Port, Polity, and Political Economy: The Archaeology of Trinidad de Nosotros, El Peten, Guatemala

Although the largest Classic Maya political capitals are frequently assumed to have served as key nodes in long-distance trade networks, empirical data supporting this contention are surprisingly limited. This study was designed to assess the role of a Late Classic Maya political capital, Motul de San José, in long-distance trade through the examination of its principal secondary center, Trinidad de Nosotros. Located on the north shore of Guatemala's Lake Petén Itzá, 2.6 km south of Motul, and at the natural gateway to the area, Trinidad was well-positioned to serve as a port and intermediary between the inland capital and long-distance trade routes. Research at Trinidad included survey; household, midden, and stratigraphic excavations; and ceramic and obsidian analyses. These investigations defined Trinidad as a moderate-sized center with an occupation extending from the Middle Preclassic to the Historical period. A complex harbor was identified, as was public plaza space more extensive than required by Trinidad's inhabitants. Midden excavations documented spectacular deposits behind Trinidad's ballcourt, indicating that it was a major locus for feasting and ritual throughout the Late Classic period. Cumulatively, these and other data support the proposition that Trinidad was a port. This study also highlights continuities at Trinidad during the latter half of the Late Classic period when Motul emerged as a political power and its sustaining area underwent demographic expansion. These developments enhanced Trinidad's port role, as indicated by additions to the harbor and greater access to obsidian. Comparison of obsidian assemblages from Trinidad and other nearby sites also indicates that Trinidad

maintained its function as a distribution center for obsidian. Elements of Trinidad's ritual economy also demonstrate continuity; lavish ballgame-related feasting, initiated prior to Motul's local ascendancy, continued under its auspices. The high level of continuity in Trinidad's port role suggests that Motul's emergent elites may have had, at most, only indirect control over long-distance trade in the Motul area. This may have resulted from the new Motul dynasts choosing to co-opt existing economic structures, rather than establish new ones. Alternatively, it may simply indicate that long-distance trade in the Maya lowlands was less centralized than previously thought.

ACKNOWLEDGEMENTS "Research is endlessly seductive; writing is hard work." Barbara Tuchman, Practicing History (1981)

Few archaeological investigations are the fruit of one person's labor alone and this study is no exception. Over the course of two field seasons at Trinidad, many hours of lab work, years of graduate study, and through the long writing process I have accrued a debt of gratitude that I shall find difficult to repay. Research at Trinidad was funded by various institutions and people. The first field season was supported by a grant from the Foundation for the Advancement of Mesoamerican Studies and funds from the Middle American Research Institute (MARI). The Amherst College John Mason Clarke Fellowship made it possible to purchase a field vehicle. The second season was supported by an NSF dissertation improvement grant (no. 0528789) and additional funds from MARI. The University of Arizona NSF-AMS Laboratory provided the radiocarbon analyses as part of their visiting scholar program. I would like to particularly thank Will Andrews for extensive MARI support and Kathe Lawton for seamlessly distributing those funds and administering my NSF account. All investigations at Trinidad were conducted under the auspices of Guatemala's IDAEH and the Departamento de Monuments Prehispánicos y Coloniales. I would like to thank Lic. Salvador López, Lic. Paulino Morales, and Licda. Yvonne Putzeys for their efforts on my behalf. I am especially grateful to Gustavo Amarra for counseling me

ii

through the permit process and to Jeanette Castellanos for serving as the project codirector for the 2005 season. From the Motul de San José Archaeological Project, I benefitted greatly from the counsel of Antonia Foias, who gave me "carte blanche" in selecting my avenue of research in the Motul area and then provided consistent support for these endeavors. Although my resulting objectives often veered overly grand in scale, the independence given me to design and implement my own program of research was much appreciated. Many other senior MSJ Project personnel helped me over the years, including Kitty Emery, Richard Terry, and Jeanette Castellanos. Field and laboratory investigations were also assisted by many people, including Ellen Spensley Moriarty, Erin Thornton, Crorey Lawton, Christina Halperin, Suzanna Yorgey, Melanie Kingsley, Ingrid Seyb, Gerson Martínez Salguero, Jorge Guzmán, Fredy Ramírez, Yovany Hernández Veliz, Andrew Wyatt, Eric Kerns, Ben Haldeman, Anita Sánchez, and many others. Most of the artifact illustrations in this dissertation were drawn by Ingrid Seyb and Luis F. Luin. Fieldwork at Trinidad would not have been possible without the support of numerous individuals in San José, El Petén. In particular, I would like to thank former Alcalde Don Julian Tesucún y Tesucún, who provided considerable assistance each season in gaining project access to Trinidad. Particular thanks are also due to the entire Alonzo Ramos family. Carlos and Luis Alonzo Ramos served at various times as ace trailblazer, master excavator, camp guard, mechanic, "cultural attaché," and, when all else failed, willing cook. We became good friends in the process. I would also like to thank Don Jorge Arturo Zac for the long and patient hours he spent sharing his considerable wisdom concerning all things San José.

iii

This study benefited from productive suggestions, ideas, or other commentaries by many people. Bruce Dahlin's enthusiasm for my interpretations of Trinidad and its location during a chance meeting in 2002 convinced me to start working there the following year. The ceramic analysis, in particular, benefitted from suggestions or assistance from many people, including Donald Forsyth, Dorie Reents-Budet, Leslie Cecil, Richard Hansen, Takeshi Inomata, and many others. The ceramic analysis was conducted in consultation with Antonia Foias, who provided innumerable suggestions as well as contributing materially to the analysis itself. I would also like to particularly thank Jeanette Castellanos, who spent countless hours with me in the IDAEH Ceramoteca, Salón 3, and other projects' labs, comparing materials and sorting out the Trinidad ceramic sequence. Greg Hodgins, Dan Healan, Timothy Pugh, Al Wesolowsky, Marcello Canuto, and many others made useful suggestions to this study. I would also like to thank a long list of people in the Tulane University Department of Anthropology where I benefitted from both wonderful teaching and a remarkable community of anthropologists. I would like to particularly thank Victoria Bricker, Harvey Bricker, Dan Healan, Bob Hill, Judith Maxwell, and John Verano for their engaging teaching and collegiality. I also want to thank my cohort at Tulane (David Hixson, Bryan Just, Crorey Lawton, Tony Ortmann, Suzanna Yorgey, and others) for their friendship and for all I learned during hours of collaborative study or animated discussions over coffee or festive beverages. I would also like to particularly thank Crorey Lawton for co-directing and co-funding the 2005 season at Trinidad and for managing everything while I convalesced from malaria.

iv

I would like to especially thank Will Andrews, my advisor, for many years of support and encouragement. Many times it looked as though I would not complete this dissertation. Only his steadfast advocacy and continued interest in my work kept this dissertation alive. For that I owe him an incalculable debt of gratitude. I also hope that he will at some point in the future relinquish his claims on my right arm, immortal soul, firstborn child, and anything else I may have promised in return for emergency funds from MARI! This study would not have been possible without the love and support of my family. My parents, Kevin and Margery Moriarty, worked hard so that I could always pursue my interests and then supported me through the trials and tribulations that sometimes followed that pursuit. Their continued and unconditional support has provided me with a wonderful definition of parenthood. I can only promise that I will do my very best to follow this example and encourage my children to follow their dreams, wherever they may lead. Finally, it is difficult to conceive of this project without thinking of my wife, Ellen Spensley Moriarty. From our first meeting in Guatemala in 2000 (thanks Antonia!) to the present, our lives and research have been interwoven. Elly directed some of the most technically difficult excavations at Trinidad, joyfully assisted in the agonizing process of cataloging artifacts at the end of each field season, and has cast a helpful eye over virtually every part of this study. Beyond archaeology and academia, Elly's faith in me has been an inspiration, and not a day goes by when I don't think I'm the luckiest boy in the world.

v

TABLE OF CONTENTS

ACKNOWLEDGEMENTS

ii

LIST OF TABLES

xvi

LIST OF FIGURES

xxiii

CHAPTER ONE. INTRODUCTION TO THE ARCHAEOLOGY OF TRINIDAD DE NOSOTROS, EL PETEN, GUATEMALA

1

CHAPTER TWO. THEORETICAL ORIENTATION

12

Introduction

12

Classic Maya Political Economy and Long-Distance Trade

13

Ports: Definitions and Attributes

18

Ancient Maya Ports

27

Summary

32

CHAPTER THREE. NATURAL AND CULTURAL SETTING

34

Introduction

34

Physical Geography, Environment, and Natural Resources in the Lake Petén Itzá Basin

34

Previous Research in the Lake Petén Itzá Basin

40

Culture-Historical Summary for the Lake Petén Itzá Basin

45

The Middle Preclassic Period

46

The Late Preclassic Period

48 vi

The Terminal Preclassic Period

50

The Early Classic Period

51

The Late Classic Period

53

The Terminal Classic Period

55

The Postclassic and Historical Periods

56

Motul de San José: History of Research and Early Conclusions

59

Conclusions

64

CHAPTER FOUR. RESEARCH DESIGN

66

Introduction

66

Previous Research at Trinidad de Nosotros

66

Identifying a Port: Geography and Geopolitics at Trinidad de Nosotros

72

Physical Setting

73

Geographic Position

78

Local and Long-Distance Trade Routes

81

Gateway to the Motul de San José Area

84

Research Design: Site Models and Expectations

85

Site Models: Coastal Sites, Transshipment Ports, and Trading Ports

86

Field and Laboratory Objectives

88

Summary of Investigations, 2003–2008

90

Field Methods

93

Excavation Methods and Recording Procedures

93

Cultural and Natural Contexts at Trinidad

98

vii

Cultural Phases at Trinidad

102

Summary

105

CHAPTER FIVE. SURVEY, MAPPING, AND SITE ORGANIZATION

106

Introduction

106

Survey and Map Preparation

108

Surface Coverage and Survey Methods

108

Total Station Mapping

110

Map Features

113

Informal Survey and Unmapped Areas

127

Final Map Preparation

129

Site Organization and Description

130

Public, Ceremonial, and Elite Architecture

130

Plazas and Public Space

142

Harbor Area Features

149

Residential Groups

155

Population Estimates and Plaza Capacities

167

Estimating Trinidad's Late Classic Sik'u' I–II Population

167

Plaza Capacity Estimates

170

Summary

172

CHAPTER SIX. THE SITE CENTER STRATIGRAPHIC TESTING PROGRAM

174

Introduction

174

Methods and Organization

175

viii

Organization of Excavation Summaries

178

Public Architecture Testing

179

Group A: Unit 1A1

179

Plaza I: Unit 1I1

184

Plaza II: Unit 1B1

186

Group F (Ballcourt): Units 1F1 and 1F2

188

Plaza IV: Unit 1F3

198

Plaza V: Units 1P1 and 1P2

200

Group E: Unit 1E1

202

Residential Group Testing

206

Group C: Units 1C1 and 4A13

206

Group G: Units 1G1–1G3 and 5A7

215

Group K: Unit 1K1

225

Group N: Unit 1N1

227

Group O: Units 1O1, 6A13, and 6D9

229

Group U: Unit 1U1

236

Group Y: Unit 1Y1

238

Summary

240

CHAPTER SEVEN. THE HARBOR AREA TESTING PROGRAM

247

Introduction

247

Methods and Organization

247

Organization of Excavation Summaries

250

Investigation of Platform EE

251

ix

Unit 2A4

251

Units 12A–12A5

256

Unit 2B2

269

Platform EE Summary

272

Investigations in Area A

275

Unit 12A6

275

Unit 2A1

282

Area A Summary

284

Investigation of Structure EE-1

286

Units 2E1–2E5

287

Structure EE-1 Summary

289

Investigation of Platform GG

291

Platform GG East: Unit 2A2

292

Platform GG East: Unit 2A5

293

Platform GG West: Unit 2A3

295

Platform GG West: Unit 12B1

301

Platform GG Summary

304

Other Harbor Area Investigations

306

Platform HH: Unit 2B1

306

Platform JJ: Unit 2C1

307

Summary for Platforms HH and JJ

309

Summary and Conclusions

310

CHAPTER EIGHT. THE SITE CENTER MIDDEN TESTING PROGRAM

x

316

Introduction

316

Methods and Organization

317

Organization of Excavation Summaries

322

Public Architecture Midden Testing

324

Plaza V: Operation 9

325

Group F (Ballcourt): Operation 10

330

Residential Group Midden Testing

358

Group C: Operations 3 and 7

358

Group U: Operation 11

367

Group Y: Operation 13

376

Group G: Operation 5

387

Group H: Operation 19

388

Group K: Operation 18

393

Group Q: Operation 16

400

Group S: Operation 15

404

Group T: Operation 14

408

Group O: Operation 8

411

Group Z: Operation 17

415

Group DD: Operation 20

419

Group KK: Operation 21

420

Summary

424

CHAPTER NINE. THE RESIDENTIAL GROUP ARCHITECTURAL TESTING PROGRAM Introduction

429

429 xi

Methods and Organization

429

Investigations in Group C

432

Units 4A1–4A15

434

Description of Structure C-1

437

Artifacts from Structure C-1

443

Investigations in Group G

444

Units 5A1–5A9

446

Description of Structure G-1

449

Artifacts from Structure G-1

453

Investigations in Group FF

454

Units 2D2–2D20

455

Description of Structure FF-1

457

Artifacts from Structure FF-1

458

Summary

460

CHAPTER TEN. CERAMICS AND CHRONOLOGY

462

Introduction

462

Trinidad de Nosotros Ceramic Analysis

463

AMS Radiocarbon Dates

471

The Trinidad de Nosotros Ceramic Sequence

480

Preclassic Period Ceramics

483

Aj Wo' Pre-Mamom Ceramic Complex

483

Ix Cha' Mamom Ceramic Complex

487

Chukan Chicanel Ceramic Complex

492

xii

Preclassic to Classic Period Transition P'ich 'Ayim Late Chicanel/Floral Park Ceramic Complex Classic Period Ceramics

501 501 505

'Ayim-tun Tzakol Ceramic Complex

505

Sik'u' I–II Tepeu Ceramic Complex

512

Yaljob'ach Tepeu 3 Ceramic Complex

540

Postclassic to Historical Period Ceramics

548

Säk-tunich New Town Ceramic Complex

548

Aj B'oj Contact and Historical Ceramic Complex

559

Summary

561

CHAPTER ELEVEN. OBSIDIAN ANALYSES

564

Introduction

564

The Motul de San José Archaeological Project Obsidian Collection

566

Classification of Obsidian Artifacts

572

Core Ratios in the Sik'u' II–Yaljob'ach Motul de San José Area

580

Core Ratios: Blades/Cores, Cores/Blades and Cores+/Blades Obsidian Conservation

581 586

Cutting-Edge-to-Mass (CE/M) Ratios

587

Average Blade Widths

591

Obsidian Source Procurement

594

XRF Analysis

594

The Obsidian Sample

597

Obsidian Source Procurement in the Motul de San José Area

598

xiii

Summary

603

CHAPTER TWELVE. CONCLUSIONS AND IMPLICATIONS

607

Introduction

607

Long-Term Developments at Trinidad de Nosotros

608

Aj Wo' Early Middle Preclassic

608

Ix Cha' Late Middle Preclassic

609

Chukan Late Preclassic

610

P'ich 'Ayim Terminal Preclassic

615

'Ayim-tun Early Classic

616

Sik'u' I–II Late Classic

617

Yaljob'ach Terminal Classic

623

Säk-tunich Postclassic

624

Aj B'oj Historical

625

Port, Polity, and Political Economy in the Sik'u' II–Yaljob'ach Motul de San José Area

626

Summary and Implications

630

APPENDIX A. EXCAVATION FORMS AND CODES

633

APPENDIX B. BURIALS AND CACHES FROM TRINIDAD DE NOSOTROS WITH NOTES ON HUMAN SKELETAL REMAINS BY ERIN E. KENNEDY-THORNTON

639

APPENDIX C. CERAMIC ANALYSIS CODES SHEETS

662

APPENDIX D. TRINIDAD DE NOSOTROS CERAMIC COMPLEXES, GROUPS, AND PRINCIPAL CONSTITUENT TYPES

677

xiv

APPENDIX E. CERAMIC WARES, GROUPS, TYPES, AND VARIETIES USED IN THE PRESENT STUDY

683

APPENDIX F. SHERD COUNTS FROM TYPE-VARIETY ANALYSIS

701

BIBLIOGRAPHY

712

xv

LIST OF TABLES

Table

Page

2.1

Harbor types defined in the World Port Index

22

4.1

Operation summary for excavations at Trinidad

95

4.2

Culture phases at Trinidad de Nosotros

103

5.1

Descriptions and areas for Trinidad Plazas I–VI

143

5.2.

Summary data for residential groups at Trinidad de Nosotros

164

5.3

Plaza capacity estimates for Trinidad de Nosotros and selected sites

171

5.4

Comparison of plaza capacities and population estimate (750 people) for Late Classic Sik'u' II Trinidad de Nosotros

172

6.1

Location and dimensions for site center test units

176

6.2

Context summary for Unit 1A1

183

6.3

Context summary for Unit 1I1

186

6.4

Context summary for Unit 1B1

188

6.5

Context summary for Unit 1F1

191

6.6

Context summary for Unit 1F2

196

6.7

Context summary for Unit 1F3

200

6.8

Context summary for Unit 1P1

201

6.9

Context summary for Unit 1P2

202

6.10

Context summary for Unit 1E1

205 xvi

6.11

Context summary for Unit 1C1

209

6.12

Context summary for Unit 4A13

214

6.13

Faunal taxa identified in Midden 1

221

6.14

Context summary for Units 1G1–1G3

222

6.15

Context summary for Unit 5A7

224

6.16

Context summary for Unit 1K1

227

6.17

Context summary for Unit 1N1

229

6.18

Context summary for Unit 1O1

232

6.19

Context summary for Unit 6A13

234

6.20

Context summary for Unit 6D9

235

6.21

Context summary for Unit 1U1

238

6.22

Context summary for Unit 1Y1

240

6.23

Summary results of site center stratigraphic testing

241

6.24

Floor sequences in Trinidad's public architectural complexes

244

7.1

Location and dimensions of harbor area excavations

249

7.2

Context summary for Unit 2A4

255

7.3

Context summary for Unit 12A1

263

7.4

Faunal taxa identified in Midden 35

267

7.5

Context summary for Units 12A2–12A5

268

7.6

Context summary for Unit 2B2

272

7.7

Sherd densities for Midden 17

278

7.8

Faunal taxa identified in Midden 17

279

7.9

Excavation volume and artifact summaries for Midden 17

280

xvii

7.10

Context summary for Unit 12A6

281

7.11

Context summary for Unit 2A1

284

7.12

Context summary for Units 2E1–2E5

289

7.13

Context summary for Unit 2A2

293

7.14

Context summary for Unit 2A5

295

7.15

Faunal taxa identified in Cache 4

298

7.16

Faunal taxa identified in Burial 2

300

7.17

Context summary for Unit 2A3

300

7.18

Context summary for Unit 12B1

304

7.19

Context summary for Unit 2B1

307

7.20

Context summary for Unit 2C1

308

7.21

Construction episodes identified in harbor area testing

311

7.22

Summary results of harbor area testing

313

8.1

Summary of midden prospecting investigations

319

8.2

Context summary for Unit 9B1

327

8.3

Context summary for Unit 9B2

328

8.4

Excavation volume and artifact summaries for Midden 9–10

330

8.5

Context summary for Units 10D1 and 10D11

335

8.6

Context summary for Unit 10D2

337

8.7

Sherd counts by ceramic group for Middens 11–13

343

8.8

Identified faunal taxa for Middens 11–13

347

8.9

Context summary for Units 10D3–10D10

354

8.10

Excavation volumes and artifact summaries for Middens 11–13

356

xviii

8.11

Context summary for Units 3C1 and 3C3

363

8.12

Context summary for Units 3C4 and 3C5

364

8.13

Context summary for Unit 3C2

365

8.14

Excavation volumes and artifact summaries for Middens 3, 4, and 6

366

8.15

Context summary for Unit 11E1

370

8.16

Context summary for Units 11E2 and 11E4

372

8.17

Context summary for Unit 11E3

375

8.18

Excavation volumes and artifact summaries for Middens 14–16

376

8.19

Faunal taxa identified in Midden 18

379

8.20

Context summary for Units 13E1 and 13E2

380

8.21

Context summary for Unit 13E4

381

8.22

Faunal taxa identified in Midden 20

383

8.23

Context summary for Units 13E3 and 13E5

386

8.24

Excavation volumes and artifact summaries for Midden 18–20

387

8.25

Context summary for Unit 19E1

391

8.26

Context summary for Unit 19E2

392

8.27

Excavation volumes and artifact summaries for Middens 29–30

393

8.28

Context summary for Unit 18E1

397

8.29

Context summary for Unit 18E2

399

8.30

Excavation volumes and artifact summaries for Middens 27–28

400

8.31

Context summary for Unit 16E1

403 xix

8.32

Excavation volumes and artifact summaries for Middens 21–24

403

8.33

Context summary for Unit 15E1

406

8.34

Context summary for Unit 15E2

407

8.35

Context summary for Unit 14C1

410

8.36

Context summary for Units 8E1 and 8E2

413

8.37

Context summary for Unit 8E3

414

8.38

Excavation volumes and artifact summaries for Midden 7–8

415

8.39

Context summary for Unit 17E1

417

8.40

Context summary for Unit 17E2

418

8.41

Excavation volumes and artifact summaries for Middens 25, 26, 31, and 32

419

8.42

Context summary for Unit 21E1

422

8.43

Context summary for Unit 21E2

424

8.44

Summary data for all Sik'u' I to Yaljob'ach phase middens at Trinidad de Nosotros

425

8.45

Combined midden data for the Sik'u' I to Yaljob'ach phases

427

8.46

Summary results of site center midden testing

428

9.1

Summary of architectural testing investigations

431

10.1

Context, material, and ceramic phase for all Trinidad AMS samples

472

10.2

475

10.3

Trinidad AMS samples: radiocarbon ages and calibrated date ranges Sherd counts by type for four Chukan phase middens

10.4

Sherd counts by type for five P'ich 'Ayim phase deposits

502

10.5

Sherd counts by type for three 'Ayim-tun phase deposits

507

xx

498

10.6

Sherd counts by type for seven Sik'u' I and Sik'u' II phase middens

538

10.7

Sherd counts by type for three Yaljob'ach phase middens near Group U

546

10.8

Sherd counts by type for two Säk-tunich phase deposits in Unit 12A6

558

11.1

Motul de San José Archaeological Project obsidian artifact counts organized by site and phase

567

11.2

All Late and Terminal Classic obsidian from the Motul de San José area sorted by context class

568

11.3

Sik'u' II–Yaljob'ach obsidian counts for the site of Motul de San José organized by area

572

11.4

Obsidian artifact classes at Trinidad de Nosotros sorted by phase with the percentage of the assemblage indicated

579

11.5

Prismatic blade to cores and cores per 100 blades in the Sik'u' II–Yaljob'ach Motul de San José Area

582

11.6

Cores+ to prismatic blades in the Sik'u' II–Yaljob'ach Motul de San José Area

584

11.7

CE/M in the Sik'u' II–Yaljob'ach Motul de San José Area

588

11.8

CE/M at Trinidad de Nosotros

589

11.9

Average blade width and thickness in the Sik'u' II–Yaljob'ach Motul de San José Area

592

11.10

Average blade widths at Trinidad de Nosotros

594

11.11

Obsidian source procurement at Trinidad de Nosotros

599

11.12

Obsidian source procurement in the Sik'u' II–Yaljob'ach Motul de San José Area

601

11.13

Ordinal ranking of Sik'u' II–Yaljob'ach sites using various obsidian indices

604

xxi

B.1

Trinidad de Nosotros burial summaries

640

B.2

Trinidad de Nosotros cache summaries

653

xxii

LIST OF FIGURES

Figure

Page

1.1

Map showing the locations for Trinidad de Nosotros, Motul de San José, and other archaeological sites in the Lake Petén Itzá Basin

2

1.2

Map of the Motul de San José area

5

1.3

Map of Trinidad de Nosotros

7

2.1

Harbor types defined in the World Port Index

21

3.1

Map showing the locations for Trinidad de Nosotros, Motul de San José, and other archaeological sites in the Lake Petén Itzá Basin

35

4.1

Map showing the locations for Trinidad de Nosotros, Motul de San José, and other archaeological sites in the Lake Petén Itzá Basin

67

4.2

Map of the Motul de San José area

71

4.3

Beach approximately 100 m east of Trinidad

73

4.4

The Trinidad harbor, beach, and inlet

74

4.5

The Trinidad lagoon and beach

75

4.6

Colorized digital elevation model (DEM) for the Lake Petén Itzá basin with major sites, drainages, and modern towns indicated

79

4.7

Trinidad's location seen from a launch on Lake Petén Itzá

80

5.1

Map of Trinidad de Nosotros

107

5.2

Three-dimensional surface map of Trinidad de Nosotros

111

xxiii

5.3

Structure E-1 and associated platform viewed from the west

115

5.4

Trinidad surface map view from the south, over Lake Petén Itzá

116

5.5

Monument 1 during excavation

121

5.6

Trinidad site center with contours and small groups removed

131

5.7

Structure A-1 viewed from the north

132

5.8

The view looking south from Structure A-1

133

5.9

Group A/D

134

5.10

The Trinidad Ballcourt and associated structures (Group F)

136

5.11

The Trinidad Ballcourt's eastern lateral structure (F-2), viewed from the northwest

137

5.12

Groups B and C from the top of Structure A-1, looking northeast

140

5.13

Groups B, C, L, and W

141

5.14

View of Plaza V from Structure E-1

146

5.15

Trinidad harbor area with principal structures indicated

150

5.16

The high, eastern end of the isleta/Platform GG, viewed from Platform EE

152

5.17

Trinidad Platform GG seen from Lake Petén Itzá with form outlined in white

153

5.18

Trinidad's principal harbor features with varying lake levels

155

5.19

MSJ Residential group classification scheme

157

5.20

MSJ Type II, Type V, and Type IV groups east of the Trinidad site center

159

5.21

MSJ Type IIIB groups south of the Trinidad site center

160

6.1

Trinidad site center with stratigraphic test units highlighted

175

xxiv

6.2

East profile of Unit 1A1

180

6.3

Eroded Sik'u' II polychrome vessel from Burial 1 (Complete Vessel 1)

182

6.4

North profile of Unit 1I1

185

6.5

North profile of Unit 1B1

187

6.6

East profile of Unit 1F1

190

6.7

Monument 1 during excavation

192

6.8

South profile of Unit 1F2

193

6.9

North profile of Unit 1F3

199

6.10

North profile of Unit 1P2

202

6.11

North profile of Unit 1E1

203

6.12

Cache 1 following excavation

204

6.13

North profile of Unit 1C1

207

6.14

Unit 1C1 Postholes 1 and 2 following excavation

210

6.15

West profile of Unit 4A13

211

6.16

North profile of Unit 1G1

216

6.17

Trinidad Burial 5

218

6.18

Sierra Red basin from Burial 5 (Complete Vessel 2)

219

6.19

Partially reconstructible Alta Mira Fluted vessel from Midden 1

220

6.20

North profile of Unit 5A7

223

6.21

East profile of Unit 1K1

226

6.22

East profile of Unit 1N1

228

6.23

West profile of Unit 1O1

230

xxv

6.24

East profile of Unit 1U1

237

6.25

East profile of Unit 1Y1

239

7.1

Trinidad harbor area with excavations indicated

248

7.2

North profile of Unit 2A4

252

7.3

Eroded Sik'u' II polychrome vessel from Burial 3 (Complete Vessel 4)

254

7.4

North-south section of Unit 12A1

257

7.5

North-south section of Units 12A1 to 12A6

258

7.6

Wall 1 and Floor 3 during excavation of Unit 12A1

261

7.7

North profile of Unit 2B2

270

7.8

North-south section of Unit 12A6

276

7.9

South profile of Unit 2A1

283

7.10

North profile of Units 2E1 to 2E5

288

7.11

South profile of Unit 2A5

294

7.12

North profile of Unit 2A3

296

7.13

Prismatic blades from Trinidad Cache 4

297

7.14

North profile of Unit 12B1

302

7.15

MSJ Project personnel digging around boulder-sized fill stones in Unit 12B1

303

8.1

Map of Trinidad with groups tested for middens indicated

318

8.2

Excavation of 50-x-50-cm midden tests along the western edge of Plaza V

320

8.3

Location of midden Units 9B1 and 9B2

325

8.4

North profile of Unit 9B1

327

xxvi

8.5

Location of midden units associated with the Trinidad Ballcourt

331

8.6

Ballcourt Locus C: north profile of Units 10D1, 10D11, and 10A19

333

8.7

Ballcourt Locus B: north profile of Unit 10D2

336

8.8

Ballcourt Locus C: north profile of Units 10D3, 10D5, and 10D7–10D9

338

8.9

Ballcourt Locus C: north profile of Unit 10D10

340

8.10

Tinaja group jar necks from Cache 3

341

8.11

Deer bone rasp from Midden 13

348

8.12

Miniature "graffiti stela" from Midden 13

349

8.13

Burial 7 during excavation

351

8.14

Eastern exterior face of Structure F-4 with eroded plaster facing removed

353

8.15

Location of midden units and other excavations around Group C and Structure B-1

359

8.16

West profile of Units 3C1 and 3C3

361

8.17

Location of midden units in Groups Q, S, T, U, and Y

367

8.18

North profile of Units 11A1 and 11E1

369

8.19

West profile of Units 11E2, 11B14, and 11E4

371

8.20

East profile of Unit 11E3

373

8.21

East profile of Units 13E1, 13D1, and 13E2

378

8.22

West profile of Units 13C4, 13E3, and 13E5

382

8.23

Cache 5 during excavation

384

8.24

Cache 5 miniatures (Complete Vessels 5–8)

384

8.25

Partial flute from Cache 5

385 xxvii

8.26

Location of midden units for Groups H and KK

389

8.27

North profile of Units 19C6 and 19E1

390

8.28

South profile of Units 19E2 and 19D5

392

8.29

Location of midden units for Group K

394

8.30

North profile of Unit 18E1

396

8.31

North profile of Unit 18E2

398

8.32

South profile of Unit 16E1

401

8.33

North profile of Units 15A1 and 15E1

405

8.34

South profile of Units 15E2 and 15B2

407

8.35

South profile of Unit 14C1

409

8.36

Location of midden units for Group O

412

8.37

North profile for Units 8E1, 8E2, and 8D4

413

8.38

Location of midden units for Group Z

416

8.39

South profile of Unit 17E1

417

8.40

North profile of Units 21E1 and 21B1

421

8.41

North profile of Units 21D9 and 21E2

423

9.1

Map of Trinidad showing groups examined by architectural testing program

430

9.2

Group C with Units 4A1–4A15 indicated

433

9.3

Plan view of Structure C-1, Units 4A1–4A12

435

9.4

North-south section of Structure C-1

436

9.5

Structure C-1 from the south-southwest

438

9.6

Structure C-1: interior face of rear wall within northern gallery

439

9.7

Structure C-1: east bench

441 xxviii

9.8

Group G with Units 5A1–5A9

445

9.9

Plan view of Structure G-1, Units 5A1–5A9

447

9.10

East-west section of Structure G-1

448

9.11

Structure G-1 during excavation

450

9.12

Group FF with Units 2D2–2D20 indicated

455

9.13

Plan view of Structure FF-1

456

9.14

Notched sherds from Structure FF-1

459

10.1

Box-and-whisker plot of pooled calibrated date ranges for all Chukan and Sik'u' I–II AMS assays from Trinidad de Nosotros

474

10.2

The Trinidad de Nosotros ceramic sequence in comparison to other sequences in the Maya lowlands

481

10.3

Rim profiles for Aj Wo' and Ix Cha' Middle Preclassic ceramic groups

486

10.4

Achiotes ceramic group rim profiles

491

10.5

Achiotes and Sierra ceramic group rim profiles

493

10.6

Sierra ceramic group rim profiles

495

10.7

Sierra and Polvero ceramic group rim profiles

496

10.8

Rim profiles for various Chukan and P'ich 'Ayim complex ceramic groups

500

10.9

Quintal and Aguila ceramic group rim profiles

508

10.10

Rim profiles for various P'ich 'Ayim and 'Ayim-tun complex ceramic groups

509

10.11

Tzakol sphere materials from Casa de las Américas

510

10.12

Cambio ceramic group rim profiles

513

10.13

Tinaja ceramic group rim profiles, part 1

515

10.14

Tinaja ceramic group rim profiles, part 2

517

xxix

10.15

Tinaja ceramic group rim profiles, part 3

518

10.16

Infierno, Azote, and unnamed white ceramic group rim profiles

520

10.17

Saxche-Palmar ceramic group rim profiles, part 1

523

10.18

Saxche-Palmar ceramic group rim profiles, part 2

524

10.19

Saxche-Palmar ceramic group rim profiles, part 3

526

10.20

Saxche-Palmar ceramic group rim profiles, part 4

528

10.21

Saxche-Palmar ceramic group rim profiles, part 5

530

10.22

Saxche-Palmar ceramic group rim profiles, part 6

531

10.23

Saxche-Palmar ceramic group rim profiles, part 7

532

10.24

Saxche-Palmar ceramic group rim profiles, part 8

533

10.25

Rim profiles for various Sik'u' I–II ceramic groups

534

10.26

Rim profiles for various Yaljob'ach and Säk-tunich complex ceramic groups

541

10.27

Rim profiles for various Säk-tunich and Aj B'oj complex ceramic groups, part 1

547

10.28

Paxcamán ceramic group rim profiles

550

10.29

Rim profiles for various Säk-tunich and Aj B'oj complex ceramic groups, part 2

551

10.30

Historical pottery from Trinidad de Nosotros

560

11.1

Map of the Motul de San José area with obsidian sample sites indicated

565

11.2

Map of Motul de San José and environs with Motul areas (Central, North, and East), Acropolis, and Chäkokot indicated

569

11.3

Prismatic blade segments from Akte, recovered as part of cache associated with Akte Stela 1

574

11.4

Prismatic blades from Trinidad Cache 4

575

xxx

11.5

Prismatic core and core fragments from Trinidad and Motul

575

11.6

Obsidian scrapers from Trinidad de Nosotros

577

11.7

Obsidian points from Trinidad de Nosotros

578

11.8

Prismatic cores to prismatic blade ratios in the Sik'u' II– Yajob'ach Motul de San José area

583

11.9

Cores+ to prismatic blade ratios in the Sik'u' II–Yajob'ach Motul de San José area

585

11.10

Total cutting-edge-to-mass (CE/M) rations for the Sik'u' II– Yajob'ach Motul de San José area

588

11.11

CE/M ratios at Trinidad de Nosotros by phase

591

11.12

Average blade widths for the Sik'u' II–Yaljob'ach Motul de San José area

593

11.13

Niton XRF analyser with sample hood attachment during the second round of XRF sourcing in 2007

595

11.14

Scatter plot of Zr and Rb base-10 logged concentrations for obsidian samples from Trinidad de Nosotros

596

11.15

Obsidian sources at Trinidad, organized by phase

600

11.16

Sik'u' II–Yaljob'ach obsidian sources for the Motul de San José area, organized by site

602

xxxi

1

CHAPTER ONE INTRODUCTION TO THE ARCHAEOLOGY OF TRINIDAD DE NOSOTROS, EL PETEN, GUATEMALA

Perhaps no subject has elicited more discussion and less harmony in Maya studies than the political economy of Classic Maya polities. Encompassing both the economics of daily life and the sociopolitical linkages binding commoners, elites, and governing royalty, these political economic structures are obviously of central importance to our understandings of ancient Maya society as a whole. Yet, despite the fundamental nature of this topic for Maya studies, relatively few investigations have sought both to directly address political economic questions and to compile local data sets appropriate to interpretation at a broader scale. This dissertation constitutes a direct attempt to examine the political economy of a single Late Classic Maya polity through the study of a specific ancient Maya port. Trinidad de Nosotros, the focus of this study, is located on the north shore of Guatemala’s Lake Petén Itzá, 35 km southwest of Tikal and 2.6 km southeast of the Late Classic ceremonial center of Motul de San José (Figure 1.1). By virtue of its physical location, at an excellent natural portage, and its strategic position, intermediate between a political capital and long-distance trade routes, Trinidad provides an unusually rich opportunity to explore the relationship between trade and politics in a small Late Classic Maya polity.

Figure 1.1. Map showing the location of Trinidad de Nosotros, Motul de San José and other archaeological sites in the Lake Petén Itzá Basin. Map coverage indicated on inset map of Maya lowlands. Map by Ellen Moriarty (2012). 2

3

In broad strokes, this study was designed to examine the political economy of the Motul de San José polity by determining the extent of Trinidad’s role in long-distance trade. Although primate political centers are frequently assumed to have served as centers of distribution for long-distance trade goods, the body of empirical evidence supporting this contention is surprisingly limited. To determine the extent to which Trinidad, a long-established port, or Motul, an emergent political capital, played a more central role in the distribution of exotic resources in the Motul polity, data from Trinidad and the Motul area are compared against three functional site models. Assessing the degree to which Late Classic Trinidad functioned as a simple coastal site, a transshipment port, or a trading port with a high level of involvement in the acquisition and distribution of long-distance trade goods provides a glimpse into the political and economic strategies of the emergent Motul dynasts. This study has obvious implications for present debates concerning both the organization of ancient Maya polities and the nature of political power in the Maya lowlands. Field investigations, conducted between 2003 and 2005, included 21 operations, excavation of more than 330 m3 of matrix, and the collection of several hundred thousand artifacts. Laboratory research from 2003 to 2007 included type-variety analysis of nearly 100,000 potsherds, eleven AMS radiocarbon assays, and detailed analysis of more than 4,000 obsidian artifacts. These investigations were augmented by numerous collaborators conducting their own programs of research. Cumulatively, the data set resulting from two seasons of work at Trinidad is voluminous and sufficient for several focused studies. The following dissertation deals exclusively with research directed by

4

the author and directly applicable to Trinidad’s developmental sequence and its involvement in long-distance trade in the Late Classic Motul de San José area. It also draws on eight seasons of research at Motul de San José and other local sites for comparative data (Figure 1.2). This chapter serves as a précis for the dissertation’s structure. The second chapter focuses on the dissertation’s principal theoretical concerns. As noted above, this study is concerned with the political and economic structures of Classic Maya society. The first half of Chapter 2 introduces current debates regarding these structures and highlights the most prominent theoretical positions. The second theoretical concern, the archaeology of ancient ports, is addressed in the second half of Chapter 2. Ports are phenomena encountered in virtually all complex societies and have been subject to varying definition and classification. The second half of Chapter 2 reviews port definitions and models, with an emphasis on their material correlates and the principal factors influencing port operation. A part of this discussion, a brief summary of the large and growing data regarding ancient Maya ports is presented. Chapter 2 concludes with a review of extant knowledge of ancient Maya trade systems. Trinidad’s utility as a possible port and study site was largely derived from its location and cultural context. Chapter 3 provides a brief synthesis of Lake Petén Itzá’s geological and cultural history. This chapter also includes a detailed discussion of longterm developments in the Motul de San José area, a small sub-unit of the lake basin. During the Late Classic period, this area formed the nucleus for a small polity. Although the historical record regarding Motul and the associated “Ik’ polity” is fragmentary and

Figure 1.2. Map of the Motul de San José Area. Map by Ellen Moriarty (2012). 5

6

incomplete, extant archaeological and epigraphic data paint a relatively clear picture regarding Motul's developmental trajectory. The short, punctuated history of Motul provides an important context and comparison for assessing Trinidad’s much longer history and role as a port. Chapter 4 presents the geologic, geographic, and historical evidence that led to Trinidad’s initial identification as a port. This chapter also introduces the research design for work at Trinidad, including a discussion of the three site models tested and summary descriptions of the programs of research initiated to test these models. Investigations also included a significant culture-historical component and the specific programs of research designed to provide primarily chronological or spatial data are described here. This chapter concludes with an overview of previous research at Trinidad and a discussion of field methods. One of the principal results of investigations at Trinidad was the preparation of a detailed site map (Figure 1.3). Chapter 5 describes the survey and mapping methods utilized at Trinidad and presents the final map. Although Trinidad is a modest-sized site, mapping revealed it to be a complex settlement with a sophisticated central precinct and elaborate public architectural complexes. This chapter presents an overview of public architectural assemblages at Trinidad, with an emphasis on the organization of public space. Further, to facilitate inter-site comparisons within the Motul area, this chapter also reviews Late Classic residential group architecture at Trinidad, Motul, and other nearby sites.

7

Figure 1.3. Map of Trinidad de Nosotros.

8

Field investigations at Trinidad included four separate programs of excavation. These are presented in Chapters 6–9. Chapter 6 describes the site center testing program. Designed to delineate the constructional history of Trinidad’s site center, this testing program included excavations in four public plazas, three public architectural complexes, and seven residential groups. This chapter includes an overview of sampling strategies, excavation methods, and recording procedures, as well as detailed summaries for each excavation unit. The concluding section of the chapter presents a brief synopsis of architectural developments in Trinidad’s site center. The second program of excavation, presented in Chapter 7, focused on delineating a series of features located in the shore area at Trinidad and identified as components within a possible harbor. Excavations there focused on three features: a large platform within a regularly flooded area and thought to have served as a harbor breakwater, a platform potentially utilized as a loading platform, and a possible dock. Additional investigations focused on the slope above the harbor area, nearby residential platforms, and the seasonally inundated area directly adjacent to the inferred harbor facilities. As with the preceding chapter, this chapter includes a review of sampling and excavation methods, as well as a unit by unit summary of excavation results. The conclusion to this chapter provides a brief overview of long-term architectural developments in the harbor area, as well as a discussion of the activities that can be inferred from artifactual data. Chapter 8 presents the results of the site center midden testing program. This program of research was designed to ensure that a robust sample of obsidian was available from Trinidad for intersite comparison. Midden excavations also provided data

9

useful in assessing household and public activities at Trinidad. A total of thirteen residential groups and two public architectural complexes were tested systematically for middens. Investigations included both small midden test units and larger follow-up excavations. Midden tests are summarized briefly by group, while larger excavations are discussed in more complete detail. Summary artifact counts for the principal artifact classes, as well as more detailed ceramic assemblage data for Late Classic and Terminal Classic deposits, are provided for each of the middens. As this program also revealed extensive information regarding the temporal occupation of groups not tested in other programs of research, this chapter concludes with a short description of culture-historical results. Investigations at Trinidad also included partial clearing of structures in three residential groups. These excavations are described in Chapter 9. Although minor in scale, these excavations provided a glimpse at architectural style at Trinidad. These investigations added to the growing dataset regarding Late Classic settlement in the Motul de San José area and facilitated the interpretation of survey data at Trinidad where mounds have been heavily trampled by cattle. Investigations at Trinidad produced a continuous record of occupation from the early Middle Preclassic period through to the modern historical era. Although this dissertation focuses on the Trinidad’s Late and Terminal Classic period occupations, the delineation of Trinidad’s occupational sequence was one of the study’s principal results. Chapter 10 reviews the methods utilized in chronological research at Trinidad, and provides an overview of the cultural phases defined for Trinidad. As the principal source

10

for temporal control was ceramics, this chapter provides a description of ceramic analyses and a detailed synopsis of Trinidad’s ceramic sequence. Further, dating methods at Trinidad included eleven AMS radiocarbon assays and the results of these analyses are also presented in Chapter 10. Assessing Trinidad’s role in long-distance trade was largely dependent on the analysis of obsidian, discussed in Chapter 11. The widespread and easily identifiable nature of obsidian made it an ideal focus for assessing Trinidad’s role in trade. To provide a significant test of Trinidad’s port function, more than 4,000 obsidian artifacts from six sites in the Motul area were analyzed. Approximately three-quarters of the obsidian assemblage came from Late Classic contexts, but materials from other periods were included to provide a context for Late Classic patterns. Obsidian analyses included morphological classification, but emphasized a variety of obsidian indices normally utilized to assess conservation and use. A robust sample of obsidian from multiple sites and time periods was also sourced using XRF analyses. This chapter concludes with a discussion of hypothesized obsidian distribution networks in the Motul de San José area. The conclusions to this study are presented in Chapter 12. The first section summarizes Trinidad’s historical development as an ancient Maya community. Trinidad’s initial settlement and subsequent growth appears to have resulted in large part as a result of its strategic position at the gateway to the Motul area. Long-term patterns in local obsidian data and a continuing interest in the construction and maintenance of the harbor area suggest that Trinidad’s residents parlayed initial advantages in access to local and long-distance trade routes into a long-term function as port for the larger Motul de

11

San José area. Secondary indicators, including the specific timing of Trinidad’s peaks in size and complexity, strongly support this interpretation. The second half of Chapter 12 addresses Trinidad’s role in trade during the Late Classic period. Both harbor area excavations and obsidian analyses support the contention that Trinidad functioned as a trading port during Motul’s Late Classic ascendancy. Further, comparison of the early and late phases of the Late Classic at Trinidad are particularly revealing, and provide clear evidence for continuity in Trinidad’s function despite dynamic local political developments. Indeed, Motul’s emergence as a political center and the corresponding local population build-up appear to have enhanced, rather than dislocated, Trinidad’s role in long-distance trade. Although this study was limited in scale, its results have obvious implications both for our understanding of Classic Maya political economy and the study of ancient Maya ports. In particular, the high level of continuity in Trinidad’s function under the auspices of the emergent Motul dynasty suggests a relatively low level of royal involvement in the material economy of the Motul area. Likewise, Trinidad’s continuing role as a trading port suggests that the political strategies of emergent dynasts, like those at Motul, may have favored the adoption and co-option of local economic systems, rather than replacement and reorganization. Finally, the apparent link between Trinidad’s economic role and large scale commensal events such as those identified for Trinidad’s ballcourt area suggests a future line of inquiry in the Maya area.

12

CHAPTER 2 THEORETICAL ORIENTATION

INTRODUCTION The principal objectives of this study were to define the role of the ancient Maya port of Trinidad de Nosotros in long-distance trade and, by doing so, to examine the political economy of a small Late Classic Maya polity. The political economic structures of Classic Maya society are among the most hotly debated topics in Maya studies. This chapter starts with a review of this debate, outlining the principal theoretical positions. The second section introduces the primary features and correlates of ports in the modern world as well as in the historical and archaeological record. Although ports constitute a diverse range of communities, they share a number of key attributes that make them excellent locations for addressing political economic questions. The third section reviews the results of previous research at ancient Maya ports and current understandings of longdistance trade in the Maya area. Although this research has been considerable, most investigations have focused on maritime ports and coastal trade routes. Inland routes and, in particular, the facilities and organization of inland trade, are much less well understood. The final section of this chapter summarizes the research opportunities provided by the study of an inland port like Trinidad de Nosotros.

13

CLASSIC MAYA POLITICAL ECONOMY AND LONG-DISTANCE TRADE Recent advances in Maya studies have revolutionized our understanding of the ancient Maya. The histories of Classic Maya cities, once thought of in general and highly speculative terms, can now be described in rich and vivid detail (Martin and Grube 2000). Studies focusing on Classic Maya rulers and their close associates have produced a wide array of insights into their day-to-day lives (Inomata and Houston 2001a, 2001b), their roles as religious and secular leaders (Demarest 1992; Freidel 1992), and even their thoughts (Houston 2000, 2001; Houston and Taube 2000). Further, increasing emphasis on Maya commoners (Johnston and Gonlin 1998; McAnany 1993; Manzanilla 1986; Wilk and Ashmore 1988) and the communities of which they were part (Bartlett and McAnany 2000; Canuto and Yaeger 2000; Joyce and Hendon 2000) has led to a greater appreciation of Classic Maya society as a pluralistic one. Yet, despite the depth and breadth of these and other advances, some of the most basic aspects of ancient Maya society remain only vaguely understood and subject to considerable ongoing debate. Perhaps no subject has elicited more discussion and less harmony than the political economy of Late Classic Maya polities (e.g., Chase and Chase 1996; Demarest 1996; Fox et al. 1996; Lucero 1999; Masson 2002; P.Rice 2009). Following Masson (2002:2) and Smith (1991:34), political economy is understood to encompass both the “ways in which economic production and exchange are manipulated to support the power of a society’s leaders” and the sociopolitical structures that underlie and sustain the political and economic systems (see also Sharer and Traxler 2006:632–635). Encompassing both the basic economics of everyday life and the sociopolitical linkages

14

binding commoners, elites, and governing royalty, political economic structures are of central importance to our understanding of Classic Maya society as a whole. Further, as the inferences drawn from this debate frame our interpretations of Classic Maya history, our understanding of Classic Maya political economics has an unusually prominent role in shaping our perspectives on long-term developments and processes in the Maya lowlands (Sharer and Traxler 2006:82–85,632–637). Perspectives on the organization of Classic Maya polities are commonly divided into two opposing camps: those supporting a strong, centralized state model and those favoring a more decentralized state model. Although models for Classic Maya political organization actually are less reified and include much more diversity than these two extremes would suggest (Chase and Chase 1996:804; Marcus 1993, 1998), it is useful to consider some of these idealized models’ attributes. Proponents of the centralized state model have emphasized the great scale and complexity of the largest Classic Maya cities (Chase and Chase 1996; Chase et al. 1990; Culbert 1991; Folan 1985, 1992, 2002), arguing that they constitute entities qualitatively distinct from the mass of smaller secondary and tertiary settlements in the surrounding landscape. In this view, the siteplanning, public works, and subsistence needs of such large population centers and their attached polities require a high level of centralized planning (Chase and Chase 1996:499). Within the strong state model, these needs are met by the governing elite who, in addition to their religious, ideological, and kin-based roles in society, also administer and manage the surrounding polity (Chase and Chase 1996). Such polities are, within strong

15

state models, typically large, territorial entities (Culbert 1991:140–144; Sharer and Golden 2004:25–26), much more in line with the regional states described by Adams (Adams and Jones 1981; see also Marcus 1993) or the super-states identified by Martin and Grube (1995, 2000), than the small polities envisaged by Mathews (1991) and others (Houston 1993). Within the strong state model, both the prestige and subsistence economies are expected to have been at least indirectly managed at the highest levels of society. Support for this perspective is found in dendritic causeway networks linking subordinate sites and surrounding areas to large capital centers (Chase and Chase 2004; Chase et al. 1990:501–502; Folan et al. 1995), the identification of central marketplaces at major centers (Becker 2003:266; Chase and Chase 2004; Coe 1967:73; Dahlin et al. 2007; Folan 1983; Folan et al. 2001; Jones 2003; McAnany 1986:289; Sharer 1988:56; Tourtellot and Sabloff 1994:88–90), large-scale public works (Puleston 1978; Scarborough 1996; Scarborough et al. 1995) and evidence for extensive diversity in economic activity at some of the largest centers in the Maya lowlands (Becker 1973; Chase and Chase 2004; Chase et al. 1990:501; Moholy-Nagy 1997). Within the strong state model, capital centers are expected to have operated, at least in part, as economic central places where the production and distribution of goods was administered directly or indirectly at central markets (Chase et al. 1990; see also Sanders and Webster 1988:525, 535–539), or as part of sponsored pilgrimage festival-fairs (Freidel 1981). The decentralized, weak state perspective on Maya society actually consists of two distinct models that emphasize the person of the ruler, rather than the state

16

administrative apparatus. In the “segmentary state” model, the state is founded upon kinship and marriage alliances constructed by the ruler (Ball 1993; Fox et al. 1996; Sharer and Golden 2004:25). In the “galactic polity” or “theater state” model, the state is constructed through the ruler’s role in ideological systems (Demarest 1992; Houston 1993). In both of these perspectives on the decentralized state, the power of the king is proportional to performance in ritual activity. The capital cities of such states are not seen as a class of entity distinct from outlying settlements, but rather as larger reflections of the same basic principles (Sharer and Golden 2004:25). Polities in both the “segmentary” and “theater” state models are correspondingly small (Fox et al. 1996; Sanders and Webster 1988; Sharer and Golden 2004:25). Within weak state models, the economic roles of the ruler and the capital city are greatly reduced from than that seen in centralized state models. The largest settlements are seen as “regal-ritual” cities (Fox 1977; Sanders and Webster 1988:529–535), and serve primarily ideological, rather than administrative, functions (Fox 1977:41). Economic activity is largely restricted to the “consumption of food, raw materials, and craft products from the outlying countryside,” while production is limited to the output of a small number of craft specialists attached to the ruler’s household (Sanders and Webster 1988:524). The weak state models have received extensive support, both from evidence for prestige good production in elite contexts (Inomata 2001, 2007; Kovacevich et al. 2001; Reents-Budet 1994), the obvious emphasis on performance in artistic depictions and contexts associated with rulership (Inomata 2006; Schele and Miller

17

1986), and the general volatility of Classic Maya polities (Demarest 1997; Houston 1993; Martin and Grube 2000). Recent syntheses and discussions of Maya political organization have also emphasized the dynamic and variable nature of Classic Maya polities (Iannone 2002; Marcus 1993, 1998; Sharer and Golden 2004), stressing the importance of assessing centralization and decentralization in temporal and spatial context (Iannone 2002:68). These perspective highlight the need for a greater emphasis on the regional or polity-level scales of analysis (Graham 1987:753; Masson 2002), as well as closer attention to the resources, commodities, and distribution systems operating within each polity or region (Graham 1987:753; Sharer and Golden 2004). Only by assessing the potential economic systems operating within each potential polity can we begin to understand the potentially complex ways in which these systems were managed at the regional level (Becker 2004; Potter and King 1995; Scarborough et al. 2003). The need for further research at the regional or polity level appears particularly acute with respect to long-distance trade and utilitarian exotics like obsidian. Although core-periphery interaction and the management of long-distance trade are often seen as critical to the elite class in Maya society (Phillips and Rathje 1977; Rathje 1971; Sharer and Traxler 2006:83–85), comparatively few projects have set out to systematically assess this situation (e.g., Aoyama 2001:346; Dreiss and Brown 1989; Ford et al. 1997; McKillop 1996). Within inland areas, recent investigations have highlighted the importance of a few key “gateway” centers for long-distance trade into the Maya lowlands (Dahlin and Ardren 2002; Demarest and Fahsen 2003). Likewise,

18

investigations at some of the largest sites (Aoyama 2001; Moholy-Nagy et al. 1984) have demonstrated that, though obsidian was widely available, some members of society had significantly greater access. Although elites are frequently assumed to have been the principal players in the distribution of exotics like obsidian, as in models favoring the centralized state, very few projects have systematically demonstrated this relationship (Aoyama 2001:354). The need for further research on the relationship between long-distance trade and Classic Maya politics is highlighted by the results of recent research along the Maya area’s maritime coast, discussed below. The emergent complexity of coastal trade networks combined with the evidence for diversity in Classic Maya political organization suggests that the organization of long-distance trade in the Maya lowlands may have been much more complex than once thought. Only by examining this organization at the local or polity level can we see how this may have played out across the Maya lowlands. This study assesses the organization of long-distance trade and political economic structures within a specific Classic Maya polity through the study of its port.

PORTS: DEFINITIONS AND ATTRIBUTES Although the term “port” is often loosely applied, it implies a specific suite of geographic and economic attributes, as well as a host of secondary correlates. At the most basic level, ports are defined as locations utilized to effect the transfer of goods or personnel between modes of transportation (Alexandersson and Norstrôm 1963; Hayuth 1982; Todd 1993:3; Weigend 1958:185). As such, ports are found wherever natural or

19

man-made routes of transportation “meet and intertwine” (Weigend 1958:185). Such locations are normally of strategic significance, and appear prominent on the local landscape. In broader economic terms, ports are defined by the role they play in connecting local and regional economies. Geographers delineate the zones of port interaction through use of the terms “port hinterland” and “port foreland.” The port hinterland is defined as “the [local] area that utilizes the port for the export or import of commodities, services, or ideas” (van Cleef 1945:257; see also Hayuth 1982; Spoehr 1960; Weigend 1958). Prior to the industrial age, port hinterlands were typically defined by physical geography and available natural transportation networks, with hinterlands often conforming to river valleys, lake basins, or other geographically delineated zones with restricted egress (van Cleef 1945:257–259; Weigend 1958). As with gateway communities (Burghardt 1971), ports are situated on the edges of their hinterlands and serve to reduce transport costs into and out of the hinterland (Hirth 1978:37–38). The port foreland, in contrast, is defined as the sum of the areas that interact with the hinterland via transportation routes intersecting at that port (Spoehr 1960; Todd 1993:3). For maritime ports, for example, the foreland is equivalent to the lands “beyond maritime space, and with which the port is connected by ocean carriers” (Weigend 1958:195). Thus, while the hinterland can be described under most circumstances as the port’s local area, the port foreland normally includes localities and areas at the regional and global levels (Todd 1993).

20

As a class of specialized community, ports are also characterized by several key features, as well as a considerable degree of variability. Harbors, above all other features, are so consistently found in ports that the two terms are often used synonymously and even present semantic problems in multiple languages (Andrews 2008). Modern ports, for example, are largely defined by the specific characteristics of the harbor facilities they provide. The World Port Index (WPI) prepared by the National Geospatial–Intelligence Agency (2011) provides a useful summary of modern ports accessible to maritime trade. Ports are classified using several dozen attributes, including associated transportation routes and services, but definitions are dominated by harbor facilities. The harbor types defined by the WPI are presented in Figure 2.1 and Table 2.1 (see also Weigend 1958:186). The WPI classifies ports using eight general harbor types. Although some ports lack harbors, with transshipment occurring over open roadsteads (Type H), most ports include some type of protected harbor. The most simple harbors are natural, where vessels are protected either by natural features (Type A), or by their location in sheltered inland zones (Types D and G). More complex types include those protected by harbor breakwaters (Type B), tidal gates (Types C and F), or artificial slips (Type E). Further, a ninth class of harbor, deemed to provide adequate protection during typhoons and other major tropical depressions, is identified throughout the index (National Geospatial– Intelligence Agency 2011). Although the WPI further classifies harbors according to size and various other attributes, the principal emphasis is on these harbor characteristics.

21

Figure 2.1. Harbor types defined in the World Port Index. Figure modified from National Geospatial-Intelligence Agency 2011:xxvi—xxvii. See Table 2.1 for descriptions.

22

Type

Table 2.1. Harbor types defined in the World Port Index (National Geospatial–Intelligence Agency 2011:xxvi-xxvii). World Port Index Definition

Type A – Coastal (Natural)

A coastal harbor sheltered from the wind and sea by virtue of its location within a natural coastal indentation or in the protective lee of an island, cape, reef, or other natural barrier.

Type B – Coastal (Breakwater)

A coastal harbor lying behind a man-made breakwater constructed to provide shelter, or supplement inadequate shelter already provided by natural sources.

Type C – Coastal (Tide Gates)

A coastal harbor, the waters of which are constrained by locks or other mechanical devices in order to provide sufficient water to float vessels at all stages of the tide.

Type D – River (Natural)

A harbor located on a river, the waters of which are not retained by any artificial means. These facilities may consist of quays or wharves parallel to the banks of the stream, or piers or jetties which extend into the stream.

Type E – River (Basins)

A river harbor in which slips for vessels have been excavated in the banks, obliquely or at right angles to the axis of the stream.

Type F – River (Tide Gates)

A river harbor, the waters of which are constrained by locks or other mechanical devices in order to provide sufficient water to float vessels at all stages of the tide.

Type G – Canal or Lake

A harbor located in the interior portion of a canal or lake that is connected with the sea by a navigable waterway.

Type H – Open Roadstead

A port which has no natural or artificial barrier to provide shelter from the wind, sea, and swell.

Although guides like the WPI are intended for modern navigational and administrative purposes, they underscore the close association between ports and harbors. This association is also reflected in the archaeology of ports. Most of the major port studies conducted in Europe and the Mediterranean basin, for example, have emphasized the investigation of harbors and directly associated facilities (Blackman 1982a, 1982b;

23

Haggai 2006; Hohlfelder et al. 1983; Karmon 1985; Kozlovskaya 2008; Loveluck and Salmon 2011; Marriner and Morhange 2007; Marriner et al. 2006; Marriner et al. 2008; Morhange et al. 2003; Milne 1985; Milne and Hobley 1981; Noureddine 2010; Raban 1985, 1987, 1991, 1995; Rougé 1966), with investigations at the harbors of Berenike (Wendrich et al. 2003), Sidon (Marriner et al. 2006), Tyre (Marriner et al. 2005; Marriner et al. 2008), and Roman London (Bateman and Milne 1983; Rogers 2011) being particularly instructive to the present study. Among many other results, these investigations demonstrated that port location or setting is heavily influenced by the technological capabilities of their builders. Technologically sophisticated cultures utilize a far more diverse suite of locations than those of less technologically sophisticated cultures. In many instances, the earliest ports start as slightly modified natural harbors. Within the Mediterranean and Black Sea basins, for example, the first ports established by early Iron Age Greeks and Phoenicians were focused on rocky bays, coves, or "pocket beaches," where natural conditions provided protection (Marriner and Morhange 2007:146; Marriner et al. 2010:22; Rougé 1966). A shift in emphasis to ports located along clastic coastlines, without the same natural protections, only occurred with the development of regional “superpowers,” like Rome and Carthage, capable of supplying the intensive labor investment and harbor technologies like hydraulic concrete required to take advantage of such locations (Brandon 1996; Marriner and Morhange 2007:146; Oleson 1988; Rougé 1966). These investigations have also revealed a considerable amount of data concerning depositional processes occurring within ancient harbors (Franco 1996; Goodman et al.

24

2009; Hohlfelder et al. 1983; Kraft et al. 2003; Marriner and Morhange 2007; Marriner et al. 2006; Marriner et al. 2008; Reinhardt et al. 1994). Marriner et al. (2010), for example, have collected extensive evidence concerning sediment deposition within ancient harbors of the Eastern Mediterranean, highlighting the impact of dredging and different types of harbor facilities on harbor basin sedimentation. Harbor basins protected by extensive artificial barriers tend to produce difference kinds of sediments than those more accessible to open waters (Marriner et al. 2010). Detailed knowledge of these processes has, in turn, facilitated the identification and interpretation of relict harbors, many of which are now completely filled with sediments and inland from open waters. Harbors are not necessarily a prerequisite for port activity in either modern or ancient examples. Many of the open roadstead (Type H) ports described in the WPI are major oil export centers, despite the fact that terminals are situated miles from the shore and well outside protected waters (Weigend 1958). Though somewhat rarer historically and in antiquity, examples of ports without harbors do appear. The medieval port of Sirāf, for example, lacked a harbor, but nonetheless served as a key entrepôt in 9th and 10th century trade in the Persian Gulf (Whitehouse 1970:143). Likewise, inland ports, especially those protected from the elements by their physical setting or servicing vessels of limited size often have more minor harbor facilities (e.g., Noah 1989). Even in these settings, however, specific landing sites are typically preferred, utilized on a regular basis, and are usually identifiable archaeologically (Ilves 2012). Historical and archaeological research also demonstrate that ports tend to have long occupational histories. This is perhaps most evident in the modern Mediterranean

25

where many of the major maritime centers that exist today have been utilized as ports for thousands of years, with, in some cases, port usage dating as far back as the Neolithic (Blackman 1982a, 1982b; Braudel 2002; Karmon 1985; Marriner et al. 2010). Whether as a result of their strategic locations (Burghardt 1971; Weigend 1958) or their access to long-distance trade routes (Blackman 1982a, 1982b; Elliot 1969; Todd 1993; Weigend 1958), port communities tend to be unusually stable settlement loci. This situation has been noted frequently by economic geographers (Weigend 1958), and is well summarized by van Cleef's (1945:257) observation that "whereas nations rise and decline or disappear entirely, ports persist.” The longevity of ports, however, does not insulate them from larger political and economic processes. In fact, because of their positioning at the interface between a local hinterland and a regional or global foreland, ports tend to be influenced by a wider range of factors outside of their control than other communities. Changes in the availability of resources at the global level, shifting trade routes, competition from other ports, and political decision making in the port's hinterland or foreland can all have dramatic effects on the operation of a specific port (e.g., Doonan 2010; Hoare 1986; Hoyle 1978; Miller 1944; Rogers 1958; Todd 1993; Wendrech et al. 2003). Weigend (1958:187) describes the "never-ending and constantly changing patterns of human activity" as a key factor in the development and decline of most ports. The dynamic history of Bordeaux provides a compelling example of the impact of broader political and economic forces on a particular port (Weigend 1955, 1958). Likely first established as a trading center during the Neolithic, Bourdeaux expanded during the

26

Iron Age through regular contact with Phoenician, Greek, and Carthaginian traders utilizing the Gironde estuary (Weigend 1955:217). Bourdeaux's role expanded under Roman rule as it became an important node in north-south trade and the center of the Roman road network in southwestern Gaul (Weigend 1955:217). Following the collapse of the Roman Empire and the destruction of its port facilities, Bourdeaux underwent a period of decline, reemerging as an important trading center during the late Middle Ages when it constituted the principal maritime link between Aquitaine and England (Weigend 1955:217–219). After a second decline following incorporation with France in 1453, Bourdeaux emerged as a major trading center during the colonization of the Americas. Weigend (1958:187) provides a synopsis of this later peak: In the seventeenth and eighteenth centuries Bordeaux was ideally located for trade and traffic between France and its possessions in the West Indies. Port life flourished, as did the entire region. Yet in the twentieth century Bordeaux finds itself in the backwater of ocean transport. There are many reasons for this change, largely beyond the control of the port and its administration. For example, Napoleon decided to encourage the growing of sugar beets in northern France, in order to reduce the country’s dependence on cane sugar, importation of which was threatened by British blockade. Bordeaux had been the principal importer of cane sugar, and many refineries were located in and near the port area, most of which were eventually liquidated. Moreover, the Industrial Revolution had its greatest impact in Northwestern Europe, and the Channel and North Sea ports became the chief terminals and ports of call for the important sea routes connecting Europe with other continents. As this example demonstrates, Bordeaux's fate as a port center has been largely determined by forces operating outside of its control. Though its position at the intersection of overland and maritime trade routes assured it of continued settlement and

27

a prominent role in local trade (Weigend 1955:217), its peaks and declines were largely brought about by outside factors. Similarly dynamic histories can be found for most ports for which detailed historical records are available (Weigend 1958:187–190). Although their strategic locations typically make them long-term focal points for local hinterland trade and excellent settlement loci, the peaks and valleys in their importance are shaped by outside factors or political decisions beyond their boundaries. This specific combination of attributes make ports unusually rich settings for exploring the dynamic interplay between economic and political processes within a defined hinterland.

ANCIENT MAYA PORTS Ports have long been recognized as important contexts for monitoring long-term patterns in interregional interaction in the Maya area. Interest in assessing ancient Maya trade networks, for example, has led to a major emphasis on port and coastal sites along the maritime margins of the Yucatan Peninsula and Belize, including Isla Cozumel (Freidel and Sabloff 1984; Rathje and Sabloff 1973), Isla Cerritos (Andrews and Gallareta 1986; Andrews et al. 1988), Xcambó (Sierra Sosa 1999), Xelhá (Navarrete 1974), Muyil (Witschey 1988, 1993), Wild Cane Cay (McKillop 1987, 1989, 1996), Ambergris Caye (Guderjan 1995a, 1995b; Guderjan and Garber 1995), Moho Cay (McKillop 1984; 2004), and many others (e.g., Dahlin et al. 1998; Garber 1985; Graham and Pendergast 1989; Masson 2000; Mock 1994). These investigations have produced

28

important cumulative insights into the operation of ancient Maya long-distance trade in coastal areas. Some of the earliest attempts to synthesize coastal trade in the Maya area focused on the port-of-trade concept, a model in which a few independent trading enclaves served as settings for the wholesale exchange of trade goods (Chapman 1957; Rathje and Sabloff 1973). This model assumes that long-distance trade was organized by a small number of elite individuals and that trade during the Classic period and earlier was focused primarily on prestige goods (Guderjan 1995a; Sabloff 1977). Recent findings, however, have led to a major resynthesis in conceptions of ancient Maya coastal trade (Andrews and Mock 2002; Dahlin and Ardren 2002; Demarest and Fahsen 2003; Graham 2002:411–412; Guderjan 1995a; McKillop 1996; McKillop and Healy 1989; O’Mansky et al. 2000). The widespread distribution of ports engaged in trade-related activities suggests that coastal trade was more decentralized than previously thought. Further, the identification of several small ports (e.g., Andrews 1990; McKillop 1996, 2004) that, despite their limited size and architectural complexity, played a major role in the distribution of long-distance trade goods has strengthened the decentralist model. In syntheses of these data McKillop (1987, 1989, 1996) and others (Graham 2002) have argued that coastal Maya trade may have been heavily motivated by economic concerns characterized by factors of “rational choice” and “maximization” of opportunities” (McKillop 1996:49-50), rather than by prestige or ruling elites. An important corollary of this perspective is that ancient Maya coastal trade may have encompassed multiple levels of society and multiple modes of organization.

29

Further, though coastal Maya trade was initially thought to have been a Postclassic phenomenon developing after the dissolution of traditional centers of power at the end of the Classic period (Sabloff 1977; Thompson 1970), the identification of earlier components at many coastal sites has demonstrated that coastal long-distance trade began as early as the Middle Preclassic (e.g., Andrews 1990; Garber 1985). This has led to an increased interest in coastal-inland interaction as a critical factor in the emergence of complex society (Freidel 1978; Freidel et al. 2002). Obsidian-sourcing studies of materials from well-defined port contexts have also provided greater resolution to generalized chronological and spatial models for coastal Maya trade routes (Andrews et al. 1989; Garber 1985; Guderjan et al. 1989; Hammond 1972, 1976; Healy et al. 1984; McKillop 1989, 1995, 1996; McKillop et al. 1988). Detailed knowledge of trade routes has been used to model interregional interaction through time and to refine our understandings of the political and cultural transformations that occurred during the Terminal Classic and Postclassic periods (Bey et al. 1997; Chase and Chase 1982; Demarest et al. 2004; Guderjan 1995b). Coastal studies have also highlighted significant variations in the size, complexity, and function of coastal Maya settlements (Andrews 1990, 2008). Although the descriptive terminologies used by investigators are not always mutually compatible, researchers have identified several functional port and coastal site types (Andrews 1990, 2008; Guderjan 1995a; Hammond 1972, 1976; McKillop 1996, 2004). Most coastal sites exchanged coastal resources in a steady, low-volume manner with nearby inland sites (Andrews 1990:162; Graham and Pendergast 1989). In some instances, particularly

30

where salt production facilities have been identified (Andrews 1983, 1984; McKillop 2002), this trade may have occurred at a much larger scale. A large number of sites also appear to have served an array of functions critical to the maintenance of long-distance trade networks including as way stations for long-distance traders (Guderjan et al. 1988), ports for transshipment between maritime and overland trade routes (Hammond 1972, 1976), and combination pilgrimage-trading centers (Andrews 1990; Freidel 1975, 1978). Recent investigations have also identified a specific class of ports, referred to as trading ports, defined as those ports that specialized in the local and regional distribution of trade goods (Andrews 1990; Hammond 1972, 1976; McKillop 1996:49–50, 2004). Trading ports are specifically distinguished from transshipment ports and ports-of-trade. Although all ports provide transshipment services, transshipment ports are utilized almost exclusively to transfer goods between terrestrial and aquatic modes of transportation (Guderjan 1995a; Guderjan et al. 1988; Todd 1993:3). Trading ports provide transshipment services, like transshipment ports, but are also engaged in the local distribution of long-distance trade goods. Further, in contrast to the port-of-trade model (Chapman 1957; Rathje and Sabloff 1973), the trading port model makes no distinctions regarding the political affiliation of the port nor any assumptions regarding the modes of exchange utilized (McKillop 1996). Several of the trading ports identified to date have been linked to specific inland political centers. Further, wholesale exchange, a key element of the port-of-trade model, has proven exceptionally difficult to identify archaeologically (Andrews 1990:163-166). In broadest terms, trading ports should be seen as highly comparable to gateway centers in

31

serving as distribution centers for long-distance trade goods and providing a broad suite of economic services related to trade (Burghardt 1971; Hirth 1978). Most studies of ancient Maya ports have focused on access to exotics as the principal marker of involvement in long-distance trade (Andrews 1990; Andrews et al. 1988; McKillop 1996:50). Exotics, as an artifact class, can be understood to encompass all goods and resources that were acquired from outside the local area (Adams 1974; Graham 1987; Hammond 1973; Webb 1974). Access to trade goods is often calculated as artifact density in relation to excavation volume or local materials such as pottery, chert, or fauna (Cobean et al. 1971; McKillop 1989, 1996; Parson 1969; Rathje 1972; P.Rice 1984; Sidrys 1977). For certain artifact classes (e.g., obsidian, non-local chert) access can be assessed by the treatment and conservation of that resource calculated by cutting-edge-to-mass ratios or average blade widths (McKillop 1989, 1996; Rovner 1976; Sheets 1978; Sheets and Muto 1972). Finally, obsidian sourcing studies can be used to determine the range of obsidian sources to which a site had access (Healy et al. 1984; McKillop 1989, 1995; Moholy-Nagy 2003b; Nelson 1989; P.Rice 1984). Trading ports are expected to be identified, in comparison to other settlements, by higher densities of exotics, less conservation of obsidian in blade production, and access to a greater range of obsidian sources. The growing perception that long-distance trade in coastal areas was organized in a heterarchical and decentralized manner corresponds well to recent, decentralized models of local economic structure in the central Maya lowlands (e.g., Demarest 1992, 1996; Fox et al. 1996; McAnany 2004; Potter and King 1995; see above). Whether

32

models of ancient Maya long-distance trade derived from coastal research are applicable to inland areas, however, remains to be determined. In many parts of the world inland and maritime trade operate following different conventions and degrees of state involvement (Adams 1974; Knapp 1993; Stargardt 1973). Further, cross-cultural research shows that maritime zones normally have cultural norms distinct from those in inland areas (Ballard et al. 2004; Cooney 2004; Cunliffe 2001; Van der Noort 2004). Although Classic Maya ceremonial centers are frequently assumed to have been centers of distribution for long-distance trade goods, comparatively few studies have examined the distribution of exotic trade goods at the local or intra-polity level (e.g., Aoyama 1995, 1996, 2001; Dreiss and Brown 1989; Ford et al. 1997; Moholy-Nagy 2003b). Although a number of inland ports and port facilities at major centers have been identified (see Andrews 2008; Pugh 2001:16; D.Rice 1996), few of these complexes have been intensively studied (Adams 1999:35; Manahan 2001; Pring and Hammond 1985; Vargas Pacheco 2001). Under these circumstances, it has been exceptionally difficult to model the economic function of Classic Maya ceremonial centers and, in broader perspective, to understand the larger economic structure of Classic Maya polities (e.g., Chase 1998; Chase and Chase 1996; Demarest 1992, 1996; Fox et al. 1996; Marcus 1993; Sanders and Webster 1988).

SUMMARY In summary, a wealth of historical and archaeological research has shown ports to be intriguing contexts for assessing issues related to long-distance trade and the operation

33

of local and regional economies. Their sensitivity to changes in policy or regional economic patterns, as well as their often long histories, make them ideal settings for exploring the nature of these changes over time. Investigations along the Maya coast have highlighted this utility, with several decades of research focusing on ports leading to reformulations of thought concerning the organization of coastal trade. Although longdistance trade constitutes a critical variable in the current debate over the political economy of Classic Maya polities, relatively few projects have sought to directly address this issue with appropriate regional data sets. The investigation of a port within the interior of the Maya lowlands, distinct from nearby political capitals, has the potential to highlight the centralized or decentralized nature of trade within the Maya lowlands.

34

CHAPTER THREE NATURAL AND CULTURAL SETTING

INTRODUCTION The archaeological site of Trinidad de Nosotros is located on the north shore of Lake Petén Itzá, approximately 35 km southwest of Tikal and 3 km northeast of the modern town of San José (Figure 3.1). This location places Trinidad within the larger Central Petén Lakes region and on the northern edge of the Lake Petén Itzá basin. This zone has a number of important traits that mark it as a distinct geographic unit within the Maya lowlands. Further, several decades of intensive research within this zone have defined an exceptionally long history of occupation. The sections that follow provide an overview of local geography and natural resources, a description of recent archaeological research, and a summary of the local culture-historical sequence. A general review of findings from recent investigations at Motul de San José are included here as background to research at Trinidad.

PHYSICAL GEOGRAPHY, ENVIRONMENT, AND NATURAL RESOURCES IN THE LAKE PETEN ITZA BASIN The Central Petén Lakes region is defined by a series of nine major lakes and numerous other small bodies of water situated along an east–west line at approximately 17˚ north latitude (Deevey et al. 1979). The lakes along this line are deep karstic

Figure 3.1. Map showing the locations for Trinidad de Nosotros, Motul de San José, and other archaeological sites in the Lake Petén Itzá Basin. Map coverage indicated on inset map of Maya lowlands. Map by Ellen Moriarty (2012). 35

36

depressions that follow a series of aligned east–west faults related to the Antillean orogenic belt (Anselmetti et al. 2006:53; D.Rice 1993:15; Vinson 1962; West 1964). The main normal fault runs along the northern edge of the lakes system and constitutes an important geologic boundary between Paleocene–Eocene marine carbonates to the north and older Cretaceous carbonates to the south (Anselmetti et al. 2006; D.Rice 1993; Vinson 1962; West 1964). On the ground, this boundary is clearly visible and marked by distinct topographic and ecological settings. To the north of the Central Petén lakes low, rolling limestone hills and intervening bajos are the dominant landforms and high canopy sub-tropical deciduous forests dominated plant communities until relatively recently. To the south, steep karstic “haystack” hills predominate, with savanna grasslands and small forest stands dispersed throughout the intervening areas (D.Rice 1993; D.Rice and P.Rice 1979; West 1964). The Central Petén Lakes region, intermediate between these two zones, is characterized by a mixture of rolling limestone topography and karstic hills, and a complex combination of semi-deciduous forests and savanna grasslands. The largest body of water within this zone is Lake Petén Itzá which covers an area of approximately 100 km2. The lake is divided into two distinct arms: a northern arm that measures approximately 30 km long by 3–4 km wide, and a southern arm that measures approximately 10 km by 1.5 km (Anselmetti et al. 2006:54). A recent bathymetric survey by the Lake Petén Itzá Scientific Drilling Project (http://plaza.ufl.edu/hodell) defined Lake Petén Itzá as the deepest lake in the Maya lowlands with a maximum depth of 160 m (Anselmetti et al. 2006). The deepest portions of the lake are found along the steep north shore of the northern arm, near the modern community of San Pedro and just east

37

of Trinidad. The southern arm of the lake is comparatively shallow at approximately 30 m deep. Seismic images and sedimentary data suggest that Lake Petén Itzá formed during the Pleistocene and has held water continuously up to the present (Anselmetti et al. 2006). The Tayasal Peninsula, a 20-x-10-km landmass, occupies an intermediate position between these two arms. The major geographic features of Lake Petén Itzá are its hydrology and the topography of its surrounding shorelines. Lake Petén Itzá is a closed basin with no surface outflow. The major lake tributaries are to the south where the shoreline shelves gradually inland. The largest of these are the Río Ixlú and the Río Ixpop, located near the southeastern corner of the northern arm of the lake. To the north, the lake catchment is shallow. With the exception of a distinct break in topography at the point where the northern arm of the lake turns southwest at Trinidad, most of the north shore is characterized by steep shorelines backed by a series of east–west running ridges and hills (see Chapter 4). Although numerous seasonal drainages feed into the lake from the north, much of the precipitation within this hilly zone drains to the north and west. The largest of these drainages, the Río K’änte’t’u’ul, runs along the northern edge of the fault-related hills lining the north shore of the lake before turning northwest and joining the Río Ixconob to become the Río Akté, a tributary of the Río San Pedro. Natural resources within the Lake Petén Itzá basin are broadly analogous to those found elsewhere in the southern lowlands, with a few important exceptions. The principal resources are lithic materials, soils, and the lake itself. Lithic resources consist primarily of localized chert nodules relatively evenly distributed across the landscape.

38

Along the northern shore of the lake, chert nodules are particularly exposed as a result of colluvial erosion, and at least one site appears to have specialized in chert tool production (Lawton 2007b). Soils within the Lake Petén Itzá basin are dominated by the Yaxá, Uaxactún, and Machicalá soil series (MAGA 2001a; Simmons et al. 1959). Although modern USDA estimates of soil fertility by Guatemala’s Ministerio de Agricultura, Ganadería y Alimentación (MAGA) rate these soils from “cultivatable with severe limitations” to “not cultivatable” (MAGA 2001b), local estimates, with subsistence milpa agriculture in mind, are more optimistic and recognize a higher degree of local variability. The soil classification system of the Itzaj Maya of San José, for example, recognizes at least six distinct soil classes, of which two—Säkni’is and Ek’lu’um—are highly regarded for traditional milpa agriculture (Atran 1993; U.Cowgill 1961, 1962; Jensen et al. 2007; Moriarty 2001, 2004b; Webb et al. 2007). The more fertile soils are generally located in well-drained uplands whereas low-lying areas are seen as having more limited agricultural potential. Agricultural manipulation of soils within the Lake Petén Itzá area by the ancient Maya appears to have been primarily extensive or bio-intensive, rather than geointensive. Although D.Rice (1996:120) has identified possible raised fields near Lake Petenxil and minor terraces near the town of San Francisco in the savannas south of Lake Petén Itzá (D.Rice 1993:37), little other evidence for geo-intensive agriculture has been identified. Instead, it appears likely that most agriculture in the Lake Petén Itzá basin was dependent on rainfall and focused on upland areas, similar in form to the model

39

proposed by Fedick and Ford (1990) for the Tikal area. Some amount of bio-intensive agriculture was almost certainly practiced as well, and the “agroforestry” model recently proposed by Atran (1993) on the basis of modern Itzaj Maya agricultural methods undoubtedly has great relevance for understanding pre-Columbian agricultural practices in the Lake Petén Itzá basin. Perhaps the most significant of the resources within the Lake Petén Itzá Basin is the lake itself. During the period of ancient Maya occupation in the Lake Petén Itzá basin, lake level is thought to have to been fairly steady, though marked by minor annual and larger episodic shifts. Yearly changes in lake level are generally between 1–2 m and depend primarily on annual variations in the amount and location of precipitation (Anselmetti et al. 2006). Although larger episodic shifts of 5–6 m are recorded and thought to have occurred with some regularity in the past, their causes are not well understood. The most recent episode occurred following the 1976 Guatemalan earthquake when, over a period of two months, lake levels rose approximately 5–6 m, inundating portions of most urban settlements around the lake. Although this particular shift in lake level was linked to tectonic activity and resultant changes in local hydrology, earlier episodes may have been caused by a wide variety of factors including changes in region-wide water tables, shifts in global climatological patterns, human-induced changes in lake catchment hydrology, and many others (Anselmetti et al. 2006; Brenner et al. 2002; D.Rice 1996:205). Although major shifts in the water level of Lake Petén Itzá undoubtedly had deleterious effects on lakeshore settlement during the pre-Columbian period, any

40

negative effects would have been more than offset by the benefits of a lacustrine ecology. The presence of a permanent source of potable water undoubtedly constituted a stabilizing factor on settlement in the basin, and potentially provided the opportunity for hand or pot irrigation, although this possibility has not been explored. Lacustrine resources also included a diverse array of fish, aquatic reptiles, and waterfowl for human consumption. Finally, as subsequent chapters will demonstrate, the lake provided a ready means of transportation and communication, linking and inter-connecting communities throughout the basin.

PREVIOUS RESEARCH IN THE LAKE PETÉN ITZÁ BASIN The first investigations in the Lake Petén Itzá basin were conducted by many of the early archaeologists in the Maya area. Teobert Maler (1908, 1910), Sylvanus Morley (1937-38), Franz Blom (Blom and La Farge 1926-1927), Cyrus Lundell (1934) and Heinrich Berlin (1955) all conducted reconnaissance in the Lake Petén Itzá basin or adjacent areas. In general, most early investigators were interested in documenting preserved Classic period monuments at Ixlú, Motul de San José, Tayasal, and other major sites. Other investigations, such as Guthe’s (1921, 1922) investigations on the Tayasal Peninsula and Borhegyi’s (1963) underwater research, focused more specifically on the search for the contact period capital of the Itzá Maya. The first rigorous culture-historical investigation conducted in the Lake Petén Itzá basin was directed by George Cowgill (1963). As part of his dissertation research, Cowgill excavated a pair of test pits in the central plaza of Flores and conducted brief

41

reconnaissance investigations at Tayasal, Motul de San José, and other sites. Cowgill used his investigations to produce the first detailed Postclassic period ceramic sequence for the Central Petén Lakes region. Although Cowgill’s original sequence has been modified and expanded by more recent studies (e.g., Cecil 2001; Chase 1979, 1983; 1984; P.Rice 1979a, 1979b, 1987a, 1996), his investigations formed an important base for all subsequent work in the Lake Petén Itzá area. The first large scale investigation in the Lake Petén Itzá basin was the University of Pennsylvania University Museum Tayasal Project directed by William R. Coe. The Tayasal Peninsula had long been regarded as the principal candidate for the location of the contact period Itzá Maya capital of Nojpetén (Guthe 1921, 1922; Maler 1910; Morley 1937-38; Reina 1966), and the principal purpose of the Tayasal Project was to locate and excavate that center (Chase 1983:12–15). Excavations focused on the sites of Tayasal, covering the western end of the peninsula, and Cenote, located approximately 10 km to the east. Additional reconnaissance investigations covered much of the Lake Petén Itzá basin and documented approximately 30 new sites (Chase 1983). Although Tayasal Project investigations were “inductively” oriented and many excavations were abandoned when Postclassic materials were absent, they represented, until relatively recently, the most intensive investigations of any site in the Lake Petén Itzá basin. When it was determined, however, that most of the major constructions on the Tayasal Peninsula dated to the Classic period or earlier, field investigations were terminated and future plans were abandoned after only one field season. Arlen Chase (1983) later synthesized the University Museum Tayasal Project data for his doctoral dissertation. Although

42

subsequent investigations have modified some of Chase’s conclusions (e.g., Chase 1976, 1982; Jones et al. 1981; P.Rice and D.Rice 1985), his work (Chase 1979, 1983, 1984) continues to be critical to understanding long-term culture-history in the Lake Petén Itzá basin. Following the Tayasal Project, the tempo and extension of research in the Lake Petén Itzá basin expanded considerably. The most significant of these investigations were a series of multidisciplinary projects directed by Don and Prudence Rice and which formed part of the larger Central Petén Historical Ecology Project (CPHEP) directed by Edward Deevey. Initial investigations focused on the Yaxhá-Sacnab basin (Deevey et al. 1979; D.Rice and P.Rice 1980a), but were eventually expanded to include Lakes Macanché, Salpetén, Quexil, and Petenxil (D.Rice and P.Rice 1980; D.Rice and P.Rice 1981a). Most of the CPHEP archaeological investigations were oriented towards survey coverage, but more intensive research was conducted at a number of sites including Muralla de Leon (D.Rice and P.Rice 1981a) and Macanché (P.Rice and D.Rice 1985). Cumulatively, these investigations produced an enormous body of data, with particular importance to understandings of ceramic development (e.g., P.Rice 1979a, 1979b, 1987), settlement (e.g., D.Rice 1976; D.Rice and P.Rice 1980a, 1990), ecology (e.g., Deevey et al. 1979; D.Rice and P.Rice 1984), and long-term cultural processes in the Central Petén lakes region (e.g., P.Rice and D.Rice 2004), with direct relevance to studies in the Lake Petén Itzá basin. Directly following on the CPHEP investigations, the Rices initiated preliminary investigations in the savannas directly south of Lake Petén Itzá (e.g., D.Rice and P.Rice

43

1979, 1980b). In total, the Rices investigations included survey coverage of approximately 30 km2 and more intensive investigations at several sites. The results of this study suggested that savanna settlement began as early as the Late Preclassic period and extended into the Late and Terminal Classic periods (D.Rice and P.Rice 1979). Although further research is needed to understand Central Petén savanna ecology and the organization of savanna settlement, these results are of particular interest for future investigations in the many small savannas found on the Tayasal Peninsula (see Chase 1983) and other sectors of the Lake Petén Itzá basin (Spensley 2007b). The most recent investigation directed by the Rices, Proyecto Maya-Colonial, was designed to examine the 17th century political geography of the Central Petén lakes region. A key component to this research has been a careful consideration of historic descriptions of the contact period political organization in the central Petén (D.Rice, P.Rice, and Jones 1993; Jones 1998) and a focus on identifying the material cultural attributes of the Kowoj branch of the Itzaj ethnic group. From 1994 to 1999, field investigations focused on the eastern end of the Lake Petén Itzá basin, the area assigned to the Kowoj by ethnohistoric documents, with intensive investigations at Ixlú (D.Rice et al. 1999) and Zacpetén (D.Rice et al. 1998; Pugh 2001, 2002-2004, 2003). The principal results of these investigations were the identification of ceramic (Cecil 2001, 2004), architectural (Pugh 2001), and mortuary (Duncan 2005) indicators for the Kowoj. More recently, Proyecto Maya-Colonial investigations have shifted to the western, or Itzaj, end of the lake, focusing on reconnaissance (D.Rice et al. 1996) and particularly on the site of Nixtún-Ch’ich’ (discussed below).

44

During the last two decades, an increasing number of other international and national investigations have been conducted in the Lake Petén Itzá area. In 1992–1993, the PRIANPEG Project, directed by Richard Hansen, excavated a series of eleven test pits on the island of Flores and conducted preliminary underwater investigations between Flores and the town of San Miguel on the Tayasal Peninsula (Hansen 1997). Although PRIANPEG investigations were discontinued after only one season, Forsyth’s (1996) report on the ceramic sequence of Flores and Marken’s (Marken and Marken 1996) underwater investigations provided important data on the island’s pre-Columbian occupation. More recently, the towns of Flores, Santa Elena, and San Benito were foci for major salvage investigations during the installation of a new urban water system. Significant quantities of artifacts, burials, and other features were recorded while construction crews tore up most of the streets in all of these towns. Lab work and an early GIS synthesis by Gámez (2006) provided the first comprehensive glimpse of the pre-Columbian settlements underlying these modern towns. Finally, within the last decade, both the Guatemalan Atlas Archaeological Project and the Proyecto Maya-Colonial began operations near the western end of Lake Petén Itzá. During the first year of work, the Atlas Project focused on preliminary reconnaissance at the many small sites between the western end of Lake Petén Itzá and Lake Sacpuy (Gerson Martinez, personal communication, 2006), but future investigations are expected to encompass most of the un-surveyed areas around the lake, with more

45

intensive investigations planned for several sites. These investigations can be expected to produce a tremendous body of data (see http://www.atlasarqueologico.com /index.html). Current Proyecto Maya-Colonial investigations are focusing on the site of NixtúnCh’ich’. Located on the Candelaria Peninsula at the western end of Lake Petén Itzá, Nixtún-Ch’ich’ is almost certainly the largest site in the lake basin. Preliminary mapping in 1994 and 1995 revealed it to cover an area of more than 2.5 km2 with more than 450 structures within its core zone. Monumental architectural assemblages include several temples more than 25 m high, two acropoli, and two ballcourts (D.Rice 1996; D.Rice et al. 1996). The largest of the ballcourts is particularly noteworthy in that it may be the second largest in the Maya lowlands, with a playing alley measuring 30 m wide by more than 100 m long (D.Rice 1996; D.Rice et al. 1996). Although Proyecto Maya-Colonial investigations are focusing on the contact period occupation of Nixtún-Ch’ich’, when it may have been occupied by Martín de Ursúa prior to the conquest of the Itzaj (Jones 1998), the site’s great size and complexity suggest that the results of Maya-Colonial investigations will have a major impact on our understanding of the culture-historical sequence in the Lake Petén Itzá basin.

CULTURE-HISTORICAL SUMMARY FOR THE LAKE PETÉN ITZÁ BASIN Previous investigations in the Lake Petén Itzá basin have produced a tremendous quantity of data. Although comparatively few sites have been studied intensively, survey coverage has been extensive and more than fifty sites have been identified and tested to some degree. In broad strokes, these studies have demonstrated that the Lake Petén Itzá

46

basin has among the most complete culture-historical sequences in the Maya lowlands, with an occupation extending from the earliest colonization of the southern lowlands through the contact period and up to the present day (Chase 1983; Jones 1998; D.Rice et al. 1996; Schwartz 1990). The following review provides a framework for later discussions of culture-historical developments at Trinidad de Nosotros. The Middle Preclassic Period The earliest documented settlement of the Lake Petén Itzá basin dates to at least as early as 800–900 B.C. Although palaeolimnological studies suggest human manipulation of the Central Petén Lakes catchments as early as 1700 B.C. (e.g., Brenner et al. 1996:58; D. Rice 1976; Rosenmeier et al. 2002), archaeological evidence for such an early occupation is presently lacking. The earliest confirmed occupations in the Lake Petén Itzá basin are associated with pre-Mamom horizon pottery similar to the early Ah Pam complex identified by Rice at Yaxhá Hill and other sites in the Yaxhá-Sacnab basin (see D.Rice 1976; P.Rice 1979a). Prudence Rice (1996) identified the first early Middle Preclassic materials in the Lake Petén Itzá basin in deep test pits at the sites of Ixlú and Nixtún-Ch’ich’. More recently, Gámez (2006) reported the discovery of pre-Mamom materials on the island of Flores during salvage operations. Finally, the largest preMamom deposits in the Lake Petén Itzá basin have come from Jeanette Castellanos’s (2007) recent investigations at the site of Buenavista-Nuevo San José. Although data concerning the pre-Mamom occupation of Lake Petén Itzá basin are too diffuse to characterize adequately, the available data highlight some potentially important patterns in early settlement. With the exception of Buenavista-Nuevo San

47

José, three sites with pre-Mamom pottery are situated in close proximity to the lake: Nixtún-Ch’ich’ on the Candelaria peninsula, surrounded on three sides by water; Flores, an island in the lake; and Ixlú, intermediate between Lake Petén Itzá and Lake Salpetén. The preliminary correlation between pre-Mamom settlement and the lake suggests that lacustrine resources may have been an important factor in early settlement location, a situation that would correlate well with earlier interpretations of pioneer settlement in Petén and Belize (Andrews 1990). Although Buenavista-Nuevo San José, situated approximately 1.0 km north of the lake, constitutes an important exception to this pattern, its location may be more closely linked to another natural resource. Just south of Buenavista, the Chäk Maman Tok’ chert outcrop appears to have been an important source of medium- to high-quality chert. Lawton’s (2006b) recent identification of intensive chert tool production at La Estrella, a site a few hundred meters south of Buenavista, highlights this possibility. In fact, prior to the modern construction of the modern barrio Nuevo San José, La Estrella and Buenavista may have constituted a single, fairly extensive site, highly involved in the production of stone tools. By the late Middle Preclassic period, settlement in the Lake Petén Itzá basin appears to have been somewhat more widespread. Although deep stratigraphic pits have been excavated in comparatively few sites, Mamom horizon pottery has been recovered in small to moderate quantities at most sites investigated, including Flores (Forsyth 1996; Gámez 2006), San Benito (Gámez 2006), Nixtún-Ch’ich’ (Chase 1983; Cowgill 1963; P.Rice 1996), Ixlú (P.Rice 1996), Buenavista-Nuevo San José (Castellanos 2006; Foias 2003), and Motul de San José (Foias 2003). Mamom materials were also recovered in

48

very small quantities in widely dispersed secondary deposits on the Tayasal Peninsula (Chase 1983). Although more sites will need to be investigated, with a greater emphasis on deep excavations, before late Middle Preclassic settlement is well understood, the available data suggest that the Mamom sphere occupation of the Lake Petén Itzá basin followed patterns set by their pre-Mamom sphere predecessors. Of the sites with appreciable Mamom occupations, only Motul de San José, located 3.5 km north of the lake, is situated outside of the immediate lakeshore area. Thus, as with the pre-Mamom occupation, we can probably surmise that late Middle Preclassic settlement strategies were heavily influenced by proximity to lacustrine resources. The Late Preclassic Period The Late Preclassic period marks the earliest cultural fluorescence in the Lake Petén Itzá basin. Chicanel sphere materials have been recovered at virtually all sites investigated to date and, at many sites, the tempo and extension of Late Preclassic construction appears higher than at any other point in time. Most sites with Middle Preclassic occupations grew significantly during the Late Preclassic and many additional sites were founded. On the Tayasal Peninsula, Late Preclassic constructions were found to be “omnipresent” (Chase 1983:30), and similar results have been noted throughout most of the lake basin (Cowgill 1963; Forsyth 1996; D.Rice et al. 1996; D.Rice et al. 1999). The marked increase in population within the Lake Petén Itzá basin appears to have been accompanied by two significant cultural developments: an increased level of sociopolitical complexity, signaled by the appearance of the first monumental

49

architectural assemblages, and the advent of a more sophisticated subsistence economy, marked by a shift in settlement strategies. Many of the largest buildings ever constructed in the Lake Petén Itzá basin date to the Late Preclassic. At Tayasal, the largest Late Preclassic constructions formed the Structure 256 Group, which consists of three structures situated atop a 12.8 m high basal platform, with the largest structure rising an additional 19.4 m from the eastern edge of the platform (Chase 1983:668). In orientation and layout, this complex is highly reminiscent of the triadic groups that form the dominant architectural tradition in the Late Preclassic north Petén (Hansen 1998). Similarly large Late Preclassic constructions have also been identified at Ixlú, Cenote, Chaltún Grande, and Paxcamán. Further, although few of the major structure groups at Nixtún-Ch’ich’ have been tested, the form and orientation of many of the largest buildings are suggestive of a Late Preclassic origin. The Late Preclassic period also marks several important divergences in basin settlement strategies. Although many lakeshore areas continued to be occupied in the Late Preclassic, some even growing to monumental size at this time, proximity to the lake appears to have been a less important factor in settlement planning than it was during earlier periods. Many of the major Late Preclassic constructions on the Tayasal Peninsula are located on the spine of the peninsula, several kilometers or more from the lake shore (Chase 1983), and many of the new settlements to the north of the lake were situated in upland areas several kilometers from the lake (Moriarty 2004). This period also marked the first settlement of the savanna zones to the north and south of the lake (D.Rice and P.Rice 1979, 1980b; Spensley 2007b).

50

The Terminal Preclassic Period The transition from the Late Preclassic to the Early Classic period in the Lake Petén Itzá basin is marked by the appearance of a Terminal Preclassic or “Protoclassic” ceramic complex related to the Floral Park ceramic sphere. Terminal Preclassic ceramics have been recovered at a somewhat restricted number of sites in the Lake Petén Itzá basin. The largest collections of materials have come from excavations on the island of Flores (Forsyth 1996; Gámez 2006), Tayasal, and Cenote (Chase 1983), and underwater investigations by Borhegyi (1963) and the PRIANPEG project (Marken and Marken 1996). Large deposits of Terminal Preclassic materials were also recovered by the Rices (D.Rice and P.Rice 1981) at the fortified site of Muralla de Leon, just east of Lake Petén Itzá in the Macanché basin. Although there are significant differences of opinion concerning use of the term “Protoclassic,” its chronological placement, and, particularly, whether "Protoclassic" or Terminal Preclassic materials constitute a “complete” ceramic complex or a specialized “ritual” complex (e.g., Brady et al. 1998; Chase 1983:31-33; Culbert 2003:58; Laporte 2003), the data from the Lake Petén Itzá basin tend to support the identification of a complete ceramic complex, transitional between the Late Preclassic Chicanel sphere and the Early Classic Tzakol sphere. On the Tayasal Peninsula, and particularly at the site of Cenote, Chase (1983) identified "Protoclassic" materials in fill deposits sequentially intermediate to Late Preclassic and Early Classic contexts. Further, on Flores, the PRIANPEG project recovered large quantities of Terminal Preclassic or "Protoclassic" materials from fill contexts (Forsyth 1996:8).

51

Although the distribution of Terminal Preclassic materials in the Lake Petén Itzá basin is much more restricted than material from earlier periods, two important developments appear to coincide with the use of these ceramics. First, this period marked the earliest appearance of carved monuments in the lake basin. Chase (1983:74) noted three monuments at the site of Cenote and an additional, possible "Protoclassic" monument at Tayasal. Second, this period seems to have marked an increased emphasis on the construction and use of E-groups within the basin. E-groups or eastern ceremonial assemblages have been securely identified at four basin sites, Cenote, Tayasal, Paxcamán, and Chächäklu’um (Chase 1983; Spensley 2007b), and eight E-groups have been identified in the savanna sites directly south of Lake Petén Itzá (Laporte and Mejía 2005). Although the E-group tradition clearly has its origins in the Preclassic (Ricketson and Ricketson 1937), University Museum excavations in the Cenote E-group documented construction stages dating to both the end of the Late Preclassic and the "Protoclassic" period (Chase 1983:150-153). Recent investigations near the Chächäklu’um E-group noted a similar pattern (Spensley 2006b). Although much further research will be necessary before the Terminal Preclassic in the Lake Petén Itzá basin is adequately understood, the appearance of the carved monument tradition and the apparent emphasis on monumental E-groups, rather than earlier architectural assemblages, suggests that important sociopolitical developments may have been underway at this time. The Early Classic Period The Early Classic period in the Lake Petén Itzá basin is marked by the appearance of ceramics affiliated with the larger Tzakol ceramic sphere. Early Classic materials have

52

been recovered at a fairly extensive list of sites, including Ixlú (P.Rice 1996), NixtúnCh’ich’ (P.Rice 1996), Flores (Forsyth 1996; Gámez 2006), San Benito (Gámez 2006), Motul de San José (Foias et al. 1999), La Estrella (Lawton 2007b), Chächäklu’um (Cowgill 1963), Cenote (Chase 1983), Paxcamán (Chase 1983), Tres Naciones (Chase 1983), and Tayasal (Chase 1983). In most cases, however, investigators have noted that Early Classic materials were only sparsely represented in collections (e.g., P.Rice 1996:292). Further, comparatively few Early Classic constructions have been identified and most materials have come from mixed fill deposits dating to later periods. Systematic test pitting programs at Ixlú (P.Rice 1996:292-293) and Motul de San José (Foias 2003; Foias et al. 1999), for example, identified only a handful of possible Early Classic construction episodes. Although some major Early Classic constructions may appear on the Tayasal peninsula (Chase 1983), they have not yet been fully investigated. Overall, the available data on the Early Classic in the Lake Petén Itzá basin paint a fairly contradictory picture. While the wide distribution of Early Classic materials seems to suggest a healthy occupation of the lake basin, the limited quantities of materials recovered and the dearth of Early Classic constructions seems to indicate a light, transitory occupation. Although, as Chase (1983:33) suggests, evidence for the Early Classic occupation may be obscured by limitations in sampling, the current data suggest the Early Classic marked a downturn in settlement and sociopolitical complexity within the Lake Petén Itzá basin.

53

The Late Classic Period The onset of the Late Classic period marks the second major cultural fluorescence in the Lake Petén Itzá basin. As with the Late Preclassic period, Late Classic materials have been recovered at virtually all tested sites. Further, many sites appear to have reached their occupational and constructional zeniths. At least four basin centers – Tayasal (Chase 1983), Nixtún-Ch’ich’(D.Rice et al. 1996; P.Rice 1996), Ixlú (D.Rice et al. 1999; P.Rice 1996), and Motul de San José (Foias 2003; Moriarty 2004b) – appear to have become major regional ceremonial centers during the Late Classic. Further, the list of large, but slightly smaller and less complex, centers is also quite extensive, and includes: Cenote, Michoacán, Chaltún Grande, and Paxcamán on the Tayasal Peninsula (Chase 1983); Pasajá, situated just west of the lake (D.Rice et al. 1996); Chächäklu’um, located just north of the lake (D.Rice et al. 1996; Spensley 2006b); and probably Jobompiche II, located just north of the east end of the lake (Ian Graham, personal communication, 2002). Although the list of tertiary and smaller settlements where Late Classic occupations have been documented is too large to include, this list includes Isla Santa Barbara, Isla Lepet, Yachul, Tres Naciones, Punta Nimá, Zacpetén, and Flores (Chase 1983; Forsyth 1996; Gámez 2006; D.Rice et al. 1996; P.Rice 1996). The Late Classic period also marked a high point in the erection of monuments in the Lake Petén Itzá basin. Stelae, datable to the Late Classic period or the first part of the Terminal Classic on stylistic grounds or by preserved long count dates, have been documented at ten sites: Motul de San José, Tayasal, Ixlú, Flores, Cenote, Yachul, Chaltún Grande, Michoacán, Santa Elena, and Zacpetén (Chase 1983; Maler 1910;

54

Morley 1937-38; D.Rice et al. 1999; P.Rice and D.Rice 2004). Additional, possible monuments have been identified at Nixtún-Ch’ich’ (personal observation, 2006), the Lake Quexil Islands, and Isla Santa Barbara in Lake Petén Itzá (Chase 1983). Just to the northwest of the basin, the sites of Akte, B'alamtun, and Aguacatal also feature Late Classic monuments (Figure 3.1) Settlement patterns during the Late Classic period largely mirror those seen during the Late Preclassic. Settlements are extensive and are found in virtually all of the available ecological niches in the lake basin. The only significant difference between the Late Preclassic and the Late Classic is a slightly greater emphasis on upland areas during the later period. Although many of the settlements in immediate shoreline and nearshore areas expanded during the Late Classic period, many new centers appear focused on the raised uplands to the north of the lake and along the spine of the Tayasal Peninsula (Chase 1983; Moriarty 2004b; D.Rice 1993). Cumulatively, the evidence for extensive settlement, the density of first- and second-order political centers, and the wide distribution of monuments suggests that the Late Classic Lake Petén Itzá basin constituted a sociopolitical arena of intricate complexity. This description appears particularly apt for the western portion of the lake basin where three major centers—Nixtún-Ch’ich’, Tayasal, and Motul de San José—and a host of smaller secondary centers are clustered within a 100 km2 area, much of which is covered by water. Within this western zone in particular, sociopolitical interactions can be expected to have been complex and highly affected by factors of political geography.

55

The Terminal Classic Period The Terminal Classic period, marked by the appearance of foreign trade wares and modal shifts in Late Classic ceramic complexes, is best described, following P.Rice and D.Rice (2004), as a period of transition and transformation in the Lake Petén Itzá basin. Most of the centers occupied during the Late Classic period continued to be occupied into the Terminal Classic. Terminal Classic materials have been recovered at virtually all tested sites, including Ixlú (P.Rice 1996), Nixtún-Ch’ich’ (P.Rice 1996), Motul de San José (Foias 2003), Flores (Forsyth 1996; Gámez 2006), and numerous sites on the Tayasal peninsula (Chase 1983). In most cases, however, the Terminal Classic marks a gradual lessening in the construction of public architecture and a reduced investment in masonry residential architecture. The Terminal Classic decline does not appear to have occurred in a uniform fashion throughout the lake basin. Chase (1983:1210) interpreted the Terminal Classic as a period of “resurgence” on the Tayasal Peninsula, after a brief decline during the second half of the Late Classic period. Further, the Proyecto Maya-Colonial has identified major Terminal Classic occupations at Macanché, Nixtún-Chi’ich’, and Ixlú, which appears to have been a major political power at this time (D.Rice et al. 1996; D.Rice et al. 1999; P.Rice 1987a). In most cases, however, these developments do not appear to have been protracted, and most sites appear to have undergone dramatic reductions in population by the end of the Terminal Classic. The most significant development of the Terminal Classic, and one with important implications for understanding later periods, was a gradual shift in basin area

56

settlement patterns. During the course of the Terminal Classic, the upland settings favored by Late Classic populations appear to have been gradually abandoned in favor of lakeshore areas and defensible positions on islands and peninsulas (Chase 1983; D.Rice 1990). P.Rice and D.Rice (2004:129-130) interpret this shift as a result of increasing political conflict and competition over scarce resources. The Postclassic and Historical Periods The Lake Petén Itzá Postclassic is normally divided into three separate chronological periods linked to ceramic complexes defined by Chase (1979, 1983, 1984) and Cowgill (1963): the Early Postclassic Chilcob complex, dating from approximately 900–1200 AD; the Middle Postclassic Cocahmut complex, dating from approximately 1200–1450 AD; and the Late Postclassic and early Historical Kauil complex, dating from approximately 1450–1700 AD (see Chase 1979, 1983, 1984; P.Rice 1996:299-305). On the ground, however, the absence of clear stratigraphy at many Postclassic sites and local variations in the distribution of key ceramic types has frequently made it difficult to separate Early, Middle, and Late Postclassic assemblages (e.g., Forsyth 1996; P.Rice 1996). As a consequence, many investigators have chosen to focus on a more generalized Postclassic period, emphasizing long-term developments throughout the Postclassic period. For simplicity, this review follows that approach. General Postclassic materials have been recovered throughout the Lake Petén Itzá basin, and virtually all tested sites have yielded some evidence for Postclassic occupation. Although most investigations in the lake basin have produced Postclassic ceramics in some quantity, particularly large collections have come from the Tayasal Peninsula

57

(Chase 1983), Flores (Cowgill 1963; Forsyth 1996; Gámez 2006), Macanché (P.Rice 1987a), Zacpetén (D.Rice et al. 1998; Pugh 2001, 2002-2004, 2003), Ixlú (D.Rice et al. 1999), and Nixtún Ch’ich’ (P.Rice 1996). Investigations at these and other sites have identified some of the basic characteristics of the Postclassic period occupation in the Lake Petén Itzá basin. Postclassic settlement patterns appear to have followed tendencies first established in the Terminal Classic period. The vast majority of Postclassic settlements identified to date have been found in immediate lakeshore or near shore areas (D.Rice and P.Rice 1990; P.Rice and D.Rice 2004). On the Tayasal Peninsula, Chase (1983:1214-1220) found Postclassic remains limited primarily to the western portion of the peninsula, near Punta Trapeche and the modern town of San Miguel. Reconnaissance investigations by the Proyecto Maya-Colonial have observed similar patterns along the north shore of the lake (D.Rice et al. 1996; Sánchez 1996). Further, many of the largest and most complex Postclassic settlements, including Flores (Cowgill 1963; Forsyth 1996; Gámez 2006; Jones 1998) and Zacpetén (Pugh 2001, 2002-2004, 2003) are situated on easily defensible islands or on peninsulas that were clearly modified for defensive purposes. The Postclassic period also seems to have marked the completion of a major shift in household architectural style. Postclassic residential constructions are normally modest in scale, with stone foundations and perishable walls. These structures are frequently situated atop low basal platforms or the terraced slopes of islands or peninsulas (D.Rice 1988; P.Rice and D.Rice 2004). Frequently, residential structures are

58

C-shaped in form with an open front (P.Rice and D.Rice 2004:132). P.Rice and D.Rice suggest (2004:132) that this style of architecture, which varies significantly from earlier architectural styles, may represent an influx of new settlers from other areas. Although public architecture is not well known within much of the lake basin, extant data also suggest an infusion of new ideas and possibly settlers. In the eastern portion of the basin, Pugh (2001) has identified a style of temple assemblage that closely resembles Late Postclassic temple assemblages at Mayapan (Proskouriakoff 1962). Although the extent to which this style is present in the western portion of the basin cannot yet be inferred, the close similarities between Central Petén complexes and those found at a Mayapán cannot be ignored. Cumulatively, the available archaeological data suggest that the Lake Petén Itzá basin was a major focus of Postclassic settlement in the Central Maya lowlands. Major shifts in settlement location and architectural arrangements, however, as well as the decrease in overall settlement size, suggest that the Postclassic period may have marked a substantial shift in both sociopolitical and community organization. Although the intrusion of new populations cannot be confirmed, it constitutes a strong possibility. Finally, the apparent emphasis on proximity to the lake and other defensible positions suggest that the area was marked by an increasing level of political competition and violence, themes that are well attested in ethnohistoric data from the contact period (e.g., Avendaño y Loyola 1987; Hellmuth 1977; Jones 1998).

59

MOTUL DE SAN JOSÉ: HISTORY OF RESEARCH AND EARLY CONCLUSIONS The archaeological site of Motul de San José is located 2.6 km north of the northwest corner of Lake Petén Itzá. Although Motul has long been thought of as an important Classic Maya center (e.g., Maler 1910; Marcus 1976), no formal investigations were conducted at the site until relatively recently. In 1998, Antonia E. Foias of Williams College and Kitty F. Emery of the University of Florida initiated the Motul de San José (MSJ) Archaeological Project. Since its inception, the principal goal of the MSJ Project has been to test models for the political and economic organization of Classic Maya states. Foias (1998, 2003) organizes these models into two opposing macromodels: the “centralist” model and the “decentralist” model. The “centralist” model (e.g., Chase and Chase 1996; Culbert 1988; Marcus 1976, 1983, 1993) posits that Classic Maya states were centralized and that economic systems were administered by ruling elites (Foias 1998). The “decentralist” model (e.g., Ball and Taschek 1991; Demarest 1992; Freidel 1986; Sabloff 1986), in contrast, argues that Classic Maya states were small in scale, with limited elite administration of economic systems, and integrated primarily through religion and public ritual (Foias 1998). Motul de San José is best known for its identification as the Late Classic Ik’ site (Marcus 1976; Reents-Budet 1994; Reents-Budet and Bishop 1989) and a secondary focus of the MSJ Project has been to confirm this identification and to explore its associated polity. The Ik’ site, named for the distinctive Ik’ glyph in its emblem, is best known from a series of inscriptions on monuments at Yaxchilan, Dos Pilas, Machaquilá, and Seibal. Texts at those sites recount an intriguing, but fragmentary history dating

60

primarily to the second half of the Late Classic period and the early part of the Terminal Classic (see Martin and Grube 2000).1 Although the monuments of Motul are poorly preserved, several partially preserved texts provide clues to the history of the Ik’ polity. On the back of Stela 1, a lengthy hieroglyphic text records the accession of an Ik’ k’uhul ajaw in 701 A.D. (9.13.9.1.17), and describes him as a yajaw, or underlord, to Hasaw Chan K’awiil I2 of Tikal. This same monument records a period ending event 10 years later at a place referred to as the ja-ik’-aj. David Stuart (personal communication, 1999, in Foias n.d.) suggests that this is almost certainly the site of Motul de San José. A considerable amount of information on the Ik’ site has also come from a set of stylistically related painted polychrome vessels commonly referred to as the “Ik’ corpus” (Reents-Budet 1994). These vessels are linked stylistically, through use of pink hues for glyphic texts, their discussion of a limited number of historical individuals, including “Lord Completion-Star” and the “Fat Cacique,” and by the presence of the Ik’ emblem. Approximately 50 Ik’ vessels have been identified to date, although the vast majority of these are in museums and private collections and lack cultural provenience. Those that have been found in archaeological investigations demonstrate a fairly wide distribution, including Tayasal (Chase 1983), Altar de Sacrificios (Adams 1971; Smith 1972) and Aguateca (Eberl 2003). Early neutron activation analyzes by Reents-Budet and Bishop (1989) provided the first secure evidence that most of the Ik’ corpus pottery was produced in the Motul de San José area. More recent neutron activation analyses of 1

The late Late Classic to Terminal Classic interval during which Motul and its associated polity reached their apogee is referred to as the Sik'u' II–Yaljob'ach interval in other parts of this study. Sik'u' II and Yaljob'ach are the phase names for the late Late Classic and Terminal Classic at Trinidad de Nosotros. 2 For consistency, the spellings of the names of Classic Maya rulers follow those published in Martin and Grube (2000).

61

ceramics and figurines from Motul have confirmed and amplified this conclusion (Reents-Budet et al. 2006). Since 1998, the MSJ Project has conducted a total of eight seasons of field research. From 1998 to 2001, investigations focused on the Motul site center. An area of approximately 2.0 km2 was surveyed and mapped (Deter-Wolf et al. 1999; Foias et al. 1998; Glaab et al. 2001; Morales et al. 2000), and most of the site was systematically test pitted (Foias 2003; Moriarty 2004). To examine architectural styles and the organization of residential space, three elite residential groups and a portion of site’s central acropolis were excavated (Castellanos 2000; Foias et al. 1999; Halperin and Deckard 2001; Moriarty 2000). The major focus of research, however, was to identify residential middens. Between 1998 and 2001, a total of 35 residential groups were systematically tested and approximately 26 middens were identified and excavated (Foias 2003). Later investigations by Christina Halperin in 2005 identified several additional middens and expanded previous midden excavations (Halperin 2006). Beginning in 2000, MSJ Project investigations expanded into Motul’s periphery. Between 2000 and 2001, three extra-site survey transects were cut to the north, south, and east of the site center (Moriarty et al. 2000; Moriarty et al. 2001). Additional reconnaissance investigations explored areas to the east, southeast, and northwest of the site center (Moriarty and Wyatt 2001). In 2002, regional investigations were expanded into the further periphery with preliminary investigations at Akté, located some 7.1 km northwest of the Motul site center (Moriarty 2002; Yorgey 2005). In 2003 and 2005, intensive investigations focused on the sites of Trinidad de Nosotros (see Chapter 4;

62

Moriarty et al. 2003; Foias et al. 2006) and Buenavista-Nuevo San José (Castellanos 2006). Finally, in 2008 Ellen Spensley directed investigations at the small site of Xilil (E.Moriarty 2012). Although the final results of MSJ Project investigations are some ways off, the first phase of research has provided fairly rich culture-historical detail (see also Foias and Emery 2012). Survey and mapping between 1998 and 2001 revealed Motul to be a small- to medium-sized Classic Maya ceremonial center, comparable in size and complexity to the “Level 9: Regional Ceremonial Center” type in the site classification scheme developed by Hammond (1975) for northern Belize. The mapped portion of Motul covers an area of approximately 1.44 km2 and includes more the 230 structures (Moriarty 2004). The core zone of Motul, covering an area of approximately 0.4 km2, includes a large residential acropolis, several large temples, a 200 m long avenue or via, and seven carved monuments. The total area actually covered by Motul, however, is probably significantly larger. Survey transects to the north, south, and east of the site center encountered contiguous settlement well beyond the mapped portion of the site, and Motul’s settlement area can be reasonably estimated at around 4.0 km2 with more than 300 structures (Moriarty 2004b). This estimate of Motul’s size places the site within the uppermost tier of settlement in the Lake Petén Itzá basin, and only Nixtún-Ch’ich’ and Tayasal, neither of which have been completely mapped, appear to be somewhat larger or more complex (see Chase 1983; D.Rice 1996). Peripheral investigations have revealed Motul to be situated within a dense network of smaller satellite centers. Although regional coverage is incomplete, survey

63

transects and informal reconnaissance investigations by MSJ Project personnel and others (e.g., D.Rice et al. 1996) have identified nine small tertiary centers and at least two larger secondary centers within a 5.0 km radius of the Motul site center (Moriarty 2004). The two secondary centers, Trinidad de Nosotros and Chächäklu’um, are located 2.6 km southeast and 5.0 km east of the Motul site center. Both of these centers include more than 100 structures and exhibit complex public architectural arrangements nearly as large in scale as those found at Motul. Smaller tertiary centers, defined by settlement areas of less than 10 ha, structure counts lower than 100, and a limited investment in public architecture, have been identified in virtually all of the zones peripheral to Motul investigated to date. Investigations, ranging from intensive excavations and formal mapping to preliminary reconnaissance and surface collection, have been conducted at six of these sites: La Estrella, Buenavista-Nuevo San José, Ox Ool, Chäkokot, K’änte’t’u’ul, Xilil, and Tikalito (e.g., Castellanos 2007; Lawton 2007b; Moriarty 2004b; Moriarty et al. 2001; Moriarty and Wyatt 2001). Ceramic analyses have revealed that Motul de San José and its satellite centers have somewhat varying occupational histories. Although the earliest occupation at Motul dates to the late Middle Preclassic, and all major periods are represented in the site’s chronology, the vast majority of ceramic materials recovered pertain to the Tepeu 2 ceramic sphere and virtually all public and residential construction at the site can be dated to the latter part of the Late Classic period (see Foias 2003). Earlier Tepeu 1 materials have been found in only two deposits (Foias, personal communication, 2006), and the later Terminal Classic occupation of the site seems to have been limited primarily to a

64

elite residential groups in the site center. The single component nature of this occupation seems to suggest that Motul’s layout and organization may have resulted from a single coherent plan affected within the site’s limited 130–150 year fluorescence. In contrast, investigations at Motul’s satellite centers have recovered evidence for longer and more accretional occupational histories. Analysis of ceramic materials from Buenavista-Nuevo San José (Castellanos 2007), Chächäklu’um (Spensley 2007b), Chäkokot (Moriarty 2004b), Xilil (E.Moriarty 2012), and Trinidad (discussed in later chapters) all documented seemingly continuous occupations beginning in the Preclassic and extending into the Postclassic. At most of these sites, the Late Preclassic and Postclassic occupations appear to have been extensive and, in several cases, included monumental architectural assemblages. Surface collections at Ox Ool and La Estrella appeared to show similar patterns (Lawton 2007b).

CONCLUSIONS Situated on the edge of the Lake Petén Itzá basin, the Motul de San José area constitutes an interesting venue for examining the interplay of politics and economics within a Late Classic Maya polity. The long and accretional history of rural sites in the Motul de San José area highlights the area's ties and continuities with the long occupied Lake Petén Itzá basin. The development of Motul de San José, however, presents an opportunity to assess how this area may have changed with the emergence of a new political power. Motul's ascendancy during the latter half of the Late Classic period appears, on the basis of epigraphic and archaeological data, to have been rapid and likely

65

involved a reconfiguring of the area's sociopolitical and economic structures. Assessing the nature of these changes sheds light on the nature of political power in the Late Classic Maya lowlands.

66

CHAPTER FOUR RESEARCH DESIGN

INTRODUCTION This chapter provides an overview of investigations at Trinidad de Nosotros. The first section reviews the history of research at Trinidad prior to the present study. As Trinidad was identified originally as a port on the basis of its location, the second section below describes the geographic and cultural features that led to this interpretation. This is followed by a brief review of the three functional site models tested at Trinidad, with the expectations for each model. The fourth section describes the two seasons of research at Trinidad directed by the author, as well as an overview of affiliated investigations. The final section provides a summary discussion of excavation methods and cultural units at Trinidad de Nosotros, as well as a short overview of the chronological phases defined for Trinidad.

PREVIOUS RESEARCH AT TRINIDAD DE NOSOTROS Although the research described here constitutes the first formal investigation of Trinidad, various archaeologists visited the site prior to the start of Motul de San José Archaeological Project investigations in 2001. As a result of its small size, the absence of carved monuments, and relatively light Postclassic occupation, however, these visits

Figure 4.1. Map showing the locations for Trinidad de Nosotros, Motul de San José, and other archaeological sites in the Lake Petén Itzá Basin. Map coverage indicated on inset map of Maya lowlands. Map by Ellen Moriarty (2012). 67

68

were brief and provided only sketchy details of the site’s size and extent. The motivation for several of these visits, however, provides an important context for understanding both the site’s location and its presumed role in the local economy. Trinidad is situated in an area of modest modern settlement (Figure 4.1), close to current routes of transportation, but it was not reported archaeologically until Arlen Chase (1983) described an earlier visit by Stanley Loten and members of the University Museum Tayasal Project to the sites of Trinidad and Chächäklu’um. Chase or Loten conflated the two sites, however, listing “Chachaclun” and “La Trinidad” as alternate names for the same settlement (Chase 1983:1168). The two sites are more than three kilometers apart, however, and situated in distinct ecological niches. Further, both toponyms are widely known and appear to have some antiquity in the San José area. It appears likely that Loten may have passed through Trinidad en route to the savanna settlement of Chächäklu’um and conflated the two in his notes. The map presented in Chase’s (1983:1187) dissertation clearly depicts Chächäklu’um and most of the accompanying description is similar to that reported by Cowgill (1963:58-59) and others (Sánchez Polo 1996) for the savanna settlement. Likewise, the listed description for Chächäklu’um’s location, “about five kilometers northeast of the modern village of San José,” closely accords with that of Playa Chächäklu'um, the beach which gives access to the trail leading up the escarpment to the site and savanna of Chächäklu’um. The description of the beach, including that of a mound “being cut into by the water of Lake Petén,” however, is slightly problematic in that no site has been identified at Playa Chächäklu’um (Chase 1983:1168). This description, however, accords well with

69

Trinidad’s Platform GG, one of the principal features in the site’s harbor and one that has been deeply impacted by the lake. As local informants would be unlikely to confuse Trinidad and Chächäklu’um, it appears likely that either Trinidad was visited as part of the trip to Chächäklu’um or Loten landed at Trinidad and traveled overland from there to Chächäklu’um. Trinidad was also likely visited earlier by George Cowgill (1963). During reconnaissance along the north shore of the lake, Cowgill (1963:59) noted a location, “several kilometers west of Playa Chachaclum,” where a small stream flows into the lake and where “the escarpment is decidedly less steep than usual for the north shore.” Cowgill (1963:59–60) noted a corn field at this location, “extremely rich in sherds,” with most dating to the Late Classic period or earlier. Although Cowgill referred to the site as Playa Buenavista, he was almost certainly referring to Trinidad. His description of the site’s location exactly corresponds to that of Trinidad, and his comments regarding the lesser steepness of the escarpment and the small stream flowing into the lake are also well matched to Trinidad. Finally, his note regarding the frequency of Late Classic and earlier materials, with only a small Postclassic component also appears to accurately reflect the settlement history of Trinidad (see below). Trinidad may also have been visited much earlier by Teobert Maler and other early explorers in the Central Petén Lakes region. En route to the site of Motul de San José, Maler (1910:132) reported traveling three kilometers east of the town of San José by canoe before turning inland via a "milpero path inland." Although Maler provides no description of the landing where he transferred from canoe, it was almost certainly that of

70

Trinidad. Trinidad is located almost exactly three kilometers northeast of San José and, prior to construction of the lake road, was most easily reached via canoe. Likewise, the path Maler reports using is almost certainly the pathway referred to locally until quite recently as the “Camino Viejo.” This pathway, discussed below, starts at Trinidad and heads north, passing within a few hundred meters of Motul’s Main Plaza. Morley and other early visitors to Motul de San José likely passed along this same route. Beginning in the 1990s, as reconnaissance around the lake increased, Trinidad became the subject of increased interest. In the early 1990s, several student archaeologists from the Centro Universitario del Petén (CUDEP) visited the site and made a brief reconnaissance (Paulino Morales, personal communication, 2003). Later, Rómulo Sánchez Polo of the Proyecto Maya-Colonial conducted a brief reconnaissance at Trinidad as part of a search for Postclassic sites along the north shore of Lake Petén Itzá (Sánchez Polo 1996). Although Sánchez felt that Trinidad was probably not a major Postclassic center, he noted it as one of the major centers on the north shore of the lake (Sánchez Polo 1996:152). Motul de San José Archaeological Project investigations at Trinidad began in 2001. Before the start of the fourth field season at Motul, Moriarty and Wyatt (2001) visited Trinidad to follow up on reports of the site from Fredy Ramirez, who had accompanied Sánchez in his 1995 visit. Moriarty and Wyatt recorded GPS positions for the site’s principal structures and collected information on the site’s local history. Local informants designated the site “La Trinidad” or “La Trinidad de Nosotros,” but noted that

Figure 4.2. Map of the Motul de San José Area. Map by Ellen Moriarty (2012). 71

72

the site’s Itzaj Maya name was Sik’u’.1 During the subsequent field season, the transect survey crew spent a day collecting surface ceramics, and Ellen Spensley and Fredy Ramirez prepared a tape-and-compass map of the site (Moriarty et al. 2001). Survey and mapping revealed Trinidad to be a complex settlement and a major secondary center within the orbit of Motul. Analysis of ceramics from surface collections pointed towards major Late Preclassic and Late Classic occupations.

IDENTIFYING A PORT: GEOGRAPHY AND GEOPOLITICS AT TRINIDAD DE NOSOTROS Trinidad was initially identified as a port on the basis of its physical setting and strategic location (Figure 4.2). Ports, as noted in Chapter 2, are defined by geography. A port’s physical setting provides access to multiple modes of transportation and its geographic position determines its viability to routes of trade and transportation. This section reviews the geographical evidence that led to Trinidad’s initial identification as a port. Evidence includes both local geography and geology, and the hypothesized routes of local and long-distance trade in the Central Petén lakes region. Further, as a port’s operation is largely determined through interaction with its trade hinterland, this section 1

The derivation of the name Sik’u’ is not well understood. Local informants provided several possible translations, including “place of water grass” and “nesting place for the white bird,” but none are well supported by linguistic research. In the dictionary of the Itzaj Maya language prepared by Hofling and Tesucún (1997) the only suggested translation for Sik’u’ is “La Trinidad.” The root sik’ is most commonly translated as “anger,” although the morpheme -u never appears in the final position in modern Itzaj Maya. It appears likely that the term Sik’u’ is a truncated version of a larger morpheme. There is some possibility that the name Sik’u’ is an ancient toponym. David Stuart’s (1985) well known reading of the Yaxhá emblem glyph as Yax-ha, the modern name of the lake adjacent to the Classic period archaeological site, provides an important example of continuity in the use of toponyms within the central Petén from the Classic period to the present. More locally, several authors have suggested that the first part of Motul de San José’s name may be derived from Mutul, the Classic period name of Tikal (Martin and Grube 2000:30; Stuart n.d. in Harrison 1999:30). Although no satisfactory translation of Sik’u’ can be provided at present, it is clear that this name along with many other modern toponyms in the San José-San Andrés area warrant further research.

73

Figure 4.3. Beach approximately 100 m east of Trinidad. Such narrow, rocky beaches are common along the north shore of Lake Petén Itzá.

also assesses the political and economic context of Trinidad’s location during the Late to Terminal Classic period. Though Trinidad’s role as a port likely played out over a much lengthier interval, the late Late Classic emergence of Motul as a local political power is expected to have had a profound impact on Trinidad’s trade hinterland and, as a consequence, Trinidad’s overall operation as a port. Physical Setting Along most of the northern shore of Lake Petén Itzá, the adjoining uplands shelve steeply downward to the lake, leaving only narrow, rocky beaches (Figure 4.3). This pattern characterizes most of the Central Petén lakes and occurs as a result of the lakes'

Figure 4.4. The Trinidad harbor, beach, and inlet. Photo taken from northwest to southeast. Modern fencing runs along the waterline of the beach. The ceiba tree rises from atop Platform EE. The islet/Platform GG is covered in dense vegetation in the upper center of the photo. Note the small launch under the ceiba tree and two people atop Platform EE for scale. 74

75

Figure 4.5. The Trinidad lagoon and beach. Photo from atop Platform EE, looking west. Most of the beach area, visible just beyond the fence posts across the lagoon, is now covered by modern fill and a restaurant.

fault-related formation processes (D.Rice 1993). At Trinidad, however, the outflow of a seasonal drainage created a wide, sandy beach backed by a gently sloping silt bed (Figures 4.4 and 4.5).2 This situation is further enhanced by the presence of a small, uplifted block of limestone bedrock just offshore from Trinidad. This small rock formation or island has the effect of creating a small inlet or lagoon, referred to as an

2

Both the beach and the drainage at Trinidad have been extensively modified in recent times. The drainage was largely sealed off by construction of the lakeshore road between San José and San Pedro, although the drainage's original path floods during particularly heavy rains. The beach at Trinidad was destroyed much more recently. Beginning in 2005, the landowners deposited 2–3 m of fill atop the beach, prior to construction of a restaurant and possible hotel.

76

ixpet-ja' in Itzaj Maya, protected from winds and waves arriving from the east and south (Figure 4.5). In technical terms, the combination of a natural inlet and a partial barrier island, which provides some protection from open waters, is usually referred to in maritime environments as a "coastal lagoon" (Marriner et al. 2010; NGIA 2011; Weigend 1958). Such locations are among the most common sites for the development of small-scale ports because they can be converted into harbors with only minor architectural investment (Marriner and Morhange 2007; Marriner et al. 2006; Marriner et al. 2010; Rougé 1966). Many of the ancient harbors of the Eastern Mediterranean, for example, started off as modified coastal lagoons during the early Iron Age, with more complex offshore constructions appearing only with the spread of the Roman empire and the development of hydraulic concrete (Marriner et al. 2010:22; Rougé 1966). The situation along the shoreline at Trinidad is further enhanced by the relatively gradual slope leading from the lagoon up to the summit of the ridge upon which the site center is located. Although this slope rises almost 40 m over a distance of only about 150 m, for an average angle of approximately 15°, many of the slopes to the east and west of Trinidad rise much more steeply from the shoreline. Cumulatively, these characteristics make Trinidad an ideal setting for landing canoes and small boats and moving inland. Local informants report that these traits led to the site's use as a transshipment point during the mid-20th century Petén chicle boom. Cargoes of chicle were shipped south from the northern Petén to Trinidad where they

77

were loaded onto barges for shipment to Flores and Santa Elena. Local informants report that barges were loaded primarily in the beach area directly adjacent to the lagoon. At the local level, the need for the protected harbor provided by Trinidad's lagoon is also clearly evident. Although the waters of Lake Petén Itzá are frequently placid, producing ideal conditions for small boats and launches, regular afternoon rainstorms are normally accompanied by strong winds. During even mild seasonal rains, these winds can produce waves sufficient to swamp small boats or canoes. Local informants report that boat travel under these conditions is dangerous, particularly for vessels heavily encumbered by goods or personnel. Informants point to a tragic accident in the 1980s when a launch returning day workers from Flores to San Andrés sank in modest waves, resulting in multiple deaths. MSJ Project personnel narrowly avoided a similar accident in 2001 when a launch overloaded with project equipment and supplies began to take on water during an afternoon rainstorm. Such conditions are exacerbated when more serious tropical depressions move through the area. Local informants report that beaches along the northern coast of the lake, from the town of San Pedro west to San José, receive particularly strong waves and heavy winds during such storms. The damage from these storms is often evident in shoreline camps and other areas all along this section of coast. Most docks and beach facilities, including those constructed of reinforced concrete, often need to be repaired or re-built following the largest storms. During the pre-Columbian era, the need for a protected harbor along the northern shore of the lake would have been even more acute. As noted in Chapter 2, the canoes

78

utilized by the ancient Maya were likely highly variable in size (Hammond 1981; Jones 1998; McKillop 2010; Thompson 1949; Tozzer 1941), with perhaps many of sufficiently small size to beach and unload along the narrow, rocky beaches that line Lake Petén Itzá's northern shore (Figure 4.3). For canoes of moderate or large size, such as those described by Avendaño during his contact period visit to the Itzá (Avendaño y Loyola 1987; Jones 1998:215), however, such beaches would have been inadequate and almost certainly would have required specialized facilities of some kind. Further, such facilities would have been particularly important if, as Hassig (1985:215) has documented for Central Mexico, canoes were utilized in the transportation of cargoes weighing hundreds or even thousands of kilograms. Geographic Position Just as Trinidad's physical setting provides it with a natural harbor, the landscape and geography directly north of the lake, beyond the immediate shoreline (see above; Figure 4.6), make Trinidad the logical point of transshipment for overland travel to the north. Along most of the northern shore, the coast is backed by a steep escarpment and secondary ridges that slow or impede inland movement. Access to the site of Chächäklu’um, located three kilometers northeast of Trinidad, provides a typical example. Although the site is located only a kilometer north of the lake, it is situated at an elevation approximately 150 m higher than that of the lake. Entering the site along the most direct route from the lake involves a relatively steep climb up the northern edge of the escarpment and a secondary ridge just below the level of the site.

Figure 4.6. Colorized digital elevation model (DEM) for the Lake Petén Itzá basin with major sites, drainages, and selected modern towns indicated. Note the slightly lower (yellow to red) area at Trinidad in comparison to higher (purple) areas to the east and west. The light yellow band running inland from just east of Trinidad is the Río Chäkokot. Image courtesy of the MSJ Project and the NASA/JPL AirSAR Program. 79

80

Trinidad, by comparison, is situated at the exact point at which the east-west trending escarpment and hills lining the lake's northern shore turn to the southwest. Trinidad is actually positioned at the exact hinge of this local topography and, in contrast to areas to the east and southwest, is not separated by steep ascents from the interior uplands. Moving inland from Trinidad is much easier than elsewhere on the lake's northern shore. This situation is graphically illustrated by the colorized digital elevation model of the Lake Petén Itzá Basin presented in Figure 4.6 (see also Figures 4.1 and 4.2). The advantages of Trinidad's position with respect to the hills backing Lake Petén Itzá's northern shore is subtle, but clearly visible from a canoe or launch on the lake (Figure 4.7). It is also well known within the community of San José. Trinidad is best

Figure 4.7. Trinidad's location seen from a launch on Lake Petén Itzá. Looking northeast from southwest. The town visible above the bow is San José. The other town, just visible on the left edge of the photo is San Andrés.

81

known locally as the southern terminus of an important trail running north from the lake. This trail (the "Camino Viejo") had some importance during the Petén chicle boom and, prior to the construction of a modern road just east of Trinidad, was the primary means of communication with the modern community of Corozal. The trail, which passes within a few hundred meters of Motul's Main Plaza, was also almost certainly the same pathway utilized by Maler (1910:132) to visit Motul in 1895. 3 Use of this trail almost certainly has some antiquity as it passes just beside the monumental heart of Motul and would have provided the most rapid access between Motul and the lake. Further, there is some possibility that a part of this trail might have been formalized by a causeway. Local informants have reported various features intermediate between Trinidad and Motul that may constitute the remains of a causeway. Informal reconnaissance by the author and other MSJ Project personnel in 2006, however, failed to locate these features. As much of the area between the two sites is covered in dense brush (guamil) and secondary growth, formal survey would be required to identify any possible features. Local and Long-Distance Trade Routes Trinidad's position may have also made it an important node in both local and long-distance trade routes. Contact period sources identify the eastern and western ports for Lake Petén Itzá as Chaltunha’ and Ch’ich’. Don Rice (1996) and Grant Jones (1998:499n55) identify Ch’ich’ as the archaeological site of Nixtún-Ch’ich’, a major center on the Candelaria Peninsula at the west end of the lake, and Chaltunha’ as a “string 3

The “Camino Viejo” continued to be used up until 2003, when an oil survey transect was cut through the same area and replaced the older pathway. According to local informants, most of the original path is now overgrown.

82

of settlements in the vicinity of the archaeological site of Ixlú." Possible port facilities and hydraulic features have been identified at both of these settlements suggesting that their roles as gateways to the lake were significant enough to require special facilities (D.Rice 1996). The quantity of trade goods passing through these two centers was likely considerable. As noted in Chapter 3, the Lake Petén Itzá basin is bounded by a large number of moderate to large centers with long histories of occupation (Figure 4.1). Contact period accounts describe the lake as a major population center, with extensive trade and interaction between different parts of the lake and external zones (Avendaño y Loyola 1987; Jones 1998; Thompson 1951). Given the great scale and number of Late Preclassic and Late Classic centers around the lake, earlier interactions were likely even more extensive than those described for the Contact period. Further, Lake Petén Itzá's central position in the Maya lowlands, as well as the greater speed and carrying capacity of canoe transportation versus porters (Chapter 2; Adams 1978; Hassig 1985; Mitchum 1994; Santone 1997) suggest that a high volume of long-distance trade goods passed through the lake, both for local consumption and beyond (see also D.Rice 1996). Nixtún-Ch'ich's position on the west end of the lake make it the logical entry point for highland goods, including obsidian, arriving via Alta Verapaz and then overland from the Río Pasión. Ixlú, in contrast, was well positioned to receive maritime imports arriving from the coast as well as highland imports arriving via the "Motagua-Caribe" route (Adams 1978; Arnauld 1990; Hammond 1972, 1976, 1978; P.Rice 1984).

83

Although Trinidad's position at the elbow of the lake's northern arm is likely less strategic than Ixlú or Nixtún-Ch'ich' in terms of controlling goods arriving in the lake basin, its central position likely it gave it local importance. For long-distance trade moving across the lake, for example, the ca. 28 km canoe journey from the east to west end of the lake, or vice versa, is likely to have required 6–9 hours or longer, depending on the cargo, number of paddlers, and other factors (Adams 1978:27–28; Hassig 1985; Santone 1997). Trinidad's location near the midpoint of this journey, as well as its excellent landing area, would have made it an ideal way-station for resting, swapping paddlers, or waiting out seasonal storms. Further, the centrality of Trinidad's position would have made it an excellent location for periodic markets or other modes of interaction for populations from both ends of the lake (Dahlin et al. 2007). Trinidad could also have served as a nexus between the Lake Petén Itzá basin and points north and northwest. The pathway running north from Trinidad makes it the logical northern egress from the lake basin to sites to the north. Further, directly north of Trinidad, the Río K'änte't'u'ul reaches within approximately 4 km of Lake Petén Itzá (Figure 4.6). Although currently only a minor seasonal stream draining the uplands north of the lake, local informants report that up until approximately two decades ago, the Río K'änte't'u'ul had a significantly greater flow than today and was navigable by canoe during part of the year. Informants attribute the decline in flow volume to local deforestation and an accompanying decline in yearly rainfall. If the Río K'änte't'u'ul had a similarly large volume during the pre-Columbian era, it could have provided a waterborne link to the northwest. Approximately 7 km

84

northwest of Trinidad, the Río K'änte't'u'ul joins with the Río Ixconob to form the Río Akte (Figure 4.2). The Río Akte continues north and then west, passing the site of Akte, before disappearing into seasonal bajos to the northwest. Although the route of the Río Akte beyond the bajo is uncertain, hydrographic maps produced by Guatemala's Instituto Geográfico Nacional (2001) depict it reaching within a few kilometers of a tributary leading into the Río San Pedro, one of the major waterways of the Maya lowlands. Such a transportation connection, if it existed, would have greatly increased the strategic significance Trinidad and the Motul de San José area as a whole. Gateway to the Motul de San José Area The final variable in assessing Trinidad's location is its position with respect to the Motul de San José area (Figure 4.2). Although this area is by no means inaccessible by other approaches, the easiest access is clearly achieved via Trinidad de Nosotros. Trinidad's position at the natural "entrance" to this zone, its location along potentially important trade routes (see above), and its position at the edge of a geographically delineated hinterland all suggest that Trinidad meets the traditional definition of a gateway community (Burghardt 1971:269–270; Hirth 1978:37–39). The emergence of an important economic center, or central place, within the hinterland of a gateway community is a classical problem in economic geography (Burghardt 1971), with relatively clear expectations, depending on the size and organization of the gateway hinterland. In situations with relatively restricted hinterlands, like the Motul de San José area, the emergence of a central place is expected

85

to lead to the gateway being "shorn" of much of its former economic significance, while still retaining its transport functions (Burghardt 1971:272). Assessing the extent to which this occurred at Trinidad during the Late to Terminal Classic period ascendancy of the Motul polity is the major proposition of this study. Trinidad's position with respect to Motul and the greater Motul de San José area is, in many ways, ideal for determining whether or not the political capital of Motul was, in fact, an economic central place. If it was, then its ascent as a central place should have been accompanied by a corresponding decline in Trinidad's role in the local economy. In contrast, if Motul played only a limited economic role, then Trinidad's function as a gateway should have remained constant or even been enhanced as a result of the increasing scale of settlement, production, and consumption within its hinterland.

RESEARCH DESIGN: SITE MODELS AND EXPECTATIONS Investigations at Trinidad de Nosotros were designed to assess its economic role in the critical Late to Terminal Classic interval when Motul de San José emerged as a major political capital. These investigations were designed, in part, to test three functional site models derived from previous research on maritime ports and coastal sites in the Maya area. The general outlines of these models, as well as their expected archaeological correlates, are presented below. Further, this section provides a review of the specific approaches and expectations for assessing Trinidad's harbor and the critical artifact class of obsidian. This sections concludes with a brief review of related culturehistorical research at Trinidad.

86

Site Models: Coastal Sites, Transshipment Ports, and Trading Ports As noted in Chapter 2, various port models have been defined for the Maya lowlands (e.g., Andrews 1990, 2008; Andrews et al. 1988; Guderjan et al. 1988; Hammond 1972, 1976; McKillop 1996, 2004; Rathje and Sabloff 1973). This study tested three site models loosely based on those defined by Andrews (1990) and applied elsewhere in the Maya lowlands (McKillop 1996, 2004; Guderjan et al. 1988; Mock 1997). The least complex model for Trinidad is that of the simple coastal site. Coastal sites, simply by virtue of their location, adjacent to bodies of water, are expected to have had access to a greater variety of resources than their inland neighbors. Most ancient Maya coastal sites exchanged coastal resources in a steady, low-volume manner with nearby inland sites (Andrews 1990:162; Graham and Pendergast 1989; Mock 1997). In some instances, particularly where salt production facilities have been identified (Andrews 1983, 1984; McKillop 2002), this trade may have occurred at a much larger scale. Overall, however, simple coastal sites are not expected to have played major roles in long-distance trade or transportation. The second model tested at Trinidad is that of the transshipment port. At their most basic level, as noted in Chapter 2, ports are locations utilized to effect the transfer of goods or personnel between modes of transportation (Alexandersson and Norstrôm 1963; Hayuth 1962; Todd 1993:3; Weigend 1958:185). Transshipment ports are important for their advantageous physical locations. Transshipment ports are concerned primarily with the provision of transportation related services, serving as transfer points or way stations

87

along long-distance trade routes (Todd 1993:3). Although, by virtue of their positions, ancient Maya transshipment ports may have had privileged access to trade goods, they are not expected to have been major points of distribution for those goods (Andrews 1990; Guderjan 1995a, 1995b; Guderjan et al. 1988; Hammond 1972, 1976; McKillop 1996). The third model tested at Trinidad is that of the trading port. As with transshipment ports, trading ports are situated at important nodes in local and longdistance trade networks and provide a suite of transportation related services (Andrews 1990; McKillop 1996). Trading ports differ, however, from transshipment port with respect to their role in trade. In contrast to transshipment ports, trading ports are expected to be much more heavily involved in the distribution of local and long-distance trade goods (Andrews 1990; Andrews et al. 1988; McKillop 1996, 2004). Specifically, trading ports should be seen as identical to gateway centers in serving as local distribution centers for trade goods (Burghardt 1971; Hirth 1978). Although a fourth model, the port-of-trade model, has also been applied in the Maya lowlands (see Chapter 2; Chapman 1957; Rathje and Sabloff 1973), this particular model was not tested in this study. Andrews (1990:163–166) notes that the port-of-trade model has proven exceptionally difficult to identify archaeologically in previous tests in the Maya lowlands. Further, the absence of textual data relating to Trinidad, as well as the close proximity of several major political centers, including Motul, make this model difficult to apply to Trinidad.

88

Field and Laboratory Objectives Field and laboratory investigations at Trinidad were guided by three overarching objectives directly or indirectly related to the testing of the three site models. The first of these goals was the identification and assessment of Trinidad's possible harbor. A wealth of cross-cultural research demonstrates that port facilities constitute a capital investment proportionate to the expected needs of the port (Hohlfelder 2000; Kim and Sachish 1986; Mayerson 1996; Mitter 1986; Robinson 1976; Weigend 1958). If Trinidad functioned as a major port of any kind, then it can be expected to have invested in port facilities. The size and extent of these facilities can be seen as reflecting the volume of goods and personnel passing through the port. Thus, the first object of field investigations was to explore Trinidad's harbor. The second objective of investigations at Trinidad was to create a robust sample of obsidian for inter-site comparisons. As the most identifiable and widely available exotic resource in artifact collections, obsidian serves as a useful proxy for the wide array of perishable, low volume, or low visibility trade goods that also may have passed through Trinidad. To provide a sample of obsidian reflecting all levels of society at Trinidad and comparable to collections from other sites in the Motul de San José area, investigations in Trinidad's site center focused on the identification and excavation of middens associated with Late to Terminal Classic residences. Middens, or secondary refuse aggregates, provide reliable indicators of household consumption and economic activity (Johnston and Gonlin 1998; Wilson 1994).

89

Analysis of obsidian focused on simple indices previously utilized in the Maya area and easily applied to inter-site comparisons: average blade widths, cutting-edge-tomass ratios, and the distribution of cores (McKillop 1989, 1996; Rovner 1976; P.Rice 1984; Sheets 1978; Sheets and Muto 1972; Sidrys 1976, 1979). Likewise, as previous research has shown that ports typically have access to a broader range of obsidian sources than other sites (Andrews et al. 1989; Guderjan et al. 1989; McKillop 1989, 1996), a large sample of obsidian from six sites in the Motul de San José area was sourced by Xray fluorescence. Although midden densities and the ratio of obsidian to other potential cutting materials have also proven useful for inter-site comparisons (McKillop 1989, 1996; P.Rice 1984), limitation in the regional sample for the Motul de San José made both approaches difficult to utilize. The midden sample from sites other than Trinidad and Motul was insufficient for meaningful comparisons. Further, Trinidad's position atop a major chert source made inter-site comparisons problematic. In addition to the theoretical concerns outlined above, the third research objective at Trinidad was shaped by basic culture-historical issues. These included the establishment of a site map and providing an overview of Trinidad’s occupational sequence. Assessing Trinidad's physical organization and its long-term development was critical to assessing its place in the larger Lake Petén Itzá basin and its relation to other sites in the Motul de San José area (Moriarty 2004b). These investigations also provided a secondary check of Trinidad's identification as a port. Ports, as noted in Chapter 2, are unusually stable settlement loci as a result of their strategic positioning and access to resources from multiple regions. If Trinidad

90

functioned as a port, then it should have a long history of occupation. This occupation should appear unusually stable in comparison to sites in its hinterland. Further, as the volume and nature of goods traveling through a port are heavily influenced by conditions in its trade hinterland, Trinidad's peaks in population and complexity should also correspond to those of its hinterland. During peak intervals, the total volume of goods passing through a port should be high, encouraging growth and investment in transportation and trade related facilities. Declines in hinterland population should be accompanied by a slackening emphasis on these facilities.

SUMMARY OF INVESTIGATIONS, 2003–2008 The first intensive investigations at Trinidad were initiated in 2003 under the auspices of the MSJ Project and with funding from the Foundation for the Advancement of Mesoamerican Studies, Inc. (FAMSI), the Middle American Research Institute, and Williams College. Field investigations, directed by the author, took place between May and August, and focused on preparing an accurate total station map of the site center and establishing the site’s chronology. An area of 35 ha was surveyed and mapped (Chapter 5), and a total of 29 test pits were excavated in various parts of the site (Chapter 6; Moriarty 2003a). To gain information on architectural style at Trinidad, parts of three residential groups (C, G, and FF) were cleared (Chapter 9; Halperin and Hernández Véliz 2003; Moriarty 2003b; Spensley 2003). Additional investigations included excavations in the harbor area (Chapter 7; Spensley 2003) and midden testing in two groups (C and G) (Chapter 8; Halperin and Hernández Véliz 2003; Moriarty 2003b).

91

The subsequent lab season, from August 2003 to July 2004, was focused on ceramic analysis. The major emphasis of research was on placing ceramics from Trinidad in regional chronological perspective. Comparisons were made to type collections from Uaxactún, Altar de Sacrificios, El Mirador, Nakbe, Flores, and other sites. A provisional ceramic sequence was established for Trinidad in consultation with Antonia Foias and Jeanette Castellanos. Additional research included analysis of figurines and ceramic artifacts by Christina Halperin, preliminary sorting of faunal materials by Erin Thornton and Kitty Emery, and a micromorphological study of plaster floors by Ellen Spensley. The second field season at Trinidad took place between May and August of 2005 under the co-direction of Moriarty and Crorey Lawton. During the first half of the season, the site map was expanded to the west and east, and four additional test pits were excavated to sample previously untested areas. The emphasis for the remainder of the field season was on systematic testing for middens and further explorations in the harbor area. Eleven residential groups and two public architectural complexes were tested for middens (Moriarty et al. 2007). These investigations identified several dozen middens that were explored by subsequent investigations. Spensley (2007a) directed the second season of excavations in the harbor area and focused on the area’s two principal features, Platforms EE and GG. To improve our understanding of architecture at Trinidad, Lawton (2007a) completely cleared Group O, a small residential group in the site center. Finally, Melanie Damour and Thad Bissett of Earth Search, Inc. conducted a preliminary

92

underwater investigation in the area just offshore from the Trinidad harbor (Damour and Bissett 2007). The final field lab season for materials recovered at Trinidad was conducted from September 2005 to April 2006. Research at Trinidad produced a tremendous quantity of artifacts and from the outset a decision was made to focus on materials from selected contexts. Artifacts from middens and stratified deposits were prioritized while materials from collapse or other mixed contexts were considered only after materials from other contexts had been assessed. Ceramics, the most numerous artifact class, were analyzed by Moriarty, Foias, and Castellanos. Lawton analyzed all chert artifacts, the second most numerous artifact class, as part of his study of chert tool production and use in the MSJ area. The author analyzed obsidian artifacts. Erin Thornton analyzed faunal remains and human osteological materials as part of her own dissertation. Christina Halperin analyzed all figurines and spindle whorls. Minor artifact classes, including groundstone, greenstone, pyrite, notched sherds, and others, were analyzed collaboratively by Moriarty, Lawton, Melanie Kingsley, and others. A more thorough modal analysis of the Late Classic ceramics of Trinidad was conducted by Spensley in 2008 (E.Moriarty 2012). Additional analyses were conducted by the author in the United States. Obsidian analyses, including metric measurements and XRF source analyses, were conducted from 2006 to 2008. In additional, the author also spent a week at the University of Arizona AMS facility in 2006 to start the radiocarbon dating of a dozen samples.

93

FIELD METHODS Field investigations at Trinidad de Nosotros included more than 680 excavation units organized into 21 separate operations. Cumulatively, these investigations covered an area of more than 440 m2 and involved excavation of more than 328 m3 of matrix. The following section describes excavation methods and recording procedures utilized during these investigations. This section also includes a brief discussion of the primary cultural and natural contexts encountered at Trinidad. Finally, this section concludes with a brief summary of the cultural phases defined for Trinidad. Detailed descriptions of the mapping project and four programs of excavation at Trinidad are provided in Chapters 5–9. Excavation Methods and Recording Procedures One of the fundamental goals of investigations at Trinidad was to provide data comparable to those acquired in previous research at Motul de San José. As such, excavation methods at Trinidad closely followed those utilized by the Motul de San José Archaeological Project. This included the use of an operation-lot organizational system similar to that employed by the Motul project. In this system, operation numbers designate the specific objectives of individual excavations (Table 4.1). Alphabetic suboperations separate operations into geographic subdivisions. Thus, all site center testpitting investigations were organized as Operation 1 and subdivided using the letter of the group tested. For example, test units placed in Group C were designated Operation 1C. As another example, household excavations in Group O (Operation 6) were subdivided into three separate sub-operations depending on the structure excavated (6A = Structure

94

O-1; 6B = Structure O-2; 6D = Structure O-3). Numeric unit designations are assigned in order of excavation within a particular sub-operation. Within individual excavation units, excavated matrices are divided into vertical or horizontal levels that correspond to natural or cultural strata (see below). The basic unit of excavation within this system is the lot. At Trinidad, arbitrary lots were used to divide natural or cultural levels vertically or horizontally. In the vast majority of excavations, levels were excavated as a series of arbitrary 10-cm vertical lots. Use of such small lots provides control over excavations and allows for the reinterpretation of associated levels after excavation. Although lots deeper than 10 cm were occasionally utilized, their use was limited primarily to excavations in deep, uniform strata (e.g., collapse, heavy platform fill), areas where previous excavations provided a solid working knowledge of local stratigraphy (e.g., midden excavations adjacent to earlier shovel tests), or 50-x-50-cm midden tests. The notation utilized in this system produces simple alphanumeric designations in the order described above (i.e., Operation, Sub-operation, Unit, Level, and Lot). Thus, the third lot excavated in the second level of unit two within Operation 2A would, with the addition of the Trinidad prefix (TRI), be described as TRI2A2-2-3. Further, to simplify sorting and analysis of excavation data each lot was assigned a unique numeric context ID following excavation (see below). Units were excavated using trowels and small archaeological handpicks. To aid in the collection of artifacts, all excavated matrices were screened using ¼-inch mesh. All artifacts and ecofacts larger than ¼-inch were collected by lot. With the exception of

95

Table 4.1. Operation summary for excavations at Trinidad.

Operation Goals and Location(s) 1 2 3 4

Site center test units (various groups). All harbor area investigations in 2003. Midden excavations in Group C. Architectural excavations in Group C.

5 6 7 8 9

Architectural excavations in Group G. Architectural excavations in Group O. Midden testing investigations in Group C. Midden testing investigations in Group O. Midden testing investigations near Plaza V.

10 11 12 13

Midden testing investigations in Group F. Midden testing investigations in Group U. All harbor area investigations in 2005. Midden testing investigations in Group Y.

14 15 16 17

Midden testing investigations in Group T. Midden testing investigations in Group S. Midden testing investigations in Group Q. Midden testing investigations in Group Z.

18 19 20 21

Midden testing investigations in Group K. Midden testing investigations in Group H. Midden testing investigations in Group DD. Midden testing investigations in Group KK.

tests excavated as part of the midden testing program, all classes of artifacts were sorted and bagged separately by lot. Midden testing operations differed only in that ceramics, chert, and groundstone from the same lot were bagged together in the field and separated later in the lab. All other artifact classes were bagged separately during these operations. To aid in the identification of small faunal and botanical remains, as well as to identify possible lithic production loci, unscreened soil samples were collected from all middens, caches, burials, and other primary deposits and from a small sample of all other contexts. Soil samples consisted of two large, 4-liter Tyvek bags of unscreened matrix.

96

These samples were later water-screened by Erin E. Kennedy-Thornton using nested 1/4, 1/8, and 1/16-inch mesh. Three small Whirlpacks of matrix were also collected from the same contexts. These were submitted to the Brigham Young University Department of Agronomy and Horticulture for phosphate, phytolith, and other chemical analyses by Dr. Richard Terry and others (Parnell et al. 2001; Terry et al. 2000; Wells et al. 2000). As with the excavation methods described above, recording procedures at Trinidad followed closely those utilized at Motul de San José. During field investigations archaeological staff completed a lot form for each lot as it was excavated (Appendix A). Basic contextual data recorded on these forms included unit location, the horizontal dimensions of the lot, and starting and ending elevations. Additional information included Munsell soil color, soil texture, the size and frequency of stone inclusions, an inventory of artifacts, and a detailed description of the lot’s cultural or natural context. Additional notes, observations, and more detailed summaries were recorded in field notebooks. Features encountered during excavations were normally excavated as separate levels and given unique context designations. Simple features such as floors, walls, and postholes were given general descriptive designations in order of discovery within a particular unit (e.g., Unit 1F2 Floor 2). These designations were later modified following excavation when stratigraphic associations with nearby units could be verified. More complex cultural features, including burials and caches, were numbered consecutively at Trinidad in order of discovery. Burials were also cataloged during excavation utilizing

97

burial forms (see Appendix A). Detailed summaries of caches and burials from Trinidad are provided in Appendix B. Photographs and drawings were prepared during and immediately after excavations. Photographs and drawings were prepared for all burials and other special features (e.g., caches), and at least one profile was drawn for every unit with dimensions of 1-x-1 m or larger. Finally, at least one corner of each excavation unit 1-x-1 m or larger was mapped into the total station grid during excavations and later plotted onto the site map. Upon completion of field investigations and subsequent laboratory research, summary excavation and artifact data were entered into an Excel spreadsheet. The final resulting document, the Trinidad context history, is organized by lot. Each of the 2,735 excavated lots and surface collections was assigned a unique context ID (e.g., TRI1A1-11 = Context 1000) and a row in the context history. Attached lot descriptions include the lot’s alphanumeric designation (e.g., TRI1A1-1-1), the location of the associated unit, a brief description of the cultural context, one or two numeric cultural context codes (see below), notes regarding features, special finds, or special contexts associated with the lot (i.e., burials or caches), and the lot’s vertical and horizontal dimensions. Two separate columns contain data pertinent to the lot’s ceramic phase. The first column includes an abbreviated summary of the ceramic phases present in the lot and the second provides the final phase as determined by the combination of ceramic data and the lot’s stratigraphic position.

98

Ceramic data were also utilized to provide a qualitative assessment of the lot’s purity (i.e., “pure,” “slightly mixed,” or “heavily mixed”). In general, the length and continuous nature of Trinidad’s occupation means that very few excavated lots contain materials dating to only one phase. Lots were assigned a “pure” rating only if they contained only a handful of sherds dating to phases other than that assigned to the lot. Lots consisting of up to approximately ten percent in-mixed materials were assigned a “slightly mixed” rating. Lots with more than approximately ten percent in-mixed materials were assigned a “heavily mixed” rating. Finally, summary descriptive data for all major artifact classes were appended to the context history for each lot. Counts and weights were appended for ceramics, chert, groundstone, obsidian, greenstone, figurines, and other major artifact classes, and more detailed summaries were added regarding stone tools, fauna, and other key artifact classes. Finally, other data (e.g., ceramic form data) are being appended as they become available. The resulting database provides a detailed overview of artifact distributions and densities for each lot. These data are particularly useful for assessing differences in context composition. Cultural and Natural Contexts at Trinidad Excavations at Trinidad encountered a wide variety of cultural and natural contexts. Contextual designations were assigned to individual lots from a list of 68 possible contexts utilized by the Motul de San José Archaeological Project and following definitions provided by Loten and Pendergast (1984) and others. This list, with the numeric codes utilized on lot forms and in the context history, is provided in Appendix

99

A. In summary, six basic classes of contexts were encountered at Trinidad: humus and natural or disturbed surface contexts; floors and cultural living surfaces; fill; wall fall or collapse; complex features; and middens. Humus contexts at Trinidad were extremely diverse. Within the site center, approximately 10–14 cm of soil has formed since the end of the Late Classic period, and humus deposits frequently included a mixture of materials from the subsequent Terminal Classic and Postclassic periods. Humus deposits also frequently showed some amount of natural or cultural disturbance. This was particularly true of the harbor area where the combination of a heavy Postclassic occupation and more recent historical usage has heavily impacted the upper 10–20 cm of most deposits. Although conditions in the site center were more favorable, trampling by cattle and frequent burning have both had major effects on preservation. Excavations at Trinidad encountered numerous floors dating to most of the major periods in the Maya chronology. The vast majority of floors at Trinidad were of durable plaster construction. Plaster floor thickness and composition differed significantly by phase (Spensley 2004), but most could be easily identified during excavation. Occasionally, the uppermost or ultimate floor in a unit was entirely eroded and could be identified only by the presence of small fragments of highly eroded stucco and the underlying ballast (see below). Only a few packed-earth floors were encountered at Trinidad and almost all dated to the site’s Middle Preclassic occupation (the Aj Wo’ and Ix Cha’ phases). Very few in situ above-floor deposits that could be interpreted as de

100

facto refuse (e.g., Joyce and Johannesen 1993) from an ultimate occupation were encountered at Trinidad and almost all came from excavations in Structure C-1. Four major classes of fill were encountered at Trinidad. Thin levels of densely packed gravel encountered directly beneath floors and other features were interpreted as ballast (Loten and Pendergast 1984:4). More general fill contexts were divided into three major classes based on the size and frequency of stone materials. Fill with a high soil content and only a small quantity of gravel and cobbles were interpreted as light fill. Light fill contexts were frequently shallow and associated with modifications to existing structures or platforms. Denser mixtures of gravel and cobbles were interpreted as medium or moderate fill. Moderate density fill deposits were the most common at Trinidad and most floors and ballast deposits were found overlying what appears to have been this basic unit of construction. The densest fill deposits, described as heavy fill, frequently included a mixture of gravels, cobbles, and boulders, often with only a light admixture of soil matrix. Heavy fill deposits were found primarily within the cores of large basal platforms. Several investigations identified rough retaining walls within heavy fill deposits, and these features were presumably utilized to consolidate heavy fill deposits during construction. Wall fall or collapse contexts were encountered primarily in architectural investigations where they constituted the bulk of contexts excavated. During excavation detailed notes were recorded regarding the size and shape of collapse materials, particularly facing stones or other key architectural elements (e.g., vault stones). These descriptions have been utilized to provide some of the architectural assessments made in

101

Chapter 9. Collapse contexts were also frequently encountered in off-mound midden testing units. In many instances, the upper 20 or 40 cm of midden tests consisted of a mixture of humus and collapse from the platform or buildings above. In most instances, the presence of collapse in upper levels contributed towards excellent preservation in underlying middens. Only occasionally did collapse deposits in off-mound contexts provide clues toward the architectural configuration of associated buildings. Complex features at Trinidad consisted primarily of burials and caches. Burials were defined as any intentional interment of human remains (Rathje 1970; Ricketson 1925). Although isolated human skeletal materials were occasionally recovered in fill or middens, most were interpreted as incidental, likely resulting from ancient Maya redeposition of earlier deposits. Burials were classified in terms of the positioning (i.e., flexed, extended, other), orientation of the head, the degree of articulation, and a provisional interpretation of the burial’s deposition (primary or secondary). Burial architecture was classified following the five-part classification system defined by Ruz Lhuillier (1965). The term cache was utilized, following Loten and Pendergast (1984:5), to refer to individual artifacts or discrete sets of “artifacts intentionally placed in specific locations, and unrelated to burials.” A principal assumption of this definition is that caches represent the material expression of ritual activities. Although only a few caches were recovered at Trinidad, several bear strong similarities to caches discussed by Coe (1965) and others (e.g., Chase and Chase 1998), and these correspondences are discussed in the attached appendix (Appendix B).

102

Finally, of all the classes of cultural contexts encountered at Trinidad, middens presented perhaps the greatest degree of variability. In general, the midden testing program was designed to identify secondary refuse aggregates, or high-density accumulations of refuse as defined by Schiffer (1972:163) and Wilson (1994). These middens, thought to have been produced by the purposeful deposition of refuse by the ancient Maya, are normally characterized by very high artifact densities, a high degree of artifact diversity (number of artifact classes present), and frequently, though not always, by excellent artifact preservation. At the other end of the spectrum, many investigations identified low-density refuse deposits or sheet midden. Most of these deposits were identified along the peripheries of residential groups within the circumambient area referred to as the “intermediate area” by Killion (1990, 1992) and as the “toft” by Hayden and Cannon (1983). At the intermediate level, numerous middens were found to have artifact densities somewhat greater than sheet middens, though not as dense or diverse as high-density middens. These were classified as moderate-density middens, and interpreted as the high points of accumulation within “toft” zone refuse deposits. Cultural Phases at Trinidad Excavations at Trinidad revealed a continuous record of occupation extending over an interval of approximately 2,750 years (ca. 800 BC–AD 1950). This interval was divided into 10 cultural phases corresponding to ceramic complexes defined through type-variety analysis (Table 4.2; see Chapter 10). Excavated contexts were assigned to a particular phase on the basis of their ceramic content and in relation to their stratigraphic position. Where complete type-variety data were available, most contexts could be

103

Table 4.2. Cultural phases at Trinidad de Nosotros.

Phase/Complex Chronological Period

Approximate Years (calibrated)

Aj B’oj Säk-tunich Yaljob’ach

Historical Postclassic Terminal Classic

AD 1525–1950 AD 950–1525 AD 850–950

Sik’u’ II Sik’u’ I ‘Ayim-tun

Late Late Classic Early Late Classic Early Classic

AD 700–850 AD 550–700 AD 200–550

P’ich ‘Ayim Chukan Ix Cha’ Aj Wo’

Terminal Preclassic Late Preclassic Late Middle Preclassic Early Middle Preclassic

AD 1–200 300 BC–AD 1 650–300 BC 800–650 BC

assigned securely to one of the ten named phases. In some instances, however, the nature of the ceramic and stratigraphic data permitted either a more precise date or necessitated a more limited estimation. One of the key problems encountered in phasing contexts appeared most consistently in humus or other disturbed deposits. Humus deposits at Trinidad frequently included a mix of ceramics from multiple phases or sherds too eroded to identify. Similar situations also occurred in some fill deposits, structural collapse, and low-density middens. In some instances, these contexts could be adequately phased on the basis of their stratigraphic affiliations; in others they could not and their phase was left as indeterminate. In contexts where insufficient diagnostic sherds were recovered, it was sometimes impossible to assign a single phase date. This situation occurred most commonly in Preclassic contexts. Although the Middle Preclassic Ix Cha’ and Late Preclassic Chukan complexes are easily separated when sufficient diagnostics are present, close similarities

104

in ware technology make the separation of body sherds alone impossible. In such contexts, where stratigraphic data did not clarify the situation, a multiphase designation was assigned (e.g., Ix Cha’–Chukan). Similar situations also occurred commonly in Late Classic deposits. Although the Sik’u’ I and Sik’u’ II phases can be separated easily when sufficient diagnostic polychromes are present, many deposits produced only small numbers of polychrome sherds. In such instances, and where other lines of evidence failed to clarify the situation, the deposit was assigned a general Sik’u’ phase date. In practice, this problem was particularly acute in small test units such as those utilized in the site center testing program. In larger clearing or midden excavations, sufficient numbers of diagnostic polychromes were normally available. In other instances, through a combination factors, it was possible to assign subphase chronological designations to certain contexts. Such instances occurred primarily in Chukan, Sik’u’ II, and Säk-tunich phase deposits (see Chapter 10). The latter portion of the Sik’u’ II Late Classic, for example, can be identified by a number of co-occurring types and forms with clearly defined temporal affiliations. Where these diagnostics appeared, it was possible isolate a context as dating to the latter portion of the Sik’u’ II phase. Subphase designations for Chukan and Säk-tunich phase deposits were more provisional, but were based on observed patterns in ceramic data, and supported by a limited number of radiocarbon dates. In all likelihood, further analysis of both the Chukan and Säk-tunich will permit the definition of discrete early and late complexes, corresponding to separate phases in Trinidad’s cultural sequence. Until these complexes

105

have been adequately defined, however, the early and late portions of these phases are simply described as such.

SUMMARY The dissertation that follows is the first synthesis of investigations at Trinidad de Nosotros. As the preceding should make clear, these investigations were a collaborative endeavor involving numerous investigators with different specialties and university affiliations. The following discussion incorporates data from many of these investigators in order to present a holistic perspective on the long-term development of Trinidad de Nosotros and evaluate its role within the Motul polity. As the following pages will show, Trinidad’s role was a complex one, combining elements of the material, political, and ritual economies. From the evidence collected during this research, it appears clear that Trinidad played the role of a trading port during the Late to Terminal Classic apogee of the Motul polity. Its residents constructed major harbor facilities, and these facilities underwent significant expansion and use during Motul's ascendancy. Further, the analysis of obsidian revealed that Trinidad was likely a major player in the production and distribution of prismatic blades during the Late to Terminal Classic interval, actually producing more evidence of such a role than the political capital of Motul. Finally, the long-term development of the site closely matches the expectations for a port and gateway, as discussed above.

106

CHAPTER 5 SURVEY, MAPPING, AND SITE ORGANIZATION

INTRODUCTION One of the principal goals of research at Trinidad de Nosotros was to produce a reliable site map and to assess its size and organization in regional perspective (Figure 5.1). This chapter presents an overview of survey and mapping procedures utilized at Trinidad, as well as short descriptions of the principal cultural and natural features identified in these investigations. Though small in comparison to major centers in the Lake Petén Itzá basin like Motul de San José, Tayasal, and Nixtún-Ch’ich, Trinidad hosts a wide range of features including a sophisticated site center, a harbor area, and extensive peripheral settlement. The organization of public and ceremonial architecture at Trinidad is particularly noteworthy, and this chapter includes detailed descriptions of the principal temples, the ballcourt, and the public plazas. This chapter also includes a brief description of the Trinidad harbor. To facilitate intersite comparisons, Trinidad’s residential groups are assessed using a “plaza plan” classification scheme developed for Motul de San José. To place these findings in context, the next-to-last section discusses population estimates for Trinidad. This chapter concludes with a brief synthesis of the site of Trinidad de Nosotros.

107

Figure 5.1. Map of Trinidad de Nosotros.

108

SURVEY AND MAP PREPARATION The site of Trinidad de Nosotros is situated within an area featuring modern pasture, scrub brush, and limited old growth forest. Large portions of the site are also situated on or adjacent to significant natural or modern features, including Lake Petén Itzá, the steep slope leading down to the lake, a deep arroyo, and two modern roads. Further, in contrast to Motul de San José, which is situated within a protected reserve, the site of Trinidad is actually divided up amongst several small landholdings, both private and municipal, each of which needed to be studied separately. This combination of factors meant that survey and mapping at Trinidad had to involve a flexible set of methods and techniques. Surface Coverage and Survey Methods Most of the Trinidad site center, as well as part of its periphery, are located within a property owned by the municipality of San José and situated just north of the San José– San Pedro road. This land was leased by the municipality to a local rancher for pasture. Although the rancher was openly hostile towards archaeology at the start of the project, an agreement was eventually reached through the municipality and IDAEH to allow work to proceed. Surface coverage within this portion of the site was divided between open pasture to the west and scrub forest to the east. Formal survey methods were employed only within areas of scrub forest. Survey trails were cut east–west through these areas at intervals of 20 m and at least one archaeologist fully explored each trail. Structures or other features identified during survey were completely cleared of undergrowth, and

109

additional trails were cut out from these features in starburst patterns to a distance of 30– 50 m to identify possible associated features. As lateral visibility was relatively clear within the area of scrub forest, these surveys are unlikely to have missed any features other than those completely obscured by leaf litter. Within the open cow pasture no formal survey was necessary. When mapping started in at the end of the dry season in 2003, regular cattle grazing had reduced grass levels to a few inches throughout the site, revealing most features in their entirety, including some that would likely not have been found in forested areas. Mapping, including contour lines, covered virtually every part of this pasture making it highly unlikely that surface features were missed. The high level of feature exposure in this pasture did not, however, prevent interpretive issues from arising. One of the most difficult problems posed by the site’s use as pasture was the obvious impact of cattle on ancient Maya architectural remains. Most of the mounds at Trinidad have very rounded edges, likely a result of cattle trampling. Thus, despite a high level of visibility, it was somewhat more difficult to assess ancient Maya architectural configurations at Trinidad than at Motul de San José or other sites in the area. Looting activity, in contrast to other parts of the Motul de San José area, had not been particularly heavy within this part of Trinidad. Although many of the most prominent structures and all of the small temples had been illegally excavated, very few of the small housemounds had been destroyed by looters. Elsewhere in the Motul de San José area it is sometimes difficult to find structures of any kind that have not been sacked.

110

South of the San José–San Pedro road, the site was covered by various types of vegetation, with some old growth forest to the west and zones of grassland and scrub brush throughout the central and eastern sections of the steep slope leading down to the lake. Survey in these areas was conducted by the same methods utilized in the site center, with trails cut at 20 m intervals through the scrub brush and forested areas. Further, all mounds and other features were cleared, with secondary survey trails radiating out in a starburst pattern. One of the most difficult issues faced in this part of the site was acquiring permission from landowners to survey their properties. Although the municipality was eager to facilitate negotiations with those holding leases to municipal property, as was the case with the land north of the San José–San Pedro road, no governmental officials would involve themselves in negotiations with private landowners. Many of the properties around Lake Petén Itzá are owned as investment properties by non-local capitaleños with no incentive to encourage archaeological research. Although permissions were eventually received from the landowners holding the three principal properties south of the road, these permissions typically came well into the field season and ended up being extremely limited in duration. Total Station Mapping Total station mapping commenced from an initial datum placed atop the summit of Structure A-1, the largest building at Trinidad de Nosotros. This datum (MS 1) was established using a Trimble GeoExplorer II GPS provided by the Tulane University Center for Archaeology. To limit the signal error, the GPS was set to create a point file

111

Figure 5.2. Three-dimensional surface map of Trinidad de Nosotros.

for one half hour under optimal satellite coverage as determined by almanac. The resulting file, recorded in WGS84 format, was subsequently processed for differential correction online using base station data supplied by Trimble. The resulting location for the Trinidad MS 1 (Northing 193,110.9; Easting 1,882,882.6; Elevation 153.0 HAE [Height Above Ellipsoid]), is highly accurate (Figure 5.2). Mapping was conducted from this point by a team of two archaeologists and one crew member using a TopCon GTS-213 total station. While one archaeologist monitored the total station, the other positioned the reflector and recorded the point number and type in the mapping notebook. Architectural points were collected at all points of architectural

112

interface as determined by surface examination and along structural edges at least every 3–5 m. Contour points were collected in a starburst pattern from each mapping station to the limits of visibility. Starting from MS 1, a total of 13,180 data points were collected from 107 mapping stations. Of these data points, approximately half (6,860) represented contour data and 43 percent (5,675) represented architectural features. The remaining points were divided among quarries (207), chultunes (24), excavation units (71), post-occupational features (174), and various operational data points (169). Cumulatively, these points provided decent coverage, with an average of at least one data point for every 35 m2 of terrain within the 0.46 km2 area surveyed and mapped. This coverage was more than sufficient for producing a contour map, with the exception of one area south of the San José–San Pedro road and southwest of Platform BB-1. Although both survey trails and total station baselines had been established, the local landowner revoked access to this area prior to completion of total station mapping. This left several small gaps in mapping coverage and prevented the incorporation of two small mound groups into the Trinidad map. To resolve resulting problems in the contour map, 160 new data points, interpolated from nearby points, were added to this area as part of the “smoothing” process during the preparation of the site map. As their locations could not be accurately estimated, the mound groups were left off the map altogether. They were, however, included within the calculations for the site population estimate (see below).

113

Map Features Survey and mapping at Trinidad described a wide range of cultural and natural features. The following sections provide short definitions, descriptions, and survey nomenclature for the most common features. Structures, Groups, and Platforms Survey and mapping within the 0.46 km2 area covered by the Trinidad map identified a total of 124 separate mounds and other surface constructions at Trinidad. These were divided between platforms (31) and structures (93). These constructions were organized into 44 architectural groups during survey. Platforms were defined as any construction that served primarily to provide an elevated surface for another construction. Structures, in contrast, were defined as any feature that included internal rooms and either masonry or perishable walls. Architectural groups were defined as sets of structures and platforms that appeared to form a coherent unit distinct from that of other nearby groups. Most such groups were organized around a small patio or plazuela. Groups were given alphabetic designations during survey in order of discovery (Group A, Group B, etc.). After designations reached the letter Z, new groups were given doubled alphabetic designations (Group AA, Group BB, etc.). Structures were designated during survey and mapping with alphanumeric codes according to their position within a group, with the first structure number normally assigned to the northernmost structure and subsequent numbers assigned in a clockwise fashion (Structure A-1, Structure A-2, etc.). Platforms, unless they were the only features within a group, were

114

simply designated by group (Platform A, etc.). If platforms were the only features within a group, they were assigned alphanumeric codes (Platform BB-1, etc.). During the preparation of the site map, an additional set of designations were assigned to each group on the basis of alphanumeric grid position within the site (Group 6M1, Group 6M2, etc.) such that each arrangement of structures could be referred to in the most accurate manner possible. In some instances, survey groups were subdivided into several separate subgroups (survey Group N = “plazuela” Groups 7N1, 7N2, and 7M1). In this alternate classification, attendant structure and platform designations changed accordingly (Structure 6M1-1, Structure 6M1-2). These newer designations are provided in Table 5.2 (below) and in the description of residential group arrangements below in order to cross-reference descriptions appearing in other reports. The size and frequency of platforms are among the most noticeable architectural features at Trinidad. Many of the groups within Trinidad’s site center are situated atop large basal platforms and many of these are substantial. Platform D rises almost 6 m above the surrounding terrain and it is only slightly larger than the platforms associated with Structure A-4 and Group E (Figure 5.3). Most of the large residential groups in Trinidad’s site center are situated atop platforms 2–3 m in height (Groups C, G, H, J, R, U, and Y), and even some of the smallest residential groups are also situated atop modest basal platforms (Groups K, O, Q, T, FF, HH, II, and KK). Platforms are also noticeable along the slope leading down to the harbor. Groups BB, CC, DD, and JJ consist exclusively of small terrace-like platforms constructed on the edge of this steep slope. These last features presumably served as platforms for perishable structures.

115

Figure 5.3. Structure E-1 and associated platform viewed from the west.

Structures at Trinidad cover a wide range of features. Depending on where one measures from, Structure A-1, the largest at the site, rises some 8.5 m above the surrounding terrain (see Figure 5.7 below). Only Structures E-1 (ca. 4.5 m high) and F-6 (ca. 4.5 m high) approach A-1 in scale or height (Figure 5.4). The vast majority of structures were encountered in residential groups and were correspondingly smaller; however, within this smaller class of constructions, there are several subclasses. Structures associated with the largest and most formal residential groups are typically large, with their mounded remains rising between 1.5 to 2.5 m high and cross-sectional widths from 8–15 m. The heights of many of these structures have also clearly been reduced by cattle trampling. Excavations in Group C (Chapter 9) indicated that many of these large structures were vaulted and, in most instances, featured multiple internal galleries.

116

Figure 5.4. Trinidad surface map viewed from the south, over Lake Petén Itzá. Note the scale of platforms and principal structures when viewed from lake.

Structures on the lower end of the residential scale varied considerably, with heights between 1.0 m and a just few cm above the modern ground surface. The larger of these constructions are likely foundation-brace structures similar to those seen in Structure G-1 and in Group O (Chapter 9). The smaller constructions, just barely peeking above the surface, were likely perishable structures, with minimal masonry architecture. Plazas and Plazuelas Survey and mapping identified large amounts of plaza space at Trinidad. A plaza was identified as any area clearly circumscribed by ancient Maya architecture, which had been leveled to some degree by ancient Maya builders, and was likely intended as space for outdoor activities (Fash 1998; Inomata 2006; Wells 2004). The largest of these areas and those associated with ceremonial architecture were described as public plazas. A total of six public plazas were identified at Trinidad (Plazas

117

I–VI). All but two of these had been carefully leveled by the ancient Maya, and all were surrounded by architectural features. These areas were presumably utilized for a wide range of public purposes including ritual or other activities, and the largest were easily capable of holding thousands of individuals (see below). The term plazuela, in contrast, was applied to formal spaces delineated strictly by residential structures (Ashmore 1981:47; Conlan and Moore 2003:59–60). All but structural isolates or the most informal residential groups included some quantity of plazuela space. In some instances, residential group plazuelas were very large (e.g., Group U), covering more space than some of the public plazas. In most groups, however, plazuela space was much more intimate in scale. These spaces were presumably utilized for household activities. Chultunes Chultunes, or “bell-shaped subterranean pits” (McAnany 1990:266; Puleston 1965), are relatively common at Trinidad. A total of 12 ovoid chultun openings were identified during survey. This total likely greatly underestimates the actual number at Trinidad, as only those without capstones were identified. Those with capstone in place would have been virtually impossible to find beneath leaf litter, humus, or collapse (e.g., Puleston 1965:26). All but three of the chultunes were close to likely residential groups. Single chultun opening were identified near Groups V, Y, Z, CC, and KK. Closely spaced, paired chultun openings were identified in Groups S and FF. These presumably lead to a single chamber or set of connected chambers. Further, in three groups (Groups CC, FF,

118

and KK), access to chultunes was through the overlying platform. In one example (Group KK), access to the chultun was by a 1.0-m deep, masonry-lined hole cut down through platform fill from the surface. Two isolated chultunes were located in the northeastern-most area of survey coverage. These chultunes were likely associated with perishable structures or mounded architecture destroyed during construction of the modern Trinidad–Corozal road which passes beside them. Unfortunately, all of these chultunes had been filled completely by the local rancher to protect his livestock. A single chultun was also identified near the western edge of Plaza V. This chultun had been completely filled by the local rancher after a horse became trapped inside it. During a short conversation, he described this chultun as being quite different from others he had encountered at Trinidad, with a wider opening, a set of stairs, and multiple chambers within. The discovery of this chultun likely contributed to a local legend, reported by numerous informants, of a “tunnel” running under central Trinidad. From the rancher’s description, this chultun sounds like an especially well-built “shoeshaped” chultun with an antechamber and multiple connected main chambers (e.g., Dahlin and Litzinger 1986:721; MacKinnon 1990:357; Puleston 1971). Such well-built chultunes, with multiple connected chambers, are not uncommon in the Motul de San José area. Investigations at Akte, for example, identified a series of at least five connected chultun chambers near the Group A-B platform (Moriarty 2004b; Yorgey and Moriarty 2012). This chultun’s position within public space, however, within public space suggests that it may have performed a function distinct from those situated within residential groups.

119

The function of chultunes is a subject of some debate (e.g., Dahlin and Litzinger 1986; McAnany 1990; Puleston 1965). Trinidad’s close proximity to a permanent water source greatly reduces the likelihood that its chultunes were utilized for water storage. Use as household storage or another function appears more likely. Dahlin and Litzinger’s (1986:730-731) suggestion that chultunes would have provided ideal facilities for fermenting alcoholic beverages would certainly fall in line with other data from Trinidad (see discussion of Trinidad ballcourt middens in Chapter 8). Quarries Survey at Trinidad identified seven small depressions that were interpreted as quarries or borrow pits. All of these features were identified along the northeastern edge of the small upland area upon which Trinidad’s site center is situated. Along the arcing edge of this area, soil development is limited and limestone bedrock is just a few centimeters below the surface. These features were also located near several small residential groups (Groups N, X, and Z). The largest of these features, just west of Structure N-4, is roughly rectangular with approximate dimensions of 8-x-18 m. The smallest pit, a roughly ovoid depression approximately 35 m north of Group X has a diameter of only about 8 m. Several of the medium-sized pits (ca. 8-x-10 m) were situated directly adjacent to residential structures (Group Z). None of these features was more than 75 cm deep, though their depth prior to erosional infilling cannot be estimated. Surface examinations of these pits were limited, but at least two of the larger pits included the kinds of vertical ledges, with possible groove marks from tools, expected

120

along the edge of a quarry. At least some of these areas were presumably mined for stone blocks or limestone sascab for plaster production. These pits may also have been used for water storage, and one or two might be the remains of filled-in collapsed chultunes (Weiss-Krejci and Sabbas 2002). An eighth possible borrow pit, several meters deep, was encountered just south of Group G. Several workers identified this is as likely related to construction of the San José–San Pedro road, and it was left off the map. Monument 1 Only a single monument was located at Trinidad, identified as such by its position at the midpoint of the south endline of the Trinidad ballcourt (Figure 5.5). Monument 1 was excavated as part of site-center test pitting (Chapter 6). These excavations determined that Monument 1 was most likely round, with a diameter of roughly 50–60 cm and a thickness of 20–25 cm. Unfortunately, for at least the last two decades the ballcourt playing alley has been utilized as nighttime shelter for a herd of 20–30 head of cattle. As a consequence of regular trampling, this monument was almost completely destroyed and was, in fact, found in several dozen pieces with no remaining evidence of decoration. Its form and dimensions could only be estimated. Several large stone blocks located in the slope between Groups G and E were identified as possible monuments. Although the poor quality of the stone and their distance from other architecture made this unlikely, their position atop the surface away from mounds or other constructions was unusual and noteworthy. They may have been stone blocks removed during looting of Structure E-1, the remains of a monument or

121

Figure 5.5. Monument 1 during excavation. Possible ballcourt marker located on south endline of Trinidad ballcourt.

other piece broken up by looters, or large construction stones moved about by various post-depositional processes. Trinidad may have had at least one additional carved monument at Trinidad. One normally reliable informant reported that when she was a child her father had shown her “a stone with figures carved on it” at Trinidad, but that someone else had later taken it away. This story presumably recounts an episode of looting at Trinidad or some other nearby site, unfortunately a common story throughout the Motul de San José area.

122

Whether the piece she describes was Monument 1 or another monument cannot be determined. Natural Features In addition to the cultural features identified above, mapping covered several natural features that are important to the site’s development and which presented particular challenges to the survey and mapping crews. One of the most critical natural features was the shoreline of Lake Petén Itzá at Trinidad. As noted in Chapter 4, characteristics of the shoreline at Trinidad are among the most likely reasons for its development as a port. Irregular changes in lake level, however, presented problems in preparation of the site map. When mapping commenced in 2003, the shoreline was situated at an elevation of approximately 97.3 m HAE. During the course of the season, however, as the lake added volume from seasonal rains, the shoreline elevation climbed to 97.8 m HAE. During the 2005 field season, initiated during an unusually dry year, the lake level was found as low as 96.6 m HAE. In addition, a quick visit to the site in 2008 found the lake level as much as a meter lower than previous calculations. These measurements, of course, contrast with local informants’ descriptions of Trinidad with lake levels as much as 2–3 m higher than present. For consistency, lake level was depicted on the Trinidad map at an elevation of 97.5 m HAE, the average of measurements collected during the initial 2003 season. As these investigations show, however, lake level at Trinidad seems to have been marked by a high level of variability.

123

A second prominent feature covered by mapping was the slope leading down from the site center to the lakefront and harbor area. This slope constitutes the upper edge of the fault escarpment forming the north shore of Lake Petén Itzá. Although Trinidad presents an unusually high level of access to the lake from inland areas, this slope is still quite steep in some areas (Figures 5.2 and 5.4). From the northern edge of the slope, just south of the San José–San Pedro road (discussed below) to the shoreline, the slope drops 33 m in elevation over a distance of between 90 and 150 m. This produces an average decline of approximately 27.5% or a slope of about 15.4°. Large sections of this slope are much steeper than the average, however, and several areas drop at a slope of 30–45° or more. Mapping these areas presented several challenges, particularly in the steepest portions of the slope directly south of Structure LL-1. Coverage along the steepest slopes could only include limited survey coverage and a lower density of contour points. As these parts of the slope were barely passable at present, it appears unlikely that significant cultural features were missed. Along the more gently sloping areas directly west and east of the steepest slopes, however, coverage was detailed and included both formal survey and high densities of contour lines. Among the most difficult problems faced in the eastern half of the area south of the San José–San Pedro road was interpreting the extent of ancient Maya modification. Although geological processes have produced natural terraces all along the north shore of Lake Petén Itzá, most of these terraces at Trinidad appeared too regular to be entirely natural in origin and most were probably modified or even constructed by the inhabitants

124

of Trinidad. In fact, virtually all of the harbor slope terraces tested by excavation (see Chapter 7) were found to be at least partially artificial. But, to avoid the risk of falsely identifying these features as artificial constructions, only those features tested by excavation and interpreted as architectural remains were identified as such in the site map. Thus, the depiction of the harbor area on the site map likely underestimates the extent of ancient Maya modification. The final prominent natural feature confronted in mapping at Trinidad is a deep arroyo visible in the northwestern quadrant of the final map. This feature is remarkably deep (ca. 8–9 m), with the highest edge to the east, closest to Trinidad. This feature was so deep that when it was first seen from the northwest it was initially interpreted as a possible defensive feature. As it passes through Trinidad from north to south, it has a steep internal slope, increasing its depth by approximately 2.5 m over just over 200 m. The geological and hydrological history of this arroyo is not well understood. During initial investigations, it was assumed that this arroyo drained the Bajo Chäkokot, a seasonal swamp located approximately 1.5 km north of Trinidad and situated between Motul de San José and the tertiary settlement of Chäkokot. Subsequent discussions with local informants, however, indicated that this bajo was drained by the Arroyo Chäkokot, located just east of Trinidad and clearly visible on satellite maps of the area (Figure 4.6). Informants reported that the unnamed arroyo at Trinidad drained part of the much larger Bajo ‘Ox-‘ool, located approximately 1.5 km to the northwest and directly south of Motul de San José. There was some disagreement in this discussion, however, and time was insufficient to ground-truth varying interpretations.

125

Questions also emerged regarding the modern flow of this arroyo. Although quite deep in the area included on the site map, the arroyo disappears into undergrowth just south of the mapped segment and does not connect directly with the slope leading down to the lake. This is curious as the beach at Trinidad was almost certainly created from this arroyo’s discharge. The area where the San José–San Pedro lakeshore road intersects the arroyo’s course, however, frequently floods during the rainy season. Further, workers were busy installing a drain pipe under the road at the start of the 2003 season, though this drain is no longer visible. Presumably part of this arroyo was obscured by construction of the road, though its course remains something of a mystery. Regardless of its course, this arroyo appears to have had some importance in shaping Trinidad settlement patterns. Only a handful of mounds were identified west or northwest of this arroyo and most were situated near the arroyo’s western edge. The vast majority of mounds were located southeast of the arroyo. This may suggest that it did serve some defensive function. A similar pattern may also hold to the east. The Arroyo Chäkokot is even more substantial, and may have formed Trinidad’s eastern boundary. Certainly, the site’s location between these two deep natural features would have provided some measure of protection to the site’s inhabitants. Post-Depositional Features In addition to the features described above, mapping and survey dealt with several modern features that have undoubtedly impacted the site. The most obvious are two modern roads that pass through Trinidad. The first of these is the lakeshore road that connects San José to the community of San Pedro. This road passes through the site west-

126

to-east just to the south of Groups G, H, and R and north of the terraced platforms that form the upper limit of the harbor area slope (Groups BB, CC, DD, LL, JJ, and KK). The second road at Trinidad connects the San José–San Pedro lakeshore road with the community of Corozal located some 4–5 km to the north. The Trinidad–Corozal road starts at a junction with the San José–San Pedro road approximately 200 m southsoutheast of Group S and runs roughly north along the eastern edge of the site map, passing between the site and a pair of isolated chultunes approximately 150 m east of Group N. The position of a modern crossroad at Trinidad, linking east–west travel along the lake with a north–south overland route, constitutes added evidence for Trinidad’s role as a port during ancient times. The Trinidad–Corozal road was constructed along the same path-of-least-resistance likely utilized by the ancient Maya. The construction of these roads, however, likely had a dramatic impact on the site’s preservation. Elsewhere in the Motul de San José area, road crews frequently utilize mounds for fill despite protective laws, and as an excuse to opportunistically loot any large mounds. The high density of mounds on both sides of the roads, but not in otherwise level areas directly adjacent to them, suggests that at least several mounds were likely destroyed in road construction. In addition to roads, several modern habitations are located at Trinidad and are likely to have obscured ancient Maya constructions. Just above the harbor area, the local property guard has constructed a small house atop Platform JJ. In the southwestern most corner of the map, the local landowner has constructed a small camp as well as a house for the property guard. Further, directly east of the harbor, just off the site map, sits the

127

hotel Casa de las Americas. The hotel is likely situated directly atop numerous ancient Maya constructions and local workers reported finding several burials during hotel construction. Finally, at least one historical residence was also located in Trinidad’s site center, although its remains could not be located with certainty. As with the roads, it is likely that the construction of these modern buildings led to the destruction of numerous ancient Maya buildings. Informal Survey and Unmapped Areas From the beginning of survey it was clear that due to the site’s large size and the mosaic of local landholdings, it would be impossible to map all of Trinidad during a two season project. To supplement survey data, informal reconnaissance investigations were carried out in the areas surrounding Trinidad. These investigations collected basic information concerning peripheral settlement and are useful primarily in estimating Trinidad’s overall size. The most significant results came from reconnaissance work to the west and northwest of Trinidad’s site center. In a field across the lowest course of the arroyo, three mound groups are visible between approximately 200–250 m west of Group G. These groups are built atop large basal platforms and are similar in form to Groups G, H, J, and R within Trinidad’s site center. At least one of these platforms had been entirely bifurcated by tractor. Surface collections made by Lawton (personal communication, 2005) encountered extensive evidence for chert tool production near the closest of these mounds.

128

Several large mounds were identified to the north of these groups, approximately 350 m northwest of the site center. This area of the site was designated Group Angel for the local informant who had guided us to the spot. Although secondary growth was nearly impenetrable at this locus, it was immediately clear that it constituted a second node to Trinidad’s site center with multiple residential groups and at least one large temple. The principal residential group at this location was a large group, nearly identical in form to Group U in the site center. These groups feature six structures arranged around a rectangular plazuela, with two structures on the east and west sides, one structure on both the north and south sides, and a small temple in the southeastern position. This group had been heavily looted. The temple at Group Angel was too covered in undergrowth to effectively explore, though it appeared to be at least as high as Structure F-6 and perhaps higher. It had, however, been heavily looted. Other nearby residential groups were not explored, though glimpses through the undergrowth suggested that at least 4–5 more groups would be revealed with further clearing. To the north of the site center, only brief reconnaissance was conducted. These investigations encountered two small residential groups similar to the formal and informal residences situated along the eastern edge of the site center (Groups S, X, Z, and AA). Investigations directly east of Trinidad were also limited. Only a small part of the field across the Trinidad–Corozal road from Trinidad was explored. No mounds were

129

encountered. Two chultunes were discovered, however, and associated architecture may have been perishable or may have been utilized in modern road fill. The grounds of the Casa de las Americas, a hotel situated just east of the Trinidad harbor, were also briefly explored. This area had been heavily modified during construction of the hotel and the hotel owner maintains a collection of pieces recovered during the digging of foundation holes. Various modern structures appear to be situated atop ancient Maya platforms and perhaps as many as 4–5 additional groups, similar to those seen on the slope above the Trinidad harbor just to the west, may have been originally found on the grounds of the hotel. In total, reconnaissance investigations suggest that Trinidad likely covers an area extending at least another 300–500 m further to the west than the mapped area, 100–200 m further to the north, and an additional 200–300 m to the east. When these crude estimates are added to the site map, the rough estimate for Trinidad’s dimensions becomes about 900 m north–south by 1,200 m east–west for a total area of just less than 1.0 km2. This further suggests that the 0.46 km2 area covered by survey and mapping constitutes roughly half of Trinidad’s total settlement area. Final Map Preparation Field mapping data were downloaded to a Microsoft Excel worksheet and gridded and plotted in Golden Software Surfer 8. Contour lines were projected at 1 m intervals and the resulting contour map with encapsulated feature points was then exported to Adobe Illustrator 10. Structures were rectified and rendered in Illustrator as conjoined line features. Modern features and other post-depositional disturbances were not

130

depicted. The resulting map constitutes an estimate of Trinidad’s final form following site abandonment.

SITE ORGANIZATION AND DESCRIPTION Although Trinidad cannot be described as a truly monumental center on the scale of the major centers around Lake Petén Itzá like Motul de San José, Tayasal, or NixtúnCh’ich', survey and mapping nonetheless revealed a large and complex settlement. The following sections provide detailed descriptions of the most prominent structures and complexes at Trinidad, as well as short summaries of harbor area features and residential groups at the site. Public, Ceremonial, and Elite Architecture Trinidad’s site center includes a substantial quantity of obviously public, ceremonial, and elite architecture (Figure 5.6). Together, these complexes form a remarkably cohesive central precinct involving at least three temples, a large ballcourt, and numerous public plazas. The following section provides short descriptions for these complexes at Trinidad de Nosotros. Structure A-1 and Group A/D The principal structure at Trinidad is an 8.5-m-high temple situated atop a 4-mhigh basal platform (Figure 5.7). Structure A-1 is the tallest structure at Trinidad and provides an uninterrupted view of Lake Petén Itzá’s northern arm and a clear line of sight to some of the principal structures on the Tayasal Peninsula across the lake some 3.5 km to the south (Figure 5.8). Structure A-1 is visible over the same distances, and serves as

131

Figure 5.6. Trinidad site center with contours and small groups removed.

an obvious landmark from vantage points around the lake. Any activities conducted there would have been seen from numerous locations within the lake basin. Structure A-1 has been heavily trampled, making it difficult to estimate its original form. In plan, however, Structure A-1 has basal dimensions of roughly 38-x-34 m, with steep sides, and a relatively small upper surface (ca. 9.5-x-8.5 m). Although no staircase was identified during survey, the south side is the steepest and presumably constitutes the front of a small temple. A small collection of stones identified on the

132

Figure 5.7. Structure A-1 viewed from the north. This is the back and lowest side of the structure.

summit of this temple may constitute the remains of a Postclassic shrine (Timothy Pugh, personal communication, 2005). Structure A-1 is flanked to the southeast and southwest by small lateral extensions (Structures A-2 and A-5) that extend out to the south. These buildings, together with Structures A-3 and A-4, bound a small plazuela situated at the southern base of Structure A-1. This plazuela connects to a second smaller plazuela to the southeast via a narrow alley between Structures A-3, A-4, and A-6. The southern end of this second plazuela is formed by Group D, a massive platform group surmounted by 3–4 structures in a large

Figure 5.8. The view looking south from Trinidad Structure A-1. The first landmass visible across the lake is the Tayasal Peninsula. The approximate location of the archaeological site of Tayasal is indicated. 133

134

Figure 5.9. Group A/D. North is at the top.

135

formal group with a central plazuela. A fourth small plazuela is also found atop the massive basal platform that underlies Structure A-4. Cumulatively, the various plazuelas attached to Structure A-1 form what appears to be a single multipatio group, designated Group A/D (Figure 5.9). Within the settlement classification system applied in this project, such groups are described as Type VI groups (see below). The only other examples in the area are found at Motul de San José and Akte (Moriarty 2004b). Group A/D varies slightly from other Type VI groups in the Motul de San José area in that its plazuelas are situated at different levels, incorporating several different basal platforms. This situation may, in part, be a result of Trinidad’s longer history, with more extant platform architecture. If these same plazuelas were situated at ground level, they would be quickly classified as a single coherent complex. The Trinidad Ballcourt and Structure F-6 Trinidad’s site center also includes a large ballcourt and associated architectural complex (Figure 5.10). The ballcourt playing alley is roughly 12 m wide and runs north– south between two 25-m long lateral structures (Structures F-1 and F-2). Both lateral structures are approximately 2.0 m high and have narrow, ridge-like summits with an overall linear form distinct from other structures at the site (Figure 5.11). Although neither of these structures has been excavated, features visible from the surface suggest that both had sloping side walls and an upper cornice. Benches could not be identified at the base of the aprons, but would presumably be obscured by collapse and humus.

136

Figure 5.10. The Trinidad ballcourt and associated structures (Group F). North is at the top.

The northern end zone is well-defined. The playing alley opens into the small plaza (Plaza III) formed by Structure L-1 to the north, the edge of the Plaza III platform to the east, and a low wall running between Structures F-6 and L-1 to the west. To the south, the end of the alley is marked by a small monument (Monument 1, see above).

137

Figure 5.11. The Trinidad ballcourt's eastern lateral structure (F-2), viewed from the northwest.

Beyond the monument, the playing alley opens out into Plaza I (see below). Two of the more unusual features of the Trinidad ballcourt are two small structures attached directly to the eastern base of the eastern lateral structure (Structure F2). These structures (F-3 and F-4) extend out into Plaza IV. Structure F-3 is approximately 5-x-3 m with a height of only about 50 cm. Structure F-4 is larger (ca. 13x-6 m), though with a height of only about 75 cm. The possible function of these structures is obscure, though excellent evidence for feasting was recovered in middens directly abutting them (see Chapters 8 and 12). The ballcourt complex also includes a rectangular structure (F-5; ca. 12-x-9 m) located 14 m south of Structure F-4. Although small (ca. 50–75 cm high), Structure F-5

138

is situated above the southwestern corner of Plaza IV and forms the eastern edge of the ballcourt’s southern end zone. Further, Structure F-5 occupies a prominent position on the east side of Plaza I. By virtue of this position, this structure very likely represents the remains of a small temple, sweatbath, stela “house,” or other ceremonial construction. The western lateral structure (Structure F-1) also has a larger structure (F-6) attached to its western face. This structure has basal dimensions of approximately 28-x21 m and reaches a height of ca. 4.5 m, making it the second tallest structure at Trinidad . The local rancher reported finding stairs on the western side of the structure while digging postholes for a fence. By virtue of its orientation, the structure was interpreted as a small temple attached to the ballcourt. From its position, facing west into Plaza II, it appears that Structure F-6 was the principal eastern temple at Trinidad and likely very important during Late Classic times. The presence of such a large and sophisticated ballcourt complex at Trinidad underscores the importance of visitors and institutionalized ritual at this settlement. Relatively few ballcourts have been found elsewhere in the Lake Petén Itzá basin. Those that have been found have been identified have been found only at the largest basin-area settlements: Tayasal, Nixtún-Ch’ich', Paxcamán, and Ixlú. Motul de San José, in contrast to its smaller and presumably politically dependent neighbor of Trinidad, does not have a ballcourt. Structure E-1 and Group E The third major complex in Trinidad’s site center is Group E, located directly west of Group D across Plaza VI (see Figure 5.6). Group E consists of two structures

139

situated atop a 3.5 m high basal platform. Structure E-1, the northernmost, is roughly pyramidal in form with basal dimensions of approximately 31-x-28 m and a height of ca. 4.5 m. At present, Structure E-1’s original form cannot be surmised because it was heavily looted at some point in the recent past, with the looters’ trench later backfilled by the local rancher. Based on its position and orientation, Structure E-1 was most likely a small temple facing to the south. The second structure in Group E is located directly southeast of Structure E-1. Structure E-2 is rectangular in form (ca. 9-x-6.5 m) and approximately 1.0 m high. This structure forms the western edge of the small plazuela in front of Structure E-1. Structure B-1 “Range" Structure Among the largest constructions at Trinidad is Structure B-1 (Figures 5.12 and 5.13). Forming the north side of Plaza II and separated from Group C only by a narrow alley, Structure B-1 is a long narrow structure, approximately 50-x-10 m with an extant height of approximately 1. 5 m. Structure B-1 is bounded to the east by Group L and to the southwest by another tall linear structure (B-3) that closes off the northwestern corner of Plaza II. A small shrine or low foundation-brace structure (Structure B-2) is located just south of its eastern end. Surface examination of Structure B-1 identified at least four depressions that likely indicate collapsed doorways on its south side. These likely provided entrance to separate vaulted rooms or galleries. These internal areas were presumably accessed by a series of low terraces similar to those seen in front of Structure C-1 (see Chapter 9). Overall, the surface form of mound B-1 is similar to that of C-1, albeit larger in scale, and

140

Figure 5.12. Groups B and C from the top of Structure A-1, looking northeast. Excavations visible to the left of center are in Structure C-1.

likely employed similar masonry techniques. Structure B-1 is a range structure, presumably utilized for an array of courtly and administrative activities as depicted in ancient Maya art (Reents-Budet 2001:199–204). Within the Motul de San José area, such range structures are found at many sites. The densest concentrations are found in the Motul de San José Acropolis, which is essentially composed of multiple superimposed range structures. Stand-alone, or unattached, range structures are also reasonably common, with one of the largest being Structure 8M6-1 at Motul, a 45-x-15 m structure located just north of the Motul Acropolis. These features also appear at some of the smallest sites in the Motul de San José area. Reconnaissance

141

Figure 5.13. Groups B, C, L, and W. North is at the top of the map.

at the site of Chäkälte (Moriarty et al. 2000), a small tertiary site located 2 km northwest of Trinidad, identified a 70-m long range structure. Similarly, Ellen Moriarty’s (2011) investigations at Xilil, another small site located 2 km northeast of Trinidad, identified a range structure more than 40 m long. The Group L “Palace” (Structures L-1 and L-2) Trinidad’s site center also includes a massive elite residential building that may constitute a two-story “palace.” Group L is located directly north of the ballcourt’s northern end zone (Plaza III) and southeast of Structure B-1, and consists of two

142

structures (L-1 and L-2) attached at right angles to form a single massive L-shaped edifice (Figure 5.13). L-shaped structural arrangements are very common in the Motul de San José area within formal residential groups (Moriarty 2004b; see below). In these groups, the largest constructions are almost always found on the west and north or west and south sides. These parts of the groups are normally interpreted as living quarters. Group L, however, constitutes an extreme example of this arrangement as its basal dimensions (Structure L-2 = 38 x 20 m) and height (ca. 4 m) dwarf other possible residential constructions at Trinidad. Although its internal form is impossible to estimate given the heavy cattle trampling in the area, its great size and elevation suggest that it was either a vaulted structure with an unusually elaborate roof or a two-storey structure. If Group L did serve as a palace or high-level elite residence, then it was wellplaced. The south façade of Structure L-1 provides a direct view into the northern end of the ballcourt and any activities conducted within the ballcourt area. Further, the west façade of Structure L-2 looks directly out into Plaza II, the most obviously ceremonial of Trinidad’s public plazas (see below). Plazas and Public Space In addition to the elite and ceremonial architectural complexes discussed above, Trinidad has an unusually large amount of public plaza space (ca. 18,674 m2) (see Figure 5.6 above). Table 5.1 provides short descriptions and dimensions for each of the public plazas at Trinidad. The following sections describe each of the six numbered plazas, as well as another area identified as having a possible public function.

143

Table 5.1. Descriptions and areas for Trinidad Plazas I–VI. Plaza Plaza Area (m2) Description 5,979

Separate, but joined plazas accessed from the south via a 3-m high staircase. Plaza I located south of the ballcourt. Plaza II located west of the ballcourt.

Plaza III

602

Located between Structure L-1 and the ballcourt playing alley; the north end zone for the ballcourt.

Plaza IV

2,646

Located to east of the Trinidad ballcourt, north of Groups U, and west of Group Y. Northern edge delineated by platform.

Plaza V

7,500

Large, slightly sloping plaza bounded by Group A to the east and long linear Structure P-3 to the west.

Plaza VI

1,947

Located between Groups A, D, and E.

Plazas I–VI (Total)

18,674

Plazas I–II

Plazas I–II The most formal and obviously ceremonial space at Trinidad is formed by Plazas I–II. Although these two plazas are conjoined and share the same sequence of floors (see Chapter 6), access patterns, shape, circumambient architecture, and micromorphological research demonstrating different construction details all suggest they should be treated separately (Spensley 2004). Plaza I is located directly south of the Trinidad ballcourt, and at least the northeastern quadrant of the plaza served as the ballcourt’s southern end zone. Plaza I is roughly square with dimensions of approximately 58 x 53 m. Access to this plaza was achieved from the south via a 30-m wide staircase rising approximately 2.5 m above

144

ground level. The exceptional width of this staircase suggests that Trinidad’s residents anticipated large crowds in Plazas I–II. In addition to the ballcourt, Plaza I is bounded to the east by Group U and to the west by Group D. Both of these groups likely constituted elite residential groups during the Sik’u’ I–II Late Classic, with Group D forming part of the larger Group A/D. A large platform (Structure I-1) forms the southern side of the plaza, next to the staircase. The stone foundation for a very small structure or shrine was also identified near the southern end of the plaza (Structure I-2). From Plaza I, visitors to central Trinidad could pass into Plaza II between the ballcourt temple (Structure F-6) and two large mounds that form part of Group A (Structures A-6 and A7). Plaza II is much more rectangular in plan (ca. 78-x-33 m) than Plaza I, and is bounded by the largest ceremonial and elite architecture at the site. The ballcourt temple (Structure F-6), the Group L “palace,” and an intervening wall serve as the east side of the plaza. The 50-m long range structure (B-1) forms the north side of the plaza, while the west side is framed by Structure B-3 and the northern half of Group A, including Structure A-1. A small shrine or other structure was also identified in the northeastern corner of this plaza (Structure B-3). Plaza III Plaza III forms the northern end zone for the Trinidad ballcourt measuring roughly 40-x-18 m. This plaza is well-delineated in all directions with the ballcourt to the south, the wall between L-1 and F-6 to the west, and a platform edge to the east. Structure L-1 forms the north side of this plaza and it appears very likely that this area

145

would have served as a plazuela associated with the Group L “palace” as well as an end zone for the ballcourt. The only formal entrance to this plaza would be through the ballcourt playing alley. Plaza IV The fourth public plaza at Trinidad is located directly east of the Trinidad ballcourt. Plaza IV is roughly rectangular in plan (ca. 63-x-50 m) and is bounded to the south and southeast by elite residential groups (Groups U and Y). To the north and northeast, the plaza is delineated by the edge off the plaza platform. The western edge of the plaza is defined by the eastern edge of the ballcourt complex, including Structures F2, F-3, F-4, and F-5. Plaza IV appears distinct from other plazas at Trinidad, both in its physical location and its pattern of access. Although test excavations determined that Plaza IV had a sequence of floor construction similar to that of Plazas I–III (Chapter 6), its upper surface is approximately 2 m lower than neighboring plazas. This lower position makes Plaza IV much more difficult to see from other parts of the site. Access to Plaza IV was also distinct from other plazas at Trinidad. While Plazas I–III, V, and VI all appear most accessible from the south, and are at least conceptually oriented towards the lake, Plaza IV is most accessible from the north. These differences all suggest that Plaza IV may have had a different function compared to other plazas at Trinidad. Plaza V Of all the plazas at Trinidad, Plaza V is the largest, least formal, and the most easily accessible. Plaza V is bounded by Group A to the east and Group E to the south.

146

Figure 5.14. View of Plaza V from Structure E-1. Figures are near the center of the plaza. The large mound to the right is Structure A-1.

The western edge of the plaza is formed by Structure P-2, a small (ca. 50 cm high) rectangular building (ca. 8-x-5 m) and Structure P-3, a very low construction that runs south from Structure P-2 for approximately 50 m. Structure P-3 is barely visible on the surface, and presumably marks the location of a low wall or other feature atop the terraced western edge of the plaza. The entrance to a chultun (discussed above) is located southeast of Structure P-2’s southern end. The northern end of Plaza V is defined by Structure P-1, a low, slightly amorphous platform that has been heavily damaged by cattle trampling. The area covered by Plaza V measures approximately 125-x-60 m for a total area of roughly 7,500 m2. This total approaches that of all other plazas found at Trinidad

147

combined. Plaza V is unusual, however, in that it lacks the long sequences of construction and leveling found in Plazas I–IV (see Chapter 6). Although testing did identify fill deposits as well as a packed earth floor (Chapter 6; see also Spensley 2004), it remains unclear if Plaza V was ever systematically leveled and plastered completely as seen elsewhere at Trinidad. The plaza actually slopes slightly downwards from east to west. Nonetheless, the area appears to have been carefully delineated and kept vacant of construction and available for other activities throughout Trinidad’s long occupation. Access to Plaza V does not appear to have been restricted in any way, though the plaza is most approachable from each of the inter-cardinal directions. Perhaps the most formal entrance is from the southeast, where those entering the plaza may have passed through Plaza VI and then along the terrace or via situated at the west edge of the Group A platform (see below). Plaza V’s tremendous size, easy access, and limited surface manipulation all suggest that it was intended for activities distinct in some way from those of other public plazas at Trinidad. Though such a space would make an ideal candidate for a marketplace (Dahlin et al. 2007), other possibilities should be considered as well (Inomata 2006; Wells 2004). Plaza VI The last of the plazas identified at Trinidad is located between Group E and Group D. This area was not identified as a plaza during survey; only the edge of a terraced platform was identified. During preparation of the site map, however, it became increasingly clear that the area directly east of Group E and north of this terrace was

148

remarkably level and free of architecture; it was, therefore, designated Plaza VI. It has not, however, been tested by excavation. Plaza VI is situated between one of Trinidad’s three largest temples (Structure E1) and a possibly elite residential group with a massive basal platform (Group D). Group A forms the northern edge of the plaza. Access to Plaza VI appears to have been from the south over the edge of the terrace or staircase that forms its southern boundary. Additional access or egress may also have been available to the north through a north– south terrace or via (described below). Public Terrace or North–South “Via” In addition to Plazas I–VI, Trinidad’s site center also includes a large public terrace or ceremonial via running ca. 60 m north along the eastern edge of Plaza V from a starting point near Plaza VI. On the ground this feature is very difficult to discern, especially in tall grass, but appears to consist of a low terraced wall running parallel with the base of the Group A platform. This terrace creates a slightly raised area approximately 10 m wide. This feature originates near the northern edge of Plaza VI. The function of this feature is obscure, though its location between Group A and Plaza V suggests its possible importance, and it almost certainly acts as the eastern edge of the plaza. Its north–south orientation is also potentially interesting. Such causeways and avenues are relatively common at major centers in the southern lowlands and may have been utilized for processions or other ritual events (Chase and Chase 2001; ReeseTaylor 2002; Schele and Freidel 1990). Several are also known among Trinidad’s neighbors. Motul de San José’s site center includes a large north–south avenue or via

149

(Moriarty 2004b), and a small, likely ceremonial, causeway was identified at the site of Akte (Yorgey and Moriarty 2012). Further, the site plan of Tayasal includes a very large north–south avenue between north-central and northwest Tayasal (Chase 1983:942). Harbor Area Features The second major area at the site, the Trinidad harbor area, includes four principal features: Platform EE, Structure EE-1, Platform GG, and Area A (Figure 5.15). The surface appearances of these features prior to excavation are described below, along with interpretations provided by informants and other observations. The final section provides a brief review of these features under conditions of slightly higher lake level. Platform EE The most obviously artificial feature within Trinidad’s harbor area is the large basal platform situated at the lowest level of the terraced slope leading down from the site center. Platform EE is roughly L-shaped, with a large central surface covering an area of approximately 1,100 m2 and a lateral arm extending an additional 65 m to the southeast. Both the southern and western edges of this platform are regular, well-defined, and easily visible at a distance; the upper northern limits to the platform are much more difficult to define as the platform grades into the hillslope. At present, the upper surface of Platform EE rises 3 m above modern lake level and approximately 2.5 m above Area A (see below). At the time of investigation, this area was being used for maize cultivation. Numerous informants also discussed fishing from this platform. Because of its close proximity to the lake and the possible harbor at Trinidad, this platform was interpreted as

150

Figure 5.15. Trinidad harbor area with principal structures indicated.

a possible loading platform or other feature associated with harbor activities. Its close proximity to current lake level, well below most other ancient Maya features around the lake, suggested a residential function was highly unlikely. Structure EE-1 Structure EE-1 is a 13.5-x-6.5 m structures that starts near the base of Platform EE and extends out to the south-southeast into Area A. At present, this feature stands only about 75 cm high. From the surface it appears as a dense mound of cobble-sized

151

stones only lightly mixed with clayey subsoil (see Chapter 7). Structure EE-1 sits just a few cm above modern lake level. Local informants report that when the lake is slightly higher (see below), Area A fills with lake water and Structure EE-1 serves as a useful platform for fishing or docking a canoe. Although it is unclear if the ancient Maya utilized docks as they are understood in modern terms (see Chapter 4), Structure EE-1’s position within Area A is certainly suggestive. Platform GG The largest feature within the harbor area is Platform GG, the 67-m long feature that forms the southwestern edge of the harbor. Platform GG consists of two clearly defined parts: the larger southeastern end (Figure 5.16), which rises 5.5 m above lake level and covers an area of approximately 22-x-19 m, and a narrower, slightly lower extension (3.0 m above lake level) that runs 45 m to the northwest. From the surface, the higher eastern end of Platform GG appears slightly irregular, and appears to be a large piece of uplifted bedrock. The lower western end of the platform, in contrast, appears remarkably regular in form and almost certainly involved some amount of ancient Maya construction (Figure 5.17). Local informants referred to Platform GG as a small peten or island (Hofling and Tesucún 1997) and most did not think it was possible that it could be an ancient Maya construction. A large looters’ pit near the eastern summit of the structure, however, suggested that at least some regarded it as an ancient Maya feature. They were, however, in agreement that its position created unusually still waters within Area A and the

152

Figure 5.16. The high, eastern end of the isleta/Platform GG, viewed from Platform EE. The person standing in the center of the photo is positioned at the northern edge of platform and within Area A. Other figures, up and to the left of center in photo, are positioned at the top of the platform.

Trinidad lagoon to the west, making it an effective breakwater regardless of origin. One informant noted that the platform’s position was particularly useful in blocking the large waves seen in Lake Petén Itzá during occasional hurricanes or tropical storms. The destructive effects of these waves are clearly evident in camps all along the northwest shore of the lake.

153

Figure 5.17. Trinidad Platform GG seen from Lake Petén Itzá, with form outlined in white. Note the regular shape of the lower western end of the platform (to the left).

Area A The last harbor area feature, Area A, is the lowlying zone between Platforms EE and GG. Area A rises just above modern lake level, and is currently filled with thick lacustrine clay. To the southeast, the lake can be accessed from Area A over a narrow, sandy beach between Platform GG and Group HH. To the west, Area A opens up into the Trinidad lagoon, referred to locally as an ixpet-ja’, or small, round lake or lagoon in Itzaj Maya (Hofling and Tesucún 1997:275), even though it connects with Lake Petén Itzá. Local informants reported that, until recent times, Area A was fully inundated and connected directly to the lake. According to informants, only the recent pattern of short rainy seasons has left this area dry and exposed, and even slight rises in lake level lead to

154

this area being innundated. When lake level rises, Area A serves both as an easy landing spot for canoes and as an excellent fishing spot favored by numerous local informants. The Harbor Area and Changes in Lake Level Although assessing the function of harbor area features was the primary objective of harbor area excavations (Chapter 7), informants’ descriptions of the relationship between lake level and various harbor area features were largely confirmed by other sources. Anecdotal data collected from around the lake, for example, confirmed that lake level has dropped several meters within the last decade. In San Benito and Flores numerous shoreline areas have been reclaimed from the lake in recent years, allowing the construction of new buildings and a ring road running around the periphery of the island of Flores. The shoreline in nearby San José is also much more extensive than in the recent past. Informants do, however, point out that water levels last peaked at higher levels only after the 1976 earthquake. Total station mapping also confirmed that Area A is low enough to be impacted by even the smallest changes in lake level. Figure 5.18 provides 50-cm contour maps for the harbor area at four different lake levels: current (as of 2003; 97.5 m HAE), and current plus 50 cm (98.0), plus 1.0 m (98.5), and plus 1.5 m (99.0). As this figure shows, even a modest increase of 50 cm is sufficient to partially fill Area A and separate Platform GG from the mainland. When lake level rises another 50 cm, Area A fills in completely. When lake level reaches 1.5 m above the modern level, all of the harbor area’s most prominent features are clearly defined, with Platform EE exposed to lake

155

Figure 5.18. Trinidad's principal harbor features with varying lake levels.

water on both the west and south sides and Structure EE-1 projecting well out into the lake. Residential Groups Survey and mapping at Trinidad identified a total of 37 architectural groups interpreted as residential. To facilitate inter-group and inter-site comparisons, residential groups were classified on the basis of a scheme developed for the Motul de San José

156

Archaeological Project (Moriarty 2004b). The following section provides a summary of this scheme (Figure 5.19). The succeeding section provides a short description for some of the principal groups at the site, as well as an overall summary of residential group distributions at Trinidad. Residential Group Classification: MSJ "Plaza Plans" The residential group classification developed for the Motul de San José area was derived in part from the Tikal classification developed by Becker (2003). In the Tikal scheme groups are classified on the basis of their “plaza plan.” This approach considers a group’s architectural arrangement as a reflection of the beliefs and decision-making of the ancient Maya builders, thus classifying groups on emic rather than potentially etic attributes (Becker 2003:256). This approach to settlement classification lends itself to field application. Most of the structures encountered during survey are covered in deep layers of soil and collapse that make it impossible to provide accurate descriptions of any building’s internal layout or construction method without excavation. Some changes were made to this scheme when applied to the Motul de San José area. Most importantly, the classification focused exclusively on residential groups. This appears warranted both by perceived functional differences between arguably residential and public ceremonial groups and by the economic emphasis of Motul de San José Archaeological Project investigations. Residential groups, or “minimum residential units,” are generally understood to represent the shared domestic quarters of a minimal social unit (Ashmore 1981). Although there are distinct epistemological problems in identifying a “household” in the archaeological record (Ashmore 1981; Ashmore and

157

Figure 5.19. MSJ residential group classification scheme.

158

Wilk 1988; Gillespie 2000; McAnany 1993), the occupants of a residential group are expected to have engaged in a variety of cooperative economic activities. Further, although the range of activities occurring in residential groups may have overlapped with those found in public or ceremonial complexes, residential groups are seen as providing a closer estimation of patterns in household economic activity. For pragmatic reasons, this classification scheme also utilizes some additional attributes of group form or architectural investment. For example, a small number of groups in the Motul de San José area were characterized by multiple conjoined plazuelas. Rather than treat each plazuela separately, classifying each according to plan, a separate type was created for multipatio groups. Likewise, some groups appeared to be characterized by extremely large basal platforms. These were treated as a separate type. Finally, in some instances, patterned differences between otherwise similar plans led to the creation of two or more subordinate types (MSJ Types IIIa and IIIb; see below). MSJ Types I–II Within the Motul de San José system, the Type I and Type II classifications designate the “informal” and “formal” residential groups found throughout the Maya lowlands and widely distributed at Trinidad and elsewhere in the Motul de San José area (Figure 5.20). Structures in Type I and II groups are generally rectangular in form and are assumed to have had primarily “non-ritual” functions (Becker 2003:259; see below). Type I groups are defined as those groups of 1–7 structures lacking a formal patio area and identified by the “constituent structures being located closer to each other than other structures or groups” (Ashmore 1981:49; Sharer 1994:474). Type II residential groups

159

Figure 5.20. MSJ Type II, Type V, and Type IV groups east of the Trinidad site center.

are defined as groups of 2–7 structures organized around a square or rectangular patio (Ashmore 1981; Sharer 1994). For simplicity, terraced platforms without visible surface constructions, but likely to have been utilized for perishable residential structures, are included as Type I groups.

160

Figure 5.21. MSJ Type IIIb groups south of Trinidad site center.

MSJ Type III The Type III classification was designed initially to deal with a set of groups at Motul de San José characterized by large autonomous platforms surmounted by only 1–2 structures. In these groups, a very considerable quantity of plazuela space on the platform’s upper surface was left vacant. Although future investigations may identify “invisible” structures atop these platforms, they were isolated as a separate class. These are referred to as Type IIIa groups.

161

Investigations at Trinidad and elsewhere also identified a class characterized by small formal groups (Type II) situated atop large basal platforms (more than 1 m high). These arrangements were classified as Type IIIb groups. Although separating groups with large basal platforms may or may not be instructive, the use of such large basal platforms is clearly selective and would have been instantly identifiable to the ancient Maya inhabitants of the area. Further, separating platform groups may help highlight specific activities associated with those kinds of groups. Freidel and Sabloff (1984:190), for example, interpreted many basal platforms encountered on Cozumel Island as storage facilities. In contrast, at Chau Hiix in Northern Belize, Cook (1997) used artifact assemblages associated with platform groups to identify a “middle class” associated with craft production activities. MSJ Types IV–V Type IV and Type V residential groups are, in most respects, identical to Type II groups and consist of 3–7 structures arranged around a common patio. Type IV and V groups are distinguished, however, by the presence of one or more “ritual” structures. “Ritual” structures are defined as those edifices that are sometimes higher than other structures in the group, roughly square in plan, and can be interpreted as temples, shrines, or oratorios (Becker 2003). Such structures were identified, following Becker (2003:59) on the basis of width to length ratios (>0.70). Unfortunately, as a result of heavy looting in the Motul de San José area, such structures can also often be identified on the basis of architectural characteristics and the remains of looted high-status burials.1

1

Almost all of the eastern temples in the MSJ area have been thoroughly looted. Exploration and cleaning of numerous looters’ trenches suggests that looters were almost always successful in locating rich tombs

162

Type V groups include a “ritual” structure on their eastern side and are similar to “Plaza Plan 2” at Tikal (Becker 2003). Type IV groups, though rare, are identified by the presence of a “ritual” structure on any side other than the east. Among the limited number of Type IV groups identified to date in the Motul de San José area, the “ritual” structure is normally situated on the west side of the group. As is the case at several other sites in the southern lowlands (e.g., Valdés and Suasnávar 1991), several variants of the standard “Plaza Plan 2” or MSJ Type V arrangement have been identified. The standard Type Va group includes a ritual structure on the midpoint of the east side (e.g., Trinidad Group C). The variant Type Vb group has a ritual structure near the southeastern corner of the group (e.g., Trinidad Group Y). MSJ Type VI The Type VI classification was created to isolate the small number of large residential groups with multiple adjoined patios. Type VI groups normally include 5–12 structures organized around 2–4 enclosed patios. In all instances the structures and patios of Type VI groups form coherent architectural complexes that cannot be systematically subdivided. For example, several of the Type VI groups in the Motul de San José area have a northern patio of Type V classification and a southern patio of Type II classification. Thus, the Type VI classification was created specifically to account for the fact that numerous large, multi-patio residential groups in the MSJ area exhibit beneath eastern temples. Screening of looters’ backdirt from a tunnel that had entered a burial chamber in the eastern temple of MSJ Group D, for example, led to the recovery of 10 jade beads, 20 pieces of a jade mosaic, 145 pieces of a shell mosaic, 98 pieces of a pyrite mosaic, and other elements of the dedicatory offering (see Foias 2001, 2003:23). Assuming the looters removed the best pieces, the associated burial appears to have been of a high or, possibly, royal status (Foias 2003). At the nearby tertiary center of Chäkokot looters also encountered what was probably an important burial in an eastern temple. Although any significant grave goods were removed, the tomb itself was an elaborate construction with a stone lintel roof and cut-stone retaining walls.

163

characteristics conforming to multiple classifications. Such groups are normally quite larger and presumably correspond to high level elite residences with at least some ceremonial component as well. MSJ Type VII An additional Type VII classification was established for residential groups exceeding Type VI in number of patios and structures, while continuing to serve an arguably residential function. To date, however, only one such group has been encountered in the Motul de San José area—the Motul de San José Acropolis. The Motul Acropolis is a massive architectural complex covering more than 83,000 m2 and including at least six separate patios (Moriarty 2004b). This complex was presumably home to the governing royalty of the Ik’ polity and their retainers. Residential Groups at Trinidad de Nosotros As noted above, survey and mapping identified 37 residential groups at Trinidad. Table 5.2 presents summary data for each of these groups organized by MSJ group type. Groups are identified both by alphabetic survey name (e.g., Group L) and plazuela group designation (e.g., 6M5). Additional group attributes are presented, including the number of structures, the orientation of the principal structures, the presence or absence of a basal platform or terrace, and the location of the largest structures in each group. As Table 5.2 demonstrates, all of the MSJ types are represented in residential settlement at Trinidad. At the upper end of the settlement classification, MSJ Type VI is represented by Group A/D. As noted above, Group A/D is a special case incorporating

164

Table 5.2. Summary data for residential groups at Trinidad de Nosotros. Survey Group

Plazuela Group(s)

A-D

U Y C S

6M6, 6M8, 6M9, & 5M2 5N1 6N7 6M2 5N3

G H J R W L Q T K O Z M KK I N

5M4 5M5 5M3 5N7 6N3 6M5 5N5 5N2 6M1 6L1 6N4 7M2 4N1 5N4 7N2

N N V W

7M1 7N1 7L1 6N1

X AA BB CC DD DD DD FF HH II JJ LL MM

6N2 6N6 5M6 4M1 5M7 4M2 4M3 4N4 3N1 4O1 4N3 4N2 5N6

Notes

Str. W-1

Strs. N-3 & N-4 Str. N-1 Str. N-2 Strs. W-2 &W-3

Str. DD-1 Str. DD-2 Str. DD-3

MSJ Type

No. of Structures

Orientation

Platform

Principal Structure(s)

VI

10

Cardinal

Basal

N

Vb Vb Va IV

6 4 4 4

Cardinal NNW–SSE Cardinal Cardinal

Basal Basal Basal --

NW W N,W N,E

IIIb IIIb IIIb IIIb IIIa II II II II II II II II I I

3 2 2 3 1 2 4 2 3 3 5 2 2 1 2

Cardinal Cardinal Cardinal Cardinal Cardinal Cardinal Cardinal Cardinal NE–SW NE–SW Cardinal NW–SE NNE–SSW Cardinal NE–SW

Basal Basal Basal Basal Basal -Terrace Terrace Terrace Terrace --Terrace Basal Terrace

W W W E E S,W S W N,W N,W N NE N -NE

I I I I

1 1 2 2

Ind. Ind. NE–SW Cardinal

-----

--N,W E

I I I I I I I I I I I I I

2 2 1 1 1 1 1 1 1 1 1 1 1

Ind. Ind. Ind. Ind. Ind. Ind. Ind. Cardinal NE–SW ESE–WSW Ind. Ind. Ind.

Terrace -Terrace Terrace Terrace Terrace Terrace Basal Basal Terrace Terrace ---

-------N S S ----

165

four connected plazuelas and three basal platforms, as well as the site’s principal structure. Although the inclusion of Structure A-1 suggests that this group had an important ceremonial function, at least one or more of the associated plazuelas is likely to have been residential, housing high-ranking members of the local sociopolitical hierarchy. Three groups were found to have eastern temples (Type V; Groups C, U, and Y). Two of these (Groups C and U) were among the largest residential groups at Trinidad, and all three are positioned prominently at the site, close or directly adjacent to major public plazas. Group C is a typical “plaza plan 2” type (Type Va) with a temple in the eastern position. This group is oriented to the cardinal directions, seated atop a large basal platform, and situated directly north of Plaza II, just behind Structure B-1. At ground level, this group is among the most prominent at the site. Groups U and Y constitute slight variants of the “plaza plan 2” type (Type Vb), with the temple in the southeastern position. Group U is the largest single patio group at the site, with six structures and a large basal platform oriented to the cardinal directions. The temple in this group is the southern-most of two structures located on the group’s east side. Group Y is slightly smaller, oriented several degrees west of north, with the temple in the southeastern corner of the group. The only Type IV group at Trinidad is Group S (5N3), one of three formal groups within a small area east of Group U and southeast of Group Y. None of the buildings in Group S are higher than 75 cm and most are quite small, though all are arranged on the cardinal directions around a small plazuela. The possible shrine or oratorio (Structure S-

166

3) is located on the south side of the group. This unusual location for a shrine is reflected in other parts of the group as well; the largest structures in the group, normally found on the north and west sides, are located on the north and east sides in this group. Settlement at Trinidad includes numerous large platform groups. Most of the platform groups at Trinidad (Groups G, H, J, and R) conform to the Type IIIb arrangement with a platform over 1 m high and a small formal group on top. Only a single group (Group W) was classified as a Type IIIa group; further, its designation is highly provisional as further investigation would likely identify additional perishable structures. All of these groups were aligned with the cardinal directions; the most prominent structures are always located on the east or west sides, never to the north or south. The majority of groups at Trinidad were classified as simple formal (Type II) or informal (Type I) residential groups. Within these classifications, however, there is a considerable degree of variability in size and composition. The two structures in Group L are among the single largest constructions at Trinidad, with two possible stories, and this group likely served as a high-level elite residence (see above). Most formal groups were much smaller, however, with structures between 0.5 and 1.0 m high. Informal residences (Type I) were particularly diverse and included many structure isolates. In many instances, these structures likely form only the visible portion of groups, with additional perishable structures impossible to identify at the surface. This category also included a large number of isolated terraces, most of which were identified

167

on the northern edge of the harbor slope. Most of these terraces are assumed to have served as basal platforms for perishable structures.

POPULATION ESTIMATES AND PLAZA CAPACITIES In assessing the community at Trinidad de Nosotros, it was important to provide an estimate for the site’s population during the Sik’u’ I–II Late Classic. Although population estimates are built on nested assumptions and produce variable results (Culbert and Rice 1990), they nonetheless provide some “real-world” context for an otherwise abstract community. Assessing Trinidad’s population was particularly important in light of the extensive plazas encountered there. Though the areal extent of Trinidad’s plazas do not compare to those found at major centers like Tikal or Copan, they do appear unusually great for a site of Trinidad’s modest size. The extent of Trinidad’s plazas is shown to be particularly great when placed in context with its estimated population. The following section provides a short description of the methods utilized to estimate Trinidad’s Late Classic population. This is followed by a discussion of plaza capacity estimates at Trinidad. The results of this comparison are discussed below and in the conclusions. Estimating Trinidad’s Late Classic Sik’u’ I–II Population To estimate Trinidad’s Late Classic Sik’u’ I–II population, this study utilized an approach derived in part from Tourtellot’s (1990) work at Seibal focusing on “structure units,” with these units corresponding to plazuela groups at Trinidad. The emphasis on

168

groups of structures, rather than structure or room counts (Culbert et al. 1990; Webster and Freter 1990), appeared particularly applicable to Trinidad settlement data. As Tourtellot (1990:85-86) noted at Seibal, structures are frequently impossible to identify beneath leaf litter or other vegetation, and estimating internal architectural divisions is prone to error without excavation. Tourtellot’s estimate of 2.72 dwelling structures per structural unit also appears to closely conform to residential patterning within the Motul de San José area. Most residential groups are composed of 2–4 structures. In most groups, 2–3 of these structures are rectangular in form and slightly larger than others. These structures are normally interpreted as dwellings, while smaller nearby structures are interpreted as ancillary structures. Using ethnohistorical data as a guide (Hellmuth 1972, 1977; Calnek 1972; Jones 1979), Tourtellot's (1990:93) estimate of a total of 11.9 inhabitants for each of these groups was applied to the Trinidad data. Determining the exact number of residential groups at Trinidad, as well as their periodicity involved various estimations and calculations. As discussed above, survey and mapping at Trinidad identified a total of 37 groups with arguably residential functions. Further, informal survey in the areas surrounding the mapped portion of Trinidad identified an additional 20 residential groups that must be added to the total. This total (57), however, fails to include additional groups within unsurveyed areas or groups that may have been destroyed by post-depositional processes (e.g., road building). First, to account for the possibility that some residential groups were destroyed during the construction of the San José–San Pedro and Trinidad–Corozal roads, the group

169

count was increased by ten percent (5.7 groups). Further, an additional ten percent (5.7 groups) was added to account for groups missed during informal survey. Combined, these additions bring the Trinidad residential group count to a total of 68 (68.4). As the number of missed structures is much more likely to have been higher than lower, no deduction was made for less than full occupancy at any one point in time. To estimate the chronological phase for these groups, residential group counts were compared to available excavation data. Stratigraphic units, midden tests, or other excavations were carried out within or adjacent to 17 of the 37 residential groups identified during survey and mapping (see Chapters 6–9). Of these, all but one small group (Group HH) produced strong evidence for a Sik’u’ I–II occupation, including plazuela floor construction or at least moderate-density circumambient midden. From this, a figure of 94.1% Late Classic periodicity was derived for Trinidad. When this factor is applied to the group count (68.4), a total of 64.4 residential groups are estimated for Sik’u’ I–II Trinidad. To complete the estimate, the Sik’u’ I–II group count (64.4) was multiplied by an estimated 11.9 people per residential unit (Tourtellot 1990:93; see above). This produced an estimated Sik’u’ I–II population of 766.4 people. Although this figure, rounded down to 750, constitutes a very rough estimate of Trinidad’s population during the Late Classic Sik’u’ I–II phases, it places Trinidad comfortably within the second tier of the local settlement hierarchy and at the third tier on the regional level. The application of the same methods to Motul de San José, for example, would produce a population estimate at least five times greater.

170

Plaza Capacity Estimates One of the most impressive aspects of Trinidad’s site center is the sheer quantity of public plaza space. As shown in Tables 5.1 and 5.3, Trinidad’s Plazas I–VI cover a combined area of more than 18,000 m2. Even if only the two most obviously ceremonial plazas (Plazas I–II) are discussed, the total plaza space is still quite extensive, covering nearly 6,000 m2. Less prominent public plazas like Plaza IV (2,646 m2) and Plaza VI (1,947 m2) add significantly to that total, while the largest plaza (Plaza V; ca. 7,500 m2) nearly equals in coverage that of all other plazas at Trinidad combined. The total extent of these plazas is notable at both the local and regional level. Trinidad’s close neighbor Motul de San José, a site many times larger and more complex than Trinidad, has only slightly more plaza space (23,000 m2). In addition, public space at Trinidad dwarfs that found at other secondary and tertiary sites in the Motul de San José area like Akte (3,778 m2), Chäkokot (4,077 m2), and Buenavista (2,870 m2). Further, although major centers like Tikal and Copan provided significantly greater space, some slightly smaller centers like Aguateca actually overlap in plaza space with that seen at Trinidad (see Inomata 2006:816). To put Trinidad’s plaza space in context, plaza areas were compared using estimates of plaza space requirements per person developed by Moore (1996) and Inomata (2006). This study uses two of the indices applied by Inomata (2006:812) to the spatial requirements of people within a plaza: 0.46 m2/person and 1.0 m2/person. Plaza capacity estimates for Trinidad and selected sites in the Motul de San José area are

171

Table 5.3. Plaza capacity estimates for Trinidad de Nosotros and selected sites in the Motul de San José area and the southern Maya lowlands. Plaza Capacity Estimates Plaza Area (m2/Person) 2 Site (m ) 0.46 m2 1.0 m2 Trinidad de Nosotros Plazas I–II Plaza III Plaza IV Plaza V Plaza VI

18,674 5,979 602 2,646 7,500 1,947

40,596 12,998 1,309 5,752 16,304 4,233

18,674 5,979 602 2,646 7,500 1,947

Motul de San José Chäkokot Buenavista Akte

23,000 4,077 2,870 3,778

50,000 8,863 6,239 8,213

23,000 4,077 2,870 3,778

Tikal Copan Aguateca

117,626 50,500 14,745

255,708 109,783 32,054

117,626 50,500 14,745

presented in Table 5.3 and compared to those calculated for Tikal, Copan, and Aguateca by Inomata (2006:811-816). Although the capacity estimates presented here have heuristic value only, the results are nonetheless instructive and more than a little surprising. Using the figure of 1.0 m2/person, Plaza I – II combined have a capacity of 5,979 people and, taken as a whole, Plazas I–VI have a capacity of 18,674 people. Uses of the less conservative figure of 0.46 m2 /person produces immensely large capacity estimates, with Plazas I–II capable of holding 12,998 people and Plazas I–VI more than 40,000 people. The potential capacities of Trinidad’s six public plazas are particularly impressive for a site of Trinidad’s relatively small size. Table 5.4 presents a direct comparison of Trinidad’s public plaza space with the population estimate discussed above for Sik’u’ I–II

172

Table 5.4. Comparison of plaza capacities and population estimate (750 people) for Late Classic Sik’u’ II Trinidad de Nosotros. Plaza Area 0.46 m2/person 1.0 m2/person 2 Plaza (m ) Capacity %Pop. Capacity %Pop. Plazas I–II Plaza III Plaza IV Plaza V Plaza VI Total (Plazas I–VI)

5,979 602 2,646 7,500 1,947 18,674

12,998 1,309 5,752 16,304 4,233 40,596

1,733.1 174.5 766.9 2,173.9 564.4 5,412.7

5,979 602 2,646 7,500 1,947 18,674

797.2 80.3 352.8 1,000.0 259.6 2,489.9

Trinidad (750 people). Plaza capacities are listed for each plaza along with the percentage of Trinidad’s total population covered by this estimate. Again, though the values presented here are an abstraction, the contrast between Trinidad’s relatively small population and its extensive public plaza space is noteworthy. Using the figures of 1.0 m2/person and 0.46 m2/person produces total plaza capacities between 24 and 40 times Trinidad’s total estimated population.

SUMMARY In summary, survey and mapping revealed Trinidad to be a modest-sized secondary center within the Motul de San José area notable primarily for the complexity of its site center and the presence of a harbor. Trinidad's site center is remarkable for both the number and extent of formal public plazas it provides. As noted above, the area of plaza greatly exceeds the potential requirements of Trinidad's residents alone, suggesting that Trinidad's builders anticipated hosting large numbers of outside visitors. Further, the presence of an elaborate ballcourt complex at the heart of the settlement places Trinidad within select company within the Lake Petén Itzá basin.

173

Ballcourts have been identified previously in the lake basin only at the large settlements of Tayasal, Nixtún-Ch'ich', Ixlú, and Paxcamán. Motul, Trinidad's nearby neighbor and political superior, does not have a ballcourt. This suggests that public rituals and events requiring a formal ballcourt were carried out at Trinidad, rather than Motul, a situation that suggests some level of cooperation between the two settlements. Finally, the presence of a harbor at Trinidad anticipates an important role for the settlement in trade and transportation entering the Motul de San José area. That several of the most prominent harbor features, particularly Platforms EE and GG, are of tremendous size suggests that this area constituted an important part of the site and key to its function within the Motul de San José area.

174

CHAPTER SIX THE SITE CENTER STRATIGRAPHIC TESTING PROGRAM

INTRODUCTION The principal goal of the site center stratigraphic-testing program was to provide an overview of Trinidad’s long-term development. Field investigations focused on Trinidad’s site center, the upper portion of the site situated atop the ridge overlooking Lake Petén Itzá and bounded to the north and west by the deep arroyo cut. This portion of Trinidad includes 30 discrete architectural groups and six public plazas and constitutes both the ceremonial and residential heart of the settlement. The stratigraphic-testing program consisted of 18 excavation units 1-x-1 m or larger placed in one-third of all architectural groups and four out of six plazas (Table 6.1; Figure 6.1). Four additional units excavated as part of the architectural-testing program (Chapter 9) are discussed here as they provide data relevant to the construction history for three groups. The following sections provide unit-by-unit descriptions of excavations. These are followed by a more detailed summary of excavation results and a perspective on long-term developments in the Trinidad site center.

175

Figure 6.1. Trinidad site center with stratigraphic test units highlighted.

METHODS AND ORGANIZATION Test locations were determined so that at least a third of all residential groups and nearly all public architectural complexes within Trinidad’s site center would be tested. Public architectural investigations focused on the site’s three largest complexes (Groups A, E, and F [the ballcourt]) and the four principal public areas (Plazas I, II, IV, and V).

176

Residential groups were selected for testing based on location, size, and plaza plan (see Chapter 5) so that the sample would represent groups of varying size and complexity, evenly distributed throughout the site center. Further, as this program of research was oriented towards long-term developments, residential group testing was also weighted towards groups with clear basal platforms. Groups without basal platforms presented less opportunity for collecting diachronic data and, further, would be assessed during midden testing (see Chapter 8).

Table 6.1. Location and dimensions for site center test units.

Group

Unit(s) Dimensions (m) Location

Plaza I Plaza II Plaza IV Plaza V

1I1 1B1 1F3 1P1 1P2

1-x-1 1-x-1 1-x-1 1-x-1 1-x-1

Ballcourt

1F1 1F2 1A1 1C1 4A13

1-x-1 2-x-2 1-x-1 1-x-1 1.5-x-1

Center of ballcourt playing alley. South end line of ballcourt playing alley. Group A Platform, just south of Str. A-1. Group C Platform, just south of Str. C-1. Beneath Str. C-1.

E G

1E1 1G1 1G2 1G3 5A7

2-x-1 1-x-1 1-x-1 1-x-1 1-x-1

Group E Platform, just south of Str. E-1. Center of Group G Platform. North extension to Unit 1G1. West extension to Unit 1G1. Beneath Str. G-1.

K N O

1K1 1N1 1O1 6A13 6D9

1-x-1 1-x-1 1-x-1 1-x-1 1-x-1

Group K Platform, just east of Str. K-3. Center of Group N Platform. Group O Platform, just south of Str. O-1. Beneath Str. O-1. Beneath Str. O-3.

U Y

1U1 1Y1

1-x-1 1-x-1

Group U Platform, just west of Str. U-2. Center of Group Y Platform.

A C

Plaza I, just north of Group I Platform. Plaza II, just west of Str. F-6. North edge of Plaza IV. Northern half of Plaza V. Southern half of Plaza V.

177

Units excavated as part of the site center testing program were placed 2–3 m in front of the principal structure in a group. This approach mirrored strategies utilized at Motul de San José and situated units in areas where moderate structural collapse was likely to improve preservation of upper floors. Alternate locations were selected only when physical barriers (e.g., trees, fences, etc.) prohibited placement, or where a principal structure could not be identified. Excavation methods included the use of arbitrary 10-cm lots and all other standard practices discussed in Chapter 4 and Appendix A. During the 2003 season all test units measured 1-x-1 m. Several units were also expanded in one or more directions to allow for recovery of burials or other features. Although these units provided a considerable amount of data and most encountered long sequences of superimposed floors, the small number of artifacts collected within such small units and the limited window each provided limited their utility in assessing Trinidad’s occupational sequence. These units were also judged to be overly small for excavations in some of Trinidad’s large basal platforms, many of which stand well over 3-m high with structural cores of loosely packed fill. As a result, several units excavated during the 2003 season had to be terminated prematurely due to the risk of collapse. During the second season a smaller number of larger units (1-x-2 or 2-x-2 m) were excavated. While some of these units were placed in new groups, additional units were added to previously tested groups known to have long sequences. In total, 18 units were excavated using these methods. The four units excavated in conjunction with the

178

architectural testing program differed from the first series only in that they were excavated through the floors of structures following clearing excavations.

ORGANIZATION OF EXCAVATION SUMMARIES Unit descriptions are divided into two sections. The first section covers excavations in public architectural complexes. The second deals with residential group excavations. A short summary, focusing on the principal cultural features identified by excavation, is provided for each unit. Floor designations are utilized for both preserved plaster and earthen constructions and for discrete fill episodes for which associated floors are absent or eroded. As the principal focus of these investigations was chronological, brief descriptions of associated ceramics are provided where applicable. Context summaries for each unit are attached as tables. These tables provide the stratigraphic depths at which features were encountered, a brief description of the feature, and the phase or phases to which they were assigned. Some features could not be dated securely, and provisional phase designations are denoted with an asterisk. Additional information is provided in attached figures. Other stratigraphic data pertaining to those groups tested by for middens or as part of the architectural-testing programs are presented in Chapters 8 and 9. The conclusion to this chapter provides a brief overview of chronological developments in the Trinidad site center. Additional summaries of burials, caches, and other excavation data are provided in the appendices.

179

PUBLIC ARCHITECTURE TESTING Testing within Trinidad’s central precinct focused on the site’s principal public architectural complexes and included excavations in Groups A, E, and F, as well as Plazas I, II, IV, and V. The following descriptions are organized spatially, starting with Group A, then moving to the eastern edge of the central precinct, before turning to areas situated on the western edge of this precinct. Unit locations are provided in Table 6.1 and Figure 6.1. Group A: Unit 1A1 Excavations in Group A consisted of a single 1-x-1-m unit placed 3 m south of Structure A-1’s central axis and within the small plaza framed by Structures A-1, A-2, and A-4 (Figure 6.1). Unit 1A1 was excavated to a depth of 2 m below the modern surface where it was terminated due to risk of collapse (Figure 6.2). Although early termination of this excavation precluded full assessment of Group A’s construction sequence, Unit 1A1 identified four episodes of construction and one burial. The uppermost floor in Unit 1A1 was identified at a depth of 0.20 m, beneath a mix of humus and collapse from Structure A-1. An abrupt transition from humus to moderately dense fill (see Chapter 4 for fill class descriptions) marked the location of an eroded floor, designated Floor 1. Ceramics within Floor 1 fill included both Sik’u’ II and Yaljob’ach diagnostics, although most of the Terminal Classic materials were recovered in the uppermost lots and were likely intrusive. A Sik’u’ II date was assigned to Floor 1. Associated Yaljob’ach materials presumably constitute refuse from activities in Group A or materials from a completely eroded later floor.

180

Figure 6.2. East profile of Unit 1A1.

Floor 2 was identified at a depth of 0.70 m. Although no physical traces of a floor were identified, a clear transition to darker soil and heavier, more densely packed fill, as well as an obvious shift in the ceramic inventory, signaled a separate construction

181

episode. The plaster floor originally present at this level likely eroded completely prior to the construction of Floor 1. A possible disturbance was noted in the southwestern quadrant of the unit. Ceramic materials associated with Floor 2 included both ‘Ayim-tun and general Sik’u’ diagnostics. Although polychromes were not present in sufficient numbers to differentiate between Sik’u’ I and II, mixing of Early Classic and Late Classic monochromes is common in Sik’u’ I deposits, and an early Late Classic Sik’u’ I was assigned. Floor 3 was identified at a depth of 0.95 m. Plaster, approximately 8 to 10 cm thick, was well preserved throughout the unit with the exception of the southwestern quadrant where a hole feature cut through the floor. Ceramics were sparse in Floor 3 fill and included a few Middle Preclassic sherds, several Late Preclassic sherds, and a handful of intrusive Sik’u’ sherds associated with the disturbed area. Although the Preclassic sherds were heavily eroded and somewhat mixed, both the presence of Chukan types and the overall thickness of the floor indicated a Chukan phase date. Floor 4 was identified at a depth of 1.32 m. As with the overlying floor, Floor 4 was in excellent condition except in the southwestern quadrant where the hole feature noted above continued. With the exception of a handful of intrusive Sik’u’ sherds associated with the disturbed area, ceramics consisted exclusively of late Middle Preclassic diagnostics and a secure Ix Cha’ phase date was assigned. Excavation of Floor 4 fill led to the identification of Trinidad Burial 1. Burial 1 was encountered beneath the disturbed area in the southwestern quadrant of the unit. Only the small portion of the burial extending into Unit 1A1 was excavated. Skeletal

182

Figure 6.3. Eroded Sik'u' II polychrome vessel from Burial 1 (Complete Vessel 1). Note "kill hole" on vessel bottom. Drawing by I.Seyb.

materials were poorly preserved, but included the left humerus, the right and left ulnae, and the left femur, as well as several unidentifiable long bone shafts (see Appendix B). Although the bones are fragmentary, their large size suggests an adult (Thornton and Moriarty 2007). From the positioning of the bones, it also appears likely that Burial 1

183

was flexed, with the body aligned north–south and facing east. In this reconstruction, the lower extremities were located slightly south and west of Unit 1A1 with the cranium just west of the excavation baulk. Although no grave architecture was located with Burial 1, a simple cist or other construction may have been unidentifiable in associated fill. Grave offerings included a complete, but eroded, orange flaring-walled polychrome bowl with a "kill hole" (Complete Vessel 1; Figure 6.3) placed inverted just east of the cranium’s posited position. Numerous fragments of river clam shell (cf. Unionidae) presumably constitute the remains of a pendant or other ornaments. Although the associated vessel was heavily eroded, its form was diagnostic for the latter half of the Late Classic, dating Burial 1 to the Sik’u’ II phase. Burial 1 was presumably inserted through Floors Two to Four during or just prior to construction of Floor 1.

Feature/Context

Table 6.2. Context summary for Unit 1A1. Depth (m) Phase Description/Notes

Humus/Collapse

0.00–0.20

Sik’u’, Yaljob’ach, Säk-tunich

Humus mixed with collapse from Structure A-1.

Floor 1

0.20–0.70

Sik’u’ II

Eroded plaza floor and associated fill.

Floor 2

0.70–0.95

Sik’u’ I

Dense, heavy fill beneath possible eroded floor.

Floor 3

0.95–1.32

Chukan

Plaster floor and associated fill.

Floor 4

1.32–2.00

Ix Cha’

Plaster floor and associated fill.

Burial 1

1.60–1.80

Sik’u’ II

Burial within earlier fill.

184

Plaza I: Unit 1I1 Test excavations in Plaza I consisted of a single 1-x-1-m unit located 3 m north of the Group I platform (Figure 6.1). Unit 1I1 was excavated to a depth of 1 m below the modern surface and documented four episodes of construction (Figure 6.4). The first floor in Unit 1I1 was encountered at a depth of approximately 0.10 m. Floor 1 was entirely eroded and was identifiable only in the profile of the unit as the upper limit of a thin fill level. Ceramics from associated fill were highly eroded but included many Late Classic diagnostics. Although a handful of well-preserved polychromes suggest a Sik’u’ II date, it was unclear whether these were intrusive from mixed humus deposits above and a general Sik’u’ I–II date was assigned. Ceramics associated with Floor 1 at the same depth in Plaza II also produced a general Sik’u’ I–II date. Floor 2 was identified at a depth of 0.30 m. As with the overlying floor, Floor 2 was heavily eroded, although a ballast layer and fragments of eroded plaster at the start of a 35-cm-thick episode of densely packed, heavy fill suggested the floor’s approximate location. Ceramics were recovered in very small quantities within this fill. Although the best preserved sherds were Late Preclassic, many eroded sherds appeared to be later, and the floor was assigned a tentative ‘Ayim-tun phase designation. Subsequent excavations in Plaza II identified an ‘Ayim-tun floor at the same approximate depth, thereby strengthening this interpretation. A third floor was encountered at a depth of 0.65 m. Floor 3 was a plaster floor, 5–10 cm thick and relatively well preserved throughout the unit. Only a thin layer of

185

Figure 6.4. North profile of Unit 1I1.

ballast separated Floor 3 from the floor below it. The intervening fill contained few sherds, all Late Preclassic, so the floor was assigned a Chukan date. The fourth and lowest plaster floor in Unit 1I1 was approximately 5 cm thick and lay directly beneath Floor 3 at a depth of 0.76 m. As with the overlying floor, Floor 4 was well preserved throughout the unit. Although many sherds were recovered in Floor 4 fill, few were diagnostic of either the Middle or Late Preclassic. Floor 4 was initially assigned a general Preclassic Ix Cha’–Chukan date; however, the identification of a pure

186

Middle Preclassic materials at the same depth in Plaza II suggested an Ix Cha’ date was most likely for Floor 4.

Feature/Context

Table 6.3. Context summary for Unit 1I1. Depth (m) Phase Description/Notes

Humus/Collapse

0.00–0.10

Säk-tunich, Sik’u’

Humus mixed with collapse from Group I platform.

Floor 1

0.10–0.30

Sik’u’

Eroded plaza floor and associated fill.

Floor 2

0.30–0.65

‘Ayim-tun*

Partially eroded plaster floor and Associated fill.

Floor 3

0.65–0.76

Chukan

Plaster floor and associated fill.

Floor 4

0.76–1.00

Ix Cha’*

Plaster floor and associated fill.

Plaza II: Unit 1B1 Test excavations in Plaza II consisted of a 1-x-1-m unit placed 3 m west of Structure F-6 (Figure 6.1). Unit 1B1 was excavated to a depth of 1 m below the modern surface and documented five episodes of construction (Figure 6.5). The first floor in Unit 1B1 was identified at a depth of 0.10 m beneath a thin layer of humus and collapse from Structure F-6. Floor 1 was entirely eroded and could be identified only by an abrupt shift from humus to fill at this depth. Ceramics recovered within associated fill were highly mixed, but included several Late Classic diagnostics and Floor 1 was assigned a general Sik’u’ I–II phase date. A second floor was identified at a depth of 0.30 m. Although Floor 2 was heavily eroded, its position was identifiable by a thin layer of eroded plaster situated atop a moderate to heavy fill episode with only a light admixture of soil. Ceramics within

187

associated fill included many secure Early Classic diagnostics and Floor 2 was assigned an ‘Ayim-tun date. The third floor in the 1B1 sequence was identified at a depth of 0.63 m. In contrast to the overlying floors, Floor 3 was well preserved and consisted of plaster approximately 5–9 cm thick. As with Unit 1I1, only a thin layer of ballast was excavated beneath Floor 3 before the next floor was encountered. Although ceramics in Floor 3 fill were comparatively small in number, all of the best preserved sherds were Late Preclassic and a Chukan phase date was assigned.

Figure 6.5. North profile of Unit 1B1.

188

The fourth floor in Unit 1B1 was encountered just below Floor 3, at a depth of approximately 0.72 m. Floor 4 was a plaster construction approximately 5 cm thick. Ceramics within associated fill included many well-preserved preserved Mamom sphere sherds, as well as a small number of pre-Mamom sherds, and the floor was assigned an Ix Cha’ phase date. A fifth floor was identified only within the last 10 cm above bedrock. Floor 5 consisted of a thin, 3–5-cm thick level of eroded stucco situated atop a light sandy fill level directly atop bedrock. Although few diagnostic sherds were recovered within associated fill, the best preserved were late Middle Preclassic in date and an Ix Cha’ phase date was assigned.

Feature/Context

Table 6.4. Context summary for Unit 1B1. Depth (m) Phase Description/Notes

Humus/Collapse

0.10–0.20

Sik’u’

Humus mixed with collapse from Structure F-6.

Floor 1

0.20–0.30

Sik’u’

Eroded plaza floor and associated fill.

Floor 2

0.30–0.63

‘Ayim-tun

Partially eroded plaster floor and Associated fill.

Floor 3

0.63–0.72

Chukan

Plaster floor and associated fill.

Floor 4

0.72–1.00

Ix Cha’

Plaster floor and associated fill.

Floor 5

1.00–1.10

Ix Cha’

Plaster floor and associated fill.

Group F (Ballcourt): Units 1F1 and 1F2 Stratigraphic excavations in ballcourt consisted of two units. The first, Unit 1F1, was a 1-x-1-m unit placed near the center of the playing alley (Figure 6.1). This unit was

189

excavated to a depth of 1.44 m below the modern surface and documented five episodes of construction (Figure 6.6). The first floor was identified at a depth of 0.10 m. Floor 1 was entirely eroded, but was identified by the appearance of eroded plaster and an abrupt transition from humus to moderate fill. Ceramic materials within associated fill consisted exclusively of Late Classic types, with few polychromes, and Floor 1 was assigned a general Sik’u’ I–II date. Floor 2 was identified at a depth of 0.40 m. Although this floor was almost entirely eroded, small sections of preserved plaster and a clear ballast level marked Floor 2’s location in the unit profile. Ceramics within associated fill were somewhat mixed and included both Early Classic and Late Classic diagnostics. The Late Classic sherds, however, were recovered only in the uppermost lots suggesting the possibility of intrusion from the overlying Floor 1 fill, and Floor 2 was assigned a tentative ‘Ayim-tun phase date. The third floor was identified at a depth of 0.80 m. Floor 3 was well preserved in the northern half of the unit and was a plaster construction approximately 8 cm thick. Only a thin layer of ballast was found between Floor 3 and the next floor in the sequence. Ceramics from this ballast level were relatively limited in quantity, but included several Late Preclassic diagnostics and Floor 3 was assigned a Chukan date. The next floor in Unit 1F1 was found directly beneath Floor 3 ballast and at a depth of 0.90 m. Floor 4 was well preserved and consisted of plaster approximately 5 cm thick. As with the previous floor, only a thin layer of ballast was found between Floor 4

190

Figure 6.6. East profile of Unit 1F1.

and the next in the sequence. Ceramics recovered within the intervening fill were relatively few in number, but included both Mamom and pre-Mamom Middle Preclassic diagnostics. Floor 4 was assigned an Ix Cha’ phase date. The final floor in the 1F1 sequence was encountered at a depth of 1 m and just beneath ballast associated with Floor 4. Floor 5 was well preserved and consisted of plaster approximately 5 cm thick. Fill associated with Floor 5 produced only a limited

191

number of sherds, but included both Mamom and pre-Mamom Middle Preclassic diagnostics and the floor was assigned an Ix Cha’ phase date.

Feature/Context

Table 6.5. Context summary for Unit 1F1. Depth (m) Phase Description/Notes

Humus/Collapse

0.00–0.10

Sik’u’

Humus and collapse from Structures F-1 and F-2.

Floor 1

0.10–0.40

Sik’u’

Eroded plaster floor and associated fill.

Floor 2

0.40–0.80

‘Ayim-tun

Partially eroded plaster floor and associated fill.

Floor 3

0.80–0.90

Chukan

Plaster floor and associated fill.

Floor 4

0.90–1.00

Ix Cha’

Plaster floor and associated fill.

Floor 5

1.00–1.44

Ix Cha’

Plaster floor and associated fill.

The second unit excavated in Group F was a 2-x-2-m unit placed on the south end line of the ballcourt playing alley (Figure 6.1). This unit, Unit 1F2, was placed so as to encompass the cluster of limestone fragments designated Monument 1 and thought to represent the remains of an entirely eroded or uncarved ballcourt marker (Figure 6.7). The secondary objectives of Unit 1F2, in addition to dating construction of the ballcourt playing alley, were to identify any preserved monument fragments and to provide a date for the monument’s placement. Unit 1F2 was excavated to a depth of 1.60 m and documented nine possible episodes of construction and one cache, as well as a series of disturbances that remain to be interpreted (Figure 6.8).

192

Figure 6.7. Monument 1 during excavation.

The first level of Unit 1F2 was excavated in a series of horizontal and vertical lots around Monument 1 to identify the position of the ultimate floor in the ballcourt playing alley and to assess its relationship to the monument. A possible eroded floor was identified at a depth of 0.20 m and designated Floor 1. The mixture of humus and fill above and below this floor produced a large number of artifacts, including more than 500 potsherds, six obsidian prismatic blade segments, two chert tools, more than 300 chert flakes, and two figurine fragments. At least some of these may have been deposited in

193

Figure 6.8. South profile of Unit 1F2.

association with Monument 1 or resulted from activities in the playing alley. The vast majority of the sherds from this episode dated to the Late Classic period, and polychromes included both Sik’u’ I and Sik’u’ II diagnostics. Floor 1 was dated to the Sik’u’ II phase, but this date is slightly tentative as the later diagnostics may be intrusive. At a depth of approximately 0.40 m, a second, heavier fill level was identified. Although no physical trace of a floor was visible at this depth, the lower fill appeared to

194

constitute a separate episode from the overlying fill, prompting the designation of Floor 2. The ceramic inventory within this fill differed slightly from that seen in overlying fill and consisted of a mix of Early Classic and Late Classic diagnostics, including numerous well-preserved Sik’u’ I polychromes. Floor 2 was assigned a Sik’u’ I date. The position of the monument with respect to these two floors was clarified following excavation. After the first lots of excavation the monument was found to be shattered into several dozen pieces, the largest of which had dimensions of approximately 40-x-20-x-25 cm (Figure 6.7). No new or clearly carved pieces were identified, although the monument seems to have been round or ovoid in its original form, with an approximate thickness of 20-25 cm. Further, the monument fragments were situated atop a separate section of fill consisting entirely of large, cobble-sized stones with little intervening matrix or artifacts. This fill was excavated separately from nearby fill and was found to be situated atop Floor 2. Assuming this section of fill was placed to serve as a foundation for the monument, its positioning atop Floor 2 and surrounded by Floor 1 fill suggests a Sik'u' II for the monument. The third floor in Unit 1F2 was identified at a depth of 0.50 m and consisted of plaster approximately 4 cm thick. Although Floor 3 was partially preserved in the western third of the unit, only isolated fragments of plaster marked the floor’s depth in other areas. Some destruction of Floor 3 may have occurred with the deposition of the fill associated with Monument 1, described above. Ceramics from within Floor 3 fill consisted primarily of Late Preclassic types, although a single large and well-preserved

195

Early Classic polychrome recovered directly beneath the preserved section of Floor 3 dated its construction to the ‘Ayim-tun phase. Beneath Floor 3, at a depth of 0.60 m, a second partially preserved plaster floor was encountered. As with the preceding floor, Floor 4 was only well-preserved in the western third of the unit and was identified by isolated fragments of plaster in the remainder of the unit. Ceramics from Floor 4 fill included both Late Preclassic and Early Classic types, although Early Classic materials were recovered only in the upper most lots and may have been intrusive from overlying fill. Floor 4 was assigned a tentative Chukan phase date. Below Floor 4, at a depth of 0.73 m, a fifth partial plaster floor was encountered. As with the preceding floors, Floor 5 was only partially preserved, but with the best preserved sections in the eastern half of the unit. Ceramics from fill directly beneath preserved portions of the floor included both Middle Preclassic and Late Preclassic diagnostics and Floor 5 was assigned a Chukan phase date. The sixth floor in the 1F2 sequence was encountered at a depth of 0.86 m and just below the thin level of fill underlying Floor 5. Floor 6 was a well-preserved plaster floor approximately 4 cm thick. Ceramic materials within the thin layer of ballast between Floor 6 and the underlying Floor 7 included several Late Preclassic diagnostics, and the floor was assigned a Chukan phase date. The seventh floor was encountered at a depth of 0.90 m and just below Floor 6 ballast. Floor 7 was a well-preserved plaster floor and approximately 6 cm thick.

196

Associated fill produced a large number of Late Preclassic diagnostics and Floor 7 was assigned a Chukan phase date. Floor 7 also produced a small cache. Cache 2 consisted of two small greenstone objects located approximately 40 cm apart along the north–south centerline of the unit. The southern object was a small broken or incomplete biconically-drilled jadeite ornament. The northern object, also likely jadeite, was a simple greenstone "chunk" weighing 5.1 g. The two objects were recovered within the uppermost level of ballast directly beneath Floor 7. The surrounding matrix was notably rich in artifacts, and the upper 8 cm of floor ballast

Feature/Context

Table 6.6. Context summary for Unit 1F2. Depth (m) Phase Description/Notes

Humus

0.10–0.20

Sik’u’ II

Humus and upper portion of fill Surrounding Monument 1.

Floor 1

0.20–0.40

Sik’u’ II

Eroded floor and associated fill.

Monument 1

0.00–0.40

Sik’u’ II

Monument fragments and underlying fill.

Floor 2

0.40–0.50

Sik’u’ I

Eroded floor and associated fill.

Floor 3

0.50–0.60

‘Ayim-tun

Partial plaster floor and associated fill.

Floor 4

0.60–0.73

Chukan*

Partial plaster floor and associated fill.

Floor 5

0.73–0.86

Chukan

Partial plaster floor and associated fill.

Floor 6

0.86–0.90

Chukan

Plaster floor and associated fill.

Floor 7

0.90–1.30

Chukan

Plaster floor and associated fill.

Cache 2

ca. 0.92

Chukan

Greenstone cache.

Floor 8

1.30–1.40

Ix Cha’*

Plaster floor, associated fill, and possible disturbance.

Floor 9

1.40–1.60

Ix Cha’

Packed earth floor or buried Ahorizon.

197

produced more than 400 potsherds, an obsidian blade, an obsidian flake, 160 chert flakes, 19 chert cores, and one chert biface. These materials likely formed part of the cache. Cache deposition also likely involved burning of some kind as the surrounding matrix was rich in well-preserved wood charcoal. AMS radiocarbon analysis of wood charcoal from this deposit produced a pooled two-sigma calibrated date range of 403–207 BC (AA72671; see Chapter 10), placing Cache 2 deposition and the construction of Floor 7 within the first half of the Chukan phase. The next floor was encountered at a depth of 1.30 m. Floor 8 was well preserved and consisted of plaster approximately 5 cm thick, although several sections showed signs of erosion or disturbance, possibly related to root action or post-depositional activities just outside the unit’s baulk. Ceramics within Floor 8 fill showed obvious signs of disturbance and included several well-preserved and obviously intrusive Late Classic sherds. Although most of the remaining sherds came from Preclassic wares, few were sufficiently preserved to provide a secure date. A tentative Ix Cha’ phase date was assigned to correspond with Floors Four and Five in the nearby Unit 1F1. Finally, at a depth of 1.40 m a 5 cm thick lens of dark, clayey soil was encountered. This level was interpreted as a packed-earth floor and given the provisional designation of Floor 9. Ceramics recovered from the underlying fill and sandy subsoil included both Mamom and pre-Mamom diagnostics and Floor 9 was assigned an Ix Cha’ phase date.

198

Plaza IV: Unit 1F3 Test excavations in Plaza IV consisted of a single 1-x-1-m unit located 3 m south of the plaza’s northern edge (Figure 6.1). Unit 1F3 was excavated to a depth of 1.80 m and identified three episodes of construction (Figure 6.9). The first floor in the 1F3 sequence was identified at a depth 0.30 m. Floor 1 was completely eroded, but was identifiable by a thin layer of eroded stucco and a clear transition from humus to fill. Ceramics recovered in Floor 1 fill were somewhat mixed, but included several Late Classic diagnostics and the floor was assigned a general Sik’u’ I–II phase date. The second floor was identified at a depth of 0.43 m. Floor 2 was highly eroded, but was originally a plaster construction approximately 4 cm thick. Ceramics from associated fill were slightly mixed, but included numerous Early Classic diagnostics and Floor 2 was assigned an ‘Ayim-tun phase date. The third floor in Unit 1F3 was encountered at a depth of 0.63 m. Floor 3 was an unusually thick plaster floor with two distinct layers of construction. The upper 5 cm of the floor consisted of dense, compacted stucco similar to other floors at Trinidad. Beneath this level, however, Floor 3 consisted of loose, unconsolidated limestone sascab approximately 35 cm thick. Fill beneath the level of sascab was more typical of other floor constructions at Trinidad and consisted of a moderate mixture of gravel and cobbles within a sandy soil matrix that gradually gave way to underlying bedrock. Ceramics from this fill consisted exclusively of Late Preclassic types and Floor 3 was assigned a Chukan phase date.

199

Figure 6.9. North profile of Unit 1F3.

200

Feature/Context

Table 6.7. Context summary for Unit 1F3. Depth (m) Phase Description/Notes

Humus

0.10–0.30

Sik’u’, Säk-tunich

Humus deposits.

Floor 1

0.30–0.43

Sik’u’

Eroded plaster floor and associated fill.

Floor 2

0.43–0.63

‘Ayim-tun

Partially eroded plaster floor and associated fill.

Floor 3

0.63–1.80

Chukan

Thick, well-preserved plaster floor, unconsolidated sascab, And associated fill.

Plaza V: Units 1P1 and 1P2 Investigation of Plaza V consisted of two 1-x-1-m units. The first unit, Unit 1P1, was placed approximately 30 m west of the Group A platform (Figure 6.1). Unit 1P1 was excavated to a depth of 0.80 m and identified two episodes of construction. The first floor was identified at a depth of 0.30 m. Although no evidence for a plaster floor was encountered, a dark soil lens was identified at this depth along with an underlying fill episode slightly lighter in color. Floor 1 likely represents the remains of an entirely eroded plaster floor or packed-earth floor. Ceramics from associated fill were heavily eroded, but included primarily Late Preclassic types. Floor 1 was assigned a Chukan phase date. Floor 2 was identified at a depth of approximately 0.50 m. Soil at this level consisted of unconsolidated limestone sascab with a light fill level beneath. Ceramics from within this fill consisted exclusively of Late Preclassic types and the floor was assigned a Chukan phase date.

201

Feature/Context

Table 6.8. Context summary for Unit 1P1. Depth (m) Phase Description/Notes

Humus

0.10–0.30

Sik’u’, Yaljob’ach, Säk-tunich

Humus deposits.

Floor 1

0.30–0.50

Chukan*

Possible eroded plaster or packed earth floor and associated fill.

Floor 2

0.50–0.80

Chukan

Possible eroded plaster floor and associated fill.

The second 1-x-1-m unit excavated in Plaza V was placed near the center of the plaza and approximately 25 m northeast of Unit 1P1 (Figure 6.1). Unit 1P2 was excavated to a depth of 0.70 m and identified two episodes of construction (Figure 6.10). The first floor in the 1P2 sequence was identified at a depth of 0.30 m. As with Floor 1 in Unit 1P1, Floor 1 in Unit 1P2 consisted of a dark soil lens with lighter colored soil fill beneath. This floor was interpreted as either an eroded plaster construction or a packed earth floor. Ceramics from associated fill were eroded, but included Late Preclassic diagnostics and the floor was assigned a Chukan phase date. As with Unit 1P1 Floor 1, however, this phase designation should be regarded as tentative. The only well-preserved floor in Unit 1P2 was identified at a depth of 0.60 m. Floor 2 was a well-preserved plaster floor with a thickness of approximately 5 cm. Fill beneath Floor 2 consisted of a thin layer of dark, rocky soil placed directly atop bedrock. Materials within associated fill were heavily eroded and no diagnostic types could be identified. Floor 2 was assigned a provisional Chukan phase date, corresponding to Unit 1P1 Floor 2.

202

Figure 6.10. North profile of Unit 1P2.

Feature/Context

Table 6.9. Context summary for Unit 1P2. Depth (m) Phase Description/Notes

Humus

0.10–0.30

Indeterminate

Humus deposits.

Floor 1

0.30–0.50

Chukan*

Possible eroded plaster or packed earth floor and associated fill.

Floor 2

0.60–0.70

Chukan

Plaster floor and associated fill.

Group E: Unit 1E1 Test excavations in Group E consisted of a single 2-x-1-m unit placed 2 m south of Structure E-1 (Figure 6.1). Unit 1E1 was excavated to a depth of 2.55 m where the unit was terminated due to risk of collapse within unconsolidated fill (Figure 6.11). Unit 1E1 documented three episodes of construction and one cache. Additional features in the upper levels of the unit may have been destroyed by looting activity on Structure E-1.

203

Figure 6.11. North profile of Unit 1E1.

The first floor in Unit 1E1 was located at a depth of 0.36 m, beneath a mixed and disturbed humus level, likely including some spoil from looters' excavations in Structure E-1. Floor 1 was almost entirely eroded, but could be identified in the unit profile by a thin layer of eroded plaster and a clear transition from humus to a fill level. Ceramics within Floor 1 fill consisted of a diverse mix of Late Preclassic, Late Classic, and

204

Figure 6.12. Cache 1 following excavation.

Terminal Classic types, including several excellent examples of Fine Orange ware, as well as several possibly intrusive Postclassic types. Floor 1 was assigned a Yaljob’ach phase date, although a slightly later, early Säk-tunich date is possible. The second floor in the 1E1 sequence was identified at a depth of 0.48 m. As with the preceding floor, Floor 2 was poorly preserved and was identified only by a thin layer of eroded plaster within the unit profile. Ceramic materials recovered within Floor 2 fill consisted exclusively of Late Preclassic and Early Classic types and an ‘Ayim-tun phase date was assigned.

205

Excavations within Floor 2 fill also produced a small cache. Cache 1 consisted of two miniature vessels placed within ballast beneath Floor 2 (Figure 6.12). The first vessel, an unslipped, flaring walled bowl (Complete Vessel 9), was found in an upright position, while the second, an unslipped plate (Complete Vessel 10), was found inverted a few cm to the north. The two vessels may have been originally placed in a lip-to-lip arrangement. Soil samples were collected from within the bowl and throughout the surrounding matrix, but no additional artifacts or ecofacts were recovered.

Feature/Context

Table 6.10. Context summary for Unit 1E1. Depth (m) Phase Description/Notes

Humus/Collapse/ Disturbance

0.10–0.36

Sik’u’, Yaljob’ach, Säk-tunich

Mixture of humus, collapse from Str. E-1, and possible looters’ disturbance.

Floor 1

0.36–0.48

Yaljob’ach

Eroded plaster floor and associated fill.

Floor 2

0.48–0.74

‘Ayim-tun

Partially eroded plaster floor and Associated fill.

Cache 1

0.56–0.68

‘Ayim-tun

Miniature vessel cache beneath Floor 2.

Floor 3

0.74–2.55

Chukan

Partially eroded plaster floor and heavy platform fill.

The final floor in the 1E1 sequence was identified at a depth of 0.74 m and consisted of a thin, poorly preserved plaster floor. Fill beneath Floor 3 consisted of a loose mixture of cobble and boulder sized fill stones with very little intervening soil matrix. A loosely conglomerated retaining wall, oriented north–south, was also identified in the western half of the unit. The thickness of this fill (at least 1.8 m), as well as its loosely consolidated nature, suggest that it was likely the principal episode of

206

construction for the Group E platform. Only sparse quantities of ceramics were recovered within Floor 3 fill and most were heavily eroded. The small preserved sherd inventory was dominated by Late Preclassic types, however, and Floor 3 was assigned a Chukan phase date.

RESIDENTIAL GROUP TESTING The residential portion of the site center testing program included test excavations in seven architectural groups: C, G, K, N, O, U, and Y. Short summaries of these excavations are presented below in the alphabetic order of group designation. Unit locations are indicated and described in Figure 6.1 and Table 6.1. Group C: Units 1C1 and 4A13 Stratigraphic excavations in Group C consisted of two units. The first unit was placed 3 m south of Structure C-1 (Figure 6.1). Unit 1C1 was excavated to a depth of 2.39 m and documented five episodes of construction, including several buried structures (Figure 6.13). The first floor in Unit 1C1 was identified at a depth of 0.30 m. Although Floor 1 was entirely eroded, its location was easily identifiable by traces of eroded plaster and a thin ballast level. This floor corresponds to the lowest exterior terrace encountered in excavations of Structure C-1 to the north (see Chapter 9). Ceramics recovered within associated fill were relatively well preserved and included many Early Classic and Late Classic diagnostics including polychromes. As the preserved polychromes dated to the first half of the Late Classic, Floor 1 was assigned a Sik’u’ I phase date.

207

Figure 6.13. North profile of Unit 1C1.

The next floor in the 1C1 sequence was identified at a depth of 0.75 m. Floor 2 was a well-preserved plaster construction approximately 5–8 cm thick. Associated fill included many Late Preclassic types and Floor 2 was assigned a Chukan phase date. Beneath Floor 2, at a depth of approximately 0.95 m, a wall was encountered oriented approximately east–west across the unit. The wall consisted of a single course

208

of roughly cut stones with approximate dimensions of 50-x-30-x-50 cm. The stones were placed end-to-end lengthwise directly atop Floor 3. These walls were interpreted as part of a possible structure and given the provisional designation of Structure C-sub-1. Floor 3 was encountered at a depth of 1.36 m and directly beneath Structure Csub-1. Floor 3 was a well-preserved plaster floor approximately 5 cm thick. Ceramics from the underlying fill included many reliable Late Preclassic diagnostics and Floor 3 and Structure C-sub-1 were assigned Chukan phase dates. Another possible structure was found beneath Floor 3 and atop Floor 4. This structure, designated Structure C-sub-2, consisted of two walls. The first wall was located in the southwestern quadrant, aligned approximately 10 degrees west of north and running from the south baulk diagonally into the west baulk. The second wall was identified in the southeastern quadrant, oriented approximately 35 degrees west of north, and running from the east baulk into the center of the unit. Both walls were two courses high and consisted of roughly shaped cobbles with approximate dimensions of 30-x-15-x20 cm. Materials collected around both walls consisted of a mix of Middle Preclassic and Late Preclassic diagnostics. As the underlying floor produced only Middle Preclassic materials, a tentative Ix Cha’ phase date was assigned. The fourth and final floor in Unit 1C1 was identified at a depth of 1.75 m and directly beneath Structure C-sub-2. Floor 4 was a plaster construction approximately 5 cm thick. Only only a thin fill level, approximately 15 cm deep, separated this floor from the underlying bedrock. Sparse quantities of ceramics were recovered in fill, but

209

Feature/Context

Table 6.11. Context summary for Unit 1C1. Depth (m) Phase Description/Notes

Humus/Collapse

0.10–0.30

Sik’u’, Yaljob’ach, Säk-tunich

Mixed humus and collapse from Structure C-1.

Floor 1

0.30–0.75

Sik’u’ I

Eroded plaster floor and associated fill.

Floor 2

0.75–0.95

Chukan

Plaster floor and associated fill.

Str. C-sub-1

0.95–1.36

Chukan

East-west structural wall situated on top of Floor 3.

Floor 3

1.36–1.75

Chukan

Plaster floor and associated fill.

Str. C-sub-2

1.55–1.75

Ix Cha’

Low, structural walls situated on top of Floor 4.

Floor 4

1.75–1.95

Ix Cha’

Plaster floor and associated fill.

Posthole 1

1.95–2.38

Aj Wo’

Ovoid posthole cut into bedrock beneath Floor 4 fill.

Posthole 2

1.95–2.39

Aj Wo’

Ovoid posthole cut into bedrock beneath Floor 4 fill.

identified types included Tierra Mojada Resist, Mars Orange, Achiotes Unslipped, and other late Middle Preclassic Mamom diagnostics, and Floor 4 was assigned an Ix Cha’ phase date. After Floor 4 fill was excavated down to bedrock, two holes were identified cutting down into bedrock in the eastern half of the unit (Figure 6.14). Both holes were roughly formed and ovoid in shape, with diameters between 35 and 40 cm and depths of approximately 45 cm. These features were tentatively interpreted as postholes (Posthole 1 and Posthole 2). As no corresponding holes were present in the overlying Floor 4, these features were interpreted as predating Floor 4 construction. Excavation of the soil matrix within these features produced small quantities of artifacts, including 30 potsherds, a chert hammerstone, and nine pieces of jute (Pachychilus indiorum) and apple

210

Figure 6.14. Unit 1C1 Postholes 1 and 2 following excavation.

snail (Pomacea flagellata) shell. Ceramics recovered from the postholes were small and eroded, but were sufficiently different from Mamom sphere materials to be assigned to the pre-Mamom Aj Wo’ complex. This early Middle Preclassic phase date places these features among the earliest identified at Trinidad (see also Chapter 8: Units 3C1 and 3C3 for another early construction in Group C), and they presumably constitute the remains of a perishable structure erected at Trinidad during its initial occupation. The second stratigraphic unit in Group C was excavated as part of architectural testing in Structure C-1 (see Chapter 9). Unit 4A13 was a 1-x-1.5-m unit excavated through the upper interior floor of Structure C-1’s central hall (Figure 6.1). Although the

211

Figure 6.15. West profile of Unit 4A13.

212

principal purpose of Unit 4A13 was to provide a date for construction of Structure C-1, it is included here as it revealed significant data concerning the construction history of Group C. Unit 4A13 was excavated through the lower interior floor of Structure C-1's southwestern gallery after the building had been cleared (see Chapter 9). As such, the first floor in the 4A13 sequence, designated Floor 1, was the well-preserved interior floor of Structure C-1 (Figure 6.15). Excavation of this floor revealed it to be an exceptionally hard plaster floor construction approximately 8 cm thick. The underlying ballast level was thick and well laid with tightly spaced gravel. Ceramics recovered within Floor 1 ballast and fill were slightly mixed and included numerous well-preserved Late Preclassic types. The latest materials, however, were Late Classic in date, including several excellent markers for the first half of the Late Classic period. Floor 1 and Structure C-1 were assigned a Sik’u’ I phase date. Beneath Floor 1 fill, at a depth of approximately 1.56 m, a section of plaster floor 75 cm wide was encountered running east–west through the middle of the unit. Excavations to the north and south of this feature encountered a mix of fill and possible collapse. After these areas were cleared, the floor was identified as the upper facing to a 40 cm high bench and the feature was designated Structure C-sub-3. The floor beneath this feature was encountered at a depth of 1.96 m. Though somewhat eroded, Floor 2 was a plaster construction approximately 5 cm thick. Floor 2 was notable in that, though level in the northern half of the unit, it was inclined at an angle of approximately 15 degrees to the south of the bench. This portion of the floor

213

may be part of an exterior molding, similar to that seen in Chukan phase constructions beneath Structure G-1 (see Chapter 9). Fill from within the bench and beneath Floor 2 consisted exclusively of Late Preclassic types and both the floor and bench were assigned Chukan phase dates. The third floor in the 4A13 sequence was encountered directly beneath Floor 2 ballast, at a depth of 2.03 m. Floor 3 was a plaster floor approximately 8 cm thick. The ceramic inventory from Floor 3 fill included many reliable Late Preclassic diagnostic sherds and Floor 3 was assigned a Chukan phase date. The fourth floor in Unit 4A13 was encountered at a depth of 2.65 m. Floor 4 was a well-preserved plaster construction approximately 4 cm thick. Ceramics in associated fill were exceptionally well preserved and likely came from midden redeposited as fill shortly after initial deposition. The ceramic inventory included many Late Preclassic sherds and the floor was assigned a Chukan phase date. An early Chukan date was further supported by AMS radiocarbon analysis of wood charcoal collected from within the upper portion of this fill. A single assay produced a pooled two-sigma calibrated date range of 357–52 BC (AA72668; see Chapter 10). At a depth of approximately 3 m, fill associated with Floor 4 gave way to a darker soil level rich in artifacts. This level was interpreted as a midden (Midden Two) situated directly atop bedrock. Artifacts included approximately 1,200 potsherds, 10 pieces of obsidian, a partial footed metate, a small jadeite pebble, and a modest quantity of bone and shell. Identified faunal specimens included dog (Canis familiaris), squirrel (Sciuridae), olive shell (Oliva sayana), apple snail (Pomacea flagellata), river clam (cf.

214

Unionidae), and two taxa of fish (Osteichthyes and cf. Cichlidae), as well as several piece of unidentified marine shell. This midden was also notably rich in wood charcoal, presumably from hearth cleaning. Ceramics from within the midden included many well-preserved Late Preclassic types, including numerous open fluted Sierra group forms similar to those described below for Midden One. Further, many of the forms in Midden Two were more delicate than those normally found in Chicanel sphere collections, and in other respects appeared transitional with earlier Middle Preclassic forms (see Chapter 10). The ceramic inventory also included numerous Middle Preclassic types and the midden was interpreted as dating to an early facet of the Chukan phase. To test this possibility, a sample of wood charcoal was submitted for AMS radiocarbon analysis. This assay produced a pooled two-sigma calibrated date range of 409–207 BC, and the midden was assigned an early Chukan phase date (AA72667; see Chapter 10).

Feature/Context

Table 6.12. Context summary for Unit 4A13. Depth (m) Phase Description/Notes

Floor 1

1.06–1.56

Sik’u’ I

Structure C-1 interior floor and associated fill.

Str. C-sub-3

1.56–1.96

Chukan

Possible bench feature.

Floor 2

1.96–2.03

Chukan

Partially eroded plaster floor beneath bench feature and associated fill.

Floor 3

2.03–2.65

Chukan

Plaster floor and associated fill.

Floor 4

2.65–3.00

(early) Chukan

Plaster floor and associated fill.

Midden 2

3.00–3.51

(early) Chukan

Moderate density midden atop bedrock.

215

Group G: Units 1G1–1G3 and 5A7 Test excavations in Group G included four units. Two of these units, 1G2 and 1G3, were extensions onto Unit 1G1 and are included below as part of the discussion of that unit. The fourth unit, Unit 5A7, was excavated as part of household excavations in Structure G-1 and is included here as it produced useful data regarding the construction sequence for Group G as a whole. Unit 1G1, the first pit in Group G, was placed approximately 4 m east of Structure G-1 and 3 m south of the northern edge of the Group G platform (Figure 6.1). Unit 1G1 was excavated to a depth of 2.90 m and identified five episodes of construction, two burials, and a high-density midden (Figure 6.16). Units 1G2 and 1G3 were later added to this unit and were excavated only down to the depth of Burial 5. The first floor in Units 1G1–1G3 was identified at a depth of 0.20 m. Floor 1 was completely eroded, but was identifiable by a shift from humus to fill and small pieces of eroded plaster. Ceramics from Floor 1 fill were very mixed. The uppermost lots included numerous Terminal Classic diagnostics, including Fine Orange and Fine Gray wares, but the latest materials in lower lots were Late Classic, including Sik'u' II diagnostics. Floor 1 was assigned a Yaljob'ach date, though the materials from lower lots may indicate a separate Sik'u' II floor. The next floor was identified at a depth of 0.66 m. Floor 2 was a well-preserved plaster floor approximately 8 cm thick. Ceramics within the underlying fill included both Early Classic and Late Classic materials. As no Late Classic polychromes were recovered in fill, Floor 2 was assigned a general Sik’u’ date, although the heavy mixing

216

Figure 6.16. North profile of Unit 1G1.

217

with Early Classic materials is common in deposits dating to the first half of the Late Classic and a Sik’u’ I date appears likely. Floor 3 was identified at a depth of 0.90 m. At this depth, fill gave way to a level of white limestone sascab, the upper 10 cm of which was densely packed and likely constituted an eroded plaster floor. Ceramics recovered within the unconsolidated sascab included exclusively Late Preclassic types and Floor 3 was assigned a Chukan phase date. In composition, this floor was very similar to Floor 3 in Plaza IV (see above). A large number of human bones were also recovered within Floor 3 fill and these were designated Burial 4, although no specific grave could be identified. Skeletal materials included small fragments from the cranium, humerus, ulna, radius, metacarpals, and phalanges, although most of the bones were too eroded to identify (Thornton and Moriarty 2007; Appendix B). Burial 4 likely represents a single, secondary burial placed within Floor 3 fill at the time of its deposition. Burial 4 was assigned a Chukan phase date. Beneath Floor 3’s white sascab fill, a level of fine grayish-yellow soil mixed with virtually no stones was identified at a depth of 1.40 m. This level was identified as fill associated with Burial 5, which was encountered 30 cm into this level (Figure 6.17). Although somewhat eroded, Burial 5 was among the most complete at Trinidad. The individual was placed in the flexed position, facing north, with the head to the west. Based on cranial characteristics, the skeleton’s robustness, heavy wear on molar enamel, and evidence for alveolar resorbtion in the mandible and maxilla, the individual is thought to be an adult male, likely more than thirty-five years old (Thornton and Moriarty

218

Figure 6.17. Trinidad Burial 5.

2007; Appendix B). Grave offerings included a large Sierra Red bowl inverted and placed over the upper torso and cranium (Complete Vessel 2; Figure 6.18) and a drilled freshwater mussel shell (Psoronaias percompressus) ornament recovered near the individual’s head. Small quantities of fish bones (Osteichthyes and Cichlidae) were recovered in soil samples and presumably formed part of the grave offerings.

219

Figure 6.18. Sierra Red basin from Burial 5 (Complete Vessel 2).

Beneath Burial 5 the yellow soil level continued for an additional 20 cm. Within this last lot of yellow soil fill, two large, roughly shaped stone blocks were encountered. Although the association of these stones was uncertain, they were given the provisional designation of Structure G-sub-2 and assigned a Chukan phase date. Beneath these stones, at a depth of approximately 2.50 m, a black, slightly clayey level, rich in artifacts, was encountered. This level continued down to bedrock. This level was interpreted as midden (Midden 1). Midden 1 produced some of the highest artifact densities encountered at Trinidad. Within the 0.40 cubic m portion of the midden

220

Figure 6.19. Partially reconstructible Alta Mira Fluted vessel from Midden 1. Drawing by I.Seyb.

excavated, more than 2,300 sherds were collected (5,755 sherds/m3). Many of these sherds were large and well-preserved and served as type specimens for the Chukan ceramic complex presented in Chapter 10. The midden was particularly rich in deep, fluted bowls classified as Alta Mira Fluted (Figure 6.19; see also Figure 10.7:B1-C2). This particular form, persistent but rare in other Chukan contexts, almost certain constitutes a Chukan phase serving vessel. Other artifacts collected in Midden 1 included several hundred chert artifacts, 48 pieces of obsidian, and more than 200 specimens of bone and shell. Most of the chert artifacts were flakes, but the inventory also included several formal tools and a number of

221

cores. The obsidian inventory included 18 prismatic blade segments, two prismatic core fragments, 10 flakes, and 8 large ovoid scrapers. Ten of these were sourced by XRF and all were found to have come from the San Martín Jilotepeque source (Chapter 11; Cecil et al. 2007). The faunal collection was particularly diverse, and included domesticated dog, white-tailed deer, peccary, rabbit, armadillo, fish, and three species of marine shell (Table 6.13).

Table 6.13. Faunal taxa identified in Midden 1.

Scientific Name

Common Name

NISP

Canis familiaris

Dog

8

Dasypus novemcinctus

Armadillo

4

Odocoileus virginianus

White-tailed deer

5

Tayassu spp.

Peccary

1

Sylvilagus spp.

Rabbit

3

Osteichthyes

Fish

3

Pomacea flagellate

Apple snail

74

Lampsilis spp.

Freshwater mussel

2

Nephronaias spp.

Freshwater mussel

1

Bivalvia (cf. Unionidae)

River clam

42

Donax denticulatus

Donax shell

1

Oliva spp.

Olive shell

2

Strombus spp.

Conch

1

Midden 1 was assigned a Chukan phase date. The presence, however, of large numbers of Middle Preclassic types and a form distribution similar to that seen in Midden 2(See Unit 4A13 above) suggested affiliation to the early facet of the Chukan complex. To test this possibility, two pieces of wood charcoal from this deposit were submitted for AMS radiocarbon analysis. These assays produced two-sigma calibrated date ranges of

222

382–175 BC and 383–180 BC (AA72669 and AA72670; see Chapter 10) and the midden was dated to the first half of the Chukan phase, possibly transitional with the preceding Ix Cha’.

Feature/Context

Table 6.14. Context summary for Units 1G1–1G3. Depth (m) Phase Description/Notes

Humus

0.10–0.20

Sik’u’, Yaljob’ach, Säk-tunich

Mixed humus deposits.

Floor 1

0.20–0.66

Yaljob’ach

Eroded plaster floor and associated fill.

Floor 2

0.66–0.90

Sik’u’ I

Partial plaster floor and associated fill.

Floor 3

0.90–1.40

Chukan

Possible floor and fill level of unconsolidated sascab.

Burial 4

0.90–1.10

Chukan

Secondary burial within Floor 3 fill.

Yellow Fill

1.40–2.50

Chukan

Yellow soil level.

Burial 5

1.70–2.30

Chukan

Burial and associated cache.

Str. G-sub-2

2.30–2.50

Chukan

Possible structural wall.

Midden 1

2.50–2.90

(early) Chukan

High-density midden atop bedrock.

The second test excavation in Group G consisted of a 1.2-x-1-m unit excavated through the ultimate levels of construction in Structure G-1. The first levels of this unit, Unit 5A7, are reported as part of household excavations in Chapter 9. The lower levels of this unit, however, provided much useful information regarding the Group G sequence (Figure 6.20). The first floor encountered beneath Structure G-1 was identified at a depth of 0.98 m. This floor was designated Floor 3, following its position with respect to floors within

223

Figure 6.20. North profile of Unit 5A7.

Structure G-1. This and the subsequent floor (Floor 4) correlate to Structure G-sub-1, a construction partially delineated in household excavations (see Chapter 9). Floor 3 was a dense plaster construction approximately 15 cm thick. Floor 3 had only a thin associated fill level as the next floor was found directly beneath it. Floor 3 was assigned a tentative Chukan phase date by stratigraphic context (see below).

224

Floor 4, encountered at a depth of 1.14 m, was a plaster construction approximately 5 cm thick. Ceramics from associated fill consisted exclusively of Late Preclassic types and Floor 4 was assigned a Chukan phase date. At a depth of approximately 1.55 m, fill gave way to a thick level of unconsolidated sascab. This level was interpreted as a poorly preserved floor and designated Floor 5. This level continued down to a depth of approximately 2.00 m where a level of very dark gray fill, with only a few large cobbles, was encountered. Further, a small round (ca. 30 cm diameter) posthole (Posthole 1) was identified cutting through Floor 5 down to fill. The ceramics mixed with the unconsolidated sascab, the dark fill level, and the posthole were Late Preclassic and all features were assigned Chukan phase dates. Floor 5 almost certainly relates to Floor 3 in Units 1G1–1G3 (see above).

Feature/Context

Table 6.15. Context summary for Unit 5A7. Depth (m) Phase Description/Notes

Floor 3

0.98–1.14

Chukan

Plaster floor.

Floor 4

1.14–1.55

Chukan

Plaster floor and associated fill.

Floor 5

1.55–2.00

Chukan

Plaster floor and fill level of unconsolidated sascab.

Posthole 1/Fill

1.57–2.52

Chukan

Dark soil level and posthole.

Str. G-sub-3

2.52–2.66

Chukan

Large cut stone block wall or possible collapse.

Beneath Floor 5 and associated deposits, five large, roughly cut blocks were found in the northern half of the unit. Because the stones were clearly worked and still maintained traces of stucco on their facing edges, they were interpreted as part of a

225

possible wall or collapse from an earlier structure. They were given the provisional designation of Structure G-sub-3, although they may be part of Structure G-sub-2 encountered at the same depth in Units 1G1–1G3. These features were identified at a depth of 2.52 to 2.66 m where the unit was terminated due to time constraints. Ceramics from the matrix surrounding these features were exclusively Late Preclassic in date and Structure G-sub-3 was assigned a Chukan phase date. Group K: Unit 1K1 Stratigraphic excavations in Group K consisted of a single 1-x-1-m unit placed 3 m south of Structure K-1 (Figure 6.1). Unit 1K1 was excavated to a depth of 1.30 m below the modern surface and documented three episodes of construction (Figure 6.21). Additional features near the modern surface may have been destroyed or concealed by looting activities on Structure K-1. The first floor in Unit 1K1 was identified at a depth of 0.30 m, beneath a level of mixed humus, collapse, and looters’ disturbance. Floor 1 was entirely eroded, but was identified by an abrupt shift to fill at this depth, as well as a marked increase in artifact density. Floor 1 may have constituted the ultimate floor construction in the Group K plazuela, although looters’ activities may have destroyed later constructions. Ceramics recovered within associated fill were slightly mixed, but included many reliable Late Classic diagnostics, though few polychromes, and the floor was assigned a general Sik’u’ I–II phase date. Approximately 20 cm beneath Floor 1 and at a depth of 0.50 m, a new, heavy fill level was encountered, much lighter in color and marked by a completely different

226

Figure 6.21. East profile of Unit 1K1.

ceramic inventory. Although no floor was identified at the start of this level, Floor 2 was presumably entirely eroded prior to the construction of Floor 1. The thickness and heavy nature of this fill episode also clearly marked it as the principal leveling episode for construction of the Group K platform. Ceramics within associated fill consisted exclusively of Late Preclassic types and this construction was assigned a Chukan date. The third floor in Unit 1K1 was identified at a depth of 1.16 m. Floor 3 was well preserved and consisted of plaster approximately 5 cm thick. Fill beneath Floor 3 was exceedingly thin and situated directly atop bedrock. Ceramics recovered within fill were

227

Feature/Context Humus/Collapse/ Disturbance Floor 1

Table 6.16. Context summary for Unit 1K1. Depth (m) Phase Description/Notes 0.00–0.30

0.30–0.50

Sik’u’ I–II

Mixture of humus, collapse from Structure K-1, and possible looters’ disturbance.

Sik’u’ I–II

Eroded plaster floor and associated fill.

Floor 2

0.50–1.16

Chukan

Possible eroded floor and Platform fill.

Floor 3

1.16–1.30

Chukan

Plaster floor and associated fill.

very eroded, but included Late Preclassic diagnostics and the floor was assigned a Chukan phase date. Group N: Unit 1N1 Stratigraphic investigations in Group N consisted of a single 1-x-1-m unit placed on the southwestern end of Structure N-1 (Figure 6.1). Unit 1N1 was excavated to a depth of 0.80 m below the modern surface and documented three episodes of construction (Figure 6.22). The first floor in Unit 1N1 was identified 0.10 m below the surface. Floor 1 was completely eroded, but its approximate position could be identified by the presence of ballast and traces of eroded plaster. Ceramics from associated fill consisted of a relatively even mix of general Sik’u’ and ‘Ayim-tun diagnostics. Although the absence of well-preserved polychromes precluded assignment of a Sik’u’ I or Sik’u’ II date, the liberal mixture of Early Classic and Late Classic materials indicated a Sik’u’ I phase date. The second floor in the 1N1 sequence, Floor 2, was encountered at a depth of 0.40 m. Although Floor 2 was largely eroded, it was identifiable by a thin layer of eroded

228

Figure 6.22. East profile of Unit 1N1.

plaster. Fill from Floor 2 produced some of the best pure ‘Ayim-tun ceramics recovered at Trinidad and its construction can be reliably dated to that period. More than 330 potsherds were recovered within the 30 cm of fill beneath Floor 2 and many of these were very large, well-preserved specimens pertaining to the Aguila, Balanza, Caribal, and Quintal groups. The large size and well-preserved nature of ceramics strongly suggests that Floor 2 fill consisted of redeposited primary midden. Beneath Floor 2 fill, a third episode of construction was identified at a depth of approximately 0.70 m. Although no plaster floor could be identified, the largest fill stones from the preceding level were situated atop a lighter and sandier fill level that appeared to represent the remains of a low platform constructed atop bedrock. Materials

229

within Floor 3 fill were heavily eroded, but consisted exclusively of Preclassic wares. As insufficient diagnostic types could be identified, a general Ix Cha’–Chukan date was assigned.

Feature/Context

Table 6.17. Context summary for Unit 1N1. Depth (m) Phase Description/Notes

Humus/Collapse

0.10–0.20

Floor 1

0.20–0.40

Sik’u’ I–II

Humus mixed with collapse from Str. N-1.

Sik’u’ I

Eroded plaster floor and associated fill.

Floor 2

0.40–0.70

‘Ayim-tun

Possible eroded plaster floor and heavy fill mixed with redeposited midden.

Floor 3

0.70–0.90

Ix Cha’–Chukan

Possible eroded plaster floor and light fill over bedrock.

Group O: Units 1O1, 6A13, and 6D9 Stratigraphic excavations in Group O consisted of three 1-x-1-m units. The first unit, Unit 1O1, was placed 2 m south of Structure O-1 and within the Group O plazuela (Figure 6.1). Unit 1O1 was excavated to a depth of 1.88 m and documented eight episodes of construction (Figure 6.23). The first floor in Unit 1O1 was encountered at a depth of 0.20 m. Floor 1 was entirely eroded, but was identified by traces of ballast and fill, as well as by clear differences in ceramics between this level and those below. Ceramics from Floor 1 fill were heavily eroded, but included many good Late Classic diagnostics and Floor 1 was assigned a general Sik’u’ I–II phase date.

230

Figure 6.23. West profile of Unit 1O1.

The second floor in the 1O1 sequence was identified at a depth of 0.30 m. Floor 2 was a partially preserved plaster floor approximately 2 cm thick. Floor 2 was situated directly atop the next floor in the sequence and was interpreted as an episode of renovation associated with the lower floor. No fill was encountered between these floors

231

and a Chukan phase date could be assigned only on the basis of fill associated with the underlying floor. The third floor, located directly beneath Floor 3 at a depth of 0.32 m, was a wellpreserved plaster floor approximately 5 cm thick. Ceramics recovered in associated fill were heavily eroded, but included several Late Preclassic diagnostics and Floor 3 was assigned a Chukan phase date. The next floor was encountered at a depth of 0.54 m. Floor 4 was encountered only in the northwestern and southwestern quadrants of the unit and it was unclear if the floor had originally extended throughout, or if it represented part of another class of construction. Ceramics from associated fill included primarily Late Preclassic diagnostics and Floor 4 was assigned a Chukan phase date. The fifth floor in Unit 1O1 was identified at a depth of 0.74 m. Floor 5 was a well-preserved plaster floor with a thickness of approximately 15 cm. Unconsolidated sascab was found beneath this floor continuing down an additional 30 to 50 cm. This level was in turn situated atop a dense, heavy fill level that likely constituted the principal construction event for the Group O platform. Ceramics from fill were slightly eroded, but included reliable Late Preclassic diagnostics and the floor was assigned a Chukan phase date. Floors 6 and 7, both plaster floors, were encountered at a depth of 1.41 m. As with Floors Two and Three, these floors were laid directly atop one another. The upper floor was approximately 3 cm thick, while the lower floor was better preserved and approximately 5 cm thick. Floor 6 likely constituted a renovation of the slightly earlier

232

Floor 7. No fill was identified between the floors, but fill beneath Floor 7 included Late Preclassic diagnostics and both floors were assigned Chukan phase dates. The final floor in the 1O1 sequence was identified at a depth of 1.59 m. Floor 8 was a plaster construction, approximately 5 cm thick and intact throughout the unit with the exception of an 11-cm-wide posthole cut through the floor in the southeastern quadrant of the unit. Floor 8 fill and the posthole were excavated separately, but produced very small quantities of artifacts. Both features were assigned provisional Ix Cha’–Chukan dates, although similarities with other early constructions at Trinidad suggest a Middle Preclassic date is most likely.

Feature/Context

Table 6.18. Context summary for Unit 1O1. Depth (m) Phase Description/Notes

Humus

0.10–0.20

Sik’u’

Mixed humus deposits.

Floor 1

0.20–0.30

Sik’u’

Eroded plaster floor and associated fill.

Floor 2

0.30–0.32

Chukan

Partially preserved plaster floor.

Floor 3

0.32–0.54

Chukan

Plaster floor and associated fill.

Floor 4

0.54–0.74

Chukan

Possible plaster floor or other construction related feature.

Floor 5

0.74–1.41

Chukan

Plaster floor and associated fill.

Floor 6

1.41–1.44

Chukan

Plaster floor.

Floor 7

1.44–1.59

Chukan

Plaster floor and associated fill.

Floor 8

1.59–1.88

Ix Cha’–Chukan

Plaster floor and associated fill.

Posthole 1

1.59–1.86

Ix Cha’–Chukan

Posthole cut through Floor 8.

The second test unit in Group O was excavated as part of household excavations in Structure O-1 (Lawton 2007a). Unit 6A13 was a 1-x-1-m unit excavated through the

233

collapse of Structure O-1. Unit 6A13 was excavated to a depth of 2.64 m and documented three episodes of construction and one buried structure (Lawton 2007a). The first floor in the 6A13 sequence consisted of an upper floor inside Structure O-1. This floor likely pertains to a late phase of construction in Structure O-1, although almost certainly not its latest. Ceramics from fill included both Late Preclassic and Early Classic diagnostics, and Floor 1 was assigned an ‘Ayim-tun phase date. Beneath Floor 1 fill, at a depth of approximately 1.30 m, a feature from an earlier structure was identified in the northern half of the unit. This feature consisted of a line of roughly cut stones running east–west across the unit with thick stucco preserved on their upper surfaces. After fill to the south was cleared, this feature was found to be 60 cm high and situated directly atop Floor 2. This feature was interpreted as a bench from a buried structure (Structure O-sub-1). Fill from within the bench consisted exclusively of Late Preclassic types and it was assigned a Chukan phase date. The second floor in Unit 6A13 was identified at a depth of 1.90 m. Floor 2 was a plaster construction approximately 5 cm thick. Very few ceramics were recovered in associated fill, but all identified sherds were Late Preclassic and Floor 2 was assigned a Chukan phase date. The final floor in Unit 6A13 was identified at a depth of 2.14 m. Floor 3 was a plaster floor approximately 5 cm thick. Fill underlying Floor 3 was situated directly atop bedrock, making this construction the earliest in the Structure O-1 sequence. Ceramics from associated fill, however, were highly eroded and included a fairly even mix of

234

Middle Preclassic and Late Preclassic diagnostics. Floor 3 was assigned an early Chukan date.

Feature/Context

Table 6.19. Context summary for Unit 6A13. Depth (m) Phase Description/Notes

Collapse

0.34–0.44

Sik’u’, Säk-tunich

Highly mixed collapse and possible fill within Str. O-1.

Floor 1

0.44–1.30

‘Ayim-tun

Possible eroded interior floor and associated fill.

Str. O-sub-1

1.30–1.90

Chukan

Possible bench feature.

Floor 2

1.90–2.14

Chukan

Plaster floor and associated fill.

Floor 3

2.14–2.62

(early) Chukan

Plaster floor and associated fill.

The third stratigraphic unit in Group O was excavated as part of household excavations in Structure O-3 (Lawton 2007a). Unit 6D9 was a 1-x-1-m unit placed along the western edge of Structure O-3. Unit 6D9 was excavated to a depth of 2.62 m and documented two episodes of construction, a midden, and a burial. The uppermost level of Unit 6D9 consisted of a mix of humus and collapse from Structure O-3 as well as a low-density midden (Midden 33) likely placed along the western edge of Structure O-3. The inventory of materials associated with the midden and the possible floor included approximately 280 potsherds, 50 chert artifacts, 4 obsidian bladelets, and one complete figurine. Ceramics within this deposit were slightly mixed, but included many diagnostic Late Classic types, including several polychrome sherds dating to the first half of the Late Classic. Cumulatively, this deposit was assigned a Sik’u’ I phase date, although some mixing with later phases may have occurred.

235

Beneath the midden, a fill level was identified at a depth of approximately 1.45 m. Although no clear floor was identified, the fill likely corresponded to an earlier eroded construction designated Floor 2. Ceramics within Floor 2 fill included Late Preclassic diagnostics and it was assigned a Chukan phase date. Between approximately 1.45 and 1.64 m, a burial was identified within Floor 2 fill. Burial 6 was highly eroded and consisted of a small number of poorly preserved human remains including fragments of the cranium, left humerus, right radius, right and left femora, right and left tibiae, and a fibula (Thornton and Moriarty 2007). No offerings were recovered and no burial architecture was identified during excavation, but a simple cyst may have been difficult to identify within fill. Ceramics from surrounding fill included primarily Late Preclassic types. But, from its limited depth, it appears most likely that Burial 6 was intruded into Floor 2 fill during the construction of a later, now eroded floor or with deposition of the overlying midden. A Sik’u’ I phase date was assigned.

Feature/Context

Table 6.20. Context summary for Unit 6D9. Depth (m) Phase Description/Notes

Humus/Collapse

0.65–0.77

Midden 33

0.77–1.45

Sik’u’ I–II

Mixed humus deposits.

Sik’u’ I

Low density midden.

Floor 2

1.45–2.34

Chukan

Possible floor and associated fill.

Burial 6

1.45–1.64

Sik’u’ I

Intrusive burial within Floor 2 fill.

Humus/Sheet Midden

2.34–2.62

Ix Cha’–Chukan

Buried humus level over bedrock mixed with low density midden.

236

Beneath Burial 6 and Floor 2, a buried humus level was identified at a depth of 2.32 m. Ceramics and other artifacts were recovered in relatively large quantities within this level and these were likely deposited as sheet midden. Ceramics included many Middle Preclassic and Late Preclassic types and a general Ix Cha’–Chukan phase date was assigned. Group U: Unit 1U1 Stratigraphic excavations in Group U consisted of a single 1-x-1-m unit placed 2 meters west of Structure U-2 (Figure 6.1). Unit 1U1 was excavated to a depth of 1.40 m below the modern surface and documented four episodes of construction (Figure 6.24). The first floor in Unit 1U1 was encountered at a depth of 0.10 m, beneath a thin layer of humus and collapse from Structure U-2. Floor 1 was almost entirely eroded, but was marked by a level of eroded plaster and underlying ballast. Ceramics from associated fill were somewhat mixed with Preclassic wares, but included many reliable Late Classic diagnostics and Floor 1 was assigned a general Sik’u’ I–II date. At a depth of 0.40 m, the light fill associated with Floor 1 gave way to a much denser fill episode mixed with a lighter soil matrix. Although no evidence for a floor construction was visible, the lower fill episode was interpreted as a separate construction and designated Floor 2. Ceramics within Floor 2 fill were highly mixed. Upper lots included both Late Preclassic and Late Classic diagnostics, while lower lots consisted exclusively of Preclassic wares. The upper Late Classic wares were treated as possibly intrusive and Floor 2 was assigned a Chukan phase date.

237

Figure 6.24. East profile of Unit 1U1.

Beneath Floor 2 fill, a second major shift in fill was identified at a depth of 0.70 m. At this point, the moderate to heavy fill beneath Floor 2 transitioned to a much heavier fill characterized by a loose conglomerate of massive cobble-to-boulder-sized fill stones mixed with very little soil matrix. This construction, designated Floor 3, appeared to constitute the principal construction phase for the Group U platform. Ceramics

238

recovered with Floor 3 fill consisted exclusively of Mamom Middle Preclassic types and it was assigned an Ix Cha’ phase date.

Feature/Context

Table 6.21. Context summary for Unit 1U1. Depth (m) Phase Description/Notes

Humus/Collapse

0.10–0.20

Sik’u’

Humus and collapse from Str. U-2.

Floor 1

0.20–0.40

Sik’u’

Eroded plaster floor and associated fill.

Floor 2

0.40–0.70

Chukan

Possible plaster floor and associated fill.

Floor 3

0.70–1.20

Ix Cha’

Heavy platform fill and possible eroded floor.

Floor 4

1.20–1.40

Aj Wo’–Ix Cha’

Packed earth floor and associated fill/subsoil.

Beneath Floor 3 platform fill and at a depth of 1.20 m, a 5-cm thick, dark soil level was identified. This feature was interpreted as a packed-earth floor similar to that found at the same depth within Unit 1F2 (see above), and given the designation of Floor 4. Fill beneath Floor 4 consisted of a sandy subsoil and a light mixture of gravel that gradually transitioned to bedrock. Ceramics recovered in association with Floor 4 included both Mamom and pre-Mamom diagnostics and the floor was assigned a general Aj Wo’–Ix Cha’ phase designation, although it likely represents the early portion of the Ix Cha’. Group Y: Unit 1Y1 Stratigraphic excavations in Group Y consisted of a single 1-x-1-m unit placed 2 m east of the southern arm of Structure Y-1 (Figure 6.1). Unit 1Y1 was excavated to a

239

Figure 6.25. East profile of Unit 1Y1.

depth of 0.40 m below the modern surface where bedrock was encountered. Only a single episode of construction was identified in this unit (Figure 6.25). Floor 1 was identified at a depth of 0.10 m below the modern surface and beneath a thin layer of humus and collapse from Structure Y-1. Floor 1 was entirely eroded and only a thin layer of ballast and eroded plaster remained to mark the probable location of the floor in the unit profile. Ceramics from associated fill were primarily Late Classic in date, although insufficient polychromes were recovered to establish a Sik’u’ I or Sik’u’ II affiliation. Bedrock was encountered directly beneath Floor 1 fill suggesting that Group Y, in contrast to many of the other residential groups at Trinidad, was built in single episode dating to the Late Classic period. The presence of Yaljob’ach and Säk-tunich diagnostics

240

within the humus zone, however, suggests that the group was occupied well into the Terminal Classic and Postclassic periods.

Feature/Context

Table 6.22. Context summary for Unit 1Y1. Depth (m) Phase Description/Notes

Humus/Collapse

0.10–0.20

Sik’u’, Yaljob’ach, Säk-tunich

Humus mixed with collapse from Str. Y-1.

Floor 1

0.20–0.50

Sik’u’

Partially eroded plaster floor and associated fill.

SUMMARY Although a more detailed synthesis of culture-historical data is presented in Chapter 12, the results of site center stratigraphic testing nonetheless provide an overview of long-term developments at Trinidad. Table 6.23 provides a group-by-group graphic summary of site center stratigraphic data. Light gray blocks indicate phases for which an occupation can be inferred by the presence of moderate quantities of ceramics mixed in later deposits. Medium gray blocks indicate phases for which at least a single episode of construction was identified (e.g., a floor or a structure). Dark gray blocks indicate phases for which two or more discrete episodes of construction were identified. These data highlight long-term developments in Trinidad’s site center and demonstrate two clear peaks in construction and occupation. All but one of the tested groups included episodes of construction dating to the Late Preclassic Chukan, and all but two were found to feature Late Classic Sik’u’ I–II constructions. These results were closely echoed in other programs of research, particularly midden testing (see Chapter 8),

241

Group Y

Group U

Group O

Group N

Group K

Group G

Group C

Group E

Plaza V

Group A

Plaza IV

Group F

Phase

Plaza II

Plaza I

Table 6.23. Summary results of site center stratigraphic testing.

Aj B'oj Säk-tunich Yaljob'ach Sik'u' I–II ‘Ayim-tun P'ich 'Ayim Chukan Ix Cha' Aj Wo'

in which Chukan and Sik’u’ I–II assemblages dominated collections. The high relative frequency of Chukan and Sik’u’ I–II deposits in stratigraphic testing, as well as the dominance of associated materials in other programs of research, suggest that Trinidad’s two principal peaks occurred during the Late Preclassic Chukan and Late Classic Sik’u’ I–II phases. Trinidad’s initial Chukan peak involved the most significant labor investments. Units in Groups C, G, K, O, and the ballcourt (Group F) all documented multiple superimposed Late Preclassic constructions. In some instances, the tempo of

242

construction appears to have been high, with three or more discrete construction episodes all dating to the Chukan phase. Further, fill episodes associated with the Chukan tend to be among the most substantial at the site, and many of the large basal platforms that characterize Trinidad’s site center were raised principally during this phase (e.g., Groups C, E, and G). Although the tempo of construction does not appear to have been quite as great during the Sik’u’ I–II phases, it is during this interval that Trinidad took on its final form. Sik’u’ I–II constructions were the latest major efforts in all but Groups E and G. In those groups where later constructions were identified, they tended to be ephemeral or additive to Late Classic constructions. Further, it should be noted that, though Sik’u’ I–II constructions appear more limited in scale than those of the Chukan, Sik’u’ I–II materials were recovered in virtually all investigations at Trinidad, and this phase of occupation likely eclipsed the Chukan in terms of total population. Both the late Middle Preclassic Ix Cha’ and Early Classic ‘Ayim-tun phases were only slightly less well represented in construction than the Chukan or Sik’u’ I–II. Seven groups included ‘Ayim-tun constructions and six featured Ix Cha’ constructions. Yet, the scale of these constructions were often slightly less than those of the Chukan or Sik’u’ I– II phases. Floors dated to the Ix Cha’ tended to be thinner with shallower fills than subsequent Chukan deposits. Likewise, ‘Ayim-tun constructions tended to be much less significant than either the earlier Chukan or later Sik’u’ I–II. The limited nature of constructions dating to these phases was also reflected in other programs of research. Although Ix Cha’ and ‘Ayim-tun phase deposits were recovered in harbor and midden

243

testing, associated materials were recovered in much more modest quantities than those associated with either of Trinidad’s two peaks. Table 6.23 also demonstrates the much more limited scale of constructions dating to the Terminal Preclassic P’ich ‘Ayim, Terminal Classic Yaljob’ach, Postclassic Säktunich, and historical Aj B’oj phases. The lack of evidence for P’ich ‘Ayim construction is particularly interesting in that midden testing identified several rich middens dating to this phase. The absence of contemporary constructions may be a result of sampling bias, a real downturn in construction, or a complex temporal relationship between the Chukan, P’ich ‘Ayim, and 'Ayim-tun ceramic complexes. The limited nature of Yaljob’ach construction is also belied by the presence of several rich middens and minor renovations to existing Sik’u’ I–II constructions (see Chapter 9). The absence of Postclassic and historical constructions is less surprising. Although Säk-tunich phase materials are found in a thin veneer throughout Trinidad, the focus of the site’s occupation had shifted to the harbor area by the Postclassic. Aj B’oj deposits were identified in only a handful of discrete loci, and Trinidad appears to have had only a minor occupation during the historical period. With the exception of the historical Aj B’oj, the early Middle Preclassic Aj Wo’ was the most limited in distribution. Materials dating to Trinidad’s earliest ceramicutilizing occupation were identified only in the deepest units, frequently directly atop bedrock, and often mixed with materials dating to the subsequent Ix Cha’. Additional deposits may also have been missed in units that were not excavated to bedrock. Aj Wo’ materials were recovered only within a tightly circumscribed area of Trinidad's site

244

Table 6.24. Floor sequences in Trinidad’s public architectural complexes.

Phase

Plaza I 1I1

Plaza II 1B1

Group F 1F1 1F2

Plaza IV 1F3

Group A 1A1

Group E 1E1

Plaza V 1P1 1P2

Aj B’oj

Säk-tunich

Yaljob'ach Fl. 1 Fl. 1 Sik’u’ I–II

Fl. 1

Fl. 1

Fl. 1

Fl. 2

Fl. 1 Fl. 1 Fl. 2

‘Ayim-tun

Fl. 2

Fl. 2

Fl. 2

Fl. 3

Fl. 2

Fl. 2

P'ich 'Ayim

Chukan

Fl. 3

Fl. 3

Fl. 3

Fl. 4 Fl. 5

Fl. 3

Fl. 3

Fl. 6 Fl. 7 Ix Cha'

Aj Wo'

Fl. 4

Fl. 4

Fl. 4

Fl. 8

Fl. 5

Fl. 5

Fl. 9

Fl. 4

Fl. 3

Fl. 1

Fl. 1

Fl. 2

Fl. 2

245

center, bounded by Groups C, F, and U, and Plazas I–II. Although the greater volume of later architecture within this area may have helped preserve these early deposits, it appears equally likely that Trinidad’s initial ceramic occupation was restricted geographically to what later became the site’s central-most precincts. The second major result of the site center testing program was the identification of clear temporal patterning in the development of public architectural complexes at Trinidad (Table 6.24). This was most evident in Trinidad’s central precincts. Testing in Plazas I–II and the ballcourt identified a sequence of four sequential construction episodes at the same approximate depths below the modern surface. The correlation of floor sequences in these areas indicates that the central-most space at Trinidad was constructed in four principal episodes starting in the Ix Cha’ Middle Preclassic, with subsequent additions in the Chukan, ‘Ayim-tun, and Sik’u’ I–II phases. Plaza IV, just to the east, appears to have been a slightly later addition, with construction beginning in the Late Preclassic and then continuing apace. Groups A and E both appear to have only slightly variant histories of construction, with starting points in the Ix Cha’ and Chukan, respectively. Although excavations in Plaza V revealed only Chukan phase floors, it was maintained as a discrete open space for the remainder of the site’s occupation. Further, middens along its western edge contained materials dating to all periods (see Chapter 8), suggesting the possibility that this plaza was utilized in later periods or that later floors, correlating to those found in other areas, were simply too eroded to identify.

246

The longevity of public architectural complexes within Trinidad’s site center, as well as within most residential groups, suggests a high degree of continuity in the site’s occupation. Further, though more detailed architectural testing would be necessary to thoroughly assess the development of monumental architecture, it appears likely that the basic form of Trinidad’s central precinct was maintained from the late Middle Preclassic onward. Although the organization of structures around this space undoubtedly shifted with time, the adherence to this basic template implies at least some continuity in perceptions and, perhaps, use of public space over time at Trinidad.

247

CHAPTER SEVEN THE HARBOR AREA TESTING PROGRAM

INTRODUCTION The principal role of any port, as noted in the research design, is to transfer goods and personnel between modes of transportation. In most ports this transfer is effected within a harbor. Testing the identification of a harbor at Trinidad and determining the extent and organization of its construction were, therefore, among the principal goals of research. Secondary objectives included describing the construction history for these features and assessing their specific functions. In total, two field seasons were conducted in the harbor area and 20 units 1-x-1 m or larger were excavated (Table 7.1; Figure 7.1). The following sections provide a brief review of harbor area excavations, a unit-by-unit description of excavations, and a detailed summary of excavation results, describing long-term developments in the harbor area.

METHODS AND ORGANIZATION Investigations in the harbor area were conducted over the course of two field seasons during which the sampling strategies and methods of investigation shifted slightly. The first season, in 2003, was largely exploratory. Trinidad’s harbor consists of an amalgam of cultural and modified natural features that, upon initial assessment, were

248

Figure 7.1. Trinidad harbor area with excavations indicated. Unit 2C1 and Platform JJ, discussed below, are located approximately 60 m north of Structure FF-1.

somewhat bewildering. The slope leading down to the harbor area the northern edge of the fault system that underlies the central Petén lakes system and is marked by a series of natural terraces that are difficult to differentiate from cultural features (see Chapters 3 and 5). Further, many of the major harbor features are of such large scale that they would, if found to be entirely manmade, constitute among the most significant constructions at Trinidad (see Figures 5.16 and 5.17). Finally, to a much greater extent than any other part of Trinidad, the harbor area has been heavily impacted by post-

249

depositional processes. Changes in lake level, as well as historical and modern use of the area as a harbor, have all contributed to the erosion and modification of ancient Maya features. Thus, the goal of the 2003 season was to determine which features constituted ancient Maya constructions and which were natural landforms.

Table 7.1. Location and dimensions of harbor area excavations.

Harbor Feature Platform EE

Unit(s)

Dimensions (m)

Location

2A4

1-x-1

12A112A5 2B2

10-x-2

2E1-2E5

8.5-x-1

Platform GG

2A2 2A5 2A3 12B1

1-x-1 1-x-1 2-x-1 2-x-2

East end of Platform GG. East end of Platform GG. Center of Platform GG. Center of Platform GG.

Area A

12A6

2-x-2

2A1

1-x-1

Northern edge of Area A, adjacent to Platform EE. Center of Area A.

Platform HH

2B1

1-x-1

Center of Platform HH.

Platform JJ

2C1

1-x-1

Northeast of Platform JJ.

Structure EE-1

1-x-1

Center of southwestern extent of Platform EE. Transverse trench cut through south edge of Platform EE. Eastern extent of Platform EE. Transverse trench cut through Str. EE-1.

The 2003 field season consisted of small excavation units, typically 1-x-1 m, in a broad sample of the harbor’s possible features. Units were placed on all of the possible platforms in the lower portion of the harbor: Platforms EE, FF, GG, and HH. More intensive follow-up excavations focused on Platform EE, the low platform that likely served as a loading platform and formed the inner wall of the aquatic (now dry) portion of the harbor (Area A). Other excavations focused on Platform GG, the massive 70-m-

250

long feature that likely served as the harbor breakwater. Finally, an 8.5-m-long trench was excavated through Structure EE-1, a possible dock attached to Platform EE and within Area A. The 2005 season of research in the harbor area was more limited in time and scale, largely as a consequence of protracted negotiations with the local landowner. Investigations, therefore, focused on the two most intriguing and possibly important features in the harbor area: Platforms EE and GG. A 12-x-2-m-long trench (Units 12A112A6) was excavated on the primary axis of Platform EE and into the northern edge of Area A. In addition, a 2-x-2-m unit was excavated at the approximate midpoint of Platform GG’s long axis. Excavation methods throughout these investigations included the use of arbitrary 10-cm lots and all other standard practices discussed previously and described in Appendix A. Unit locations are provided in Figure 7.1, burials and caches are described in Appendix B, and context summaries are attached as tables. The conclusion to this chapter presents a brief reconstruction of the harbor’s development and use over time.

ORGANIZATION OF EXCAVATION SUMMARIES Unit descriptions, provided below, are divided into six sections. The first four sections deal with investigations in the harbor area’s four principal features: Platform EE, Area A, Structure EE-1, and Platform GG. As these were thought to be the key features within the harbor, they were the most heavily investigated. The last section provides brief summaries of the more limited investigations on Platform HH and near Platform JJ.

251

Unit descriptions focus on the most important features encountered during these excavations. As with the preceding chapter, floor designations are utilized to separate discrete episodes of construction, whether or not preserved floors were present.

INVESTIGATION OF PLATFORM EE The first feature tested in harbor area investigations was Platform EE, the ca. 40x-40-m feature situated at the base of the slope leading down to the lake and forming the inner wall of the aquatic portion of the harbor. This feature is among the most regular in the harbor area and was obviously one of its major components. This platform was wellsituated to have served as a loading and unloading platform for canoes arriving at or departing from the Trinidad harbor, and its upper surface would have provided extensive space for organizational activities. Unit 2A4 Investigation of Platform EE included seven excavation units. The first, Unit 2A4, was a 1-x-1-m unit placed near the south-central portion of the platform’s welldefined western end, approximately 3 m north of its southern edge (Figure 7.1). This unit was subsequently expanded to dimensions of 2-x-1 m to fully expose Burial 3. Unit 2A4 was excavated to bedrock at 1.60 m (Figure 7.2). Excavation confirmed that Platform EE was an artificial feature, documented three discrete episodes of construction, and recovered a burial. The uppermost floor in Unit 2A4 was identified at a depth of 0.30 m. Floor 1 was heavily eroded and consisted solely of scattered fragments of eroded plaster at the

252

Figure 7.2. North profile of Unit 2A4.

transition from humus to a light fill episode. Only modest quantities of ceramics were recovered within the light fill associated with this floor, but many diagnostic Postclassic sherds were identified, and Floor 1 was assigned a general Säk-tunich-phase date. Subsequent excavations to the east (see below: Unit 12A1), however, suggested that an early Säk-tunich date is very likely.

253

The next floor was encountered at a depth of 0.60 m. Floor 2 was a wellpreserved plaster floor approximately 5 cm thick. Fill included a thin ballast layer mixed with a dense, clayey matrix. The underlying fill consisted of a mixture of cobble-sized fill stones and a silty soil matrix similar to Classic-period fills in the site center (see Chapter 6). Ceramics within ballast and fill included both Early Classic and Late Classic monochromes and unslipped materials, but very few polychromes. Floor 2 was initially assigned a general Sik’u’ I–II date; however, the presence of clear Sik’u’ II diagnostics within Burial 3 beneath this floor (see below) provided a secure Sik’u’ II-phase date. At a depth of approximately 0.90 m, the cobble-sized fill associated with Floor 2 gave way to a much heavier fill consisting of cobble- and boulder-sized fill stones mixed with little soil matrix. In composition, this fill was very similar to heavy Preclassic platform fills within the site center (see Chapter 6). The ceramic content of this fill episode also differed significantly from overlying deposits and included only Late Preclassic types. This episode of construction was interpreted as the fill for a Chukanphase construction that was either destroyed or eroded prior to the construction of the Floor 2 platform. This level was given the designation of Floor 3 and dated to the Chukan phase. Within Floor 3 fill, a burial was also encountered. Burial 3 was a formal grave burial, bounded by roughly shaped vertical slab walls and capped by a large rectangular block. The body was oriented north-south in the lateral flexed position, facing east with the cranium to the north. Skeletal materials were heavily eroded, likely as a result of periodic inundation and drying, but all skeletal regions were represented by at least a few

254

Figure 7.3. Eroded Sik'u' II polychrome vessel from Burial 3 (Complete Vessel 4). Note "kill hole" on vessel bottom. Drawing by I.Seyb.

preserved specimens. Cranial, dental, and other skeletal characteristics indicated the individual was likely an adult male middle-aged or older. Evidence for extensive tabular shaping (following Comas 1960), similar to that seen in Burial 7 (see Chapter 8 and

255

Appendix B), was also observed on the individual’s frontal bone (Thornton and Moriarty 2007). Grave offerings included two complete vessels. The first, an eroded polychrome tripod plate with flared walls (Complete Vessel 3), was found inverted over the cranium and upper torso. The second vessel, was an eroded polychrome bowl with flared walls (Complete Vessel 4; Figure 7.3), was found inverted, placed just east of the cranium, and beneath the plate. Both vessels were terminated, the small bowl with a drilled “kill hole” in its base, and the plate by having its tripod supports knocked off. A complete obsidian prismatic blade from the El Chayal source was recovered between the individual’s right tibia and fibula. Small quantities of fish bones, including several species of catfish (e.g., B .marinus) and tropical gar (A. tropicus) were also recovered in float samples collected around the cranium. Although some of these specimens may have been introduced with lakeshore sediments, they may also have formed part of the burial offering.

Feature/Context

Table 7.2. Context summary for Unit 2A4. Depth (m) Phase Description/Notes

Humus

0.10–0.30

Säk-tunich

Mix of humus, fill, and possible collapse.

Floor 1

0.30–0.60

early Säk-tunich

Eroded plaster floor and associated fill.

Floor 2

0.60–0.90

Sik’u’ II

Plaster floor and associated fill.

Burial 3

0.90–1.30

Sik’u’ II

Formal grave burial and associated cache.

Floor 3

0.90–1.60

Chukan

Heavy fill episode and possible Floor.

256

The slips on associated cache vessels were too eroded to identify to the type level, but the forms for both were diagnostic for the latter half of the Late Classic period and a Sik’u’ II date was assigned. Burial 3 was presumably inserted into Floor 3 fill during the construction of Floor 2. Floor 2 was also found to slope downward to the south directly above the location of Burial 3. This slope may have resulted from fill subsidence within Burial 3. Units 12A1-12A5 The second set of excavations on Platform EE consisted of five 2-x-2-m units (12A1–12A5). These units were placed in a line running north to south from the upper platform surface down its sloping southern side into Area A (Figure 7.1). A sixth unit (12A6) was excavated at the south end of this line, but is discussed below in the section dealing with Area A. The purpose of these units was to supplement chronological data collected in Unit 2A4 and to assess the architecture of Platform EE. These units were located near the midpoint of the platform’s well-defined southern edge. The selected area appeared from the surface to be slightly disturbed by modern activity, but factors in unit placement included fresh milpa, newly planted banana trees, and other modern features that had to be avoided. Further, due to time constraints, only Unit 12A1, the northernmost unit, could be excavated to a significant depth. Excavations in other units were shallow and designed to follow the uppermost features identified in Unit 12A1. Unit 12A1 was excavated to a depth of 1.90 m and encountered five episodes of construction, a buried structure, and a level of activity-related deposits (Figures 7.4 and 7.5). The uppermost surface lot of this unit, as well as those of Units 12A2-12A5 (see

257

Figure 7.4. North-south section of Unit 12A1. See Figure 7.5 for additional features.

below), was among the most mixed and heavily disturbed contexts at Trinidad. Modern and historical usage of the harbor area undoubtedly accounted for some of this disturbance. Surface deposits produced an artifact mixture including Postclassic ceramics, historical artifacts like colonial glass, square-cut nails, and a small-caliber

Figure 7.5. North-south section of Units 12A1 to 12A6. Terraces 1 to 4 and Structure EE-sub-1 indicated.

258

259

musket ball, and modern nails, glass, and rum bottle tops. The historical artifacts may have been associated with a late floor or activity surface that could not be identified within such disturbed deposits. Just beneath the surface, however, a clearer sequence of constructions could be identified. The first floor in Unit 12A1 was encountered at a depth of approximately 0.23 m. Floor 1 was entirely eroded, but was identified by the concentration of small bits of eroded plaster at this depth. These pieces of plaster were visible throughout the unit, but terminated in the southern half of the unit near a possible wall associated with Structure EE-sub-1 (see Units 12A2-12A5 below). Beneath the floor only a light fill episode was identified. This was heavily mixed with a clayey soil matrix likely gathered from the lake or swampy deposits to the west. Large quantities of Postclassic ceramics were recovered in fill and a Säk-tunich phase date was assigned. Further, the prevalence of Paxcamán Red among monochrome sherds and the near absence of Augustine group types provided a late Säk-tunich phase date. The second floor in Unit 12A1 was encountered in the western half of the unit at a depth of 0.60 m. Floor 2 was heavily eroded and its depth could be estimated only by the presence of flecks of eroded plaster. A light ballast level was also encountered directly beneath 0.60 m. The ballast, however, quickly gave way to a much denser fill, consisting of large cobble- and small boulder-sized stones. Many of the stones recovered within this fill appeared to have been roughly shaped and likely came from a collapsed structure directly east of the floor (Structure EE-sub-1; see below). Ceramics in fill were highly mixed and included types pertaining to all periods from the Late Preclassic through

260

Postclassic. The earlier materials were interpreted as inclusions within the collapse reutilized as fill and a Postclassic Säk-tunich phase date was assigned. Further, as the type Augustine Red dominated the Postclassic assemblage, an early Säk-tunich phase date was assigned. Beneath Floor 2 fill and at a depth of approximately 0.94 m, small sections of plaster were identified at various points within the unit. The presence of an eroded floor (Floor 3) at this depth was indicated both by the presence of this plaster and better preserved sections of floor at the same depth in Unit 12A2 just to the south. This part of Floor 3 was likely partially destroyed by erosion and one of several large ceiba tree roots that cut through the excavated area. Fill beneath Floor 3 consisted of a dark silty soil matrix mixed with numerous gravel- to cobble-sized stones. Some degree of mixing between fill associated with Floors Two and Three was evident in the ceramic content of Floor 3 fill. Some upper fill lots included Terminal Classic and Postclassic materials, while lower lots produced consistently Late Classic assemblages. The mixing in upper lots almost certainly resulted from the bioturbation discussed above, and a Late Classic date was assigned based on lower lots. Further, the presence of good Sik’u’ II diagnostics within the lower fill dated construction to the second half of the Late Classic period. The date for Floor 3 also provided a date for a 75-cm-wide wall running northsouth through the eastern half of Unit 12A1 and originally encountered at a depth of approximately 0.40 m (Figure 7.6). Designated Wall 1, this feature consisted of three courses of unshaped stones stacked to a height of approximately 50 cm, oriented

261

Figure 7.6. Wall 1 and Floor 3 during excavation of Unit 12A1.

approximately five degrees west of magnetic north-south, and running from the north baulk of Unit 12A1 into Unit 12A2 to the south (see below). This wall was seated atop Floor 3 and was assigned a Sik’u’ II phase date. Part of a similarly constructed wall was also visible within the western baulk of the unit. The function of these walls, even following excavation of Units 12A2–12A5, is unclear. Their width, however, is suggestive of a foundation-brace structure of at least moderate height as seen elsewhere at Trinidad (see Chapter 9). Beneath Floor 3 fill, at a depth of approximately 1.30 m, a dark brownish gray clayey soil level, rich in all manner of artifacts and sand-sized white inclusions, was

262

encountered. Stones were notably absent from this level. This level was interpreted as midden (TRI Midden 34) resulting from activities atop Floor 4 (see below). Alternatively, this level may have resulted from a short-term rise in lake level, with the midden being produced by the same processes as the much better documented Midden 17 in Area A (below). This deposit continued to a depth of 1.60 m where Floor 4 was identified. The upper half of the midden level (Midden 34a) included a mix of ceramic types, with the Late Classic Sik’u’ II particularly well represented, but also included a few Terminal Classic and Postclassic sherds. The Yaljob'ach and Säk-tunich materials were interpreted as intrusive, resulting from the bioturbation noted above. The lower half of the midden (Midden 34b) contained ceramics dating to no later than the first half of the Late Classic period. The upper portion of the midden was assigned at Sik’u’ II phase date, while the lower lot was assigned a Sik’u’ I date. Floor 4 was encountered at an approximate depth of 1.60 m. Floor 4 was entirely eroded, but was identifiable by a rapid shift from the dark midden soil described above to a moderately dense fill deposit mixed within a light grayish brown silty soil matrix. Materials within this fill included a mixture of Early Classic and Late Classic monochromes, and Floor 4 construction was assigned a Sik’u’ I date. This floor presumably eroded during the period of heavy use or inundation indicated by the overlying midden. Finally, beneath Floor 4 fill, at a depth of approximately 1.80 m, a new light fill episode was encountered. This episode consisted of a dull yellow silty soil matrix mixed with light quantities of gravel and cobbles. No floor was encountered in association with

263

Feature/Context

Table 7.3. Context summary for Unit 12A1. Depth (m) Phase Description/Notes

Humus

0.10–0.23

Säk-tunich – Aj B’oj

Mixed humus deposits.

Floor 1

0.23–0.60

(late) Säk-tunich

Eroded floor and associated fill.

Floor 2

0.60-0.94

(early) Säk-tunich

Partially preserved plaster floor and associated fill.

Wall 1/ Str. EE-sub-1

0.44–0.94

Sik’u’ II

Wall 1 feature situated atop Floor 3.

Floor 3

0.94–1.30

Sik’u’ II

Partially preserved plaster floor and associated ill.

Midden 34a

1.30–1.40

Sik’u’ II

Midden or activity area.

Midden 34b

1.40–1.60

Sik’u’ I

Midden or activity area atop eroded floor.

Floor 4

1.60–1.80

Sik’u’ I

Eroded plaster floor and associated fill.

Floor 5

1.80–1.90

Chukan– P’ich ‘Ayim

Light fill episode for low, possible platform construction.

this level, but the level was interpreted as fill for a separate, earlier platform and the designation of Floor 5 was assigned. Materials within Floor 5 fill included many Late Preclassic types, along with a handful of possible Early Classic or Terminal Preclassic polychromes. Although this construction most likely dated to the Late Preclassic Chukan phase, the inclusion of possible Classic period markers suggests a transitional P’ich ‘Ayim date is possible. Following excavation of Unit 12A1, four 2-x-2-m units were placed in a line running south-southwest from Unit 12A1’s southern baulk (Figures 7.1 and 7.5). Due to time constraints, these units could not be excavated to bedrock. Instead, excavations followed the upper surface of features first identified in Unit 12A1 in association with Floor 3. These features were also visible from the surface as a series of terraces leading

264

down from the upper surface of Platform EE into Area A. Following excavation of Units 12A1-12A5, these features were found to be the principal construction episode for Platform EE and were designated Structure EE-sub-1. The following section provides brief descriptions for the recovery contexts in Units 12A2-12A5, as well a short summary of Structure EE-sub-1’s appearance following excavation. Within all four units, the uppermost level consisted of a dense mixture of very dark grayish brown clayey humus and collapse, presumably from Structure EE-sub-1 or later constructions. Floors One and Two from Unit 12A1 could not be identified in these units. Either no corresponding floors were constructed on the slope leading down to Area A, or they were destroyed as a result of erosion or other post-depositional processes (see above). Collapse consisted of a dense mixture of small gravel- and cobble-sized fill stones, as well as larger, roughly worked stone blocks. Ceramics and other artifacts were recovered in large numbers throughout this level. Ceramics were very mixed and included types from all periods from the Late Preclassic through to the Postclassic. Postclassic materials were recovered in particularly large numbers, and most lots were assigned general Säk-tunich phase dates. Some of the uppermost lots, however, included square-cut nails or other historical artifacts and were assigned Aj B’oj phase dates. Beneath this mixed level of humus and collapse, a series of four terraces was identified. The uppermost terrace (Terrace 1) consisted of a 2.2 m long southern extension of Floor 3 from Unit 12A1’s southern baulk into Unit 12A2 and 12A3. This terrace likely constitutes the uppermost surface of Platform EE during Sik'u' II times (see below). Plaster was visible at an approximate depth of 0.94 m throughout Unit 12A2.

265

Structure EE-sub-1 Wall 1 (described above) continued atop this level of floor approximately 1.4 m south of Unit 12A1’s southern baulk. The southern end of the wall was marked by a large rectangular cut-stone block. This level of terrace ended just to the south at a pile of roughly cut stone blocks that led downward to a lower terrace (Terrace 2). These stones likely constitute fill exposed by the collapse of the terrace facing stones. Although excavations terminated just below the surface of Terrace 1, its position with respect to Floor 3 in Unit 12A1 provided a Sik'u' II date. The next terrace (Terrace 2) was identified just to the south and approximately 60 cm below Terrace 1. Eroded plaster was encountered at an approximate depth of 1.60 m, but the terrace floor appeared to be otherwise completely eroded. The cut stones that presumably formed the terrace face between Terrace 1 and Terrace 2 were encountered within the collapse just above the approximate depth of the terrace floor. A second wall, designated Wall 2 from Structure EE-sub-2, was also identified atop this level of terrace exactly oriented with Wall 1 above, though slightly to the west of that feature.. Although Wall 2 was only partially explored, it may constitute a southwestern extension of Wall 1 or a slightly earlier construction in the same location. As excavations terminated just below the badly eroded terrace floor and Wall 2, no secure date can be provided for either feature. If, however, as appears likely, Terrace 2 was built as part of the same episode of construction as Structure EE-sub-1 and Terrace 1, then it dates to Sik’u’ II times. Although Terrace 2 may constitute the exposed portion of an earlier platform built at the Platform EE locus, the close similarity in the size of

266

Terraces 1 to 4 suggest a single episode of construction. As the uppermost feature dates to Sik'u' II times, a provisional Sik'u' II date is assigned to other terraces as well. The third terrace (Terrace 3) was identified just to the south and approximately 60 cm lower than Terrace 2. The facing stones between Terrace 2 and Terrace 3 were still largely in place, and consisted of roughly cut rectangular blocks. The position of Terrace 3’s floor was indicated only by a gradual transition to fill at an approximate depth of 2.10 m. If this terrace originally included a plaster surface, it was likely eroded by wave action or alternating periods of inundation and drying. Materials associated with this terrace were also exceptionally mixed, suggestive of various post-depositional processes and no secure date can be provided. If this feature was constructed contemporaneous with the overlying terraces, a Sik’u’ II date would be inferred. The fourth terrace (Terrace 4) was identified just to the south of Terrace 3. The facing between Terrace 3 and Terrace 4 was similar to that between Terraces 2 and 3, with the same approximate height of 60 cm. The southern edge of the terrace was poorly defined, but was likely the lowest edge of the platform above Area A. Although the terrace floor was completely eroded, a transition to fill was identified between 2.80 and 3.00 m. Further, the humus and collapse situated on top of the terrace was extremely mixed, likely resulting from wave action and other post-depositional processes. The date of this terrace likely corresponds to that of Terrace 3. Finally, during the clearing of these features, a thin (1-cm thick or less) level of artifact- and charcoal-rich dark brown clayey silt was encountered in various loci directly atop the approximate level of floors on Terraces One through Four. Artifacts recovered

267

Table 7.4. Faunal taxa identified in Midden 35.

Scientific Name

Common Name

NISP

Mazama spp.

Brocket deer

2

Odocoileus virginianus

White-tailed deer

7

Dasypus novemcinctus

Armadillo

13

Tayassu spp.

Peccary

1

Agouti paca

Paca

1

Didelphis spp.

Opossoum

1

Aves

Bird

2

Testudines

Turtle

38

Dermatemys mawii

Giant river turtle

3

Trachemys scripta

Slider turtle

1

Staurotypus triporcatus

Giant musk turtle

1

Rhinoclemmys areolata

Furrowed wood turtle

1

Unionidae

Freshwater mussel

1

Lampsilis spp.

Freshwater mussel

2

Pomacea flagellata

Apple snail

22

within these rich lenses included 990 potsherds, 32 pieces of obsidian, 249 chert artifacts, a partial metate, and more than 130 specimens of bone and shell (see Table 7.4). This level was interpreted as an activity-related deposit and designated Midden 35. Artifacts recovered from this lens presumably constitute the detritus of activities occurring on the terraces themselves and near Structure EE-sub-1. These activities also likely produced the high-density middens situated at the base of Platform EE within Area A (see Unit 12A6 and Midden 17 below). Although the ceramic collection from these deposits was insufficiently large to assess the function of the ceramic assemblage, other artifact assemblages demonstrated some interesting patterning. The faunal collection was particularly notable for the large number of turtle remains, while the chert assemblage

268

was among the largest recovered in the harbor area, and marked by a relatively high (240x-3) ratio of flakes to finished tools. At least some of these materials presumably resulted from the processing of faunal resources adjacent to the lake, while others may have resulted from spillage during portage operations.

Feature/Context

Table 7.5. Context summary for Units 12A2–12A5. Depth (m) Phase Description/Notes

Humus/Collapse

Various

Säk-tunich– Aj B’oj

Midden 35

Various

Sik’u’ II, Yaljob’ach, (early) Säk-tunich

Mixed humus and collapse deposits. directly atop Terraces 1-4.

0.44-0.94

Sik’u’ II

Wall feature situated atop Terrace 1.

ca. 0.94

Sik’u’ II

Partially preserved plastered terrace.

1.35-1.60

Sik’u’ II

Wall feature situated atop Terrace 2.

Terrace 2

ca. 1.60

Sik’u’ II*

Eroded floor on terrace.

Terrace 3

ca. 2.10

Sik’u’ II*

Eroded floor on terrace.

Terrace 4

ca. 2.80

Sik’u’ II*

Eroded floor on terrace.

Wall 1/ Str. EE-sub-1 Terrace 1 Wall 2/ Str. EE-sub-2

Ceramics within these deposits were very mixed, but included only types dating to the Sik’u’ II Late Classic or later. The Terminal Classic Yaljob’ach was represented by several clear fine paste sherds, and the early Säk-tunich was identified by the high incidence of Augustine group types. A similar mix of materials was found in the highest density portion of the midden (Midden 17) found at the base of Platform EE and within Area A (see below). This interval, from the Sik’u’ II Late Classic through to the early Säk-tunich Postclassic, appears to have marked the high-point in usage of the harbor area.

269

Unit 2B2 The third excavation on Platform EE consisted of a single 1-x-1-m unit (Unit 2B2) placed on the platform’s eastern extension, approximately 40 m east-southeast of Unit 12A1 (Figure 7.1). Unit 2B2 was placed in this location to determine if the eastern end of Platform EE was an artificial construction and, if so, to correlate its construction to the more heavily investigated western end of the of the platform. Unit 2B2 was excavated to a depth of 1.65 m and documented four floor constructions and one buried structure or retaining wall (Figure 7.7). The uppermost level in Unit 2B2 was a dark brown humus level rich in artifacts. No floor was encountered, but the density of artifacts suggested an eroded floor or possible living surface. Ceramics from this level were very eroded, but consisted exclusively of Postclassic types. Several of the forms present, however, appeared slightly variant from Säk-tunich complex norms, suggesting the possibility of mixing with later Aj B’oj materials, and a transitional Säk-tunich-Aj B’oj phase date was assigned. The first floor in Unit 2B2 was identified at a depth of 0.30 m. Floor 1 was completely eroded and was identifiable only by a light fill level extending from this depth down to the better preserved Floor 2. No traces of preserved plaster were noted. Ceramics within Floor 1 fill were eroded and slightly mixed, but included many reliable Postclassic diagnostics. Further, the type Paxcamán Red dominated monochrome collections, suggesting a late Säk-tunich phase date. The next floor in the 2B2 sequence was identified at a depth of 0.37 m. Floor 2 was partially preserved and was identified by a thin layer of eroded plaster over a light

270

Figure 7.7. North profile of Unit 2B2.

ballast level. Beneath ballast, fill was composed of a sparse mix of cobble-sized fill stones and a brownish black, silty soil matrix, likely collected from nearby lake deposits. Ceramics from Floor 2 fill were heavily mixed and included numerous Classic and Preclassic diagnostics. Postclassic types dominated the collection, however, with a

271

particularly high portion coming from the Augustine and Trapeche groups, and an early Säk-tunich phase date was assigned. At a depth of 1.10 m a level of white, powdery soil was encountered. This layer was interpreted as the remains of an eroded plaster floor and designated Floor 3. Fill beneath this layer contained only small quantities of stones and was probably a rough earthen fill. Although only a handful of potsherds were recovered within this light fill, all dated to the Late Classic period. Floor 3 was assigned a general Sik’u’ I–II date. At a depth of 1.30 m the light fill associated with Floor 3 continued, but with a much higher artifact density, a slight increase in the size and number of stones, and a notable increase in clay content. Although no evidence for a plaster floor was identified, a chronological shift in ceramics (see below) and a clear change in the soil matrix suggested that this level constituted a separate construction and was designated Floor 4. Further, the edge of a low wall, oriented approximately west-northwest to east-southeast was encountered in the southern half of the unit. Only three aligned stones one coarse high were identified in the limited excavation area, but all appeared to be worked and were interpreted as facing stones. This feature was designated Str. EE-sub-3, although it likely constitutes an exterior retaining wall for the platform fill described above. In contrast to previous levels, excavations in Floor 4 fill and along the edge of Structure EEsub-3 produced copious quantities of well-preserved ceramics. All of the identifiable materials were Late Preclassic and both Floor 4 and Structure EE-sub-3 were assigned Chukan phase dates.

272

Feature/Context

Table 7.6. Context summary for Unit 2B2. Depth (m) Phase Description/Notes

Humus

0.10-0.30

Säk-tunich– Aj B’oj

Mixed humus deposits.

Floor 1

0.30-0.37

(late) Säk-tunich

Completely eroded floor and associated fill.

Floor 2

0.37-1.10

(early) Säk-tunich

Partially eroded plaster floor and light fill deposit.

Floor 3

1.10-1.30

Sik’u’ I–II

Partially eroded floor and associated fill.

Floor 4

1.30-1.60

Chukan

Possible floor and associated fill.

Str. EE-sub-3

1.38-1.65

Chukan

Buried structure or retaining wall.

Platform EE Summary Investigation of Platform EE identified four discrete episodes of construction and one modification or repair. The earliest of the construction episodes dated to the Late Preclassic Chukan phase. Evidence for this episode was identified under both the eastern and western ends of the platform. To the east, excavations revealed a 35-cm-high platform with a possible cut-stone retaining wall along its southern, lake-facing side (Structure EE-sub-3). A similar, 70-cm high construction, marked by large platform fill, was identified in Unit 2A4 (Floor 3). A lighter level of fill was also identified in an area later modified in Unit 12A1 (Floor 5). There is no evidence that these constructions constituted part of a single coherent platform, but all were situated at the same approximate depth atop bedrock with no evidence for antecedent construction, and a single episode appears likely. No secure construction episodes were identified for the subsequent P’ich ‘Ayim or ‘Ayim-tun phases, but a minor modification was identified for the Sik’u’ I Late

273

Classic. This modification was identified securely only in Unit 12A1. Excavations there located a 20-cm thick floor fill (Unit 12A1 Floor 4) placed directly atop the Chukan phase construction (Unit 12A1 Floor 5). This construction appears to have been a minor modification of the existing platform, possibly built as part of a repair to the lower platform. The Sik’u’ I Late Classic also appears to have been accompanied by extensive usage of the platform. Midden (Midden 34) dating to the Sik’u’ I and early Sik’u’ II phases collected atop this level of construction. During the subsequent Sik’u’ II phase, Platform EE underwent a dramatic transformation. Construction dating to this interval was most clearly visible in the western end of the platform (Units 2A4 and 12A1–12A5), where excavations identified a 36-cm episode of floor fill overlying the existing platform. In Unit 2A4, an elaborate grave burial (Burial 3) accompanied platform construction. A similar floor also appears to have been placed over the eastern extent of the platform, though its dating is less certain. Cumulatively, these constructions raised Platform EE height to ca. 1.0 m above the bedrock surface and approximately 1.5 m above current water levels. The Sik’u’ II platform also likely featured four terraces leading to the waterfront. These were identified in Units 12A2–12A5, but presumably extended the length of its southern side. The uppermost terrace was only ca. 75-cm wide, while the lower terraces were approximately 1.20 m in width. All of the terraces were separated by ca. 60-cm high facings constructed with roughly cut stones. Refuse accumulated along the tops of these terraces (Midden 35) included materials dating to the Sik’u’ II, Yaljob’ach, and

274

early Säk-tunich phases. These steps presumably provided both easy access to the aquatic portion of the harbor and served as an area for open-air activities. Finally, at least one large superstructure was built atop the Sik’u’ II platform. Structure EE-sub-1 was identified only in Units 12A1–12A2. Evidence for this structure consisted of a 75-cm-thick wall running north-south across the upper surface of the Sik’u’ II platform. Additional, possibly associated walls were identified in the western baulk and just to the south (Structure EE-sub-2). The wall was constructed in a manner very similar to that of the interior north-south spine wall of Structure C-1 within the site center (Chapter 9), although with less regularly shaped facing stones. This wall presumably formed the base layer for a high foundation-brace structure situated atop Platform EE. No specific function can be suggested for this structure, though its proximity to the water is intriguing. The final two episodes of construction date to the Säk-tunich phase. The first of these dated to the early Säk-tunich phase and consisted of an approximately 35-cm-thick floor level constructed directly atop the Sik’u’ II platform. This construction was most visible in the platform’s western extent, particularly in Unit 12A1 where collapse and/-or other debris from Structure EE-sub-1 were re-utilized as platform fill. To the east, construction fill was much lighter, but deeper, raising the platform ca. 75 cm. The final episode of construction for Platform EE consisted of a minor modification made during the late Säk-tunich phase, possibly extending into the early Aj B’oj. Light fill levels dating to the Late Postclassic or early Historical periods were identified near the surface in two of the three excavation loci (Units 12A1 and 2B2).

275

Although floors associated with this fill were entirely eroded, they presumably constituted the ultimate modification of Platform EE.

INVESTIGATIONS IN AREA A The second zone investigated as part of harbor area testing was the low lying area between Platform EE and Platform GG referred to as Area A (Figure 7.1). Although currently slightly above lake level, local informants indicate that this area was inundated up until very recently and was utilized as the aquatic portion of the harbor during the twentieth century Petén chicle boom. As noted in Chapter 3, the record of water level fluctuations for Lake Petén Itzá is only partially understood; the great depth of the lake prevents fine-grained assessments of lake level through palaeolimnology for any one particular point in time (e.g., Anselmetti et al. 2006). Nevertheless, it is likely that this area was underwater during various phases of Trinidad’s occupation and would have provided an ideal, protected access to Platform EE and other harbor facilities. Excavations in Area A were designed to examine the stratigraphic and depositional history of this area and, if possible, to identify phases during which it was underwater and utilized as part of Trinidad’s harbor. Unit 12A6 Some of the most detailed data regarding Area A came from Unit 12A6, the southernmost of six aligned units excavated from the top of Platform EE south into Area A (see Units 12A1–12A5 above). This unit is included here as it provided data primarily of use in assessing Area A. Unit 12A6 was excavated to a depth of 2.30 m and identified

276

Figure 7.8. North-south section of Unit 12A6. See Figure 7.5 for associated units.

an episode of collapse as well as a rich, stratified midden mixed with dense lacustrine deposits (Figure 7.8). The first two levels in Unit 12A6 consisted of a mixture of humus, collapse, and midden. Collapse included numerous moderate to large stones that presumably fell from the various constructions on Platform EE and identified in Units 12A1–12A5 (see above). Ceramics and other artifacts were recovered in tremendous quantities throughout these

277

levels, suggesting at least some degree of mixing with the underlying midden (see below). Artifacts consisted of approximately 900 potsherds, seven obsidian bladelets, and more than 200 faunal specimens, including white-tailed deer (O.virginianus), brocket deer (Mazama spp.), opossum (Didelphis spp.), armadillo (D.novemcinctus), rabbit (Sylvilagus spp.), peccary (Tayassu spp.), various taxa of fish (Cichlidae), and three species of turtle (D. mawii, S.triporcatus, and Emydidae spp.). The ceramic inventory included numerous Postclassic diagnostics, particularly Paxcamán and Pozo group sherds, and this level was assigned a late Säk-tunich phase date. Beneath collapse, and beginning at a depth of approximately 1.25 m, a new level of dark brown to gray silty clay, with very few stone inclusions, was encountered. Ceramics and other artifacts were recovered in moderate to high densities throughout this level, which was interpreted as midden (TRI Midden 17). Excavations determined that this midden continued to a depth of at least 2.30 m, where the unit was terminated due to steady flooding from the water table. Ceramic analyses and two AMS radiocarbon assays determined that this midden was stratified and contained deposits dating to all phases from the P’ich ‘Ayim Terminal Preclassic up to the early Säk-tunich Postclassic. Artifact densities showed only slight signs of diminishing when groundwater flooding forced the termination of this unit, and it appears likely that deeper excavations would reveal earlier midden components. The uppermost level of midden (Midden 17a), from approximately 1.25 to 1.60 m, included primarily Postclassic materials. Further, the types present included a relatively even mix of Augustine, Trapeche, and Paxcamán group types suggesting that

278

this portion of the midden dated to the early part of the Säk-tunich phase. This date was later confirmed by AMS radiocarbon analysis of brocket deer bone (Mazama spp.) from the middle lot of this level, which produced a calibrated two-sigma date range of AD 1035–1218 (AA72662; see Chapter 10).

Table 7.7. Sherd densities for Midden 17. Depth (m)

Midden Level

1.25-1.60

3

Sherd Count

Density (sherds/m3)

Phase

Vol. (m )

17a

early Säk-tunich

1.40

849

606.4

1.60-1.70

17b

Yaljob’ach

0.40

721

1,802.5

1.70-1.90

17c

Sik’u’ II

0.80

1,573

1,966.3

1.90-2.00

17d

Sik’u’ I

0.20

281

1,405.0

2.00-2.20

17e

‘Ayim-tun

0.40

512

1,280.0

2.20-2.30

17f

P’ich ‘Ayim–‘Ayim-tun

0.20

60

300

Immediately beneath this level, from approximately 1.60 to 1.70 m, the ceramic inventory underwent a dramatic shift. Although Postclassic types were still present in small quantities, the majority of the sherds dated to the general Sik’u’ I–II and Yaljob’ach phases. Terminal Classic markers included several Fine Orange and Fine Gray ware sherds, including one large piece from a Fine Orange ware jar. This level (Midden 17b) was assigned a Yaljob’ach phase date, although the presence of Postclassic materials suggests some degree of mixing or a transitional Terminal Classic to Postclassic date. This level also produced some evidence for burning in association with midden deposition. Patches of wood charcoal and ash were found throughout the level. From approximately 1.70 to 1.90 m, the ceramic content of the midden shifted again, this time towards Sik’u’ II Late Classic materials. This level (Midden 17c) was

279

the densest part of the midden sequence and produced more than 1,500 sherds (see Table 7.7). This portion of the midden also included large quantities of wood charcoal, presumably indicative of extensive burning contemporaneous with midden deposition. AMS radiocarbon analysis of wood charcoal from this level produced a calibrated twosigma date of AD 670 – 891 (AA72660; see Chapter 10).

Table 7.8. Faunal taxa identified in Midden 17.

Scientific Name

Common Name

NISP

Mazama spp.

Brocket deer

2

Odocoileus virginianus

White-tailed deer

5

Dasypus novemcinctus

Armadillo

1

Tayassu spp.

Peccary

1

Agouti/Dasyprocta

Paca/Agouti

3

Aves

Bird

3

Dermatemys mawii

Giant river turtle

2

Trachemys scripta

Slider turtle

1

Cheloniidae

Sea turtle

1

Bivalvia (cf. Unionidae)

River clam

7

Pomacea flagellate

Apple snail

49

Decapoda

Crab

1

Atrocosteus tropicus

Tropical gar

1

Osteichthyes

Fish

13

From 1.90 to 2.00 m, artifacts were recovered more sparsely and the soil matrix shifted to a light gray brown color, transitional between the overlying and underlying levels. This level (Midden 17d) also contained a much lower charcoal content than the preceding level. Ceramics within this level (Midden 17d) dated to the Sik’u’ I Late Classic.

280

Table 7.9. Excavation volume and artifact summaries for Midden 17. Artifact Class

Count by Midden 17c 17d Sik’u’ II Sik’u’ I

17a Säk-tunich

17b Yaljob’ach

0.80

0.40

0.80

0.20

0.40

17f P’ich ‘Ayim 0.20

Ceramics Count (f) Weight (Kg) Weight/Count Count/m3 Weight/m3 Unique Rims

1,021 6.780 0.0066 1,276.25 8.48 85

721 5.780 0.0080 1,802.50 14.45 59

1,573 19.617 0.0125 1,966.25 24.52 122

281 3.721 0.0132 1,405.00 18.61 31

512 5.384 0.0105 1,280.00 13.46 51

60 0.961 0.0160 300.00 4.81 5

Chert (Total) Flakes Cores Formal Tools Hammerstones

80 76 2 -2

39 35 3 1 --

120 110 5 4 1

19 17 2 ---

40 38 -2 1

0 -----

Obsidian (Total) Prismatic Bladelets

9 9

3 3

7 7

2 2

0 --

0 --

Groundstone (Total) Manos Metates

2 2 --

1 -1

2 2 --

0 ---

0 ---

0 ---

Ceramic Artifacts Figurines Sherd Disks Notched Sherds Spindle Whorls

6 -2 3 1

1 1 ----

6 4 2 ---

0 -----

1 -1 ---

0 -----

Phase Excavation Vol. (m3)

17e ‘Ayim-tun

From 2.00 to 2.30 m, where groundwater flooding forced the termination of the unit, a final midden level was encountered. This level featured two shifts in the ceramic content as well as a clear change in the surrounding soil matrix. Overlying midden deposits were mixed within a dark brownish black matrix of clay with a high silt content. The soil matrix within this lower portion of the midden, however, was gray and marked by a much more fine-grained clay texture. This transition presumably denotes the shift from a high- to a low-energy depositional environment within Area A (see below). Ceramics recovered within this level consisted primarily of Early Classic types and the

281

upper 0.20 m of this level (Midden 17e) was assigned an ‘Ayim-tun phase date. Several fragments from unusual flanged polychromes and a partial mammiform support, however, were recovered within the final lot excavated prior to termination of the unit, and this portion of the level (Midden 17f) was assigned a tentative Terminal Preclassic P’ich ‘Ayim phase date.

Feature/Context

Table 7.10. Context summary for Unit 12A6. Depth (m) Phase Description/Notes

Humus/Collapse

0.77–1.25

late Säk-tunich

Humus mixed with collapse from structures on Platform EE.

Midden 17a

1.25–1.60

early Säk-tunich

High-density midden.

Midden 17b

1.60–1.70

Yaljob’ach – Säk-tunich

High-density midden mixed with charcoal deposit.

Midden 17c

1.70–1.90

Sik’u’ II

High-density midden mixed with charcoal deposit.

Midden 17d

1.90–2.00

Sik’u’ I

High-density midden.

Midden 17e

2.00–2.20

‘Ayim-tun

High-density midden mixed fine clay.

Midden 17f

2.20–2.30

P’ich ‘Ayim‘Ayim-tun

Low-density midden mixed fine clay.

In addition to ceramics, the Unit 12A6 midden produced a diverse artifact assemblage (Table 7.9), including 22 obsidian blades, a groundstone metate, several notched sherds, two sherd disks, fragments from five figurines, and more than 300 chert artifacts. This midden also produced a rich faunal assemblage, including approximately 165 identified specimens of shell and bone (Table 7.8). The inventory of identified species was notable for the diversity of the mammal taxa represented, with two species of deer, armadillo, peccary, agouti, and several unidentified species of bird present. The

282

Postclassic portion of the midden also included a partial carapace from an unidentified species of sea turtle. Most fish remains recovered in this deposit were in float samples from the burned portion of the Late Classic midden level. Several of the fish bones were blackened and at least this portion of the midden may have been burned prior to deposition to alleviate a noisome problem of rotting fish remains. Unit 2A1 The other unit excavated in Area A (Unit 2A1) was placed approximately five m south of Platform EE and 12 m west of Unit 12A6 (Figure 7.1). Unit 2A1 was a 1-x-1-m unit excavated to a depth of approximately 0.95 m, where it breached the water table and had to be terminated due to steady flooding (Figure 7.9). Although Unit 2A1 did not identify any cultural features, it provided much useful information regarding the stratigraphy of Area A. The uppermost level of Unit 2A1 consisted of a mixture of brown clayey humus and numerous small stones. The stone content of this level was interpreted as colluvial debris or collapse deposited in Area A subsequent to its abandonment as a harbor. Ceramics recovered in this level were mixed, heavily eroded, and Postclassic in date. At a depth of approximately 0.57 m, the stone content of the unit dropped off completely, and a new level of dark brown to dark brownish black, slightly silty clay was encountered. Ceramics were recovered in relatively high frequencies throughout this level and demonstrated some evidence for stratification. The uppermost lot consisted of a mix of Terminal Classic and Postclassic diagnostics. Although ceramic data were somewhat limited, the mix of Paxcamán, Augustine, Fulano, and Trapeche group

283

Figure 7.9. South profile of Unit 2A1.

materials closely resembles the distribution seen in the upper portion of the Unit 12A6 midden, and this deposit likely also dates to the first part of the Säk-tunich phase. Lower lots were dominated by Late Classic types, including several Sik’u’ II diagnostic polychromes, with smaller numbers of Late Preclassic and Early Classic sherds mixed in. Cumulatively, this level was interpreted as sheet midden (Midden 36a), likely deposited contemporaneous and in the same manner as materials in the upper portion of the Midden 17.

284

Beneath this level, beginning at a depth of approximately 0.85 m, a new level of fine dark gray clay was encountered. As with the preceding level, no stones were present. This level correlates with the lowest levels of Midden 17 described above (Middens 17d to 17f), and also presumably marks a shift in the depositional environment of Area A. Ceramics mixed within this level were interpreted as a low-density continuation of the overlying sheet midden (Midden 36b). Further, although the ceramic content of this level was greatly reduced, numerous diagnostic sherds were recovered. These consisted of a relatively even mix of Late Classic and Early Classic types, suggesting a Sik’u’ I or transitional ‘Ayim-tun–Sik’u’ I phase date.

Feature/Context

Table 7.11. Context summary for Unit 2A1. Depth (m) Phase Description/Notes

Humus/Colluvium

0.10–0.57

Säk-tunich

Mixture of humus, colluvial debris, and possible collapse.

Midden 36a/ Clay

0.57–0.85

Sik’u’ II – Säktunich

Moderate density midden mixed with clay.

Midden 36b/ Fine Clay

0.85–0.95

‘Ayim-tun – Sik’u’ I

Low density midden mixed with fine clay.

Area A Summary Although Area A was less extensively investigated than Platform EE, investigations there nonetheless provided key data for understanding Trinidad’s ancient harbor. First, at the most basic level, the identification of stratified midden mixed with the lacustrine clays undergirding Area A suggests that this area was at least seasonally inundated for a major portion of Trinidad’s occupation. On the basis of ceramic data

285

from Middens 17 and 36, the interval when these clays developed extends at least from the Terminal Preclassic P’ich ‘Ayim through to the early Säk-tunich Postclassic. During this interval, Area A presumably provided an easy locus for the disposal of refuse building up along the edge of Platform EE and other harbor features. Ceramic data further indicate that the build-up of colluvial debris, architectural collapse, and silt atop these clays dates to the late Säk-tunich Postclassic and historical Aj B’oj phases. This interval was almost certainly marked by a slight drop in lake level and a decrease in activity within the harbor area. Although the population of the harbor area likely peaked in the Postclassic, this interval marked a major downturn in overall activity at the site. Under these conditions, Area A was allowed to silt over. The internal stratigraphy of the lacustrine clays in which the middens were recovered also provides clear evidence for a major change in the depositional environment of Area A. The deepest clay excavated within the two units discussed above (2A1 and 12A6) was gray and finely textured. The overlying clay, however, was dark brown and less fine-grained, with a high silt content. As smaller particles tend to settle in stiller water (Goldberg and Macphail 2006: Figure 5.2), such a shift, from fine-grained clays to more silty clays, appears indicative of a change from a low energy depositional environment to a higher energy one. Associated ceramics indicate that the lower clays developed during the P’ich ‘Ayim through ‘Ayim-tun phases, and possibly as late as the Sik’u’ I Late Classic. The upper clays, in contrast, date to the period starting with the Sik’u’ II phase and continuing to the early Säk-tunich.

286

Although changes in lake level may partially account for this shift in depositional environment, human activity almost certainly contributed as well. This depositional shift occurred at the same time as the single largest episode of construction on Platform EE (see above). Further, the evidence for activity-related debris on the Platform EE terraces (Midden 35), as well as artifact densities within Midden 17, all suggest that the interval from the Sik’u’ II Late Classic to the early Säk-tunich Postclassic was the peak period of use for the harbor area. The Sik'u' II construction of Structure EE-sub-1, as well as the four terraces leading down from that structure may have subtly shifted the shoreline, creating slightly more turbulent waters within Area A. Further, the increasing numbers of canoes passing through the harbor and the large amount of activity occurring on the Platform EE terraces may have increased the sediment load and allowed for coarser particles to settle out of suspension.

INVESTIGATON OF STRUCTURE EE-1 The next feature investigated as part of the harbor area testing program was Structure EE-1, the low, 12-m long platform that extends south from Platform EE into Area A (Figure 7.1). This feature is the only construction within Area A and presently stands only about 0.80 m above modern lake level. Its position within Area, as well as reports by local informants of its use during periods of slightly higher-than-present lake level, suggest it was a dock. To examine Structure EE-1’s possible function, a 9.5-x-1-m trench was excavated east-west across the structure’s principal axis. This excavation was divided into five units (2E1-2E5). Excavations reached a maximum depth of 1.04 m and

287

identified only one episode of construction. These excavations were terminated due to time constraints, however, and additional episodes of construction may await further research. Units 2E1-2E5 The uppermost level within Units 2E1-2E5 (Figure 7.10) consisted of dark brown clayey humus mixed with large quantities of gravel- to cobble-sized stones. This level was interpreted as humus mixed with collapse from Structure EE-1 or, more likely, a later structure at the same locus. Ceramics and other artifacts were recovered in great quantities throughout this level. Ceramics were slightly mixed and included numerous Preclassic, Early Classic, Late Classic, and Terminal Classic types. Postclassic types dominated, however, and a Säk-tunich phase date was assigned. Further, the prevalence of the type Paxcamán Red throughout this level suggested a late Säk-tunich date. At a depth of approximately 0.33 m, a partially preserved plaster floor was encountered in the central-most units of the trench. This floor constituted the upper floor of Structure EE-1 and extended approximately 3.60 m east-west. Subsequent excavations revealed that this floor was situated atop a low platform delineated to the east and west by 40–60-cm-high retaining walls constructed of unmodified cobbles. No walls were identified above the surface of the floor and Structure EE-1, but these may have collapsed to either side of the structure. Brief exploratory excavations into Floor 1 fill collected a sufficient quantity of ceramics to provide a general Säk-tunich phase date for Structure EE-1’s construction, although an early Säk-tunich date appears most likely.

Figure 7.10. North profile of Units 2E1 to 2E5.

288

289

Large quantities of loose gravel and cobble-sized stones were also noted to either side of Structure EE-1, with particularly dense debris located to the platform’s eastern side. None of the stones appeared to have been worked. These materials were interpreted as collapse, although their structural origin is unknown and they may constitute the visible remains of a late Säk-tunich cobble construction situated directly atop Structure EE-1. Finally, it should also be noted that collapse above and adjacent to Structure EE-1 produced great quantities of broken Postclassic censers. Although relatively few of these materials could be identified to the type level, censers included effigy, impressed, chambered, and bowl forms. Many censer fragments from these contexts also featured patches of eroded stucco or other calcareous substances. The presence of these forms suggested a variety of alternate functions for Structure EE-1 during the Säk-tunich phase (see below).

Feature/Context

Table 7.12. Context summary for Units 2E1–2E5. Depth (m) Phase Description/Notes

Humus/Collapse

Various

(late) Säk-tunich

Humus mixed with dense collapse.

Floor 1/Str. EE-1

0.33-0.77

(early) Säk-tunich

Low platform with partially preserved plaster floor.

Structure EE-1 Summary Investigation of Structure EE-1 revealed a construction most likely dating to the early Säk-tunich Postclassic. This construction appears to have been simple in form, extending approximately 12.0 m south from the southern edge of Platform EE, with an

290

east-west width of 3.6 m. No exterior walls were identified in these excavations, although they may have been of poor quality and unidentifiable within surrounding collapse. If this structure bore no masonry facings, then in its final early Säk-tunich form it would have provided a platform area of approximately 43.0 m, raised 50-75 cm above Area A. Although only a single clear episode of construction was identified, at least one and possibly other episodes are hinted at by the considerable collapse overlying Structure EE-1. The large cobbles, in particular, that surround the platform may constitute the remains of the ultimate episode of construction, likely dating to the late Säk-tunich Postclassic. Likewise, the densities of Chukan, ‘Ayim-tun, Sik’u’ I–II, and Yaljob’ach phase materials within surrounding collapse is suggestive of earlier constructions at this locus, perhaps mirroring those on nearby Platform EE. These constructions were presumably buried by the early Säk-tunich phase construction. The limited nature of these investigations makes it difficult to assess Structure EE-1’s possible function. In its rectangular shape and overall dimensions, Structure EE-1 does appear similar in form to several Postclassic residential structures within the harbor area (see Structures FF-1, HH-1, and II-1). Structure EE-1’s positioning at the lowest point of the entire harbor area, however, contrasts with typical Postclassic settlement strategies. Elsewhere at Trinidad, Postclassic structures are located atop platforms and low hillocks. Structure EE-1’s low-lying location suggests an alternative function. Such a location, adjacent to the aquatic portion of the harbor, as well as the possible absence of

291

exterior facings, suggests that the structure may have been utilized for other activities, with use as a dock one possibility. It should also be noted that the large numbers of censer fragments within associated fill are suggestive of another function as well. Although censers undoubtedly served several purposes, their presence in large numbers is often indicative of a ritual context (D.Chase 1988; P.Rice 1999). Use of Structure EE-1 as a shrine would not be out of line for a dock or harbor, as shoreline shrines are present elsewhere in Mesoamerica and appear to have been a regular feature at many pre-Columbian ports (Andrews 1990).

INVESTIGATION OF PLATFORM GG One of the most prominent features in the harbor area is Platform GG, the 70-mlong construction that forms the southern boundary of the harbor (Figure 7.1). This feature helps separate the Trinidad lagoon from Lake Petén Itzá and provides a secure landing area utilized by local informants. Further, during periods of slightly higher lake level, when Area A was inundated, this feature would have served as a breakwater protecting activities occurring within the harbor’s confines. Among the most important questions about this feature were whether it was an artificial construction and, if so, its date of construction. Excavations on Platform GG were divided into two sets of units. The first two 1x-1-m units (Units 2A2 and 2A5) were excavated on the higher eastern end of Platform

292

GG. The second two units, measuring 1.5-x-1 m (Unit 2A3) and 2-x-2 m (Unit 12B1), were excavated on the lower western end of the platform. Platform GG East: Unit 2A2 The first excavation on Platform GG was a 1-x-1-m unit placed on the platform’s eastern end. This end of the platform rises some 2 m higher than the western end, and from the surface, appears to be a steep, natural feature. Unit 2A2 was placed at the approximate midpoint of the eastern portion of the platform. This unit was excavated to a depth of 0.60 m and identified a single episode of construction. The uppermost level of Unit 2A2 consisted of a dark brown humus level mixed with large quantities of gravel and cobble-sized stones. The stone content from this level may have come from collapse or looters’ activities to the west (see below). Ceramics within this level included only Postclassic types and a general Säk-tunich phase date was assigned. At a depth of approximately 0.30 m small fragments of eroded plaster were recovered throughout the unit and atop a shallow layer of ballast. This floor was designated Floor 1. Ceramics within Floor 1 fill were dominated by Postclassic types, particularly Augustine Red, and an early Säk-tunich phase date was assigned. Beneath Floor 1 fill gave way to a pale brown subsoil mixed with numerous gravel-sized stones and very few artifacts. This level presumably constitutes the unmodified portion of Platform GG’s eastern end. Ceramics were small and eroded, and were presumably introduced by natural processes.

293

Feature/Context

Table 7.13. Context summary for Unit 2A2. Depth (m) Phase Description/Notes

Humus

0.10-0.03

Säk-tunich

Humus mixed with possible collapse.

Floor 1

0.30-0.40

(early) Säk-tunich

Partially preserved plaster floor and associated fill.

Subsoil

0.40-0.60

Indeterminate

Unmodified subsoil.

Platform GG East: Unit 2A5 The second unit excavated on the eastern end of Platform GG (Unit 2A5) was placed approximately 5 m northwest of Unit 2A2, and just east of a looters’ pit cut into the platform’s upper surface. No evidence was available regarding the reason for looting an otherwise feature-less locus; however, the presence of a possible shrine or other feature of interest should not be overlooked. Unit 2A5 was placed in this location to date the floor first identified in Unit 2A2 and also visible in the looters’ pit, and to document any additional episodes of construction. Unit 2A5 was excavated to a depth of 0.50 m and identified a single episode of construction (Figure 7.11). The uppermost level in Unit 2A5 consisted dark brown humus mixed with small quantities of gravel-sized stones. Associated ceramics were mixed and included materials from the Early Classic and Late Classic. The latest, however, dated to the Postclassic and a general Säk-tunich phase date was assigned. At a depth of approximately 0.29 m, a thin, but well-preserved plaster floor was encountered throughout the unit. This floor was designated Floor 1 and corresponded closely with Floor 1 in Unit 2A2. Floor fill was light, consisting of numerous small gravel-sized stones mixed with a dark brown matrix. Only a small number of sherds

294

Figure 7.11. South profile of Unit 2A5.

were recovered within fill. These were mixed and included mostly Early Classic pieces. An early Säk-tunich phase date was assigned to accord with Floor 1 in Unit 2A2. Beneath Floor 1, fill gave way to a pale brown subsoil mixed with numerous small stones and few large pieces of sascab. This level presumably constitutes the unmodified portion of Platform GG’s eastern end. Interestingly, however, sherds identified within this level included only Early Classic types. An ‘Ayim-tun phase date was assigned, although these materials were presumably introduced through natural processes.

295

Feature/Context

Table 7.14. Context summary for Unit 2A5. Depth (m) Phase Description/Notes

Humus

0.10-0.29

Säk-tunich

Humus.

Floor 1

0.29-0.42

(early) Säk-tunich

Partially preserved plaster floor and associated fill.

Subsoil

0.42-0.50

‘Ayim-tun*

Subsoil/possible fill.

Platform GG West: Unit 2A3 The first unit excavated on the lower, western end of Platform GG was a 1.5-x-1m unit (Unit 2A3) placed near the center of the platform (Figure 7.1). This unit was designed to determine if the lower, more regularly shaped western portion of the platform was an artificial feature and to date its construction. Unit 2A3 was excavated to a depth of 1.03 m and identified a cache, a burial, and a major episode of construction (Figure 7.12). The uppermost level of Unit 2A3 was a black clayey humus level mixed with copious quantities of artifacts, including several prismatic bladelets, a complete chert point, and a figurine fragment. Very few stones were encountered in this level, nor was there any evidence for a floor or other construction. Ceramics were highly mixed and included materials from all periods. But, Postclassic types, particularly Paxcamán Red, dominated and a late Säk-tunich phase date was assigned. At a depth of approximately 0.30 m, a new level of brownish gray clayey soil, mixed with numerous small to medium sized stones, was encountered. This level was interpreted as fill. Although no evidence for a plaster floor was encountered, the shift from humus to fill was clear at this depth. This fill episode likely constituted the ultimate

296

Figure 7.12. North profile of Unit 2A3.

phase of construction for the platform. Further, the dense concentration of artifacts directly above this level provided further evidence for the presence of a floor. This construction was designated Floor 1. Ceramics were recovered in great quantities throughout Floor 1 fill. Although the collection was slightly mixed, with numerous Late Preclassic and Late Classic sherds present, Postclassic materials dominated and a Säk-tunich phase date was assigned. Further, as the vast majority of red-slipped materials pertained to the Augustine group, with very few Paxcamán materials present, this construction can be dated to the first half of the Säk-tunich Postclassic.

297

Figure 7.13. Prismatic blades from Trinidad Cache 4. Drawing by I.Seyb.

At the end of this fill level, an obsidian cache was encountered in the western half of the unit (Figure 7.13). Cache 4 consisted of five complete and one nearly complete prismatic blades. The blades were oriented longitudinally northeast-southwest, with their lateral edges aligned running east to west and their dorsal surfaces facing down. All of the blades were from the Ixtepeque source and likely came from early in the core reduction process. One of the blades was pristine while the other five showed evidence for light to moderate bilateral use-wear.

298

Although no other artifacts were recovered during excavation of Cache 4, a float sample collected from the surrounding matrix produced more than 340 specimens of bone and shell (Table 7.15). Most specimens were too eroded to identify, but taxa included armadillo (D. novemcinctus), brocket deer (Mazama spp.), and an unknown species of ray (Rajiformes). This deposit also produced more than 180 fish bones, seventy fragments of freshwater mollusk shell, and a variety of freshwater gastropods. Although the freshwater component to the faunal assemblage may have been included within lake deposits re-utilized as fill, float samples collected elsewhere in the harbor area, with the exception of Burial 2 (below), did not produce similar rich zooarchaeological inventories.

Table 7.15. Faunal taxa identified in Cache 4.

Scientific Name

Common Name

NISP

Mazama spp.

Brocket deer

1

Dasypus novemcinctus

Armadillo

4

Mollusca

Freshwater mollusk

70

Pachychilus indiorum

Jute

1

Pomacea flagellata

Apple snail

17

Rajiformes

Ray

1

Atrocosteus tropicus

Tropical gar

1

Cichlidae

Cichlid fish

7

Osteichthyes

Fish

175

No ceramics were recovered with Cache 4, but its placement with respect to Floor 1 fill and the associated Burial 2 (see below) provides an early Säk-tunich phase date. Beneath Floor 1 fill and Cache 4, a new fill episode was encountered. The most notable aspect of this lower fill level was a remarkable increase in the size of fill stones.

299

Below approximately 0.60 m fill consisted of massive stones, most too large to be removed. Small and medium-sized stones, mixed with a yellowish brown, slightly clayey soil matrix, were tightly packed in the interstices between these large stones. The large size of stones and dense nature of this fill effectively precluded further excavation beneath a depth of approximately 1.03 m. This level presumably represents the principal stage of construction for Platform GG's western arm, or, minimally, a major modification to an existing natural feature. Unfortunately, ceramics recovered within fill were highly mixed and included materials from the Late Preclassic, Late Classic, and Postclassic. Although Augustine group materials were quite common, it was unclear whether or not these materials were introduced in association with Burial 2 (see below). As a consequence, the date of this fill level is unknown. During excavation of this lower fill level, a burial was encountered approximately 6 cm below Cache 4. Burial 2 was an informal cyst burial, placed in a small void between two of the massive fill stones described above and capped with two unworked boulders. The individual was placed in the lateral flexed position with the head to the north and the body facing to the southeast. Skeletal remains in Burial 2 were very poorly preserved. Most bones were stained, heavily eroded and highly fragmented, and all long bones were missing epiphyses. Other more fragile remains such as the vertebrae and pelvis were absent. With these exceptions, most major skeletal regions (i.e., cranium, upper limb, lower limb, hands and feet) were represented in the burial by at least a few elements. The burial also showed clear signs of natural disturbance, likely root action, as the cranium and mandible were found 15 cm apart, with several isolated teeth in the

300

Table 7.16. Faunal taxa identified in Burial 2.

Scientific Name

Common Name

NISP

Synbranchidae

Swamp eel

5

Synbranchus marmoratus

Mottled swamp eel

1

Mollusca

Freshwater mollusk

67

Pomacea flagellata

Apple snail

3

Atrocosteus tropicus

Tropical gar

1

Aridae

Catfish

1

Cichlidae

Cichlid fish

9

Osteichthyes

Fish

78

intervening area. The burial’s poor preservation precluded the precise determination of the individual’s age or sex, but the size of the skeleton suggests that the individual was an adult (Thornton and Moriarty 2007; see Appendix B).

Feature/Context

Table 7.17. Context summary for Unit 2A3. Depth (m) Phase Description/Notes

Humus

0.10–0.30

(late) Säk-tunich

Mixed humus deposit.

Floor 1/Light Fill

0.30–0.64

(early) Säk-tunich

Light fill episode for possible eroded plaster or packed-earth floor.

Cache 4

0.64

(early) Säk-tunich

Obsidian cache.

Burial 2

0.70–0.80

(early) Säk-tunich

Informal cyst burial.

Heavy Fill

0.64–1.03

Indeterminate (poss. early Säktunich)

Massive platform fill.

Other than Cache 4, no grave offerings were recovered in association with Burial 2. As with the cache, however, float samples produced massive quantities of shell and bone (Table 7.16), primarily from freshwater species, including more than ninety fish

301

bones, several freshwater gastropods, and hundreds of unidentified mammal specimens. Identified fish remains included tropical gar (A.tropicus), mottled swamp eel (S.marmoratus), and catfish (Ariidae). As with Cache 4, these materials may have been included within lake deposits utilized as fill. Several of the fish bones were blackened, however, likely as a result of burning, and at least some of the remains were presumably included as part of the grave offering. Small quantities of potsherds were also recovered in association with the burial. Although these were slightly mixed, Augustine group materials predominated and Burial 2 was dated to the first half of the Postclassic Säktunich phase. Platform GG West: Unit 12B1 The final excavation on Platform GG (Unit 12B1) was placed 2 m south of Unit 2A3. The purpose of this unit was to examine the heavy fill identified in Unit 2A3 and to date its construction. The unit was given dimensions of 2-x-2 m in hopes of providing more space to remove boulder-sized fill stones. Unit 12B1 was excavated to depth of 2.12 m and provided some additional data regarding the Platform GG fill (Figure 7.14). The uppermost level of Unit 12B1 was a black clayey humus level similar to that in Unit 2A3. Further, as with the earlier unit, artifacts were recovered in moderate to high density throughout humus, and artifacts included 13 obsidian bladelets, approximately 300 potsherds, and several notched sherds likely utilized as net sinkers. This level also produced trace quantities of fauna, although in densities much lower than those seen in Unit 2A3. Ceramics collected in this level were slightly mixed, but were dominated by Postclassic types and a general Säk-tunich phase date was assigned.

302

Figure 7.14. North profile of Unit 12B1.

As with Unit 2A3, a light fill consisting of small- and medium-sized stones within a brownish gray clayey soil matrix was encountered at a depth of 0.30 m. This level was designated Floor 1. Further, as with Unit 2A3, ceramics from fill consisted primarily of Augustine group types and the floor was dated to the early Säk-tunich phase. The first evidence for the heavy fill level discussed above was encountered at a depth of 0.60 m. As with Unit 2A3, this level consisted of massive stones mixed with a

303

Figure 7.15. MSJ Project personnel digging around boulder-sized fill stones in Unit 12B1.

yellowish-brown clay matrix. Removal of these stones was difficult and, in some instances, impossible (Figure 7.15). This level was excavated to an approximate depth of 2.12 m, where further excavations would have been impossible without the use of ropes and pulleys. It was impossible to determine the ultimate depth of this fill or to see if earlier fill levels lay beneath it.

304

Ceramics and other artifacts were recovered in this level to 1.40 m. Because of the density of the surrounding clay matrix and lack of evidence for organic processes, it is difficult to imagine that these materials were intrusive. Ceramics, however, were fragmentary and difficult to date. Identified materials included Preclassic, Early Classic, Late Classic, and Postclassic types. This fill was assigned a tentative early Säk-tunich date. It is unclear, however, if this dated the entirety of Platform GG or only a small section of its western end (see discussion below).

Feature/Context

Table 7.18. Context summary for Unit 12B1. Depth (m) Phase Description/Notes

Humus

0.10–0.30

Säk-tunich

Mixed humus deposit.

Floor 1/Light Fill

0.30-0.60

(early) Säk-tunich

Light fill episode for possible eroded plaster or packed-earth floor.

Heavy Fill

0.60-2.12

(early) Säk-tunich*

Massive platform fill.

Platform GG Summary Although investigation of Platform GG did not resolve all of the issues surrounding its construction or use, it produced some important data for understanding harbor area developments. Among the most important results was the determination that Platform GG was at least partially artificial. Excavations on the higher eastern end of the platform determined that this part of the platform was a small islet or section of uplifted bedrock that was later modified to provide a flat platform surface. This portion of the platform was later floored. Although no excavations were conducted on the sloping side

305

of this end of the platform, these modifications also presumably included the addition of stairs or another means of access. More importantly, these investigations also determined that at least a major part of the lower, more regularly shaped western end of the platform was artificial. Although the extent of the natural feature upon which subsequent construction was likely based remains unknown, its final form is substantial. Even if it were a heavily modified natural feature, the tremendous size (ca. 50-x-15 m) and volume (ca. 2,250 m3) of Platform GG's western end, as well as the massive boulders utilized for fill, all indicate that its construction was a significant labor investment for Trinidad’s residents. The great size of this feature, greatly exceeding all nearby constructions, also indicates its probable importance in the harbor area. Assessing the function of Platform GG is difficult. Its position along the outer edge of the hypothesized harbor (Area A) suggests it was a breakwater or mole protecting other constructions within the harbor from the large waves in Lake Petén Itzá during seasonal rains or tropical storms. Further, the presence of a plaster floor atop the higher eastern end of the platform suggests that this construction may have had multiple uses. Dating the construction of Platform GG’s western end is problematic. Although ceramic data from excavations (Units 2A3 and 12B1) indicate an early Säk-tunich date for construction, ceramics were sparse and very eroded. The scale of this feature also greatly exceeds that of all other early Postclassic features at Trinidad combined. While this imbalance does not exclude the possibility of an early Postclassic date for construction, it makes it appear unlikely. Further investigations would likely reveal

306

additional episodes of construction or clarify present data. On the basis of developments elsewhere at Trinidad, a Chukan or Sik’u’ II date appears most plausible, though still largely conjectural.

OTHER HARBOR AREA INVESTIGATIONS In addition to investigations in the harbor’s largest features, discussed above, exploratory excavations were conducted at two other loci within the harbor area. Platform HH: Unit 2B1 The first of these investigations focused on Platform HH, a small platform located atop a low rise near the east end of the harbor and eroding into Lake Petén Itzá along its southern edge. Unit 2B1, a 1-x-1-m unit, was placed near the platform’s western end and was designed to delineate the group’s construction sequence (Figure 7.1). This excavation identified a single episode of construction. The first level encountered in Unit 2B1 was a dark brown humus level mixed with small quantities of artifacts and stones, possibly collapse from Structure HH-1. Ceramics were heavily eroded but consisted exclusively of Postclassic types. A general Säk-tunich phase date was assigned. Beneath humus, at a depth of approximately 0.20 m, a light fill episode was encountered. The fill episode consisted of many gravel to cobble-sized stones mixed with loose, silty soil matrix. No floor was identified during excavations, but this fill presumably constituted the principal construction episode for Platform HH and the designation Floor 1 was assigned. Ceramics within fill included a handful of Late

307

Preclassic and Late Classic sherds, but consisted primarily of Postclassic types and a general Säk-tunich phase date was assigned. Beneath Floor 1 fill, the unit transitioned to a very light subsoil mixed with small quantities of artifacts, likely deposited as sheet midden. The handful of preserved sherds indicated a Säk-tunich phase date.

Feature/Context

Table 7.19. Context summary for Unit 2B1. Depth (m) Phase Description/Notes

Humus/Collapse

0.10–0.20

Säk-tunich

Humus mixed with possible collapse.

Floor 1

0.20–0.50

Säk-tunich

Eroded floor and light fill.

Subsoil/Sheet Midden

0.50–0.70

Säk-tunich

Subsoil mixed with possible sheet midden.

Platform JJ: Unit 2C1 The final excavation in the harbor area was located just off the northeastern edge of Platform JJ, a large basal platform located midway up the slope from the harbor to the site center. This platform currently supports a modern house occupied by Trinidad’s caretaker. The only unit (Unit 2C1) excavated in this vicinity was placed approximately 10 m northeast of the platform’s northeastern corner. Unit location was derived from discussions with the caretaker, who indicated that excavation for a nearby outhouse had revealed deep deposits rich in artifacts. This unit was excavated to a depth of 1.07 m. Although no features were identified, three distinct levels of cultural materials were documented. The chronological affiliations for the lowest two levels were particularly interesting in that they dated to phases poorly represented elsewhere in the harbor area.

308

The uppermost level in Unit 2C1 consisted of a mixture of dark brown silty humus and a light mixture of artifacts. This level was interpreted as low density sheet midden resulting from activities on Platform JJ. Artifacts within this uppermost level were exceptionally eroded and only a small number could be identified to the type level. Of these, the majority were Late Classic types and this level of midden was assigned a general Sik’u’ I–II phase date. At a depth of approximately 0.55 m, soil matrix became much more clayey in texture and artifacts became somewhat more common than the previous level. As with the previous level, this level was interpreted as low density midden. Ceramics within this level were relatively well preserved, and included primarily Late Preclassic types. The inventory included, however, a small number of unusual orange pieces and several forms likely transitional to the Classic period. The orange pieces were identified as possible examples of Iberia Orange and the level was dated to the P’ich ‘Ayim Terminal Preclassic phase.

Feature/Context

Table 7.20. Context summary for Unit 2C1. Depth (m) Phase Description/Notes

Humus/Sheet Midden

0.10–0.55

Sik’u’ I–II

Mixture of humus and sheet midden.

Sheet Midden

0.55–0.75

P’ich ‘Ayim

Sheet midden mixed with clayey subsoil.

Subsoil/Sheet Midden

0.75–1.07

Ix Cha’

Light subsoil mixed with scattered sheet midden.

309

Finally, at a depth of approximately 0.75 m the soil matrix became much drier and less cohesive than the overlying soils and artifact densities decreased significantly. This level was identified as a sub-soil just above bedrock, and associated artifacts were interpreted as light sheet midden, presumably from nearby activities. Interestingly, the ceramic inventory included only Ix Cha’ Mamom specimens, making this deposit the only late Middle Preclassic context in the harbor area. Summary for Platforms HH and JJ Although these investigations produced comparatively minor results, they nonetheless provided important information for understanding the harbor area’s longterm development. For example, investigation of Platform HH identified only a single Postclassic episode of construction. This makes Group HH one of the few constructions at Trinidad to produce no evidence for earlier construction. The Postclassic date for Group HH also suggests that other similar groups nearby (e.g., Group II and others) date to the same interval. Investigations near Platform JJ were important primarily in that they produced the only evidence for an Ix Cha’ Middle Preclassic occupation as well as the single richest deposit of Terminal Preclassic P’ich ‘Ayim materials recovered within the harbor area. Although a handful of possible Mamom sherds were recovered in other harbor excavations, their numbers were very limited. Likewise, the sample of Terminal Preclassic materials was limited to a few specimens collected near in the lowest level of Midden 17 (see above). The low frequency of these materials within general collections

310

elsewhere in the harbor area suggests that lake levels may have been significantly different during these periods.

SUMMARY AND CONCLUSIONS Two seasons1 of investigation in the Trinidad harbor produced substantial results. Among the most important of these was the determination that Trinidad’s harbor is an almost entirely artificial construction. Testing of Platforms EE and GG as well as Structure EE-1 determined that they are either partially or completely manmade features, and at least two of these features were found to have been built or modified during multiple phases (Table 7.21). Further, the presence of middens and other occupational debris, as well changes in the depositional environment of Area A, indicate that Trinidad’s harbor had a long history of occupation and use (Table 7.22). The following section provides an outline overview of harbor area developments. The earliest evidence for construction or significant activity in the harbor area dates to the Late Preclassic Chukan phase. Harbor excavations produced no Aj Wo’ preMamom materials and only a handful of Ix Cha’ Mamom sherds; the most significant deposit of late Middle Preclassic materials was recovered near Platform JJ, located Sadly, these investigations are likely to be the last research conducted in Trinidad’s harbor. Just months after the end of the 2005 season the owner of the beach just west of the features described here, and a zone likely utilized as part of the Trinidad harbor, began construction of a restaurant. Most of the area of potential interest was covered by more than 4 m of fill. Similar plans have been made for the harbor area described above, and a recent Google Earth image shows a modern structure atop the western end of Platform GG. Although local landowners and government officials have been apprised of the certain loss of cultural patrimony should Trinidad’s harbor be developed, it nonetheless appears likely that this area will soon be destroyed. 1

311

Table 7.21. Construction episodes identified in harbor area testing. Phase

2A4

Platform EE 12A1 2B2

Str. EE-1 2E1-2E5

Platform GG 2A2, 2A5 2A3, 12B1

Plt. HH 2B1

Aj B’oj ______ Fl. 1

______ Fl. 1 ______ Fl. 1

Säk-tunich ______ Fl. 1

______ Fl. 2

______ Fl. 2

______ Fl. 2

______ Fl. 3

______ Fl.3

Yaljob'ach

Sik’u’ I–II ______ Fl. 4

‘Ayim-tun

P'ich 'Ayim

Chukan

Ix Cha'

Aj Wo'

Fl. 3

______ Fl. 5

______ Fl. 4

______ Fl. 1

______ Fl. 1

______ Fl. 1

312

almost 100 m north of and 25 m vertically from Area A. The harbor area was either little utilized during the Middle Preclassic or lake levels differed significantly from those seen at present or during most of Trinidad’s occupation. The Chukan phase occupation of the harbor area also appears to have been minor. Evidence for construction dating to this phase consists of a single, low platform constructed directly atop bedrock beneath Platform EE. No middens could be assigned to this phase. Evidence for this phase in the harbor sequence, however, may be underrepresented as excavations within Area A were terminated prior to reaching levels where Chukan phase deposits may have been located. Likewise, although little is known of the initial Chukan construction, the identification of a well-constructed stone retaining wall associated with the Chukan phase construction of Platform EE hints at the possibility of more sophisticated constructions elsewhere within the harbor. Evidence for human activity in the harbor area increased slightly during the subsequent P’ich ‘Ayim, ‘Ayim-tun, and Sik’u’ I phases. Trace quantities of P’ich ‘Ayim materials were recovered near the upper surface of the Chukan phase construction of Platform EE and the lowest level of the largest midden in Area A. Evidence for an ‘Ayim-tun occupation came primarily from the Area A midden and indicated a slight build-up over the preceding phase. Evidence for activity in the harbor area increased slightly during the Sik’u’ I Late Classic. Although the only evidence for Sik’u’ I construction was a modification or repair to Platform EE, Sik’u’ I refuse was recovered both atop this construction and within the Area A midden.

313

Table 7.22. Summary results of harbor area testing.2 Phase

Platform EE

Area A

Structure EE-1

Platform GG

Platform HH

Platform JJ

Aj B'oj Säk-tunich Yaljob'ach Sik’u’ I–II ‘Ayim-tun P'ich 'Ayim Chukan Ix Cha' Aj Wo'

The onset of the Sik’u’ II Late Classic marked the beginning of the peak interval in harbor area construction and use. This interval appears to have been initiated with a major renovation of Platform EE, raising the platform an additional 35–50 cm and adding a series of terraces to its southern side. A large and well-built foundation-brace structure was also added to the top of the platform. Although the function of this structure cannot be surmised, its quality and location at the inner edge of the harbor suggests an integral role in harbor area activities. Likewise, the addition of the terraced steps suggests a substantial addition to the functional activity space available within the harbor area.

2

For an explanation of table shading, see Table 6.23.

314

The Sik’u’ II phase was also marked by a clear increase in evidence for activity within the harbor area. Midden scattered across the top of the Platform EE terraces date to the interval between the Sik’u’ II and early Säk-tunich phases, and Area A middens increased in density starting with the Sik’u’ II Late Classic. This interval of intense activity also appears to have had a dramatic effect on deposition within Area A, transforming a low-energy environment to a high-energy one. This period of intense activity continued through Terminal Classic Yaljob’ach and culminated during the first half of the Säk-tunich phase. Early Postclassic constructions were identified for Platform EE and Structure EE-1, and Groups FF and HH appear to have been occupied. Further, Platform GG, the single largest feature within the harbor area was heavily modified during this phase. Evidence for activities associated with these episodes of construction was also recovered throughout the harbor area, as both surface remains and within the Area A midden. During the subsequent late Säk-tunich and Aj B’oj phases, the harbor area became the principal settlement zone for the site of Trinidad as most other areas were abandoned. Evidence for activity dating to the late Postclassic and historical phases is abundant. Refuse dating to these phases is scattered across the tops of most harbor features, and construction episodes dating to this interval were identified for both Platform EE and, possibly, Structure EE-1. Likewise, Groups HH, FF, and other possible residential groups within the harbor area appear to have been occupied at this time. Despite this evidence for activity, use of the harbor area appears to have shifted significantly during this period. Silt, gravel and cobble-sized colluvial debris, and other

315

materials were allowed to collect within Area A, raising the bed of this area significantly and likely ending its use as an aquatic harbor. The build-up of materials in this area may signal a dramatic shift in the function of Trinidad or a change in lake level. Although this area was later re-utilized to transship loads of chicle, its form had changed significantly from its original design.

316

CHAPTER EIGHT THE SITE CENTER MIDDEN TESTING PROGRAM

INTRODUCTION The third program of research at Trinidad was designed to identify and collect a robust sample of midden from Late Classic Sik’u’ I–II and Terminal Classic Yaljob’ach contexts in Trinidad’s site center for intra- and intersite comparison. Investigations focused on 13 residential groups and two public architectural complexes. Residential groups were selected on the basis of their chronological affiliation and spatial organization so as to provide a sample reflecting the breadth of Trinidad’s Late and Terminal Classic sociopolitical hierarchy. Public architectural complexes (Plaza V and the Trinidad Ballcourt) were selected for more specific reasons related to their potential functions within Trinidad’s ritual and material economy. Overall, this program of research was the most intensive at Trinidad, involving most project personnel for a full field season and including more than 500 50-x-50-cm midden tests and 42 excavation units with dimensions of 1-x-1 m or larger (Table 8.1). Cumulatively, these excavations moved more than 117 m3 of matrix, located 28 middens, and produced the lion’s share of artifacts recovered at Trinidad. The following sections describe midden prospecting methods, review the results of prospecting in each group, and provide midden-by-midden excavation summaries. The conclusions to this chapter discuss the general results of

317

midden research at Trinidad, with comments on patterns in the midden assemblages and chronological data.

METHODS AND ORGANIZATION The site center testing program included systematic midden testing in thirteen residential groups and two public architectural complexes (Table 8.1, Figure 8.1). Residential groups were selected for testing on the basis of their surface configuration and temporal occupation, so as to provide a cross-section of sociopolitical status during the Late Classic Sik’u’ I–II and Terminal Classic Yaljob’ach phases. Individual groups were selected to represent each of the four principal residential group types present at Trinidad (Types I, II, IIIb, and V; see Chapter 5, Moriarty 2004b). The sample included all three Type V groups at Trinidad, two Type IIIb groups, six Type II groups, and two Type I arrangements. The three Type V groups, the largest and most elaborate residential groups at Trinidad, are thought to represent the upper end of the sociopolitical spectrum, while the smaller, less elaborate Type I and II groups are inferred to represent the lower end (see Chapter 5). The Type IIIb groups were included as a possible mid-level within the settlement hierarchy and a source of comparison for Type IIIa groups at Motul de San José. All of these groups were inferred to have been occupied during the Late Classic Sik’u’ I–II or Terminal Classic Yaljob’ach phases on the basis of nearby test pits, surface collections, inspection of looters’ trenches, surface configuration, or other criteria. Cumulatively, these thirteen groups constitute 36 percent of the residential groups visible within the mapped portion of Trinidad.

318

Figure 8.1. Map of Trinidad with groups tested for middens indicated.

319

The ballcourt (Group F) and Plaza V, in contrast, were selected to address specific questions regarding their role in the material and ritual economy of Late Classic Trinidad. As noted in the research design and Chapter 5, both complexes are unusual features for a site of Trinidad’s modest size: Plaza V for providing an inordinate amount of public space and the ballcourt for its location at a secondary political center. Testing of Plaza V was designed to explore the possibility that it served as a setting for a marketplace or other activities, with debris from regular cleaning gathering along its margins (Dahlin et al. 2007; Inomata 2006; Wells 2004). The ballcourt area was tested to locate evidence for ballgame related feasting as discussed by J. G. Fox (1996) and others (Fash and Fash 2007), as well as to explore potential connections between the ballgame and ancient Maya trade.

Group(s)

Table 8.1. Summary of midden prospecting investigations. MSJ Group Op.(s) Midden Tests Midden Results 3 Type No. Vol. (m ) No. Identified Midden ID

Plaza V

Public Plaza

9A

51

7.16

2

9,10

F

Ballcourt

10A-C

55

8.40

3

11,12,13

C U Y

Va Vb Vb

7C 11A-D 13A-D

42 58 37

5.18 8.50 5.45

3 3 3

3,4,6 14,15,16 18,19,20

G H

IIIb IIIb

5B-E 19B-D

53 23

8.20 3.50

0 2

N/A 29,30

K O Q S T Z

II II II II II II

18A-D 8A-D 16A-C 15A-D 14A-B 17A-D

27 24 15 33 15 29

3.66 3.17 2.30 4.55 2.25 3.65

2 2 1 2 1 2

27,28 7,8 24 22,23 21 25,26

DD KK

I I

20B-D 21B-D

25 26

3.75 3.95

0 2

N/A 31,32

320

Figure 8.2. Excavation of 50-x-50-cm midden tests along the western edge of Plaza V.

Midden prospecting investigations followed standardized procedures modified from those first utilized in research at Motul de San José (Yorgey et al. 1999). First, lines of 50-x-50-cm midden test pits were established around the periphery of each group (Figure 8.2). Pits were placed along the lower edge of the group, as close as possible to exterior structural walls or the group’s basal platform, so as to fall within the “intermediate” or “toft” area (Deal 1985; Hayden and Cannon 1983; Johnston and Gonlin 1998; Killion 1990, 1992). Proximity to architecture also provided better preservation in identified middens, with collapse or later deposits effectively sealing the middens and creating a more neutral pH than typical of local soils. Pits were spaced at intervals of 3–4 m, depending on the size and orientation of the group. Additional pits were placed near

321

structure corners, between closely spaced groups, downhill from groups, or within other areas subjectively judged to be high probability for midden deposition. Midden tests were excavated using hand picks and trowels. Excavation methods followed protocols slightly different from those utilized in larger units. To proceed efficiently, midden test units were excavated in arbitrary 20-cm vertical lots unless architectural features were encountered. Further, to permit the simultaneous excavation of up to six pits per archaeologist, recording procedures were simplified, and only data concerning soil conditions, artifact content, and cultural context were recorded. Midden tests were excavated to a depth of 60–80 cm below the surface unless bedrock or intact architectural features were encountered. Continuation of units to this depth was the principal distinction from earlier MSJ Project midden prospecting operations. The increased depth, as well as positioning units close to structures, was found to be particularly useful in identifying middens covered by collapse and only visible well below the surface. For example, several of the highest density middens were identified more than 60 cm below the surface. All matrices from these investigations were screened through ¼ -inch mesh and all artifacts and ecofacts were collected. Using these procedures, a team consisting of an archaeologist and 4–6 excavators was capable of testing a large platform group in 2–3 days. As noted in the research design (see Chapter 4), these investigations were designed to identify high-density middens or “secondary refuse aggregates” associated with each group (see also Rosenswig 2009; Schiffer 1972, 1987; Stark and Ohnersorgen 2001; Wilson 1994). At Trinidad, these deposits were identified on the basis of artifact density, diversity, and preservation, as well as by basic contextual indicators. In general,

322

these deposits were characterized by higher artifact densities (e.g., potsherds/m3), greater artifact diversity (more artifact classes represented), and better artifact preservation (e.g., larger sherds, more refits) than other nearby deposits. Contextual clues often included a very low stone content (in comparison to collapse or fill), slightly darker soils (also in comparison to collapse or fill), and very little evidence for chronological mixing within the ceramic assemblage. The presence of large numbers of faunal remains also often tipped off excavators to the presence of a midden. Several of the richest deposits provided evidence for discrete deposition, including the presence of a pit feature or an associated cache. Once middens were identified, larger excavation units were placed directly adjacent to or between the most promising midden test units. Most of the follow-up units measured 1-x-1 m, although excavations were expanded adjacent to the most productive units where time permitted. All follow-up excavations followed standardized MSJ Project procedures, including the use of arbitrary 10-cm lots within natural or cultural levels, the collection of all artifacts and ecofacts, the screening of all matrix through ¼inch mesh, standardized soil sampling, and all recording procedures (Appendix A). Where possible, most of these units were continued to bedrock to supplement chronological data recovered in the site center testing program (Chapter 6) or to identify earlier midden levels.

ORGANIZATION OF EXCAVATION SUMMARIES The results of midden investigations are presented below by architectural group. Investigations in public architectural complexes are discussed first, followed by

323

residential groups. Each section provides a short summary of midden prospecting investigations followed by more detailed syntheses of larger excavation units. Although most of the midden tests identified “sheet” or low-density “open-air” midden (Rosenswig 2009; Stark and Ohnersorgen 2001; Wilson 1994), time constraints prevented full analysis of all materials from 50-x-50-cm midden test units, as well as full description here. Context and artifacts summaries are provided as tables for each of the larger excavation units. The final paragraphs in each section review the principal results of midden testing within each group. To simplify discussion here and elsewhere, each midden is treated as a discrete cultural feature and assigned a unique numeric designation (Midden 1, Midden 2, etc.). Further, to organize discussion of stratified deposits, each chronological level of a multicomponent midden is also assigned an alphabetic designation in order of recovery (Midden 1a, 1b, etc.). Dates were assigned on the basis of ceramic analysis and limited radiocarbon assays. Although midden ceramic assemblages were generally much easier to work with than those recovered in tertiary contexts, some midden levels could be dated only to general phases. This was particularly true for some Late Classic deposits; those lacking well-preserved Tepeu 1 or Tepeu 2 sphere polychromes or other indicators were dated to the general Sik’u’ phase, rather than the more specific Sik’u’ I or Sik’u’ II. Summary volumetric data and artifact counts are appended in tables for most deposits. For simplicity, midden densities are described in terms of sherd density (count/m3 or kg/m3), while preservation is discussed on the basis of average sherd weight (kg/count). All rims were refitted during analysis and the total number of unique rims is used as an indicator for the maximum number of vessels (MaxNV) in the deposit (Orton

324

et al. 1993:171-173; Rice 1987b:290-293). Chert, obsidian, groundstone, ceramic artifacts, and special finds are presented following classifications utilized by the MSJ Project. Faunal data are presented as the number of individual specimens (NISP); with some exceptions (e.g., Aves), only fauna identified to the order level or better are listed in attached tables. Additional artifact data are presented in Chapters 10–11. Finally, it should be noted that describing the richness or density of specific middens often presents semantic problems. For simplicity, the descriptions below follow a classification derived exclusively from the density of ceramic potsherds per unit excavation (count/m3). Deposits producing more than 1,500 potsherds per m3 were described as high-density. Those producing fewer than 1,000 potsherds per m3 were described as low-density. Deposits producing from 1,000 to 1,500 potsherds per m3 were described as moderate-density. Although this system is obviously less than perfect and fails to account for artifact diversity, preservation, or the density of other artifacts (e.g., middens with few potsherds, but with dense concentrations of chert debitage), it does correspond to observable patterns in the ceramic density data.

PUBLIC ARCHITECTURE MIDDEN TESTING Midden testing within Trinidad’s site center included two public architectural complexes: Plaza V and the ballcourt (Group F). As noted above, investigations in these complexes were designed to assess specific hypotheses about the kinds of activities taking place in their confines.

325

Figure 8.3. Location of midden Units 9B1 and 9B2. See Figure 8.1 for scale.

Plaza V: Operation 9 Midden Prospecting: Operation 9A Midden prospecting around Plaza V focused on the gentle slope directly west of the plaza and downhill from Structures P-2 and P-3. Prospecting included three lines of 17 midden test pits running north-to-south (Operation 9A: Units 9A1–9A51). The first

326

line of pits started 3 m northwest of Structure P-2 and continued south to a point approximately 8 m south-southwest of Structure P-3 (Units 9A1–9A17). Pits were placed at 4 m intervals and subsequent rows were placed 4 m (Units 9A18–9A34) and 8 m (9A35–9A51) to the west, with the second row offset 2 m to the south. Moderate quantities of artifacts, particularly obsidian, chert, and Late Preclassic Chukan and Late Classic Sik’u’ I–II potsherds, were recovered throughout these investigations, most in contexts suggestive of sheet midden deposition. No obvious high-density deposits were identified. Two areas, however, produced slightly greater artifact concentrations and were identified as moderate-density midden (Figure 8.3). The first of these (Midden 9) was situated 8 m west of Structure P-3’s midpoint, while the second (Midden 10) was found 12 m southwest of Structure P-2. Midden 9: Unit 9B1 Investigation of the first deposit associated with Plaza V (Midden 9) consisted of a single 1-x-1-m unit (Unit 9B1) placed between the midden tests 9A24 and 9A25. Unit 9B1 was excavated to a depth of 0.70 and collected a 0.40 m3 sample of moderate-density midden. The uppermost level of Unit 9B1 consisted of black, slightly clayey humus mixed with small quantities of artifacts and a few small to medium-sized stones. The stones were interpreted as collapse from Structure P-3 or fill from another eroded construction episode along Plaza V’s western edge. Ceramics were very eroded and the inventory included both Late Preclassic Chukan and Late Classic Sik’u’ diagnostics. This level was assigned a general Sik’u’ date as the materials were too poorly preserved to identify either Sik’u’ I or Sik’u’ II diagnostics.

327

Figure 8.4. North profile of Unit 9B1.

Midden 9 was encountered at a depth of approximately 0.30 m. At that depth, the soil matrix became dark grayish brown in color and much sandier in texture. Artifact frequencies also increased substantially. Numerous small stones also appeared in this level, though in insufficient density to be interpreted as fill. Although artifacts were recovered in only low to moderate densities, the midden assemblage included 480 potsherds, 315 chert pieces, and 6 obsidian artifacts. Ceramics consisted exclusively of Late Preclassic diagnostics and Midden 9 was assigned a Chukan phase date.

Feature/Context

Table 8.2. Context summary for Unit 9B1. Depth (m) Phase Description/Notes

Humus/Collapse/ Midden

0.10–0.30

Sik’u’

Humus mixed with midden and possible collapse or fill from Plaza V.

Midden 9

0.30–0.70

Chukan

Low-density midden.

328

Midden 10: Unit 9B2 Investigation of the second deposit associated with Plaza V (Midden 10) also consisted of a single 1-x-1-m unit (Unit 9B2). Unit 9B2 was placed amidst a series of moderate-density midden tests, approximately 12 m southwest of Structure P-2. This unit was excavated to a depth of 0.60 m and yielded a 0.40 m3 sample of moderate- to highdensity midden. The uppermost level in Unit 9B2 consisted of humus identical in color and texture to that seen in Unit 9B1 (see above). Artifacts were recovered only in trace quantities within this level. The only preserved sherds dated to the Late Preclassic period, but a general Late Classic Sik’u’ date was assigned on the basis of context (see above).

Feature/Context

Table 8.3. Context summary for Unit 9B2. Depth (m) Phase Description/Notes

Humus

0.10–0.20

Sik’u’

Humus.

Midden 10a

0.20–0.50

Sik’u’

High-density midden.

Midden 10b

0.50–0.60

Chukan

Low- to moderate-density midden.

Beginning at a depth of approximately 0.20 m the soil matrix became dark grayish brown in color and slightly sandy in texture. Artifact densities also increased notably and this level was interpreted as midden. The inventory of artifacts from this deposit (Midden 10) included 650 potsherds, 464 chert artifacts, 6 pieces of obsidian, and various ceramic objects. Ceramics within Midden 10 also provided clear evidence for at least two separate periods of deposition. The uppermost level of the deposit (Midden 10a) was dominated by general Late Classic Sik’u’ diagnostics, while the lowest portion (Midden

329

10b) consisted exclusively of Late Preclassic Chukan materials. Middens 10a and 10b presumably resulted from activities occurring along the western edge of Plaza V. Midden Summary for Plaza V Investigations along the western edge of Plaza V failed to identify large or notably rich high-density middens. Both of the middens identified during these investigations (Middens 9 and 10) appeared to constitute the highest density areas within a long zone of low- to moderate-density sheet midden spread along the gentle slope below the western edge of the plaza. Artifact data from the midden follow-up excavations closely follow patterns observed during midden prospecting. Ceramics were heavily eroded, but dated almost exclusively to the Late Preclassic Chukan and Late Classic Sik’u’ I–II phases. Both chert and obsidian were recovered in moderate to large quantities throughout these investigations. Chert densities were similar to those encountered in nearby residential Groups G and O (see Chapters 6 and 9; Lawton 2007a), and presumably resulted from the processing of materials from nearby chert beds. The obsidian collection, including 35 artifacts from midden prospecting, was notable in that it included numerous large flakes and chunks, likely left over from core preparation and processing (see Chapter 11). Although these investigations failed to identify the kinds of high-density middens expected for a large marketplace, the presence of such extensive sheet midden along one boundary of Trinidad’s largest plaza is certainly interesting. Further, the content of sheet midden associated with Plaza V appears distinct from that of the ballcourt middens (see below), and much closer in to that seen in nearby residential groups (see below). This is

330

precisely the distribution expected for a marketplace (e.g., Hirth 1998), though more research would be required to further test this possibility. Further, the ceramic content of these deposits indicates that activities in Plaza V occurred primarily during the site’s two principal peaks in occupation and operation as a port (see Chapter 12).

Table 8.4. Excavation volume and artifact summaries for Middens 9–10. Artifact Class Count by Midden 9 10a 10b Phase Chukan Sik’u’ Chukan Excavation Vol. (m3) 0.40 0.30 0.10 Ceramics Count (f) Weight (kg) Weight/Count Count/m3 Weight/m3 MaxNV

480 2.158 0.0045 1,200.00 5.40 21

484 2.644 0.0055 1,613.33 8.81 31

175 0.800 0.0046 1,750.00 8.00 12

302 265 36 1

326 320 4 2

138 132 5 1

Obsidian (Total) Prismatic Bladelets Chunk Flakes

6 4 -2

3 2 -1

3 1 1 1

Ceramic Artifacts (Total) Sherd Disks Notched Sherds

0 ---

2 2 --

1 -1

Chert (Total) Flakes Cores Tools

Group F (Ballcourt): Operation 10 Midden Prospecting: Operations 10A–10C Midden prospecting in Group F focused on the ballcourt complex’s eastern perimeter and included a total of 55 pits. Pits were divided into three suboperations. The first suboperation of 38 pits (Operation 10A) was initiated with a line of 14 pits (Units

331

Figure 8.5. Location of midden units associated with the Trinidad ballcourt.

10A1–10A14) along the eastern edge of Structure F-2, the ballcourt’s eastern lateral structure. This initial line of pits, placed at 3 m intervals, ran from the midpoint of Structure F-2’s northern side, around its northeastern corner, down its eastern edge, and around to the midpoint of its southern side. A second line of 17 pits (Units 10A1510A31) followed the first, offset 3 m out from the initial line and crossing over Structures F-3 and F-4, the two small appendage structures attached to Structure F-2’s eastern side (Figure 8.5). A third line of 7 pits (10A32–10A38) was established 3 m east of the

332

second line, between the eastern ends of Structures F-3 and F-4. The second suboperation (Operation 10B) consisted of seven pits (Units 10B1–10B7) excavated around the eastern and southern edges of Structure F-4. The final seven pits (Operation 10C; Units 10C1–10C7) were excavated around the eastern periphery of Structure F-5, the small rectangular structure directly south of Structure F-2, starting near the midpoint of its northern side and ending at the midpoint of its southern side. With the exception of those pits placed near Structure F-5 (Units 10C1–10C7) or within heavy collapse directly against the base of Structure F-2 (Units 10A1–10A14), midden prospecting tests in the ballcourt area encountered rich refuse deposits. Midden assemblages were particularly noteworthy in containing numerous large and wellpreserved Sik’u’ I and Sik’u’ II potsherds, and for the remarkable diversity of collections, with almost all artifact classes represented throughout these investigations. Within the generally high-density deposits encountered by these investigations, three specific loci stood out. The first deposit (Locus A, Midden 11) was identified along the eastern edge of Structure F-3. The second deposit (Locus B, Midden 12) was identified near the center of the small area (part of Plaza IV) framed by Structures F-2, F-3, and F-4. The third deposit (Locus C, Midden 13) was identified along the northeastern corner and eastern edge of Structure F-4. These three loci were selected for further investigation. Midden 11 (Locus A): Units 10D1 and 10D11 The northernmost midden locus, Midden 11, was investigated by two 1-x-1-m units (Units 10D1 and 10D11) placed directly east and southwest of shovel test 10A19 and immediately adjacent to Structure F-3. Units 10D1 and 10D11 (Figure 8.6) were

333

Figure 8.6. Ballcourt Locus A: north profile of Units 10D1, 10D11, and 10A19.

excavated to a depth of 1.32 m, yielded a 0.90 m3 sample of high-density midden, and identified an early construction level in Plaza IV. The uppermost level of Units 10D1 and 10D11 consisted of very dark grayish brown silty humus mixed with large to medium-sized stone collapse from Structure F-3. Artifact densities increased notably within the lowest lots of this level, indicating some mixing with the underlying midden. Ceramics consisted exclusively of Late Classic diagnostics and a Sik’u’ II date was assigned on the basis of this level’s stratigraphic position with respect to Midden 11 (see below).

334

At a depth of approximately 0.50 m, collapse gave way to a level of fine, dark grayish brown silty soil, almost no stones, and large quantities of artifacts. This level, which continued to a depth of ca. 0.90 m, was designated Midden 11. The artifact inventory from this level included 1,463 potsherds, 356 chert artifacts, 9 obsidian pieces, 11 partial figurines, 5 sherd disks, a spindle whorl, a mano, a metate fragment, and a handful of faunal specimens. Although these totals place Midden 11 within the moderate- to high-density range, they likely constitute only a very small sample of the overall deposit. The positive midden test at this locus (Unit 10A19), running just north of Units 10D1 and 10D11, produced the highest ceramic density of all middens at Trinidad, with potsherds collected at a rate of approximately 6,600 per m3. Such high artifact densities were not seen in either of the units described here. As the areas to the south of Midden Test10A19 and Units 10D1 and 10D11 were well covered by midden tests and other excavations, process of elimination suggests that the highest density portion of this deposit likely lies just to the north of these units. Unfortunately, time constraints prevented further investigation of this locus. Beneath Midden 11 and beginning at a depth of approximately 0.90 m, the remains of a plaster floor and its associated fill were encountered. Although the floor was completely eroded, small bits of plaster were present throughout the unit and the transition from midden to floor fill was relatively clear in the unit profile. Ceramics were very mixed, particularly in the uppermost lots of fill. Beneath the initial lots, however, only Late Preclassic and Middle Preclassic diagnostics were identified and a Chukan phase date was assigned. This floor was presumably the same Plaza IV floor (Floor 3) encountered at a depth of 0.63 m in Unit 1F3 some 20 m to the northeast (see Chapter 6).

335

The upper ‘Ayim-tun and Sik’u’ floors also identified in that unit (Floors 1 and 2) and in part of Locus C (see below) were not present in Units 10D1 and 10D11, and were presumably destroyed during the deposition of Midden 11.

Table 8.5. Context summary for Units 10D1 and 10D11. Feature/Context Depth (m) Phase Description/Notes Humus/Collapse

0.10–0.50

Sik’u’ II

Humus mixed with collapse from Str. F-3.

Midden 11

0.50–0.90

Sik’u’ II

Moderate- to high-density midden.

Floor 3

0.90–1.32

Chukan

Eroded plaster floor and associated fill.

Midden 12 (Locus B): Unit 10D2 Investigation of the second ballcourt midden locus (Midden 12) consisted of a single 1-x-1-m unit (Unit 10D2). This unit was placed ca. 8 m east of Structure F-2, near the midpoint of the north-south area bounded by Structures F-2, F-3, and F-4. The unit was also placed directly adjacent to midden test 10A23. Unit 10D2 was excavated to a depth of 1.03 m and yielded a 0.70 m3 sample of low- to moderate-density midden (Figure 8.7). The uppermost level in Unit 10D2 consisted of very dark brown slightly clayey silt mixed with moderate quantities of artifacts. This level was interpreted as humus mixed with materials from the underlying midden. Ceramics within this level included numerous large and well-preserved Late Classic diagnostics, but few intact polychromes, and a general Late Classic Sik’u’ date was assigned.

336

Figure 8.7. Ballcourt Locus B: north profile of Unit 10D2.

Midden 12 was encountered beginning at a depth of 0.23 m when the soil became dark brown in color and artifact densities increased substantially. Although artifacts were collected in only light to moderate quantities, the midden inventory included 587 potsherds, 134 chert pieces, 2 figurine fragments, a spindle whorl, and a small faunal sample. Ceramics included many well-preserved Late Classic diagnostics, but insufficient polychromes to date the deposit to either the Sik’u’ I or Sik’u’ II phases. The difficulty in dating this particular deposit likely resulted in part from depositional processes. In contrast to Middens 11 and 13, artifacts in Midden 12 were smaller, more eroded, and presumably deposited more irregularly (“tracked” into the area), rather than in a handful of discrete episodes.

337

At a depth of 0.93 m, a level of light to moderately rocky fill was encountered. Although no traces of plaster were encountered, this level closely correlated with the stratigraphic position of Plaza IV Floor 3 identified in Units 10D1 and 10D11 (above) and test pit 1F3 (see Chapter 6). As with the above units, associated ceramics were dominated by Late Preclassic Chukan phase diagnostics. Further, as with Units 10D1 and 10D11 above, no traces of Plaza IV Floors 1 or 2, identified in Unit 1F3 and at Locus C (see below), were encountered in Unit 10D2. These floors were presumably destroyed during midden deposition or did not originally extend into the area between Structures F3 and F-4.

Feature/Context

Table 8.6. Context summary for Unit 10D2. Depth (m) Phase Description/Notes

Humus/Midden

0.10–0.23

Sik’u’

Humus mixed with materials from underlying midden.

Midden 12

0.23–0.93

Sik’u’

Low- to moderate-density midden.

Floor 3

0.93–1.03

Chukan

Eroded plaster floor and fill.

Midden 13 (Locus C): Units 10D3–10D10 Of all the middens excavated at Trinidad, Locus C (Midden 13) of the ballcourt middens was the most thoroughly investigated and produced the most spectacular results. Follow-up excavations at this locus included 8 units (Units 10D3–10D10) with dimensions of 0.70-x-0.60 m or greater, with the largest (Unit 10D10) covering an area of 3.0-x-1.5 m (Figures 8.8 and 8.9). All of these units were placed directly east and northeast of Structure F-4, within a ca. 25 m2 zone where high-density deposits were identified in multiple midden tests (Units 10A37, 10A38, 10B3, and 10B4).

Figure 8.8. Ballcourt Locus C: north profile of Units 10D3, 10D5, and 10D7–10D9.

338

339

Cumulatively, follow-up excavations covered an area of nearly 9 m2 up to a maximum depth of 2.28 m, yielded a sample of approximately 8.50 m3 of high-density stratified midden, and revealed several important cultural features, including the eastern face of Structure F-4. The uppermost level of Units 10D3–10D10 consisted of a dense mixture of humus, collapse, and midden. Soil within this level was dark grayish brown in color, slightly clayey in texture, and mixed with moderate quantities of charcoal, presumably from hearth cleaning or possibly from modern or historical milpa burning at the site. Collapse was encountered primarily within the units closest to Structure F-4 (Units 10D8–10D10), and included carefully cut rectangular facing stones from the structure’s eastern wall. Artifacts were recovered in large quantities throughout this level, and the final artifact inventory from collapse included more than 1,500 potsherds, 440 chert artifacts, 20 obsidian pieces, a partial groundstone metate, 8 figurine fragments, a small faunal sample, and various ceramic artifacts and ornaments. The density of artifacts within this level indicated either extensive mixing with the underlying deposits (see below) or the presence of a separate level of midden directly at the surface and mixed with humus. Ceramics from this level tend to support the latter interpretation. Although most of the polychromes were identical to those found in the Sik’u’ II portion of the underlying midden (Midden 13a), a handful of unusual plate modes, as well as forms possibly derivative of Chablekal group imports (Chapter 10), suggest a late Sik’u’ II date, perhaps on the cusp of the Terminal Classic. The absence of any fine paste wares, however, precluded the assignment of a pure Yaljob’ach phase date.

Figure 8.9. Ballcourt Locus C: north profile of Unit 10D10.

340

341

Figure 8.10. Tinaja group jar necks from Cache 3.

Directly beneath this level, and at the interface with the underlying midden, a cache was identified. Cache 3 consisted of two Tinaja group jar rims, complete from the neck up, placed directly adjacent to each other in an upright position (Figure 8.10). Interestingly, the two most important rim modes for the Sik’u’ I–II and Yaljob’ach complexes were represented. The first rim was characterized by the outcurved neck typical of the general Late Classic Sik’u’ complex, while the second had the everted rim more characteristic of the very late Sik’u’ II or the Terminal Classic Yaljob’ach. Many other artifacts were recovered in close proximity, but could not be separated from surrounding midden. Cache 3 may have been placed to roughly mark the location of Burial 7, located some 1.5 m to the southwest and 0.50 m below the level of the cache (see below), or constituted part of an offering sealing the underlying midden.

342

Midden 13 was first encountered at a depth of between 0.40 and 0.65 m throughout the Locus C units. At this depth the stone content of the units dropped off completely, the soil matrix shifted to a fine, dark brown silt with rare white, sand-sized inclusions, and artifact densities increased substantially. This level continued down to between 1.30 and 2.28 m within an obvious hole cut through earlier features as part of midden deposition. In the westernmost units closest to Structure F-4, the hole terminated at the level of Plaza IV Floor 1 (ca. 1.30 m), a Sik’u’ phase construction first identified in Unit 1F3 (see Chapter 6). Although this floor was completely eroded, a clear transition from midden to fill was visible in unit profiles and in association with Burial 7 (see below). Just 1.00 m east of Structure F-4, however, no trace of Floor 1 could be identified and the midden continued down to the level of Plaza IV Floor 3 (ca. 1.60 m), a Chukan phase construction also identified in Unit 1F3. Both Plaza IV Floors 1 and 2 were presumably destroyed to create a hole for midden deposition. In addition, just 1.20 m further to the east, Floor 3 had also been destroyed and removed, with the eastern portion of the midden continuing down to a depth of at least 2.28 m, well below the upper surface of Plaza IV Floor 3. The midden’s horizontal dimensions could not be estimated. Neither the northern- or southern-most units identified a clear edge to the deposit. Further, many of the best preserved and most spectacular finds came from the last lots excavated in the eastern-most units of Locus C, suggesting that much more extensive deposits remain to be explored. Artifacts were recovered in massive quantities throughout Midden 13. The ceramic collection alone included 16,396 potsherds weighing more than 220 kilograms. Many of the sherds were large and well-preserved, and collections included four

343 Table 8.7. Sherd counts by ceramic group for Middens 11–13. Count (%Total) by Midden Ceramic Group 11 12 13a 13b Cambio Saxche-Palmar Tinaja Infierno Azote Tialipa Unnamed White Unnamed Stucco Other Eroded Total

433 (24.8) 154 (8.8) 251 (14.4) 1 (