to the nearest 10 years.

Even at 2σ (95% probability), neither of the two Early Formative dates overlap with any of the other dates. However, as the previously reported dates for the site are from different contexts, this might be meaningless – the two Early Formative Period dates may accurately reflect the earliest known occupation of the site and need not

287

overlap with dates from other contexts. In this regard, the two ceramic charred residue dates are important as they are both from the same contexts (Cut 1 and Cut 3H) as the two Early Formative wood charcoal dates.

As discussed in Chapter six, section 6.9, the Beta-312072 (CR-83) AMS date obtained on ceramic charred residue is younger than the wood charcoal date (TX-9241), even though the CR-83 sherd was stratigraphically below (from Level 5) the wood charcoal (from Level 4) and both are from Cut 3H. In Chapter 6, section 6.9, I calculated the discrepancy between the Cut 3H, level 4 and 5 dates as being about 424 to 844 years, but with recalibrating the previously reported dates, as shown in Table 7.1, the age difference can perhaps be reduced to 360 to 630 years using the 2 sigma calibrated age ranges. The fact remains, however, that the Level 4 wood charcoal date is older than the

Level 5 ceramic charred residue date. As such, the accuracy of the TX-9241 Early

Formative date may be suspect. Grieder’s plan and profile drawings of the Cut 3 excavations do not shed light on the issue, as units H, J and K are not illustrated (only units A, C, E, G and F are shown in profile) (Grieder et al. 2009:13). All of the animal bone AMS (AA series) radiocarbon dates for Cut 3G are younger than the Cut 3H dates

(including Beta-312072) and do not overlap, even at 2 sigma, suggesting that some contexts of Cut 3H are indeed older than some contexts of Cut 3G. It is possible that a systematic error is responsible for the AA series animal bone dates being consistently younger than the wood charcoal (and ceramic charred residue dates). A systematic error might occur if the AA series animal bone dates were obtained on whole bone rather than the bone collagen fraction, as the collagen (protein) fraction does not always preserve

288

well (but is less subject to external contamination) (Bowman 1990:29-30), and this is what Grieder suggests (Grieder et al. 2009:22), but does not explicitly state. The other possibility, of course, is that the Beta-312072 is incorrect for the context (or, alternatively, both are incorrect). In Chapter 6, section 6.9, I argued that one of the ceramic charred residue AMS dates from Cerro Narrío, Unit 3, was not valid for the context from which it supposedly derived, showing that mistakes can be made. In fact, as additional dates from the same and other levels of Cut 3H were not determined, it is impossible to say which of the dates for Cut 3H are accurate.

From Cut 1 at Chaullabamba, two dates are available for comparison: 1) TX-8439

(ca. 3010 to 1880 Cal BC, 2σ) from Level 4; and 2) Beta-312071 (ca. 1290 to 1130 Cal

BC, 2σ) from Level 3. Even at the two sigma level, the TX-8439 wood charcoal date does not overlap with the Beta-312071 (CR-83) ceramic charred residue date. However, in this case, the sherd was apparently recovered one level above the wood charcoal and so the dates are stratigraphically consistent. The result is that Cut 1 Level 3 appears to be ca.

579 to 1074 years younger (comparing 1σ age ranges), or 820 to 1720 years younger if the 2σ age ranges are used. The plan and profile drawings for Cut 1 seem to show that

Level 3 and the position of the Level 4 wood charcoal (TX-8439) were at about the same depth (Grieder et al. 2009:10), but the site has a complicated history of construction making it entirely possible that the two dates are both accurate – only further dated samples from Cut 1 will resolve the issue. In this regard, and owing to the potential stratigraphically inconsistent Early Formative and Late Formative Period dates from Cut

3H, it seems that renewed investigations at Chaullabamba are warranted. New

289

investigations are also important as Chaullabamba, as discussed in Chapters 3 and 4, is a key site in understanding not only the regional chronology of the Cuenca and surrounding basins, but the highlands as a whole. Future investigations should include multiple radiocarbon age determinations of samples from stratigraphic contexts to assess the internal consistency of age determinations.

7.3.2 Assessing the Ceramic Charred Residue AMS Radiocarbon Dates: Trapichillo and La Vega – the Catamayo Sites.

As discussed in Chapters 4 (section 4.4) and 6 (sections 6.6 and 6.9), Catamayo

Phase A has been previously dated to the terminal Early Formative Period (2000 – 1400

BC) and Phase B to the Middle to Late Formative Periods (1200 – 900 BC) (Guffroy

2004:85-86, 91). For the Trapichillo site, one previous radiocarbon date is available.

During the 1979-1981 excavations on the lower shelf of the hill where the Trapichillo site is located, a ceramic vessel was recovered that was filled halfway with ashes and burned plant fragments. A sample of this was submitted for radiocarbon assay. The date reported is 3480 ± 90 BP, 1845 to 1771 Cal BC (Guffroy 2004:31), although it is not noted whether this is a 1 or 2 sigma calibrated date (but based on the spread it is likely 1 sigma). Recalibrating the date with OxCal 4.1 (Bronk Ramsey 2009), using IntCal09 calibration database (Reimer et al. 2009), rounded to the nearest 10 years, results in 1920 to 1680 Cal BC (1σ) and 2040 to 1530 Cal BC (2σ), confirming that the vessel contents, representing a secure context, date to the terminal Early Formative Period. In Chapter 6, section 6.9 I discussed the fact that the CR-69 sherd I sampled for adhering interior charred residues, which had been designated as Catamayo Phase A, was most likely 290

miscataloged in some way (incorrectly assigned to Phase A), because the AMS radiocarbon date from the charred residue for the sherd was anomalously too young. The

CR-69 date (Beta-312070, Cal AD 680 to 870), is far too young to be associated with

Catamayo Phase A at Trapichillo, which, based on the seemingly secure context of the date reported by Guffroy, seems to solidly date some contexts at Trapichillo to the terminal Early Formative Period (designated Catamayo Phase A). The CR-70 sherd from

Trapichillo was designated as Catamayo Phase B, and the AMS radiocarbon date on the charred residues (Beta-315533, Cal BC 1190 to 980) from this sherd is consistent with

Catamayo Phase B and the Middle to Late Formative Periods.

Although the La Vega site also has Catamayo Phase A contexts, based on the examination of ceramics from a systematic surface collection in 1980, Guffroy (2004:37) notes that Catamayo Phase A sherds were very scarce (5% were Phase B, 35% Phase C, and 60% Phase D). The CR-72 sherd from La Vega sampled for charred residues originated from Zone CV, where in Level 5 a quadrangular structure (known as structure

2, also in Zone CIV) was excavated, and all of the associated material “corresponds to

Catamayo Phase B” (Guffroy 2004:55). A small hearth was present in the interior of the structure, and a radiocarbon date of 2900 ± 60 BP is reported for “carbonized plants” associated with the hearth (Guffroy 2004:56-57). Recalibrating the date with OxCal 4.1

(Bronk Ramsey 2009), using IntCal09 calibration database (Reimer et al. 2009), rounded to the nearest 10 years, results in 1210 to 1000 Cal BC (1σ) and 1290 to 910 Cal BC (2σ), which is consistent with Catamayo Phase B. An AMS radiocarbon assay on the CR-72

(Beta-312077, ca. 1190 to 1000 Cal BC, 2σ) ceramic charred residue sample from Zone

291

CV, Levels 6-7, shows that the sherd is contemporaneous with the Catamayo Phase B quadrangular structure 2 and hearth. These dates are consistent with the Middle (1430-

830 Cal BC) to Late Formative (1300-300 Cal BC) Periods (Zeidler 2008:460).

In conclusion, although the Catamayo Phase A context for Trapichillo was not confirmed by the dates obtained on a dated ceramic charred residue sample, it seems likely, based on the date (and especially the secure, interior vessel context) previously reported by Guffroy, that Trapichillo has a (terminal) Early Formative Period occupation.

This would indicate that while the CR-69 Beta-312070 may be precise, it is not an accurate reflection of Catamayo Phase A. In contrast, the other dates obtained on ceramic charred residues for Trapichillo (CR-70) and La Vega (CR-72) are both precise and accurate. It is unfortunate, however, that none of the sherds tested for starch granules date to Catamayo Phase A and the Early Formative Period. As with Chaullabamba, I suggest the Catamayo sites (especially Trapichillo) should be reinvestigated to confirm the Early

Formative Period designation.

7.3.3 Assessing Highland verses Coastal Formative Period Dates: Potential Sources of Error.

As I presented in Chapter 6, section 6.9, none of the highland archaeological site ceramic charred residue AMS radiocarbon dates are earlier than the Middle to Late

Formative Period. I also mentioned that one may argue that all of the dates are consistent with the Late Formative Period, as only the CR-89 Chaullabamba sherd date (Beta-

312072) barely breaches the Middle to Late Formative Period overlap (and thus may be considered as dating to the Middle Formative Period). Because of the reasons stated in 292

sections 7.3.1 and 7.3.2, however, it does not necessarily follow that that the few ceramic charred residue AMS dates obtained for this dissertation overthrow previous radiocarbon dates for Chaullabamba and the Catamayo sites (especially).

As noted previously, all organic samples, be they wood or other carbonized plant remains, ceramic charred residues, faunal or human bone, or shell, are subject to potential contamination of older and/or younger carbon. Only careful interpretation (and detailed publication) of radiocarbon determinations within the context of the site and region can overcome sample-by-sample biases. A similar statement may be made for (and errors may be compounded by) fractionation, whereby the lighter carbon isotopes (12C and 13C) are taken up preferentially, as compared to 14C, and fixed in the tissues of living organisms. Not all radiocarbon dates are corrected for isotopic fractionation, and, in some cases where they are, a standard value is used (rather than corrected by the actual δ13C of the sample material) (Bowman 1990:20-22). Even these two examples of potential errors may make comparing dates problematic if not considered. Relating radiocarbon dates obtained 30, 20, or even 10 years ago to more recent determinations may be biased by differences within and between radiocarbon laboratory procedures over time, refinements in radiocarbon measurement methods (and precision), and the use of different calibration curves. All of these factors make comparing radiocarbon dates between highland archaeological sites, and among highland and coastal archaeological sites, in Ecuador problematic (as for any other world region).

One of the aims of this dissertation was to assess the timing of highland Formative

Period food production, especially during the Early Formative Period, and in this regard

293

the dates obtained fail to confirm that any of my results (excluding those from the Santa

Ana-La Florida site) are contemporaneous with the Coastal Early Formative Period. This eliminates the possibility that I can assess the nature of Early Formative Period highland food production and cultural interaction, and make comparisons with similar developments on the coast, based on my data alone. I am still troubled, however, with the notion that the Highland Formative Period is delayed in comparison to the coast. In this regard, as discussed above, Trapichillo (and probably other sites in the Catamayo Valley) and Chaullabamba probably have terminal Early Formative Period occupations, although perhaps the dating is somewhat ambiguous for Chaullabamba. It is important to remember that Cotocollao in the Quito Valley, a site where I was unable to obtain reliable samples for analysis, has an Early Formative Period occupation, as discussed in Chapter

3, section 3.2.2. The Cotocollao village was not an ephemeral occupation but lasted over

1000 years (Villalba 1988:40), beginning about 1800 Cal BC and ending by 400 Cal BC

(by the Pululahua volcanic eruption). Therefore, Cotocollao overlaps with the terminal

Early Formative, Middle Formative, and Late Formative Periods. For consistency in comparing the Early Formative Period dates for Cotocollao to the other highland sites discussed in this chapter, Table 7.2 shows the contexts for the seven earliest radiocarbon dates (from 55 dates in total) reported by Villalba (1988:242-243) and recalibrated with

OxCal 4.1. As the recalibrated dates in Table 7.2 show, GX-4768, GX-7210, and GX-

8323 all fall within the Early Formative Period, concurrent with Valdivia 6-8b (2100 to

1450 Cal BC, at 1σ) (Zeidler 2003:519). As such, botanical remains associated with these contexts must also be considered, along with the results obtained from starch analyses

294

and complementary data from other sites, in order to have a balanced view of highland

Formative Period plant use.

Table 7.2. Earliest Reported Radiocarbon Dates for the Cotocollao Site: Contexts and Recalibration.

Measured Age1 Calibrated2 Age Calibrated2 Age Lab number Cut Provenience Context (RCYBP) BC (1σ) BC (2σ) GX-4768 F23 op. 1 DP post hole 3495 ± 210 ca. 2130 to 1530 ca. 2470 to 1390 GX-7210 F50 Dpiso floor hearth 3340 ± 135 ca. 1870 to 1450 ca. 2020 to 1310 GX-8323 F6 op. 2 B-D floor hearth 3310 ± 150 ca. 1860 to 1420 ca. 2010 to 1260 GX-4766 F19 5C15 cemetary 3135 ± 165 ca. 1620 to 1130 ca. 1870 to 930 GX-7207 F37 BC14 post hole 3100 ± 140 ca. 1520 to 1130 ca. 1690 to 970 GX-7206 F37 C10 floor hearth 3030 ± 135 ca. 1440 to 1050 ca. 1610 to 910 GX-5057 F23 op. 1 1-5B5/B6 habitation floor 3000 ± 150 ca. 1420 to 1040 ca. 1610 to 840 1As reported in Villalba (1988:242-243). Villalba does not indicate whether these dates were corrected for isotopic fractionation. 2Calibrated dates determined withOxCal 4.1 (Bronk Ramsey 2009), using IntCal09 calibration database (Reimer et al. 2009), rounded to the nearest 10 years. Multiple calibration ranges resulted and so the end-limits of the oldest and youngest ranges are given as a single range, preceeded by "ca."

Although not all of the contexts from which botanical remains (phytoliths, charred macroremains, and pollen) were recovered are from the Early Formative Period contexts, some are. Charred remains of chocho/tarwi (Lupinus mutabilis) and two types of

Phaseolus seeds (probably two varieties of common bean) were from the same context as

the GX-8323 sample, a hearth, and therefore likely date to ca. 2010 to 1260 Cal BC (2σ)

(Villalba 1988:333). Similarly, charred macroremains of chocho/tarwi and one of the

Phaseolus types were also identified from the same context as the GX-7210 hearth, which dates to ca. 2020 to 1310 Cal BC. Pollen grains identified as Zea mays, Solanum tuberosum (domesticated potato), Solanum spp., Chenopodium quinoa (quinoa), Lupinus mutabilis, and Oxalis tuberosa (oca) were recovered from the floor sediment of the structure dated by the GX-4768 post-hole (Villalba 1988:340), ca. 2470 to 1390 Cal BC.

295

I assume that, because both genera and species level identification are reported, the pollen analyst was able to differentiate domesticated from wild/weedy taxa. The differentiation of C. quinoa pollen from C. hircinum, and C. petiolare, as well as amaranths (see section 7.4.1.2 below) is important and we must assume that the 152 pollen grains (Villalba 1988:340) allowed for a definitive identification of quinoa.

Therefore, these data show that domesticated species of beans, chocho/tarwi, quinoa, oca, maize and potato were present at the Cotocollao site by at least the terminal Early

Formative Period. In addition, charred remains of a monocotolydenous tuber, identified as “probably achira (Canna edulis)” (Villalba 1988:333) are from the same context as the

GX-7207 and GX-7206 samples, and likely date to ca. 1690 to 910 Cal BC, showing that this lowland domesticated species was also present by at least the Late Formative Period.

In summary, it is difficult to compare absolute dates from different sites and regions because not all dates are reported with the same degree of detail. Compounding the situation is that, as inevitable, radiocarbon dates have been obtained from different labs over the past 30 years or so, making disparities between analytical and calibration procedures likely. In an attempt to resolve these and other biases, I recalibrated the measured radiocarbon dates for Chaullabamba, the Trapichillo and La Vega sites in the

Catamayo Valley, as well as the Cotocollao site, showing that Chaullabamba, Trapichillo and Cotocollao all have Early Formative Period components. Although the starch granule analysis results realized for this dissertation may not date earlier than the Middle to Late

Formative Period, the botanical evidence from Cotocollao is important, as will be discussed below, as it show that a suite of domesticated crops are present by at least the

296

terminal Early Formative Period in the highlands. Irrespective of the potential problems in comparing radiocarbon dates within and between sites (and regions), the preponderance of radiocarbon dates does seem to indicate that the Formative Period in the highlands was delayed in comparison to the coast (as discussed in Chapter 3). While archaeological investigations in the highlands (and other regions) have seriously lagged behind those for the coast, and so much more work is needed before final judgements can be pronounced, I think it important that renewed investigations at key highland sites and regions, such as Chaullabamba (and the Cuenca and nearby valleys), the Catamayo

Valley, and Cotocollao-related sites in the Quito region, are undertaken to specifically address chronological issues of the Ecuadorian highland Formative Period.

7.4 Paleoethnobotany of the Highlands and Eastern Lowlands during the Ecuadorian Formative Period

One of the major aims of the starch granule analysis was to determine the types of starchy plants that were utilized at the highland, and the one eastern lowland, archaeological sites investigated for this dissertation. Table 7.2 shows a compilation of the starch granule types from the sites investigated. With respect to absolute counts, it is clearly apparent that maize starch granules (even if the cf. Zea mays identifications are excluded) represent the overwhelming majority of starch granules observed, representing

44% (57.9% if cf. Zea mays identifications included) of the overall assemblage, and is the only starch granule type identified from all samples (100% ubiquity). As cautioned previously, however, these data do not allow one to infer the absolute or relative abundance (in this regard I mean the original volumes of plant matter) of one species 297

over another, as there are reasons that starch granules can be over or underrepresented in assemblages. Further, these biases are not static in relation to each other and likely vary between samples, sample types, and contexts. For example, do maize starch granules always preserve by a set factor (1.5x, 2x, or what-have-you) more than starch granules of other species, from different sample types, different contexts, etc.?

Similarly, the absence of starch granules, even from species that produce abundant starch in the plant parts utilized, is not necessarily “evidence of absence.” The number of samples analyzed from La Chimba, for instance, was less than the number from Cerro

Narrío (eight verses 16, respectively), and an even greater discrepancy occurred in the number of samples that produced positive starch results (4/8 from La Chimba verses

12/16 from Cerro Narrío) and this factor alone may have skewed the types and numbers of starch granules recovered. In any event, these issues are reviewed in section 7.2.4 above, and need not be repeated here, but the abundance of maize starch granules as compared to other types does not mean that maize was being cultivated, processed and consumed in far greater amounts than other starchy plant species. As a consequence, I consider the most important information realized from the starch analysis results to be the types of plant species identified and what these data can reveal with respect to food production strategies and long-distance cultural interaction.

298

Table 7.3. Starch Granule Analysis Results from all Archaeological Sites Investigated.

) ll ) s (a u l l e s o g le r e n / a s u r e l m n a l t a ta . w l a a u h n o a a c l g . . e p d r u i p P l p n c r u n p . . . s k r G i p a p p f. p u p p t e s s c s n te h n h u ri p p . p a y u (c s s s p s / s c e p C T a e a u r m e e p m s d l a L S a t a s o m y e t 1 2 u r a e o t m c 1 6 1 1 c o e n s u r a g S i c o i b a a h e e e e s c ih a l n e m c l s a m n r o a p p p p p o b a la e Z a m r t y y y y a i o a a x a a o a p o h f. e t Archaeological Site T T T T C D F I M M O S T c Z D S T Cerro Narrío 5 45 43 54 61 192 36 18 454 Chaullabamba 2 9 10 27 6 2 50 1 3 110 299 La Chimba 12 26 14 88 9 11 160 La Vega 35 5 12 11 16 29 70 10 2 190 Santa Ana - La Florida 141 49 1 2 8 6 3 34 15 22 25 118 341 22 7 794 Tajamar 43 8 20 7 59 153 16 5 311 Trapichillo 4 5 13 20 6 39 3 2 92 Totals by type 225851 2 206 9034512290932528993397482111 Ubiquity 71.4 85.7 14.3 14.3 28.6 14.3 85.7 14.3 42.9 14.3 42.9 57.1 14.3 100 100 Relative Frequency 10.7 4.0 0.0 0.1 0.9 0.3 4.3 1.6 2.4 1.0 4.3 4.4 1.2 13.7 44.2 4.6 2.3

299

7.4.1 Fine-tuning the Starch Granule Taxonomic Classifications

In order to further discuss issues of food production and cultural interaction, however, it is important to examine the plant identifications made by starch analysis and assess if the classifications can be enhanced by considering complementary data. As discussed in section 7.4.1 above and in Chapter 6, in addition to comparison to a reference collection, the starch granule types recovered and very conservatively identified were, in almost all cases, consistent in every way with published description and micrographs that indicate that most of the starch granules can be identified as domesticated species. Here I will provide further support.

In addition to the plant identifications realized from starch analysis, the results of botanical analyses from the Cotocollao site are highly important for three reasons. First, as shown in Table 7.2 and discussed in section 7.3.3 above, an examination of the earliest absolute dates shows that Cotocollao has components that date firmly within the terminal

Early Formative Period. Second, botanical evidence from Cotocollao can be correlated to not only one, but all three of the earliest dates for Cotocollao. The third, and possibly most important, point, in so far as corroborating the results of the starch granule analysis, is that both macrobotanical and pollen analysis show that domesticated, and not wild, species of beans, chocho/tarwi, quinoa, oca, maize and potato were being cultivated and utilized at Cotocollao during the terminal Early Formative Period (ca. 2470 to 1260 Cal

BC).

In Chapter 6, and as discussed above in section 7.2.3, I employed a very conservative assemblage-based approach to identifying starch granules to different taxa

300

in my analysis. Due to the high diversity of native plant species present in Ecuador, in most cases, I identified the unknown starch granule types to the genus level only, despite that the starch granule in question was consistent with my reference collection for a domesticated species as well as published descriptions. Also, it is important to highlight that the published descriptions referred to in Chapter 6 are based on extensive comparison of the starch granule morphology of domesticated species to related species, genera, and families. Only one of the starch granule types was restricted to family-level identification (Fabaceae). Although I cannot be certain, and as discussed in Chapter 6, section 6.2, it is most likely that all of the Fabaceae starch granules identified are from one or more species in the Phaseolus genus. As noted in Chapter 6, all of the starch granule types identified were consistent in morphology and other characteristics for domesticated species (excluding Theobroma, where I do not have specimens for all possible species, and previous descriptions are few). The low number of Fabaceae starch granules recovered from some samples decreased the confidence for making the species- level identification. Genus-level classifications were assigned to starch granules identified as Dioscorea, Ipomoea, Maranta, Oxalis, Solanum, and Theobroma. The

Capsicum spp. identification does not indicate that the chili peppers that contributed starch granules to the assemblages from SALF and La Vega were native (wild) species, as discussed in Chapter 6, section 6.2, but that one or more of the domesticated species

(C. baccatum, C. chinense, and/or C. pubescens) may be represented. Zea mays and

Manihot esculenta were also identified to species. As discussed above, however, multiple lines of evidence (from macrobotanical and pollen analysis) show that domesticated plant

301

species were present at Cotocollao. As the Cotocollao site potentially has the earliest dates in the highlands (see sections 7.3.1, 7.3.2, and 7.3.3, and Tables 7.1, 7.2, above, and

Chapter 3), and has evidence for domesticated plants dating to the earliest occupation, the

Cotocollao botanical record can be used to enhance the starch granule identifications.

This means that the starch granule types identified as Fabaceae (cf. Phaseolus) are likely domesticated Phaseolus spp. cf. vulgaris (common bean) and/or lunatus (lima bean), the

Oxalis spp. starch granules are likely domesticated Oxalis tuberosa (oca), and the

Solanum spp. identifications are probably domesticated Solanum tuberosum (potato). Due to preservation issues, as discussed in Chapter 5, macrobotanical remains are rare (unless charred) in Ecuador. However, because Peru and Chile have a much drier climate, especially along the desert coast, botanical evidence from archaeological sites in these regions can be used to further support starch granule species-level (for almost all taxa) identifications.

The earliest record for domesticated Phaseolus in South America comes from

Chilca, along the southern coast of Peru, where P. lunatus (lima bean) desiccated pods were directly (AMS) dated to ca. 4490 to 4360 Cal BC, 2σ (Kaplan and Lynch

1999:266). The earliest macrobotanical occurrence of P. vulgaris (common bean) is from

Guiterrero Cave in the central highlands of Peru, where desiccated seeds were directly dated by AMS C14 to ca. 3030 to 2890 Cal BC (Kaplan and Lynch 1999:265). While these are the earliest, minimum dates for macrobotanical remains of domesticated

Phaseolus, they are not isolated examples, and numerous other directly dated specimens in Peru and Chile range in age from ca. 1880 to 385 Cal BC (as well as specimens dating

302

into the common era) (Kaplan and Lynch 1999:255-256). The earliest microbotanical evidence for Phaseolus comes from the Ñanchoc Valley in northern (western slopes of coastal) Peru, where Piperno and Dillehay (2008:19623) argue that starch granules recovered from human teeth calculus samples (with associated, as well as direct, dates on human bone that range from ca. 6260 to 5020 Cal BC, 2σ) are from Phaseolus seeds, probably P. lunatus. Therefore, the occurrence of domesticated Phaseolus over a broad region, encompassing northern coastal Chile, the coast and highlands of Peru, to the northern Ecuadorian highlands (at Cotocollao), at archaeological sites that predate, or are concurrent with, the samples from SALF, La Vega, Trapichillo, Chaullabamba, Cerro

Narrío, and Tajamar, strongly support the conclusion that the starch granules identified as

Fabaceae cf. Phaseolus are from domesticated P. vulgaris or P. lunatus. However, both

Phaseolus augusti and Phaseolus polyanthus are wild relatives of domesticated

Phaseolus and are still cultivated and consumed today in Ecuador (Rios et al. 2007:461).

Therefore, while I believe the starch granules are from Phaseolus spp., I cannot be certain of which species, or if a combination of species, are represented in the starch granule assemblages.

Evidence supporting that the Solanum spp. starch granules identified in samples from Chaullabamba, Cerro Narrío, Tajamar and La Chimba are likely S. tuberosum

(domesticated potato) also comes from coastal Peru. Domesticated potato macroremains, dating to as early as 2250 BC (therefore, roughly contemporaneous with Cotocollao), have been securely identified from the Casma Valley (Ugent 1994; Ugent et al. 1982) on the Peruvian coast, as discussed above. The Casma Valley evidence for domesticated

303

Manihot esculenta (Ugent 1994; Ugent et al. 1986) and Ipomoea batatas (Ugent 1994;

Ugent et al. 1981) also support the starch granule identifications of domesticated manioc/yuca at SALF, Trapichillo and La Vega, and increase the confidence that the starches identified as Ipomoea spp. at SALF are from domesticated (I. batatas) sweet potatoes as well. Starch granules recovered from charred ceramic residues, that were directly dated to 3350 to 3010 Cal BC (2σ), were also identified as domesticated manioc, arrowroot (Maranta arundinacea), chili peppers, and maize, (supported by phytoliths of arrowroot and maize from the same samples) from the coastal Ecuadorian Early

Formative site of Loma Alta (Zarrillo et al. 2008) and Real Alto (Chandler-Ezell et al.

2006). Therefore, based on starch granule evidence from directly-dated ceramic charred residues and phytolith evidence from stone tools dated by association to similar contexts, domesticated arrowroot is also present on coastal Ecuador by at least 3000 Cal BC.

For the most part, I have restricted the complementary evidence to directly-dated macrobotanical specimens to show that domesticated species of Phaseolus, Oxalis,

Solanum, Manihot, Ipomoea, and Maranta are the likely sources of the starch granules recovered from the samples analyzed for this dissertation. There is also abundant microbotanical evidence from Ecuador as well, as reviewed in Chapter 3 and discussed above, to support domesticated species presence at archaeological sites that predate or are concurrent with the sites investigated by starch granule analysis. With respect to maize, as reviewed in Chapter 3, microbotanical evidence from coastal and highland Ecuador

(alone) also show that its presence in Ecuador predates or is concurrent with the Early

Formative (SALF, Cotocollao) to Late Formative (La Vega, Trapichillo, Chaullabamba,

304

Cerro Narrío, Tajamar and La Chimba – based the charred ceramic dates alone) occupations of these sites. It is generally accepted, based on archaeological, phytogeographical, and genetic studies (and other techniques), that maize was first brought under domestication in southwestern Mexico, perhaps as early as 9000-7000 BP, and wild populations of maize’s closest relatives (the teosinte grasses) are not present in

South America (e.g., Matsuoka et al. 2002; Piperno and Flannery 2001). Therefore any occurrences of maize in South America are representative of domesticated Zea mays ssp. mays, and offer insight into the timing and route of dispersal of this important crop from

Mexico into South America. Importantly, numerous macrobotanical maize remains from

Paredones and Huaca Prieta, located on the north coast of Peru, have recently been directly-dated. The earliest dates range from ca. 4825 to 3371 Cal BC (2σ) (Grobman et al. 2012), unequivocally establishing maize’s presence in South America, in close geographical proximity to Ecuador, by the onset of the Early Formative Period (4400 to

3800 Cal BC) (Zeidler 2003:519).

I cannot say with absolute certainty that starch granules from wild species are not present in the samples I analyzed. Indeed, wild species of Dioscorea, as well as

Theobroma and/or Herrania, may have contributed all or some of the starch granules identified to these taxa. Wild species of both Dioscorea and Theobroma are used in

Amazonia today (e.g., Cabral Velho et al. 1990; Chu and Figueiredo-Ribeiro 1991). As well, in a study examining the use and management of wild edible plants in southern

Ecuador, Van den Eynden (2004a) found that not only were 354 species used (none of the starch granule taxa I identified are present), but that in certain areas (where little natural

305

vegetation remains), such as the dry central zone of Loja province and the high western

Andes, many wild plants are managed without necessarily being cultivated or domesticated. Active wild plant resource management by farmers has likely occurred for centuries (and probably much longer) within agricultural areas, explaining why they have not, like many other species and much of the natural vegetation, been extirpated (see

Chapter 2, section 2.4, discussion on Ecuador’s vegetation). Further, these actively- managed resources not only have local economic value, but are also present for all ecological zones (mainly determined by elevation) (Van den Eynden 2004a:126-132).

However, based especially on pollen and macrobotanical evidence from Cotocollao, many domesticated crop species were present in the highlands by the terminal Early

Formative Period, supporting my contention that most of the taxa identified through starch analysis are probably domesticates. I have also presented further evidence showing the antiquity of domesticated plants in the region, to support that domesticated plant species are likely represented in the starch granule assemblages for the sites under consideration.

The botanical data from Cotocollao, coupled with the record from Santa Ana-La

Florida, are critically important in understanding: 1) inter-regional cultural interaction, because of their early dates; and 2) the development of Formative Period plant food production in both the Ecuadorian highlands and eastern lowlands. In order to appreciate both of these phenomena, I will first examine the ecological requirements for the plant species identified, their regions of domestication, and their nutritional attributes (section

7.4.2). From these results I will show (section 7.4.3) that the plant species identified, and

306

their regions of origin, suggest extra-local cultural interaction occurred between the highlands and regions to the south, particularly through the Inter-Andean corridor, and not primarily with coastal Ecuadorian groups. I will also argue that the nutritional complementarities of the plant species identified, as discussed in section 7.4.4, suggest knowledge gained over a long period of time. The nature of agricultural strategies, as shown by the types of species present, at the different archaeological sites (and between broad regions), is discussed in more detail in section 7.5.

7.4.2 The Geographical Origins, Ecological Requirements, and Nutritional Aspects of the Identified Plants.

The variety of plant species identified through starch analysis and by previous botanical research for the highlands and eastern lowlands of Ecuador, allows me to discuss aspects related to human adaptation, nutrition, and cultural interaction.

Adaptation to the varied environments present within and between the major ecological divisions – coast, highlands, and eastern lowlands – speak to the knowledge different groups possessed of their local natural environment, and how to best manage these environments. Learning what combinations of foods result in a healthy diet, especially if the diet is low in meat protein, also suggests a long period of knowledge-acquisition and development. Therefore, I propose that such detailed awareness must have been gained over a considerable period of time, allowing for complex agricultural production systems to be archaeologically evident by the terminal Early Formative Period in the highlands, and earlier in the eastern lowlands at SALF. The following discussion presents the plant species identified for the highlands and eastern lowlands, their geographical origins of 307

domestication and ecological requirements, and some aspects of their nutritional benefits.

Table 7.4 shows the botanical taxa identified for the highlands and the eastern and western slopes, as well as the cultural periods they are associated with. Although I include the identification of tree fruit and unknown tuber macrobotanical remains in

Table 7.4, I do not discuss them, or the cf. Canna edulis (achira), in detail below. Table

7.5 lists the elevational range of the plant species and their proposed regions of domestication. In Table 7.5 I have placed the domesticated species first, followed by possible con-generic species. Figure 7.2 shows a map illustrating the proposed regions of domestication for the plant species, based on the most current evidence, as discussed below for each of the species. Figure 7.2 also exemplifies why South America was classified as a “noncenter” (Harlan 1971), as opposed to a “center” (Vavilov 1951), of plant domestication. By observing and documenting the geographical patterns of diversity among crops around the world, Vavilov (1951:20) proposed “eight independent centers of origin of the world’s most important cultivated plants,” with the centers being the core areas (areas of highest diversity) where plants were domesticated and dispersed from. Harlan (1971:174), though, suggested that centers of species diversity do not necessarily represent centers of domestication. A “noncenter” is where plant domestication took place over a geographically broad area, with multiple domestication events occurring in multiple areas.

I also prepared topographic and 3-Dimensional perspective maps with ArcGIS1 showing the elevations present within a 5 km catchment area for each of the sites. These are shown in Appendix A, and will be discussed in more detail, including why a 5 km

308

catchment area size was determined to be appropriate, in section 7.5 below. Villalba

(1988:326, Figure 167) also shows an approximate 6 km resource catchment area for

Cotocollao and indicates that all of the crop species identified could have been grown in the immediate vicinity of the site.

309

Table 7.4. Plant Taxa Identified from the Ecuadorian Highlands and Eastern and Western Slopes

) ) se a n ) o ) nts de m in s a ) u u me p. su rb ) ro cirp a sp m q S r b sa be u r Frag m ) tu t ube lus spp.) ero odi o Pi . iu o b m p / t/T p p p se . u u o d o a inus mays ulenta en spp. sp ssy sp c pp is t a Rin Ph s l lan Ch t m Go ea Lup Zea t es a So Cyperus m i u ( r e ( o ta ( ( ro Cultural Period for ic n m h n (Ox e b Fru u ize ( ra g known Ro ps Canna edulis osco a ani ca Site Taxa Identified Ca cf. Cotto Di Leg Lupine ( M M Ma O Potato Quinoa ( Sed Theo Tree Un Northern Highlands Cotocollao Early Formative C C, P C, P C, Ph, P P P P C C Middle/Late Formative C C C C, Ph C C La Chimba Late Formative C C C C, S C, S C, S C Integration Period C C C C C Lake San Pablo Early Formative Ph, P Tajamar Late Formative S S S S Lake Yambo Late Formative P

Southern Highlands 310 Cerro Narrío1 Late Formative S S S S Chaullabamba Middle/Late Formative S S S S Laguna Chorreras2 Preceramic P Late Formative P Pirincay3 Late Formative Ph C, Ph

Far Southern Highlands La Vega Middle/Late Formative S S S S Trapichillo Late Formative S S S

Eastern Slopes Santa Ana - La Florida Early Formative S S S S S S S Lake Ayauchi Early Formative Ph, P

Western Slopes Nueva Era Late Formative C C C C C = Carbonized plant remains, P = Pollen grains, Ph = Phytoliths, S = Starch Data compiled from Athens 1998:161-164; Bruhns 2003:156; Bush et al. 1989:303-305; Colinvaux et al. 1988:95; Pearsall 2003a:217-232; Piperno and Pearsall 1998:243-261; Villalba 1988:330- 340. 1Pearsall and Piperno (1990:332) report "cob fragments recovered from the Cerro Narrio site, dated to ca. 2000 B.C.". Only the Late Formative starch results are shown. 2Hansen et al. (2003:102, 106) report "a single grain of Zea mays was found at the 7000 cal. yr BP level" but becomes more frequent after 4000 cal. yr. BP. 3Although the botanical data for Pirincay are reported here, the dating and contexts for the botanical data are not reported in sufficient detail (Bruhns 2003:156). 310

Table 7.5. Elevation Range and Region of Domestication of Plant Species

Species Proposed Region(s) of Domestication Elevation Range1 Capsicum baccatum Lowland Bolivia 0-1000 C. chinense N. Amazon Basin 0-500 C. pubescens E. Andean Slopes 1500-3000 C. lycianthoides Native Andean 1000-3000

Chenopodium quinoa Native Andean / southern Andes 2000-3500 C. hircinum Native Andean 3000-3500 C. petiolare Native Andean 1500-3500

Dioscorea trifida N. Amazon Basin 0-500 D. chimborazensis Endemic to Chimborazo 1000-2000 D. coriacea Native Andean 0-500 and 3000-3500 D. crotalariifolia Native Andean 0-500 D. lehmannii Native Andean 1000-1500 and 2000-2500 D. megacarpa Native Amazonian 0-500 D. nicolasensis Native Amazonian 0-500 D. pilosiuscula Native Andean 1000-2500 D. samydea Native Amazonian 0-500 D. sulcata Native Andean 1000-1500

Ipomoea batatas NW South America 0-2000

Lupinus mutabilis Native Andean / Northwest slopes or southern 800-3500 highlands, Peru

Manihot esculenta Southern margin of Amazon Basin 0-2300

Maranta arundinacea N. Amazon Basin 0-1000

Oxalis tuberosa Bolivia / NW Argentina 2500-4000

Phaseolus lunatus Western Andes N Peru / Ecuador 0-3000* Phaseolus vulgaris "B" Northern Andes 0-3000* Phaseolus vulgaris "nuñas, popping-type" Mid-elevation S. Andes 1800-3000* Phaseolus vulgaris "pole type" Mid-elevation S. Andes 1000-3000* Phaseolus augusti Native Andean 2500-3000 Phaseolus polyanthus Native Andean 1000-2500

Solanum tuberosum Central Andes 0-1000 and 2500-4500

Theobroma cacao SE Ecuador/NE Peru 0-1000 T. bicolor Cultivated Amazonian 0-1000 T. glaucum Native Amazonian 0-500 T. grandiflorum Cultivated Amazonian 0-500 T. subincanum Native Coastal and Amazonia 0-1000

Herrania cuatrecasana Native Amazonian 0-500 H. dugandii Native Amazonian 0-500 H. kofanorum Native Amazonian 0-500 H. mariae or H. nycterodendron Native Andean and Amazonia 0-1000/1500 H. nitida Native Amazonian 0-500

Zea mays SW Mexico 0-3100 1Data compiled from Jørgensen and León-Yánez 1999:409-411, 438-439, 480-481, 558, 776-778, 901-902, 909-918, 920- 921; National Research Council 1989:189, 198-199; Sauer 1993:59. *Jørgensen and León-Yánez (1999:480) report the elevation distributions as 0-500 masl. The upper limit reported seems far too low. I suspect this is due to the tendancy to grow Phaseolus spp. at lower elevations, while introduced legumes, including peas (Pisum sativum, 2500-3000 masl) and especially broad/fava beans (Vicia faba, 2000-3500 masl) are grown in the highlands today (Jørgensen and León-Yánez 1999:480, 483). Therefore, Jørgensen and León-Yánez report the modern distribution for the Andean domesticated Phaseolus beans. Hargrave (2006a, 2006b, 2006c) and Sauer (1993:59) report elevation tolerances as ranging from 0-3000 masl, but that nuñas especially grow best at 2500 masl.

311

Figure 7.2. Proposed regions of domestication for plant species discussed in the text. These areas are based on phytogeography and/or genetic analyses. The specific areas delineated may be somewhat larger or smaller than how they are represented. In addition, only maize is shown for Mesoamerica, but the region was an independent center of domestication for some plants shown for South America, such as sweet potato, Chenopodium, and chili peppers, as well as others not shown, such as squashes, in addition to other plants not present in South America (such as sunflowers).

7.4.2.1 Chili Peppers: Capsicum spp. (Solanaceae)

There are five domesticated species of Capsicum including C. annuum, C.

frutescens, C. chinense, C. baccatum, and C. pubescens. Tables 7.5 shows the native

Capsicum species reported for Ecuador (including wild species) and their current elevation ranges, and Figure 7.2 shows the proposed areas of domestication. Today chili

312

peppers are cultivated and used as vegetables, ornamental plants, medicine, and as a spice, in both temperate and tropical areas (Clement et al. 2010:84). Aside from adding flavour to enliven an otherwise bland diet of starchy root crops, beans, and maize, chili peppers are also rich in Vitamin A and C (Coe 1994:60-62; USDA 2011). When consumed with plant sources of iron (such as beans), iron uptake is increased because of the presence of Vitamin C (Zarrillo et al. 2008:5009). While chilies did not provide protein or major calories to the diet of Pre-Columbian people, they were (Coe 1994:62), and still are, very significant in Latin American cuisine.

Although the genus likely originated in Bolivia (Macleod et al. 1982), the exact domestication regions and routes of dispersal for the five economically important domesticated chilies are speculative. Phytogeography, cytogenetics and molecular studies, and archaeological evidence have led to probable regions of domestication for these species, however. Three regions are thought to be independent centers of Capsicum domestication: Mesoamerica; the tropical (Amazonian) lowlands of South America; and the Andes (Pickersgill 2007:935). C. annuum, C. chinense and C. frutescens form a complex and may have arisen from closely related and widely distributed wild and weedy species (Heiser 1995; Pickersgill 1988; Walsh and Hoot 2001:1414-1418), but their postulated regions of domestication are not the same. C. annuum is the most widely cultivated species on a world-wide scale and contains a diverse array of sweet and hot types (Hancock 2004:238), (including bell, jalapeño and cayenne) and was probably first domesticated in northern Mesoamerica (Pickersgill 1971:684-688; 1988:383), more precisely in upland east-central Mexico (Loaiza-Figueroa et al. 1989:185-188). C. chinense (cumari, marupi, habañero, scotch bonnet), C. frutescens (cayenne, tabasco), C.

313

baccatum (girl’s finger, chili, ají), and C. pubescens (rocoto) are primarily used as spices

(Coe 1994: 60-62; Hancock 2004:238), and their centers of origins are not as clear

(Clement et al. 2010:84-85). C. chinense was likely domesticated in northern lowland

Amazonia, where the greatest variability is seen, C. frutescens in the Caribbean,

Mesoamerica and/or northern Amazonia, C. baccatum in lowland Bolivia, and C. pubescens in the mid-elevation Andes (Pickersgill 1988:383; Pickersgill and Heiser

1977:821-823; Piperno and Pearsall 1998:152-154). C. pubescens is the most strongly isolated from the other species and is also the most cold-tolerant, occurring along the

Andean chain from Bolivia to Colombia (Hancock 2004:239; Loaiza-Figueroa et al.

1989; Pickersgill and Heiser 1977:822).

Capsicum spp. (domesticated) starch granules recovered from sediments, charred ceramic residues, and flaked and groundstone tool residues, document the use and geographic extent of chili peppers at archaeological sites from Peru, Ecuador, Brazil,

Bolivia, Panama, and the Caribbean dating back to ca. 4400 to 5300 Cal BP (Dickau et al.

2012; Duncan et al. 2009; Perry et al. 2007; Zarrillo et al. 2008), with the earliest dates derived from Loma Alta and Real Alto (Early Formative). Macroremains of both C. baccatum and C. chinense fruits that date to about 4000 yr. BP were recovered from

Huaca Prieta and Punta Grande on the Peruvian coast (Pickersgill 1969). Therefore, although it may be tempting to presume that the Capsicum spp. starch granules recovered from La Vega and SALF (Table 7.4) are C. pubescens based on region of probable domestication, the presence of C. baccatum and C. chinense on the Peruvian coast (and on the other side of the Andes from their proposed regions of domestication), and approximately concurrently at La Vega (but post-dating SALF), precludes making such a

314

conclusion. Consequently, as noted previously, any (and possibly more than one) of the domesticated Capsicum spp. may be represented in the starch assemblages from La Vega and SALF, although C. annuum, and C. frutescens can most likely be excluded based on regions of domestication. C. annuum is grown today in Ecuador between 0-3000 masl, but was likely dispersed by Europeans from Mexico to South America, and C. frutescens is reportedly not grown today in Ecuador (Jørgensen and León-Yánez 1999:901-902).

Based on current elevational distributions (cultivation practices), as shown in Table 7.5,

C. chinense is grown at elevations below 500 masl, while C. pubescence is grown today between 1500 and 3000 masl (Jørgensen and León-Yánez 1999:901-902) and neither La

Vega (ca. 1200 masl) nor SALF (ca. 1040 masl) is represented by current cultivation distributions. Furthermore, C. chinense is better adapted to hot and humid conditions

(Pickersgill 2007:930), and while the Catamayo Valley may be warm, it is not humid. C. pubescence – rocoto – certainly has a greater toleration for colder temperatures, and was likely domesticated within the region encompassing both the Catamayo Valley and

SALF. However, it does not tolerate the heat well, such as at SALF and La Vega, and so it may not be the prime candidate (elevations greater than 1500 masl are located within a

5 km catchment area for both La Vega and SALF, though; see section 7.4.3 below and

Appendix A). A more likely candidate, despite originating from the south in lowland

Bolivia, may be C. baccatum, as, despite being primarily a lowland species, it is grown today up to 1100 masl (National Research Council 1989:198-199). C. baccatum, ají, is usually very hot, but sweet varieties are also available and used, and it is widely popular in the Andes. It is used especially to make a sauce, also called ají (National Research

Council 1989:198-199), usually together with other ingredients, such as onions, tamarillo,

315

cilantro and/or chocho (ají recipes vary between regions), that is used as a condiment and to marinate uncooked seafood (shrimps, scallops, fish, etc.) in a dish called ceviche. As such, it may be that the chilies grown at La Vega and SALF (Tables 7.3 and 7.4) may be either locally (Andean) domesticated C. pubescence, C. chinense (possibly grown at lower elevations and imported, although the climate may have been warmer and wetter in the past, as discussed in Chapter 4), or, more likely (based on growing requirements), the exotic C. baccatum, with the latter indicating some form of interaction with groups to the south prior to its appearance in SE Ecuador by at least the Early Formative Period.

7.4.2.2 Quinoa: Chenopodium quinoa (Chenopodiaceae)

Chenopods were independently domesticated in the eastern United States (C. berlandieri ssp. jonesianum), Mesoamerica (C. berlandieri ssp. nuttalliae), and the

Andes (C. quinoa and C. pallidicaule) (Bruno 2006:32; Smith 2006). While chenopods remain an important part of the Andean diet today, they are only a minor crop in Mexico

(mostly used as a green vegetable), and the eastern North American domesticate became extinct in prehistoric times (Bruno 2006:32; Pickersgill and Heiser 1997:809; Smith

2006). It is ironic, then, that less is known about the Andean domesticated chenopods (C. quinoa and C. pallidicaule) than their less used and extinct North American cousins. The principal pseudo-cereal (it is not a true cereal as it is not a grass, but is used much like a cereal) grown in the Andes is quinoa (C. quinoa) and was probably domesticated in the southern Andes from wild C. hircinum (ajara) (Wilson 1990:98-99). Although C. hircinum occurs in Ecuador today (Table 7.5), it is not cultivated but occurs as an agricultural weed – and thus it was probably dispersed along with quinoa as a

316

“monophyletic, possibly co-evolving unit” (Ruas et al. 1999:26), as first suggested by

Wilson (1990:98-99). Wilson (1990:99) relates that farmers recognize and remove ajara from quinoa fields before they flower, but that the plants, which produce black seeds instead of the pale seeds seen in quinoa, can escape detection to produce progeny that are then sown with the next crop. Archaeological studies at Chiripa also suggest that weeding and careful seed selection occurred in the past (Bruno and Whitehead 2003:350-355). As noted previously (section 7.3.3 above, and Chapter 3 Table 3.2) C. quinoa pollen has been identified from terminal Early Formative deposits at Cotocollao (Table 7.4) and it is assumed that the species-specific identification is correct. Cañihua/kañawa (C. pallicaule) is a cold and drought tolerant semi-domesticated chenopod grown in Peru above 3600 masl, where few crops can survive (Gade 1970; Pickersgill and Heiser

1997:809), and is not grown in Ecuador today (and not likely to have been in the past as, being a wild/weedy species, it would likely still remain there as a feral variety).

Like Capsicum and the other domesticated species considered, the timing of quinoa’s domestication and routes of dispersal mainly rely on archaeological evidence.

Bruno (2006:42-43) securely identified domesticated quinoa seeds at Chiripa (southern

Lake Titicaca basin, Bolivia) that were subsequently directly dated by AMS to 1500 BC

(ca. 3500 BP), establishing a minimum date for quinoa cultivation at the site. While this is the earliest directly-dated quinoa reported, earlier associated dates for quinoa are present in Peru. Thin-testa large-size domesticated quinoa seeds (either C. quinoa and/or C. pallidicaule) were present in all levels at Panaulauca Cave (central Andes, Junín puna), dated by association to as early as 3000 BC, and were used as food (i.e., they were not present in animal dung or puna grass-mat samples analyzed) (Pearsall 1988:104-105,

317

1989:320-322). A single carbonized chenopod seed, possibly C. quinoa, was recovered from hearth contents dated to between ca. 8000 and 7500 Cal BP at site CA-09-27 in the

Ñanchoc Valley (north coast of Peru) (Dillehay et al. 2007:1891). More abundant (and thus more securely identified) desiccated and carbonized chenopod seeds (argued to be C. quinoa), were recovered from Tierra Blanca sites in the same region, dating to 7000 to

4500 Cal BP (Dillehay et al. 2007:1891). Bruno and Whitehead (2003) and Bruno (2006) suggest that C. quinoa was domesticated in the Lake Titicaca and Lake Junín regions of southern Peru (and Wilson also places the domestication origin in southern Peru, as noted earlier; see also Figure 7.2) by at least 5000 – 4000 Cal BP. From these data, then, we might conclude that domesticated C. quinoa dispersed quite quickly after its initial domestication in southern Peru (or was domesticated earlier than the archaeological record presently establishes), being established in highland Ecuador at Cotocollao (Table

7.4) by the terminal Early Formative Period (ca. 4420 to 3340 Cal BP). This indicates that cultural interaction was occurring within the Andean corridor, between groups in

Ecuador and Peru, leading to transmission of this crop from the southern Peruvian highlands where it was domesticated, to the north.

As mentioned earlier, quinoa is still an important food in the Andes today (Figure

7.3), and is also gaining popularity world-wide. It is made into flour and baked or steamed as various types of bread, used as a breakfast cereal, fermented to make beer

(chicha) used in soups and stews, or cooked and consumed like rice (Cieza de León

1864:143-144; Coe 1994:181-182; National Research Council 1989:149). Most varieties contain bitter-tasting saponins in the outer seed-coat layers, and, therefore, the seeds either need to be milled to remove the seed coat or washed to remove the saponins prior

318

to consumption. The effort needed would have been worthwhile for prehispanic populations, though, as the seeds are high in protein (more than twice that of common cereal grains) and contain a better amino acid balance than maize, where lysine is absent and protein is minimal. Moreover, in addition to lysine, the other essential amino acids methionine and cystine are present in quinoa, but are deficient in beans. Quinoa is also an excellent source of carbohydrates as the seeds are composed of about 58 to 68% starch

(National Research Council 1989:153) (the starch granules are very small and unlikely to be identified by standard light microscopy).

Figure 7.3. Quinoa (Chenopodium quinoa) growing in highland Ecuador. Near Riobamba, July 2009.

The bitterness of unwashed quinoa seeds makes them resistant to insect and rodent infestation during storage (Dillehay et al. 2007:1891), which is an added benefit of this crop. Quinoa is grown in Ecuador today at elevations between 2000 and 3500 masl

(Table 7.5) and so its presence at Cotocollao (ca. 2700 masl) is not surprising. Villalba’s

(1988:326, Figure 167) resource use map for Cotocollao, the single highland site where

319

quinoa has been identified, shows that the crop could have been grown in the immediate area of the site. Future macrobotanical (flotation) analysis of sediments at other highland archaeological sites, within the elevation limits of quinoa cultivation, will likely result in further identifications of quinoa.

7.4.2.3 Yams: Dioscorea spp. (Dioscoreaceae)

As shown in Table 7.5, in addition to the domesticated American yam (Dioscorea trifida), nine other Dioscorea species are present in Ecuador east of the Andes below

2500 masl, and 33 species are found in Ecuador in total. As discussed in Chapter 6, section 6.2, I limited the identification of starch granules to the genus level because of an absence of information on whether any of these species, other than the domesticated yam, were cultivated and/or collected and consumed (although none are presently reported to be), and also because I did not have comparative samples for any of the other species.

Although it may be likely that the starch granules from SALF (Tables 7.3 and 7.4) originated from domesticated yam, I am reluctant to make a species identification based on the few (n=6) starch granules recovered (Table 6.1, Chapter 6, and Table 7.4), mainly because I still cannot exclude the nine species present in Ecuador east of the Andes.

Piperno (2006b:62) also states that “yam [starch] grains... are genus specific.”

Little is known about the origin, diversity, genetics and phylogeny of domesticated yam (Dioscorea trifida), despite that it remains an important crop in lowland regions

(Bousalem et al. 2006:440; Piperno and Pearsall 1998:163). Based on phytogeography and ecology of Dioscorea trifida and related species, Piperno and Pearsall (1998:163) suggest that Dioscorea trifida was first domesticated in the seasonally dry, low elevation

320

tropical forests covering southern Venezuela, southern Guiana, and northern Brazil

(Piperno and Pearsall 59-61) (see Table 7.5 and Figure 7.2). Many domesticated crops are polyploids, with three or more sets of chromosomes (as opposed to most animals and plants which are diploid, with two sets of chromosomes), and polyploidy can make identifying the wild ancestors of a domesticate very complex (Doebley 1992:209-211).

Polyploidy can occur naturally, however, so there is no clear universal link between polyploidy and domestication (Emshwiller 2006:157). So little is known regarding the genetics of Dioscorea and related species, that only recently has the base chromosome number of Dioscorea trifida (x=20) been determined, making it a tetrasome (4n) with 80 chromosomes (it was previously assumed that it was an octoploid, 8n) (Bousalem et al.

2006:449). Bousalem and colleagues (2006:449) also found that cultivars collected and tested from diverse regions and ethnic groups in French Guyana showed no significant genetic variation nor did they show a complex of different polyploid levels, as is often the case for species with high chromosome numbers propagated clonally, suggesting stability in the cultivated pool. This, then, may support the hypothesis that Dioscorea trifida was indeed domesticated along the northern margins of the Amazon Basin. Starch granules identified as originating from domesticated Dioscorea trifida have been recovered from

Panamanian archaeological contexts at Aguadulce dating to ca. 3300 BC (Piperno

2006b:62; Piperno et al. 2000) and at Zapotal between 2500 and 1800 BC (Dickau

2005:298-299; 2010:113; Dickau et al. 2007:3653-3656). These data suggest that

Dioscorea trifida was domesticated and being dispersed by at least ca. 3300 BC (i.e., approximately contemporaneous with SALF).

321

Like arrowroot (Maranta arundinacea), achira (Canna edulis), cocoyam

(Xanthosoma spp.), and llerén (Calathea spp.), Dioscorea trifida is a monocot, its USO is high in starch, and it is propagated by simply replanting a portion of its root (Calathea requires a complete rootstock) (Piperno and Pearsall 1998:110-112, 115). Yams are also high in potassium and vitamin A (USDA 2011). Its cultivation distribution in Ecuador is noted to be below 500 masl (Table 7.5). Therefore, as SALF is located at ca. 1040 masl, the crop was either grown at a lower altitude and brought to the site, or a suitably warm niche was located near SALF. Paleoenvironmental reconstruction (Chapter 4, section 4.2) also showed that the climate may have been warmer than it is today, and the lowest elevation within a 5 km catchment area for SALF is 880 masl (Appendix A), so it may have been locally grown. Irrespective of this, if the starch granules are from domesticated yam, its presence at SALF indicates that the crop was dispersed to SE Ecuador (from the northern Amazon Basin margins) by at least 3500 to 3300 Cal BC, based on the AMS radiocarbon dates shown in Tables 6.1 and 6.8 (Chapter 6).

7.4.2.4 Sweet Potato: Ipomoea batatas (Convolvulaceae)

Unlike yams, arrowroot, achira, cocoyam, and llerén, sweet potatoes are dicots (as is manioc) and are propagated by stem cuttings (Piperno and Pearsall 1998:120). It is a vine-like herb, and where both manioc and sweet potato are present today, sweet potato assumes a secondary role to manioc (Piperno and Pearsall 1998:126). Little is known about the geographical and botanical origins of sweet potato, mainly because these issues have been overshadowed in the past by the greater interest in sweet potatoes’ antiquity and introduction into the Pacific (Pearsall and Piperno 1998:126; Pickersgill and Heiser

322

1977:817-818). In possibly the most definitive genetic study to date, Roullier and colleagues (2011) analysed sweet potato landraces (329 accessions) from Mexico to Peru for both chloroplast and nuclear microsatellite markers. Both types of markers support that two geographically restricted genepools are present, one in northwestern South

America and one in the Caribbean and Central American region. The divergence between these two genepools is probably ancient, predating domestication events. Previous studies have favoured a single domestication event in the lowland tropics east of the Andes (see

Pickersgill and Heiser 1977:816-817), possibly in the northern margins of the Amazon

Basin from the wild diploid I. trifida (see Piperno and Pearsall 1998:126-128). Genetic analyses, however, indicate at least two independent domestication events (probably from wild tetraploid I. batatas, and not the diploid I. trifida) – one in the lowlands of

Central/Caribbean America and one in the lowlands of northwestern South America, with the domesticated crops being dispersed from these regions across tropical America

(Roullier et al. 2011:3972-3975).

Archaeological evidence, as discussed previously, indicates that domesticated I. batatas is present in the Casma Valley of coastal Peru by 2250 to 295 B.C. (Ugent 1994;

Ugent et al. 1981). These dates are more recent than the earliest dates for SALF (the only site where sweet potatoes were identified by starch analysis – Tables 7.3 and 7.5), and

SALF is within the now proposed region of domestication for I. batatas. In Figure 7.2, I only show the proposed South American region of sweet potato domestication (and not the Central/Caribbean America location). As shown in Table 7.5, sweet potatoes are grown in Ecuador from sea level to 2000 m, and this range encompasses the elevation of

SALF, as also shown in Appendix A for SALF’s 5 km catchment area. While sweet

323

potatoes, like all the root crops, are deficient in protein, they are high in easily digestible carbohydrates (35%), relatively high in calcium and vitamin C, and are an outstanding source of Vitamin A (USDA 2011). They are also easy to grow, and require minimal preparation (baking, roasting, boiling) (Piperno and Pearsall 1998:120); they were a favorite baby food for all my children.

7.4.2.5 Chocho/Tarwi: Lupinus mutabilis (Fabaceae)

Chocho, as it is known in Ecuador (tarwi in Peru), is an overlooked crop (Figure

7.4). Not only does it have potential to add greatly to international markets (it is now becoming popular in Europe and Australia) (Eastwood and Hughes 2008:373; National

Research Council 1989:181), it has received little attention archaeologically, despite that it was, and remains, an important crop in the mid to high elevation Andes from

Bolivia/northern Chile to Venezuela. The Old World domesticate Lupinus albus (the

Mediterranean “lupini” bean) is listed as one of the crops that produces no to few phytoliths of taxonomic significance (Piperno 2006a:48, Table 2), and Korstanje and

Babot (2007:64) report that only non-diagnostic phytoliths are found in the seeds of

Lupinus mutabilis and none in the pod. While Korstanje and Babot (2007:64) characterize the seeds starch granules, I do not feel that these are sufficiently diagnostic

(spherical to oval rounded granules ranging from 8 to 40μ in size) to enable a species identification until more research is conducted. Therefore, in the absence of microbotanical markers, identifying chocho from archaeological contexts to trace its domestication origin has relied on macrobotanical remains and pollen analysis. As such,

324

questions such as when, where, how many times, and from what progenitor(s) Lupinus mutabilis was domesticated are largely unknown.

The first hypothesis on Lupinus mutabilis origins saw that it arose as a hybrid between L. douglasii and L. ornatus (two North American species), with seeds then transported to South America by humans (Kazimierski and Nowacki 1961, as cited in

Eastwood and Hughes 2008:3373-3374). This scenario has been rejected because it is based on plant morphological characters that are shared by many western New World

Lupinus species, and recent phylogenetic studies support that L. mutabilis is part of a large group of closely-related species, all of which are restricted to the Andes (Hughes and Eastwood 2008; Drummond 2008). The second, more recent hypothesis suggests that

Lupinus mutabilis was domesticated near Cuzco, Peru, because purportedly wild forms of

Lupinus mutabilis are found there and there is a close morphological similarity between

L. mutabilis and L. praestabilis (Blanco 1982, 1984, 1986; Tapia and Vargas 1982). More recently, Eastwood and Hughes (2008) question that L. mutabilis was domesticated from

L. praestabilis, and that its origin was the southern Peruvian Andes. They point out that

L. praestabilis shows some striking morphological differences from L. mutabilis and that the greatest difficulty in understanding the origins of domesticated species is hampered by “uncertainty about species delimitation and the chaotic taxonomy of the Andean species” (Eastwood and Hughes 2008:374). Indeed, in the Catalogue of Vascular Plants of Ecuador, Jørgensen and León-Yánez (1999:476) state that, although many Lupinus species have been described for Ecuador, they believe that most of these are “dubious, pending a much-needed revision of Lupinus for all of Andean South America.” Hughes and Eastwood (2006) have revised the 480 Andean Lupinus species to a new estimate of

325

85 species (and their geographical boundaries) based on field and herbarium work. They have now combined these results with molecular phylogenetic analysis to conclude that:

1) L. mutabilis does not occur in natural vegetation in Bolivia, Peru and Ecuador, is only known in cultivation, and, occasionally, immediately after in cultivated fields left to fallow; 2) of four possible progenitor species, L. piurensis is the most similar (based on morphological characters of leaves, flowers and fruits), being almost impossible to separate from L. mutabilis without examining the fruiting material; and 3) four species

(including L. piurensis) group together with L. mutabilis based on genetic analysis, and, of these, L. piurensis is genetically the most (although the resolution is not strong) similar to L. mutabilis (Eastwood and Hughes 2008:374-375). Based on these results, they suggest that L. mutabilis was domesticated only once, and that, if L. piurensis is its progenitor, then it was likely domesticated on the western slopes of the Andes in northern

Peru, between 1650 and 3300 masl (Eastwood and Hughes 2008:375). While Eastwood and Hughes (2008:377) are suitably cautious in their interpretations and conclusions, they suggest that further genetic testing should focus on L. piurensis as the most likely progenitor. Despite that further work is certainly required, I feel that the revision of the

Lupinus species, coupled with the genetic analysis, makes it more likely that the proposed domestication origin is northwestern Peru rather than the southern Peruvian Andes.

However, in both Table 7.5 and in Figure 7.2 I leave the latter region as a possibility.

As noted previously, (Chapter 3, Table 3.2, section 7.3.3 above, and Table 7.4) carbonized macroremains and pollen identified as Lupinus mutabilis date to at least 2010 to 1260 Cal BC (2σ, recalibrated date) at Cotocollao, and Lupinus spp. were recovered from the Late Formative through Integration Period deposits at La Chimba. It is actually

326

based on the results from Cotocollao that Eastwood and Hughes (2008:377, Table 2) propose that L. mutabilis was domesticated around that time period.

Figure 7.4. Lupinus mutabilis growing near Quito, Ecuador. Picture taken May, 2010.

Another problem in determining the history of domesticated L. mutabilis is that macrobotanical remains are often identified as simply Lupinus spp., or simply subsumed under “Legume” (e.g., Pearsall 2003:230-231; Hastorf 1993). This is not a criticism of previous paleobotanical work, but is likely due to prior uncertainty of Lupinus taxonomy and in defining secure seed morphological characteristics for identification. I think that L. mutabilis was probably domesticated much earlier, and that, with a revised understanding of L. mutabilis seed size morphology and variation, coupled with a concerted effort to

327

recover macrobotanical remains from archaeological sites and more secure identifications, further (and earlier) evidence will come to light.

As a food crop, chocho is exceedingly useful. Chocho is exceptionally nutritious in that it is high in protein (about 40-50%), including the vital amino acids lysine and cystine (as mentioned previously, maize is deficient in lysine), and is high in oil (about

20%), which is lacking in Phaseolus (National Research Council 1989:181). However, chocho seeds only contain about 25-30% of the required amino acid methionine, but, as mentioned earlier, quinoa does contain adequate methionine. The oil is also relatively rich in unsaturated fatty acids, including the nutritionally vital linoleic acid. Chocho, then, is an important source of fat where animal meat does not make a large contribution to overall diet and other plant foods are starch-rich and/or lacking in complementary amino and fatty acids (such as Phaseolus). As a “pioneer” species, chocho can be grown on marginal soils, and, as a leguminous plant, its strong taproot not only fixes nitrogen but also loosens the soil, both of which lead to soil improvement. It is often planted at the end of crop rotation cycles for these reasons (Hastorf 1993:115; National Research

Council 1998:186). In addition to being adaptable to poor soils, chocho is also tolerant of drought, frost, and many pests, and Hastorf (1993:115) notes that it requires no weeding.

One of the only drawbacks of chocho is that the seed coats contain alkaloids that make them inedible without processing, but the alkaloids are water soluble, and can be relatively easily washed out (Hastorf 1993:115-116; National Research Council

1998:186). Hastorf (1993:116) explains that, once mature, the stalks are cut off close to the ground, the plants are then beaten against a flat surface to remove the seeds, the seeds are winnowed from the pods, and the seeds are then dried in the sun and stored in sacks

328

or ceramic jars. Thus, like quinoa, chocho seeds are naturally resistant to insect and pest infestation during storage. Before being consumed, the seeds are leached of the alkaloids

“in cold running water for approximately ten days before they can be consumed raw, roasted, or boiled” (Hastorf 1993:116). The fact that the seeds can be consumed raw or roasted is also a benefit at high altitudes where water boils at a much lower temperature.

Indeed, based on the excellent nutritional profile, relative ease of cultivation and processing, soil improvement benefits, and good storability, it is surprising that more chocho has not been identified in the archaeological record. Hopefully, as mentioned previously, this situation may change if the paucity of past identifications has been more a consequence of uncertainty in taxonomic morphological characteristics than a real absence of chocho in the archaeological record. Villalba’s (1988:326, Figure 167) resource use map for Cotocollao shows that the chocho (Figure 7.4) could have been grown in the immediate area, and Appendix A shows the same for La Chimba.

Importantly, if domesticated Lupinus mutabilis was the species present at Cotocollao, this indicates some type of cultural interaction (direct or, more likely, indirect) between

Ecuadorian highland groups and groups located along the northern Peruvian coastal slopes where the most recent genetic analysis indicates it was domesticated.

7.4.2.6 Manioc: Manihot esculenta (Euphorbiaceae)

As Isendahl (2011:452) points out, with a worldwide annual production of 184 million tons, which feeds about half a billion people in Latin America, Africa, and Asia, manioc is the most important crop today that originated from the lowland Neotropics. On a worldwide scale, manioc is the sixth largest food crop produced (Clement et al.

329

2010:76). Historical and ethnographic data show that manioc was a major staple and key component of Neotropical subsistence and political economies (Isendahl 2011:452).

Despite this importance, manioc has been extremely difficult to identify archaeologically.

While plant remains do not preserve well in Neotropical contexts, identifying the presence of manioc has proven to be a particularly intractable case. Manioc, when it does flower, produces pollen that is carried by insects, and so it is less represented in sediments than wind-pollinated species (Piperno and Pearsall 1998:37). In addition, although diagnostic phytoliths of manioc’s root secretory cells have been defined

(Chandler Ezell et al. 2006), they are neither abundant nor robust. Finally, as manioc is grown by vegetative propagation, seeds (when they are produced) have little opportunity to enter the archaeological record, and charred macroremains from cooking accidents or discard are rare and difficult to identify (Lentz 1999:11; Piperno and Pearsall 1998:33).

Therefore, as with all plants that were potentially used in the past, and especially those that leave few vestiges of their former presence, the dearth of manioc in the archaeological record cannot be used to indicate a deficiency in past use, nor, when it is present, its relative importance to overall diet. In recent years, starch analyses of archaeological samples have added greatly to documenting the past use of manioc, and, importantly, genetic analyses have been used to reveal the relationship between Manihot species and the origins of M. esculenta.

Christian Isendahl (2011) provides a recent and thorough review of the domestication and early dispersal of manioc, and so any treatment I can provide here will be shallow in comparison. The details that are important to the present discussion are the region of manioc’s domestication, the timing of its dispersal, and its ecological

330

requirements. With respect to the first point, several genetic studies have revealed that M. esculenta was domesticated only once, from Manihot esculenta ssp. flabellifolia in the transition zone between the lowland rainforest and the savannah scrub vegetation of the southwestern Amazon Basin (Allem 1999; Carvalho and Schaal 2001; Olsen and Schaal

1999, 2001, 2006; Olsen 2002; Roa et al 1997), as shown in Table 7.5 and illustrated in

Figure 7.2. In sections 7.2.3 and 7.4.1, I reviewed macrobotanical evidence for manioc remains from the Peruvian coast that place the crop there by 2250 BC, and starch granule evidence from charred ceramic residues that place manioc at Loma Alta on the

Ecuadorian coast by 3350 to 3010 BC. Earlier evidence also comes from the Zaña Valley of the northern Peruvian coast where a tuber skin fragment, dated by associated wood charcoal to 7000 BC, was recovered from Quebrada de Las Pircas and identified by starch analysis as domesticated manioc (Piperno and Pearsall 1998:207-208; Dillehay et al. 2007:1892; Piperno and Dillehay 2008; Rossen et al. 1996:395, 400). These results show that domesticated manioc had been dispersed from its region of origin and was being cultivated in the general region of SALF and the Catamayo Valley (Tables 7.3 and

7.4) well prior to the evidence presented in this dissertation.

Manioc is a woody shrub that can grow to heights of 1-4 meters (see Figure 4.13,

Chapter 4). A single shrub can produce five to ten tuberous roots, some as large as 1 m in length and 2 kg in weight. It is undemanding to grow, and very productive (Isendahl

2011:454). The tuberous roots are very high in starch (energy) and potassium (USDA

2011), which is important in hot climates where dehydration can be an issue. It is usually propagated by stem cuttings, which produces a more vigorous plant that sprouts much more quickly than seed propagation (Alves 2002:67), and ensures that the plants will be

331

true to its parent (Ng and Ng 2002:168). The principal limiting factor in manioc cultivation is that the plant does not tolerate excessive soil moisture (Sauer 1993:58), but agricultural techniques such as ridging and constructing drainage canals can overcome this limitation in most areas. Manioc is also extremely tolerant of pests, nutrient-poor acidic soils, and periods of drought (it can also remain in the ground for long periods of time before being harvested, as a means of “storage”), but does not tolerate frost (Lathrap

1970:49). It does best where temperature extremes do not exceed 16° to 38°C (Alves

2002:79-80). Therefore, as shown in Table 7.3, its distribution is limited to areas below

2300 masl (Sauer 1993:59). These, as well as other aspects of manioc, may explain why it became, and remains, a major staple in lowland Neotropical regions (Piperno and

Pearsall 1998:124), and a major crop worldwide today, despite the toxicity of some varieties. Aside from being deficient in protein, all varieties of manioc contain cyanogenic glucosides, that, when the tuber tissue is damaged, are hydrolyzed to hydrocyanic acid (also known as prussic acid) (Carneiro 2000:67; Wilson 2003:404).

Manioc varieties are often classified as “sweet” (Figure 7.5) and “bitter” based on the relative amount of the cyanogenic glucosides present. The sweet varieties can be eaten without additional processing, either boiled (Figure 7.5) or roasted, while the bitter varieties must be processed by grating or mashing to remove the juice, washing/rinsing/straining, and drying to produce flour (which can be stored) before being consumed. The dichotomy between sweet and bitter varieties is probably oversimplified, as a continuum exists in the relative amounts of cyanogenic glucosides present in the numerous manioc varieties (e.g., Balyejusa Kizito et al. 2007), but today the most toxic varieties are primarily grown in Amazonia and the Antilles in areas where they form the

332

major staple, whereas in areas where maize is dominant, sweet manioc is grown (Hawkes

1989:486; Piperno and Pearsall 1998:124; Sauer 1950:508; Wilson 2003). While I cannot determine whether sweet or bitter manioc (or a variety of these) was being grown at

SALF and in the Catamayo Valley, bitter manioc is not grown in Ecuador today. The ecological requirements of manioc, coupled with its modern cultivation in the Catamayo

Valley and around SALF, demonstrate that the crop could have been grown within the 5 km catchment areas for these sites (Appendix A). The domestication origin of Manihot esculenta in the southern margins of the Amazon Basin, and its presence on the Peruvian and Ecuadorian coast prior to the earliest dates for both SALF and the Catamayo Valley, indicate that the crop had already been dispersed to the region and it is likely that it was being grown earlier in, at least, the Ecuadorian lowlands, than SALF attests to.

333

Figure 7.5. Meal of freshly harvested boiled sweet manioc, fresh papaya, wild tomato salad, and, the ever present, introduced rice. Huaorani village, Napo river region, Ecuador, March 2010.

7.4.2.7 Arrowroot: Maranta arundinacea (Marantaceae)

Arrowroot is another crop whose domestication is shrouded in mystery. There is even debate about whether or not it should be considered a domestic species. The other

Marantaceae species cultivated for its edible tubers is Calathea spp., particularly C. allouia (llerén) (Piperno and Pearsall 1998:115). Maranta arundinacea (the cultivated arrowroot) is considered indigenous to northern South America and the Caribbean

334

(Purseglove 1972), and has been reported growing wild in Brazil, northern South

America, and perhaps Central America (Sturtevant 2009:184), where it has been observed in Panama (Piperno and Pearsall 1998:115). Therefore, based on the phytogeography of the wild populations Maranta arundinacea and related species, an origin of “cultivation” for the species, as shown in Figure 7.2, is proposed to be the northern Amazon Basin, but suitable seasonally dry habitats (leading to the formation of large subterranean storage organs) are also located west of the Andes on the Ecuadorian coast, the southern Amazon Basin, and in parts of Central America (Pearsall 2008:105-

106, and Figure 7.1; Piperno and Pearsall 1998:115, 164-165). Arrowroot starch granules and/or phytoliths have been recovered from Real Alto (Chandler-Ezell et al. 2006) and

Loma Alta (Zarrillo et al. 2008) on the Ecuadorian coast dating to the Early Formative

Period.

Arrowroot rhizomes are high in starch, and the plant is propagated by simply replanting a rhizome tip (Piperno and Pearsall 1998:110-112, 115), but the rhizomes are covered by tough “scales” and so thorough maceration or grinding is required to release the starch (Piperno and Pearsall 1998:115). Although the plant requires high rainfall

(1500-2000 mm per year) and grows best on sandy loams, it, like manioc, cannot tolerate excessive soil moisture (Piperno and Pearsall 1998:115). While the rhizomes are low in protein, Maranta arundinacea starch is very easily digestible and a good source of potassium, and the plant has many medicinal uses (Sturtevant 2009:185-189), which may have led to the importance of this crop in lowland tropical agricultural systems – as a weaning/convalescent food and a source of medicine. The ecological conditions found around SALF (ca. 3000 mm annual precipitation, with many well-drained slope areas),

335

then, are suitable for the cultivation of arrowroot (as well as manioc and yams, which have similar requirements), as shown in Table 7.5 and Appendix A. Maranta arundinacea is present in Ecuador today between 0 and 1000 masl. The presence of arrowroot at SALF, indicates that the crop was dispersed to southeastern Ecuador by at least the Early Formative, and to the Ecuadorian coast by at least then as well, if it was first cultivated in the northern Amazon Basin. Arrowroot, may, however, have originated to the west of the Andes, or, perhaps more likely, in more than one area.

7.4.2.8 Oca: Oxalis tuberosa (Oxalidaceae)

As a root crop, oca is only overshadowed in importance by the potato (Solanum tuberosum) in the Andean highlands. Unlike the potato, however, which has been distributed world-wide to become the fourth largest crop (National Research Council

1989:83), oca is not well-known outside of its native home. Like Dioscorea, discussed previously, and potatoes, to be discussed below, oca is a polyploid crop. Oxalis tuberosa originated from morphologically similar species, informally known as the “Oxalis tuberosa alliance”, which are found through the central and northern Andes (Emshwiller

2002). Most of these alliance species are diploids that do not form tubers, but four tuber- bearing wild populations have been found in four areas from northwestern Argentina to central Peru (Emshwiller et al. 2009:1839). Based on the most recent genetic analysis, two of the four wild populations are the most likely genome donors of domesticated

Oxalis tuberosa. These are O. chicligastensis from northwestern Argentina, and an as yet unnamed population from the eastern Andean slopes of Bolivia, designated BolW/T

(Emshwiller et al. 2009:1846-1847). These regions are designated in Table 7.5, although

336

I have retained an overlap into SE Peru for Figure 7.2. While further genetic analysis is required to understand the complexity of Oxalis tuberosa development, what is clear is:

1) that the crop was first domesticated much to the south of Ecuador; and 2) that the wild

Oxalis species in Ecuador do not produce tubers, nor did they contribute to oca’s domestication. These data indicate, then, that cultural interaction, allowing for oca to be transmitted north to Ecuador, was most likely occurring within the Inter-Andean corridor.

Oca is easy to propagate (vegetatively) and is very tolerant of poor soils and harsh climates, where it can produce double the yield of potatoes, but is susceptible to heat and bacterial infection when grown in the humid tropical lowland regions (National Research

Council 1989:83, 85). Oxalis tuberosa cultivation in the Andes extends from Venezuela to Argentina and remains a staple for indigenous people living between 3000 and 4000 masl. It is prepared in a number of ways: boiled, steamed, baked/roasted, fried, added to soups and stews, and a few types are eaten raw (National Research Council 1989:83-85).

Some of the “bitter” varieties contain more oxalic acid than the “sweet” varieties, but the tubers are almost always placed in the sun for a few days during which time they become sweet, and the amount of glucose can almost double. The bitterest varieties are almost always converted to a dry, storable product (kalla, cavi or caya) by the same freeze- drying chuño process used for potatoes – soaking the tubers in water, repeatedly exposing them to freezing night temperatures, and stomping on them to remove the water (Hastorf

1993:114; National Research Council 1989:86). Because of the great diversity in oca varieties, nutrition levels vary, but in general they are composed of 11-22% carbohydrates and some varieties contain more than 9% protein (with a good balance of amino acids), which is excellent for a root crop (National Research Council 1989:87).

337

Oca tubers also contains more iron and vitamin C (ascorbic acid) than potatoes, but less calories per weight (oca is 70-80% moisture) (Hastorf 1993:113; National Research

Council 1989:87). Apparently the freeze-dried oca product contains more protein, minerals and niacin than potato chuño (Coe 1994:183). These qualities make oca an excellent complement to other high-elevation crops, and so it is not surprising that it had been dispersed to the Ecuadorian highlands by at least the Formative Period. As shown in

Table 7.5, its modern distribution in Ecuador is from 2500 to 4000 masl, and these elevations are found within the 5 km catchment areas (Appendix A) for the sites where oca starch granules, carbonized macroremains and/or pollen (Tables 7.3 and 7.4) are shown it to be present, including La Chimba, Tajamar, Cerro Narrío, Chaullabamba, and

Cotocollao (as shown in Villalba 1988:326, Figure 167).

7.4.2.9 Beans: Phaseolus spp. (Fabaceae)

Species of Phaseolus are grown in Ecuador today up to 3000 masl, as shown in

Table 7.5. Phaseolus vulgaris (the common bean) and Phaseolus lunatus (the lima bean) are domesticated species, while Phaseolus augusti and Phaseolus polyanthus are both cultivated wild relatives. Rios and colleagues (2007:461) report that the Saraguro

(southern highlands, Loja province, ca. 2500 masl) cultivate both P. augusti and P. polyanthus to eat in soups or with maize, and a tea is also made with P. polyanthus leaves and flowers (with flowers of other plants) to treat fevers. As discussed in section 7.4.1, the earliest record for domesticated Phaseolus in South America comes from Chilca, along the southern coast of Peru, where desiccated domesticated P. lunatus pods have been directly dated to ca. 4490 to 4360 Cal BC, while directly-dated macrobotanical

338

remains of P. vulgaris from Guiterrero Cave in the central highlands of Peru show that common beans were domesticated by at least 3030 to 2890 Cal BC (Kaplan and Lynch

1999:265-266). Although charred macroremains of more than one type of Phaseolus species have been identified from Cotocollao and La Chimba (see sections 7.3.3 and

7.4.1 above, Chapter 3, Table 3.2, and Table 7.5), they are not definitively identified to species, although pollen grains identified as P. vulgaris were recovered from the terminal

Early Formative contexts at Cotocollao (Villalba 1988:340). Still, I cannot exclude the possibility that wild species were being cultivated and consumed, even after domesticated

Phaseolus species became available. Therefore, all of the Phaseolus spp. identified by starch analysis from the highland archaeological sites and from SALF on the eastern slopes may be, and likely are, from more than one Phaseolus spp., possibly both domesticated and wild species. The reason these distinctions are important is to assess cultural interaction.

As shown in Figure 7.2, Phaseolus vulgaris was independently domesticated in two different geographic locations. Wild beans of the mid-elevation Andes were domesticated to produce the small-seeded P. vulagris common beans (Columbia, Central America, and

Mexico), which contain the “B” form of phaseolin protein, and large seeded P. vulgaris common beans (southern Andes). Phaseolus lunatus was also independently domesticated in two regions from wild lima beans that grow in the low to mid elevations and range from Argentina to Colombia, resulting in the large seeded lima beans of the western Andes of Ecuador, northern Peru, and the small seeded (sieve type) of Central

America and northern South America (see Chacón et al. 2005; Gepts and Debouk 1991;

Gutiérrez-Salgado et al. 1995; Pearsall 2008:108; Piperno and Pearsall 1998:134-139;

339

Chacón and Pickersgill 2005). In Figure 7.2 I show the Phaseolus regions of domestication for South American only. From the following discussion, and as shown in

Figure 2, what is apparent is that not only were some species of Phaseolus domesticated in, or in very close proximity to, the region of study (highlands and eastern lowlands of

Ecuador), but that wild Phaseolus spp. use cannot be excluded. This makes it impossible to infer extra-local cultural interaction based on the presence of Phaseolus, especially from the starch granule assemblage. As such, the most important information to be discussed regarding the identification of Phaseolus spp. are its cultivation requirements and nutritional aspects.

Most present day Andean people (as likely did their ancestors) depend on beans as an essential source of nutrients, fibre, and protein (National Research Council 1989:173).

As the note for Table 7.5 indicates, European domesticated fava beans (Vicia faba) and peas (Pisum sativum) are the most important beans grown in highland Ecuador today (as also discussed in Chapter 2, section 2.5), with the common and limas now generally grown below 500 masl. Hastorf (1993:118-119) also relates that, when the hardier fava beans and peas arrived in the Andes in the early years of the Spanish conquest, they were adapted to the Andean climate where they yield more per unit area than lupine

(chocho/tarwi) and Phaseolus, and thus they became the most important legumes grown in highland regions. During Cieza de León travels in Ecuador and Peru from 1532 to

1550, he states (in describing the valleys near Quito), “the pulses of Spain grow abundantly, and all other provisions may be had that man requires” (Cieza de León

1864:142). In the past, however, Phaseolus was likely grown up to 3000 masl, and chocho up to 3500 masl, and they were probably the main sources of protein in pre-

340

Columbian diets. Being leguminous plants Phaseolus shares the same soil improvement, nitrogen-fixing, qualities as chocho. In addition to providing the essential amino acids lysine and tryptophan, which are deficient in maize, Phaseolus beans are also a good source of iron, vitamin C, potassium, niacin, and, especially vitamin A (USDA 2011).

Therefore, maize and legumes (including chocho) complement each other and constitute a nutritionally complete food (e.g., Zarrillo et al. 2008).

One of the Phaseolus vulgaris varieties I want to highlight are nuñas. Nuñas are still grown in Ecuador and Peru above 2500 masl, mostly for home consumption

(National Research Council 1989:173). Although nuñas look much like common bean, they have a quality that makes them especially suitable for highland economies – their seed coat is hard. When nuñas are heated in a little oil they burst open after only 5-10 minutes and require no further processing before eating (they are delicious and taste somewhat like peanuts), much like popcorn and toasted maize (tostado, similar to corn nuts). The major benefit of nuñas, as well as for popped and toasted maize, is that they can be prepared and eaten without the need for the long boiling times usually required.

This is especially important at high elevations where water boils at a lower temperature and cooking times are greatly increased, resulting in the greater use of fuel, which may have been difficult to obtain at higher elevations (National Research Council 1989:173).

Irrespective of how beans were prepared, and whether they were native or exotic varieties, the 5 km catchment area maps (Appendix A) for the sites where Phaseolus spp. starch granules, carbonized macroremains and/or pollen have shown it to be present indicate that the crop could have been grown in close proximity to all of the sites, including La Chimba, Tajamar, Cotocollao (as shown in Villalba 1988:326, Figure 167),

341

Cerro Narrío, Chaullabamba, the Catamayo Valley (Trapichillo and La Vega), and SALF

(Table 7.4). I have excluded results from Pirincay, as the botanical identifications made for the site are neither well-reported nor are the contexts clearly defined. It is probably not a coincidence that, through the presence pollen grains, carbonized macroremains, phytoliths, starch granules, or a combination of these, Phaseolus has been identified at almost all (save Nueva Era) of the highland archaeological sites where botanical information is available, as is maize (100% ubiquity).

7.4.2.10 Potato: Solanum tuberosum (Solanaceae)

The potato perhaps epitomizes the ingenuity of indigenous Andean pre-Columbian plant husbandry. Through a geographically vast and long-standing process of human selection, thousands of varieties of Solanum tuberosum were developed that thrive in the multitude of microenvironments that the Andes present. Indigenous farmers developed types not only suited to local environments, but also for preference of other qualities, and it is reported that up to 200 different varieties of “papa”, as it is known in the Andes, may be grown in a single field (National Research Council 1989:93). The reasons that a diversity of potatoes are grown, even in single fields, are that both farmers and researchers report that there is a desire for different types based on flavours and uses, but also as a means to ensure a harvest, even if some varieties planted may be lost by frost and/or pathogens (Hastorf 1993:111; National Research Council 1989:100). Today, individual Andean farmers maintain diversity by selecting for frost tolerance, pest resistance, and taste, and by exchanging seeds (although not stated, I believe true botanical seeds to reduce the transmission of potato virus) through a network established

342

among neighbors and villages (Hastorf 1993:111). Unlike oca, which has seen little worldwide acceptance, potatoes reached Europe by the late 16th century and today are one of the world’s staple carbohydrate sources (National Research Council 1989:103).

Previously, I reviewed (sections 7.2.3 and 7.4.1) the antiquity of potato, showing that macroremains and pollen (section 7.3.3), identified as domesticated S. tuberosum, have been recovered from archaeological contexts in Peru (and in Ecuador at Cotocollao) that predate the starch granule identification of S. tuberosum from La Chimba, Tajamar, Cerro

Narrío, and Chaullabamba. The use of Solanum species is also very ancient, as Ugent

(1997; Ugent et al. 1987) identified wild potato (S. maglia) macroremains at Monte

Verde (Chile), dating to about 12,500 BP. As such, I will focus here on the origin of the domesticated potato, as well as its ecological requirements and nutritional benefits.

Like other polyploid crops, the cultivated potato (ploidy levels include diploid, triploid, tetraploid, and pentaploid), S. tuberosum, has a complicated genetic history.

Andean Solanum tuberosum varieties (landraces) are distributed from northern Argentina in the south to western Venezuela in the north, as noted above, but the progenitor species of Solanum tuberosum have long been disputed. A complex of wild, tuber-bearing taxa, called the S. brevicaule complex, that are found in two regions, the “northern” complex

(southern Peru) and the “southern” complex (Bolivia), are agreed to be the progenitor population(s) from which S. tuberosum originated, either by complex hybridization or by multiple origins from both the southern and northern complexes (Spooner et al.

2005:14694). To clarify this situation, Spooner and colleagues (2005) sampled and genotyped 362 wild tuber and non-tuber bearing relatives of the potato, that included 264 wild species representative of the southern and northern S. brevicaule complex. Their

343

results show that: 1) there was a single domestication for S. tuberosum; and 2) S. tuberosum derived from the “northern” (southern Peru) S. brevicaule complex only

(Spooner et al. 2005:14698-14699). The origin of domesticated S. tuberosum in southern highland Peru is reflected in Table 7.5 and Figure 7.2, and indicates that domesticated potato had dispersed to highland Ecuador through the Inter-Andean corridor by at least the terminal Early Formative Period (Table 7.4).

Figure 7.6. Cañari women harvesting potatoes with a foot plow. Photo courtesy of Diego Castro, 2012.

As mentioned previously, potato varieties suitable for almost all Andean environments were developed. This is reflected in Table 7.5, which shows that Solanum tuberosum is grown in Ecuador today up to 4500 masl (the humid, steep, and undeveloped montaña region is basically the only environment not represented). As

344

shown by the 5 km catchment area maps (Appendix A) for La Chimba, Cotocollao (as shown in Villalba 1988:326, Figure 167), Tajamar, Cerro Narrío, and Chaullabamba, where Solanum tuberosum has been shown to be present (Table 7.4), the crop could have been grown.

Hastorf (1993:112) relates that, because potatoes require more nutrients, they are usually planted just after a fallow phase, even though they have similar production cycles as oca (discussed above) and other Andean tubers (such as ulluco and mashwa). Seed potatoes are planted one pace apart in rows, harvested (often still with a foot plow, Figure

7.6), some five to six months after planting (depending on the variety), sorted by size and stored in a cool, dark place whole or (especially the high elevation types) processed into chuño, as described above, which can be stored almost indefinitely (Hastorf 1993:112-

113; National Research Council 1989:99-100). Potatoes must be cooked (or made into chuño, which is then soaked, cooked in soups and stews, or ground and made into cakes) prior to consumption (boiled, baked, roasted, fried) because they contain toxic glycoalkaloids (Hastorf 1993:112). Potatoes (with the skin on), aside from being high in carbohydrates (starch), are also excellent sources of vitamin C, calcium, and potassium, and chuño yields even higher calcium and calories than whole potatoes (Hastorf

1993:111). The diversity of potato varieties developed, their nutrient profile, and the storage potential of chuño (making them a lightweight yet nutritious “trail” food that is high in energy), demonstrate why the papa is the most economically important crop in the

Andes today (Hastorf 1993:110), and was also probably so in the past as well.

345

7.4.2.11 Cacao/Chocolate: Theobroma spp. / Herrania spp. (Malvaceae)

Linnaeus named the cacao/chocolate (cocoa) tree the “food of the gods”,

Theobroma, to honor the beliefs of its divine origin held by the Aztecs (Clement et al.

2010:78). Due to the importance cacao played in Aztec and Maya economic (serving as a unit of currency), political (an important item of tribute), and ritual life (as part of feasting and other rituals to cement political and social ties) (Coe and Coe 2007:33-106;

Vail 2009:3), and the absence of similar use and importance in South America, until recently, the prevailing view was that Theobroma cacao was domesticated in Central

America by the Maya by ca. 2450 BP, or even as early as 3350 BP based on linguistic evidence (Brown 2009:85-89). However, the earliest botanical studies suggested that the

Central American population of T. cacao was dispersed there from the eastern edges of the Andes in the Upper Amazon, where it originated (Cheesman 1944). In fact, recent genetic analysis has provided strong support for a single domestication of T. cacao in the

Upper Amazon (eastern and southeastern Ecuador/northeastern Peru), and that the

Central American “criollo” type cacao probably originated from a few plants transported by humans from South America (Motamayor et al. 2002, 2003, 2008:2). Therefore, the most recent phylogenetic evidence places T. cacao domestication in the region of Santa

Ana-La Florida (Figure 7.2), where I have tentatively identified Theobroma spp. starch granules from the residues of two stirrup-spout bottles, stone bowls, and a ceramic vessel

(Table 6.1, Chapter 6 and Tables 7.3 and 7.4), all recovered from tombs (Table 5.2,

Chapter 5). As cautioned in Chapter 6, section 6.2, and above in section 7.2.3, the identification of Theobroma and/or Herrania spp. are tentative until further work can be conducted (underway) to confirm or deny the diagnostic specificity of the starch granule

346

morphotype defined in Chapter 6, section 6.2. Table 7.5 shows that: 1) T. cacao and wild

Theobroma species grow within elevation ranges that encompass SALF (ca. 1040 masl, see Appendix A); and, 2) although only one species of Herrania (H. mariae is suspected to be the same species as H. nycterodendron) grows within the elevation ranges found in a 5 km radius around SALF, the distribution of Herrania may have been different in the past, and/or the climate may have been warmer in the past than today (see Chapter 4).

Figure 7.7. Theobroma cacao tree with pod growing in the Santa Ana-La Florida site garden. Picture taken May, 2010.

347

As the identification of Theobroma remains tentative, I will not discuss its ecological requirements and uses at length. Both Theobroma and Herrania are strictly tropical trees, they are limited to lowland areas where rainfall is at least 1200 to 2000 mm per year, they require well-drained deep soils, and they do not tolerate intense sunlight

(being understory trees) (Cobley 1976:208). The flowers, and hence the large fruit (pods) that contain the seeds (often called beans), sprout directly from the bark of the main trunk as well as from older, leafless branches (Figure 7.7). Each pod contains a mass of seeds embedded in a mucilaginous pulp (Figure 7.8). The pulp is very sweet, being composed of almost 50% sugars (Rogez et al. 2004:382), and in South America, the pods are often used just for this purpose – split open, the sweet pulp is sucked off of the seeds, and then the seeds are spat out. I find the pulp has a pleasing flavour and is very refreshing.

Figure 7.8. Theobroma cacao pod split open to show seeds and pulp. Santa Ana-La Florida site garden, May 2010.

348

In Brazil, Theobroma grandiflorum, the capuaçu tree, is grown specifically to use the juicy pulp to flavour ice-cream and make jam, juice and liquor (Rogez et al. 2004:

381). These last two uses – to make a fresh or fermented drink from the pod pulp – may be how Theobroma and Herrania species have always been used in the South American tropics, and may have been the first uses for T. cacao when it was dispersed to Central

America (Henderson and Joyce 2006), This hypothesis is similar to the “maize stalk sugar hypothesis”, which proposes that maize’s (and teosinte’s) initial use may have been for its sweet stalk (as well as its sweet greens), rather than its starchy seeds (Blake

2006:68-69; Smalley and Blake 2003), to make a fermented alcoholic beverage important in facilitating social relations. There is also a very practical reason why plant foods of all sorts may have been made into juice and fermented, that being increased preservation.

Even low alcohol content will prevent the growth of mold (fungi) and other microorganisms and increase storage life (Campbell-Platt 2000:736-739), which is particularly important in the hot and humid lowlands.

In contrast, chocolate, as it is (and was) used in Central America, is made from the seeds and is an elaborate process of fermenting the seeds (with pulp attached), drying, roasting, removing the husks, and grinding the seeds into a paste, which is then boiled with added water and frothed to make the chocolate drink (with spices, especially chilli peppers, and even maize flour added) (Coe and Coe 1996:22-25; Dreiss and Edgar

Greenhill 2008:106). Chemical analysis for theobromine, the main methylxanthines

(alkaloid) found in cacao seeds and pulp, has identified it in absorbed pottery residues from Mesoamerica (Henderson et al. 2007; Powis et al. 2011) and as far north as the

American Southwest (Washburn et al. 2011). The earliest dates for cacao use in

349

Mesoamerica (based on theobromine traces in pottery), ca. 1800 BC (Powis et al. 2007,

2011:8597), and ca. 1300 BC (Henderson et al. 2007:18938), fit in reasonably well with linguistic analysis that indicates the appearance of the proto-Zapotecan word for cacao dating to ca. 1400 BC (Brown 2009:85-89). Henderson and colleagues (2007:1838-

18939) correlate the shapes of the earliest of the theobromine-positive vessels as being consistent with cacao pulp juice or fermentation, and the later vessel shapes to the alcoholic-free chocolate beverage made from fermented seeds, as described above. The earliest dates for cacao (Powis et al. 2007, 2011:8597-8798), however, are associated with theobromine-positive vessel shapes and forms that are suitable for a variety of purposes and storage related to cacao use, and not necessarily with liquid made from cacao pod pulp (whether fermented or not) or a liquid made from cacao seeds. Cacao pod pulp, in addition to being sweet and juicy, is also a source of, in particular, the amino acids asparagine and glutamine, the fatty acids palmitic, oleic and α-linolenic, and the minerals copper, potassium, zinc, magnesium and phosphorus (Rogez et al. 2004:382).

The seed nutrients are similar, but are also very high in iron (Rucker 2009:944-945).

Cacao pulp and seeds are low in vitamins, but when consumed with maize and pigmented fruits and USOs, these foods are complementary and provide a balanced source of both proteins (amino acids) and vitamins (Rucker 2009:944-945). Despite that we may not (as yet) know why T. cacao spread from South America north (i.e., in what way, or ways, was it initially used in both regions), the dates from SALF are sufficiently early to predate cacao’s appearance in Mesoamerica. It is interesting that starch granules of maize, especially, and chili peppers, reminiscent of the Mesoamerican chocolate recipes,

350

were recovered from the same contexts as the possible T. cacao starch granules (Table

6.1, Chapter 6).

7.4.2.12 Maize: Zea mays (Poaceae)

If the potato symbolizes Andean plant husbandry ingenuity, then maize embodies the same for the pre-Columbian Americas as a whole. By the time Europeans arrived in the Americas, maize was being grown as an important staple from Canada to Argentina.

The countless varieties developed for different latitudes, altitudes, day length, maturation length, ear size, seed characteristics, and culinary uses (Hancock 2004:180) are a testament to the changes possible under human selection. While both maize and beans were the most widely-dispersed crops in the Americas (likely because the two crops nutritionally complement each other), only maize attained, almost universally, ritual and symbolic importance. As a result of maize’s dietary, social, political and ideological importance, as well as intense interest in its origin, evolution, and dispersal, it has received more attention than any other plant domesticated in the Americas. As noted previously, it is now almost unanimously accepted that Zea mays ssp. mays was domesticated only once, in the Balsas Valley region of southwestern Mexico, from a single progenitor, the wild teosinte Zea mays ssp. parviglumis, possibly as early as 9000

BP based on the mutation rate of genetic microsatellites (Figure 7.2) (Matsuoka et al.

2002). The timing of maize’s dispersal to South America, its route of dispersal, and its initial uses have also been topics of debate. There is now a convincing body of evidence to support that maize was dispersed from Mexico to Ecuador and Peru by the early mid-

Holocene, where it was initially part of a broad spectrum diet, gaining importance in both

351

diet and ritual use in Ecuador by the Middle to Late Formative Periods (e.g., see

Grobman et al. 2012; Pohl et al. 2007; Zarrillo et al. 2008, for an overview of these issues, and Chapter 3). Maize, along with native Andean and introduced crops, is grown in the highlands today up to its elevational limit of ca. 3100 masl (Table 7.5, Figure 7.9) and remains important in both ritual practice and diet (although today, everywhere in

Ecuador, rice is the most consumed carbohydrate, served in large quantities with almost every meal, including breakfast).

Maize is an annual grass and its vigorous growth and yields are attributed to its C4 photosynthetic pathway (Langer and Hill 1991:118-119). Maize varieties, like other

Andean crops (such as oca and potato), were developed to suite every microclimate

(aside from very dry and/or high elevations) of the Andes, as well as for seed size, colour, taste, use and pathogen resistance (Hastorf 1993:116). Like other crops as well, the different uses of maize result in a variety of types being grown in the same region

(suitably separated by planting time or physical distance to eliminate cross-pollination).

As Hastorf (1993:116-117) describes, maize is planted in furrows and two or three seeds are placed 1 cm apart and covered with soil. The fields require some weeding and hoeing during the six to seven month growth period, after which the entire plant is cut off near the ground, the cobs are removed (and the stalks used for fuel or fodder), and the cobs are braided together and dried in the sun (Figure 7.9). The seeds for the next planting are then removed and the braided, dried cobs are stored (usually hung from rafters). Maize is prepared in various ways depending on the variety. It can be popped or toasted (as described previously), boiled alone or in soups and stews, ground into a coarse meal or flour and prepared in different ways, or sprouted to brew chicha (beer).

352

Figure 7.9. Maize drying near Saraguro, Loja. Photo taken July, 2008.

Maize is present at all of the sites in the region (Table 7.4). As well, and as shown in Appendix A, and for Cotocollao in Villalba (1988:326, Figure 167), maize can be grown within a 5 km area of all of these sites. From a nutritional standpoint maize is high in calories, primarily owing to its high starch content, and fiber. The main dietary limitation of maize is its low protein content, being particularly deficient in the amino acids lysine and tryptophan.

353

7.4.3 Cultural Interaction

As discussed above, the regions of origin for the plant species identified allow for consideration of cultural interaction. It has been suggested that food production and socio-political complexity in the highlands emerged only during the Late Formative

Period and was stimulated by contact with coastal groups (Bruhns 2003). Previous botanical analysis shows that a highland agricultural system, including oca, quinoa, chocho, and potatoes, and a sedentary lifeway, was present at Cotocollao by at least the terminal Early Formative Period. Although the ceramic charred C14 AMS dates (Table

6.8, Chapter 6) did not confirm the Early Formative dates previously reported for

Chaullabamba (Chapter 4, section 4.3) and the Catamayo Valley sites (Chapter 4, section

4.4), as I have argued above in section 7.3, these highland sites/regions may indeed have

Early Formative occupations. The previously reported dates for Santa Ana-La Florida

(Chapter 4, section 4.2, Table 4.1), located on the eastern slopes of the Andes, as well as the directly dated ceramic charred residue sample (Table 6.8, Chapter 6), show that the

Mayo-Chinchipe Early Formative Period culture of the eastern Andean slopes was contemporaneous with, but distinct from, Early Formative coastal Valdivia (as discussed in Chapter 4, section 4.2). The lake core pollen records for the highlands and eastern lowlands (Table 7.4) also show a greater history of plant cultivation, in particular maize, than the thus-far known archaeobotanical record attests to. While none of the highland starch analysis samples dated to earlier than the Middle to Late Formative Period, some inferences regarding the directionality of cultural interaction can be made based on the plant species identified and their regions of origin.

354

The widespread occurance of Solanum tuberosum, and Oxalis tuberosa at highland

Formative Period archaeological sites, including Cotocollao, La Chimba, Tajamar, Cerro

Narrío, and Chaullabamba, and the presence of Chenopodium quinoa at Cotocollao, show that Ecuadorian highland groups were not only familiar with the cultivation requirements of these crops, but that cultural interaction had occured within the highlands through the

Inter-Andean corridor (Figure 7.10). As potato, oca, and quinoa were all domesticated in the southern Andes (Figure 7.2), the directionality of interaction was from the south to the north, and the crops were probably transmitted from group to group, eventually reaching Ecuador by at least the terminal Early Formative Period. Other species, such as

Phaseolus and Lupinus mutabilis (chocho) suggest multiple possibilities for cultural interaction. As argued, it is probable that more than one species of Phaseolus was cultivated at highland sites (and possibly at SALF as well), and, in fact, the macrobotanical analysis for Cotocollao (Villalba 1988:322, Table 170) indicates that two types were present at Cotocollao, a “Type A” and a “Type B”. Therefore, based on the origins of domestication for Phaseolus (Figure 7.2), they may be local (P. vulgaris), have originated from the north (P. vulagaris “B” phaseolin), and/or have originated from the western Andean slopes of northern Peru/southwestern Ecuador (P. lunatis). Lupinus mutablis also suggests interaction towards the western slopes of the Andes in northern

Peru, as the most recent genetic analysis and revision for Lupinus spp. shows that it was likely domesticated there (Figure 7.2). Zea mays presents the greatest challenge for reconstructing a route of transmition into highland Ecuador.

355

Figure 7.10. Major cultural interaction spheres based on botanical data.

Although maize was first domesticated in southwestern Mexico (Figure 7.2) the route(s) of its dispersal into South America, and highland Ecuador, are not known. With respect to the antiquity of maize in the highlands, although pollen grains, phytoliths,

356

starch granules, and macrobotanical remains show that maize was ubiquitous at archaeolgical sites in the highlands (and at Nueva Era on the western slopes), from the terminal Early Formative onwards, the lake core pollen data suggest an earlier presence

(Table 7.4). Lake San Pablo has maize pollen and charcoal at the base of the core dating to the Early Formative (Table 7.4, and Chapter 3, section 3.2.2), which suggests the presence of maize in the northern highlands prior to the Early Formative Period, and, although the evidence is not strong, the Laguna Chorreras lake core record suggests the same for the southern highlands. On the eastern side of the Andes, maize starch granules are securely dated to at least Valdivia 1b (Table 6.8, Chapter 6), and this is also supported by the Lake Ayauchi pollen record (Table 7.4). From the western side of the Andes, maize phytoliths occur in deposits at the Las Vegas (preceramic) site and date to before

7000 Cal BP (Chapter 3, section 3.1.1), and directly-dated cobs show that maize was present by at least 6700 BP from preceramic (and early ceramic) contexts at Huaca Prieta and Paredones on the north coast of Peru (Grobman et al. 2011). As reviewed in Chapter

3, maize is ubiquitous at coastal Formative Period sites in Ecuador as well.

Lathrap and colleagues (1975:20-21) hypothesized that maize spread from group to group southward from the Rio Balsas in Mexico through the tropical lowlands of Central

America and along the eastern tropical base of the Colombian, Ecuadorian and Peruvian

Andes, from here entering the highlands and coastal regions of Ecuador and Peru.

Pearsall (1977:44) also suggested that early maize spread into South America by root- crop agriculturalists living in the low to mid-altitude, seasonally moist, regions along the eastern side of the Andes, or through the lower altitude Inter-Andean valleys. The dates for maize from coastal sites are (thus far) earlier than dates documenting maize on the

357

eastern slopes of the Andes and from the highlands, but this does not necessarily mean that maize entered the highlands by way of the coast, although this conclusion is supported by current evidence. The overall lack of investigations for the highlands and eastern slopes/lowlands may contribute to a bias for early maize dates. When the Early

Formative Period Valdivia ceramics were found to predate ceramic production in Peru by at least 1000 years, this came as a shock to archaeologists. The evidence was always there, it had just not been revealed yet due to the lack of investigations in Ecuador as compared to Peru. Therefore, despite that the preponderance of current evidence (earliest dates) points to the Pacific coast as the source of maize, it is also possible that maize was introduced to the Ecuadorian highlands from the north through the Inter-Andean corridor, or by way of the eastern slopes, and/or by more than one route, but only further research in these little-studied areas will sort these possibilities out.

For the Catamayo Valley sites (La Vega and Trapichillo) and SALF, cultural contact, as documented by plant species present and their origins of domestication, is less clear. The only non-maize crop that can be considered “exotic” to the region (Figure 7.2) is manioc (Manihot esculenta), which was domesticated in the southern margins of the

Amazon Basin. However, as discussed above (sections 7.2.3, 7.4.1, and 7.4.1.6) domesticated manioc is present on the northern Peruvian coast by 7000 BC, and so its presence on the eastern side of the Andes and in the Catamayo Valley is likely more ancient than current evidence indicates, and it may have already been a “local” crop by the time La Vega, Trapichillo, and SALF were occupied. The other possible exotic species (Figure 7.2) might be Capsicum baccatum and/or C. chinense, although C. pubescens (local) cannot be excluded as a possibility. Like manioc, domesticated chili

358

pepper use may predate the occupations of SALF and La Vega, and all of these species may have been locally known. The chili pepper starch granules identified from Loma

Alta and Real Alto (Perry et al. 2007, and especially Zarrillo et al. 2008) are basically contemporaneous with SALF (Table 6.1, Chapter 6, Chapter 3, section 3.2.1). Arrowroot

(Maranta arundinacea) and yam (Dioscorea trifida) may also be exotic (Figure 7.2), but it is possible that local species may have been used. Theobroma cacao (cacao) was also domesticated in the region of SALF, as was Ipomoea batatas (sweet potato) (Figure 7.2).

As such, cultural interaction outside of the regions of SALF and the Catamayo Valley are not well supported by the botanical evidence. However, interaction between these two regions (the Catamayo Valley and the eastern slopes of the Andes, Figure 7.10) is indicated by the co-occurrence of Capsicum spp. and Manihot esculenta. Indeed, the starch analysis results from the Catamayo Valley sites suggest a link to a lowland agricultural system, which will be discussed further in section 7.5. This is why these sites are grouped separately in Table 7.4 as “Far Southern Highlands.”

In looking at Figure 7.2, and based on the most recent genetic evidence for the origins of the domesticated plant species considered in this dissertation, what is striking is the pattern of convergence and overlap located in southern Ecuador and northern Peru for the origin of several domesticated plant species. It is not surprising, then, that some of the earliest evidence for domesticated plant use comes from these regions. Importantly, many of the food plants utilized at coastal Early Formative sites have their domestication origin to the east, in the Amazon Basin (and margins) and low to mid-elevation tropics (Piperno and Pearsall 1998; Pearsall 2008), including manioc, chili peppers, llerén, arrowroot, and achira (Pearsall 2003). The southern Ecuadorian and northern Peruvian highlands have

359

long been considered a crucial area for trade routes between the eastern lowlands and the

Pacific coast due to the lower elevation than other highland areas (Braun 1982, see also

Chapter 2, and Valdez 2008). The complexity evident at SALF, concurrent with, yet distinct from, coastal Ecuador Early Formative villages, is a further testament to the unique opportunity for inter-zonal interaction presented by the physical environment of southern Ecuador. It is therefore likely that, if not in the Catamayo Valley, then elsewhere in the far southern highlands, early archaeological sites are yet to be found that unite these three zones. These three areas – the southern Ecuadorian eastern slopes/lowlands, the far southern highlands, and the coastal lowlands – also appear united on botanical grounds, in that a lowland agricultural system seemed to be practiced, albeit with regional variation (and based on an incomplete botanical record). Yet at the same time they are culturally distinct from one another based on, at the very least, ceramic styles and settlement plans (see Chapters 3 and 4). The botanical data suggest an east– west axis of interaction and even an east to west directionality for the spread of some of the plant species (Figure 7.10). In this regard, it is interesting that, as the 3-D projection for SALF (Appendix A) illustrates well, the natural valley corridors that run north-south and from the site to the west, towards the highlands and the coast, and the modern road that connects Palanda (and SALF) to the rest of Ecuador leads directly to the Catamayo

Valley, perhaps following an ancient trail. Distinction is also seen between the far southern highlands and the rest of highland Ecuador based on botanical evidence alone.

The highlands, north of the Catamayo Valley, show a north–south orientation for cultural interaction (Figure 7.10), while an east-west axis is not well supported by the botanical evidence alone. Directionality from the southern highlands of Peru to the north into

360

Ecuador is suggested by the botanical data. Undoubtedly interaction was occurring between all regions of the highlands, the coast, and the eastern lowlands, but what the botanical data highlight is that a highland agricultural system was being practiced north of the Catamayo Valley during the Formative Period. These topics – highland and lowland agricultural systems – will be discussed further in section 7.5.

7.4.4 Why Nutrition Matters

Throughout the preceding discussion on the crop species, I have considered the nutritional and, especially, protein content (amino acids) of the edible plant parts. This is because, as mentioned previously, Andean diets are high in carbohydrate-rich plants, but low in meat, and this may have been the situation in the past as well. Therefore, obtaining adequate protein and iron from plant foods present challenges and can lead to serious health consequences. Protein-energy malnutrition (PEM) is still a major concern and cause of death, especially in children, in many countries in the world. PEM (commonly called kwashiorkor) is a range of pathological conditions, characterized by symptoms such as increased infections, low body weight, muscle wasting, diarrhea and dehydration, hair loss, and mental changes due to a deficiency of protein and energy (DeMaeyer

1976:23-26). Children, adolescents, and nursing women are particularly susceptible to

PEM, and it is also associated with potassium, magnesium, vitamin A, and iron deficiency anemia (DeMaeyer 1976:33). Pellagra is a nutritional disease associated with low niacin intake. The disease is characterized by gastrointestinal problems, dermatitis, mental disturbances, and, if left untreated, may progress to death (Barakat 1976:126).

Fortunately humans are capable of transforming the amino acid tryptophan into niacin,

361

but maize is deficient in tryptophan. As such, people who consume a high-maize diet (to the exclusion of other sources of tryptophan and/or niacin) are vulnerable to developing pellagra and, without adequate supplementation, PEM. These individuals are at serious risk for increased morbidity and mortality.

It is a myth that an all-plant protein diet is inferior to one where meat is consumed.

Complementary mixtures of plant proteins can provide a complete and well-balanced source of amino acids (Young and Pellet 1994). Therefore, while I cannot quantify the relative abundance of the different plant species identified based on the starch granule

(and other types) of analysis, I can assess their importance to an overall balanced diet. In general, energy (calories) is not an issue because most of the plants used are high- carbohydrate sources. Adequate iron and protein, with a balance of essential amino acids, are the main limitations. Of the plants foods identified, maize and beans (Phaseolus) are ubiquitous at all sites, and, when consumed together produce an adequate, but not ideal, diet that is high in energy, fibre, iron, potassium, and vitamins A and C. Vitamin C is heat sensitive, however, so the long cooking times required for dried beans can nullify this benefit and also lead to a decreased absorption of the iron present in beans. In this regard, eating toasted nuñas, because they are not cooked for very long, has an added nutritional benefit. For the highland sites above 2000 masl (La Chimba, Tajamar, Cotocollao, Cerro

Narrío, and Chaullabamba), the addition of quinoa to the diet provides added protein, including the amino acids methionine and cystine, which are lacking in beans. Chocho is complementary as well because it is high in oil and protein, including the amino acids lysine and cystine, with cystine lacking in beans. Oca is also high in protein, with a good balance of amino acids, but is especially complementary because it is high in iron, niacin,

362

and vitamin C (cooking does not necessarily destroy all of the vitamin C in a food source, depending on content, cooking time, and whether the cooking water is drained off).

Potatoes add additional vitamin C, calcium, and potassium, and, if made into chuño, the calcium content is even more concentrated. Not to be forgotten is that tree fruit rinds and pit fragments have also been identified from highland sites (Table 3.2, Chapter 3 and

Table 7.4), and were probably more common than the limited macrobotanical investigations have revealed. These would have been an additional and important source of vitamins (including vitamin C).

There has only been one study published that has examined human stable isotopes from the Ecuadorian highlands. Ubelaker and colleagues (1995) performed stable isotope analysis on high-status (shaft tomb burials) and low-status individuals from the La

Florida site (not to be confused with SALF) located near Quito. The age at death and sexing determinations resulted in a sample of 32 individuals ranging in age from about 7 to 50 years. Interestingly, all of the individuals were healthy, and no significant difference in the consumption of animal protein was found between the high- and low- status individuals, or between males and females (Ubelaker et al. 1995:408-409).

Additionally, all the individuals showed a diet high in maize, but the high-status individuals showed an increased consumption, interpreted as differential access to chicha.

The sample from La Florida dates to ca. 100 to 400 AD (Ubelaker at al. 1995:403)

(Regional Development Period), which postdates the Formative Period under consideration. However, if cultural complexity increases over time (as indicated by the transition between the Formative and Regional Development Periods), then one would expect that, if differential access to adequate protein was to be evident, it would be later

363

in time rather than earlier in time. The evidence from La Florida, however, does not show a differential access to protein. The population was healthy, and so we might assume the same for the preceding Formative Period, especially as the diet does not seem to be limited to a few staple, protein deficient, crops. During the 1540’s, when Cieza de León was travelling through the highlands of Ecuador, he said this about the indigenous people of Latacunga (south of Quito): “These Indians eat early in the morning... After they have eaten their maize, with meat or fish, they pass the rest of the day in drinking chicha, or wine made from maize, always holding the cup in their hands” (Cieza de León 1864:151-

152).

For the sites below ca. 1500 masl, other foods are complementary to a diet of maize and beans. Chili peppers are a source of vitamin C, and as they can be consumed fresh, or added at the end of cooking, the vitamin C content remains high. Manioc, other than providing a high amount of energy, is low in protein (as are all of tropical root crops identified). Yams are important because they add calcium and vitamin A. Sweet potatoes are also a source of calcium as well as vitamin C, and, especially, vitamin A; they are also easily digestible. Arrowroot is also easily digestible, high in potassium, and has many medicinal benefits. Juice made from cacao pulp is exceptionally high in amino acids and minerals, and, when combined with maize and pigmented fruits and vegetables, contributes to a well-balanced, nutritious diet. Many other fruits were undoubtedly consumed as well, and meat (I include fish and other aquatic sources as meat) protein was also surely part of the diet. Peanuts, and leguminous tree seeds (e.g., Inga) were probably also important sources of protein.

364

It is beyond the scope of this dissertation to enter into a protracted discussion on the long-standing “environmental limitations” and “protein thesis” debates regarding

Amazonian cultural development. In brief, it has been postulated that environmental limitations (Gross 1975; Harris 1984; Meggers 1954, 1971; Ross 1978; Steward 1949;

Steward and Faron 1959), primarily the nutrient-poor Amazonian soils and other limiting factors, led to less-complex cultures that shifted their settlements frequently. It was also postulated that higher protein availability along the large Amazonian rivers resulted in higher population densities and more developed cultures there, while, away from the rivers, game (and obviously aquatic protein sources), was scarce, and so social units were small and their settlements semi-permanent (e.g., Lathrap 1968, 1970). However,

Carneiro (1960), and others (e.g., Balée 1988, 1989, 1993, Beckerman 1979; DeBoer et al. 1996; Denevan 1992, 2001; Erickson 1995, 2008; Heckenberger 1998, Heckenberger et al. 2003; Myers 1992, 2004; Neves and Petersen 2006; Oliver 2001, 2008; Piperno and

Pearsall 1998; Redman 1999; Roosevelt 1999; Woods and McCann 1999) have produced a growing body of evidence that environmental limitations were less of a restraint than previously thought, as the tropical forest landscape and vegetation was managed, altered, and “domesticated”, and, in some areas, soil fertility was improved, so that large, sedentary and complex societies were present in Amazonia, even in upland areas, in the past (see definitions in Chapter 1 for “landscape domestication” and “agroforesty”).

Undoubtedly many other plant foots were cultivated, managed, and/or collected and contributed to the diet in both the higher and lower elevation regions. While meat may or may not have contributed considerably to the diet in the past, an overall lack of meat would not have been overtly harmful if plant protein and iron sources were

365

adequate. What I want to highlight is that, while protein is low in highland and tropical root crops as well as in maize, beans (Phaseolus) are not the only complementary source of iron and essential amino acids (protein), and many other foods contribute to a healthy overall diet. I should also point out that I am not suggesting that maize and beans formed the staple crops, as other combinations of plant foods can produce an equally nutritious diet. In fact, maize and Phaseolus are not necessary in high-elevation diets above 3000 masl, as chocho, oca, quinoa and potatoes together are adequate. I do not feel that the plant food combinations that result in a healthy diet, especially if the diet is low in meat protein, came about by chance, but rather were the result of human selection over a long period of time. This suggests a long period of knowledge-acquisition and development and I propose that such detailed awareness must have been gained over a considerable period of time, as was the knowledge that allowed for complex agricultural production systems to be developed in different environments – lowland, mid-elevation, and high- elevation. The results of these wisdoms are the village sites that become archaeologically evident by at least the terminal Early Formative Period in the highlands, and earlier in the eastern lowlands at SALF.

7.5 The Use of Ecological Zones and the Nature of Agricultural Production Systems

There are patterns evident, probably based on site elevation, to the plant species identified by starch analysis, as discussed above in section 7.4.3 (see also Table 7.4). As such, and despite that there are weaknesses in comparing the relative frequency and ubiquity of the different starch granule types identified, bar graphs help to visualize these patterns. Table 7.5 shows the starch analysis results for Cerro Narrío, Chaullabamba, La

366

Chimba and Tajamar, with these sites all located above 2000 masl. Table 7.6 shows the starch analysis results for La Vega, Trapichillo, and SALF, which are all located at elevations below 1500 masl. Figure 7.11 shows the same results as Tables 7.5 and 7.6 in bar graph form. For both the tables and graphs, I have excluded all of the unknown starch granule types. Most of the earliest identified domesticated plants in the highlands are

Andean domesticates, suggesting an in situ development of food production.

As Tables 7.5 and 7.6 and Figure 7.11 show, high-elevation crops (Oxalis tuberosa, with 75% ubiquity, and Solanum tuberosum, with 100% ubiquity) are present at Cerro

Narrío, Chaullabamba, La Chimba and Tajamar, and these sites are all located above

2000 masl. Although I did not recover starch granules of oca from the (limited) samples I analyzed from La Chimba, carbonized macroremains identified as Oxalis tuberosa were recovered from samples dating to the Late Formative Period at La Chimba (Pearsall

2003:217-232), as discussed in Chapter 3 (see Table 3.2). In addition, the high elevation crops Lupinus mutabilis (chocho/tarwi) and Chenopodium quinoa (quinoa) were identified by carbonized macroremains and pollen at Cotocollao (Villalba 1988:330-340), as shown in Table 3.2, Chapter 3, and discussed above in section 7.4.2, with Cotocollao located at ca. 2700 masl (Villalba 1988:39, figure 13). Therefore, it appears that a high- elevation complex of plant species (as defined and discussed in Chapter 4), as well mid- elevation maize and beans, was being utilized at Cerro Narrío, Chaullabamba,

Cotocollao, La Chimba and Tajamar.

367

Table 7.5. Starch Granule Analysis Results for Sites Located Above 2000 masl.

ules . osa erosum Gran spp h tarc olus s tuber S ea mays se Z ha . Archaeological Site P Oxali Solanumc ftub Zea maysTotal Cerro Narrío 45 43 54 61 192 395 Chaullabamba 10 27 6 2 50 95 La Chimba 26 14 88 128 Tajamar 8 20 7 59 153 247 Totals by type 63 90 93 136 483 865 Ubiquity 75 75 100 100 100 Relative Frequency 7.3 10.4 10.8 15.7 55.8

Table 7.6. Starch Granule Analysis Results for Sites Located Below 1500 masl.

a ace s ta . . . . p p p p tata ulen sp sp sp sp a c a m ea t es m ys u r olus o ro co se ih b a n ranta arundineo Zea mays h a a ea ma Archaeological Site Capsic Dios P Ipomoea bM M Th cf. Z Total Starch Granules La Vega 12 11 16 29 70 138 Santa Ana-La Florida 8 6 3 34 15 22 25 118 341 572 Trapichillo 13 20 6 39 78 Totals by type 20 6 27 34 51 22 25 153 450 788 Ubiquity 67 33 100 33 100 33 33 100 100 Relative Frequency 2.5 0.8 3.4 4.3 6.5 2.8 3.2 19.4 57.1

368

100

50 Ubiquity

40 Relative Frequency

100 90 80 70 60 50 40 30 20 10 0

Figure 7.11. Ubiquity and Relative Frequency of Plants Identified by Starch Granule Analysis. The top graph shows results for sites located >2000 masl: Cerro Narrío, Chaullabamba, La Chimba, and Tajamar. The bottom graph shows results for sites located <1500 masl: Trapichillo, La Vega, and Santa Ana-La Florida.

369

In contrast, the sites located below 1500 masl, La Vega, SALF and Trapichillo show species suited to warm lowland locations (a lowland complex/agricultural system, see Chapter 4), such as Capsicum spp., Dioscorea spp., Ipomoea batatas, Manihot esculenta, Maranta arundinacea, and Theobroma spp. Zea mays and Phaseolus spp. are present at all of the sites, regardless of elevation (Phaseolus was not identified by starch granule analysis at La Chimba, but carbonized macroremains of Phaseolus spp. are present at the site dating to the Late Formative occupation, as shown in Table 3.3,

Chapter 3). As mentioned previously, some taxa may not be represented at all sites simply due to sampling bias, but there does seem to be a pattern of plant species presence and land-use based on site elevation (>2000 and <1500 masl). In order to assess this it is worthwhile to explore ecological zones present in the local region for each of the sites, and I also want to clarify why a 5 km catchment area size is being used.

Cieza de León (1864:144), in describing the Qhapaq Ñan (the north-south Inca

Royal Road), which he travelled over in the late 1540’s from Quito to Cuzco, remarks that “on these roads were pleasant and beautiful lodgings [tampus] and palaces every three or four leagues”. These distances are about one day’s travel apart (depending on the terrain) (Hyslop 1984:300; McEwan 2006:119). Spain had two measurements distance in the 16th century: the legua legal (ca. 4.2 km) was primarily used in juridical matters such as land grants; and the legua común (ca. 5.6 km), a more general and older system used to indicate the distance one travelled, with one league indicating the distance a person could walk in an hour (Chardon 1980:295). Chardon (1980) argues that educated travellers to the Americas used the legua común, and Hyslop (1984:300) also uses a distance of 5 km to estimate that, in general, tampus were located about 15-25 km (three to five leagues)

370

apart. From these data we can assume then that a person can travel about 20 km in a day

(without returning to the starting point) and about 5 km per hour in the Andes when travelling on established paths. Therefore, I have set a conservative 5-km boundary area around each site in order to assess the types of ecological zones (based on elevation alone) that a person can walk to in an hour. This would allow time for travel (on paths) to a cultivated field, at least an 8 hour work-day, and return travel to the village, all within a

10-12 hour period. Hastorf (1993:148-149) also used a 5-km boundary, although for slightly different reasons, for her land-use catchment analysis of sites located in the Jauja region of Peru.

The topographic maps in Appendix A show the elevations present within a 5 km catchment area for each of the sites and the 3-D projections (also in Appendix A) help to visualize the surrounding landscapes. Unfortunately, detailed climatic data (temperature and precipitation differences) and soil information were not available for the site catchment areas, and so I could not integrate these data into the GIS maps. Consequently,

I can only infer the presence (and differential use) of ecological zones based on elevation differences shown for each of the 5 km site catchment areas, with the added information that local environmental conditions for all of the sites were either similar to current conditions, or possibly warmer and wetter (as discussed in Chapter 4). What the maps clearly show is that large areas, above 2500 masl, best suited to growing high-elevation crops (potatoes, oca, chocho/tarwi, and quinoa, see sections 7.4.2.1 to 7.4.2.12 and Table

7.5), are located within a 5 km catchment area for Cerro Narrío, Chaullabamba, La

Chimba, and Tajamar (as noted previously, the ca. 6 km catchment area for Cotocollao is similar).

371

As discussed above, maize can be grown up to 3100 meters but does much better at lower elevations, as do beans, and many suitable habitats allowing for their cultivation are located in the valley bottom as well as on the lower elevation valley slopes for Cerro

Narrío, Chaullabamba, and Tajamar. The La Chimba site is the one possible exception to this pattern, as, despite that the site overlooks a broad expanse of valley floor, most of the land is above 3100 masl (Appendix A). While maize and beans can be grown at this elevation, it may be that those crops were grown a little further to the southwest (this is best shown in the 3-D projection for La Chimba, Appendix A), outside of the 5 km site catchment areas, as they would be much more productive at lower elevations. In this regard, it is interesting that, as discussed in Chapter 4, section 4.6, Stahl and Athens

(2001) convincingly argue that the faunal assemblage suggests that jerked meat (ch’arki) was prepared at the site for trade. Perhaps, then, in addition to the evidence present at the site for extra-local trade (including Spondylus and Strombus) (Athens 1995), more localized trade of ch’arki for maize and beans may have occurred. The 3-D projection for

La Chimba (Appendix A) shows that the site appears to be strategically located for easy access to the higher-elevation páramo as well as lower elevation cultivation areas. Cerro

Narrío also seems to be strategically located midway between higher (better for growing potatoes and oca) and lower elevations, based on the 3-D projection map (Appendix A), but is at a lower elevation than La Chimba, and not above the maximum limit for maize and bean cultivation. The topographic maps, and especially the 3-D projections, for

Tajamar and Chaullabamba (Appendix A), show that these sites are located at the lowest

(or close to) elevations within the 5 km catchment areas, best suited to growing maize and beans, in contrast to La Chimba and Cerro Narrío. Other factors undoubtedly

372

contributed to settlement strategies for these areas, but the importance of crop requirements should not be overlooked.

As discussed in Chapter 4, sections 4.3 (Chaullabamba), 4.5 (Cerro Narrío), 4.6

(La Chimba), and 4.7 (Tajamar), paleoenvironmental reconstructions for these site regions (Cotocollao can be grouped together with Tajamar) show modern climatic conditions, and all of the high-elevation crop species identified by starch analysis or by previous botanical investigations for these sites (oca, potato, chocho/tarwi and quinoa) are part of the modern agricultural system (as discussed in Chapter 2, section 4.2). What this may indicate, then, is that a complex highland agricultural and land-use system was utilized at Cerro Narrío, Chaullabamba, Cotocollao, La Chimba, and Tajamar whereby crops were cultivated in niches best suited to their production, as is similar to today (see

Chapter 2, section 2.5 and Chapter 4). These sites, all located higher than 2000 masl, also correspond to the tierra fría (temperate, 2000-3400 masl) climatic zone defined for

Ecuador (see Chapter 2, Table 2.1), and a highland agricultural complex is best suited to this zone. In addition, regardless of the physical origin of introduced maize, extensive knowledge of the multitude of microenvironments available, and how to best manage this landscape, would have been required for successful farming strategies in the highlands.

This suggests longstanding local knowledge. Based on the site catchment area maps, we may assume that not all crops, especially the high-elevation species, were being grown immediately adjacent to each of the sites, but that they were grown in fields located within the 5 km (one day work-return) catchment areas, with La Chimba being the possible exception.

373

One of the most important adaptations human groups have made to the Andean environment is the complementary use of different ecological zones. The vertical archipelago model, as defined and based on historic research of the Aymara in the southern Peruvian highlands (Murra 1968, 1972), states that ethnic groups have adapted to the diverse Andean environment by developing economic and social structures to control and access ecological zones that are widely dispersed, both spatially and vertically. As Hastorf (1993:186) explains, this land-use strategy occurs because no one zone can provide “the proper dietary mix”, therefore labourers sometimes travel hours or days to access vertically dispersed zones in order for the community to produce a variety of different crops. Brush’s (1977) ethnographic study in Uchucmarca, located in the northern Peruvian Andes, is similar, but describes the use of continuous and compacted ecological zones, as does Hastorf’s (1993) study for the Sausa in the Jauja Valley (central

Peruvian Andes). This “compact verticality” model (Brush 1976), whereby contiguous localized microzones within a 5 km area were utilized to produce the majority (if not all) of the agricultural diet (Hastorf 1993:186), is probably most applicable to all of the sites investigated in this dissertation, but particularly for Cerro Narrío (Figure 7.11),

Chaullabamba, La Chimba, Tajamar, and Cotocollao.

374

Figure 7.11. Modern “Compact Vertical” agricultural system near Cerro Narrío. The patchwork quilt of fields produce different crops such as maize and potatoes (not all fields are monocropped), as well as introduced broad/fava beans, barley, carrots, cabbages, etc., in microenvironments suited to their production based on varying factors of elevation, slope, aspect, soil fertility, moisture and temperature. Picture taken July, 2009.

As Hastorf (1993:29) notes, features such as proximity, varying temperature and moisture regimes, soil fertility, and crop life cycles will affect the use and management of different ecological zones. I do not have soil fertility, temperature and moisture data to assess variability within the 5 km catchment areas, but it is likely that suitable microhabitats within 5 km of the sites, based on elevation alone, were available to allow for: 1) the production of a variety crops, including high-elevation tubers, beans

(chocho/tarwi) and pseudocereals (quinoa), 2) crop rotation and field fallowing to maintain soil fertility, and 3) the continuous production of crops. These strategies, in addition to storage, would guard against famine in the event of single or multicrop failure

(Denevan 2001:299). Although I did not conduct a formal study, my personal observation

375

of the agricultural practices for the immediate areas surrounding Cerro Narrío (see Figure

7.11) and Chaullabamba supports this view. Today, most of the cultivated land (in and around Cerro Narrío, Chaullabamba, and the Catamayo Valley) are ordinary, rainfed fields, and where landform modification is found, it is usually simple canal irrigation with field ditches.

In contrast, the topographic maps (Appendix A) for Trapichillo and La Vega

(shown on one map) and Santa Ana-La Florida, show that abundant areas below 1500 masl are located in the 5 km site catchment areas. Elevations below 1500 masl are best suited to growing the lowland plant species identified at these sites, as discussed above

(sections 7.4.1.1 to 7.4.1.12 and as shown in Table 7.5), and the elevations of these sites correspond to the tierra templada (subtropical, 800-2000 masl) climatic zone (see Chapter

2, Table 2.1). As mentioned previously, SALF is located at a transition zone, between the cloud forest/montaña and lowlands (tierra caliente, tropical, climatic zone: 0-800 masl).

Both the Catamayo Valley sites and SALF are located in the valley bottoms (as illustrated by the topographic and 3-D projections in Appendix A), which would be the warmest areas. In addition, land above 2500 masl, as shown on the topographic maps in Appendix

A, which would be more suitable for growing high-elevation crops in particular, are almost entirely absent within the 5 km catchment areas of Trapichillo/La Vega and

SALF, and these species were not represented in my samples. As noted previously, there may be a sampling bias explaining the absence of starch granules for some species, with this especially important for the Catamayo sites (10 samples in total for Trapichillo and

La Vega, all charred ceramic residues and all positive for starch). A sampling bias is less likely for SALF as 27 samples (including multiple samples from some artifacts) were

376

analyzed from several contexts (including ceramic charred residues as well as residues from ceramic bottles and stone bowls), and 11 of the 19 different contexts tested positive for starch (Chapter 6, section 6.2). Therefore, I cannot say with certainty that high- elevation species were never present at Trapichillo/La Vega and/or SALF. Nevertheless, the lowland species identified are consistent with local cultivation within a 5 km catchment area for these sites based solely on elevation. Indeed, as discussed in Chapter

4, section 4.2 for SALF and section 4.4 for Trapichillo and La Vega, local climatic conditions were either very similar to modern conditions, or perhaps warmer and wetter, at the time the sites were occupied. As well, and as also discussed in Chapter 4, section

4.4, the Catamayo Valley is a major producer of sugar cane and manioc today, by employing canal irrigation and field ditches, as the valley is very warm despite its highland location. SALF (Chapter 4, section 4.2) is located on the eastern Andean slopes where the cloud forest transitions to the lowlands, and where lowland crops are grown today (see Figures 4.3, 4.4, 4.6, 4.13 and 4.14, Chapter 4, and Figure 7.5 below). I have observed swidden agriculture and agroforesty in the close vicinity of SALF, and most of the field types are ordinary and rainfed. Therefore the lowland plant species, and a lowland agricultural complex (see Chapter 4), identified from these sites are not surprising based on local environmental conditions. Collectively then, the paleoclimatic data, location of the sites within the tierra templada, subtropical, climatic zone, and modern agricultural practices, add further support for the suggestion that Trapichillo and

La Vega show an affinity for a low-elevation plant complex and agricultural system, as does, to an even greater extent and not surprisingly, Santa Ana-La Florida.

377

7.6 Chapter Summary

In section 7.2 I discussed methodological issues involved in starch granule analysis.

Numerous lines of evidence, as well as the extensive precautions taken to avoid incidental starch contamination, show that the starch granules recovered from the artifact residues’ are a result of use. I also discussed the importance of including blank control samples with archaeological samples in starch analysis, and that these procedures should be “standard practice” whenever ancient starch analysis is undertaken. While the methods for sampling were successful in recovering starch granules from almost all of the samples analyzed, some refinements could be made and are suggested, such as retaining and testing the distilled/deionized water used to wash previously excavated stone tools. The method for removing residues from the interior of intact stirrup-spout bottles was simple, yet has not been previously reported. For future analysis I suggest that more samples be analyzed to test the efficacy the sonication method. An assemblage-based approach is the most conservative method of identifying unknown starch granules to different taxa, and why I chose this method. The identifications were hampered, however, by the bewildering number of plant species present in Ecuador, and so further work is required to confirm some of the identifications made, in particular the specificity of the

Theobroma cacao starch granule form I identify as being potentially diagnostic. Methods of quantification in starch granule analysis are particularly problematic. I argue that even relative frequency and ubiquity measures are as likely to be biased as absolute counts.

Cultural and non-cultural factors, such as different processing, use and storage patterns, differential preservation (based on a host of factors), and sampling strategies, will bias the numbers and types of starch granules recovered, making quantification within and

378

between archaeological sites problematic. As well, starch analysis alone cannot provide a detailed list of past plant resources as not all plants utilized produce storage starch (or not in the plant parts targeted for use). Quantification of the relative abundance of C3 and C4 foods based on stable isotope assays of ceramic charred food residues is also fraught with problems. In particular, maize use can be easily masked in the δ13C signature when C3 resources are cooked in the same pot, thereby resulting in false negatives. The differential production of starch between plant species and the difficulty in quantifying, and even detecting, maize use by stable isotope measurements, highlight why it is important to use different analytical techniques (pollen, phytolith, macrobotanical, absorbed organic residues, and even ancient DNA) when attempting to reconstruct the history of plant use in a region. While my analysis focused on starch granules to recover information on the use of USO’s from the samples, I used the results of past botanical analysis to provide a more inclusive overview of past plant use for the Formative Period.

In order to include the botanical results from Cotocollao realized from previous analysis, I re-evaluated the radiocarbon dates reported for the site. In doing so I was able to confirm that that the earliest occupation of the site, and the botanical remains, date to the terminal Early Formative Period. I also discussed the difference between accuracy and precision with respect to radiocarbon dating. While the dates I obtained from ceramic charred residues from Chaullabamba and the Catamayo Valley sites may be precise, they may not be accurate (and are probably not) in so far as overturning established chronologies for these sites that have been shown to begin in the Early Formative Period.

Further, I also discuss the difficulty in comparing radiocarbon dates for Ecuador as a whole, and that, at the very least, renewed investigations in key regions of the highlands

379

(Cotocollao sites in the Quito region, Chaullabamba, and the Catamayo Valley) should be undertaken specifically to address chronological issues.

Based on multiple lines of evidence, in particular previous archaeological evidence for domesticated plant use, I “fine tune” the very conservative taxonomic identification I made by starch analysis. I argue that several of the taxa identified, in particular Phaseolus, Oxalis tuberosa, and Solanum tuberosum, starch granules are most likely indicative of domesticated species. The current botanical record for the highlands, eastern and western slopes, is then discussed with a focus on the ecological requirements of the plant species identified, their regions of origin, and their nutritional profiles. The regions of origin for the plant species was then used to discuss cultural interaction. The conclusions reached were that:

1. The presence of the Andean domesticated highland crops – oca, quinoa, and

potatoes – at Cotocollao by at least the terminal Early Formative Period, suggests

that these plant species may be present even earlier than the present record attests.

Their abrupt appearance at Cotocollao, far removed from their regions of origin to

the south, coupled with a suite of other crops (including maize), supports this

hypothesis;

2. The plant species from identified from highland sites to the north of the Catamayo

Valley (Chaullabamba, Cerro Narrío, Cotocollao, Tajamar, and La Chimba) also

show that cultural interaction had occurred with groups to the south, through the

inter-Andean corridor. Oca, potato, and quinoa were all domesticated far to the

south, and so a south-north axis of cultural interaction and, at least as much as the

botanical record can indicate, directionality is suggested. Other plant species

380

indicate multiple possibilities for cultural interaction, including the western

northern slopes of northern Peru;

3. Although maize is ubiquitous at the sites investigated, and current evidence shows

that it is present earlier on the coast of Ecuador and northern Peru, the route of its

dispersal to the eastern slopes of the Andes (SALF) and into the highlands, is not

clear. It may be that maize was introduced to the highlands from the coast, but

lake core pollen evidence is also suggestive of an inter-Andean corridor

transmission from the north (Lake San Pablo) or the eastern slopes of the Andes

(Lake Ayauchi). It is possible that maize entered highland Ecuador by multiple

routes;

4. The far southern highland sites in the Catamayo Valley and SALF on the eastern

slopes of the Andes show a predominantly east-west axis of cultural interaction.

Although manioc is the strongest candidate as an “exotic” plant as it originates

from the southern margins of the Amazon Basin, it is likely that it was present in

the region prior to the occupations at SALF and in the Catamayo Valley. This is

also the case for the other species identified, and, if confirmed, the identification

of Theobroma cacao may represent the earliest evidence for its use (also setting a

minimum date for its domestication);

5. The fact that SALF, the Catamayo Valley sites, and coastal Ecuador share an

affinity for lowland cultivation and an east-west axis of cultural interaction is

likely a result of geography. The southern highlands of Ecuador and northern

highlands of Peru are the lowest point in the Andes, and, therefore, provide a

381

natural corridor for cultural interaction and the movement of plants and other

products between the eastern lowlands and the Pacific coast;

Based on the nutritional profiles of the major plant species identified I also argue that these species are complementary in providing a balanced and healthy diet. Indeed, while I am not suggesting that the Andean people were vegetarians, they could have lived healthy lives without a great deal of meat in their diets. Even though the results of human bone stable isotope analysis from La Florida date to a slightly more recent time period, they indicate that highland people had a healthy diet with adequate, and non-differential, access to animal protein. At all of the sites, maize and beans could have provided an adequate complement of calories, iron, and protein, especially with the addition of other plant foods, such as fruits. The plant species also show a pattern of complementarity based on elevation. For sites above 2000 masl oca, potatoes, quinoa, chocho, and nuñas, can provide a more nutritious diet than one based primarily on maize and beans.

Toasted/popped maize and beans can conserve fuel as well as maintain the vitamin C needed to aid in iron uptake. It is interesting, and probably not coincidental, that the type of maize found in the preceramic deposits at Paredones and Huaca Prieta was the popcorn variety (Grobman et al. 2012), and so toasting maize and beans may indicate one early method of preparing these foods (which, prior to the introduction of ceramics, may make sense). At even higher elevations, above 3000 masl, where maize and beans cannot grow, oca, chocho, quinoa, and potatoes are adequate, and the production of oca and potato freeze-dried products increases their storage life, concentrates their nutrients, and makes them light and easily transported. The lack of protein in the diet is one of the major limitations in the tropical lowlands, especially away from the rivers, because

382

aquatic resources are no longer available and game is scarce. These facts, as well as others, have led to many hypotheses regarding past population densities, settlement patterns and longevity, and limitations on cultural development imposed by the tropical forest environment. As noted, evidence is now emerging that the environmental limitations of the Amazonian lowlands did not exceed humans’ ability for ingenious solutions. Further, the availability plant protein in the tropical lowlands has been given short-shrift in the past. While the staple root crops are indeed low in protein, other plant foods, such as beans, drinks made from fruits (and the pulp, as the case for cacao), tree and other legume sources (such as Inga and peanuts) can provide all the essential amino acids required for a healthy diet, especially if maize is consumed as well. The earliest directly-dated remains of peanut (8640-8435 Cal BP, 2σ) are from the Zaña valley from northern Peru (Dillehay et al. 2007:1892), indicating great antiquity for its use.

Finally, I also discuss, in particular, the nature of the highland agricultural system.

The highland sites located north of the Catamayo Valley are all located above 2000 masl and show that plant species that are part of the high-elevation complex, as well mid- elevation maize and beans, were cultivated at Cerro Narrío, Chaullabamba, Cotocollao,

La Chimba and Tajamar. Based on the catchment area maps, all of these crops could have been grown within 5 km of each site, indicating that people could have left the village, worked in their fields, and returned home the same day. This suggests that a “compact verticality” model (Brush 1976), whereby contiguous localized microzones within a 5 km area were utilized to produce the majority (if not all) of the agricultural diet (Hastorf

1993:186), was practiced at Cerro Narrío, Chaullabamba, La Chimba, Tajamar, and

Cotocollao. The 5 km catchment area maps for the Catamayo Valley (Trapichillo and La

383

Vega), and Santa Ana-La Florida also indicate that local production for the plant species identified was possible, and that a low-elevation agricultural system was practiced, especially at SALF. Collectively, these results signify a greater antiquity of plant cultivation than the terminal Early Formative Period in the highlands and the Early

Formative Period on the eastern slopes to acquire and/or domesticate the crops and to learn the ecological requirements of each. The apparent abrupt appearance of an advanced agricultural system in both regions, and the sophisticated knowledge needed to learn what foods are complementary in producing a healthy diet, both indicate a much greater antiquity than the Early Formative Period for plant food production in the

Ecuadorian highlands and eastern slopes.

Notes

1. The topographic maps were created by using individual DEM (Digital Elevation Map)

.asc (ascii format) files. These files were created by Marc Souris (2011b) and available for free download and scientific use The .asc files were transformed to ESRI grid format in order to use them in ArcGIS. Using ArcGIS the ESRI files were then merged to individual DEM raster files to create a single DEM for all of Ecuador. The spatial reference/coordinate system information of the DEM was then fixed to bring the DEM

Ecuador to its correct spatial location. Contours were then created with 50 and 100 meter intervals using the Spatial Analyst extension of ArcGIS and hillshade was was added to the DEM in order to create a shaded relief. The site locations, contours and hillshade shaded relief (with 60% transparency) were overlaid on the DEM to create maps with the

PSAD 1956 UTM Zone17S projection.

384

Chapter Eight: Conclusions

8.1 Introduction

The purpose of this research was to investigate the timing and nature of the highland Formative Period in Ecuador, and to test the hypothesis, suggested by others, that the highland Formative Period, supported by maize agriculture, was stimulated (i.e., diffused from) coastal Ecuador. The primary botanical data were obtained from starch granule analysis of stone tool and ceramic charred residues, and chronology was addressed by directly-dating ceramic charred residue samples by AMS radiocarbon assays. Six highland sites were investigated: La Chimba and Tajamar, located in the northern highlands; Cerro Narrío and Chaullabamba from the southern highlands; and La

Vega and Trapichillo, both located in the Catamayo Valley in the far southern highlands.

The Santa Ana-La Florida site, located on the southeastern slopes of the Andes was also included as a contrast and supplement to previous botanical analyses from highland and coastal Ecuador. The main questions being asked were: 1) What is the antiquity of domesticated plant use, including maize, in the highlands; and 2) what do these data tell us about the stimuli to increased socio-cultural complexity during the Formative Period in highland Ecuador. Based on the results of analysis, coupled with previous botanical and other analysis from both the highlands and coastal Ecuador, the following presents the conclusions reached along with suggestions for future research.

8.2 Starch Granule Analysis Conclusions

The application of starch granule analysis in archaeological research has increased in use over the past 30 years, since the pioneer studies conducted by Donald Ugent in

385

coastal Peru (Ugent et al. 1981; Ugent et al. 1982; Ugent et al. 1984; Ugent et al. 1986).

As methodologies, sampling strategies, identification, and quantification methods are still being developed in ancient starch analysis, I discussed these at length in Chapters 5 and

7. Confirmation of previous methods for recovering starch granules from stone tool, stone bowl, and ceramic charred residues were realized. Previous methods for recovering starch granules from stone tool residues, principally by sonication, were applied in a novel way, as discussed in Chapter 5, to sample a ceramic sherd and intact stirrup-spout vessels for residues. These methods proved successful as starch granules were recovered. Further, the matched artifact residue samples allowed quantification methods, relative frequency and presence analysis (ubiquity), to be applied and showed that the sonication technique was successful in recovering use-related starch granules, as discussed in Chapters 6 and

7. Other methodological issues, primarily the possibility for modern starch to contaminate archaeological samples, were discussed in Chapters 5 and 7. Although starch analysts employ methods to address modern starch contamination in their analyses, the problem, as a whole, has only recently been discussed as a stand-alone topic (Laurence et al. 2011). This is an under-discussed issue in ancient starch analysis and I have made suggestions to reduce the possibility for incidental starch to come in contact with artifacts and their residues during excavation and laboratory procedures (Chapter 5). I also strongly advocate that blank control samples be used through all steps of sampling and lab analysis to assess the “starch hygiene” procedures used, and to allow for any problems to be pinpointed to various steps during sampling and lab processing procedures, so they can be addressed (Chapter 7). Further work is also required to understand the patterning of starch granules in the archaeological record, and how

386

cultural use, differential preservation, and sampling techniques skew the results obtained from analysis. Additional research to address these matters will allow for progress to be made in quantifying the results of starch granule analyses. As discussed in Chapters 5 and

7, analyses complementary to starch granule analysis, including other paleobotanical techniques such as macrobotanical, phytolith and pollen analysis, as well as stable isotope and other chemical analyses of food residues and human (and animal) remains, ancient

DNA, and soil chemistry analysis, to name a few, can and should be performed to complement and enhance the picture of past plant use. Ancient starch research is now being conducted in almost all world regions to study, not only the past use of plants, but also much broader processes including human cultural, social, and economic development. These were the larger aims of my research.

8.3 Conclusions Regarding Highland Formative Period Plant Use and Timing

In Chapter 6, I identified starch granules to taxa centered on morphological and other characteristics, employing an assemblage-based approach. Starch granules were classified to species, genera or family, in a highly conservative way, despite that, in almost all cases, the starch granules were consistent with both comparative specimens and published descriptions for starch granules of domesticated species. As discussed in

Chapter 7, it is important to remember that, while my comparative collection does not include all possible species for the plant taxa identified, the published descriptions I refer to in Chapter 6 are based on examining starch granule assemblages from a very broad and inclusive number of species for the highlands and tropical lowlands. By incorporating evidence documenting domesticated plant use from Cotocollao, in particular, and other

387

regional archaeological sites that predate the archaeological sites from which the starch analysis samples originate, I argue in Chapter 7 that most of the starch granules identified are indeed from domesticated species.

The basic results realized from the starch analysis are important in that the types of plant foods utilized at the different archaeological sites investigated expand our knowledge regarding the geographic extent and temporal range of species previously identified at highland sites. Legumes/beans (Phaseolus spp.), maize (Zea mays), oca

(Oxalis tuberosa), and potato (Solanum tuberosa) use in the highlands was previously documented at Cotocollao (terminal Late Formative) and La Chimba (Late Formative), both located in the northern highlands (see Chapters 3, 4, and 7). Maize and beans have also been documented at Pirincay in the southern highlands, but the specific contexts for these identifications have not been published (see Chapters 3 and 7), but likely date to at least the Late Formative Period. An evaluation of previous absolute dates documented for

Cotocollao confirmed that domesticated beans, maize, oca, chocho (Lupinus mutabilis), quinoa (Chenopodium quinoa), and other unidentified USOs and tree fruits, were part of a “fully-formed” agricultural system by the terminal Early Formative Period, as noted by others (e.g., Bruhns 2003; Pearsall 2003; Villalba 1988; Zeidler 2008). Previous botanical analysis at La Chimba, dating to the Late Formative Period, also documents an agricultural system where maize, beans, potatoes and chocho were cultivated. These previous results are important in establishing a minimum date for the appearance of domesticated plant species in the highlands, and that plant food production contributed to the overall economy of highland Formative villages.

388

The results of my analysis show that beans, maize, potatoes and oca were also present at the northern highland site of Tajamar, as well as Cerro Narrío and

Chaullabamba in the southern highlands. Thus, the use of beans, maize, potatoes and oca was geographically widespread in highland Ecuador from at least the terminal Early

Formative (certainly at Cotocollao), and continued through to the Late Formative Period, as evidenced at La Chimba, Tajamar, Cotocollao, Cerro Narrío, and Chaullabamba

(Middle Formative Period as well). An earlier presence for maize, likely prior to the

Early Formative Period, in the northern highlands is suggested by pollen and phytolith analysis from the Lake San Pablo core, as discussed in Chapters 3 and 7. The Lake

Yambo (northern highlands) sediment core pollen record also shows a strong signature for regional maize cultivation by at least the Late Formative Period (consistently present from the bottom of the core upward) (Colinvaux et al. 1988:95). Similarly, in the southern highlands, near Chaullabamba, regional maize cultivation is evidenced during the Middle to Late Formative Period (ca. 4000 Cal BP) by the Laguna Chorreras sediment core pollen sequence, and possibly much earlier (Hansen et al. 2003:106), although evidence for the latter is weak.

In addition to beans and maize, the domesticated chili peppers (Capsicum spp.) and manioc (Manihot esculenta), not previously reported for highland Ecuador, were present by at least the Late Formative Period (Middle Formative Period as well) at La Vega and

Trapichillo, located in the far southern highlands in the Catamayo Valley. The plant species identified at Santa Ana-La Florida constitute the first Formative Period botanical record for Ecuador documented from an archaeological site east of the Andes. In addition to chili peppers, maize, and manioc, which were also identified from the Catamayo

389

Valley sites, Dioscorea spp. (yams), Ipomoea batatas (sweet potatoes), Maranta arundinacea (arrowroot), and possibly Theobroma cacao (cacao) and/or Herrania spp.

(monkey cacao) were identified in either domestic (cooking pot residues) and/or ceremonial contexts (stirrup-spout bottles and elaborately carved stone bowls recovered from high-status tombs) from Santa Ana-La Florida. The SALF botanical results are critically important as they date to Valdivia 1b-3 of the Early Formative Period, based on a directly-dated ceramic charred residue sample and previous absolute dates for the site

(as discussed in Chapters 4, 6, and 7). Therefore, the contexts at SALF, from where the plant species identified in this analysis originate, are contemporaneous with the Early

Formative Period sites of coastal Ecuador. An Early Formative presence of maize on the eastern base of the Ecuadorian Andes is also shown by the presence of maize pollen and phytoliths in the Lake Ayauchi sediment core, as discussed in Chapters 3 and 7.

While the AMS radiocarbon dates obtained from Chaullabamba and the Catamayo sites (La Vega and Trapichillo) failed to confirm the Early Formative Period initiation of these site occupations, I argued in Chapter 7 that the ceramic charred residue AMS dates cannot be used in isolation from other evidence (previous dates and contexts) to determine overall chronology for the archaeological sites. Trapichillo, in particular, and

Chaullabamba very likely date to at least the terminal Early Formative Period, even if my samples do not accurately reflect that. The AMS radiocarbon dates determined for Cerro

Narrío, where only one previous date was available from an unreported context, and

Tajamar, where no previous dates have been reported, as discussed in Chapters 3 and 4, establish minimum dates for these site occupations. The Cerro Narrío site is thought to have an Early Formative period component, based on ceramic analysis and the one

390

previous date reported, but our limited excavations at the site did not reveal ceramics that could be assigned to the “Early Narrio” phase (Collier and Murra 1943). Therefore, in this regard, our excavations at Cerro Narrío confirmed the “Late Narrío” phase (ceramic analysis is ongoing) for the site and established a Late Formative Period date for the phase, but we did not uncover any components (and associated artifacts) that were representative of Collier and Murra’s Early Narrío phase. This does not mean that Early

Narrío is not a separate phase from Late Narrío. Indeed, our lack of finding Early Narrío wares in the site excavations (all dated to the Late Formative Period), suggests that the two phases are indeed separate in time, but that we were not fortuitous in finding Early

Narrío contexts during our excavations at this very large site. The question of an Early

Formative Period occupation of Cerro Narrío, then, remains an enigma.

Returning to my original research questions, collectively, the botanical data from highland Formative Period sites show that high-elevation Andean domesticated crops, such as oca, potatoes, chocho and quinoa are present by at least the terminal Early

Formative Period. Maize and beans are present at the Cotocollao site by at least the terminal Early Formative and at all other highland sites by at least the Middle to Late

Formative. The lake core sediment pollen records show an Early Formative, and probably earlier, presence of maize in the highlands as well. The absence of markers for other crops in lake core pollen records cannot be taken as evidence of absence for cultivation of other crops prior to the Early Formative Period, however, as often extended counts are required to identify species that do not produce pollen as abundantly (Pearsall 2000:304-

305). For almost all lake sediment pollen analyses conducted in the Ecuadorian

391

highlands, save Lake San Pablo, other questions were the focus of the research, so it is unexpected (and fortuitous) that even maize pollen has been identified.

8.4 Conclusions Regarding the Stimulus to Formative Period Complexity in Highland Ecuador

My original research questions on the stimulus to complexity in the highlands of

Ecuador were based on archaeological and botanical evidence. The precocious appearance of Formative villages in the highlands, supported by a fully-developed agricultural system, has led to the suggestion that cultural complexity in the highlands diffused from coastal Ecuador, and that maize, also introduced from the coastal lowlands, was the crop that fueled this transition. Previous botanical research at highland archaeological sites showed that maize was only one of the crops present when Formative

Period village sites appear, as discussed above and in previous chapters. To me, these data pointed to other hypotheses. First, the sudden appearance a multi-crop agricultural system suggested a greater antiquity of plant food production in the highlands, regardless of whether maize was or was not part of that system and regardless of whether or not plant cultivation was occurring at nucleated villages. The presence of Andean domesticated crops also suggested that plant cultivation in the highlands had a greater antiquity than the Formative Period, and their presence in Formative village sites also suggested cultural interaction within the Inter-Andean corridor, leading to the in situ development of cultural complexity, and not stimulus primarily from the coast. Recent excavations at the Santa Ana-La Florida site, on the eastern slopes of the Andes, had shown it to be contemporaneous with the coastal Early Formative Period sites. This also

392

suggested to me that coastal Ecuador was not the only “hotspot” in Ecuador for a

Formative Village lifeway. Indeed, the remarkable cultural complexity at SALF, as evidenced by the sophisticated ceramics and settlement plan, is distinct from coastal

Ecuador. The stirrup-spout bottle form may have been introduced to coastal Ecuador, the highlands, and into Peru from the eastern lowlands, as this vessel form is earliest at SALF

(see Chapter 4 and Valdez 2008:884). Has my research, then, answered more of the hypotheses I proposed to test?

I used the botanical record for highland Ecuador, based on starch granule analysis and by previous research, to explore the cultural interaction spheres present in the highlands, principally by determining the geographic origins for the plants species. The ubiquitous occurance of Solanum tuberosum and Oxalis tuberosa at highland Formative

Period archaeological sites, including Cotocollao, La Chimba, Tajamar, Cerro Narrío, and Chaullabamba, and the presence of Chenopodium quinoa at Cotocollao, show that

Ecuadorian highland groups were not only familiar with the cultivation requirements of these crops, but that cultural interaction had occured within the highlands through the

Inter-Andean corridor. As potato, oca, and quinoa were all domesticated in the southern

Andes, the directionality of interaction was from the south to the north, and the crops reached Ecuador by at least the terminal Early Formative Period. Other species, such as

Phaseolus and Lupinus mutabilis (chocho) suggest multiple possibile directions for cultural interaction. Based on the origins of domestication for Phaseolus, they may be local (P. vulgaris), have originated from the north (P. vulagaris “B” phaseolin), and/or have originated from the western Andean slopes of northern Peru/southwestern Ecuador

(P. lunatis). Lupinus mutablis also suggests interaction towards the western slopes of the

393

Andes in northern Peru. Zea mays presents the greatest challenge for reconstructing a route of transmition into highland Ecuador, as discussed above and in Chapter 7.

Current evidence from coastal Formative Period sites, as well as the preceramic Las

Vegas (coastal Ecuador), Huaca Prieta, and Paradones (north coast of Peru) sites, are earlier than dates documenting maize on the eastern slopes of the Andes and from the highlands. I argue, though, that this does not necessarily mean that maize entered the highlands by way of the coast, much like the early dates for manioc in coastal Peru do not indicate that the crop was domesticated there. Preservation and the intensity of archaeological research are greatest in this region. Very little archaeological research, let alone botanical research, has occurred in the Ecuadorian eastern slopes and lowlands, and so evidence for early maize may be realized there. Most importantly, however, is that maize is present earlier at SALF, on the eastern slopes of the Andes, than in highland

Ecuador. Therefore, it is just as likely that maize reached the highlands by way of the east, as previously suggested by Lathrap and colleagues (1975) and by Pearsall (1977), as from the coast. Further, although I cannot deny with certainty the hypothesis that maize diffused to the highlands from coastal Ecuador to fuel the highland Formative Period, based on the research presented in this dissertation, it seems unlikely as: 1) maize was only one of several crops being grown; and 2) time would have been required to develop high-elevation varieties of maize prior to its widespread cultivation throughout the highlands during the Formative Period.

Cultural interaction spheres in the far southern highlands, based solely on the botanical species identified and their origins, are more difficult to determine. The evidence from the Catamayo Valley sites (La Vega and Trapichillo), as well as SALF,

394

show that the only non-maize crop that can be considered “exotic” to the region is manioc (Manihot esculenta), which was domesticated in the southern margins of the

Amazon Basin. In Chapter 7, I argued that manioc, as well as chili peppers, arrowroot, yams, and sweet potato (and cacao), were likely “local” crops by the time La Vega,

Trapichillo, and SALF were occupied. Therefore, based on the botanical evidence, the cultivation of these crops may have a greater antiquity in the region than currently known. The co-occurrence of chili peppers and manioc at SALF and in the Catamayo

Valley suggests cultural interaction occurred between these two regions. The similarity seen in ceramic decorative motifs and other types of artifacts (lime mortars), as discussed in Chapter 4, also supports this interpretation. Importantly, as discussed in Chapters 1, 3, and 7, many of the food plants identified from coastal Early Formative sites have their domestication origin to the east, in the Amazon Basin (and margins) also supporting an east-west axis of cultural interaction for the far southern highlands.

I also discussed, based on the most recent genetic evidence for the origins of the domesticated plant species considered in this dissertation, the pattern of convergence and overlap located in southern Ecuador and northern Peru for the origin of several domesticated plant species. It is not surprising, then, that some of the earliest evidence for domesticated plant use comes from these regions. Nor is it surprising that cultural interaction (including the transmission of plant species and knowledge related to their use and management) between the eastern lowlands, highlands, and Pacific coast, occurred in the far southern highlands of Ecuador and the northern highlands of Peru, due to the topographic privilege of this region - the lowest part of the Andes. The fact that manioc and chili peppers were being cultivated in the far southern highlands by the Middle to

395

Late Formative Period suggests that earlier evidence for these and other crops may be found in the region. Moreover, the presence and cultivation of these crops, as well as the distinction of the Catamayo settlement plan from the eastern (SALF) and coastal lowlands, as discussed in Chapter 4, and the distinctions, yet similarities of the Catamayo ceramics between these regions, shows that the highland people were active participants in trans-Andean interaction spheres. As discussed in Chapter 7, undoubtedly interaction was occurring between all regions of the highlands, the coast, and the eastern lowlands, and further archaeological evidence needs to be integrated into these discussions.

8.5 Conclusions Regarding the Nature of Highland Plant Food Production

Based on the ecological requirement of the plant species present from the highland archaeological sites and from Santa Ana-La Florida, I contend that, during the Formative

Period, a highland, “compact vertical”, agricultural system was practiced north of the

Catamayo Valley during the Formative Period, at Chaullabamba, Cerro Narrío,

Cotocollao, Tajamar, and La Chimba. In contrast, a lowland agricultural system was practiced in the Catamayo Valley and at SALF. While a lowland agricultural complex is not surprising at SALF, my research is the first to suggest that, at least at Trapichillo and

La Vega in the Catamayo Valley, a lowland agricultural system was practiced in the far southern highlands. The sophistication of these adaptive and agricultural strategies signify: 1) an in situ development of highland root-crop cultivation; 2) a greater antiquity of plant cultivation than the Formative Period to acquire and/or domesticate the crops and to learn the ecological requirements of each; and 3) a greater antiquity of maize

396

cultivation in the highlands as maize cultivars needed to be developed that were suited to higher elevations (as discussed above).

In Chapter 7, I also discuss the nutritional benefits and deficiencies of the crops identified. Undoubtedly many other plant foots were cultivated, managed, and/or collected and contributed to the diet in both the higher elevation sites (those north of the

Catamayo Valley) and lower elevation (Catamayo Valley and SALF). While meat may or may not have contributed considerably to the diet in the past, an overall lack of meat would not have been overtly harmful if plant protein and iron sources were adequate.

While protein is low in both highland and tropical root crops, as well as in maize, beans

(Phaseolus) are not the only complementary source of iron and essential amino acids

(protein), and many other foods can contribute to a healthy overall diet. I also suggest that maize and beans need not have formed the staple crops, as other combinations of plant foods can produce an equally nutritious diet. In particular, at high-elevations, above

3000-3100 masl and the limit for maize and bean cultivation, chocho, oca, quinoa and potatoes together provide an adequate diet. I also put forward that the plant food combinations that result in a healthy diet, especially if the diet is low in meat protein, were the result a long period of knowledge-acquisition and development, Thus, I propose that such detailed awareness must have been gained over a considerable period of time.

The sophisticated knowledge of the nutritional complementarities of the plant foods utilized, the variety of ways to prepare (and in some cases detoxify) them for consumption, as well as the complex agricultural production systems allowing for their production in different environments, attest to the wisdom of the highland Formative

397

Period people and do not support previous suggestions that cultural complexity in the highlands was a result of external (coastal origin) stimuli.

8.6 Final Thoughts and Future Research

The primary data used to identify plants in the archaeological record from the sites

I investigated in my dissertation were based primarily on starch analyses of ground stone tool and charred ceramic cooking-pot residues. While my results have considerably advanced our knowledge regarding the timing and nature of plant use and food production, as well as cultural interaction and development in the Ecuadorian Andes, starch analysis alone cannot provide a thorough inventory of economic plant usage. My samples and analyses were biased towards starchy plants that were either ground and/or cooked. Many plants may have been used that either do not store energy as starch, or the plant part utilized does not contain starch, and were, thus, “invisible” to my analyses. In order to overcome this bias and to provide a more thorough overview of highland

Formative Period plant use, I integrated the results from previous research. The botanical record from the Cotocollao site was critical to this dissertation. As noted in Chapter 5, I have retained the samples analyzed for starch granules for other analyses, including phytolith analysis, to address this bias. I also have additional sediment samples, tool and ceramic residue samples, and macrobotanical samples recovered from water flotation for some of the sites investigated, which, when analyzed may greatly add to the results from this dissertation.

The Theobroma identification clearly requires more work to confirm or deny this classification. If confirmed, this may mark the earliest instance of cacao use in the

398

archaeological record in South America. I feel that chocho/tarwi (Lupinus mutabilis) has been overlooked and needs to be better studied, both by examining macrobotanical collection of “legumes” from previous investigations to look for missed identifications, and by careful analysis in future excavations. Expanding my comparative collection of

Andean and Ecuadorian plant species to include more wild species will also help to support the identifications made.

In chapter 7, I also discussed chronological issues for the highland Formative

Period. I suggested that further work to specifically address the timing of the highland

Formative be renewed at key sites, such as Chaullabamba, the Catamayo Valley, the valleys in the Quito area (for Cotocollao-related sites), and, despite the long history of looting and disturbance, at Cerro Narrío as well. Future research should also include surveying areas above 3000 masl. If the cultivation of high-elevation Andean crops preceded maize cultivation in the highlands, then sites should be located in higher elevations better suited to their production. More archaeological research needs to be conducted in the central highlands as well. As discussed in Chapter 3, it is ironic that in the early 20th century the chronology developed for the central highlands formed the basis for understanding all other regions of Ecuador, yet, today, it is the least understood area of the highlands for, especially, the Formative Period. It is also apparent that much more effort needs to be directed towards survey and excavation on the eastern slopes of

Ecuador. The question of maize’s dispersal to the highlands and coastal regions, as well as the full nature of the Ecuadorian Formative Period, cannot be adequately addressed when there is a lacuna of knowledge for this region.

399

Stable isotope analysis of human remains is also much understudied in highland

Ecuador. As discussed in Chapter 7, I cannot infer the relative abundance of the different plant species, nor their relative contribution to the overall diet, based on starch granule analysis. Stable isotope analysis of food residues is almost equally problematic. Both the

Cotocollao and Chaullabamba site excavations resulted in the recovery of human remains, and therefore undertaking stable isotope analysis will greatly aid in determining the importance of maize in the diet of Formative Period people.

As the “genetic revolution” continues to redraw the origins of domesticated plant species in South America, it is more apparent than ever that the southern regions of

Ecuador and northern regions of Peru was a nexus for early plant use, domestication, and cultural interaction, and more research needs to be directed in this regions, especially in the far southern highlands and slopes (both east and west) of Ecuador.

I approached this study from a human behavioural and historical ecological framework, and used the results of the starch granule and previous botanical analyses, as well as the local ecological settings of the archaeological sites, to make inferences about the human-modified landscapes and the food-production strategies used. Further work can be realized in this area as well. By integrating climatic and soil data into the basic

GIS topographic maps, as well as field sampling, where possible for some of the sites, a much better understanding of the local landscapes and agricultural strategies employed will be realized. While landscape studies should be expanded, they should not exclude the people associated with them. Although I integrate historical accounts by Cieza de

León in my discussions, in future research I would like to incorporate more ethnohistoric and ethnographic information. The highland indigenous people altered the land with

400

small-scale intensive agricultural techniques over many generations. The built, anthropogenic environment we see today represents the landscape investment of hundreds of generations of farmers and is a reflection of their rich indigenous knowledge, and these wisdoms needs to be integrated in future research.

When Formative village sites were first discovered in the 1950’s in southwestern

Ecuador that predated similar developments in Mexico and Peru, the crucibles of later state civilizations, diffusionary models were invoked to explain the precocious appearance of social complexity found there. Models of transoceanic origins, external cultural diffusion, and migration have been largely abandoned to explain social complexity in Early Formative coastal Ecuador. In order to fully understand developments occurring on the coast and in the highlands during both the Preceramic and

Formative Periods, we must shed the myopic view of the highlands as being a region not involved in cultural interaction until the Late Formative, and then only as a recipient, and not a contributor, to earlier or contemporary developments.

The Andean region provides an exceptional and important chapter in the course of human history as it is the only high-elevation area of the world where plants were domesticated. As such, farming in the Ecuadorian highlands cannot be understood as an

“offshoot” of a lowland tropical system, and so we must gain a better understanding of the progression to agriculture in the highlands. While my research contributes to this knowledge, there is much more to be done. We need to intensify the search for

Preceramic and Formative sites in both the highlands and eastern lowlands to allow for a more thorough understanding of when plant cultivation begins, the types of plants grown and agricultural strategies employed, and the changes in social formations and cultural

401

interactions that occurred over time. In conclusion, we should no longer think of the

Ecuadorian highlands as a recipient of social, technological and economic innovations from the coast. Rather, we should attempt to understand the Ecuadorian highlands as a unique environment that was involved in multiple interaction spheres, with its own trajectory to complexity. To apply diffusionary models to explain the adoption of agriculture and the emergence of cultural complexity in the Ecuadorian highlands is just as inappropriate here as it was for the Ecuadorian coast. It also denies the surviving highland cultures a history of innovation and cultural sophistication that their predecessors achieved.

402

References Cited

Allem, Antonio C. 1999 The Closest Wild Relative of Cassava (Manihot esculenta Crantz). Euphytica 107:123-133.

Alves, Alfredo A. C. 2002 Cassava Botany and Physiology. In Cassava: Biology, Production and Utilization, edited by R.J. Hillocks, J.M. Thresh, and A.C. Belloti, pp. 67-89. CABI, Wallingford and New York.

Arellano Lopéz, A. Jorge 1994 Loma Pucara: a Formative site in Cebadas Valley, Ecuador. National Geographic Research & Exploration 10(1):118-120.

Athens, J. Stephen 1978 Formative Period Occupations in the Highlands of Northern Ecuador: A comment on Meyers. American Antiquity 43:493-496. 1990a The site of Pumpuesta and the Pastaza Phase in southwestern lowland Ecuador. Ñawpa Pacha 24:111-124. 1990b Prehistoric Agricultural Expansion and Population Growth in Northern Highland Ecuador: Interim Report for 1989 Fieldwork. International Archaeological Research Institute, Honolulu. 1995 Relaciones interregionales prehistóricas en el norte de los Andes: evidencia del sitio La Chimba en el Ecuador septentrional. In Perspectivas Regionales en la Arqueología del Suroccidente de Colombia y Norte del Ecuador, edited by C. Gnecco, pp. 3-29. Editorial Universidad del Cauca, Popayán, Colombia. 1998 Volcanism and Archaeology in the Northern Highlands of Ecuador. In Actividad Volcánica y Pueblos Precolombinos en el Ecuador. Edited by Patricia Mothes, pp. 157- 189. Abya-Yala, Quito.

403

Athens, J. Stephen, and Alan J. Osborn 1974 Archaeological Investigations at Two Ceramic Period Sites in the Highlands of Northern Ecuador: Two Preliminary Reports. Instituto Otavaleño de Antropología, Otavalo, EC.

Baillie, I.C. 1996 Soils of the Humid Tropics. In The Tropical Rainforest: an ecological study, Second edition. Richards, Paul, W., with contributions by R.P.D. Walsh, I.C. Baillie, and P. Greig, pp. 256-286. Cambridge University Press, Cambridge.

Balée, William 1988 Indigenous adaptations to Amazonian palm forests. Principes 32(2):47-54. 1989 The culture of Amazonian forests. Advances in Economic Botany 7:1-21. 1993 Indigenous transformation of Amazonian forests: An example from Maranhão, Brazil. L’Homme 33(2-4):231-254.

Bakker, Johan, Marcela Moscol Olivera, and Henry Hooghiemstra 2008 Holocene environmental change at the upper forest line in northern Ecuador. The Holocene 18:877-893.

Balme, Jane and Wendy E. Beck 2002 Starch and Charcoal: Useful Measures of Activity Areas in Archaeological Rock Shelters. Journal of Archaeological Science 29:157-166.

Balsev, Henrik 1988 Distribution Patterns of Ecuadorian Plant Species. Taxon 37:567-577.

Balyejusa Kizito, Elizabeth, Linley Chiwona-Karltun, Thomas Egwang, Martin Fregene, and Anna Westerbergh

404

2007 Genetic Diversity and Variety Composition of Cassava on Small-scale Farms in Uganda: An Interdisciplinary Study Using Genetic Markers and Farm Interviews. Genetica 130:301-318.

Banks, W. and C. T. Greenwood 1975 Starch and its Components. Edinburgh University Press, Edinburgh.

Barakat, M. R. 1976 Pellagra. In Nutrition and Preventive Medicine: The major deficiency syndromes, epidemiology, and approaches to control, edited by G.H. Beaton and J.M. Bengoa, pp. 126-135. World Health Organization, Geneva.

Barthlott, W., J. Mutke, M.D. Rafiqpoor, G. Kier, and H. Kreft 2005 Global centers of vascular plant diversity. Nova Acta Leopoldina 92:61-83.

Barton, Huw, Phillip J. Piper, Ryan Rabett, and Ian Reeds 2009 Composite hunting techniques from the Terminal Pleistocene and Early Holocene, Niah Cave, Borneo. Journal of Archaeological Science 36:1708-1714.

Barton, Huw, Robin Torrence, and Richard Fullagar 1998 Clues to Stone Tool Function Reexamined: Comparing Starch Grain Frequencies Used and Unused Obsidian Artefacts. Journal of Archaeological Science 25:1231-1238.

Basile, David Giovanni 1974 Tillers of the Andes: Farmers and Farming in the Quito Basin. Studies in Geography No. 8. University of North Carolina at Chapel Hill, Department of Geography, Chapel Hill, NC.

Bayer, C., M. F. Fay, A. Y. De Bruijn, V. Savolainen, C. M. Morton, K. Kubitzki, W. S. Alverson, and M. W. Chase

405

1999 Support for an expanded family concept of Malvaceae within a recircumscribed order Malvales: a combined analysis of plastid atpB and rbcl DNA sequences. Botanical Journal of the Linnean Society 129:267-303.

Beckerman, Stephen 1979 The abundance of protein in Amazonia: A reply to Gross. American Anthropologist 81:533-560.

Bell, Robert E. 1965 Investigaciones arqueológicas en el sitio del El Inga, Ecuador. Casa de la Cultura Ecuatoriana, Quito.

Bennet, Wendell C. 1946 Excavations in the Cuenca Region, Ecuador. Yale University Publications in Anthropology, numbers 35-36. Yale University Press, New Haven.

Bertoft, E. 2004 Analysing starch structure. In Starch in Food: Structure, function and applications, edited by A-C. Eliasson, pp. 57-96. Woodhead and CRC Press, Boca Raton.

Blake, Michael 2006 Dating the Initial Spread of Zea mays. In Histories of Maize: multidisciplinary approaches to the prehistory, linguistics, biogeography, domestication, and evolution of maize, edited by J.E. Staller, R.H. Tykot, and B.F. Benz, pp. 55-72. Academic Press, Burlington, MA.

Blanco, G. O. 1982 Genetic variability of tarwi (Lupinus mutabilis Sweet). In Proceedings of the First International Lupin Conference, pp. 33-50. Lima, Peru.

406

1984 Lupinus mutabilis genetic resources in South America. In Proceedings of the Third International Lupin Conference, pp. 159-164. La Rochelle, France. 1986 Wild stock of Lupinus mutabilis Sweet. In Proceedings of the Fourth International Lupin Conference, pp. 159-164. Geraldton, Australia.

Bou, Lizet Rached, Consuelo A. de Vizcarrondo, Alicia M. Rincón, and Fanny Padilla 2006 Evaluación de harinas y almidones de mapuey (Dioscorea trifida), variedades blanco y morado. Archivos Latinoamericanos de Nutricion 56(4):375-383.

Bousalem, M., G. Arnau, I. Hochu, R. Arnolin, V. Viader, S. Santoni, and J. David 2006 Microsatellite segregation analysis and cytogenetic evidence for tetrasomic inheritance in the American yam Dioscorea trifida and a new basic chromosome number in the Dioscorea. Theoretical and Applied Genetics 113:439-451.

Bowman, Sheridan 1990 Radiocarbon Dating. University of California Press, Berkeley.

Braun, Robert 1982 Cerro Narrío reanalyzed: The Formative as Seen from the Southern Ecuadorian Highlands. In Primer Simposio de Correlaciones Antropológicas Andino- Mesoamericano, edited by Jorge G. Marcos and Presley Norton, pp. 41–119. Escuela Superior Politécnica del Litoral, Guayaquil.

Bronk Ramsey, C. 2009 Bayesian analysis of radiocarbon dates. Radiocarbon 51(1):1023-1045.

Brown, C. H. 2009 Development of Agriculture in Prehistoric Mesoamerica: The linguistic evidence. In Pre-Columbian Foodways, edited by J.E. Staller and M. Carrasco, pp. 71-107. Springer, Berlin.

407

Brubaker, C. L., and J. F. Wendel 1994 Reevaluating the origin of domesticated cotton (Gossypium hirsutum Malvaceae) using nuclear restriction fragment length polymorphisms (RLFP). American Journal of Botany 81:1309-1326.

Bruhns, Karen Olsen 1989 Intercambio entre la Costa y la Sierra en el Formativo Tardío: Nuevas Evidencias del Azuay. In Relaciones Interculturales en el Area Ecuatorial del Pacífico Durante la Epoca Precolombina, edited by J.F. Bouchard and M. Guinea, pp. 57–74. British Archaeological Reports, International Series 503, Oxford. 2003 Social and Cultural Development in the Ecuadorian Highlands and Eastern Lowlands During the Formative. In Archaeology of Formative Ecuador, edited by J. S. Raymond and R. L. Burger, pp. 125-176. Dumbarton Oaks, Washington. D.C.

Bruhns, Karen Olsen, James H. Burton, and George R. Miller 1990 Excavations at Pirincay in the Paute Valley of Ecuador, 1985-1988. Antiquity 64:221-233.

Bruno, Maria C. 2006 A Morphological Approach to Documenting the Domestication of Chenopodium in the Andes. In Documenting Domestication: New Genetic and Archaeological Paradigms, edited by M. Zeder, E. Emshwiller, D. Bradley, and B.D. Smith, pp. 32-45. University of California Press, Berkeley and Los Angeles.

Bruno, Maria C., and William T. Whitehead 2003 Chenopodium Cultivation and Formative Period Agriculture at Chiripa, Bolivia. Latin American Antiquity 14:339-355.

Brush, Stephen B. 1976 Man’s use of an Andean ecosystem. Human Ecology 4(2):147-146.

408

1977 Mountain, Field and Family, the Economy and Human Ecology of an Andean Valley. University of Pennsylvania Press, Philadelphia.

Bryant, Vaughn M., Jr. 1989 Botanical Remains in Archaeological Sites. In Interdisciplinary Workshop of the Physical-Chemical-Biological Processes Affecting Archaeological Sites, compiled by Christopher C. Methewson, pp. 85-115. Environmental Impact Research Program Contract Report E11-89-1. U.S. Army Corps of Engineers, Washington, D.C.

Buckalew, James O., Maurice James, Lisa Scott, and Paul Reed 1998 Water Resources Assessment of Ecuador. US Army Corps of Engineers, US Army Topographic Engineering Center. Electronic document, http://www.sam.usace.army.mil/en/wra/Ecuador/Ecuador%20WRA%20English.pdf, accessed March 2, 2012.

Buitrón, Aníbal 1952 The Sierra. In Ecuador: Andean Mosaic, edited by R. Blomberg, pp. 128-139. Hugo Gebers Förlag, Stockholm.

Burleigh, Richard, and Andrew Hewson 1979 British Museum Natural Radiocarbon Measurements XI. Radiocarbon 21(3):339- 352.

Burleigh, Richard, Andrew Hewson, and Nigel Meeks 1977 British Museum Natural Radiocarbon Measurements IX. Radiocarbon 19(2):143- 160.

Bush, M.B., J.A. Hanselman, and H. Hooghiemstra 2007 Andean montane forests and climate change. In Tropical Rainforest Responses to Climate Change, edited by M.B. Bush and J.R. Flenley, pp. 33-54. Springer, New York.

409

Bush, Mark B., Dolores R. Piperno, and Paul A. Colinvaux 1989 A 6,000 year history of Amazonian maize cultivation. Nature 340:303-305.

Cabral Velho, Christiane, Anna Whipkey, and Jules Janick 1990 Cupuassu: A New Beverage Crop for Brazil. In Advances in New Crops, edited by J. Janick and J.E. Simon, pp. 372-375. Timber Press, Portland.

Calvert, Paul 1997 The structure of starch. Nature 389:338-339.

Campbell, Neil A., Jane B. Reece and Lawrence G. Mitchell 1999 Biology. Fifth edition. Benjamin/Cummings, Menlo Park, CA.

Campbell-Platt, Geoffrey 2000 Fermented Foods: Origins and Applications. In Encyclopedia of Food Microbiology, Volume 2, edited by R.K. Robinson, pp. 736-739. Academic Press, San Diego.

Carneiro, Robert L. 1960 Slash and burn agriculture: a closer look at its implications for settlement patterns. In Men and Cultures, edited by A. Wallace, pp. 229-232. University of Pennsylvania Press, Philadelphia. 2000 The Evolution of the Tipití: A Study in the Process of Invention. In Cultural Evolution: Contemporary Viewpoints, edited by G.M. Feinman and L. Manzanilla, pp. 61-93. Kluwer Academic, New York.

Carvalho, Luis J. C. B., and Barbara A. Schaal 2001 Assessing Genetic Diversity in the Cassava (Manihot esculenta Crantz) Germplasm Collection in Brazil Using PCR-based Markers. Euphytica 120:133-142.

410

Caviedes, César, and Gregory Knapp 1995 Strategies for Agricultural Subsistence. In South America, edited by C. Caviedes and G. Knapp, pp. 162-183. Prentiss Hall, Englewood Cliffs, NJ.

Cerruto Torrico, Noelia E. 2010 Defining and Defending Andean Water Territories: a case of the community of La Chimba, in the Páramo of Cayambe, Ecuador. Unpublished M.Sc. thesis, Wageningen University, Wageningen, Netherlands.

Chacón, S. M. I., B. Pickersgill, and D. G. Debouck 2005 Domestication patterns in common bean (Phaseolus vulgaris L.) and the origin of the Mesoamerican and Andean cultivated races. Theoretical and Applied Genetics 110:432-444.

Chandler-Ezell, Karol, and Deborah M. Pearsall 2003 “Piggyback” microfossil processing: Joint starch and phytolith sampling from stone tools. The Phytolitharien 15:2-8.

Chandler-Ezell, Karol, Deborah M. Pearsall, and James Zeidler 2006 Root and Tuber Phytoliths and Starch Grains Document Manioc (Manihot esculenta), Arrowroot (Maranta arundinacea), and Llerén (Calathea sp.) at the Real Alto Site, Ecuador. Economic Botany 60:103-120.

Chardon, Roland 1980 The Elusive Spanish League: A Problem of Measurement in Sixteenth-Century New Spain. Hispanic American Historical Review 60(2):294-302.

Cheesman, E. E. 1944 Notes on the nomenclature, classification and possible relationships of cocoa populations. Tropical Agriculture 21:144-159.

411

Chu, Edison P., and R.C.L. Figueiredo-Ribeiro 1991 Native and Exotic Species of Dioscorea Used as Food in Brazil. Economic Botany 45:467-479.

Cieza de León, Pedro de 1864 La Crónica del Perú. The travels of Pedro de Cieza de León, A.D. 1532-50, contained in the first part of his Chronicle of Peru. Translated and edited by Clements R. Markham, The Hakluyt Society, no. 33, London.

Cisneros, F., P. Torres and J. Feyen 2001 Quantitative analysis of the performance of sprinkler irrigation in the southern Sierra of Ecuador, on slopes steeper than 12%. Anales de la Universidad de Cuenca – Ciencias Naturales 2:39-56.

Clawson, David L. 2000 Latin America and the Caribbean: Lands and People, Second edition. McGraw Hill, Boston.

Clement, Charles R., Michelly de Cristo-Araújo, Geo Coppens d’Eechenbrugge, Alessandro Alves Pereira, and Doriane Picanço-Rodrigues 2010 Origin and Domestication of Native Amazonian Crops. Diversity 2:72-106.

Cobley, Leslie S. 1976 An Introduction to the Botany of Tropical Crops, second edition. Revised by W.M. Steele, Longman, London and New York.

Cochrane, T.T., L.G. Sánchez, L.G. de Azevedo, J.A. Porras, and C.L. Garver 1985 Land in Tropical America. Vol. 1: A Guide to Climate, Landscapes, and Soils for Agronomists in Amazonia, the Andean Piedmont, Central Brazil and Orinoco. Centro Internacional de Agricultura Tropical (CIAT), Cali, Colombia.

412

Coe, Sophie D. 1994 Americas First Cuisines. University of Texas Press, Austin.

Coe, Sophie D., and Michael D. Coe 2007 The True History of Chocolate, second edition. Thames and Hudson, London.

Collier, Donald 1946 The Archaeology of Ecuador. In Handbook of South American Indians, Bulletin no. 143, volume 2. Bureau of American Ethnology, Washington, D.C.

Collier, Donald, and John V. Murra 1943 Survey and Excavations in Southern Ecuador. Field Museum of Natural History Anthropological Series 35, Publication 528, Chicago.

Collins, Matthew, and Les Copeland 2011 Ancient Starch: Cooked or just old? Letter to Proceedings of the National Academy of Sciences of the United States of America 108:E145.

Colinvaux, P. A., K. Olson, and Kam-Biu Liu 1988 Late-Glacial and Holocene Pollen Diagrams from Two Endoheic Lakes of the Inter-Andean Plateau of Ecuador. Review of Palaeobotany and Palynology 55:83-99.

Cortella, Alicia R. and Maria L. Pochettino 1994 Starch Grain Analysis as a Microscopic Diagnostic Feature in the Identification of Plant Material. Economic Botany 48:171-181.

Crumley, Carole L. 1994 Historical Ecology: A Multidimensioinal Ecological Orientation. In Historical Ecology: Cultural Knowledge and Changing Landscapes, edited by C.L. Crumley, pp. 1- 16. School for Advanced Research Press, Santa Fe.

413

Czaja, A. Th. 1978 Structure of Starch Grains and the Classification of Vascular Plant Families. Taxon 27:463-470.

Dave, J., M.H. Wilcox, and M. Kellett 1999 Glove powder: implications for infection control. Journal of Hospital Infection 42:283-285.

DeBoer, Warren R., Keith Kintigh, and Arthur Rostoker 1996 Ceramic seriation and settlement reoccupation in lowland South America. Latin American Antiquity 7:263-278.

DeMaeyer, E. M. 1976 Protein-energy Malnutrition. In Nutrition and Preventive Medicine: The major deficiency syndromes, epidemiology, and approaches to control, edited by G.H. Beaton and J.M. Bengoa, pp. 23-54. World Health Organization, Geneva.

Deneven, William M. 1992 The pristine myth: the landscape of the Americas in 1492. Annals of the Association of American Geographers 82:369-385. 2001 Cultivated Landscapes of Native Amazonia and the Andes. Oxford University Press, Oxford.

DeNiro, M.J. and S. Epstein 1978 Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42:495-506.

414

Dickau, Ruth 2005 Resource use, crop dispersals, and the transition to agriculture in prehistoric Panama: Evidence from starch grains and macroremains. Unpublished Ph.D. dissertation, Department of Anthropology, Temple University, Philadelphia.

Dickau, Ruth, Maria C. Bruno, José Iriarte, Heiko Prümers, Carla Jaimes Betancourt, Irene Holst, and Francis E. Mayle 2012-06-14 Diversity of cultivars and other plant resources used at habitation sites in the Llanos de Mojos, Beni, Bolivia: evidence from macrobotanical remains, starch grains, and phytoliths. Journal of Archaeological Science 39:357-370.

Dickau, Ruth, Anthony J. Ranere, and Richard G. Cooke 2007 Starch grain evidence for the preceramic dispersals of maize and root crops into tropical dry and humid forests of Panama. Proceedings of the National Academy of Sciences of the United States of America 104:3651-3656.

Dillehay, Tom D., Jack Rossen, Thomas C. Andres, and David E. Williams 2007 Preceramic Adoption of Peanut, Squash, and Cotton in Northern Peru. Science 316:1890-1893.

Doebley, J. F. 1992 Molecular systematics and crop evolution. In Molecular Systematics of Plants, edited by D. Soltis, P. Soltis and J. Doyle, pp. 202-222. Chapman-Hall, New York.

Domínguez, Victoria, Rosalba Chacón, Fernando Mejía, Francisco Sánchez, and Alexia Ibarra 2003 Identificacíon de Zonas Arqueológicas en el Area Urbana de Quito – Fondo de Salvamento del Patrimonio Cultural del Distrito Metropolitan de Quito. Report on file, Instituto National de Patrimonio Cultural, Quito, EC.

415

Donald, M.A. 2004 Understanding starch structure and functionality. In Starch in Food, edited by Ann-Charlotte Eliasson, pp. 156-184. Woodhead, Cambridge.

Dreiss, Meredith L., and Sharon Edgar Greenhill 2008 Chocolate: Pathway to the Gods. The University of Arizona Press, Tucson.

Drummond, C. S. 2008 Diversification of Lupinus (Leguminoseae) in the western New World: derived evolution of perennial life history and colonisation of montane habitats. Molecular Phylogenetics and Evolution 48:408-421.

Duncan, Neil A,, Deborah M. Pearsall, and Robert A. Benfer, Jr. 2009 Gourd and squash artifacts yield starch grains of feasting foods from preceramic Peru. Proceedings of the National Academy of Sciences of the United States of America 106:13202-13206.

Dunnell, Robert C. 1989 Aspects of the application of evolutionary theory in archaeology. In Archaeological Thought in America, edited by C. Lamberg-Karlovsky, pp. 35-49. Cambridge University Press, New York.

Eastwood, Ruth J., and Colin E. Hughes 2008 Origins of Domestication of Lupinus mutabilis in the Andes. In Lupins for Health and Wealth: Proceedings of the 12th International Lupin Conference, edited by J.A. Palta and J.B. Berger, pp. 373-379. International Lupin Association, Canterbury, NZ.

416

Eckhoff, Steven R., and Stanley A. Watson 2009 Corn and Sorghum Starches: Production. In Starch Chemistry and Technology, Third edition, edited by J. BeMiller and R. Whistler, pp. 373-439. Academic Press, Boston.

Emshwiller, Eve 2002 Biogeography of the Oxalis tuberosa alliance. Botanical Review 68:128-152. 2006 Origins of Polyploid Crops: The Example of the Octoploid Tuber Crop Oxalis tuberosa. In Documenting Domestication: New Genetic and Archaeological Paradigms, edited by M. Zeder, E. Emshwiller, D. Bradley, and B.D. Smith, pp. 153168. University of California Press, Berkeley and Los Angeles.

Emshwiller, Eve, Terra Theim, Alfredo Grau, Victor Nina, and Franz Terrazas 2009 Origins of domestication and polyploidy in oca (Oxalis tuberosa; Oxalidaceae. 3. AFLP data of oca and four wild, tuber-bearing taxa. American Journal of Botany 2009:1839-1848.

Erickson, Clark 1995 Archaeological perspectives on ancient landscapes of the Llanos de Mojos in the Bolivian Amazon. In Archaeology in the American Tropics: Current Analytical Methods and Applications, edited by Peter W. Stahl, pp. 66-95. Cambridge University Press, Cambridge. 2006 The domesticated landscapes of the Bolivian Amazon. In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands, edited by W. Balée and C. Erickson, pp. 187-234. Columbia University Press, New York. 2008 Amazonia: The Historical Ecology of a Domesticated Landscape. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 157-183. Springer, New York.

417

Esau, Katherine 1965 Vascular Differentiation in Plants. Holt, Rinehart & Winston, New York.

Estrada, Emilio 1956 Valdivia, un Sitio Arqueológico Formativo en el Provincia del Guayas, Ecuador. Museo Arqueológico Victor Emilio Estrada, Publicación No. 1, Guayaquil. 1958 Las Culturas pre-clásicas, formativas o arcaicas del Ecuador. Museo Arqueológico Victor Emilio Estrada, Publicación No. 5, Guayaquil.

Evans, Clifford, and Betty J. Meggers 1954 Informe preliminary sobre las investigaciones arqueológicas realizidas en el cuenca del Guayas, Ecuador. Cuadernos de Historia y Arqueología 4(12):307-336. 1957 Formative Period Cultures in the Guayas Basin, Coastal Ecuador. American Antiquity 22:235-247.

Evans, Lance 2010 Excavation Report: Cerro Narrío 2009 Field Season (FS-09): Unit 7. Report on file, Department of Archaeology, University of Calgary, Calgary.

Ferdon, Edwin N. 1940 The Archaeological Survey of Ecuador. El Palacio 47:137-145.

Ford, Richard 1979 Paleoethnobotany in American Archaeology. In Advances in Archaeological Method and Theory, edited by M. Schiffer, pp. 285-336. vol. 2. Academic Press, Orlando.

Frisch, Johan Dalgas, and Christian Dalgas Frisch 2004 Aves Brasileiras e Plantas que as atraem, 3rd edition. Dalgas Ecotec, São Paulo.

418

Gade, Daniel W. 1970 Ethnobotany of cañihua (Chenopodium pallidicaule), rustic seed crop of the Altiplano. Economic Botany 24:55-61.

Gallegos, Rocío Alarcón 1988 Etnobotanica de los Quichuas de la Amazonia Ecuatoriana. Museos del Banco Central del Ecuador, Guayaquil.

Gentry, A.H. 1988 Changes in plant community diversity and floristic composition on environmental and geographical gradients. Annals of the Missouri Botanical Gardens 75:1-34.

Gepts, P., and D. Debouck 1991 Origin, domestication and evolution of the common bean (Phaseolus vulgaris L.) In Common Beans: research for crop improvement, edited by A. Schoonhoven and and O. Voyest, pp. 7-53, CABI, Wallingford, UK.

Gomis, Dominique 1999 La ceramic Formativa Tardía de la sierra austral del Ecuador. In Formativo sudamericano: una revaluación, edited by P. Ledergerber-Crespo, pp. 139-159. Abya- Yala, Quito, EC.

Gott, Beth, Huw Barton, Delwen Samuel, and Robin Torrence 2006 Biology of Starch. In Ancient Starch Research. Edited by R. Torrence and H. Barton, pp. 35-45. Left Coast Press, Walnut Creek, CA.

Grieder, Terence, James D. Farmer, Antonio Carrillo, and Brandford M. Jones 2002 Art and Prestige among noble houses of the equatorial Andes. In Andean Archaeology II: Art, Landscape and Society, edited by Helaine Silverman and William H. Isbell, pp. 157-177. Kluwer Academic, New York.

419

Grieder, Terrence, with James D. Farmer, David V. Hill, Peter W. Stahl, and Douglas H. Ubelaker 2009 Art and Archaeology of Chaullabamba, Ecuador. University of Texas Press, Austin.

Grobman, Alexander, Duccio Bonavia, Tom D. Dillehay, Dolores R. Piperno, José Iriarte, and Irene Holst 2012 Preceramic maize from Paredones and Huaca Prieta, Peru. Proceedings of the National Academy of Sciences of the United States of America 109:1755-1759.

Gross, Daniel 1975 Protein capture and cultural development in the Amazon Basin. American Anthropologist 77:526-549.

Guffroy, Jean 2004 Catamayo Precolumbino: Investigaciones arqueológicas en la provincia de Loja (Ecuador). IRD, IFEA, UTPL, BCE, Loja.

Guilbot, André and Christiane Mercier 1985 Starch. In The Polysaccharides, Vol. 3, edited by G. O. Aspinall, pp. 210- 282. Academic Press, London.

Gutiérrez-Salgado, A., P. Gepts, and P. Debouck 1995 Evidence for two gene pools of the lima bean, Phaseolus lunatus L., in the Americas. Genetic Resources and Crop Evolution 42:15-28.

Hall, M.L., P. Samaniego, J.L. Le Pennec, and J.B. Johnson 2008 Ecuadorian Andes volcanism: A review of Late Pliocene to present activity. Journal of Volcanology and Geothermal Research 176:1-6.

420

Hancock, James F. 2004 Plant Evolution and the Origin of Crop Species. Second ed. CABI, Cambridge, MA.

Hansen, B.C.S., D.T. Rodbell, G.O. Seltzer, B. Leon, K.R. Young, and M. Abbott 2003 Late-Glacial and Holocene vegetational history from two sites in the western Cordillera of southwestern Ecuador. Palaeo 194:79-108.

Hardy, Karen, Tony Blakeney, Les Copeland, Jennifer Kirkham, Richard Wrangham, and Matthew Collins 2009 Starch granules, dental calculus and new perspectives on ancient diet. Journal of Archaeological Science 36:248-255.

Hargrave, Bob 2006a Phaseolus lunatus Plant Information Sheet. Educational Resources for Hunger Organization (ECHO), Fort Meyers. 2006b Phaseolus vulgaris (pole type) Plant Information Sheet. Educational Resources for Hunger Organization (ECHO), Fort Meyers. 2006c Phaseolus vulgaris (Nuñas, popping beans) Plant Information Sheet. Educational Resources for Hunger Organization (ECHO), Fort Meyers.

Harlan, Jack R. 1971 Agricultural Origins: centers and noncenters. Science 174:468-474. 1975 Crops and Man. American Society of Agronomy, Madison, WI.

Harris, Marvin 1984 Animal capture and Yanomamo warfare: Retrospect and new evidence. Journal of Anthropological Research 40:183-201.

421

Hart, Thomas C. 2011 Evaluating the usefulness of phytoliths and starch grains found on survey artifacts. Journal of Archaeological Science 38:3244-3253.

Hart, John P., William A. Lovis, J. K. Schulenberg, and Gerald R. Urquhart 2007 Paleodietary implications from stable carbon isotope analysis of experimental cooking residues. Journal of Archaeological Science 34:804-813.

Hart, John P., Gerald R. Urquhart, Roberts S. Feranec, and William A. Lovis 2009 Non-linear relationship between bulk δ13C and percent maize in carbonized cooking residues and the potential of false-negatives in detecting maize. Journal of Archaeological Science 36:2206-2212.

Haslam, Michael 2004 The decomposition of starch grains in soils: implications for archaeological residue analyses. Journal of Archaeological Science 31:1715-1734.

Hastorf, Christine A. 1993 Agriculture and the Onset of Political Inequality before the Inka. Cambridge University Press, Cambridge.

Hastorf, Christine A. and Virginia S. Popper 1988 Current Paleoethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains. University of Chicago Press, Chicago.

Hawkes, J. G. 1989 The domestication of roots and tubers in the American tropics. In Foraging and Farming: The Evolution of Plant Exploitation, edited by D.R. Harris and G.C. Hillman, pp. 481-503. Unwin Hyman, London.

422

Hawkes, K., J. F. O'Connell and L. Rogers 1997 The behavioral ecology of modern hunter-gatherers and human evolution. Trends in Ecological Evolution 12:29-32.

Heckenberger, Michael J. 1998 Manioc agriculture and sedentism in Amazonia: the Upper Xingu example. Antiquity 72:633-848.

Heckenberger, Michael, Afukaka Kuikuro, Urissapá Tabata Kuikuro, J. Christian Russel, Morgan Schmidt, Carlos Fausto, and Bruna Franchetto 2003 Amazonia 1492: pristine forest or cultural parkland? Science 301:1710-1714.

Heiser, C. B. Jr. 1995 Peppers. Capsicum (Solanaceae). In Evolution of Crop Plants, second ed., edited by J. Smartt and N.W. Simmonds, pp. 449-451. Longman Scientific and Technical, London.

Henderson, John S., and Rosemary A. Joyce 2006 Brewing distinction: The development of cacao beverages in formative Mesoamerica. In Chocolate in Mesoamerica: A culture history of cacao, edited by C.L. McNeil, pp. 140-153. University of Florida Press, Gainesville.

Henderson, John S., Rosemary A. Joyce, Gretchen R. Hall, W. Jeffrey Hurst, and Patrick E. McGovern 2007 Chemical and archaeological evidence for the earliest cacao beverages. Proceedings of the National Academy of Sciences of the United States of America 104:18937-18940.

423

Henry, Amanda G., Holly F. Hudson, and Dolores R. Piperno 2009 Changes in starch grain morphologies from cooking. Journal of Archaeological Science 36:915-922.

Henry, Amanda G., and Dolores R. Piperno 2008 Using plant microfossils from dental calculus to recover human diet: a case study from Tell al-Raqā’i, Syria. Journal of Archaeological Science 35:1943-1950.

Hernández-Lauzardo, Ana Niurka, Guadalupe Méndez-Montealvo, Miguel Gerardo Velázquez del Valle, Javier Solorza-Feria, and Luis Arturo Bello-Pérez 2004 Isolation and Partial Characterization of Mexican Oxalis tuberosa Starch. Starch/Stärke 56:357-363.

Hickey, Michael, and Clive King 2000 The Cambridge Illustrated Glossary of Botanical Terms. Cambridge University Press, Cambridge.

Hill, Betsy D. 1975 A New Chronology of the Valdivia Ceramic Complex from the Coastal Zone of Guayas Province, Ecuador. Ñawpa Pacha 10/12:1-39.

Holdridge, L. R., Laurence V. Teesdale, J. Edson Myer, Elbert L. Little Jr., Eugene F. Horn, and José Marrero 1947 The Forests of Western and Central Ecuador. Forest Service, United States Department of Agriculture, Washington, D. C.

Horrocks, Mark, and Simon B. Best 2004 Analysis of Plant Microfossils in Early European Latrine Fills from Russel, Northern New Zealand. International Journal of Historical Archaeology 8(4):267-280.

424

Horrocks, Mark, Matthew Campbell, and Rebecca Conway 2007 A short note on starch and xylem of sweet potato (Ipomoea batatas) and bracken (Pteridium esculentum) in archaeological deposits from northern New Zealand. Journal of Archaeological Science 34:1441-1448.

Hughes, C.E., and R.J. Eastwood 2006 Island radiation on a continental scale: exceptional rates of species diversification after uplift of the Andes. Proceedings of the National Academy of Sciences of the United States of America 103:10334-10339.

Hyslop, John 1984 The Inka Road System. Academic Press, New York.

Idrovo Urigüen, Jaime 1999 El Formativo en la Sierra Ecuatoriana. In Formativo Sudamericano: Una Revaluación, edited by P. Ledergerber-Crespo, pp. 114-123. Abya-Yala, Quito.

Iqbal, M. J., O. U. K. Reddy, K. M. El-Zik, and A. E. Pepper 2001 A genetic bottleneck in the ‘evolution under domestication’ of upland cotton Gossypium hirsutum L. examined using DNA fingerprinting. Theoretical and Applied Genetics 103:547-554.

Isendahl, Christian 2011 The domestication and early spread of manioc (Manihot esculenta Crantz): a brief synthesis. Latin American Antiquity 22:452-468.

Isaacson, John S., and James A. Zeidler 1998 Accidental History: Volcanic Activity and the End of the Formative in Northwestern Ecuador. In Actividad Volcánica y Pueblos Precolombinos en el Ecuador, edited by P. Mothes, pp.41-72. Abya-Yala, Quito.

425

James, Douglas W., Jack Preiss, and Alan D. Elbein 1985 Biosynthesis of Polysaccharides. In The Polysaccharides, Vol. 3, edited by G. O. Aspinall, pp. 107-207. Academic Press, London.

Jara, Holguer, and Alfredo Santamaria 2009 Atlas Arqueológico: Distrito Metropolitano de Quito, Volume 1. FONSAL, Quito, EC.

Jijón y Caamaño, Jacinto 1927 Puruhá: contribución al conocimiento de los aborígenes de la provincia del Chimborazo. Two vols. Boletín de la Academia Nacional de Historia, Quito.

Jetz, W., D. S. Wilcove, and A. P. Dobson 2007 Projected impacts of climate and land-use change on the global diversity of birds. PLoS Biology 5:1211-1219.

Jørgensen, Peter M., and Susana León-Yánez 1999 Catalogue of the Vascular Plants of Ecuador. Missouri Botanical Garden Press, St. Louis.

Jørgensen, Peter M., Carmen Ulloa Ulloa, Blanca León, Susana León-Yánez, Stephen G. Beck, Michael Nee, James L. Zarucchi, Marcela Celis, Rodrigo Bernal, and Robbert Gradstein 2011 Regional Patterns of Vascular Plant Diversity and Endemism. In Climate Change and Biodiversity in the Tropical Andes, edited by S.K. Herzog, R. Martinez, P.M. Jørgensen, and H. Tiessen, pp. 192-203. IAI and SCOPE, São José dos Campos, Brazil.

Jokisch, Brad D. 2002 Migration and Agricultural Change: the Case of Smallholder Agriculture in Highland Ecuador. Human Ecology 30:523-550.

426

Josse, Carmen, Francisco Cuesta, Gonzalo Navarro, Víctor Barrena, María Teresa Becerra, Edersson Cabrera, Eulogio Chacón-Moreno, Wanderley Ferreira, Manuel Peralvo, José Saito, Antonio Tovar, and Luis G. Naranjo 2011 Physical Geography and Ecosystems in the Tropical Andes. In Climate Change and Biodiversity in the Tropical Andes. Edited by S.K. Herzog, R. Martinez, P.M. Jørgensen, and H. Tiessen, pp. 152-169. IAI and SCOPE, São José dos Campos, Brazil.

Kadane, Joseph B. 1988 Possible Statistical Contributions to Paleoethnobotany. In Current Paleoethnobotany, edited by Christine A. Hastorf and Virginia S. Popper, pp. 206-214. University of Chicago Press, Chicago.

Kaplan, Lawrence, and Thomas F. Lynch 1999 Phaseolus (Fabaceae) in Archaeology: AMS Radiocarbon Dates and Their Significance for Pre-Columbian Agriculture. Economic Botany 53:261-272.

Katzenberg, M. Anne, Henry P. Schwarcz, Martin Knyf, and F. Jerome Melbye 1995 Stable Isotope Evidence for Maize Horticulture and Paleodiet in Southern Ontario, Canada. American Antiquity 60:335-350.

Kesler, C. C., and W. G. Bechtel 1953 Physical Methods of Characterising Starch. In Starch and its Derivatives, vol.2, Third ed., edited by J. A. Radley, pp. 402-437. Chapman and Hall, London.

Kessler, Michael, Jon-Arvid Grynes, Stephen R. P, Halloy, Jürgen Kluge, Thorsten Krömer, Blanca León, Manuel J. Macía, and Kenneth R. Young 2011 Gradients of Plant Diversity: Local Pattern and Processes. In Climate Change and Biodiversity in the Tropical Andes, edited by S.K. Herzog, R. Martinez, P.M. Jørgensen, and H. Tiessen, pp. 97-109. IAI and SCOPE, São José dos Campos, Brazil.

427

Korstanje, M.A., and M.P. Babot 2007 Microfossils characterization from South American Andean economic plants. In Plants, People and Places: Recent Studies in Phytolith Analysis, edited by M. Madella and D. Zurro, pp. 41-72. Oxbow, Oxford.

Knapp, Sandra 2002 The Solanaceae and Drugs and Medicine: a Natural History of the Potato Family. Electronic document, http://www.fathom.com/feature/122038/index.html, accessed Feb.18, 2012. Fathom Archive, Columbia University. 2005 The Solanaceae as Food: a Natural History of the Potato Family. Electronic document, http://www.nhm.ac.uk/resources-rx/files/16feat_nat_hist_potato_family- 3062.pdf, accessed Feb. 18, 2012. Natural History Museum, London.

Krebs, J. R. and N. B. Davies 1993 An Introduction to Behavioural Ecology. 3rd ed. Blackwell, Oxford.

Langer, R. M. H., and G. D. Hill 1991 Agricultural Plants, second edition. Cambridge University Press, Cambridge.

Lathrap, Donald W. 1968 The hunting economies of the tropical forest zone of South America: an attempt at historical perspective. In Man the Hunter, edited by R.B. Lee and I. DeVore, pp. 23-29. Aldine, Chicago. 1970 The Upper Amazon. Praeger, New York.

Lathrap, Donald W., Donald Collier, and Helen Chandra 1975 Ancient Ecuador: Culture, Clay and Creativity, 3000-300 B.C. Field Museum of Natural History, Chicago.

428

Lathrap, Donald W., Jorge G. Marcos, James A. Zeidler 1977 Real Alto, and Ancient Ceremonial Center. Archaeology 30(1):2-13.

Laurence, Andrew R., Alston V. Thoms, Vaughn M. Bryant, and Cassandra McDonough 2011 Airborne Starch Granules as a Potential Contamination Source at Archaeological Sites. Journal of Ethnobiology 31(2):213-232.

Laurière, Michel, Peter Gorner, Isabel Bouchez-Mahiout, Richard Wrobel, Christine Breton, Jean-Franxois Fabries, and Dominique Choudat 2008 Physical and Biochemical Properties of Airborne Flour Particles Involved in Occupational Asthma. The Annals of Occupational Hygiene 2008:1-11.

Lentfer, Carol, Michael Therin, and Robin Torrence 2002 Starch grains and environmental reconstruction: a modern test case for West New Britain, Papua New Guinea. Journal of Archaeological Science 29: 687-698.

Lentz, David L. 1999 Plant Resources of the Ancient Maya: The Paleoethnobotanical Evidence. In Reconstructing Ancient Maya Diet, edited by C.D. White, pp. 3-18. University of Utah Press, Salt Lake City.

Lippi, Ronald D. 2003 Appendix B: Formative Period Chronology for the Northern and Central Highlands of Ecuador. In The Archaeology of Formative Ecuador, edited by J.S. Raymond and R.L. Burger, pp. 529-537. Dumbarton Oaks, Washington, D.C.

Lisińska, Graźyna, and Wacław Leszczyński 1989 Potato Science and Technology. Elsevier Science Publishing, New York, NY.

429

Loaiza-Figueroa, F., K. Ritland, J. A. Laborde Cancino, S. D. Tanksley 1989 Patterns of genetic variation of the genus Capsicum (Solanaceae) in Mexico. Plant Systematics and Evolution 165:159-188.

Lovis, William A., Gerald R. Urquhart, Maria E. Raviele, and John P. Hart 2011 Hardwood ash nixtamalization may lead to false negatives for the presence of maize by depleted bulk δ13C in carbonized residues. Journal of Archaeological Science 38:2726-2730.

Loy, Thomas H. 1994 Methods in the Analysis of Starch Residues on Prehistoric Stone Tools. In Tropical Archaeobotany: Applications and New Developments, edited by J. G. Hather, pp. 86-114. Routledge, London.

Loy, Thomas and Huw Barton 2006 Post-excavation contamination and measures for prevention. In Ancient Starch Research. Edited by R. Torrence and H. Barton, pp. 165-167. Left Coast Press, Walnut Creek, CA.

Loy, Thomas H., M. Spriggs, and S. Wickler 1992 Direct evidence for human use of plants 28,000 years ago: Starch residues on stone artefacts from the northern Solomon Islands. Antiquity 66:898-912.

Lisińska, Grażyna and Wacław Leszczyński 1989 Potato Science and Processing. Elsevier Science Publishing, New York.

Lynch, Thomas F. 1983 The Paleo-Indians. In Ancient South Americans, edited by J. Jennings, pp. 87-137. W. H. Freeman, New York. 1988 Chobshi Cave in Retrospect. Andean Past 2:1-32.

430

MacLeod, M. J., W. H. Eshbaugh, and S. I. Guttman 1982 Early evolution of chili peppers (Capsicum). Economic Botany 36:361-368.

MacMasters, M. M. 1964 Microscopic Techniques for Determining Starch Granule Properties. In Methods in Carbohydrate Chemistry, edited by R. L. Whistler. Vol. IV Starch. Academic Press Inc., New York.

Maine, M.J., C.P. Carroll, H.E. Stewart, R.M. Solomon, and R.L. Wastie 1993 Disease resistance in Solanum phujera and diploid and tetraploid S. tuberosum x S. phujera hybrids. Potato Research 36:21-28.

Malainey, Mary 2011 Food-Related Pottery Residues. In A Consumer’s Guide to Archaeological Science: Analytical Techniques, Manuals in Archaeological Method, Theory and Technique, Part 3, edited by M. Malainey, pp. 319-331. Springer, New York.

Marcos, Jorge G. 2003 A Reassessment of the Ecuadorian Formative. In Archaeology of Formative Ecuador, edited by J. S. Raymond and R. L. Burger, pp. 7-32. Dumbarton Oaks, Washington, D.C.

Marcos, Jorge G., and Adam Michzynski 1996 Good Dates and Bad Dates in Ecuador; Radiocarbon Samples and Archaeological Excavations: A Commentary Based on the “Valdivia Absolute Chronology.” Proceedings of the Third Annual Latin American Congress of the University Varsovia, Varsovia and Biskupin-Wenecja. Andean Archaeological Mission of the Institute of Archaeology, Warsaw University. http://www.maa.uw.edu.pl/obp/MICH_MAR.HTM

431

Masucci, Maria A. 2008 Early Regional Polities of Coastal Ecuador. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 489-503. Springer, New York.

Matsuoka, Yoshihiro, Yves Vigouroux, Major M. Goodman. Jesus Sanchez G., Edward Buckler, and John Doebley 2002 A single domestication for maize shown by multilocus microsatellite genotyping. Proceedings of the National Academy of Sciences of the United States of America 99:6080-6084.

Mayer-Oakes, William J. 1986 El Inga, a Paleo-Indian Site in the Sierra of Northern Ecuador. Transactions of the American Philosophical Society 76(4):1-235.

McEwan, Colin, and Florencio Delgado-Espinoza 2008 Late Pre-Hispanic Polities of Coastal Ecuador. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 505-525. Springer, New York.

McEwan, Gordon F. 2006 The Incas: New Perspectives. W.W. Norton and Company, New York.

Meggers, Betty J. 1954 Environmental limitations on the development of culture. American Anthropologist 56:801-824. 1966 Ecuador. Praeger, New York.

Meggers, Betty J., Clifford Evans, and Emilio Estrada 1965 The Early Formative Period of Coastal Ecuador: The Valdivia and Machalilla Phases. Smithsonian Contributions to Anthropology 1. Smithsonian Institution Press, Washington, D.C.

432

Mercader, Julio 2009 Mozambican Grass Seed Consumption during the Middle Stone Age. Science 326:1680-1683.

Mercader, Julio, Tim Bennet, and Mussa Raja 2008 Middle Stone Age starch acquisition in the Niassa Rift, Mozambique. Quaternary Research 70:283-300.

Meyers, Thomas P. 1976 Formative Period occupations in the highlands of northern Ecuador. American Antiquity 41:353-360. 1978 Formative Period Occupations in the highlands of northern Ecuador: rejoinder to Athens. American Antiquity 43:497-500.

Miller, E. V. 1959 Agricultural Ecuador. Geographical Review 49:183-207.

Miller, Naomi 1988 Ratios in Paleoethnobotanical Analysis. In Current Paleoethnobotany, edited by Christine A. Hastorf and Virginia S. Popper, pp. 72-85. University of Chicago Press, Chicago.

Minnis, Paul E. 1989 Prehistoric Diet in the Northern Southwest: Macroplant Remains from Four Corners Feces. American Antiquity 54:543-563.

Mittermeier, R. A., P. Robles Gil, M. Hoffman, J. Pilgrim, T. Brooks, C. J. Mittermeier, J. Lamoreux, and G. A. B. da Fonseca 2004 Hotspots Revisited. CEMEX, Mexico City.

433

Moorthy, S. N. 2004 Tropical sources of starch. Starch in Food: Structure, function and applications, edited by A-C. Eliasson, pp. 321-359. Woodhead and CRC Press, Boca Raton.

Mörner, Magnus 1985 The Andean Past: Land, Society and Conflicts. Columbia University Press, New York.

Moss, G. E. 1976 The Microscopy of Starch. In Examination and Analysis of Starch and Starch Products, edited by J. A. Radley, pp.1-32. Applied Science, Essex.

Motamayor, Juan C., Phillipe Lachenaud, Jay Wallace da Silva e Mota, Ray Loor, David N. Kuhn, J. Steven Brown, and Raymond J. Schnell 2008 Geographic and Genetic Population Differentiation off the Amazonian Chocolate Tree (Theobroma cacao L). PLoS ONE 3(10):1-8.

Motamayor, J. C., A. M. Risterucci, P. A. Lopez, C. F. Ortiz, A. Moreno and C. Lanaud 2002 Cacao domestication I: the origin of cacao cultivated by the Mayas. Heredity 89:380-386.

Motamayor, Juan C., A. M. Risterucci, M. Heath, and C. Lanaud 2003 Cacao domestication II: progenitor germplasm of the Trinitario cacao cultivar. Heredity 91:322-330.

Moy, Christopher M., Geoffrey O. Seltzer, Donald T. Rodbell, and David M. Anderson 2002 Variability of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature 420:162-165.

434

Murra, John V. 1968 An Aymara kingdom in 1567. Ethnohistory 15(2):115-151. 1972 El control vertical de un maximo de pisos ecologicos en la economia de las sociedades andinas. In Visita de la Provincia de Leon de Huánuco II, edited by J. Murra, pp. 428-476. Universidad Hermilio Valdizan, Huánuco.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. de Fonseca, and J. Kent 2000 Biodiversity hotspots for conservation priorities. Nature 403 :853-398.

Myers, Thomas P. 1992 Agricultural Limitations on the Amazon in Theory and Practice. World Archaeology 24 :82-97. 2004 Dark earth in the upper Amazon. In Explorations in Amazonian Dark Earths, edited by B. Glaser and W. Woods, pp. 67-94. Springer, New York.

National Research Council 1989 Lost Crops of the Incas: Little-Known Plants of the Andes with Promise for Worldwide Cultivation. National Academy Press, Washington, D.C.

Neves, Eduardo G., and James B. Petersen 2006 Political economy and pre-columbian landscape transformation in Central Amazonia. In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands, edited by W. Balée and C.L. Erickson, pp. 279-310. Columbia University Press, New York.

Newsom, S.W.B. and Peggy Shaw 1997a Airborne Particles from Latex Gloves in the Hospital Environment. European Journal of Surgery 597:31-33. 1997b A survey of starch particle counts in the hospital environment in relation to the use of powdered latex gloves. Occupational Medicine 47(3):155-158.

435

Ng, N. Q., and S. Y. C. Ng 2002 Genetic Resources and Conservation. In Cassava: Biology, Production, and Utilization, edited by R.J. Hillocks, J.M. Thresh, and A.C. Bellotti, pp. 167-177, CABI, Wallingford and New York.

Niemann, Holger, and Hermann Behling 2008 Late Quaternary vegetation, climate, and fire dynamics inferred from the El Tiro record in the southeast Ecuadorian Andes. Journal of Quaternary Science 23(3):203-212. 2009 Late Pleistocene and Holocene environmental change inferred from the Cocha Caranga sediment and soil records in the southeast Ecuadorian Andes. Palaeogeography, Palaeoclimatology, Palaeoecology 276:1-14.

Niemann, Holger, Torsten Haberzettl, and Hermann Behling 2009 Holocene climate variability and vegetation dynamics inferred from the (11700 cal. yr BP) Laguna Rabadilla de vaca sediment record, southeastern Ecuadorian Andes. The Holocene 19(2):307-316.

Oliver, José O. 2001 The Archaeology of Forest Foraging and Agricultural Production in Amazonia. In Unknown Amazon: Culture in Nature in Ancient Brazil, edited by C. McEwan, C. Barreto, and E. Neves, pp. 50-85. British Museum Press, London. 2008 The Archaeology of Agriculture in Ancient Amazonia. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 185-216. Springer, New York.

Olsen, Kenneth M. 2002 Population history of Manihot esculenta (Euphorbiaceae) inferred from nuclear DNA sequences. Molecular Ecology 11:901-911.

436

Olsen, Kenneth M., and Barbara A. Schaal 1999 Evidence of the Origin of Cassava: Phylogeography of Manihot esculenta. Proceedings of the National Academy of Sciences of the United States of America 96:5586-5591. 2001 Microsatellite Variation in Cassava (Manihot esculenta, Euphorbiaceae) and its Wild Relatives: Further Evidence for a Southern Amazonian Origin of Domestication. American Journal of Botany 88:131-142. 2006 DNA Sequence Data and Inferences on Cassava’s Origin of Domestication. In Documenting Domestication: New Genetic and Archaeological Paradigms, edited by M. Zeder, E. Emshwiller, D. Bradley, and B.D. Smith, pp. 32-45. University of California Press, Berkeley and Los Angeles.

Oyuela-Caycedo, Augusto, Peter W. Stahl, and J. Scott Raymond 2010 Cerro Narrío y Max Uhle: El arqueólogo como agent del desarollo de la arqueología Ecuatoriana. In Max Uhle (1856-1944) Evaluaciones de Sus Investigaciones y Obras, edited by Peter Kaulicke, Manuela Fischer, Peter Masson and Gregor Wolff, pp. 359-377. Fondo Editorial de la Pontificia Universidad Católica del Perú, Lima.

Parr, J. F. 2002 The Identification of Xanthorrhoea Resins by Starch Morphology: Prospect for Archaeological and Taxonomic Applications. Economic Botany 53(3):260-270.

Parsons, James J. 1982 The Northern Andean Environment. Mountain Research and Development 2:253- 262.

Pearsall, Deborah M. 1977 Early movement of maize between Mesoamerica and South America. Journal of the Steward Anthropological Society 9:41-75.

437

1988 Interpreting the Meaning of Macroremain Abundance: The Impact of Source and Context. In Current Paleoethnobotany, edited by Christine A. Hastorf and Virginia S. Popper, pp. 97-118. University of Chicago Press, Chicago. 1989 Adaptation of prehistoric hunter-gatherers to the high Andes: the changing role of plant resources. In Foraging and Farming: The Evolution of Plant Exploitation, edited by D.R. Harris and G. Hillman, pp. 318-332. Unwin Hyman, London. 1992 The Origins of Plant Cultivation in South America. In The Origins of Agriculture: an International Perspective, edited by C. W. Cowan and P. J. Watson, pp. 173-206. Smithsonian Institution Press, Washington, DC. 1995 ‘Doing’ paleoethnobotany in the tropical lowlands: adaptation and innovation in methodology. In Archaeology in the Lowland American Tropics. Edited by P.W. Stahl, pp. 113-129. Cambridge University Press, Cambridge. 2000 Paleoethnobotany: a Handbook of Procedures. Second ed. Academic Press, San Diego. 2003a Plant Food Resources of the Ecuadorian Formative: An Overview and Comparison to the Central Andes. In Archaeology of Formative Ecuador, edited by J. S. Raymond and R. L. Burger, pp. 213-258. Dumbarton Oaks, Washington, D.C. 2003b Integrating biological data: Phytoliths and starch grains, health and diet, at Real Alto, Ecuador. In Phytolith and Starch Research in the Australian-Pacific Regions: the state of the art, edited by D.M. Hart and L.A. Wallis, pp.187-200. Pandanus, Canberra. 2004 Plants and People in Ancient Ecuador: the Ethnobotany of the Jama River Valley. Wadsworth, Belmont, CA. 2008 Plant Domestication and the Shift to Agriculture in the Andes. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 105-120. Springer, New York.

Pearsall, Deborah M., Karol Chandler-Ezell and James A. Zeidler 2004 Maize in ancient Ecuador: results of residue analysis of stone tools from the Real Alto site. Journal of Archaeological Science 31:423-442.

438

Pearsall, Deborah M. and Dolores R. Piperno 1990 Antiquity of Maize Cultivation in Ecuador: Summary and Reevaluation of the Evidence. American Antiquity 55:324-337.

Pearsall, Deborah M., and James A. Zeidler 1994 Regional environment, cultural chronology, and prehistoric subsistence in northern Manabí. In Regional Archaeology in Northern Manabí, Ecuador, Volume 1: Environment, Cultural Chronology, and Prehistoric Subsistence in the Jama River Valley, edited by J.A. Zeidler and D.M. Pearsall, pp. 200-215. University of Pittsburgh Memoirs in Latin American Archaeology, No. 8. University of Pittsburgh, Pittsburgh.

Percy, R. G., and J. F. Wendel 1990 Allozyme evidence for the origin and diversification of Gossypium barbadense L. Theoretical and Applied Genetics 79:529-542.

Pérez, Serge, Paul M. Baldwin, and Daniel J. Gallant 2009 Structural Features of Starch Granules I. In Starch: Chemistry and Technology, Third edition. Edited by J. BeMiller and R. Whistler, pp. 147-192. Academic Press, Burlington, MA.

Perry, Linda 2001 Prehispanic Subsistence in the Middle Orinoco Basin: Starch Analyses Yield New Evidence. Unpublished Ph.D. dissertation, Southern Illinois University at Carbondale, Carbondale. 2002 Starch Granule Size and the Domestication of Manioc (Manihot esculenta) and Sweet Potato (Ipomoea batatas). Economic Botany 56(4):335-349.

Perry, L., R. Dickau, S. Zarrillo, I. Holst, D. M. Pearsall, D. R. Piperno, M. J. Berman, R. G. Cooke, K. Rademaker, A. J. Ranere, J. S. Raymond, D. H. Sandweiss, F. Scaramelli, K. Tarble, and J. A. Zeidler

439

2007 Starch Fossils and the Domestication and Dispersal of Chili Peppers (Capsicum spp. L.) in the Americas. Science 315:986-988.

Perry-Castañeda Library Map Collection 1973 Perry- Castañeda Library Map Collection, Courtesy of the University of Texas Libraries, The University of Texas at Austin, Map No. 501177, http://www.lib.utexas.edu/maps/ecuador.html, accessed February 27, 2011.

Pickersgill, Barbara 1969 The archaeological record of chili peppers (Capsicum spp.) and the sequence of plant domestication in Peru. American Antiquity 34:54-61. 1971 Relationship between weedy and cultivated forms in some species of chili pepper (genus Capsicum). Evolution 25:683-691. 1988 The genus Capsicum: A multidisciplinary approach to the taxonomy of cultivated and wild plants. Biologische Zentrablatt 107:381-389. 2007 Domestication of plants in the Americas: insights from Mendelian and molecular genetics. Annals of Botany 100:925-940.

Pickersgill, B., and C. B. Heiser 1977 Origins and distribution of plants domesticated in the New World tropics. In Origins of Agriculture, edited by C.A. Reed, pp. 803-835. Mouton, The Hague.

Piperno, Dolores 1988 Phytolith Analysis: An archaeological and geological perspective. Academic Press, San Diego. 1998 Paleoethnobotany from microfossils in the Neotropics: new insights into ancient plant use and agricultural origins. Journal of World Prehistory 12:393-449. 2006a Phytoliths: A Comprehensive Guide for Archaeologists and Paleoecologists. Altamira, Lanham, MD.

440

2006b Identifying Manioc (Manihot esculenta Crantz) and Other Root Crops in Pre- Columbian Tropical America through Starch Grain Analysis. In Documenting Domestication: New Genetic and Archaeological Paradigms, edited by M. Zeder, pp. 46- 67. University of California Press, Berkeley and Los Angeles.

Piperno, Dolores R., Thomas C, Andres, and Karen E. Stothert 2000 Phytoliths in Cucurbita and other Neotropical Cucurbitaceae and their occurrence in early archaeological sites from the Lowland American Tropics. Journal of Archaeological Science 27:193-208.

Piperno, Dolores R., and Tom D. Dillehay 2008 Starch grains on human teeth reveal early broad crop diet in northern Peru. Proceedings of the National Academy of Sciences of the United States of America 105:19622-19627.

Piperno, D. R., and K. V. Flannery 2001 The earliest archaeological maize (Zea mays L.) from highland Mexico: New accelerator mass spectrometry dates and their implications. Proceedings of the National Academy of Sciences of the United States of America 98:2101-2103.

Piperno, Dolores R. and Irene Holst 1998 The Presence of Starch Grains on Prehistoric Stone Tools from the Humid Neotropics: Indications of Early Tuber Use and Agriculture in Panama. Journal of Archaeological Science 25:765-776.

Piperno, Dolores R. and Deborah M. Pearsall 1998 The Origins of Agriculture in the Lowland Neotropics. Academic Press, San Diego.

441

Piperno, D. R., A. J. Ranere, I. Holst and P. Hansell 2000 Starch Grains reveal early root crop horticulture in the Panamanian tropical forest. Nature 407:894-897.

Piperno, Dolores R., and Karen E. Stothert 2003 Phytolith Evidence for Early Holocene Cucurbita Domestication in Southwest Ecuador. Science 299:1054-1057.

Pohl, Mary. E. D., Dolores R. Piperno, Kevin O. Pope, and John G. Jones 2007 Microfossil evidence for pre-Columbian maize dispersals in the Neotropics from San Andrés, Tabasco, Mexico. Proceedings of the National Academy of Sciences of the United States of America 104:6870-6875.

Popper, Virginia S. 1988 Selecting Quantitative Measurements in Paleoethnobotany. In Current Paleoethnobotany, edited by Christine A. Hastorf and Virginia S. Popper, pp. 53-71. University of Chicago Press, Chicago.

Powis, Terry G., W. Jeffrey Hurst, Maria del Carmen Rodriguez, Ponciano Ortiz C., Michael Blake, David Cheetham, Michael D. Coe, and John G. Hodgson 2007 Oldest chocolate in the New World. Antiquity 81 (314): project galley online http://www.antiquity.ac.uk/projgall/powis/index.html.

Powis, Terry G., Ann Cyphers, Nilesh W. Gaikwad, Louis Grivetti, and Kong Cheong 2011 Cacao use and the San Lorenzo Olmec. Proceedings of the National Academy of Sciences of the United States of America108:8595-8600.

442

Pozorski, Sheila, and Thomas Pozorski 2008 Early Cultural Complexity on the Coast of Peru. In Handbook of South American Archaeology, edited by Helaine Silverman and William H. Isbell, pp. 607-631. Springer, New York.

Preiss, Jack 2009 Biochemistry and Molecular Biology of Starch Synthesis. In Starch: Chemistry and Technology, Third edition. Edited by J. BeMiller and R. Whistler, pp. 83-148. Academic Press, Burlington, MA.

Purseglove, J. W. 1972 Tropical Crops. Monocotyledons. Longham, London.

Quon, Sonia 2003 Appendix G: Report of Residue Analysis on Ground Stone Tools from EgPn- 612. In Montreux Partnership Lands HRIA for Portions of NE1/4 9-24-2-W5M, and HRM 1 for EgPn-612, Calgary, Alberta, Final Report, by Charles Ramsay and Allyson Ramsay. Report on file, Archaeological Survey of Alberta, Edmonton.

Radley, J. A. 1953 Starch and its Derivatives. Chapman and Hall, London. 1976 Physical Methods of Characterising Starch. In Examination and Analysis of Starch and Starch Products, edited by J. A. Radley, pp. 91-131. Applied Science Publishers Ltd, London.

Ramsay, Paul M. and E.R.B. Oxley 2001 An Assessment of Aboveground Net Primary Productivity in Andean Grasslands of Central Ecuador. Mountain Research and Development 21:161-167.

443

Raviele, Maria 2011 Experimental assessment of maize phytoliths and starch taphonomy in carbonized cooking residues. Journal of Archaeological Science 38:2708-2713.

Raymond, J. Scott 1993 Ceremonialism in the Early Formative of Ecuador. In El Mundo Ceremonial Andino, edited by L. Milliones and Y. Onuki, pp. 23-43. Senri Ethnological Series, Vol. 37. National Museum of Ethnology, Osaka. 1998 Beginnings of sedentism in the lowlands of Northwestern South America. In Recent Advances in the Archaeology of the Northern Andes, edited by A. Oyuela Caycedo and J. Scott Raymond, Vol. 39, pp. 10-19. The Institute of Archaeology, UCLA, Los Angeles. 2003a Social Formations in the Western Lowlands of Ecuador during the Early Formative. In The Archaeology of Formative Ecuador, edited by J.S. Raymond and R.L. Burger, pp. 33-67. Dumbarton Oaks, Washington, D.C. 2003b Appendix D: Formative Period Chronology for the Southern Highlands of Ecuador. In The Archaeology of Formative Ecuador, edited by J.S. Raymond and R.L. Burger, pp. 547-551. Dumbarton Oaks, Washington, D.C. 2008 The Process of Sedentism in Northwestern South America. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 79-90. Springer, New York.

Raymond, J. Scott, and Richard L. Burger 2003 The Archaeology of Formative Ecuador. Dumbarton Oaks, Washington, D.C.

Raymond, J. Scott and Florencio Delgado 2008-2009 Rescuing Cerro Narrío, Cañar Province, Ecuador: Report on Investigations at Cerro Narrío, Ecuador July 1 to August 15, 2008. Project Grant Report to Dumbarton Oaks, Washington, D.C.

444

Raymond, J. Scott, Jorge G. Marcos, Donald Lathrap 1980 Evidence of Early Formative Settlement in the Guayas Basin, Ecuador. Current Anthropology 21:700-701.

Reber, E.A, S.N. Dudd, N. J. van der Merwe, and R. P. Evershed 2003 Direct detection of maize in pottery residues via compound specific carbon isotope analysis. Antiquity 78:682-691.

Reber, E. A., and R. P. Evershed 2004 How did Mississippians prepare maize? The application of compound-specific carbon isotope analysis to absorbed pottery residues from several Mississippi Valley sites. Archaeometry 46:19-33.

Redman, Charles 1999 Human Impact on Ancient Environments. University of Arizona Press, Tucson.

Rehren, Thilo, and Mathilde Temme 1994 Pre-Colombian gold processing at Putishío, South Ecuador: the archaeomettalurgical evidence. In Archaeometry of Pre-Columbian Sites and Artifact: proceedings of a symposium organized by the UCLA Institute of Archaeology and the Getty Conservation Institute, Los Angeles, California, March 23-27, 1992, edited by D.A. Scott and P. Meyers, pp. 267-284. The Getty Conservation Institute, Los Angeles.

Reichert, Edward Tyson 1913 The Differentiation and Specificity of Starches in Relation to Genera, Species, etc. 2 vols. Carnegie Institute of Washington, Washington, D.C.

Reimer, P. J., M. G. L. Baillie, E. Bard, A. Bayliss, J. W. Beck, P. G. Blackwell, C. Bronk Ramsey, C. E. Buck, G. S. Burr, R. L. Edwards, M. Friedrich, P. M. Grootes, T. P. Guilderson, I. Hajdas, T. J. Heaton, A. G. Hogg, K. A. Hughen, K. F. Kaiser, B. Kromer,

445

F. G. McCormac, S. W. Manning, R. W. Reimer, D. A. Richards, J. R. Southon, S. Talamo, C. S. M. Turney, J. van der Plicht, and C. E. Weyhenmeyer 2009 IntCal09 and Marine09 radiocarbon age calibration curves, 0-50,000 years cal BP. Radiocarbon 51(4):1111-1150.

Rindos, David 1984 The Origins of Agriculture: An Evolutionary Perspective. Academic Press, Orlando.

Rios, Montserrat, Rodrigo de la Cruz, and Arturo Mora 2008 Conocimiento Traditional y Plantas Útiles del Ecuador: Saberes y Prácticas. Abya-Yala, Quito.

Rios, Montserrat, Michael J. Kozial, Henrik Borgtoft Pedersen, and Gabriela Granda 2007 Useful Plants of Ecuador: Applications, Challenges, and Perspectives. Abya- Yala, Quito.

Roa, Ana Carolina, Maria M. Maya, Myriam Cristina Duque, Joseph Tohme, Antonio C. Allem, and Meridith W. Bonierbale 1997 AFLP Analysis and Relationships among Cassava and other Manihot Species. Theoretical and Applied Genetics 95:741-750.

Rodbell, Donald T., Geoffrey O. Seltzer, David M. Anderson, Mark B. Abbott, David B. Enfield, and Jeremy H. Newman 1999 An ~15,000-Year Record of El Niño-Driven Alluviation in Southwestern Ecuador. Science 283:516-520.

Rogez, Hervé, Raphaële Buxant, Eric Mignolet, Jesus N. S. Souza, Evaldo M. Silva, and Yvan Larondelle

446

2004 Chemical composition of the pulp of three typical Amazonian fruits, araça-boi (Eugenia stipitata), bacuri (Platonia insignis) and capuaçu (Theobroma grandiflorum). European Food Research and Technology 218:380-384.

Rollenbeck, R., P. Fabian and J. Bendix 2006 Precipitation Dynamics and Chemical Properties in Tropical Mountain Forests of Ecuador. Advances in Geosciences 6:73-76.

Rollins, Harold B., James B. Richardson III, and Daniel H. Sandweiss 1986 The birth of El Niño: Geoarchaeological evidence and implications. Geoarchaeology 1(1):3-15.

Roosevelt, Anna C. 1999 The development of prehistoric complex societies: Amazonia, a tropical forest. In Complex Polities in the Ancient Tropical World, edited by E.A. Bacus and L.J. Lucero, pp. 13-34, Archaeological Papers of the American Anthropological Association, Number 9, Arlington, VA.

Ross, Eric 1978 Food taboos, diet, and hunting strategy: The adaptation to animals in Amazonian cultural ecology. Current Anthropology 19:1-36.

Rossen, Jack, Tom D. Dillehay, and Donald Ugent 1996 Ancient Cultigens or Modern Intrusions?: Evaluating Plant Remains in an Andean Case Study. Journal of Archaeological Science 23:391-407.

Rostoker, Arthur 2003 Appendix C: Formative Period Chronology for Eastern Ecuador. In The Archaeology of Formative Ecuador, edited by J.S. Raymond and R.L. Burger, pp. 539- 545. Dumbarton Oaks, Washington, D.C.

447

Roullier, C., G. Rossel, D. Tay, D. McKey, and V. Lebot 2011 Combining chloroplast and nuclear microsatellites to investigate origin and dispersal of New World sweet potato landraces. Molecular Ecology 20:3963-3977.

Rovner, Irwin 1983 Plant Opal Phytolith Analysis: major advances in archaeobotanical research. In Advances in Archaeological Method and Theory. Vol. 6, edited by Michael B. Schiffer, pp. 225-226. Academic Press, New York.

Ruas, P. M., A. Bonifacio, C. F. Ruas, D. J. Fairbanks, and W. R. Andersen 1999 Genetic relationship among 19 accessions of six species on Chenopodium L., by random amplified polymorphic DNA fragments (RAPD). Euphytica 105:25-32.

Rucker, Robert 2009 Nutritional Properties of Cocoa. In Chocolate: History, Culture, and Heritage, edited by L.E. Grivetti and H-Y Shapiro, pp. 943-946. John Wiley & Sons, Hoboken.

Sahlins, Marshall 1972 Stone Age Economics. Aldine, Chicago.

Salazar, Ernesto 2008 Pre-Columbian Mound Complexes in the Upano River Valley, Lowland Ecuador. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 263-278. Springer, New York.

Samuel, Delwen 1996 Investigation of ancient Egyptian baking and brewing methods by correlative microscopy. Science 273:488-498.

448

Sanders, Donald H. 1990 Statistics: A Fresh Approach. Fourth edition. McGraw-Hill, New York, NY.

Sandweiss, Daniel H., and James B. Richardson III 2008 Early Andean Environments. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 93-104. Springer, New York.

Sandweiss, Daniel H., Kirk A. Maasch, Richard L. Burger, James B. Richardson III, Harold Rollins, and Amy Clement 2001 Variations in Holocene El Niño frequencies: Climate records and cultural consequences in ancient Peru. Geology 29:603-606.

Sandweiss, Daniel H., James B. Richardson III, Elizabeth J. Reitz, Harold Rollins, and Kirk Maasch 1996 Geoarchaeological Evidence from Peru for a 5000 Years B.P. Onset of El Niño. Science 273:1531-1533.

Sarmiento, G. 1986 Ecologically crucial features of climate in high tropical mountains. In High Altitude Tropical Biogeography, edited by F. Vuilleumier and F. Monasterio, pp. 11-45. Oxford University Press, Oxford.

Sauer, Carl O. 1950 Cultivated Plants of South and Central America. In Handbook of South American Indians, Volume 6, edited by J.H. Steward, pp. 487-543. Smithsonian Institution, Washington, D.C. 1963 Geography Of South America, Part 4. Geography and Plant and Animal Resources. In Handbook of South American Indians., edited by J. H. Steward, pp. 319- 344. vol. 6, Physical Anthropology, Linguistics and Cultural Geography of South American Indians. Cooper Square, New York.

449

Sauer, Jonathan D. 1993 Historical Geography of Crop Plants: A Select Roster. CRC Press, Boca Raton.

Schmieder, R.L., and P.G. Keeney 1980 Characterization and Quantification of Starch in Cocoa Beans and Chocolate Products. Journal of Food Science 45:555-563.

Schultes, R. E., and A. Hoffman 1992 Plants of the Gods: their sacred, healing and hallucinogenic powers. Healing Arts Press, Rochester, VT.

Schwarcz, H.P., J. Melbye, M.A. Katzenberg and M. Knyf 1985 Stable isotopes in human skeletons of southern Ontario: reconstructing paleodiet. Journal of Archaeological Sciences 12:187-206.

Schwartz, Deborah and Roy L. Whistler 2009 History and Future of Starch. In Starch: Chemistry and Technology, Third edition. Edited by J. BeMiller and R. Whistler, pp. 1-10. Academic Press, Burlington, MA.

Seidemann, J. 1966 Stärke-Atlas: Grundlagen der Stärke-Mikroscopie und Beschreibung der Wichtigsten Stärkearten. Paul Parey, Berlin.

Shannon, Jack C., Douglas L. Garwood, and Charles D. Boyer 2009 Genetics and Physiology of Starch Development. In Starch: Chemistry and Technology, Third edition. Edited by J. BeMiller and R. Whistler, pp. 23-82. Academic Press, Burlington, MA.

450

Smalley, John, and Michael Blake 2003 Sweet Beginnings: Stalk Sugar and the Domestication of Maize. Current Anthropology 44:675-703.

Smith, Bruce D. 1998 The Emergence of Agriculture. Paperback ed. Scientific American Library, W.H. Freeman, New York. 2006 Eastern North America as an independent center of plant domestication. Proceedings of the National Academy of Sciences of the United States of America 103:12223-12228.

Smith, E. A. and B. Winterhalder 1992 Natural selection and decision making: some fundamental principles. In Evolutionary Ecology and Human Behavior, edited by E. A. Smith and B. Winterhalder, pp. 25-60. Aldine, New York.

Snyder, Rodney, Bradley Foliart Olsen, and Laura Pallas Brindle 2009 From Stone Metates to Steel Mills: the evolution of chocolate manufacturing. In Chocolate: History, Culture and Heritage. Edited by L.E. Grivetti and H-Y Shapiro, pp. 661-623. John Wiley and Sons, Hoboken, New York.

Souris, Marc 2011a Digital Landscape Map of Ecuador, Institut de recherche pour le développement (IRD). Electronic document, http://www.rsgis.ait.ac.th/~souris/Ecuador%20Landscapes.htm, accessed February 28, 2011. 2011b Continental Ecuador 30m Resolution per Pixel DEM Maps. Electronic files, http://www.rsgis.ait.ac.th/~souris/ecuador.htm, accessed March 22, 2011.

451

Spooner, David, Karen McLean, Gavin Ramsay, Robbie Waugh, and Glenn J. Byan 2005 A single domestication for potato based on multilocus amplified fragment length polymorphism genotyping. Proceedings of the National Academy of Sciences of the United States of America 102:14694-14699.

Stahl, Peter W. 2005 Selective faunal provisioning in the southern highlands of Formative Ecuador. Latin American Antiquity 16:313-328. Stahl, Peter W. 2009 Zooarchaeology. In Art and Archaeology of Challuabamba, Ecuador. Edited by Terence Grieder, pp. 165-206. University of Texas Press, Austin.

Stahl, Peter W., and J. Stephen Athens 2001 A High Elevation Zooarchaeological Assemblage from the Northern Andes of Ecuador. Journal of Field Archaeology 28:161-176.

Staller John E. 2001 Reassessing the developmental and chronological relationships of the Formative of coastal Ecuador. Journal of World Prehistory 15(2):193-256.

Staller, J. E. and R. G. Thompson 2002 A Multidisciplinary Approach to Understanding the Initial Introduction of Maize into Coastal Ecuador. Journal of Archaeological Science 29:33-50.

Steward, Julian H. 1949 South American cultures: an interpretive summary. In Handbook of South American Indians, Bulletin 143, edited by J.H. Steward, pp. 669-772. Bureau of American Ethnology, Washington, D.C.

452

Steward, Julian H., and Louis C. Faron 1959 Native Peoples of South America, McGraw-Hill, New York.

Stothert, Karen E. 1983 Review of the early preceramic complexes of the Santa Elena Peninsula, Ecuador. American Antiquity 48:122-127. 1985 The Preceramic Las Vegas Culture of Coastal Ecuador. American Antiquity 50:613-637.

Stothert, Karen E., Dolores R. Piperno, and Thomas C. Andres 2003 Terminal Pleistocene/Early Holocene human adaptation in coastal Ecuador: the Las Vegas Evidence. Quaternary International 109-110:23-43.

Sturtevant, William C. 2009 History and ethnography of some West Indian starches. In The Domestication and Exploitation of Plants and Animals, second printing, edited by P.J. Ucko and G.W. Dimbleby, pp. 177-200, Transaction Publishers, New Brunswick, NJ.

Tapia, M., and C. Vargas 1982 Wild lupines of the Andes of southern Peru. In Proceedings of the First International Lupin Conference, pp. 159-164. Lima, Peru.

Temme, Mathilde 1982 Excavaciones en el sitio precerámico de Cubilán, Ecuador. Miscelanea Antropológica Ecuatoriana 2:135-164.

Tester, Richard F. and John Karkalas 2001 The Effects of Environmental Conditions on the Structural Features and Physico- chemical Properties of Starches. Starch/Stärke 53:513-519.

453

Thomas, David J. and William A. Atwell 1997 Starches. Eagan Press, St. Paul, MN.

Torrence, Robin 2006 Starch and Archaeology. In Ancient Starch Research. Edited by R. Torrence and H. Barton, pp. 17-34. Left Coast Press, Walnut Creek, CA.

Torrence, R., R. Wright, and R. Conway 2004 Identification of starch granules using image analysis and multivariate techniques. Journal of Archaeological Science 31:519-532.

Torres, Fernando G., Omar P. Troncoso, Diego A. Díaz, and Evelin Amaya 2011 Morphological and thermal characterization of native starches from Andean crops. Starch/Stärke 63:381-389.

Turcotte, E. L., and R. G. Percy 1990 Genetics of kidney seed in Gossypium barbadense L. Crop Science 30:384-386.

Tykot, Robert H., and John E. Staller 2002 The Importance of Early Maize Agriculture in Coastal Ecuador: New Data from La Emerenciana. Current Anthropology 43:666-677.

Ubelaker, Douglas H. 1980 Human skeletal remains from site OGSE-80, a preceramic site on the Santa Elena Peninsula, Coastal Ecuador. Journal of the Washington Academy of Sciences 70(1):3-24.

Ubelaker, D. H., M. A. Katzenberg, and L. G. Doyen 1995 Status and Diet in Precontact Highland Ecuador. American Journal of Physical Anthropology 97:403-411.

454

Ugent, Donald 1994 Chemosystematics in Archaeology: A preliminary study of the use of chromatography and spectrophotometry in the identification of four prehistoric root crop species from the desert coast of Peru. In Tropical Archaeobotany: Applications and new developments, edited by J.G. Hather, pp. 215-216. Routledge, London. 1997 The tuberous plant remains of Monte Verde. In Monte Verde: A Late Pleistocene Settlement in Chile: The Archaeological Context and Interpretations, vol. 2, edited by T.D. Dillehay, pp. 903-910. Smithsonian Institution Press, Washington, DC.

Ugent, Donald, Tom Dillehay, and Carlos Ramirez 1987 Potato remains from a Late-Pleistocene Settlement in South-central Chile. Economic Botany 41:401-415.

Ugent, D., Sheila Pozorski and Thomas Pozorski 1981 Prehistoric remains of the sweet potato from the Casma Valley of Peru. Phytologia 49:401-415. 1982 Archaeological potato tuber remains from the Casma Valley of Peru. Economic Botany 36:182-192. 1984 New Evidence for Ancient Cultivation of Canna edulis in Peru. Economic Botany 38(4):417-432. 1986 Archaeological manioc (Manihot) from coastal Peru. Economic Botany 40:78- 102.

Uhle, Max 1922 Influencias Mayas en el Alto Ecuador. Boletín de la Academia Nacional de Historia 4:205–240.

USDA 2011 Composition of Foods: Raw, Processed, Prepared. USDA National Nutrient Database for Standard Reference, Release 24. U.S. Department of Agriculture,

455

Agricultural Research Service, Beltsville Human Nutrition Research Center, Nutrient Data Laboratory, Beltsville, MD.

Vail, Gabrielle 2009 Cacao use in Yucatán among the Pre-Hispanic Maya. In Chocolate: History, Culture, and Heritage, edited by L.E. Gravetti and H-Y Shapiro, pp. 3-15. John Wiley & Sons, Hoboken.

Valdez, Francisco 2008 Inter-zonal Relationships in Ecuador. In Handbook of South American Archaeology, edited by Helaine Silverman and William H. Isbell, pp. 865-888. Springer, New York.

Valdez, Francisco, Jean Guffroy, Geoffrey de Saulieu, Julio Hurtado, and Alexandra Yépez 2005 Découverte d’un site cérémoniel formatif sur le versant oriental des Andes. Comptes Rendus de l’Academie des Sciences de l’Institut de France. Paleovol 4:369-374.

Valencia, Renato, Robin B. Foster, Gorky Villa, Richard Condit, Jens-Christian Svenning, Consuelo Henández, Katya Romoleroux, Elizabeth Losos, Else Magårds, and Henrik Balsev 2004 Tree species distribution and local habitat variation in the Amazon: large forest plot in eastern Ecuador. Journal of Ecology 92:214-229.

Van den Eynden, Veerle 2004a Regional and ecological variations of wild edible plants in southern Ecuador. Lyonia 7(2):123-132. 2004b Use and management of edible non-crop plants in southern Ecuador. Unpublished Ph.D. dissertation, University of Gent, Belgium.

456

van der Merwe, N. J., J. A. Lee-Thorp, and J. S. Raymond 1993 Light, stable isotopes, and the subsistence base of Formative cultures at Valdivia, Ecuador. In Prehistoric Human Bone: Archaeology at the Molecular Level, edited by J.B. Lambert and G. Grupe, pp. 63-97. Springer-Verlag, Berlin. van der Merwe, N.J. and J.C. Vogel 1978 13C content of human collagen as a measure of prehistoric diet in Woodland North America. Nature 276:815-816. van Reeuwijk, L. P. 1998 Guidelines for Quality Management in Soil and Plant Laboratories. FAO Soils Bulletin – 74. FAO and ISRIC, Rome.

Vavilov, N. I. 1951 The phylogentic basis of plant breeding. In The Origin, Variation, Immunity and Breeding of Cultivated Plants, pp. 13-54, translated by K.S. Chester. The Ronals Press Company, New York.

Villalba, Marcelo 1988 Cotocollao: una aldea formativa del valle de Quito. Miscelanea Antropologica Ecuatoriana 2, Museo del Banco Central del Ecuador, Quito.

Vuille, Mathias, Raymond S. Bradley and Frank Keimig 2000 Climate Variability in the Andes of Ecuador and Its Relation to Tropical Pacific and Atlantic Sea Surface Temperature Anomalies. Journal of Climate 13:2520-2535.

Walsh, B. M., and S. B. Hoot 2001 Phylogenetic relationships of Capsicum (Solanacea) using DNA sequences from two non-coding regions: The chloroplast atpb-rbcl spacer region and nuclear waxy introns. International Journal of Plant Science 162:1409-1418.

457

Washburn, Dorothy K., William N. Washburn, and Petia A. Shipkova 2011 The prehistoric drug trade: widespread consumption of cacao in Ancestral Pueblo and Hohokam communities in the American Southwest. Journal of Archaeological Science 38:1634-1640.

Wassenaar, T., P. Gerber, P. H. Verburg, M. Rosales, M. Ibrahim, and H. Steinfeld 2007 Projecting land use change in the Neotropics: the geography of pasture expansion into forest. Global Environmental Change 17:86-104.

Weatherforcaster 2012 Weather and Averages – Cayambe, Pichincha, Ecuador. Weatherforcaster.eu. Electronic Document, http://www.weatherforecaster.eu/weather/cayambe/ecuador/ecxx0022#clima, accessed March 7, 2012.

Weber, Lydia, Norbert U. Haase, and Meinolf G. Lindhauer 2002 Potato Starch – Quality and Analytical Aspects. In Starch and Starch Containing Origins: Structure Properties and New Technologies, edited by V.P. Yuryev, A. Cesário, and W.J. Bergthaller, pp. 189-198. Nova Science, New York.

Weiner, Stephen 2010 Microarchaeology: Beyond the Visible Archaeological Record. Cambridge University Press, Cambridge.

Weng, C., M.B. Bush, J.S. Athens 2002 Holocene climate change and hydrarch succession in lowland Amazonia Ecuador. Review of Palaeobotany and Palynology 120(1-2):73-90.

458

Westengen, Ola T., Zósimo Huamán, and Manfred Heun 2005 Genetic diversity and geographic pattern in early South American cotton domestication. Theoretical and Applied Genetics 110:392-402.

White, P. J., and A. Tziotis 2004 New corn starches. In Starch in Food: Structure, function and applications, edited by A-C. Eliasson, pp. 295-320. Woodhead and CRC Press, Boca Raton.

Wilson, Hugh D. 1990 Quinua and relatives (Chenopodium sect. Chenopodium subsect. Cellulata). Economic Botany 42:478-494.

Wilson, Julie, Karen Hardy, Richard Allen, Les Copeland, Richard Wrangham, and Matthew Collins 2010 Automated classification of starch granules using supervised pattern recognition of morphological properties. Journal of Archaeological Science 37:594-604.

Wilson, Warren 2003 Cassava (Manihot esculenta Crantz), Cyanogenic Potential, and Predation in Northwestern Amazonia: The Tukanoan Perspective. Human Ecology 31:403-416.

Winsborough, Barbara, Izumi Shimada, Lee A. Newsom, John G. Jones, and Rafael A. Segura 2012 Paleoenvironmental catastrophies on the Peruvian coast revealed in lagoon sediment cores from Pachacamac. Journal of Archaeological Science 39:602-614.

Woods, W., and J. M. McCann 1999 The anthropogenic origin and persistence of Amazonian Dark Earths. Yearbook Conference of Latin American Geographers 25:7-23.

459

World Trade Press 1998-2011 Ecuador Maps, World Trade Press. Electronic Document, http://www.worldtraderef.com/wtr_nl/WTR_site/maps_new.asp?cid=43, accessed March 8, 2012.

Wright, Patti J. 2010 Methodological Issues in Paleoethnobotany: a consideration of issues, methods, and cases. In Integrating Zooarchaeology and Paleoethnobotany: a consideration of issues, methods, and cases. Edited by A.M. VanDerwarker and T.M. Peres, pp. 37-64. Springer, New York.

Yang, Xiao Yan, and LePing Jiang 2010 Starch grain analysis reveals ancient diet at Kuahuqiao site, Zhejiang Province. Chinese Science Bulletin 55:1150-1156.

Yang, Xioyan, Jianping Zhang, Linda Perry, Zhikun Ma, Zhiwei Wan, Mingqi Li, Xianmin Dao, and Houyuan Lu 2012 From the modern to the archaeological: starch grains from millets and their wild relatives in China. Journal of Archaeological Science. 39:247-254.

Young, E. 1953 Photomicrographs. In Starch and its Derivatives, vol.2, Third ed., edited by J. A. Radley, pp. 443-445. Chapman and Hall, London.

Young, Kenneth 2011 Introduction to Andean Geographies. In Climate Change and Biodiversity in the Tropical Andes, edited by S.K. Herzog, R. Martinez, P.M. Jørgensen, and H. Tiessen, pp. 128-140. IAI and SCOPE, São José dos Campos, Brazil.

460

Young, Vernon R., and Peter L. Pellet 1994 Plant proteins in relation to human protein and amino acid nutrition. American Journal of Clinical Nutrition 59(Supplement):1203S-1212S.

Zarrillo, Sonia 2004 A Preliminary Analysis of Carbonized Residues for Starch Grains on Ceramics from the Loma Alta Site, Ecuador. Unpublished B.Sc. Honours Thesis, Department of Archaeology, University of Calgary, Calgary.

Zarrillo, Sonia and Ruth Dickau 2008 Methods for controlling modern starch contamination in the laboratory. Unpublished results on file, University of Calgary, Calgary.

Zarrillo, Sonia and Brian Kooyman 2006 Evidence for Berry and Maize Processing on the Canadian Plains from Starch Grain Analysis. American Antiquity 71(3):473-499.

Zarrillo, Sonia, Deborah M. Pearsall, J. Scott Raymond, Mary Ann Tisdale, and Dugane J. Quon 2008 Directly dated starch residues document early formative maize (Zea mays L.) in Tropical Ecuador. The Proceedings of the National Academy of Sciences of the United States of America. 105(13):5006-5011.

Zarrillo, Sonia and Francisco Valdez 2010 Corn and the Origins of Socio-Political Complexity in Highland Ecuador. Paper presented at the 33rd annual meeting of the Society of Ethnobiology, Victoria, BC. 2011 Evidencias del cultivo de maiz y otras plantas en la ceja de selva oriental ecuatoriana. Paper presented at Coloquio International - Arqueología regional en la Amazonia occidental: temáticas, resultados y políticas, Quito.

461

Zeidler, James A. 1994 Archaeological testing in the middle Jama Valley. In Regional Archaeology in Northern Manabí, Ecuador, Volume 1: Environment, Cultural Chronology, and Prehistoric Subsistence in the Jama River Valley, edited by J.A. Zeidler and D.M. Pearsall, pp. 71-98. University of Pittsburgh Memoirs in Latin American Archaeology, No. 8. University of Pittsburgh, Pittsburgh. 1995 Archaeological survey and site discovery in the forested Neotropics. In Archaeology in the Lowland American Tropics: Current Analytical Methods and Recent Applications, edited by P.W. Stahl, pp. 7-41. Cambridge University Press, Cambridge. 2003 Appendix A: Formative Period Chronology for the Coast and Western Lowlands of Ecuador. In Archaeology of Formative Ecuador, edited by J. S. Raymond and R. L. Burger, pp. 487-528. Dumbarton Oaks Research Library and Collection, Washington, D.C. 2008 The Ecuadorian Formative. In Handbook of South American Archaeology, edited by H. Silverman and W.H. Isbell, pp. 459-488. Springer, New York.

Zeidler, James A. and John S. Isaacson 2003 Settlement Process and Historical Contingency in the Western Ecuadorian Formative. In Archaeology of Formative Ecuador, edited by J. S. Raymond and R. L. Burger, pp. 69-123. Dumbarton Oaks, Washington, D.C.

Zeidler, James A., and Deborah M. Pearsall 1994 Regional Archaeology in Northern Manabí, Ecuador, Volume 1: Environment, Cultural Chronology, and Prehistoric Subsistence in the Jama River Valley. University of Pittsburgh Memoirs in Latin American Archaeology, No. 8. University of Pittsburgh, Pittsburgh.

Zevallos, Carlos M., Walton C. Galinat, Donald W. Lathrap, Earl R. Leng, Jorge G. Marcos, Kathleen M. Klumpp 1977 The San Pablo Corn Kernel and Its Friends. Science 196:385-389.

462

Zimmerer, Karl S. 1999 Overlapping Patchwork of Mountain Agriculture in Peru and Bolivia: Toward a Regional-Global Landscape Model. Human Ecology 27:135-165.

463

APPENDIX A: FIVE KILOMETER CATCHMENT AREA TOPOGRAPHIC AND

3-D PROJECTION MAPS FOR LA CHIMBA, TAJAMAR, CERRO NARRÍO,

CHAULLABAMBA, THE CATAMAYO VALLEY, AND SANTA ANA-LA

FLORIDA ARCHAEOLOGICAL SITES

464

465

Three-Dimensional Projection for the La Chimba Site. The triangle indicates the archaeological sites, and points to north. The 5 km catchment area is delimited by the blue circle.

466

467

Three-Dimensional Projection for the Tajamar Site. The triangle indicates the archaeological sites, and points to north. The 5 km catchment area is delimited by the blue circle.

468

469

Three-Dimensional Projection for the Cerro Narrío Site. The triangle indicates the archaeological sites, and points to north. The 5 km catchment area is delimited by the blue circle.

470

471

Three-Dimensional Projection for the Cerro Narrío Site. The triangle indicates the archaeological sites, and points to north. The 5 km catchment area is delimited by the blue circle.

472

473

Three-Dimensional Projection for the Catamayo Valley. The right triangle is the Trapichillo site, and the left arrow is the La Vega site. The triangles point north. The 5 km catchment area is delimited by the blue circle.

474

475

Three-Dimensional Projection for the Santa Ana-La Florida Site. The triangle indicates the archaeological sites, and points to north. The 5 km catchment area is delimited by the blue circle.

476